Journal of Chromatography A, 1458 (2016) 145149

Contents lists available at ScienceDirect

Journal of Chromatography A
journal homepage: www.elsevier.com/locate/chroma

Short communication

Optimization of a single phase method for lipid extraction from milk

Zhiqian Liu a, , Simone Rochfort a,b , Benjamin G. Cocks a,b
a
Biosciences Research, Department of Economic Development, Jobs, Transport and Resources, AgriBio, 5 Ring Road, Bundoora, Victoria 3083, Australia
b
School of Applied Systems Biology, La Trobe University, Bundoora, Victoria 3083, Australia

a r t i c l e i n f o a b s t r a c t

Article history: For LCMS-based lipidomic analysis of milk, total lipid extraction from raw milk is generally conducted
Received 16 March 2016 with Folch or Bligh and Dyer methods; both methods are based on two-phase partition of lipids, and
Received in revised form 19 May 2016 thus time-consuming. In this work, three solvent systems for one-phase extraction of milk lipids were
Accepted 16 June 2016
compared with the standard Folch method. Two of the solvent systems (butanol/methanol, 3:1 and 1:1)
Available online 17 June 2016
previously tested for lipid extraction from plasma were found to provide adequate extraction for polar
lipids, but incomplete extraction for triglycerides, especially highly lipophilic species. By contrast, our
Keywords:
newly designed solvent mix composed of butanol, methanol and chloroform (at a 3:5:4 ratio) provided
Milk
Lipid similar extraction efciency for triglycerides and higher yield for some of the phospholipids, as compared
One-phase extraction to the Folch method. This new one-phase extraction method is very simple yet comprehensive and thus
Liquid chromatographymass spectrometry suitable for high throughput lipid analysis of milk samples.
2016 Elsevier B.V. All rights reserved.

1. Introduction
Fatty acid composition and the isomeric form of glycerolipids
in milk determines the absorption and metabolism of milk fat by
the human body, and has a profound inuence on the property and
quality of dairy products [1,2]. Liquid chromatography coupled to
mass spectrometry (LCMS) is the prevalent technique for the char-
acterization of lipids at the molecular species level. Generally, prior
to LCMS analysis, total lipid is extracted from milk samples using
Folch or Bligh and Dyer methods [3,4]. Both methods are based
on the use of a chloroform-methanol-water mix to create a two-
phase system and the partition of the lipid fraction in the organic
(chloroform) phase. At least two cycles of extraction are generally
required to provide adequate recovery. In addition, removing chlo-
roform and reconstituting the extracted lipids in a different solvent
system is frequently needed before LCMS analysis. Consequently,
the entire lipid extraction procedure is tedious and low throughput.
Simpler lipid extraction procedures have been tested. A one-
phase extraction method was rst tested with plasma samples
using chloroform/methanol (2:1) added to plasma at a 20:1 ratio
[5]. After extraction, this method required the removal of the
extraction solvent and reconstitution of the sample in a different
solvent system to give satisfactory chromatographic performance.
Recently, Alshehry et al. [6] tested two more one-phase solvent
systems (butanol/methanol, 3:1 and 1:1) for extracting plasma
lipids and found that butanol/methanol (1:1) mixed with plasma
(at a 10:1 ratio) resulted in efcient extraction of all major lipid
classes and the extracted sample could be analysed directly by
LCMS. This prompted us to search for a one-phase lipid extraction
method suitable for milk samples. We report here the com-
parison of the two previously described solvents systems (i.e.
butanol/methanol at 1:1 and 3:1 ratios) and a new one-phase sol-
vent mix (butanol/methanol/chloroform, 3:5:4) with the standard
Folch method in lipid extraction efciency from raw milk.
2. Materials and methods
2.1. Chemicals and reagents
Three triglyceride (TAG) standards tricaprin (TAG 30:0),
tripentadecanoin (TAG 45:0) and tristearin (TAG 54:0), and
ve polar lipid standards phosphatidylcholine (PC 32:0), phos-
phatidylethanolamine (PE 32:0), phosphatidylserine (PS 36:4),
phosphatidylinositol (PI 34:2) and sphingomyelin (SM 36:1) were
purchased from Sigma-Aldrich. Solvents used for lipid extraction
and mobile phase prepration were of chromatographic grade and
were from Merck (methanol, butanol and acetonitrile) and Sigma-
Aldrich (chloroform and isopropanol). Ammonium formate, used
as mobile phase additive, was of analytical grade (Sigma-Aldrich).

2.2. Lipid extraction from milk

Corresponding author. One bulked full cream milk sample obtained from the local
E-mail address: Zhiqian.liu@ecodev.vic.gov.au (Z. Liu). market was used for method comparison. Three one-phase lipid

http://dx.doi.org/10.1016/j.chroma.2016.06.055
0021-9673/ 2016 Elsevier B.V. All rights reserved.
146 Z. Liu et al. / J. Chromatogr. A 1458 (2016) 145149

RT: 0.00 - 25.00 SM: 3B

100
TAG 54:0
50

0
100

50 TAG 45:0
0
100

50 TAG 30:0
0
100

50 PC 32:0
0
100

50 PE 32:0
0
100

50 PS 36:4
0
100

50 PI 34:2
0
100

50 SM 36:1
0
0 2 4 6 8 10 12 14 16 18 20 22 24
Time (min)

Fig. 1. LCMS prole of spiked triglyceride and polar lipid standards after extraction by our new solvent mix butanol/methanol/chloroform (3:5:4).

extraction solvent systems (butanol/methanol at 3:1 and 1:1 ratios,
and butanol/methanol/chloroform at a 3:5:4 ratio) were compared
with the classic two-phase Folch method. For all the three one-
phase extraction protocols, one mL of solvent mix was added to
100 L of 3-fold diluted (in water) milk. The mixture was shaken by
vortex for 20 s, sonicated for 20 min and then centrifuged for 15 min
(15,000g). The supernatant was transferred to an injection vial and
analysed directly by RP-LCMS. In the case of the Folch method,
milk lipid was extracted twice by chloroform/methanol (2:1). The
combined organic phase was dried under a stream of nitrogen
and the extract was reconstituted in isopropanol/chloroform (2:1)
before LCMS analysis.
The lipid extraction efciency of the four methods was com-
pared based on: 1) the percentage recovery of three TAG (TAG 30:0,
TAG 45:0 and TAG 54:0) and ve polar lipid (PC 32:0, PE 32:0, PS
36:4, PI 34:2 and SM 36:1) standards spiked to the milk sample; 2)
the abundance (ion intensity) of representative TAG and polar lipid
species of milk, immediately after extraction and after a 3-day stor-
age at 12 C. Both the recovery of spiked authentic standards and
the extraction yield of milk native lipid species were determined
with three replicates for each extraction method.
2.3. Liquid chromatographymass spectrometry
Chromatographic separation of lipids was achieved using an
Acquity UPLC HSS T3 column (100 2.1 mm, 1.8 m, Waters) on
an Agilent 1290 Innity HPLC system. The column compartment
was maintained at 50 C and the auto-sampler at 12 C. The mobile
phase was composed of acetonitrile/water (60:40, v/v) contain-
ing 10 mM ammonium formate (A) and acetonitrile/isopropanol
(10:90, v/v) containing 10 mM ammonium formate (B). The ow
rate was 0.28 mL/min with a gradient elution of 20100% B over
20 min. The injection volume was 5 L. The lipid detection was by
LTQ-Orbitrap mass spectrometer (Thermo Scientic) operated in
electrospray ionization (ESI) positive or negative Fourier transform
mode (resolution: 60,000 for both positive and negative modes).
Identication of lipid species present in milk was performed using
accurate mass of parent ions (5 ppm) and MS2 spectra as described
previously [7,8]. Quantication of selected lipid species was based
on peak area of parent ions extracted from the full scan spectrum
using Xcalibur software (Thermo Scientic).
3. Results and discussion
The chromatographic performance of the eight spiked lipid
standards extracted by the three one-phase solvent systems is com-
parable, all displaying consistent retention time and similar peak
shape for the same lipid species. This indicates that the presence
of chloroform in our new one-phase solvent mix (one third by vol-
ume) had no adverse effect. Indeed, without sample reconstitution
(i.e. direct injection of the extract), this new one-phase solvent mix
generated satisfactory peak shape for all the three TAG and ve
 Z. Liu et al. / J. Chromatogr. A 1458 (2016) 145149 147

Butanol/methanol(1:1) Butanol/methanol(3:1) New solvent mix Folch method

120

Recovery (%) 100

80

60

40

20

0
TAG 30:0 TAG 45:0 TAG 54:0 PC 32:0 PE 32:0 PS 36:4 PI 34:2 SM 36:1

Lipid species
Fig. 2. Recovery of spiked triglyceride and polar lipid standards by four different extraction methods. Error bars are standard deviation (n = 3).

Table 1
List of milk lipid species surveyed.

Lipid class Formula CN:DBa Ion quantied m/z (calculated)

Triglycerides C33 H62 O6 30:0 [M+NH4 ]+ 572.4890

C35 H66 O6 32:0 [M+NH4 ]+ 600.5203
C37 H70 O6 34:0 [M+NH4 ]+ 628.5516
C39 H74 O6 36:0 [M+NH4 ]+ 656.5829
C41 H78 O6 38:0 [M+NH4 ]+ 684.6142
C43 H82 O6 40:0 [M+NH4 ]+ 712.6455
C45 H86 O6 42:0 [M+NH4 ]+ 740.6768
C47 H90 O6 44:0 [M+NH4 ]+ 768.7081
C49 H94 O6 46:0 [M+NH4 ]+ 796.7397
C51 H98 O6 48:0 [M+NH4 ]+ 824.7707
C53 H102 O6 50:0 [M+NH4 ]+ 852.8020
C55 H106 O6 52:0 [M+NH4 ]+ 880.8333
C57 H110 O6 54:0 [M+NH4 ]+ 908.8646
C59 H114 O6 56:0 [M+NH4 ]+ 936.8959
C61 H118 O6 58:0 [M+NH4 ]+ 964.9272

PC C38 H76 NPO8 30:0 [M+H]+ 706.5387

C42 H82 NPO8 34:1 [M+H]+ 760.5856
C44 H84 NPO8 36:2 [M+H]+ 786.6013

SM C39 H79 N2 PO6 34:1 [M+H]+ 703.5754

C44 H89 N2 PO6 39:1 [M+H]+ 773.6537
C46 H93 N2 PO6 41:1 [M+H]+ 801.6850

PE C39 H74 NPO8 34:2 [M+H]+ 716.5230

C41 H78 NPO8 36:2 [M+H]+ 744.5543
C41 H74 NPO8 36:4 [M+H]+ 740.5230

PS C42 H78 NPO10 36:2 [M+H]+ 788.5442

C44 H78 NPO10 38:4 [M+H]+ 812.5442

PI C43 H81 PO13 34:1 [MH] 835.5337

C45 H85 PO13 36:1 [MH] 863.5650
a
Total carbon number:double bond number of all fatty acids.

polar lipid standards spiked in milk prior to extraction (Fig. 1). By
contrast, direct injection of lipid fraction extracted by the Folch
solvent mix (chloroform/methanol, 2:1) could lead to severe peak
splitting and broadening (data not shown). Clearly, reconstitution
of the sample in a different solvent system is necessary for the Folch
method.
The extraction efciency of the three one-phase methods and
the standard Folch method was compared in the rst instance by
measuring the recovery of representative TAG and polar lipid stan-
dards. Overall, lower recovery (<90%) was obtained for the PC and
SM species regardless of the extraction method, whereas the recov-
ery of the PS, PI and two of the TAG species (TAG 45:0 and TAG 54:0)
varied with the extraction methods (Fig. 2). The recovery of all the
ve polar lipid species was very close across the three one-phase
extraction methods and it is worth noting that all the three one-
phase methods provided a higher recovery than the Folch method
for the PS and PI species. However, in the case of the TAG species,
a clear difference was observed across the three one-phase meth-
ods for TAG 45:0 and TAG 54:0; our new one-phase solvent mix
achieved similar results compared to the Folch method, whereas
the other two one-phase methods sustained a lower recovery for
one or both species (Fig. 2).
Bovine milk contains over one hundred TAG species with vari-
able abundance and lipophilicity [911]. In order to fully compare
the performance of the three one-phase methods with the Folch
method in extracting milk TAG, the ion intensity (reective of
148 Z. Liu et al. / J. Chromatogr. A 1458 (2016) 145149

Butanol/methanol (1:1) Butanol/methanol (3:1) New solvent mix Folch method

1800
1600
1400 A
Ion intensity
1200
1000
800
600
400
200
0
30:0 32:0 34:0 36:0 38:0 40:0 42:0 44:0 46:0 48:0 50:0 52:0 54:0 56:0 58:0

TAG species

1800 B
1600
1400
Ion intensity

1200
1000
800
600
400
200
0
30:0 32:0 34:0 36:0 38:0 40:0 42:0 44:0 46:0 48:0 50:0 52:0 54:0 56:0 58:0

TAG species
50

40
C
Ion intensity

30

20

10

0
PC-30:0 PC-34:1 PC-36:2 PS-36:2 PS-38:4 SM-34:1 SM-39:1 SM-41:1 PE-34:2 PE-36:2 PE-36:4 PI-34:1 PI-36:2

Polar lipid species

50
D
40
Ion intensity

30

20

10

0
PC-30:0 PC-34:1 PC-36:2 PS-36:2 PS-38:4 SM-34:1 SM-39:1 SM-41:1 PE-34:2 PE-36:2 PE-36:4 PI-34:1 PI-36:2

Polar lipid species

Fig. 3. Ion intensity of representative triglyceride species (A & B) and polar lipid species (C & D) extracted from raw milk by four different method. A & C: samples are analysed
immediately after extraction; B & D: samples are analysed after three days storage (at 12 C). Error bars are standard deviation (n = 3).

extraction yield) of 15 TAG species known to be present in milk and their identity was conrmed by accurate mass of parent ions
was surveyed after extraction by each of the four methods. These and MS2 spectra (Table 1).
15 TAG species were selected to cover a wide range of lipophilicity Overall, the extraction efciency of the four methods for the
native TAG species is in agreement with the recovery rate observed
 Z. Liu et al. / J. Chromatogr. A 1458 (2016) 145149
on spiked TAG standards. Butanol/methanol (1:1) mix produced
comparable results to the Folch method for TAG species of low
to medium total (acyl) carbon number (i.e. between 30 and 40),
but the extraction efciency was substantially reduced for TAG
species of higher lipophilicity (total acyl carbon number > 40), and
the higher the total carbon number, the lower the ion intensity
(Fig. 3A). Butanol/methanol (3:1) was more efcient for extract-
ing TAG compared to butanol/methanol (1:1), but reduced yield
was also observed for highly lipophilic TAG species (acyl carbon
number > 46) as compared to the Folch method. By contrast, our
new solvent mix showed similar extraction efciency as the Folch
method for all the TAG species surveyed. When these same sam-
ples were analysed after 3 days storage in the sample tray (kept at
12 C), only the new solvent mix produced satisfactory results com-
pared to the more complex Folch method, whereas a huge reduction
in ion intensity was recorded with the two butanol/methanol sys-
tems for all TAG species having a total carbon number > 42 (Fig. 3B).
Indeed, precipitation was observed in the butanol/methanol (1:1)
system a few hours after being placed in the sample tray and after
24 h in the butanol/methanol (3:1) system.
A large number of polar lipids species have also been identi-
ed in bovine milk previously, most of them belonging to PC, PE,
SM, PI and PS classes [12,13]. The efciency of the four methods in
extracting native polar lipids from milk was also compared. A total
of 13 species from the ve major polar lipid classes were surveyed
(Table 1). Overall, all the three one-phase extraction methods pro-
duced very similar results for polar lipids; the extraction efciency
of these methods was similar to that of the Folch method for PC, SM
and PE, but was clearly better than the latter for PI and PS (Fig. 3C),
conrming the nding observed in the recovery test using authen-
tic standards. For all the three one-phase extraction methods, ion
intensity of the polar lipid species did not decline after a 3-day stor-
age (Fig. 3D), indicating that polar lipids are stable in these solvent
systems.
Folch method is probably the most widely used protocol for
total lipid extraction from raw milk [1315]. As stated in Alshehry
et al. [6], a key disadvantage of this method is that the collection of
the lower phase is cumbersome resulting in increased processing
time and reduced throughput. Therefore a simpler lipid extrac-
tion method is needed in milk lipidomics research and the single
phase extraction approach is an attractive alternative. Alshehry
et al. [6] reported that butanol/methanol (1:1) was efcient for
extracting all major lipid classes from plasma in a single phase sys-
tem. However, this method proved to be unsuitable for total lipid
extraction from raw milk, due to poor recovery of TAGs. Interest-
ingly, butanol/methanol (3:1) was as effective as the Folch method
for most of the TAG species surveyed, except highly lipophilic ones,
whereas our new solvent mix produced satisfactory results for
all the representative TAG species. It should be pointed out that
although only 15 TAG species with saturated fatty acids were sur-
veyed in this test, the lipophilicity range of these molecules covers
all the essential TAG species found in milk. Consequently, our new
one-phase method is expected to be reliable for extracting all TAG
species from milk. It is worth noting, however, that all the three
one-phase methods are adequate for polar lipid extraction from
milk.
For large scale lipidomics studies, simultaneous preparation of
a large number of samples is a preferred practice. Due to the low-
throughput nature of LCMS analysis, it is common that samples
are kept in the cooled sample tray for 23 days prior to injec-
tion. Therefore, the stability of lipids in the extraction solvents is
149
important for a robust one-phase extraction method. In this regard,
our new solvent mix displayed a clear advantage over the two
butanol-methanol-based methods for milk lipid extraction. The
sharp decline of ion intensity after storage in the butanol/methanol
mix is believe to be related to the lower solubility of lipophilic
TAG species in these solvent systems as compared to our new
butanol/methanol/chloroform mix.
4. Conclusion
In conclusion, a new single phase method based on the use of
a butanol/methanol/chloroform mix (3:5:4) has been developed
for the extraction of milk lipids. Compared to the standard Folch
method, the new protocol shows similar extraction efciency for
TAG species of varying lipophilicity, but produces higher yield for
polar lipids. This single phase extraction method, which requires
no solvent removal and reconstitution steps, is very simple and is
expected to signicantly increase the throughput of sample prepa-
ration in milk lipidomics research.
References
[1] T. Karupaiah, K. Sundram, Effects of stereospecic positioning of fatty acids in
triacylglycerol structures in native and randomized fats: a review of their
nutritional implications, Nutr. Metab. 4 (2007) 16.
[2] K. Smet, J. de Block, P. van der Meeren, K. Raes, K. Dewettinck, K. Coudijzer,
Inuence of milk fatty acid composition and process parameters on the
quality of ice cream, Dairy Sci. Technol. 90 (2010) 431447.
[3] L. Folch, M. Lees, G.H.S. Stanley, A simple method for the isolation and
purication of total lipids from animal tissue, J. Biol. Chem. 726 (1957)
497509.
[4] E.G. Bligh, W.J. Dyer, A rapid method of total lipid extraction and purication,
Can. J. Biochem. Physiol. 37 (1959) 911917.
[5] P.J. Meikle, G. Wong, D. Tsorotes, C.K. Barlow, J.M. Weir, M.J. Christopher, G.L.
MacIntosh, B. Goudey, L. Stern, A. Kowalczyk, I. Haviv, A.J. White, A.M. Dart,
S.J. Duffy, G.L. Jennings, B.A. Kingwell, Plasma lipidomic analysis of stable and
unstable coronary artery disease, Arterioscler. Thromb. Vasc. Biol. 31 (2011)
27232732.
[6] Z.H. Alshehry, C.K. Barlow, J.M. Weir, Y. Zhou, M.J. McConville, P.J. Meikle, An
efcient single phase method for the extraction of plasma lipids, Metabolites
5 (2015) 389403.
[7] Z. Liu, B. Cocks, A. Patel, A. Oglobline, G. Richardson, S. Rochfort, Identication
and quantication of phosphatidylinositol in infant formulas by liquid
chromatographymass spectrometry, Food Chem. 205 (2016) 178186.
[8] Z. Liu, B. Cocks, S. Rochfort, Comparison of molecular species distribution of
DHA-containing triacylglycerols in milk and different infant formulas by
liquid chromatographymass spectrometry, J. Agric. Food Chem. 64 (2016)
21342144.
[9] J. Gresti, M. Bugaut, C. Maniongui, J. Bezard, Composition of molecular species
of triacylglycerols in bovine milk fat, J. Dairy Sci. 76 (1993) 18501869.
[10] H.R. Mottram, R.P. Evershed, Elucidation of the composition of bovine milk fat
triacylglycerols using high-performance liquid chromatography-atmospheric
pressure chemical ionisation mass spectrometry, J. Chromatogr. A 926 (2001)
239253.
[11] Q. Zhou, B. Gao, X. Zhang, Y. Xu, H. Shi, L. Yu, Chemical proling of
triacylglycerols and diacylglycerols in cow milk fat by ultra-performance
convergence chromatography combined with a quadrupole time-of-ight
mass spectrometry, Food Chem. 143 (2014) 199204.
[12] V.C. Trenerry, G. Akbaridoust, T. Plozza, S. Rochfort, W.J. Wales, M. Auldist, S.
Ajlouni, Ultra-high-performance liquid chromatography-ion trap mass
spectrometry characterisation of milk polar lipids from dairy cows fed
different diets, Food Chem. 141 (2013) 14511460.
[13] Z. Liu, P. Moate, B.G. Cocks, S. Rochfort, Comprehensive polar lipid
identication and quantication in milk by liquid chromatographymass
spectrometry, J. Chromatogr. B 978 (2015) 95102.
[14] A. Avalli, G. Contarini, Determination of phospholipids in dairy products by
SPE/HPLC/ELSD, J. Chromatogr. A 1071 (2005) 185190.
[15] P. Donato, F. Cacciola, F. Cichello, M. Russo, P. Dugo, L. Mondello,
Determination of phospholipids in milk samples by means of hydrophilic
interaction liquid chromatography coupled to evaporative light scattering
and mass spectrometry detection, J. Chromatogr. A 1218 (2011) 64766482.