Subscriber access provided by The Library | University of Bath
Food and Beverage Chemistry/Biochemistry
Fingerprinting of Phospholipids Molecular Species from Human Milk and Infant Formula Using HILIC-ESI-IT-TOF-MS and Discriminatory Analysis by Principal Component Analysis Chenyu Jiang, Baokai Ma, Shuang Song, Oi-Ming Lai, and Ling-Zhi Cheong J. Agric. Food Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jafc.8b01393 • Publication Date (Web): 14 Jun 2018 Downloaded from http://pubs.acs.org on June 17, 2018
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 31
Journal of Agricultural and Food Chemistry
Fingerprinting of Phospholipids Molecular Species from Human Milk and Infant Formula Using HILIC-ESI-IT-TOF-MS and Discriminatory Analysis by Principal Component Analysis
Chenyu Jianga, Baokai Mab, Shuang Songc*, Oi-Ming Laid,e, Ling-Zhi Cheonga*
a
Department of Food Science and Engineering, School of Marine Science, Ningbo
University, Ningbo 315211,China; b
School of Life and Sciences, Shanghai University , Shanghai, 200444,China.
c
National Institute for Nutrition and Health, Chinese Center for Disease Control and
Prevention, Beijing 100050, China d
Department of Bioprocess Technology, Faculty of Biotechnology & Bimolecular
Sciences, Universiti Putra Malaysia, 43400 UPM Serdang, Selangor, Malaysia. e
Institute of Bioscience, Universiti Putra Malaysia, 43400 UPM Serdang, Selangor,
Malaysia.
_________________________ Corresponding author: Ling-Zhi Cheong, Tel: +86-574-87608368, Fax: +86– (0)574-87608368.
E-mail:
[email protected],
Song
Shuang,
Tel:
+86-10-66237158. Email:
[email protected] 1
ACS Paragon Plus Environment
Journal of Agricultural and Food Chemistry
Page 2 of 31
1
ABSTRACT
2
Phospholipids composition in the milk fat globule membrane (MFGM) fluctuates
3
during entire lactation period in order to suit the growing needs of newborn infants.
4
Present study elucidated and relatively quantified phospholipids molecular species
5
extracted from human milk (HM), mature human milk (MHM) and infant formulas
6
(with or without MFGM supplementation) using HILIC-ESI-IT-TOF-MS system.
7
Principal component analysis was used to clarify the differences between
8
phospholipids composition in HM, MHM and infant formulas. HM and MHM
9
contained high concentrations of sphingomyeline (HM: 107.61µg/mL, MHM: 227.18
10
µg/mL), phosphatidylcholine (HM: 59.96 µg/mL, MHM: 50.77 µg/mL) and
11
phosphatidylethanolamine (PE) (HM: 25.24 µg/mL, MHM: 31.76 µg/mL). Significant
12
concentrations ( 3 months exclusive breastfeeding) and human milk samples
92
(including colostrum, transitional milk and mature milk) were collected from healthy
93
Chinese women in Huangpu city (Guangdong Province). All human milk samples
94
were collected from full expression of one breast using electronic milk pump and
95
while the baby was fed on the opposite breast. This is to assure sample homogeneity
96
as milk quality and quantity have been found to vary during one feeding and whether
97
or not milk is expressed by a pump or by the baby. All the collected human milk
98
samples were stored at -80 °C until further analysis. Written informed consent was
99
obtained from all participants and the study was approved by ethics committee of the
100
National Institute of Nutrition and Food Safety, China CDC. Infant formulas (2 liquid 5
ACS Paragon Plus Environment
Journal of Agricultural and Food Chemistry
Page 6 of 31
101
formula and 2 powder formula) were purchased from local supermarkets in Ningbo,
102
China. PLs standards including phosphatidylinositol (16:0/16:0), phosphatidylglycerol
103
(14:0/14:0), phosphatidylethanolamine (14:0/14:0), phosphatidic acid (14:0/14:0),
104
phosphatidylserine (14:0/14:0), phosphatidylcholine (14:0/14:0), sphingomyelin
105
(d18:1/12:0), lysophosphatidylglycerol (14:0), lysophosphatidylethanolamine (14:0),
106
lysophosphatidylinositol (16:0), lysophosphatidylcholine (17:0) and SPLASH
107
LIPIDOMIX (isotopic internal standard mixture) were purchased from Avanti Polar
108
Lipids (USA). Other reagents and chemicals used were either of HPLC or MS grade
109
(chloroform, methanol, acetonitrile, hexane, ammonium formate and formic acid) and
110
purchased from Sigma Aldrich (USA). Ultra-pure water was obtained from Milli-Q
111
system (Millipore, USA).
112
Milk samples preparation and extraction of PL from milk samples. Infant
113
formula (1 g powder formula) was dissolved in lukewarm water (7.0 mL) before PL
114
extraction. Meanwhile, human milk and liquid formula without prior pretreatment
115
were used directly for PL extraction. PL were extracted from 100 µL of milk samples
116
according to previously described method.27 Three parallel samples were prepared for
117
each milk samples to reduce variation caused by sampling.
118
HILIC-ESI-IT-TOF-MS for separation, identification and quantification of
119
PL in milk samples. HPLC system (LC-20AB, Shimadzu Corporation, Japan)
120
equipped with binary pump, online degasser, autosampler and thermostatic column
121
oven was coupled online to an Ion Trap-Time of Flight mass spectrometer (IT-TOF
122
MS, Shimadzu Corporation, Japan) equipped with an Electrospray Ionization (ESI)
123
source for milk PL separation, identification and quantification.
124
Separation of PL in milk samples. The extracted PL were separated on a
125
CORTECS HILIC Column (2.7 µm, 2.1 mm × 150 mm, Waters, USA). CORTECS 6
ACS Paragon Plus Environment
Page 7 of 31
Journal of Agricultural and Food Chemistry
126
HILIC Columns is a high-efficiency column based on solid-core particle that are
127
designed to retain extremely polar analytes such as phospholipids. It has been reported
128
for excellent separation of plasma phospholipids.28 Separation of PL in milk samples
129
was conducted according to our previously described method.27 Column temperature
130
was set at 40 °C. Eluent A and eluent B were water and acetonitrile/water (95/5),
131
respectively. Both eluents contained 10mM ammonium formate and 0.1% formic acid.
132
Flow rate of the eluent was set at 0.3 mL/min. The gradient elution was as follows: 0
133
min, 100% B; 40 min, 90% B; 40.5 min, 70% B; 50 min, 70% B; 50.5 min, 100% B;
134
60 min, 100% B. No significant carryover of lipids was observed in blank injection
135
(10 µL methanol).
136
Identification and quantification of PL in milk samples. Identification and
137
quantification of milk PL were conducted under negative electrospray ionization
138
mode. Instrument parameters were set as follows: Nebulizer gas flow: 1.5 L/min;
139
interface (probe) voltage: 3.5 kV; CDL temperature: 200 °C; heat block temperature:
140
200 °C; detector voltage: 1.70 kV; IT zone vacuum: 1.7×10-2 Pa; Flight tube
141
temperature: 40 °C.
142
MS data were collected in the range of 400-1000 m/z with 30 ms of ion
143
accumulation in ion trap; meanwhile, MS2 data were collected in the range of 100-800
144
m/z with 50 ms of ion accumulation and 50-100% of collision energy. MS3 data were
145
collected in the range of 100-700 m/z with 50 ms of ion accumulation and 50-100% of
146
collision energy. All the data were collected at a loop time of 0.82 s. Extracted milk
147
PL were identified by comparing obtained m/z values with calculated exact masses
148
using LIPID MAPS Structure Database (LMSD).
149
Extracted milk PLs were relatively quantified using internal standard method.
150
Calibration curves were constructed using PL standards in series of concentrations 7
ACS Paragon Plus Environment
Journal of Agricultural and Food Chemistry
Page 8 of 31
151
spiked with constant concentration of SPLASH LIPIDOMIX. Peak area of the
152
internal standards and milk PL were integrated from extracted ion chromatograms
153
(EICs). Final concentration of PL was expressed in mean ± standard deviations
154
(ng/mL).
155
Statistical analysis. Discriminatory analysis of the PL composition from human
156
milk, mature human milk and four different types of infant formulas were analyzed by
157
PCA using IBM SPSS. PCA is a statistical procedure which converts a set of
158
observations of possibly correlated variables into a set of values of linearly
159
uncorrelated variables called principal components (PCs). Each PC is defined by a
160
vector known as the eigenvector of the variance-covariance matrix. The first PC (PC1)
161
expresses the most messages of original variances and the following is the PC2. Much
162
of the total variability in the data set can be expressed by only the first several PCs to
163
entail data reduction. Each loading of variables was used for the contribution of the
164
original variables to the PC.29, 30 To achieve visual discrimination of human milk,
165
mature human milk and four different types of infant formulas, the PCA plots mapped
166
variables (n=161) and eighteen samples through loadings and scores in dimension
167
determined by the significant PCs. The loading plots represent the contribution of
168
important variables to the PCs and the score plots indicate the differences of the milk
169
PL.
170
RESULTS AND DISCUSSION
171
Separation of milk PLs using HILIC-ESI-IT-TOF-MS system. Hydrophilic
172
stationary phase and reversed-phase mobile phase (usually acetonitrile and small
173
amount of water) are commonly employed in HILIC-ESI-IT-TOF-MS system. The
174
polar mobile phase form a water-enriched layer on the surface of the polar stationary
175
phase. This water-enriched stationary layer and organic solvent-enriched mobile phase 8
ACS Paragon Plus Environment
Page 9 of 31
Journal of Agricultural and Food Chemistry
176
have different solvency for PLs resulting in distribution of PLs between these two
177
layers. The water-enriched stationary layer retain hydrophilic molecules; hence, PLs
178
are eluted in order of increasing hydrophilicity of their headgroups.25,
179
elution mode can prevent co-elution of PL molecules with similar m/z values but
180
different head groups which are commonly encountered in reversed phase liquid
181
chromatography (RPLC) that separates PL according to chain length and unsaturation
182
degree of the hydrophobic acyl moiety.33 In addition, unlike normal phase liquid
183
chromatography (NPLC) which uses water-isopropanol-hexane elution systems,
184
HILIC uses reversed-phase eluent system which can ensure good retention stability
185
and ionization efficiency in ESI.27, 33 In terms of detection, MS detection has the
186
advantage over evaporative light-scattering detection (ELSD) technique as it is
187
capable of identifying and quantifying individual molecular species of PLs.34 The
188
aforementioned reasons underlined the advantages of HILIC-ESI-IT-TOF-MS system
189
in accurate identification and quantification of milk PLs as compared to other
190
analytical methods (thin-layer chromatography, 31P nuclear magnetic resonance
191
spectroscopy, etc.)
31, 32
Such
192
Figure 1 shows the total ion chromatograms (TICs) of PLs in human milk, mature
193
human milk and four different types of infant formulas. All the different classes of
194
milk PLs with the exception of LPGs and PEs were efficiently separated and eluted
195
within 40 min. LPGs and PEs have odd and even m/z value and thus can be easily
196
distinguished in mass spectra. As aforementioned, PLs molecular species within the
197
same class were eluted was according to increasing polarity (decreasing acyl chain
198
length and increasing degree of unsaturation) which is in consistent with previously
199
reported findings.31 According to our previous findings,27 buffer modifiers namely a
200
mixture of 0.1% formic acid and 10mM ammonium formate can be used to improve 9
ACS Paragon Plus Environment
Journal of Agricultural and Food Chemistry
Page 10 of 31
201
the peak shape of neutral PLs (PEs, PCs, and SMs) and anion PLs (PI, PG, and
202
especially PS). In present study, a satisfactory tailing factor (0.80-1.15) can be
203
achieved by using similar buffer modifiers.
204
Fragmentation principle of milk PLs. Present study found all the characteristics
205
fragments required for structural elucidation of PLs present in human milk, mature
206
human milk and infant formulas can be collected under negative electrospray
207
ionization mode. In our previous study, we have compared the fragmentation of PLs
208
under both positive and negative ionization mode. Although higher intensities (about
209
twice) could be achieved under positive ionization mode, either characteristic
210
fragments or fatty acyl identification fragments was not available under ESI+ mode
211
for most PL species. In contrast, all the necessary information for structure
212
identification could be collected under ESI- mode. Moreover, lower background noise
213
and better baseline separation was observed under ESI- mode. Thus, the negative
214
mode was used for PLs analysis in present work.28 Herein, SM (d18:1/22:0), PC
215
(18:0/18:2) and PE (18:0/18:2) detected most abundantly in human milk and mature
216
human milk; and PA (16:0/18:2) detected most abundantly in infant formulas were
217
selected to elucidate fragmentation principle of PLs under negative electrospray
218
ionization mode.
219
Figure 2A shows the multistage mass spectra of SM (d18:1/22:0). Under the
220
negative ESI mode, SM (d18:1/22:0) at m/z ion 831.65 demonstrated a neutral loss of
221
60.02 between [M+HCOO]- and [M-CH3]-. This could be the characteristic fragment
222
of choline head group. The neutral loss of 322.30 between [M-CH3]- and
223
[M-CH3-R2]- provided information of the fatty acyl moiety attached to sphingosine
224
which can be attributed to the loss of docosanoic acid (C22:0). The remaining fatty
225
acyl moiety could be further calculated from [M-CH3-R2]- fragment by subtracting 10
ACS Paragon Plus Environment
Page 11 of 31
Journal of Agricultural and Food Chemistry
226
the remaining head group which was identified as C18:1. Similar fragmentation law
227
can be applied for PC (18:0/18:2) which also contains a choline headgroup (Figure
228
2B). PC (18:0/18:2) at m/z ion 830.59 also demonstrated a neutral loss of 60.02
229
between [M+HCOO]- and [M-CH3]- indicating characteristic fragment of choline
230
head group. Signals at m/z ions 279.28 in MS3 can be attributed to linoleic acid on
231
sn-2 and should represent [M-CH3-R2]-. Meanwhile, the fatty acyl structure on sn-1
232
could be calculated from [M-CH3-R2]- fragment by subtracting the remaining head
233
group and the glycerol backbone which was identified as C18:0.
234
As for PE (18:0/18:2) at m/z ion 742.53, PE firstly lost one fatty acyl chain
235
between MS and MS2. The neutral loss of 262.22 between MS and MS2 can be
236
attributed to the loss of linoleic acid. Fatty acyl substituent at sn-2 was sterically more
237
favorable to dissociate as compared to that at sn-1 due to the phosphate charge site
238
closer to the fatty acyl moiety at sn-2.35, 36 Therefore, the m/z ion at 480.30 in MS2
239
and the 283.27 in MS3 should represent [M-H-R2+OH]− and [R1−H]− respectively.
240
The neutral loss of 197.05 between [M-H-R2]- and [R1-H] should represents
241
glycerolphosphoethanolamine and could be used as characteristic fragment of PE
242
(Figure 2C). Another PLs with amine-containing head groups, PS also demonstrated
243
similar fragmentation law to that of PE. The characteristic fragment for PS is NL
244
87.03. In addition, [M-H-R2]- is replaced by [M-H-Ser]- and [M-H-Ser-R2]- (Figure
245
S1). Figure 2D shows the multistage mass spectra of PA(16:0/18:2). PA (16:0/18:2) at
246
ions
671.46
produced
[M-H-R2]-
and
[M-H-R2-Glycerol]-
in
MS2.
247
m/z
248
[M-H-R2-Glycerol]- was then used as second precursor ion to produce [R1-H]- in MS3.
249
The neutral loss of water (17.99) between [M+H-R2]- and [M+H-R2-Glycerol]- could
250
be used for characterization of PA. Fragmentation law of other anionic PLs namely PI 11
ACS Paragon Plus Environment
Journal of Agricultural and Food Chemistry
Page 12 of 31
251
(Figure S2) and PG (Figure S3) is very similar to that of PA. Except that, a neutral
252
loss of 162.05 was used as a characteristic fragments for PI.
253
Structural elucidation and quantification of PLs in human milk and infant
254
formula. Table S1 shows PLs molecular classes identified and relatively quantified in
255
human milk, mature human milk and infant formulas. A total of 161 types of PLs were
256
detected in human milk, mature human milk and infant formulas. Figure 3 shows the
257
total molecular species within each PL classes in human milk, mature human milk and
258
different infant formulas. Both samples of human milk and mature human milk were
259
found to have a diverse species of PE (45 species) and SM (ranged from 17-20
260
species). Meanwhile, all the different brands of infant formula analyzed contained
261
diverse species of PE (ranged from 26 to 38 species), PC (ranged from 20 to 22
262
species) and SM (ranged from 19 to 22 species). IF 1, IF 2 and IF 4 were found to
263
have more diverse species of lysoPLs and PA (ranged from 28 to 30 species) in
264
comparison to human milk samples (11 species) and IF 3 (17 species).
265
All the different PL classes were relatively quantified according to their
266
corresponding standard curves (RSD of retention time ≤ 0.15%, peak area ratio ≤ 5%).
267
Table 1 shows the calibration equations and R2 values of PL standards and their
268
corresponding internal standards. PLs concentration varied significantly across the
269
different samples of human milk and infant formulas. As shown in Table S1, total PLs
270
concentrations in human milk and mature human milk were found to be 406.45 and
271
228.60 µg/mL, respectively. This is in agreement with values previously reported in
272
cohort studies involving Asian22, 26, 37 and European lactating mothers20. Similar to
273
findings in present study, higher total PLs concentrations in colostrum and transitional
274
milk as compared to mature human milk have been previously reported. It has been
275
postulated that biosynthesis of PLs in mammary gland diminishes during later stages 12
ACS Paragon Plus Environment
Page 13 of 31
Journal of Agricultural and Food Chemistry
276
of lactation resulting in significant decrease in the MFGM thickness.38, 39 Meanwhile,
277
total PLs concentrations in the different infant formulas varied significantly (IL 1:
278
942.75 µg/mL, IL 2: 2274.33 µg/mL, IL 3: 150.48µg/mL; IL 4 – 958.18 µg/mL). The
279
significantly (P