Lipid composition and structural characteristics of bovine, caprine and human milk fat globules

International Dairy Journal 56 (2016) 64e73

Contents lists available at ScienceDirect

International Dairy Journal journal homepage: www.elsevier.com/locate/idairyj

Lipid composition and structural characteristics of bovine, caprine and human milk fat globules Yunping Yao a, Guozhong Zhao a, Jingying Xiang b, Xiaoqiang Zou a, Qingzhe Jin a, Xingguo Wang a, * a

State Key Laboratory of Food Science and Technology, Synergetic Innovation Centre of Food Safety and Nutrition, School of Food Science and Technology, Jiangnan University, 1800 Lihu Road, Wuxi 214122, Jiangsu Province, PR China Wuxi Maternity and Child Health Care Hospital, Wuxi 212422, PR China

b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 5 November 2015 Received in revised form 28 December 2015 Accepted 28 December 2015 Available online 8 January 2016

Milk fat is widely accepted to be the major nutrient in human milk. Commercial infant formulas are usually based on mammalian milk such as bovine or caprine milk, but the differences in milk fat globules (MFGs) between human, bovine and caprine milk remain unclear. We showed that saturated fatty acid content was higher in bovine and caprine MFGs (>60%) than in human MFG (<40%), but that content of the unsaturated fatty acids C18:2 in human MFG was >7 times higher than in bovine and caprine MFGs. The cholesterol content of human milk was ~20% higher than that of bovine and caprine milk. Triacylglycerol molecular species and polar lipids also differed between bovine, caprine and human MFGs. Confocal laser scanning microscopy images of MFGs revealed that the shape of the liquid-ordered domains varied by species. © 2016 Elsevier Ltd. All rights reserved.

1. Introduction Human milk is an important nutritional resource for newborn infants and provides 40e60% of energy from lipids (Abrahamse et al., 2012). Although human milk is recommended as the optimal food for infants, not every mother is able to breastfeed her baby, for a variety of reasons. In such cases, commercial infant formulas are the best substitutes to fulfil the nutritional needs of infants. Bovine and caprine milk and ingredients derived therefrom are the main ingredients in the infant formulas currently on the market (Prosser, McLaren, Frost, Agnew, & Lowry, 2008). However, the composition and structure of milk fat globules (MFGs) differ between bovine, caprine and human milk. It is thus necessary to study the MFGs of these types of milk, to explore a more suitable breast milk substitute than the currently available products. MFGs are composed of a core of triacylglycerols (TAGs) and an MFG membrane (MFGM) consisting of proteins, glycoproteins, enzymes, phospholipids and sterols. The nutritional and functional characteristics of MFGs depend not only on their chemical composition, but also on their lateral membrane organisations

(Garcia-Saez, Chiantia, & Schwille, 2007). Recently, confocal laser scanning microscopy (CLSM), using lipophilic fluorescent probes, has been used to investigate the features of MFGs (Gallier, Gragson, Jimenez-Flores, & Everett, 2010b). The microstructure of MFGs is associated with the phospholipid and fatty acid composition. Thus, the different proportions of phospholipids and the distributions of fatty acids in each phospholipid among the species could result in significant differences in the microstructure and nutritional aspects of MFGs. Therefore, the structure and, physical and chemical properties of bovine, caprine and human milk should be thoroughly studied. The purpose of this study was to compare the lipid composition of bovine, caprine and human MFGs on the basis of fatty acids of TAG and polar lipids, fatty acids of individual polar lipids, TAG, polar lipids, and sterols. The size distribution, zeta potential and microstructure of the MFGs were also evaluated. Such comparative studies may contribute to the improvement of infant formula. 2. Materials and methods 2.1. Samples and reagents

* Corresponding author. Tel.: þ86 510 85876799. E-mail address: [email protected] (X. Wang). http://dx.doi.org/10.1016/j.idairyj.2015.12.013 0958-6946/© 2016 Elsevier Ltd. All rights reserved.

Bovine milk samples from 10 Chinese Holstein cows (mature milk, 30e60 d in lactation, third quarter of 2015) were kindly

Y. Yao et al. / International Dairy Journal 56 (2016) 64e73

provided by a local producer (Tian zi Dairy Industry Co., Ltd., Wuxi, Jiangsu, China). Ten pure-bred Boer caprine milk samples (mature milk, 30e60 d in lactation, third quarter of 2015) were purchased from Zhengxing animal husbandary Co., Ltd. (Zhejiang, China). Ten human milk samples (mature milk, 16e30 d in lactation, third quarter of 2015) were kindly donated by healthy Chinese woman, 28e32 y old, in the Wuxi Maternal and Child Health Hospital. The mothers had been well informed before participating in this study, which was approved by the medical ethics committee of Wuxi Maternal and Child Health Hospital (ethics approval number 2013022). The milk samples were characterised within 24 h after collection for particle size, zeta-potential and CLSM. Other samples were then stored at 20  C until further chemical analysis. The 37component fatty acid methyl ester (FAME), cholesterol, phospholipids standards, boron triflouride-methanol solution and the high performance thin-layer chromatography (HPTLC) plates were purchased from SigmaeAldrich (St. Louis, MO, USA). All other reagents used were all of high-performance liquid chromatograph purity (Sinopharm Chemical Reagent Co., Ltd., Beijing, China). 2.2. Analysis of lipid composition 2.2.1. Milk fat extraction Milk fat was extracted following the method described by Folch, Lees, and Sloane-Stanley (1957). Briefly, 10 mL of milk was mixed with 200 mL of chloroform: methanol (2:1, v/v). The mixture was shaken for 15 min and then centrifuged at 3500  g for 5 min. The clear lower chloroform layer was mixed with 50 mL 0.86% NaCl solution. The organic phase was evaporated by rotary vacuum evaporator. The extracted lipid was stored at 20  C until further analysis. 2.2.2. Analysis of fatty acids of total lipids and polar lipids For fatty acid analysis of total lipids (Christopherson & Glass, 1969), milk fat (20 mg) was added to a sealable tube and hexane (2 mL) and methanolic KOH (2 M, 500 mL) were added. The mixture was vortexed for 5 min, after which then deionised water (5 mL) was added for extraction. After shaking, the upper layer was collected and dried over anhydrous sodium sulphate, and the resulting FAME solution (1 mL) was analysed by gas chromatography (GC). Non-polar and polar fractions were separated by HPTLC with a solvent system of hexane/diethyl ether/acetic acid (80:20:1, by vol) (Benoit et al., 2010). The baseline band containing the polar lipids fraction was scraped off the plates and placed into screw-capped tubes. Then, polar lipids were extracted three times with the mixture of chloroform/methanol/water (5:5:1, by vol). Fatty acid methyl esters of phospholipids were used to characterise the longchain fatty acids. BF3 (14% in methanol, 500 mL) solution was added for methylation, and the screw-capped tubes were kept at 100  C for 90 min. Hexane (600 mL) and deionised water (1.5 mL) were subsequently added, after which the mixture was centrifuged at 2100  g for 10 min at 20  C, and the solvent phase was collected and dried over anhydrous sodium sulphate. After centrifugation (2100  g for 10 min), the upper layer was injected into GC for fatty acid analysis. The analysis of fatty acid composition was performed with an Agilent 7820A GC (Agilent Technologies Inc., Palo Alto, CA, USA) equipped with an autosampler, a flame ionisation detector, and a Trace TR-FAME capillary column (60 m  0.25 mm  0.2 mm, ThermoFisher Scientific, Waltham, MA, USA). The temperatures of injector and detector were 230 and 250  C respectively. Nitrogen carrier gas at 1 mL min1 was used and the split ratio was 1:100. The oven temperature was held at 60  C for 3 min, then raised to 175  C at 5  C min1 and held for 15 min at this temperature,

65

followed by an increase to 220  C at 2  C min1 and held for 10 min (Yao, Zhao, Zou, Huang, & Wang, 2015). 2.2.3. Fatty acid analysis of the different polar lipids species A HPTLC method was carried out to separate the phospholipids classes. The mobile phase was composed of methyl acetate/isopropanol/chloroform/methanol/0.25% (w/v) KCl (25:25:25:10:9, by vol). Iodine vapour was applied overnight for the revelation of spots. The different phospholipid species were scraped off and fatty acid content was determined as described above for total polar lipids. The fatty acids of sphingomyelin were converted to the corresponding methyl ester by treatment with methanolic HCl (0.5 M, 400 mL) at 80  C for 20 h (Sanchez-Juanes, Alonso, Zancada, & Hueso, 2009). The procedure was then continued as described above for total fatty acids analysis. 2.2.4. Triacylglycerol analysis The separation and identification of TAG was performed as previously described by Zou et al. (2012a). TAG species were separated by reverse-phase high-performance liquid chromatography (RP-HPLC) and the identification of TAG was carried out on a HPLCeatmospheric pressure chemical ionisation mass spectrometry (HPLC-APCIeMS). 2.2.5. Polar lipid composition Polar lipid composition was determined according to the method described by Rombaut, Camp, and Dewettinck (2005). Separation of the polar lipids was carried out on a HPLC equipped with an evaporative light scattering detector (ELSD). A silica column (4.6 mm  250 mm, 5 mm particle size; Phenomenex, Inc., Torrance, CA, USA) was used in this study. One hundred milligrams milk fat was dissolved with 1 mL chloroform/methanol (88:12, v/v) and transferred into capped test tubes for HPLC analysis. Nitrogen was used as the nebulising gas at a flow rate of 1 L min1, and the evaporating temperature was 85  C. The set of the program was followed as Zou et al. (2012a). 2.2.6. Sterols content The sterols in all samples were determined according to the method of Fraga et al. (2000). Milk samples (100 mg) were saponified in capped tubes with 2 M KOH by heating at 85  C for 1 h, and the unsaponifiable fraction was extracted with hexane. The hexane phase was then dried and silylated by 400 mL N,O-Bis (trimethylsilyl) trifluoroacetamide þ1% trimethylchlorosilane (BSTFA þ TMCS) at 70  C for 30 min; the residue was dissolved in 1 mL of hexane, and 1 mL of product was analysed by gas chromatography-mass spectrometry (GCeMS) equipped with a DB-5 MS capillary column (30 m; 0.25 mm i.d., 0.52 mm film thickness; Agilent Corp.). The carrier gas was helium, with a flow rate of 1 mL min1 and the split ratio was 1:50. The oven temperature was held at 150  C for 1 min and then raised to 300  C at a rate of 10  C min1 and then held for 15 min at 300  C. Scan time and mass range were 1 s and 50e500 (m/z), respectively (Li et al., 2011). 2.3. Physical and structural analysis 2.3.1. Particle size The particle size distribution of milk samples was determined by laser light scattering using a Mastersizer 2000 (Malvern Instruments, Malvern, UK), equipped with an He/Ne laser (l ¼ 633 nm) and an electroluminescent diode (l ¼ 466 nm). The refractive index of milk fat was taken to be 1.460 at 466 nm and 1.458 at 633 nm, and the aqueous phase was 1.33. The apparatus and method for particle size analysis have been described in detail by Michalski, Briard, and Michel (2001).

66

Y. Yao et al. / International Dairy Journal 56 (2016) 64e73

2.3.3. Microstructural analysis The milk sample for CLSM observation was prepared as previously reported (Lopez, Madec, & Jimenez-Flores, 2010), with some modifications. The microstructure of MFGs was analysed with a confocal microscope (Zeiss LSM 710, Carl Zeiss Meditec AG, Jena, Germany). The observations were performed using a 63  1.4 oil immersion objective. The neutral lipids were stained by adding 100 mL of Nile Red (42 mg mL1 in acetone) to 0.5 mL of milk. The phospholipids of the MFGM were labelled with N-(lissamine rhodamine B sulfonyl)1,2-dioleoyl -phosphatidylethanolamine (Rh-DOPE) (1 mg mL1 in chloroform). The Rh-DOPE solution (about 20 mL) was added to the milk (0.5 mL). After labelling, the samples were kept at room temperature for 30 min, then 100 mL stained milk was slowly mixed with 100 mL agarose (5 g L1 and stored at 45  C). Then, one drop of the milk stained with the fluorescent dyes were deposited onto the glass and observed on the microscope.

are precursors of long-chain PUFA. Long-chain PUFAs play an important role in early human growth and development (Koletzko & Rodriguez-Palmero, 1999). To highlight the preferential localisation of certain fatty acids in polar lipids of the different species, our study compared the fatty acid profile in polar lipids with that in total lipids (Table 1). The percentages of unsaturated fatty acids of polar lipids in bovine (42.9%) and caprine (38.7%) milk were higher than those in total lipids. The unsaturated fatty acids of polar lipids in human milk (37.1%) were lower than those in total lipids (P < 0.05, Fig. 1). This indicates that more unsaturated fatty acids exist in the MFGM of bovine and caprine milk than in the MFGM of human milk. The TAG core of MFGs in human milk had more unsaturated fatty acids than that in bovine and caprine milk. The SFA content in human MFGM was 62.9%, a result slightly higher than the value of 57% reported by Sala-Vila, Castellote, Rodriguez-Palmero,
pez-Sabater (2005). The different percentages of Campoy, and Lo SFA observed in our study could be a result of diet, region and climate. The majority of SFA in the polar lipids of milk fat comprised long-chain fatty acids (C16:0 and C18:0), whereas the most abundant unsaturated fatty acid was C18:1 (Table 1). The C18:0 content was significantly higher in the polar lipids of human milk than in total milk fats. This may be due to human milk having the highest content of sphingomyelin (SM), which is mainly composed of SFA.

2.4. Statistical analysis

3.2. Fatty acids composition of individual polar lipid species

The experiments were carried out three times and the results were expressed as the mean ± standard deviation. The data were subjected to analysis of variance (ANOVA) using the Statistical Analysis Software (Version 9.0, SAS Institute, Inc., Cary, NC, USA). Results were considered statistically significant at P < 0.05.

The values found for the fatty acids of each polar lipid species in the bovine, caprine and human milk were significantly different (Fig. 2). SM was the polar lipid with the most SFA in bovine milk. Phosphatidylcholine (PC), phosphatidylserine (PS) and phosphatidylinositol (PI) had intermediate values of SFA. In contrast, phosphatidylethanolamine (PE) had the lowest content of SFA. The distribution of the fatty acids of PE, SM and PI in caprine milk was similar to the fatty acid profile of polar lipids in bovine milk, but differences existed in the profiles of PC and PS. The unsaturated fatty acids represented in PE, PS and PC in human milk (>40%) were higher than those in bovine and caprine milk. Contrastingly, more SFA were found in PI and SM in human milk (Fig. 2). PE had the highest content of MUFA and PUFA and the lowest content of SFA among the polar lipids of bovine, caprine and human milk. PE was also reported to be the most unsaturated of the phospholipids by Gallier, Gragson, Cabral, Jimenez-Flores, and Everett (2010a). C18:1 and C18:2 were the major fatty acids of the PE fraction. The PE of human milk had a higher content of C18:2 than that of bovine or caprine milk. C16:0 and C18:2 were the main fatty acids in PC among the human milk polar lipids, but C16:0 and C18:1 were the major fatty acids in the bovine and caprine milk polar lipids. The fatty acids contents of PI and PS had very similar profiles in the three types of milk. SM contained high amounts of SFA. C16:0 together with long-chain SFA (22:0, 23:0 and 24:0) were the major fatty acids in bovine and caprine milk. However, the most abundant fatty acids of SM in human milk were C18:0, C22:0 and C24:0. Until recently, all infant formulas lacked MFGM material because this fraction was lost during regular dairy processing. One study found that, infant cognitive scores were significantly higher in the group that was fed with formula containing MFGM than in the standard-formula group, but there were no differences in cognitive scores between the MFGM group and the €f, Hernell, Lo € nnerdal, & breastfed group (Timby, Domello

2.3.2. Zeta-potential The apparent zeta-potential of MFG was measured by using a Zetasizer 2000 (Malvern Instruments). Samples were prepared by suspending 10 mL milk in 10 mL buffer (20 mM imidazole, 50 mM NaCl, 5 mM CaCl2, pH 7.0), and zeta-potential was measured at 25  C as previously described by Michalski, Michel, Sainmont, and Briard (2002).

3. Results and discussion 3.1. Fatty acid composition of total lipids and polar lipids The milk of different mammals shows specific differences in the fatty acid composition. The relative contents of the fatty acids in bovine, caprine and human milk fat are shown in Table 1, with the fatty acids grouped by type as saturated fatty acids (SFA), monounsaturated fatty acids (MUFA) and polyunsaturated fatty acids (PUFA). The results (Fig. 1) show that bovine and caprine milk contained a higher percentage of SFA (>60%) compared with human milk (<40%). A lower proportion of SFA in milk fat seems to be favourable for human health because of their native role in arteriosclerosis (Pfeuffer & Schrezenmeir, 2000). For SFA, C16:0, C18:0 and C14:0 were present in the highest proportions in bovine and caprine milk, whereas the main SFA in human milk were C12:0, C14:0 and C16:0. C18:1 was the predominant MUFA in the milk fat, accounting for 22.2%, 31.7% and 29.7% of total lipids in bovine, caprine and human milk, respectively. Human milk fat contained a significantly higher amount of PUFA (28.2%) compared with bovine (4.4%) and caprine (3.9%) milk fat (P < 0.05). The content of PUFA in human milk fat in our study was somewhat higher than that reported in North America and Europe
pez-Sabater, & (Bahrami & Rahimi, 2005; De la Presa-Owens, Lo Rivero-Urgell, 1996; Scopesi et al., 2001). This is probably related to the higher consumption of vegetable oil (mainly soybean oil) among local women, which can result in the high content of C18:2 in human milk fat (Wan, Wang, Xu, Geng, & Zhang, 2010). The predominant PUFAs of human milk were C18:2 and C18:3, which

Y. Yao et al. / International Dairy Journal 56 (2016) 64e73

67

Table 1 Fatty acid composition of the total lipid and phospholipid fraction of bovine, caprine and human milk.a Fatty acids

Total lipids

Phospholipids

Bovine milk C4:0 C6:0 C8:0 C10:0 C11:0 C12:0 C13:0 C14:0 C14:1 C15:0 C15:1 C16:0 C16:1 C17:0 C18:0 C18:1T C18:1 C18:2 C18:3n-6 C18:3n-3 C20:0 C21:0 C20:2 C20:3n-6 C20:4 C20:3n-3 C22:0 C22:1 C20:5 C23:0 C22:2 C22:4 C24:0

0.95 1.07 0.96 2.78 0.07 3.71 0.14 12.33 1.39 1.38 0.02 34.04 1.86 0.75 11.19 0.41 22.15 3.31 0.07 0.29 0.21 0.04 0b 0.15 0b 0.02 0.08 0.02 0.14 0.05 0.01 0c 0.06

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

Caprine milk a

0.03 0.03a 0.02b 0.06b 0.01a 0.13b 0.01a 0.11a 0.07a 0.04a 0.00a 0.21a 0.05a 0.04b 0.07b 0.03b 0.12c 0.06b 0.01c 0.02b 0.01b 0.01b

± 0.04b ± ± ± ± ± ±

0.00b 0.01c 0.00b 0.02a 0.01b 0.00b

± 0.01b

0.88 1.00 1.14 3.58 0.05 1.61 0b 5.48 0.10 1.28 0b 30.31 1.06 0.91 15.4 0.52 31.73 3.27 0.24 0.31 0.47 0.13 0b 0.05 0b 0.02 0.17 0.04 0b 0.08 0b 0.04 0.08

± ± ± ± ± ±

b

0.03 0.02a 0.03a 0.08a 0.00b 0.08c

± 0.14b ± 0.00b ± 0.06b ± ± ± ± ± ± ± ± ± ± ±

0.63b 0.05b 0.07a 0.08a 0.04a 0.15a 0.04b 0.00a 0.02b 0.04a 0.01a

± 0.01c ± 0.00b ± 0.02b ± 0.01a ± 0.01a ± 0.01b ± 0.02b

Human milk 0.02 0.05 0.16 1.20 0c 5.59 0b 5.55 0.11 0.18 0b 21.41 1.97 0.30 4.27 0.16 29.65 22.85 0.16 1.60 0.21 0c 1.06 0.92 1.07 0.18 0.45 0.20 0.13 0.03 0.26 0.51 0.11

± ± ± ±

Bovine milk c

0.00 0.01b 0.03c 0.07c

± 0.05a ± 0.07b ± 0.01b ± 0.02c ± ± ± ± ± ± ± ± ± ±

0.71c 0.08a 0.03c 0.12c 0.03c 0.61b 0.18a 0.01b 0.06a 0.02b

± ± ± ± ± ± ± ± ± ± ±

0.05a 0.04a 0.04a 0.03a 0.02a 0.02b 0.02a 0.00c 0.03a 0.03a 0.01a

0.14 0.15 0.19 0.58 0.11 1.17 0.15 6.75 0.46 1.10 0.19 27.22 1.77 0.83 16.11 0.78 26.99 9.63 0.10 0.63 0.60 0.15 0b 0.63 0 0.06 0.65 0.26 0.63 0.96 0.09 0c 0.48

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

Caprine milk c

0.02 0.03a 0.02a 0.04a 0.01a 0.04a 0.01a 0.08a 0.02b 0.07a 0.01a 0.13c 0.05a 0.03a 0.22b 0.03a 0.10a 0.06b 0.01a 0.02a 0.03c 0.01b

± 0.04a ± ± ± ± ± ±

0.01b 0.02c 0.01c 0.04a 0.07b 0.01a

± 0.03c

0.29 0.10 0.11 0.40 0b 0.50 0b 1.43 0.16 0.38 0b 29.45 0.66 0.62 23.67 0.80 26.26 7.58 0.06 0.54 0.84 0.28 0b 0.21 0 0.10 1.03 0.46 0b 1.16 0b 0.36 1.00

± ± ± ±

b

0.02 0.01b 0.02b 0.03b

± 0.02b ± 0.04b ± 0.01c ± 0.01b ± ± ± ± ± ± ± ± ± ± ±

0.34b 0.03b 0.02b 0.15a 0.02a 0.07a 0.05c 0.01b 0.02b 0.04b 0.01a

± 0.02c ± 0.01a ± 0.03a ± 0.01b ± 0.05a ± 0.01b ± 0.02b

Human milk 0.39 0.15 0.09 0.31 0b 0.21 0b 1.40 0.56 0.24 0b 31.52 0.55 0.24 23.66 0b 13.79 12.04 0.13 0c 0.98 0c 0.52 0.46 0 0c 0.80 1.13 0b 0.52 0b 2.53 2.38

± ± ± ±

0.05a 0.01a 0.01b 0.02c

± 0.02c ± 0.03b ± 0.02a ± 0.01b ± ± ± ±

0.39a 0.04c 0.03c 0.16a

± 0.07b ± 0.06a ± 0.02a ± 0.06a ± 0.03a ± 0.04b

± 0.04b ± 0.03a ± 0.04c ± 0.13a ± 0.06a

a Values are means ± standard deviation and are expressed as mass %; calculations were based on 10 samples with three replicate measurement per milk sample. Different superscript letters in a row indicate significant differences for a lipid class (P < 0.05).

€ f, 2014). The specific structure of most phospholipids in Domello human milk may be of functional significance for the infant. Therefore the functional significance of polar lipid species in human milk should be investigated further to improve infant health and development.

Fig. 1. Comparison of the fatty acids distribution (in mass %: , polyunsaturated; ,, monounsaturated; -, saturated) of total lipids (TL) and phospholipids (PL) in bovine milk (BM), caprine milk (CM) and human milk (HM). Total values are relative to fatty acids reported in Table 1.

3.3. Triacylglycerol composition TAGs make up ~98% of the lipids in mammalian milk and are composed of a glycerol backbone to which three fatty acids are esterified. Although all species secrete milk fat similarly, the TAG core is uniquely arranged to optimise digestion, absorption, nutritional and lipid metabolism for the new born. Some studies showed that milk TAG structure affects different aspects of infant health, such as the absorption of fatty acids and calcium, intestinal flora and bone health (Bar-Yoseph, Lifshitz, & Cohen, 2013). The TAG composition of the bovine, caprine and human milk fat is shown in Table 2. The TAG composition of bovine milk fat contained a wide range of short and medium-chain TAG species compared with that of the caprine and human MFGs (Fig. 3). PMBu, PMCp/POBu and, PPBu/ SMBu were the main TAG species in bovine milk, which was in accordance with the previous studies (Smiddy, Huppertz, & van Ruth, 2012; Zou et al., 2012b). In caprine milk fat, the main TAG species were MMLa/PMCa, POS and PPS, along with minor longchain TAG such as MLaO/POCa, SMLa, MLaO/POCa, OOCa and PoPO/OOL. Smiddy et al. (2012) also observed a maximum at CN40; however, Fontecha, Diaz, Fraga, and Juarez (1998) reported this maximum at CN42 in caprine milk. The major TAGs in human milk fat were POO and POL, and other TAG with palmitic acid (16:0) such as POS, PPL, PPO and PSS were also present. Similar findings were
, and Lo
pez Sabater reported by Morera Pons, Castellote Bargallo (1998). In human milk, palmitic acid (16:0) is predominantly esterified in the sn-2 position of TAG molecules (Martin, Bougnoux, Antoine, Lanson, & Couet, 1993), whereas C16:0 is primarily

68

Y. Yao et al. / International Dairy Journal 56 (2016) 64e73

▫

Fig. 2. Fatty acid composition (mass %) of the different phospholipids in bovine (-), caprine ( ) and human ( ) milk. Values are means ± standard deviation (n ¼ 10); different letters in a group indicate significant differences (P < 0.05).

esterified in the sn-1 and sn-2 positions in both bovine and caprine milk (Blasi et al., 2008). The location of C16:0 at the sn-2 position of TAG in human milk fat increases the absorption of fatty acids and decreases the loss of calcium in the infant (Maduko, Akoh, & Park, 2007). Thus, the TAG of infant formulas based on bovine and caprine milk with added vegetable oils should be modified to simulate the TAG composition of human milk fat for better absorption. 3.4. Polar lipids composition The profile of the polar lipids was similar in the three species of mammalian milk, although the content may differ due to the state of lactation, feed factors and genetic factors. The average polar lipids contents of bovine (229 mg L1) and human (225 mg L1) milk were significantly lower than that of caprine milk (288.1 mg L1; P < 0.05).

The relative proportion of each species of polar lipids is listed in Table 3. There are different profiles of polar lipids between bovine, caprine and human milk. Among the different polar lipid species, SM was more abundant in human milk, and PC was less abundant relative to that in bovine and caprine milk. This was consistent with the existing literature on class distribution of human milk polar lipids, although the relative percentage of SM was less than that in previous reports (more than 30%) (Benoit et al., 2010; Giuffrida et al., 2013; Lopez & Menard, 2011; Zou et al., 2012b). PC was the main polar lipid (33.1 and 31.6%, respectively) in bovine and caprine milk, followed by SM (25.4% and 25.0%, respectively) and PE (23.4% and 19.9%, respectively), respectively. However, Rodriguez-Alcala and Fontecha (2010) reported that the predominant phospholipid was PE followed by PC and SM in bovine and caprine milk. The differences were probably due to the breed of animals, stage of lactation, diet, and environmental and seasonal factors (Lopez et al. 2008). Although polar lipids have key functions in signal

Y. Yao et al. / International Dairy Journal 56 (2016) 64e73

69

Table 2 Triacylglyceride compositions of bovine, caprine and human milk fat.a TAG

ECN

CN

ND

Bovine milk

LaOBu MMBu PLaBu BuOM/MMCp/PLBu PMBu MOCp PMCp/POBu PPBu/SMBu MCaLa POCp/PCaLa SOBu PSBu/PoLaLa OLaCa/PPCp SOBu/SMCp OLaLa MMLa/PMCa LLaO MLaO/POCa PMLa/SLaLa OOCa MLaO/POCa SMLa LaOO PoOM POLa/MMO PSCa PoPO/OOL SMM POL PPL PPM OOO POO PPO SOO POS PPS PSS

32 32 32 34 34 36 36 36 36 38 38 38 38 38 40 40 40 42 42 42 42 44 44 44 44 44 46 46 46 46 46 48 48 48 50 50 50 52

34 32 32 36/34/38 34 38 36/38 36 36 40/38 40 38/40 40/38 40/38 42 40 48 44 42 46 44 46 48 48 46 44 50/54 46 52 50 46 54 52 50 54 52 50 52

1 0 0 1/0/2 0 1 0/1 0 0 1/0 1 0/1 1/0 1/0 1 0 4 1 0 2 1 1 2 2 1 0 2/4 0 3 2 0 3 2 1 2 1 0 0

0.90 1.14 1.51 4.22 6.19 3.52 9.71 14.02 e 5.51 5.22 5.65 e e e e e 2.22 e e e 1.12 e 2.61 2.51 e e e e e 3.23 3.43 2.81 6.77 3.56 5.09 6.32 1.60

± ± ± ± ± ± ± ±

0.05 0.12 0.13 0.21a 0.23 0.30a 0.42 0.13

± 0.18 ± 0.20 ± 0.18

± 0.11b

± 0.04b ± 0.19a ± 0.14b

± ± ± ± ± ± ± ±

0.15a 0.18b 0.20c 0.28a 0.34a 0.42c 0.28b 0.13b

Caprine milk

Human milk

e e e 0.69 e 2.41 e e e e e e 1.19 4.17 e 12.25 e 9.68 e 5.24 5.43 5.98 e 2.18 1.67 2.69 5.75 e e e 2.39 2.66 3.49 3.95 3.32 11.02 10.51 1.14

e e e e e e e e 0.35 e e e e e 0.80 0.42 1.30 2.23 2.04 e e e 4.60 e 4.96 e e 6.47 13.55 6.16 e 6.57 27.04 4.54 1.61 13.45 e 3.91

± 0.07b ± 0.08b

± 0.15 ± 0.09 ± 0.23a ± 0.17a ± 0.13 ± 0.12 ± 0.22a ± ± ± ±

± ± ± ± ± ± ± ±

0.09b 0.16c 0.21 0.12

0.20b 0.11c 0.12b 0.11c 0.07b 0.06b 0.46a 0.04c

± 0.04

± ± ± ± ±

0.11 0.05b 0.08 0.09b 0.10

± 0.60 ± 0.17a

± 0.13 ± 0.20 ± 0.12 ± ± ± ± ±

0.10a 0.16a 0.08b 0.07c 0.18a

± 0.09a

a

Abbreviations are: ECN, equivalent carbon number; CN, total carbon number; ND, number of double bonds; Bu, butyric acid; Cp, caproic acid; C, caprylic acid; Ca, capric acid; La, lauric acid; M, myristic acid; P, palmitic acid; Po, palmitoleic acid; S, stearic acid; O, oleic acid; L, linoleic acid. Values are means ± standard deviation and are based on measurements on 10 samples with three replicate measurement per milk sample (mass %). Different superscript letters in a row indicate significant differences (P < 0.05).

Table 3 Polar lipids composition of bovine, caprine and human milk.a Polar lipids (%)

Bovine milk

PE PI PS PC SM

23.42 8.97 9.07 33.12 25.40

± ± ± ± ±

0.13b 0.06b 0.07c 0.21a 0.19b

Caprine milk 19.92 9.37 14.03 31.64 25.04

± ± ± ± ±

0.10c 0.06a 0.05a 0.21b 0.17b

Human milk 25.33 7.85 13.12 24.39 29.28

± ± ± ± ±

0.14a 0.07c 0.03b 0.12c 0.14a

a Abbreviations are: PE, phosphatidylethanolamine; PI, phosphatidylinositol; PS, phosphatidylserine; PC, phosphatidylcholine; SM, sphingomyelin. Values are means ± standard deviation (n ¼ 10); different superscript letters in a row indicate significant differences (P < 0.05).

transduction that affect important cellular functions (Contarini & Povolo, 2013; Slotte, 2013), there is presently no requirement for polar lipids content in infant formula. Thus, polar lipids should also be added to infant formulas as a source of long-chain polyunsaturated fatty acids. 3.5. Cholesterol and its precursors

▫

Fig. 3. Distribution (mol %) of triacylglycerols in bovine (-), caprine ( ) and human ( ) milk fat according to equivalent carbon number (ECN). Values are means ± standard deviation (n ¼ 10); different letters in a group indicate significant differences (P < 0.05).

Sterols are a small fraction of the total milk fats that are particularly important for human nutrition. Bovine, caprine and human milk fats contained a similar profile of sterol fractions,

70

Y. Yao et al. / International Dairy Journal 56 (2016) 64e73

▫

Fig. 4. Sterol contents of bovine (-), caprine ( ) and human ( ) milk (mg 100 g1 lipid). Values are means ± standard deviation (n ¼ 10); different letters in a group indicate significant differences (P < 0.05).

bovine, caprine and human MFGs. The size parameters of the MFGs calculated from the size distribution of the different species are shown in Table 4. Laser light scattering measurements showed that the size distributions of the bovine, caprine and human MFGs ranged from 1.49 to 15.55 mm, 1.26e9.25 mm and 0.31e26.16 mm respectively. The average diameter of the fat globules in caprine milk (3.64 ± 0.33 mm) was smaller than that in bovine milk (4.89 ± 0.17 mm) and in human milk (4.53 ± 0.18 mm). The size measurements for bovine and human milk were in agreement with previously reported results, whereas the size for caprine milk was slightly larger than that in other reports (El-Zeini, 2006; Park, Juarez, Ramos, & Haenlein, 2007). The differences in the size parameters observed in this study could be due to various factors such as animal breed, diet, genetics and lactation. The absolute value of the zeta potential in caprine milk was obviously higher than that in bovine and human milk (Table 4) The difference between the milk species in zeta-potential could be explained by the differences in the polar lipids and proteins located in the MFG trilayer membrane and the minerals present in the aqueous phase which results in uncommon surface active properties. 3.7. Microstructure of bovine, caprine and human milk fat globules

with a large amount of cholesterol and minor quantities of other components (Fig. 4). Small quantities of squalene, desmosterol, lathosterol, lanosterol and two phytosterols (stigmasterol and bsitosterol) were found in these milk species. One peak had a relative retention time similar to that of campesterol, but the mass spectra were too poor to make definite assignments. The cholesterol contents in bovine, caprine and human milk were 293, 247 and 351 mg 100 g1 milk fat, respectively. A high content of cholesterol in infants could affect long-term cholesterol metabolism for reducing the risk of cardiovascular disease in adult life (Owen et al., 2008). We found that desmosterol was quantitatively the main minor sterol in human milk, consistent with the data reported by Benoit et al. (2010). In contrast, lathosterol was the main minor sterol in bovine and caprine milk. Compared with human milk, desmosterol and lanosterol were not detected in bovine and caprine milk. These minor sterols offer beneficial bioactive properties such as that of lanosterol, which might be associated with the prevention of colon cancer, and phytosterols, which might reduce cardiovascular risk (Fassbender et al., 2008). 3.6. Size distribution and apparent zeta-potential The fat globule size distributions and apparent zeta potentials in all three species were determined for the purpose of characterizing the structural characteristics and surface properties of

Table 4 Size parameters and apparent zeta-potential of bovine, caprine and human milk.a Size parameters

Bovine milk

Caprine milk

Human milk

D4,3 (mm) D3,2 (mm) zeta-potential（mV）

4.89 ± 0.17a 4.04 ± 0.16a 10.84 ± 0.66b

3.64 ± 0.33b 3.21 ± 0.12b 12.54 ± 0.39a

4.53 ± 0.18a 3.35 ± 0.21b 7.32 ± 0.65c

a Values are means ± standard deviation, n ¼ 10; different superscript letters in a row indicate significant differences (P < 0.05).

The characteristics of the bovine, caprine and human MFG microstructure in this study are presented in Fig. 5. The MFGs were spherical in shape and were dispersed in the aqueous phase of milk. The overlay of the differential interference contrast microscopy and fluorescent CLSM images showed that the TAGs were located in the core of the MFGs. The lateral organisation of polar lipids in the MFGM was investigated using CLSM with fluorescent probe Rh-DOPE. Fig. 6 shows that the fluorescence of Rh-DOPE was located at the periphery of the MFG, and the interior (mainly composed of TAG) was not labelled by Rh-DOPE. The Rh-DOPE fluorescent probe showed a variety of distribution patterns in the MFGM. Some fat globules were completely labelled by the probe, whereas in others, both fluorescent and non-fluorescent areas were visible at the periphery. Non-fluorescent areas were also present on the surface of the MFG, as indicated by the white arrows. Two hypotheses might explain these areas: (i) part of the bilayer is lost from the MFG in bovine milk after secretion by the secretory cells (Evers et al., 2008); and (ii) the fluorescent dye is not allowed into the areas (Lopez et al., 2010). It is well-accepted from the research of Lopez et al. (2010) that the nonfluorescent areas are liquid-ordered domains (lipid raft), which are mainly composed of SM together with cholesterols surrounded by the matrix of phospholipids (PC, PI, PE and, PS) in the liquid-disordered phase. Three dimensional observations allowed the characterisation of the liquid-ordered domains present in the outer bilayer of the MFGM (Fig. 7). These domains were mostly circular shapes of different size on the bovine MFGM but were more irregularly shaped non-fluorescent regions on the caprine MFGM. Both circular shapes and irregularly shaped domains were present on the surface of the human MFGM. This means that the different liquidordered shapes in the MFGM among species is mainly due to the different fatty acids composition of the phospholipids. The liquidordered domains present in the outer bilayer of the MFGM could potentially be involved in digestion, and in interaction with path-

Y. Yao et al. / International Dairy Journal 56 (2016) 64e73

71

Fig. 5. Microstructure of bovine (A1,2,3), caprine (B1,2,3) and human (C1,2,3) milk fat globules observed using CLSM with triacylglycerols stained with a Nile Red fluorescent probe (A1, B1, and C1), optical microscopy with differential interferential contrast (A2, B2 and C2), and the overlay images (C1, C2 and C3). Scale bar ¼ 10 mm.

Fig. 6. Confocal laser scanning micrograph showing (A) the emission fluorescence of Rh-DOPE fluorescent dye detected using confocal laser scanning microscopy (green), (B) fat globules observed using differential interferential contrast, and (C) the overlay of images A and B showing the fluorescent dye integrated in the milk fat globule membrane surrounding fat globules, which are characterised in their equatorial plane. Scale bar ¼ 2 mm. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

72

Y. Yao et al. / International Dairy Journal 56 (2016) 64e73

Fig. 7. 3-D microscopy images showing the heterogeneous distribution of polar lipids within the bovine (A), caprine (B) and human (C) milk fat globule membrane. Scale bar ¼ 5 mm.

ogens and viruses in the gut, so the function of the special membrane structure between different species will require further study. 4. Conclusion Our study provides comprehensive information on bovine, caprine and human MFGs based on their individual lipid composition and microstructural characteristics. It also shows that the differences in lipid composition maintained a close connection with the microstructure of the MFGs. The results of this study provide increased knowledge about milk from three species that will be useful in the manufacture of milk powder. We suggest that manufacturers of milk powder should focus on the composition of milk lipids to support infant growth but also to mimic the MFG structure of human milk. Acknowledgements This work was supported by the Jiangsu Provincial Natural Science Foundation (BK20140149) and Fundamental Research Funds for the Central Universities (JUSRP11439). References Abrahamse, E., Minekus, M., van Aken, G. A., van de Heijning, B., Knol, J., Bartke, N., et al. (2012). Development of the digestive system e experimental challenges and approaches of infant lipid digestion. Food Digestion, 3, 63e77. Bahrami, G., & Rahimi, Z. (2005). Fatty acid composition of human milk in Western Iran. European Journal of Clinical Nutrition, 59, 494e497. Bar-Yoseph, F., Lifshitz, Y., & Cohen, T. (2013). Review of sn-2 palmitate oil implications for infant health. Prostaglandins, Leukotrienes and Essential Fatty Acids, 89, 139e143. Benoit, B., Fauquant, C., Daira, P., Peretti, N., Guichardant, M., & Michalski, M. C. (2010). Phospholipid species and minor sterols in French human milks. Food Chemistry, 120, 684e691. Blasi, F., Montesano, D., De Angelis, M., Maurizi, A., Ventura, F., Cossignani, L., et al. (2008). Results of stereospecific analysis of triacylglycerol fraction from donkey, cow, ewe, goat and buffalo milk. Journal of Food Composition and Analysis, 21, 1e7. Christopherson, S. W., & Glass, R. L. (1969). Preparation of milk fat methyl esters by alcoholysis in an essentially nonalcoholic solution. Journal of Dairy Science, 52, 1289e1290. Contarini, G., & Povolo, M. (2013). Phospholipids in milk fat: composition, biological and technological significance, and analytical strategies. International Journal of Molecular Sciences, 14, 2808e2831.
pez-Sabater, M., & Rivero-Urgell, M. (1996). Fatty acid De la Presa-Owens, S., Lo composition of human milk in Spain. Journal of Pediatric Gastroenterology and Nutrition, 22, 180e185. El-Zeini, H. M. (2006). Microstructure, rheological and geometrical properties of fat globules of milk from different animal species. Polish Journal of Food Nutrition Sciences, 15/56, 147e154. Evers, J. M., Haverkamp, R. G., Holroyd, S. E., Jameson, G. B., Mackenzie, D. D. S., & McCarthy, O. J. (2008). Heterogeneity of milk fat globule membrane structure

and composition as observed using fluorescence microscopy techniques. International Dairy Journal, 18, 1081e1089. € nig, J., Walter, S., et al. Fassbender, K., Lütjohann, D., Dik, M. G., Bremmer, M., Ko (2008). Moderately elevated plant sterol levels are associated with reduced cardiovascular riskdthe LASA study. Atherosclerosis, 196, 283e288. Folch, J., Lees, M., & Sloane-Stanley, G. (1957). A simple method for the isolation and purification of total lipids from animal tissues. Journal of Biological Chemistry, 226, 497e509. Fontecha, J., Diaz, V., Fraga, M., & Juarez, M. (1998). Triglyceride analysis by gas chromatography in assessment of authenticity of goat milk fat. Journal of the American Oil Chemists Society, 75, 1893e1896. Fraga, M., Fontecha, J., Lozada, L., Martínez-castro, I., & Ju
arez, M. (2000). Composition of the sterol fraction of caprine milk fat by gas chromatography and mass spectrometry. Journal of Dairy Research, 67, 437e441. Gallier, S., Gragson, D., Cabral, C., Jimenez-Flores, R., & Everett, D. W. (2010a). Composition and fatty acid distribution of bovine milk phospholipids from processed milk products. Journal of Agricultural and Food Chemistry, 58, 10503e10511. Gallier, S., Gragson, D., Jimenez-Flores, R., & Everett, D. (2010b). Using confocal laser scanning microscopy to probe the milk fat globule membrane and associated proteins. Journal of Agricultural and Food Chemistry, 58, 4250e4257. Garcia-Saez, A. J., Chiantia, S., & Schwille, P. (2007). Effect of line tension on the lateral organization of lipid membranes. Journal of Biological Chemistry, 282, 33537e33544. Giuffrida, F., Cruz-Hernandez, C., Fluck, B., Tavazzi, I., Thakkar, S. K., Destaillats, F., et al. (2013). Quantification of phospholipids classes in human milk. Lipids, 48, 1051e1058. Koletzko, B., & Rodriguez-Palmero, M. (1999). Polyunsaturated fatty acids in human milk and their role in early infant development. Journal of Mammary Gland Biology and Neoplasia, 4, 269e284. Li, C., Yao, Y., Zhao, G., Cheng, W., Liu, H., Liu, C., et al. (2011). Comparison and analysis of fatty acids, sterols, and tocopherols in eight vegetable oils. Journal of Agricultural and Food Chemistry, 59, 12493e12498. Lopez, C., Briard-Bion, V., Menard, O., Rousseau, F., Pradel, P., & Besle, J.-M. (2008). Phospholipid, sphingolipid, and fatty acid compositions of the milk fat globule membrane are modified by diet. Journal of Agricultural and Food Chemistry, 56, 5226e5236. Lopez, C., Madec, M. N., & Jimenez-Flores, R. (2010). Lipid rafts in the bovine milk fat globule membrane revealed by the lateral segregation of phospholipids and heterogeneous distribution of glycoproteins. Food Chemistry, 120, 22e33. Lopez, C., & Menard, O. (2011). Human milk fat globules: polar lipid composition and in situ structural investigations revealing the heterogeneous distribution of proteins and the lateral segregation of sphingomyelin in the biological membrane. Colloids and Surfaces B: Biointerfaces, 83, 29e41. Maduko, C. O., Akoh, C. C., & Park, Y. W. (2007). Enzymatic production of infant milk fat analogs containing palmitic acid: optimization of reactions by response surface methodology. Journal of Dairy Science, 90, 2147e2154. Martin, J.-C., Bougnoux, P., Antoine, J.-M., Lanson, M., & Couet, C. (1993). Triacylglycerol structure of human colostrum and mature milk. Lipids, 28, 637e643. Michalski, M. C., Briard, V., & Michel, F. (2001). Optical parameters of milk fat globules for laser light scattering measurements. Lait, 81, 787e796. Michalski, M. C., Michel, F., Sainmont, D., & Briard, V. (2002). Apparent z-potential as a tool to assess mechanical damages to the milk fat globule membrane. Colloids and Surfaces B: Biointerfaces, 23, 23e30.
, A. I., & Lo
pez Sabater, M. C. (1998). Analysis of Morera Pons, S., Castellote Bargallo human milk triacylglycerols by high-performance liquid chromatography with light-scattering detection. Journal of Chromatography A, 823, 475e482. Owen, C. G., Whincup, P. H., Kaye, S. J., Martin, R. M., Smith, G. D., Cook, D. G., et al. (2008). Does initial breastfeeding lead to lower blood cholesterol in adult life? A quantitative review of the evidence. American Journal of Clinical Nutrition, 88, 305e314.

Y. Yao et al. / International Dairy Journal 56 (2016) 64e73 Park, Y. W., Juarez, M., Ramos, M., & Haenlein, G. F. W. (2007). Physico-chemical characteristics of goat and sheep milk. Small Ruminant Research, 68, 88e113. Pfeuffer, M., & Schrezenmeir, J. (2000). Bioactive substances in milk with properties decreasing risk of cardiovascular diseases. British Journal of Nutrition, 84, 155e159. Prosser, C. G., McLaren, R. D., Frost, D., Agnew, M., & Lowry, D. J. (2008). Composition of the non-protein nitrogen fraction of goat whole milk powder and goat milkbased infant and follow-on formulae. International Journal of Food Sciences and Nutrition, 59, 123e133. Rodriguez-Alcala, L. M., & Fontecha, J. (2010). Major lipid classes separation of buttermilk, and cows, goats and ewes milk by high performance liquid chromatography with an evaporative light scattering detector focused on the phospholipid fraction. Journal of Chromatography A, 1217, 3063e3066. Rombaut, R., Camp, J. V., & Dewettinck, K. (2005). Analysis of phospho- and sphingolipids in dairy products by a new HPLC method. Journal of Dairy Science, 88, 482e488.
pezSala-Vila, A., Castellote, A. I., Rodriguez-Palmero, M., Campoy, C., & Lo Sabater, M. C. (2005). Lipid composition in human breast milk from Granada (Spain): changes during lactation. Nutrition, 21, 467e473. Sanchez-Juanes, F., Alonso, J. M., Zancada, L., & Hueso, P. (2009). Distribution and fatty acid content of phospholipids from bovine milk and bovine milk fat globule membranes. International Dairy Journal, 19, 273e278. Scopesi, F., Ciangherotti, S., Lantieri, P., Risso, D., Bertini, I., Campone, F., et al. (2001). Maternal dietary PUFAs intake and human milk content relationships during the first month of lactation. Clinical Nutrition, 20, 393e397.

73

Slotte, J. P. (2013). Biological functions of sphingomyelins. Progress in Lipid Research, 52, 424e437. Smiddy, M. A., Huppertz, T., & van Ruth, S. M. (2012). Triacylglycerol and melting profiles of milk fat from several species. International Dairy Journal, 24, 64e69. €f, E., Hernell, O., Lo €nnerdal, B., & Domello €f, M. (2014). NeuroTimby, N., Domello development, nutrition, and growth until 12 mo of age in infants fed a lowenergy, low-protein formula supplemented with bovine milk fat globule membranes: a randomized controlled trial. American Journal of Clinical Nutrition, 99, 860e868. Wan, Z. X., Wang, X. L., Xu, L., Geng, Q., & Zhang, Y. (2010). Lipid content and fatty acids composition of mature human milk in rural North China. British Journal of Nutrition, 103, 913e916. Yao, Y., Zhao, G., Zou, X., Huang, L., & Wang, X. (2015). Microstructural and lipid composition changes in milk fat globules during milk powder manufacture. RSC Advances, 5, 62638e62646. Zou, X., Guo, Z., Huang, J., Jin, Q., Cheong, L. Z., Wang, X., et al. (2012b). Human milk fat globules from different stages of lactation: a lipid composition analysis and microstructure characterization. Journal of Agricultural and Food Chemistry, 60, 7158e7167. Zou, X. Q., Huang, J. H., Jin, Q. Z., Guo, Z., Liu, Y. F., Cheong, L. Z., et al. (2012a). Model for human milk fat substitute evaluation based on triacylglycerol composition profile. Journal of Agricultural and Food Chemistry, 61, 167e175.