Journal of Food Composition and Analysis 86 (2020) 103386

Contents lists available at ScienceDirect

Journal of Food Composition and Analysis

journal homepage: www.elsevier.com/locate/jfca

Original Research Article

Comprehensive characterization of neutral and polar lipids of buttermilk T

from different sources and its milk fat globule membrane isolates
María Visitación Calvoa,1, María Carmen Martín-Hernándeza,1, Alba García-Serranoa,
María Pilar Castro-Gómeza, Loreto Alonso-Miravallesa, Rosa García-Martína,
Javier Megino-Telloa, Leocadio Alonsob, Javier Fontechaa,*
a
Institute of Food Science Research (CIAL, CSIC-UAM), Food Lipid Biomarkers and Health Group. 28049, Madrid, Spain
b
Dairy Research Institute of Asturias (IPLA-CSIC), 33300 Asturias, Spain

A R T I C LE I N FO A B S T R A C T

Keywords: Buttermilk (BM) has recently received much attention as a source of milk fat globule membrane (MFGM) due to
Buttermilk its high polar lipid (PL) content, which gives it functional properties and health benefits. Nevertheless BM can be
Butter oil obtained from two different technological processes, i) the BM from the butter oil industry (BM1) and ii) BM
Milk fat globule membrane from the butter manufacture (BM2). Neutral lipids (NL) and PL characterization from both BM and their MFGM
Polar lipids
isolated fractions have been qualitatively and quantitatively characterized including the individual phospho- and
Lipid classes
sphingolipids by HPLC-ELSD as well as triacylglycerols (TAG) and cholesterol and FAME by GC-FID. The results
Triacilglycerides
Neutral lipids revealed that while BM (either from BM1 or BM2) had a PL fraction of 12–16 %, the MFGM isolated fraction
Phosphatidylethanolamine contained about 40 % of PL with major presence of phosphatidylethanolamine (PE), phosphatidylcholine (PC)
Phosphatidylcholine and sphingomyelin (SM). Besides, MFGM showed a significant increase in the medium molecular weight TAG
Sphingomyelin and cholesterol and almost two fold amount of PUFA content than BM due mainly to linoleic acid. The results of
this study provide deeper information on the lipid composition of BM and MFGM isolated fractions of great
importance for the design of dietary supplements with potential beneficial effects on human health.

1. Introduction by specific proteins (between 25–70 %) and a complex mixture of lipids

Milk contains lipid droplets covered with a tri-layered membrane,
called the milk fat globule membrane (MFGM), which is rich in polar
lipids (PL) as phospho- and sphingolipids and proteins that have been
associated with important biological functions and health benefits (see
reviews by Spitsberg, 2005; Singh, 2006; Ward et al., 2006; El-Loly,
2011; Guerin et al., 2017; Ortega-Anaya and Jimenez-Flores, 2019).
The main milk PLs are the glycerophospholipids as phosphatidyletha-
nolamine (PE), phosphatidylcholine (PC), phoshatidylserine (PS) and
phosphatidylinositol (PI) and the sphingolipid, sphingomyelin (SM).
There is evidence of cholesterol-lowering, anti-inflammatory, anti-
neurodegenerative activities associated with the bioactive PLs of the
MFGM (Dewettinck et al., 2008; Contarini and Povolo, 2013; Castro-
Gómez et al., 2014). For this reason there has been an increasing in-
terest in the isolation and characterization of the MFGM, but its com-
position is variable and still not fully elucidated. MFGM is constituted
(between 30–75 %) from which a high proportion are PL (26–90 %)
(Rodríguez-Alcalá and Fontecha, 2010; Castro-Gómez et al., 2014;
Holzmüller and Kulozik, 2016a). This high variability in the reported
composition of the MFGM are probably due to differences in the iso-
lation procedures used (Rombaut and Dewettinck, 2006; Dewettinck
et al., 2008), milk maturation and treatment, origin and the application
of different isolation methods to handle the raw milk and separate the
cream (Mather, 2011).
Buttermilk (BM) is a by-product obtained either from the butter
manufacturing or from the butter oil process and is mainly used in the
food industry because of its emulsifying capacity and as animal feed.
Only a limited amount is used for human consumption as fermented
milks and also added in the formulation of traditional milk products
(Lambert et al., 2016). BM composition is similar to that of skim milk,
including all the water soluble components of cream such as proteins,
lactose and minerals. However, it also contains small fat globules and a

⁎
Corresponding author.
E-mail addresses: mv.calvo@csic.es (M.V. Calvo), mc.martin@csic.es (M.C. Martín-Hernández), albamaria.garcia.serrano@csic.es (A. García-Serrano),
mpilar.castro.gomez@gmail.com (M.P. Castro-Gómez), loretoalonso2@hotmail.com (L. Alonso-Miravalles), rosagm92.rga@gmail.com (R. García-Martín),
javier.megino@csic.es (J. Megino-Tello), lalonso@ipla.csic.es (L. Alonso), j.fontecha@csic.es (J. Fontecha).
1
These authors contributed equally to this work and share the first authorship.

https://doi.org/10.1016/j.jfca.2019.103386
Received 23 July 2019; Received in revised form 13 November 2019; Accepted 1 December 2019
Available online 02 December 2019
0889-1575/ © 2019 Elsevier Inc. All rights reserved.
M.V. Calvo, et al. Journal of Food Composition and Analysis 86 (2020) 103386

high content of MFGM fragments released upon the churning of cream.

The content of PL in BM is around 6 times higher than that of whole
milk and therefore, it can be considered as an ideal source of MFGM
(Vanderghem et al., 2010). Several methods for isolation of MFGM and
the obtention of MFGM from enriched product as BM differently pro-
cessed have been described (Haddadian et al., 2018). Many attempts
have been made using micro- and ultrafiltration to remove most of the
minerals, lactose, non-protein nitrogen and small proteins (Gassi et al.,
2016; Holzmüller and Kulozik, 2016a). Both acid and sweet coagulation
of caseins prior microfiltration have been used to eliminate the caseins
(Corredig et al., 2003; Rombaut et al., 2006). Enzymatic treatment to
hydrolyze proteins and addition of calcium to precipitate MFGM have
also been explored (Rombaut and Dewettinck, 2007; Rombaut et al.,
2007; Holzmüller and Kulozik, 2016b; Holzmüller et al., 2016). The
main problem encountered with most of isolation methods is that
caseins and MFGM fragments have similar size and isoelectric points,
which makes it very difficult to separate them. The best results can be
obtained by dissociation of milk caseins with sodium citrate followed Fig. 1. Scheme for cream processing to obtain the buttermilk (BM1, BM2 and
by ultracentrifugation (Corredig and Dalgleish, 1997). Although this FBM1) as well as the isolation of the MFGM fractions characterized in this
approach may not be suitable to be used in a large scale in the food study.
industry, it is very useful for obtaining MFGM free of milk proteins
(caseins and whey proteins) and other milk components (as minerals (99.5 %), the TAG standards trinanoin and tridecanoin, the FFA stan-
and lactose), for further isolation and purification. dards, phosphatidylcholine (PC), phosphatidylinositol (PI), phosphati-
For lipid characterization of BM and MFGM, it is essential to have a dylserine (PS), phosphatidylethanolamine (PE) and sphingomyelin
fat extraction method able to extract all lipid classes including not only (SM) standards as well as 5α-cholestane and cholesterol (Chol), were
the neutral lipids but also the PLs. For this type of matrix, the best supplied by Sigma-Aldrich (St. Louis, MO, USA). Reference samples
extraction procedure is the classical method described by Folch et al. with known composition as butter fat BCR-164 and BCR-519 (EU
(1957), but it is high time-consuming and requires using a large amount Commission; Brussels, Belgium) were from Fedelco Inc. (Madrid,
of toxic solvents (as chloroform). It would obviously be preferable to Spain).
use low toxicity solvents but this does not allow an efficient extraction
of polar lipids (Hara and Radin, 1978). Therefore, care should be taken
2.2. Preparation of BM enriched in MFGM fractions by microfiltration and
to use a small quantity of less toxic solvents (as dichloromethane) and
diafiltration
minimize the time of analysis as described by Castro-Gómez et al.
(2014).
Microfiltered BM was obtained by using a filtration pilot unit (MFS-
While studies describing the MFGM lipid composition obtained ei-
1, Tetra Pak S.A., Filtration Systems, Aarhus, Denmark) equipped with
ther from milk or from BM or from cheese whey are relatively abundant
ceramic membrane (1P19-40, Pall Corporation, New York, US) with
in the literature, there is scarce information related to the composition
pore size of 0.15 μm was used. The temperature was set at 50 °C and the
of MFGM obtained from the aqueous whey released during cream
transmembrane pressure was maintained at a constant 0.2 MPa during
process to produce butter oil. Therefore, the major objective of this
filtration. Prior to filtration the membrane was flushed with 10 L of
work was carried out the characterization of the fat composition and
Milli-Q water. For each filtration experiment 4 L of the 10 % powder
the neutral and polar lipids of BM from different origin (butter manu-
buttermilk reconstituted in Milli-Q water was used. Sample was filtered
facture or butter oil process) and its MFGM isolates, with special in-
until 25 % of the initial volume of retentate was obtained. Milli-Q water
terest on the phospholipids and sphingolipids content. The molecular
was added to the retentate for diafiltration in a continuous mode until 5
species of triacylglicerides, cholesterol content and fatty acid compo-
volumes of the starting material were collected as permeate. Retentate
sition have been also described.
was frozen for and stored a -35 °C until analyses.

2. Materials and methods

2.1. Materials and reagents
BM powder 1 (BM1) was obtained from the whey released after
butter oil process of pasteurized cream, evaporated and spray dried to
produce powder. The BM1 was provided by a dairy company (Reny
Picot, Asturias, Spain). The BM powder 2 (BM2) was obtained from the
aqueous phase released during churning of cream into butter and after
freeze dried to produce powder. BM2 was provided from a dairy com-
pany (Innolact, Galicia, Spain). Besides, two powder products com-
mercialized as enriched in MFGM and enriched in PL (A and B re-
spectively) were also analyzed for comparison purposes (Fig. 1).
All solvents were HPLC grade. Dichloromethane (DCM), chloroform,
hexane, methanol, isooctane, isopropanol, dimethylformamide, and
acetonitrile were purchased from Labscan (Dublin, Ireland). Sodium
sulphate anhydrous and sodium carbonate were obtained from Panreac
(Barcelona, Spain). Methyl tert-buthyl ether (MTBE) was supplied by
VWR International Eurolab S.L. (Barcelona, Spain). Sodium methoxide
(95 %), sodium citrate dehydrate, formic acid (98 %), triethylamine
2.3. Isolation of MFGM fraction from BM
The isolation of MFGM was performed using the method described
by Corredig and Dalgleish (1997) with some modifications. BM pow-
ders (BM1 and BM2) were reconstituted in Milli-Q water 10 % (w/v)
and stirred for 15 min at room temperature. Afterwards, 2 % (w/v) of
sodium citrate dihydrate was added to the mixture and incubated using
two different conditions: a) 1 h at room temperature under agitation
and b) 18 h at 4 °C. The incubated samples were then centrifuged at
70,000 x g, 45 min at 15 °C in a Beckman Optima L-70 preparative
Ultracentrifuge (Beckman Coulter, Inc. U.S.A.) and keep in ice during
30 min. Supernatant was carefully discarded to avoid contamination of
the pellet that was washed with Milli-Q water, freeze-dried and kept
frozen at -35 °C for further analyses.
2.4. Analytical methods
2.4.1. Fat extraction
Fat was extracted using two different methods: the classical method

2
M.V. Calvo, et al. Journal of Food Composition and Analysis 86 (2020) 103386

Table 1
Assessment of the fat content (g/100 g product) and content of neutral lipids (NL) and polar lipids (PL) (g/100 g fat) of buttermilk sample 1 (BM1) and microfiltered
(FBM1) and commercial products (A and B) using the Folch (F) (Folch et al., 1957) and Löfgren (L) (Löfgren et al., 2012) fat extraction methods. The values are
shown as mean ± standard deviation.
BM1 FBM1 A B

F (n = 5) L (n = 12) F (n = 8) L (n = 8) F (n = 4) L (n = 7) F (n = 4) L (n = 6)

a
Fat content 6.96 ± 0.38 6.95 ± 1.08 18.47 ± 1.87 20.94 ± 1.02 14.90 ± 1.44 16.70 ± 2.21 24.19 ± 3.17 28.33 ± 1.56b
NL 83.61 ± 1.08 84.46 ± 1.67 82.17 ± 1.06 83.66 ± 1.35 85.41 ± 1.66 85.18 ± 1.09 61.74 ± 1.19 59.48 ± 1.82
PL 16.09 ± 0.38 14.5 ± 3.58 17.85 ± 2.09 16.08 ± 1.26 14.42 ± 0.97 14.72 ± 1.78 39.02 ± 3.33 40.52 ± 2.94

a, b
Different superscripts mean significant differences (p < 0.05) between the fat extraction methods (F and L) for a given sample and parameter.

Table 2
Content of fat and total polar lipids (PL) and lipid classes including individual neutral lipids and phospho- and sphingolipids (g/100 g fat) of buttermilks (BM1 and
BM2) and milk fat globule membranes isolates from BM1 and BM2 (MFGM-BM1 and MFGM-BM2 respectively), and commercial products A and B. The values are
shown as mean ± standard deviation.
BM1 MFGM-BM1 (n = 6) BM2 MFGM-BM2 (n = 6) A B
(n = 12) (n = 3) (n = 7) (n = 6)

Fat content 6.95 ± 1.08a 10.54 ± 0.44c 4.98 ± 0.07a 10.56 ± 0.51c 16.70 ± 2.21 28.33 ± 1.56
Total PL 14.50 ± 3.58a 40.30 ± 3.90c 12.69 ± 0.76a 39.14 ± 2.91c 14.72 ± 1.78 40.52 ± 2.94
TAG 69.34 ± 5.99a 44.38 ± 4.42c 67.28 ± 0.91a 36.08 ± 6.88c 71.51 ± 3.95 52.87 ± 4.41
DAG 13.25 ± 4.81a 11.26 ± 3.55b 17.50 ± 0.25 17.98 ± 4.31 9.47 ± 2.01 5.09 ± 1.39
FFA + Chol 1.46 ± 0.44a 2.98 ± 1.30c 2.36 ± 0.18a 5.36 ± 1.45c 3.82 ± 0.64 0.76 ± 0.24
MAG 0.07 ± 0.03a 0.30 ± 0.17c 0.14 ± 0.02a 0.79 ± 0.32c 0.06 ± 0.02 0.03 ± 0.01
GlucCer 0.08 ± 0.03a 0.19 ± 0.03c n.d. 0.19 ± 0.04c 0.12 ± 0.01 0.23 ± 0.08
LacCer 0.26 ± 0.07a 0.60 ± 0.13c 0.03 ± 0.02a 0.45 ± 0.13c 0.20 ± 0.04 0.50 ± 0.14
PE 3.44 ± 0.37a 13.07 ± 0.97c 4.50 ± 0.12a 14.29 ± 1.38c 5.71 ± 0.66 13.37 ± 1.18
PI 0.95 ± 0.52a 0.47 ± 0.18b 0.45 ± 0.20 0.51 ± 0.29 0.72 ± 0.19 2.13 ± 0.27
PS 3.28 ± 1.17a 5.53 ± 1.11c 2.22 ± 0.42a 5.10 ± 0.92c 1.22 ± 0.22 5.63 ± 0.41
PC 4.74 ± 0.56a 13.83 ± 1.39c 4.07 ± 0.09a 12.87 ± 0.96c 4.24 ± 0.49 12.90 ± 0.96
SM 2.09 ± 1.16a 7.41 ± 0.90c 1.45 ± 0.13a 6.38 ± 0.69c 2.82 ± 0.37 6.49 ± 0.88

Triacylglycerols (TAG); Diacylglycerols (DAG); Free Fatty Acids + Cholesterol (FFA + Chol); Monoacylglycerols (MAG); Phosphatidylethanolamine (PE),
Phosphatidylcholine (PC), Phoshatidylserine (PS) and Phosphatidylinositol (PI) and Sphingomyelin (SM). Glucosylceramide (GlucCer); Lactosylceramide (LacCer);
n.d.: not detectable.
a, b
Different superscripts mean significant differences (p < 0.05) between a buttermilk and their corresponding MFGM isolate, for a given lipid component.
a, c
Different superscripts mean significant differences (p < 0.01) between a buttermilk and their corresponding MFGM isolate, for a given lipid component.

described by Folch et al. (1957) and the simplified method described by
Löfgren et al. (2012) with some modifications based on the optimiza-
tion of solvent/ratio sample. Briefly, to 300 μl sample (or 60 mg if
powder, dissolved in 300 μl of Milli-Q) 2490 μl of methanol was then
added and vortex. The mixture was then slightly shaken mechanically
for 10 min at room temperature. Then 4980 μl of DCM was added,
vortex and shook for 20 min. Afterwards, 1500 μl of 20 mM acetic acid
was added and again slightly shaken for 10 min. The mixture was
centrifuged at 3200 rpm for 5 min at 4 °C. The bottom organic phase
(DCM) was carefully removed with a Pasteur pipette. The upper me-
thanol phase was washed again with 2490 μl of DCM, shaken and
centrifuged and removed and combined with that previously collected
and passed through syringe filters PVDF membrane, pore size 0.45 μm
(Sigma-Aldrich)containing approx. 100 mg of anhydrous sodium sul-
fate. Lipid extracts were collected in amber vials, flushed with nitrogen
to dryness, weighted and stored at -35 °C until further chromatographic
analysis.
2.4.2. Lipid classes by HPLC-ELSD
Separation of lipid classes was accomplished in an HPLC system
(model 1260; Agilent Technologies Inc. Palo Alto, CA, USA) coupled
with an ELSD (SEDEX 85 model; Sedere SAS, Alfortville Cedex, France)
using prefiltered compressed air as the nebulizing gas at a pressure of
350 kPa at 60 °C; and the gain was set at 3. Two columns in serie (250
× 4.5 mm Zorbax Rx-SIL column with 5-μm particle diameter; Agilent
Technologies Inc.) and a pre-column with the same packing were used.
Before analysis, samples were dissolved in CH2Cl2 (at 5 or 30 mg/mL)
and 50 μL was injected after column equilibration at 40 °C. Solvent
gradient was as detailed in Castro-Gómez et al. (2014). Samples and
standards were analyzed under the same conditions, using solvents
freshly prepared. Assays were carried out in triplicate.
2.4.3. Triacylglycerides and cholesterol determination by GC-FID
TAG and Chol analysis in samples was performed following
Fontecha, et al. (2005), on a Clarus 400 GC (PerkinElmer Ltd.,
Beaconsfield, UK) equipped with an automatic split/splitless injector
and a flame ionization detector. An Rtx-65TAG fused-silica capillary
column (30 m × 0.25 mm i.d. × 0.1-μm film thickness; Restek Corp.,
Bellefonte, PA) was used. Experimental chromatographic conditions
were carried out with a temperature program as follows: 120 °C held for
30 s, 10 °C/min to 220 °C and held for 30 s, and 6 °C/min to 350 °C and
held for 30 min. Injector and flame ionization detector temperatures
were 355 and 370 °C, respectively. Helium was used as carrier gas (172
kPa) and the injection volume was 0.5 μL of dilutions of fat (30 mg/mL)
in hexane. For qualitative and quantitative analysis of TAG, response
factors were calculated using an anhydrous milk fat(reference material
BCR-519) of known TAG and Chol composition, and glyceryl trinanoate
as internal standard (100 μL; 1 mg/mL).
2.4.4. Determination of fatty acid methyl esters (FAME) by GC-FID
FAMEs were prepared by direct transesterification using methanolic
sodium methoxide (5 % solution) according to the method described by
(Golay et al., 2006).
GC-FID analysis was carried out on an Autosystem chromatograph
(Perkin Elmer, Beaconsfield, UK) fitted with a VF-23 ms, fused-silica
capillary column (30 m x 0.25 mm i.d. x 0.25 μm film thickness, Varian,
Middelburg, Netherlands) and FID. The column was held at 60 °C for 1
min after injection, temperature-programmed at 10 °C/min to 130 °C,

3
M.V. Calvo, et al. Journal of Food Composition and Analysis 86 (2020) 103386

Fig. 2. Lipid classes profile by HPLC-ELSD including the neutral (NL) and polar lipids (PL) as well as the individual phospholipids and sphingolipids of buttermilk
samples (BM1 and BM2) and milk fat globule membranes isolates from BM1 and BM2 (MFGM-BM1 and MFGM-BM2 respectively), and commercial products A and B.
TAGs: Triacylglycerols; DAGs: Diacylglycerols; FFA + Chol: Free Fatty Acids + Cholesterol; MAG: Monoacylglycerols; GlucCer: Glucosylceramide; LacCer:
Lactosylceramide; PE: Phosphatidylethanolamine, PC: Phosphatidylcholine, PS: Phoshatidylserine; PI: Phosphatidylinositol; SM: Sphingomyelin.

then temperature-programmed at 3 °C/min to 170 °C and last ramp at
10 °C/min to 230 °C and held there for 5 min. Helium was the carrier
gas with a column inlet pressure set at 20 psig and a split ratio of 1:20.
The injection volume was 0.5 μl. Total run time was of 32 min. The
injector and detector temperatures were set at 250 °C and 270 °C, re-
spectively. For FAME determination and quantification, an anhydrous
milk fat (reference material BCR-164) (Fedelco Inc., Madrid, Spain) and
tritridecanoine as internal standard (Sigma, St. Louis, MO) were em-
ployed. All the analyses were carried out at least in triplicate.
2.4.5. Statistical analysis
Data are expressed as means ± SD. Statistical analysis of difference
was assessed by non-parametric test (U-Mann Whitney test). All sta-
tistical analyses were performed using the SPSS package (SPSS 22.0 for
Windows, SPSS Inc., Chicago, IL). The criterion for significance was
P < 0.05 for all comparisons.

4
M.V. Calvo, et al. Journal of Food Composition and Analysis 86 (2020) 103386

Table 3
Fatty acid methyl esters (FAME) content (g/100 g fat) of buttermilk samples (BM1 and BM2) and their milk fat globule membrane isolates (MFGM-BM1 and MFGM-
BM2 respectively) and of commercial products A and B. The values are shown as mean ± standard deviation.
FAME (g/100 g fat) BM1 MFGM-BM1 BM2 MFGM-BM2 A B
(n = 12) (n = 6) (n = 3) (n = 6) (n = 7) (n = 6)

C4:0 1.93 ± 0.26a 1.65 ± 0.07b 2.06 ± 0.23c 1.54 ± 0.23d 1.59 ± 0.27 1.41 ± 0.13
C6:0 1.29 ± 0.08a 1.06 ± 0.07b 1.36 ± 0.08c 0.96 ± 0.12d 1.31 ± 0.31 0.91 ± 0.13
C8:0 0.72 ± 0.05a 0.64 ± 0.04b 0.73 ± 0.01c 0.59 ± 0.09d 0.79 ± 0.19 0.50 ± 0.05
C10:0 1.79 ± 0.07a 1.45 ± 0.06b 1.92 ± 0.02c 1.45 ± 0.16d 1.90 ± 0.40 1.25 ± 0.10
C12:0 2.41 ± 0.31a 1.95 ± 0.13b 2.64 ± 0.16 2.13 ± 0.29 2.69 ± 0.34 1.88 ± 0.16
C14:0 9.04 ± 0.55a 7.62 ± 0.36b 9.76 ± 0.20c 8.22 ± 0.43d 9.48 ± 0.75 6.97 ± 0.18
C16:0 27.95 ± 1.55a 24.87 ± 0.60b 30.08 ± 0.57c 27.46 ± 0.56d 29.16 ± 0.52 25.52 ± 0.16
C18:0 10.31 ± 1.12 9.86 ± 0.35 9.92 ± 0.10 10.22 ± 0.46 10.02 ± 0.32 11.78 ± 0.35
C24:0 0.53 ± 0.18a 0.26 ± 0.03b 0.15 ± 0.21 0.32 ± 0.09 0.17 ± 0.05 0.19 ± 0.06
Total SFA 58.70 ± 3.69a 52.34 ± 1.36b 61.82 ± 2.50c 55.74 ± 1.23d 60.92 ± 1.62 53.64 ± 0.48

C14:1 c9 1.06 ± 0.21 0.89 ± 0.05 1.10 ± 0.24 0.96 ± 0.15 1.05 ± 0.18 0.88 ± 0.08
C16:1 c9 1.29 ± 0.16 1.32 ± 0.07 1.49 ± 0.02 1.54 ± 0.18 0.49 ± 0.09 1.46 ± 0.04
C18:1 c9 23.53 ± 2.26a 26.18 ± 0.34b 22.89 ± 0.29c 27.52 ± 0.75d 21.27 ± 0.47 27.57 ± 0.24
Total cis C18:1 27.06 ± 2.76a 29.47 ± 0.96b 25.36 ± 1.01c 29.00 ± 0.86d 24.91 ± 0.92 30.71 ± 0.61
C18:1 t10 1.61 ± 0.36a 0.91 ± 0.14b 0.74 ± 1.04c 0.42 ± 0.35d 0.68 ± 0.03 0.97 ± 0.14
C18:1 t11 (TVA) 1.57 ± 0.33 1.48 ± 0.30 1.03 ± 0.38 0.76 ± 0.13 1.43 ± 0.30 1.43 ± 0.22
Total trans C18:1 4.00 ± 1.06a 2.40 ± 0.26b 2.05 ± 1.82 1.24 ± 0.43 3.11 ± 0.15 2.75 ± 0.07
Total MUFA 33.50 ± 3.33 34.31 ± 1.11 30.48 ± 2.56 32.87 ± 0.86 30.27 ± 1.06 36.17 ± 0.58

C18:2 c9.c12 (LA) 4.22 ± 0.88a 7.56 ± 0.45b 3.69 ± 0.13c 6.47 ± 0.37d 3.31 ± 0.11 5.56 ± 0.18
C18:3 (ALA) 0.77 ± 0.28 0.91 ± 0.16 0.93 ± 0.12 0.91 ± 0.12 0.69 ± 0.20 0.76 ± 0.15
C18:2 c9.t11 (RA) 0.96 ± 0.51 0.99 ± 0.17 1.09 ± 0.32 1.03 ± 0.25 0.77 ± 0.07 0.89 ± 0.18
C20:3 n6 (DGLA) 0.42 ± 0.15 0.71 ± 0.13 0.48 ± 0.01 0.60 ± 0.16 0.34 ± 0.23 0.64 ± 0.10
C20:4 n6 (AA) 0.44 ± 0.02b 0.67 ± 0.07a 0.38 ± 0.01d 0.61 ± 0.19c 0.55 ± 0.08 0.61 ± 0.15
C20:5 n3 (EPA) 0.39 ± 0.08 0.29 ± 0.04 0.27 ± 0.11 0.33 ± 0.15 0.36 ± 0.10 0.18 ± 0.08
C22:6 n3 (DHA) 0.31 ± 0.11 0.34 ± 0.04 0.44 ± 0.02 0.48 ± 0.07 0.48 ± 0.16 0.45 ± 0.04
Total PUFA 7.80 ± 0.35a 13.35 ± 0.53b 7.70 ± 0.06c 11.39 ± 0.85d 8.81 ± 0.59 10.19 ± 0.18

Total CLA 0.96 ± 0.51 1.21 ± 0.34 1.09 ± 0.32 1.37 ± 0.33 0.77 ± 0.07 1.16 ± 0.16
Total n3 1.76 ± 0.11a 2.96 ± 0.22b 1.72 ± 0.08 2.07 ± 0.27 2.36 ± 0.39 2.22 ± 0.29
Total n6 5.08 ± 0.75a 9.18 ± 0.53b 4.89 ± 0.34c 7.95 ± 0.52d 4.84 ± 0.34 6.81 ± 0.27
n6/n3 2.88 ± 0.25 3.11 ± 0.24 2.84 ± 0.33c 3.90 ± 0.55d 2.09 ± 0.33 3.11 ± 0.53
SCFA 5.97 ± 0.70a 5.03 ± 0.21b 6.29 ± 0.01c 4.66 ± 0.63d 5.91 ± 0.84 4.26 ± 0.38
MCFA 14.14 ± 2.07 11.95 ± 0.57 15.10 ± 1.53 12.74 ± 0.92 15.40 ± 0.94 11.15 ± 0.42
LCFA 79.89 ± 2.77a 83.03 ± 0.72b 78.61 ± 1.52 82.59 ± 1.49 78.69 ± 1.78 84.59 ± 0.68
C12+C14+C16 39.40 ± 2.42a 34.44 ± 1.09b 42.49 ± 0.93c 37.81 ± 1.15d 41.33 ± 1.41 34.37 ± 0.50

Short-chain fatty acids (SCFAs); Medium-chain fatty acids (MCFA); Long-chain fatty acids (LCFA)a, b Different superscripts mean significant differences (p < 0.05)
between columns BM1 and MFGM−BM1 for a given FA.
c, d
Different superscripts mean significant differences (p < 0.05) between columns BM2 and MFGM−BM2 for a given FA.

3. Results and discussion subsequent extraction and characterization of lipids.

3.1. Comparison of lipid extraction methods 3.2. Optimization of MFGM isolation method

Table 1 shows the fat and the lipid classes content including the
neutral and PL in samples of BM and microfiltered FBM1 and the
commercial products A and B. Two fat extraction methods were assayed
in order to determine the most suitable procedure to extract and ana-
lyze all the lipid classes: the classical method as described by Folch
et al. (1957) (F) and the method described by Löfgren et al. (2012) (L).
Both methods showed to be very efficient in the extraction of lipid
classes from the different matrices and no differences were found for
the total neutral lipids (NL) and PL content. The amount of NL and PL
obtained by the two procedures was comparable. Nevertheless, re-
garding the fat content obtained with both methods, in the BM samples
no significant differences were found after extractions although there
was a tendency to increase yield efficiency with the L method. This was
revealed especially for the commercial ingredient B where the fat
content was significantly higher when the extraction method L was
used (24.2 % against 28.3 % for the method F versus L, respectively).
Therefore, the results showed that the L method performs a complete
extraction and higher yield of total lipids from BM with the advantage
that this method uses solvents with less toxicity and in less volume and
less time needed than the classical F method. Therefore, in this work,
the L method has been selected as the appropriate procedure for the
For isolation of the MFGM fraction by ultracentrifugation it has
been reported that the incubation of milk samples with sodium citrate 2
% (w/v) leads to increase dissociation of caseins from casein micelles
during 18 h at 4 °C which allows isolating the MFGM practically pure in
the ultracentrifugation sediment. Due to the much lower presence of
caseins in the BM samples than in milk, incubation of BM samples with
2 % (w/v) sodium citrate but during 1 h at room temperature was also
examined. The amount of the isolated pellet that could be detected after
the ultracentrifugation at the two incubation periods assayed was
comparable, indicating that the dissociation and solubilization of the
micelle caseins was complete in both cases. Also, the results showed
that the lipid composition of the isolated MFGM was similar in the two
conditions therefore the shorter incubation time was chosen as suitable
for the MFGM isolation.
3.3. Composition of neutral and polar lipid classes
Table 2 shows the fat and the lipid classes content including the
neutral and PL as well as the individual phospholipids and sphingoli-
pids of the two samples of BM from different origin either from butter
oil process or butter manufacture (BM1 and BM2 respectively) and their

5
M.V. Calvo, et al. Journal of Food Composition and Analysis 86 (2020) 103386

Table 4
Molecular species of triacylglycerols content (TAG, g/100 g fat) according to carbon number chain length (CN) of buttermilk samples (BM1 and BM2) and their milk
fat globule membrane isolates (MFGM-BM1 and MFGM-BM2 respectively) and commercial products A and B. The values are shown as mean ± standard deviation.
TAG (%) BM1 MFGM-BM1 BM2 MFGM-BM2 A B
(n = 12) (n = 6) (n = 3) (n = 6) (n = 7) (n = 6)

CN24 0.11 ± 0.04a 0.20 ± 0.07b 0.46 ± 0.19 0.35 ± 0.08 0.24 ± 0.13 0.28 ± 0.01
CN26 0.44 ± 0.04 0.40 ± 0.09 0.82 ± 0.24 0.56 ± 0.11 0.41 ± 0.19 0.42 ± 0.11
CN28 0.75 ± 0.05a 1.21 ± 0.18b 1.10 ± 0.19c 1.60 ± 0.16d 0.86 ± 0.21 0.87 ± 0.14
CN30 1.89 ± 0.17a 2.65 ± 0.22b 2.71 ± 0.17 2.87 ± 0.22 1.72 ± 0.32 1.84 ± 0.12
CN32 2.94 ± 0.11a 3.46 ± 0.32b 4.38 ± 0.19 4.63 ± 0.35 2.77 ± 0.12 3.33 ± 0.08
CN34 6.03 ± 0.27 6.42 ± 0.46 7.63 ± 0.17 8.05 ± 0.59 6.35 ± 0.26 5.27 ± 0.88
CN36 9.52 ± 0.77 10.52 ± 0.85 10.53 ± 1.02 10.02 ± 1.14 10.29 ± 0.68 9.95 ± 1.42
CN38 12.50 ± 0.96a 19.24 ± 2.32b 12.48 ± 1.56c 21.12 ± 1.23d 12.06 ± 1.23 16.26 ± 0.09
CN40 12.24 ± 0.35 12.31 ± 0.89 11.35 ± 0.38c 12.17 ± 0.71d 11.68 ± 0.22 16.08 ± 0.45
CN42 6.20 ± 0.23a 5.26 ± 0.34b 5.90 ± 0.13c 4.78 ± 0.21d 7.24 ± 0.54 5.49 ± 0.04
CN44 5.98 ± 0.30a 4.92 ± 0.38b 5.89 ± 0.24c 4.63 ± 0.46d 7.17 ± 0.63 5.41 ± 0.16
CN46 6.35 ± 0.17a 5.23 ± 0.35b 6.47 ± 0.45c 4.81 ± 0.21d 8.01 ± 0.43 6.06 ± 0.26
CN48 8.19 ± 0.27a 6.87 ± 0.43b 7.83 ± 0.29c 6.39 ± 0.25d 9.55 ± 0.21 7.70 ± 0.01
CN50 10.30 ± 0.44a 7.82 ± 0.73b 9.22 ± 0.44c 7.12 ± 0.08d 10.17 ± 1.08 9.16 ± 0.27
CN52 10.11 ± 0.61a 8.05 ± 0.46b 8.69 ± 0.26c 6.74 ± 0.55d 7.60 ± 0.70 8.11 ± 0.42
CN54 6.45 ± 0.91a 5.46 ± 0.54b 4.56 ± 0.26 4.17 ± 1.33 3.90 ± 0.32 3.76 ± 0.73

LMW-TAG 12.16 ± 0.24a 14.33 ± 0.78b 17.09 ± 0.51 18.05 ± 0.94 12.34 ± 0.70 12.01 ± 0.94
MMW-TAG 46.44 ± 0.89a 52.24 ± 0.73b 46.15 ± 1.43c 52.72 ± 0.88d 48.43 ± 1.05 53.20 ± 1.40
HMW-TAG 41.40 ± 1.12a 33.42 ± 0.85b 36.76 ± 0.95c 29.23 ± 1.59d 39.23 ± 0.92 34.79 ± 0.45

Low Molecular weight LMW-TAG (CN26-CN34); Medium Molecular weight MMW-TAG (CN36-CN44) and High molecular weight HMW-TAG (CN46-CN54)a, b

Different superscripts mean significant differences (p < 0.05) between columns BM1 and MFGM−BM1 for a given TAG.
c, d
Different superscripts mean significant differences (p < 0.05) between columns BM2 and MFGM−BM2 for a given TAG.

MFGM isolates (from BM1 and BM2) and the commercial samples (A
and B). As expected, most lipid classes analyzed were quantitatively
and qualitatively different between BM and its corresponding MFGM
isolated fraction as seen in Fig. 2. Regarding the fat content, the BM1
from the butter oil process contains 30 % more fat than BM2 obtained
from cream churning (7 % vs. 5 %) due to that high yield of the an-
hydrous butter oil while the MFGM isolates comprise almost twice as
high as the BM samples. In the commercial samples and especially for
the B sample, the fat content was almost three times higher than in the
MFGM isolates (10 vs. 28 %).
For the neutral lipids, it can be observed that the percentage of TAG
in the isolated MFGM was clearly lower, than its BM (p < 0.001).
Nevertheless the DAG percentages remain similar to that in the MFGM
fraction. The product B showed lower content of TAG and DAG (52.87
and 5.09 g/100 g fat, respectively) than that of ingredient A (71.51 and
9.47 g/100 g fat, respectively). This is due to the fact that during the
process of isolation of the MFGM by ultracentrifugation a greater con-
tent of milk fat associated to the MFGM is separated which mainly
corresponds to neutral fat. On the other hand, the Chol and other minor
components, but described as highly biological active compounds such
as GlucCer and LacCer increased significantly in the MFGMs compared
with that from its BM, due most likely because are mostly associated
with the MFGM as a constituent of the membrane (Mather, 2011). Once
more, as expected, the total PL increased around 3 times in the MFGM
isolated fractions in comparison to that from BM samples due to the
important presence of glycerophospho- and sphingolipids in the MFGM.
The amount of total PL in the MFGM fractions was similar to that found
in the commercial product B and higher than that of product A which
presented a similar content to BM samples. Individual PL (PE, PS, PC
and SM) increased in the MFGM fraction between two a three times,
with the exception of PI, which concentration remained similar to that
BM. This could be due because the PI, as located in the inner mono-
layer, is less strongly attached to the membrane and it was lost during
the isolation procedure. Values found for the individual PL in MFGM
fractions are in the same range than those found for product B. On the
other hand the BM1 and BM2 investigated in this study have a similar
amount of individual PL than those from product A. Those results in-
dicate that the fat composition of product B is similar to that of MFGM
isolate, while in the product A is close to that of BM.
When comparing the results of this study with the data reported in
the literature, it appears that the PL/fat ratio determined in the MFGM
of our study (about 40 %) was slightly lower than that calculated by
Gassi et al. (2016) (around 47 %) and higher than the values reported
by Dewettinck et al. (2008); Le et al. (2011) and Mather (2011) (be-
tween 25–33 %). Regarding the individual PL, the proportion of them in
the isolated MFGM of this study was in the same range as those reported
previously in the literature (Zheng et al., 2014; Gassi et al., 2016), in
which PE and PC are the major, followed by SM and PS and PI are
present in a lower concentration.
3.4. Fatty acid methyl esters composition
In general terms, no differences between BM1 and BM2 were found,
as regards the FAME composition (Table 3). Isolation of the MFGM
fractions led to some remarkable changes in the lipid profile and both
MFGM isolates exhibited a similar trend. Thus, a significant decrease
(p < 0.05) in the concentration of the most of saturated FA (SFA) and
therefore in its total content occurred in MFGM fractions which was
likely linked to the lower presence of NL in these samples (Table 2).
Although some authors reported the absence of short FA (C4-C8) in the
PL fractions (Sánchez-Juanes et al., 2009; Zancada et al., 2013), we
found such FAs in the isolated MFGM fractions, which could be related
to the high content of NL (Table 2). The total MUFA content did not
substantially differ among BM and MFGM fractions. Nevertheless the
oleic acid (C18:1c9) concentration significantly increased in the MFGM
isolates, reaching levels of around 29 %. A remarkable reduction (1.6-
fold) of the trans C18:1 isomers also occurred, mainly due to the lower
presence of C18:1 t10 (p < 0.05). It should be noted that the level of
C18:1 t11 (trans vaccenic acid; TVA), a physiological precursor to ru-
menic acid (RA) the major CLA isomer, also decreased. However such
decrease was not significant and neither was the increase of CLA.
BM1 and BM2 showed a similar PUFA content (around 7.7 %), but
in MFGM isolates samples, this value significantly increased until reach
13.35 % and 11.39 % in MFGM-BM1 and MFGM-BM2, respectively.
This was mainly caused by the high increase (p < 0.05) of the linoleic
acid (C18:2 c9, c12; LA) concentration in the MFGM fractions. Castro-
Gómez et al., 2016 attributed this higher PUFA content to the compo-
sition of MFGM fractions, which have a low TAG and high PL content.

6
M.V. Calvo, et al.
Fig. 3. Cholesterol (Chol) content (g/100 g fat) in: (a) BM1 and MFGM-BM1,
(b) BM2 and MFGM-BM2, and (c) commercial ingredients A and B.
However, it is important to highlight that there were no major changes
in the content of α-linolenic acid (C18:3; ALA) between BM and MFGM
isolates, so the n-6/n-3 ratio (nutritionally important) increased in the
MFGM fractions. These results support the idea that MFGM-enriched
samples, would be an interesting source of PL that provide essential
PUFA in a favorable proportion for human nutrition, as previously
7
Journal of Food Composition and Analysis 86 (2020) 103386
reported by López et al., 2017.
On the other hand, when comparing with the two commercial in-
gredients studied, it could be observed that the FAME composition of
product B was quite similar to that in the MFGM isolated fractions;
meanwhile the lipid profile of product A did not substantially differ
from those in BM.
3.5. Triacylglycerides and cholesterol determination
TAG profile of BM1 and BM2 and their corresponding MFGM iso-
lated fractions are shown in Table 4. Data from the commercial pro-
ducts A and B were also included for comparison. Sixteen TAG groups
were identified by the carbon number chain length (from CN24 to
CN54) as previously reported by Fontecha et al., 2005. Both BM and
MFGM fractions exhibited a bimodal TAG distribution with two
maxima located at CN38-CN40 and between CN50-CN52 (Table 4).
This profile was in agreement with that reported by Castro-Gómez et al.
(2017) for BM, although these authors did not found TAG shorter than
CN28. The remarkable increment (p < 0.05) of the CN38 concentration
observed in MFGM isolates may be related with their high presence of
PL and to the preferential presence of long SFA towards the TAG frac-
tion in those samples. At this respect, Castro-Gómez et al. (2017) re-
ported that among saturated species determined by UPLC/QTof-MS, the
most abundant TAG included those of the CN38 group which corre-
sponded with the molecular specie (16:0/16:0/C6:0). The pre-
dominance of low molecular weight TAG (LMW-TAG) both in BM and
in MFGM fractions could be explained by the presence of other lipid
classes as DAG, MAG, and PL that co-elute with these TAGs as described
Castro-Gómez et al. (2017). Even though FAME analysis revealed a high
presence of long-chain FA (LCFA) in MFGM fractions (Table 3), the
amount of high molecular weight TAG (HMW-TAG) significantly de-
crease in these MFGM samples (Table 4), which could be indicating that
the LCFA, mainly unsaturated, were located preferentially in the PL
fraction, as previously observed by other authors (Sánchez-Juanes
et al., 2009; Castro-Gómez et al., 2015, 2017).
As regards the Chol content (Fig. 3), although it was higher in BM2
than in BM1 (almost double), in both cases in the isolated MFGM
fractions reaching much higher values than BM samples which would
confirm the data showed in Table 2. This result was expected since Chol
is a key constituent of the MFGM, mainly located in the outer bilayer,
where interact with SM to form liquid ordered domains, also called
lipid "rafts", whose biological functions have not yet been elucidated
(Simons and Vaz, 2004; Gallier et al., 2014; Singh and Gallier, 2017).
Concerning the commercial products A and B, again the TAG profile
for sample A was closer to that found in BM samples, whereas the
product B was similar to that of the isolated MFGM fractions. Contrary
to what would be expected, the Chol concentration in B was 3-fold
lower than in A, suggesting that the commercial sample B is a polar
lipid enriched product with a very low membrane presence or that the
Chol has been loss throughout its manufacturing process.
4. Conclusions
The BM1 and BM2 obtained from different technological process
show different fat content and neutral and polar lipid content than their
MFGM isolated fractions. Through microfiltration system and further
ultracentrifugation process, MFGM fractions can be isolated and ana-
lyzed by different chromatographic technics. MFGM isolated fractions
are an excellent source of bioactive lipid components. While powdered
BM had a low fat and PL content, the powdered MFGM isolated fraction
contains twice fat content but 4 fold PL than BM with major presence of
PE, PC and SM. MFGM isolated fraction, although it has much less TAG
content than BM shows an increase in medium molecular weight TAG
and two fold PUFA due mainly to n-6 FA in particular to linoleic acid.
The cholesterol content is also twice as high in the MFGM fractions
although it is in line with the expected presence in cell membranes.
M.V. Calvo, et al.
This study provides deeper information on the lipid composition of
BM from different technological origin and their MFGM isolates of great
importance for the design of dietary supplements with potential bene-
ficial effects on human health.
Contributions
María Pilar Castro-Gómez, Alba García-Serrano, Loreto Alonso-
Miravalles, Rosa García-Martín and Javier Megino, contributed with the
analytical assays and to data acquisition and interpretation. María
Carmen Martín-Hernandez, María Visitación Calvo, Leocadio Alonso
and Javier Fontecha contributed with the conception and design of the
study, and interpretation of data. All authors have collaborated in the
drafting and revision of the article, and have approved the final version
of the manuscript for submission.
Funding
This study was supported in part by the Spanish Ministry of Science,
Innovation and Universities. Project AGL2017-87884 (MINECO/AEI/
FEDER, UE).
Acknowledgement
The samples of this study have been kindly provided by the Spanish
dairy companies: Reny Picot and Innolact.
References
Castro-Gómez, M.P., Rodriguez-Alcalá, L.M., Calvo, M.V., Romero, J., Mendiola, J.A.,
Ibañez, E., Fontecha, J., 2014. Total milk fat extraction and quantification of polar
and neutral lipids of cow, goat, and ewe milk by using a pressurized liquid system and
chromatographic techniques. J. Dairy Sci. 97, 6719–6728.
Castro-Gómez, P., Garcia-Serrano, A., Visioli, F., Fontecha, J., 2015. Relevance of dietary
glycerophospholipids and sphingolipids to human health. Prostaglandins Leukot.
Essent. Fatty Acids 101, 41–51.
Castro-Gómez, P., Rodríguez-Alcalá, L.M., Monteiro, K.M., Ruiz, A.L.T.G., Carvalho, J.E.,
Fontecha, J., 2016. Antiproliferative activity of buttermilk lipid fractions isolated
using food grade and non-food grade solvents on human cancer cell lines. Food Chem.
212, 695–702.
Castro-Gómez, P., Montero, O., Fontecha, J., 2017. In-depth lipidomic analysis of mole-
cular species of triacylglycerides, diacylglycerides, glycerophospholipids, and
sphingolipids of buttermilk by GC-MS/FID, HPLC-ELSD, and UPLC-QToF-MS. Int. J.
Mol. Sci. 18, 605.
Contarini, G., Povolo, M., 2013. Phospholipids in milk fat: composition, biological and
technological significance, and analytical strategies. Int. J. Mol. Sci. 14, 2808–2831.
Corredig, M., Dalgleish, D.G., 1997. Isolates from industrial buttermilk: emulsifying
properties of materials derived from the milk fat globule membrane. J. Agric. Food
Chem. 45, 4595–4600.
Corredig, M., Roesch, R.R., Dalgleish, D.G., 2003. Production of a novel ingredient from
buttermilk. J. Dairy Sci. 86, 2744–2750.
Dewettinck, K., Rombaut, R., Thienpont, N., Le, T.T., Messens, K., Van Camp, J., 2008.
Nutritional and technological aspects of milk fat globule membrane material. Int.
Dairy J. 18, 436–457.
El-Loly, M.M., 2011. Composition, properties and nutritional aspects of milk fat globule
membrane. A review. Polish J. Food Nutr. Sci. 61, 7–32.
Folch, J., Lees, M., Sloane Stanley, G.H., 1957. A simple method for the isolation and
purification of total lipids from animal tissues. J. Biol. Chem. 226, 497–509.
Fontecha, J., Goudjil, H., Ríos, J.J., Fraga, M.J., Juárez, M., 2005. Identity of the major
triacylglycerols in ovine milk fat. Int. Dairy J. 15, 1217–1224.
Gallier, S., Laudscher, A., Jiménez-Florez, R., 2014. The milk fat globule membrane:
structure, methodology for its study, and functionality. In: Boland, M., Golding, M.,
Singh, H. (Eds.), Food Structures, Digestion and Health. Academic press, San Diego,
CA, USA, pp. 107–142 ISBN 978-0-12-404610-8.
Gassi, J.Y., Blot, M., Beaucher, E., Robert, B., Leconte, N., Camier, B., Rousseau, F.,
Bourlieu, C., Jardin, J., Briard-Bion, V., Lambert, S., Gésan-Guiziou, G., Lopez, C.,
Gaucheron, F., 2016. Preparation and characterization of a milk polar lipids enriched
ingredient from fresh industrial liquid butter serum: combination of physicochemical
modifications and technological treatments. Int. Dairy J. 52, 26–34.
Journal of Food Composition and Analysis 86 (2020) 103386
Golay, P.-A., Dionisi, F., Hug, B., Giuffrida, F., Destaillats, F., 2006. Direct quantification
of fatty acids in dairy powders with special emphasis on trans fatty acid content. Food
Chem. 101, 1115–1120.
Guerin, J., Burgain, J., Gomand, F., Scher, J., Gaiani, C., 2017. Milk fat globule membrane
glycoproteins: valuable ingredients for lactic acid bacteria encapsulation. Crit. Rev.
Food Sci. Nutr. https://doi.org/10.1080/10408398.2017.1386158.
Haddadian, Z., Eyres, G.T., Bremer, P., Everett, D.W., 2018. Polar lipid composition of the
milk fat globule membrane in buttermilk made using various cream churning con-
ditions or isolated from commercial samples. Int. Dairy J. 81, 138–142.
Hara, A., Radin, N.S., 1978. Lipid extraction of tissues with a low-toxicity solvent. Anal.
Biochem. 90, 420–426.
Holzmüller, W., Kulozik, U., 2016a. Technical difficulties and future challenges in iso-
lating membrane material from milk fat globules in industrial settings. A critical
review. Int. Dairy J. 61, 51–66.
Holzmüller, W., Kulozik, U., 2016b. Isolation of milk fat globule membrane (MFGM)
material by coagulation and diafiltration of buttermilk. Int. Dairy J. 63, 88–91.
Holzmüller, W., Gmach, O., Griebel, A., Kulozik, U., 2016. Casein precipitation by acid
and rennet coagulation of buttermilk: impact of pH and temperature on the isolation
of milk fat globule membrane proteins. Int. Dairy J. 63, 115–123.
Lambert, S., Leconte, N., Blot, M., Rousseau, F., Robert, B., Camier, B., Gassi, J.-Y., Cauty,
C., Lopez, C., Gésan-Guiziou, G., 2016. The lipid content and microstructure of in-
dustrial whole buttermilk and butter serum affect the efficiency of skimming. Food
Res. Int. 83, 121–130.
Le, T.T., van Camp, J., Pascual, P.A.L., Meesen, G., Thienpont, N., Messens, K.,
Dewettinck, K., 2011. Physical properties and microstructure of yoghurt enriched
with milk fat globule membrane material. Int. Dairy J. 21 (10), 798–805.
Löfgren, L., Ståhlman, M., Forsberg, G., Saarinen, S., Nilsson, R., Hansson, G.I., 2012. The
BUME method: a novel automated chloroform-free 96-well total lipid extraction
method for blood plasma. J. Lipid Res. 53, 1690–1700.
López, C., Blot, M., Briard-Bion, V., Cirié, C., Graulet, B., 2017. Butter serums and but-
termilks as sources of bioactive lipids from the milk fat globule membrane: differ-
ences in their lipid composition and potentialities of cow diet to increase n-3 PUFA.
Food Res. Int. 100, 864–872.
Mather, I.H., 2011. Milk fat globule membrane. In: In: Fuquay, J.W., Fox, P.F.,
McSweeney, P.L.J. (Eds.), Encyclopedia of Dairy Sciences Vol. 3. Academic Press, San
Diego, CA, USA, pp. 680–690. https://doi.org/10.1016/B978-0-12-374407-4.
00337-X.
Ortega-Anaya, J., Jimenez-Flores, R., 2019. Symposium review: The relevance of bovine
milk phospholipids in human nutrition—Evidence of the effect on infant gut and
brain development. Journal of Dairy Science 102 (3), 2738–2748. https://doi.org/10.
3168/jds.2018-15342.
Rodríguez-Alcalá, 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. J.
Chromatogr. A 1217, 3063–3066.
Rombaut, R., Dewettinck, K., 2006. Properties, analysis and purification of milk polar
lipids. Int. Dairy J. 16, 1362–1373.
Rombaut, R., Dejonckheere, V., Dewettinck, K., 2006. Microfiltration of butter serum
upon casein micelle destabilization. J. Dairy Sci. 89, 1915–1925.
Rombaut, R., Dewettinck, K., 2007. Thermocalcic aggregation of milk fat globule mem-
brane fragments from acid buttermilk cheese whey. J. Dairy Sci. 90, 2665–2674.
Rombaut, R., Dejonckheere, V., Dewettinck, K., 2007. Filtration of milk fat globule
membrane fragments from acid buttermilk cheese whey. J. Dairy Sci. 90, 1662–1673.
Sánchez-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.
Int. Dairy J. 19, 273–278.
Simons, K., Vaz, W.L.C., 2004. Model systems, lipid rafts, and cell membranes. Annu. Rev.
Biophys. Biomol. Struct. 33, 269–295.
Singh, H., 2006. The milk fat globule membrane-A biophysical system for food applica-
tions. Curr. Opin. Colloid Interface Sci. 11, 154–163.
Singh, H., Gallier, S., 2017. Nature’s complex emulsion: the fat globules of milk. Food
Hydrocoll. 68, 81–89.
Spitsberg, V.L., 2005. Invited review: bovine milk fat globule membrane as a potential
nutraceutical. J. Dairy Sci. 88, 2289–2294.
Vanderghem, C., Bodson, P., Danthine, S., Paquot, M., Deroanne, C., Blecker, C., 2010.
Milk fat globule membrane and buttermilks: from composition to valorization.
Biotechnologie, Agronomie, Société et Environnement 14, 485–500.
Ward, R.E., German, J.B., Corredig, M., 2006. Composition, applications, fractionation,
technological and nutritional significance of milk fat globule membrane material. In:
Fox, P.F., McSweeney, P.L.H. (Eds.), Advanced Dairy Chemistry: Vol 2 Lipids.
Springer, Boston, MA, USA, pp. 213–244 ISBN 978-0-387-28813-0.
Zancada, L., Pérez-Díez, F., Sánchez-Juanes, F., Alonso, J.M., García-Pardo, L.A., Hueso,
P., 2013. Phospholipid classes and fatty acid composition of ewe’s and goat’s milk.
Grasas y Aceites 64, 304–310.
Zheng, H., Jiménez-Flores, R., Gragson, D., Everett, D.W., 2014. Phospholipid archi-
tecture of the bovine milk fat globule membrane using giant unilamellar vesicles as a
model. J. Agric. Food Chem. 62, 3236–3243.