Talanta 50 (2000) 1211 1222

www.elsevier.com/locate/talanta

Speciation of iron in breast milk and infant formulas whey

by size exclusion chromatography-high performance liquid
chromatography and electrothermal atomic absorption
spectrometry
P. Bermejo a,*, E. Pena a, R. Domnguez a, A. Bermejo a, J.M. Fraga b,
J.A. Cocho c
a

Department of Analytical Chemistry, Nutrition and Bromatology, Faculty of Chemistry, Uni6ersity of Santiago de Compostela,
A6da de las Ciencias, 15705 Galeras SN, Spain
b
Department of Pediatrics, Faculty of Medicine, Uni6ersity of Santiago de Compostela, Galeras SN,
15705 Santiago de Compostella, Spain
c
Laboratory of Metabolic Disorders, Complejo Uni6ersitario, Uni6ersity of Santiago de Compostela, Galeras SN,
15705 Santiago de Compostela, Spain
Received 5 January 1999; received in revised form 1 June 1999; accepted 23 July 1999

Abstract
Speciation of iron in milk was carried out by high performance liquid chromatography (HPLC) and electrothermal
atomic absorption spectrometry (ETAAS). Milk whey was obtained and low molecular weight protein separation was
performed by size exclusion chromatography (SEC) with a TSK Gel SW glass guard (Waters) pre-column and a
TSK-Gel G2000 glass (Toso Haas) column. After studying water as a possible mobile phase, this mobile phase was
carefully selected in order to avoid alterations of the sample and to make subsequent iron determination in the protein
fractions easier by ETAAS. The proposed method is sensitive (limit of detection [LOD] and LOQ 1.4 and 4.7 mg l 1,
respectively) and precise (relative standard deviation [RSD] B 10%). Iron is principally found in the proteins of 3 and
76 kDa in breast milk, and it is irregularly distributed in infant formulas. 2000 Elsevier Science B.V. All rights
reserved.
Keywords: Iron speciation; Milk; SEC-HPLC; ETAAS

1. Introduction

* Corresponding author. Tel.: +34-981-591-079; fax: +34981-595-012.

E-mail address: qn1956@usc.es (P. Bermejo)

Iron deficiency anemia (IDA) is a priority nutritional problem in industrialized as well as developing countries. Besides anemia per se, tissue iron
deficiency may lead to a defect in learning capac-

0039-9140/00/$ - see front matter 2000 Elsevier Science B.V. All rights reserved.
PII: S 0 0 3 9 - 9 1 4 0 ( 9 9 ) 0 0 2 3 3 - 7

1212

P. Bermejo et al. / Talanta 50 (2000) 12111222

ity, cognitive performance and abnormalities in

psychomotor development in infants and preschoolers. This deficiency also effects work performance in adults and an increase in the frequency
of low birth weight, prematurity, and prenatal
mortality in pregnancy can be found [1]. Because
of the risk of cancer and heart disease in individuals with high iron stores [2,3] it is not recommended to supply iron to individuals who do not
require it. A high priority must be assigned to the
prevention of iron depletion among infants [4],
and so, iron fortification has been widely used in
the industrialized world.
Breast milk and/or milk formulas are the main
nutrient fluids of newborn infants. Trace elements
are usually added to infant formulas as inorganic
salts, whereas in milk, these elements are bound
to different compounds, which affect bioavailability. In order to carry out a rational supplementation, breast milk is used as a reference to evaluate
the nutritional content of alternative formulas,
assuming that the composition of breast milk may
satisfy the growing demands of healthy infants
during the early months of life [5]. It is interesting
to know how the highest breast milk bioavailability might depend on the distribution among the
different milk proteins and how inorganic salts
are found in infant formulas after being added.
Few reports can be found in the literature.
Previous results have been obtained by means of
size exclusion chromatography-high performance
liquid chromatography (SEC-HPLC) and inductive coupled plasma-atomic emission spectrometry
(ICP-AES) [5 7]. Negretti de Bratter et al. [8]
used instrumental neutron activation analysis
(INAA) as the reference method for the quality
control of the shape of the element profiles (Se,
Fe, Zn) obtained with ICP-AES after chromatographic separation. When the sensitivity of ICPAES is not sufficient ICP-MS may be used.
However, instruments for these techniques are
expensive and not available in many laboratories.
Based on a cost-benefit analysis, atomic absorption spectrometry (AAS) seems to be the preferred
method. The detection limits obtained by AAS
methods, requiring a small sample volume, are
low enough to allow the determination of most of
the trace elements in speciation. Consequently,

the use of AAS, in particular electrothermal

atomic absorption spectrometry (ETAAS) needs
to be considered [9,10]. Thus, in an earlier study
Fransson and Lonnerdal [11] determined the distribution of iron among various fractions of
breast milk by HPLC, ultrafiltration and AAS.
The samples were freeze dried, ashed at 600C for
5 h, and dissolved in 1:1:1 HCl/HNO3/H2O before
analysis.
In this work, iron speciation in milk was carried
out by SEC-HPLC and ETAAS. Milk fat and
casein micelles were removed and the milk whey
obtained was chromatographied. Special attention
was paid to the column and mobile phase selection, when looking for the best separation in the
lowest range of molecular weights, where iron
presence had been reported [58,11].
Water was studied as mobile phase [12,13].
Then, the mobile phase was carefully selected,
salinity was decreased, in order to avoid contaminations, stability problems of the organometal
complex, undesired interactions with the sample
to a great extent and interferences of our mobile
phase in the direct determination of iron in
protein fractions by ETAAS.

2. Experimental

2.1. Apparatus
Milk whey was obtained using an ultracentrifuge L8-Beckmann with a SW-40 rotor.
A Crison pH-meter equipped with a combined
electrode gas-calomel INGOLD U455-Ag 7.0
DIN pH 014 was utilized to determine pH of
mobile phases.
For the chromatographic separation of whey
proteins, a High Performance Liquid Chromatograph 625-LC System (Waters, USA)
equipped with a TSK gel SW glass guard precolumn, 4 cm 8 mm, and a TSK gel G 2000
glass, 30 cm 8 mm (Toso Haas, Japan) column
was employed. The column has a selected silicabased packing, derivatized using the glycol ether
function containing spherical 10 mm particles with
pore sizes of 13 nm. The Waters 625 LC system is
a non-metallic HPLC solvent delivery system for

P. Bermejo et al. / Talanta 50 (2000) 12111222

liquid chromatography applications where wetted

surface material is made of inert, polymeric materials and flexible PEEK tubing. Peaks were detected by an UV detector model 486 (Waters,
USA) at 254 nm, and acquisition and processing
of data were performed by the Millennium Chromatography Manager System, version 2.15 (Waters, USA) programme.
Iron measurement in milk and milk whey was
carried out using an Atomic Absorption Spectrophotometer Perkin Elmer model 5000 (Perkin
Elmer, Germany) at 248.3 nm, a hollow cathode
lamp operating at 30 mA, a slit of 0.2 nm, and an
acetylene air-flame.
Iron measurement in protein fractions was
carried out using an Atomic Absorption
Spectrophotometer 1100 B (Perkin Elmer, Germany) with a hollow cathode lamp and a
HGA-700 graphite furnace (Perkin Elmer,
Germany). The instrument was fitted with
an AS-70 autosampler. Pyrolitic graphite tubes
with Lvov platforms were used throughout
the course of this study. Instrument settings
for Fe determination are summarized in Table
1.

2.2. Reagents
All solutions were performed with ultrapure
water, specific resistivity 18 MV cm, from a MilliQ purification system (Millipore).

1213

In order to prepare mobile phases, different

reactives (ammonium nitrate, NH4NO3, Suprapur
(Merck, Germany), ammonia solution, NH3,
Suprapur (Merck, Germany), sodium azide, NaN3
(Sigma, St. Louis, MO) were used. The calibration column was performed using protein standards ribonuclease A, ovalbumine, albumine,
aldolase, coming from HMW Gel Filtration calibration Kit and LMW Gel Filtration Calibration
Kit (Pharmacia Biotech, USA), and cianocobalamine (Sigma).
A previously diluted stock solution of iron 1 g
l 1 (Merck, Germany), was employed to optimize
ETAAS temperature programme and to obtain
the calibration graphs. Finally, magnesium nitrate, MgNO3, Suprapur (Merck, Germany), was
assayed as a possible modifier.
A Reference Material A-11 non-fat milk of the
International Atomic Energy Agency (IAEA) with
a certified iron content was used.

2.3. Procedure
2.3.1. Sampling
Breast milk samples were collected following
strict precautions in order to minimize contamination and avoid alterations. The samples were collected by hoc trained personnel in polyethylene
flasks using a motorized pump. Care was paid to
avoid touching the inner wall of the device or
flask.

Table 1
Instrumental conditions and furnace programme for iron determination in protein fractions
Step

T (C)

t (s)
Ramp

Hold

Dry
Pyrolysis1
Pyrolysis2
Atomization
Clean

100
700
1350
2400
2600

15
10
5
0
1

30
25
20
3
3

Wavelength (nm)
Background corrector
Lamp (mA)
Signal processing
Sample volume

248.3
Deuterium
30
Peak area
20 ml

Purge gas
Graphite tubes
Injection volume (ml)
Slit width

Argon
Pyrolitic
20
0.2 nm

Gas flow (ml min1)

Signal read out

300
300
300
0
300

P. Bermejo et al. / Talanta 50 (2000) 12111222

1214

the retention time of the different peaks. A second

injection of 100 ml was performed afterwards and
the eluent emerging from the UVA-VIS detector
was collected in different fractions according to
the number of different chromatographic peaks
obtained for each sample.

Fig. 1. Total procedure.

With regard to infant formulas, commercially

available, solutions were prepared by dissolving
milk powder using ultrapure water, following the
manufacturers instructions.
Containers and covers (polyethylene) were kept
in nitric acid for at least 48 h, rinsed three times
with ultrapure water and maintained dried until
used. Samples were stored at 20C until treatments were performed.

2.3.2. Iron determination in milk and milk whey

Total values of iron concentration in milk and
milk whey were directly determined by an AAS
flame [14] using a high performance nebulizer.
The addition procedures were always used.
2.3.3. Sample preparation: ultracentrifugation
Milk samples were ultracentrifuged [1416] at
31 000 rpm (160 000 g) for 60 min, with 1 min
acceleration and 1 min deceleration times. Milk
whey was taken out with a micropipette after fat
separation. The lower phase (casein micelles) remained in the bottom of the tube.

2.3.4.2. Iron determination. Iron was directly determined in the collected fractions by ETAAS,
without addition of a chemical modifier. Calibration was performed using standards of iron diluted with the mobile phase at concentrations of
0, 10, 20, 30 mg l 1. The volume of sample
injected was 20 ml. As the volumes of fractions are
known (0.51.5 ml), the iron concentration was
given as ng fraction 1.

3. Results and discussion

3.1. Water as a mobile phase

Water was assayed to find a simple mobile
phase which avoids interferences in iron measurement and undesirable interactions with the sample. Chromatograms of an infant formula were
compared, using only water as the mobile phase
and using a solution of the bactericide sodium
azide 0.05% m V 1 [6], these results are shown in
Fig. 2. Negative peaks and signal distortions appeared indicating that the aforementioned compound must be taken out of the mobile phase.
Nevertheless to avoid the rapid deterioration of
the column the use of the sodium azide was
proposed only in the washing programme of the
column.

3.2. Chromatographic conditions

2.3.4. Iron speciation
2.3.4.1. Chromatographic separation. Milk wheys
were filtered using Millex GV13 0.22 mm sterile
units (Millipore, France), and 100 ml were injected
in the chromatographic system, with a flow rate
of 1 ml min 1. Measurement wavelength was 254
nm (Fig. 1). As each milk can present a different
protein profile it was always necessary to perform
a first injection to know the protein profile and

The term non size effects in SEC includes

attractive interactions, such as ion exchange and
hydrophobic binding, which tend to increase the
elution volumes of solutes, and increase forces of
electrostatic repulsion with the opposite effect. A
balance must be made between the need to increase ionic strength to reduce ionic electrostatic
interactions and to decrease ionic strength to limit
hydrophobic interaction [17].

P. Bermejo et al. / Talanta 50 (2000) 12111222

Early in the development of electrothermal atomization Lvov [18] found that molecular absorption by alkali halides was one of the major
causes of interferences, when these halides are
present in a concentration four or five orders of
magnitude higher than the analyte element. Due
to the very high sensitivity of the graphite furnace
technique, this relationship is very quickly
reached. For this reason the utilization of sodium
chloride at concentrations of 0.05 0.3 M as one
of the components of the mobile phase, as other
authors using ICP-AES [5 8] had reported, was
not possible here.
Ammonium nitrate concentration was optimized in the range of 0.1 0.4 M using a protein
mixture (0.02 mg of cianocobalamine, 0.19 mg
ribonuclease A, 0.21 mg of ovalbumine, 0.19 mg
of albumin, 0.77 mg of aldolase). Using ammonium nitrate 0.2 M the best separation at higher
retention times (lower molecular weights) was obtained. This fact was observed because of the
peak appearance corresponding to the aggregate
which elutes slightly before the true peak of ribonuclease A (retention time= 15 min, Fig. 3).
Different pHs around neutrality were assayed (pH
6.1, 6.7, 7.3) without observing significant changes
in the elution. Ammonia solution was added to
obtain pH 6.7, reported as milk pH [19]. The final
composition of mobile phase was 0.2 M NH4NO3
and 3.2410 4 M NH3.
Using this mobile phase the best peak definition
was obtained, as can be seen in the chro-

1215

matograms shown in Fig. 4, for a breast milk

sample and the same infant formula used in the
chromatograms of Fig. 2.

3.3. Column calibration

In order to know the molecular weight of the
proteins found in the milk whey, the column was
calibrated using a series of standard proteins with
certified molecular weight. These proteins were
chromatographied and the retention times obtained as well as the molecular weight are shown
in Table 2.
The equation obtained to column calibration
was:
Log MW (kDa)= 0.276+ 7.591tR (min)
with a correlation coefficient r=0.974.

3.4. Precision in the protein separation

The final aim of this work is to determine iron
in the different proteins of the milk whey, and for
this reason it is necessary to know the repeatibility
of the protein retention times in order to be able
to separate the protein fractions. To perform this
study ten replicates of a breast milk sample and
12 replicates of an infant formula sample were
chromatographed; the results obtained for the
different proteins found in both types of milk are
shown in Table 3. The retention times for all
proteins were constant, and the relative standard

Fig. 2. Chromatograms demonstrating the negative effect of sodium azide addition (used as bactericide) in the elution. Mobile phase:
(1) water; (2) sodium azide 0.05% (w/v).

P. Bermejo et al. / Talanta 50 (2000) 12111222

1216

Fig. 3. Effect of different concentrations of ammonium nitrate in the elution of a standard protein mixture.

Fig. 4. Chromatograms: (1) breast milk sample; (2) infant formula sample.

deviation (RSD) for the absorbance expressed in

height or area peak mode can be considered good
in all cases. An R.S.D. of 12.3% was obtained for
the peak corresponding to 2 kDa, this value could
be produced because of being the nearest peak to
the separation limit of the column.

3.5. Graphite furnace programme

The chemical modifier proposed for the iron
determination in ETAAS is magnesium nitrate.

Table 2
Column calibration
Standard protein MW (kDa)

Retention time (min)

Cianocobalamine
1.355
Ribonuclease A
13.700
Ovalbumine
43.000
Albumine
67.000
Aldolase
158.000

15.5
13.6
10.5
9.6
8.9

P. Bermejo et al. / Talanta 50 (2000) 12111222

1217

Table 3
Precision in protein separation
Breast milk (n = 10)

Infant Formula (n =12)

MW (kDa)

MW (kDa)

101
78
53
14
3
2

RSD (%)
Retention time

Peak height

Peak area

0.0
0.0
0.0
0.0
0.0
0.0

2.3
0.8
2.0
1.0
9.1
9.0

2.9
2.4
5.0
0.6
5.6
8.3

To optimize the pyrolysis and atomization temperatures an aqueous iron standard solution of 10
mg l 1 with Mg(NO3)2 and without Mg(NO3)2
were used. No important improvement in the iron
stabilization was observed. For this reason and to
avoid risk of contamination and to shorten the
analytical procedure we propose the elimination
of the use of the Mg(NO3)2 and thus the direct
introduction of the protein fractions become possible. The optimum pyrolisis and atomization
temperatures were 1350 and 2400C, respectively.
An intermediate pyrolisis step at 700C was necessary to follow a complete mineralization of the
sample. Finally a cleaning step at 2600C was
introduced to avoid possible memory effects. The
instrumental conditions and the furnace programme are summarized in Table 1.

3.6. Calibration
Solutions prepared in mobile phase with iron
standard concentrations between 0 and 30 mg l 1
were used to determine a calibration curve. The
equation obtained was:
A= 4.48 10 3 [Fe] 5.0 10 3

r =0.999

where A is the absorbance (peak area) and [Fe] is

the iron concentration expressed in mg l 1.

3.7. Sensiti6ity
The limit of detection (LOD) is defined as 3
S.D./m and the limit of quantification is given by

25
15
4
3.5
2

RSD (%)
Retention time

Peak height

Peak area

0.0
0.0
0.0
0.0
0.0

2.2
1.4
3.7
2.7
3.8

1.6
0.9
3.7
4.0
12.3

10 S.D./m (m= slope of the calibration graph and

S.D.= the within-run standard deviation of the
blank signals). The values based on ten replicates
of the blank were 1.4 and 4.7 mg l 1, respectively.
The characteristic mass (m0) defined as the mass
of analyte that provides an integral absorbance of
0.0044 for an aliquot sample of 20 ml was 20.4 pg.

3.8. Ironprotein complex stability

To know the stability of the Feprotein complexes a study about the time effects on the chromatographic protein separation was performed. A
milk whey sample (infant formula) was chromatographed at different times after preparation
1.02.03.04.0 and 24 h later. The results are
shown in Fig. 5(1)). It can be seen that only the
absorbance of the protein corresponding to 2 kDa
changes with the time and this absorbance was
constant after 24 h. For this reason, the performance of chromatographic separation 24 h after
the milk preparation was proposed.
On the other hand to know the effect of the
sample freezing, the same sample was chromatographed before and after freezing, both chromatograms are shown in Fig. 5(2). It can be seen
that the freezing did not affect to the protein
separation.
In the same way the iron determination in the
different protein fractions was performed before
and after freezing. The levels of iron obtained are
in Table 4 and it can be seen that there are no
significant differences due to the sample freezing.

P. Bermejo et al. / Talanta 50 (2000) 12111222

1218

3.9. Precision in the iron determination

ing both types of milk. The iron content of the

fractions obtained for a breast milk sample (six
replicates) and for infant formula milk (nine replicates) was determined using the proposed proce-

The within-run precision (RSD) of the method

(instrumental and matrix factors) was studied us-

Fig. 5. (1) Chromatograms of a sample after ultracentrifugation: (a) 1 h later; (b) 2 h; (c) 3 h; (d) 4 h; (e) 24 h. 2) Chromatograms
of a sample stabilized: (a) after ultracentrifugation; (b) after defreezing.
Table 4
Iron distribution in fractions of an infant formula sample before and after freezing
MW (kDa)

Before freezing

321
59
38
15
3
2

After freezing

[Fe] (ng fraction1)

[Fe] (ng fraction1)

74.790.6
110.39 1.6
207.09 0.0
162.49 3.5
46.190.8
19.690.1

12.0
17.8
33.4
26.2
7.4
3.2

86.5 9 1.8
106.7 95.4
211.2 91.6
150.4 9 0.0
39.6 90.8
23.4 9 0.1

14.0
17.3
34.2
24.3
6.4
3.8

Table 5
Precision in iron determination by electrothermal atomic absorption spectrometry (ETAAS)
Breast milk (n = 6)

Infant formula (n =9)

MW (kDa)

Fe (ng fraction1)

S.D.

RSD (%)

MW (kDa)

Fe (ng fraction1)

S.D.

RSD (%)

101
78
53
14
3
2

1.6
11.2
14.4
NDa
3.4
NDa

0.1
1.0
1.0

0.4

6.3
9.0
7.0

11.8

25
15
4
3.5
2

26.5
19.8
109.2
144.3
22.8

1.5
1.4
8.8
9.2
1.6

5.7
7.1
8.1
6.4
6.9

ND, not detected.

P. Bermejo et al. / Talanta 50 (2000) 12111222

Table 6
Mass balance study

[Fe]fractions
(ng 100 ml1)
[Fe]milk (ng 1000 ml1)
[Fe]whey (mg ml1)
[Fe]fractions
100 (%)
[Fe]whey

Before freezing
[Fe]
(ng fraction1)

After freezing
[Fe]
(ng fraction1)

619.9

617.6

1150.0 9 0.0
720.09 20.0
86.1

86.1

dure. The results obtained are shown in Table 5.

The values of RSD (%) obtained are acceptable in
all cases because this precision includes all the
analytical procedure: the chromatographic separation, the fraction collection and the iron
determination.

3.10. Accuracy for the Fe determination in total

milk and milk whey
The accuracy of the total iron concentration in
milk was performed using a Certified Reference
Material, non-fat milk A-11 of the IAEA with a
certified content of 3.65 90.76 mg Fe g 1. This
reference material was prepared at 15% (W/V)
and the Fe content was determined using the
addition procedure. The results obtained for five
replicates were 3.04 90.17 mg Fe g 1. On the
other hand the recovery of the method was studied measuring iron added to whole milk, due to
the fact that this sample has a more complex
matrix than milk whey. Different amounts of Fe
added to the milk sample were studied, the results
obtained were 100.0, 96.0 and 102.0% for 0.50,
1.00 and 1.5 mg ml 1 of Fe added.

3.11. Mass balance study

To check the accuracy in the iron determination
in the protein fraction a study about the mass
balance in a milk sample was performed. The
sample was studied after its collection and after
freezing and each sample was chromatographed
by duplicate. The iron levels for the total milk, the
milk whey and for different fractions are shown in

1219

Table 6. It can be seen that the iron mass balances

always approached to 100% (86.1%), and it can be
concluded that there is no contamination or loss
problems.

3.12. Applications
The proposed method has been applied to the
study of the iron distribution in the milk whey
proteins of ten infant formulas and ten breast
milk samples. The infant formula samples of
different brands were prepared at the concentrations suggested by the manufacturer, whereas
breast milk was individual samples from women
living in Galicia, North West of Spain with most
of them corresponding to the first stage of lactation.
The protein identification of the milk whey is
usually performed by comparison of the retention
times of standards and samples, but in case of an
incomplete separation metals can not be clearly
attributed to proteins [13], or may be attributed to
the wrong one. To avoid this problem the iron
found in a protein fraction was only associated
with the molecular weight of the protein obtained
in the calibration column.
The iron in the protein fractions was determined by ETAAS using the established conditions
at least twice. The iron concentrations in milk and
milk whey are reported in Table 7.
Previous researchers concluded that a more
careful speciation should be performed in the
range of low molecular weight compounds. First,
Fransson and Lonnerdal [11] demonstrated using
ultrafiltration (membrane molecular weight cutoff of 15 kDa) and SEC that a considerable
fraction of the iron was bound to LMW compounds in breast milk. Bratter and al. [5] found
iron in a fraction of HMW (\ 600 kDa) and in
the fraction of LMW (910 kDa) eluting together
with citrate. Suzuki et al. [6] studied daily changes
in components of breast milk with number of
lactation days. They found iron in one or two
peaks, whose elutions times coincided with the
transferrin peak and citrate peak time. However,
their chromatographic separation seems to be
worse. Negretti de Bratter [8] studied a wide range
of MWs protein obtaining a complicate chro-

1220

P. Bermejo et al. / Talanta 50 (2000) 12111222

Fig. 6. Iron distribution among protein fractions: (1) breast milk; (2) infant formula. Concentration is expressed as ng Fe fraction 1.

P. Bermejo et al. / Talanta 50 (2000) 12111222

1221

Table 7
Iron concentration in milk and milk whey
Breast milk
Sample

Infant formula
Milk

Milk whey

Sample

[Fe] (mg ml1)

1
2
3
4
5
6
7
8
9
10
a

0.289 0.01
0.239 0.00
0.609 0.01
0.339 0.00
0.199 0.00
0.239 0.00
0.229 0.00
a
0.289 0.01
0.249 0.00

Milk

Milk whey

[Fe] (mg ml1)

0.229 0.00
0.129 0.00
0.41 9 0.00
0.249 0.00
0.179 0.00
0.17 9 0.00
0.139 0.00
a
0.179 0.00
0.229 0.00

1
2
3
4
5
6
7
8
9
10

3.6 90.30
5.8 90.10
5.7 90.10
4.2 90.10
3.3 90.10
3.1 90.30
5.7 90.10
6.3 90.30
4.4 90.60
2.9 90.10

1.8 90.00
1.6 90.00
1.6 90.10
1.1 90.10
2.0 90.20
1.8 90.00
1.8 90.10
1.4 90.10
1.5 90.00
1.5 90.00

Results not available.

matographic pattern and concluded that columns

with a separation range B250 kDa should be
used for a detailed human milk speciation. The
chromatographic patterns are similar to that obtained by Coni et al.[7]. They found a homogeneity of iron distribution in protein fractions of
infant formulas, whereas there is an increase in
the amounts of elements in those intermediate
fractions of mature breast milk that are related to
substances with molecular weights ranging between 10 and 100 kDa. Underlying this distribution, an important role is played by the
differences in percentage of these proteins in the
composition of cow, cow-based formulas and human milk. Colostrum shows a different element
composition from that of mature milk.
It was found that the behaviour of the infant
formulas is very different and the iron distribution
in the proteins is not regular. This can be explained by the different chemical composition of
the formulas, which are prepared in different proportions of whey, proteins and minerals according
to the needs of the infants. For the studied breast
milks, the iron was bound principally to the
proteins corresponding to molecular weights of 3
and 76 kDa (Fig. 6), these results agree with the
results obtained by other authors.

4. Conclusions
A speciation method for iron in milk by
HPLC and ETAAS has been developed in order
to compare the iron distribution among low
molecular weight proteins in breast milk and
infant formulas. With the use of a simple
mobile phase a good protein separation was
obtained, and moreover its use does not present problems in the iron determination by
ETAAS.
The results obtained showed a different
iron distribution in the proteins of the milk
whey of infant formulas and human milk,
and this can be important in the iron bioavailability of both types of milk. For this reason
it is necessary to continue these studies in order to prepare new infant formulas where
the iron distribution is more similar to human
milk.

Acknowledgements
This work was partially supported by the research project 1997, CEO12 Laboratorios Ordesa,
Barcelona, Spain.

1222

P. Bermejo et al. / Talanta 50 (2000) 12111222

References
[1] J.D. Cook, B.S. Skikne, R.V. Baynes, Progress in Iron
Research, Advances in Experimental Medicine and Biology, Plenum Press, New York, 1994.
[2] R.G. Stevens, D.Y. Jones, M.S. Micorzi, P.R. Taylor, N.
Engl. J. Med. 319 (1988) 1047.
[3] T.J. Salonen, K. Nyyssonen, H. Korpela, J. Tnomilehto,
R. Seppanen, R. Salonen, Circulation 86 (1992)
803.
[4] E.E. Ziegler, S.J. Fomon, Nutr. Rev. 54 (1996) 348.
[5] P. Bratter, B. Gercken, U. Rosick, A. Tomiak, Analytical
Chemistry in Medicine and Biology, vol. 5, Walter de
Gruyter, Berlin, 1988, p. 145.
[6] K.T. Suzuki, H. Tamagawa, S. Hirano, E. Kobayashi, K.
Takahashi, N. Shimojo, Biol. Trace Elem. Res. 28 (1991)
109.
[7] E. Coni, A. Alimonti, A. Bocca, F. La Torre, D. Pizzuti,
S. Caroli, Element Speciation in Bioinorganic Chemistry,
Wiley, Chichester, 1966.

[8] V.E. Negretti de Bratter, S. Recknagel, D. Gawlik, Fresenius J. Anal. Chem. 353 (1995) 137.
[9] Y. Havezov, Fresenius J. Anal. Chem. 355 (1996) 452.
[10] A.K. Das, R. Chakraborty, Fresenius J. Anal. Chem. 357
(1997) 1.
[11] G.B. Fransson, B. Lonnerdal, J. Pediatr. 96 (1980) 380.
[12] B. Michalke, D.C. Munch, P.S. Schramel, J. Trace Elem.
Electrolytes Health Dis. 5 (1991) 251.
[13] B. Michalke, Fresenius J. Anal. Chem. 350 (1994) 2.
[14] P. Bermejo, R. Domnguez, A. Bermejo, Talanta 45
(1997) 325.
[15] M. Martin, F. Jacobs, J. Brushmiller, J. Nutr. 14 (1984)
869.
[16] A. Conti, L. Napolitano, J. Libratori, Milchwissenschaft
38 (1983) 392.
[17] M.E. Himmel, J.O. Baker, Size Exclusion Chromatography of Proteins, Handbook of Size Exclusion Chromatography, Marcel Dekker, New York, 1995.
[18] B.V. Lvov, Spectrochim. Acta 24B (1969) 53.
[19] F.M. Luquet, Leche y productos lacteos. Vaca. Oveja.
Cabra, Societe Scientifique dHygiene Alimentarie, Ed.
Acribia S.A., 1991.