Short communication

Quantication of lactose using ion-pair RP-HPLC during enzymatic lactose

hydrolysis of skim milk
Sarah Erich, Theresa Anzmann, Lutz Fischer

University of Hohenheim, Institute of Food Science and Biotechnology, Department of Biotechnology, Garbenstr. 25, 70599 Stuttgart, Germany
a r t i c l e i n f o
Article history:
Received 21 February 2012
Received in revised form 2 July 2012
Accepted 11 July 2012
Available online 20 July 2012
Keywords:
Enzymatic lactose hydrolysis
Ion-pair RP-HPLC
b-galactosidase
Lactose-free foods
a b s t r a c t
The correct labelling of dairy foods as lactose-free requires a suitably sensitive and valid analytical
method for the quantication of lactose in complex food matrices. Thus, an ion-pair RP-HPLC method
for the simultaneous determination of lactose, glucose and galactose in original skim milk was investi-
gated. The samples derived from an enzymatic lactose hydrolysis approach (0.5 L) using the commercial
b-galactosidase Godo-YNL2. After derivatisation with p-aminobenzoic acid and sodium cyanoborohy-
dride, the samples were injected on a RP-C
18
column. Tetrabutylammonium hydrogen sulphate was used
as the ion-pair reagent in the eluent system. The sugars were quantied using photometric- (UV; 303 nm)
and uorescence-detection (k
ex
313 nm, k
em
358 nm). The overall run time was 27 min. The limits of
detection (LOD) were estimated at 2 mg L
1
(UV detection) and at 0.13 mg L
1
(uorescence detection).
The limits of quantication were 6 mg L
1
(UV detection) and 0.45 mg L
1
(uorescence detection). Thus,
this analytical method is suitable for sensitive lactose quantication in milk systems of less than
10 mg L
1
.
2012 Elsevier Ltd. All rights reserved.
1. Introduction
Approximately 70% of the adult world population seems to be
lactose intolerant, causing gastrointestinal problems after the con-
sumption of foods containing a certain amount of lactose (Harju,
1987). Thus, the dairy industry has successfully commercialised
lactose-free or lactose-poor dairy products. The GDCh (Gesell-
schaft Deutscher Chemiker e.V.) recommends criteria for this type
of lactose declaration (Positionspapier GDCh, 2005). According to
these criteria, the lactose content of products with the most desired
label (lactose-free) must be lower than 100 mg lactose kg
1
/L
1
,
of the dairy product (0.01%). To monitor such a low lactose content
in dairy products, an accurate, robust and sensitive analytical meth-
od for lactose quantication needs to be available. Commonly used
methods are high performance liquid chromatography (HPLC) and
gas chromatography (GC), always coupled with a suitable detection
system (Cardelle-Cobas, Martnez-Villaluenga, Sanz, & Montilla,
2009; Ferreira, Gomes, & Ferreira, 1998). Other commonly used
sensitive methods for the quantication of saccharides include
high-performance anion-exchange chromatography (HPAEC) with
pulsed amperometric detection (PAD) and capillary electrophoresis
(CE) with electrochemical detection (Cataldi, Angelotti, & Bianco,
2003; Zhang, Cao, & Ye, 2001). Another method for the rapid
quantication of lactose in milk was described by Colinas, Barrera,
& Blanco, 2006. However, with this method, only the lactose con-
tent of the milk could be determined. In case of GC analysis, the sug-
ars have to be chemically converted into volatile and stable
derivatives (e.g., trimethylsilyl derivatives) prior to injection, which
is generally laborious and time consuming (Medeiros & Simoneit,
2007; Valero, Villamiel, Sanz, & Martnez-Castro, 2000).
Thus, the method of choice seems to be HPLC. If a refractive
index (RI) detector is used, the analysis is straightforward but not
very sensitive (Binder, 1980; Xinmin, Ruili, Zhihau, Yuanghong, &
Tingfu, 2008; Chvez-Servn, Castellote, & Lpez-Sabater, 2004).
Chvez-Servn et al. (2004) described a limit of detection of
250 mg L
1
and a limit of quantication of 380 mg L
1
for their
lactose analysis. Using an evaporative light scattering detector
(ELSD), the limit of detection was increased by factor of 10 (La Pera,
Di Bella, Magnesi, Lo Turco, & Dugo, 2007; Wei & Ding, 2000). With
this detection system, Lucena, Gallego, Crdenas, and Valcrcel
(2003) investigated milk samples for quality control and found a
limit of detection for lactose of 300 mg L
1
.
Another possible approach for sugar detection using HPLC anal-
ysis is the application of a photometric or uorometric detection
system. In these cases the carbohydrates must be tagged with
chromophores or uorophores before quantication (Anumula,
1994; Strydom, 1994).
For example, Meyer, Raba, and Fischer (2001) reported an inter-
esting method for the analysis of sugars in soil and landll samples,
using an ion-pair RP-HPLC with pre-column derivatisation, coupled
with an UV- or uorescence-detector. Various solvent mixtures
and gradient proles were tested to obtain a sufcient resolution.
0308-8146/$ - see front matter 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.foodchem.2012.07.059

Corresponding author. Tel.: +49 711 459 23018; fax: +49 711 459 24267.
E-mail addresses: s.gulan@uni-hohenheim.de (S. Erich), t.anzmann@
uni-hohenheim.de (T. Anzmann), lutz.scher@uni-hohenheim.de (L. Fischer).
Food Chemistry 135 (2012) 23932396
Contents lists available at SciVerse ScienceDirect
Food Chemistry
j our nal homepage: www. el sevi er . com/ l ocat e/ f oodchem
Fischer, Wacht, and Meyer (2003) advanced the method of Meyer
et al. and achieved the best results using p-aminobenzoic acid for
pre-column derivatisation and tetrabutylammonium hydrogen
sulphate (TBAHSO
4
) as the ion pair reagent. For lactose analysis,
a limit of detection of 0.03 mg L
1
was measured using the uores-
cence detection system.
The objective of our study was to analyse the lactose content in
foods such as milk at very lowquantities. Thus, an ion-pair RP-HPLC
method, which was coupled with UV- and uorescence-detection,
was chosen. The lactose hydrolysis in milk was performed with
the commercial enzyme preparation Godo-YNL2, containing the
b-galactosidase from Kluyveromyces lactis.
2. Materials and methods
2.1. Chemicals
All chemicals used were of analytical reagent grade. Glucose,
lactose, p-aminobenzoic acid, orthophosphoric acid (85%), sodium
cyanoborohydride, triethanolamine hydrochloride and magnesium
sulphate heptahydrate were obtained from Merck (Darmstadt,
Germany). Galactose, perchloric acid and tetrabutylammonium
hydrogen sulphate (TBAHSO
4
) were obtained from SigmaAldrich
(Steinheim, Germany). Dimethyl sulfoxide (DMSO) was purchased
from Fluka (Buchs, Switzerland), and potassium hydroxide and
acetic acid were purchased from Carl-Roth (Karlsruhe, Germany).
Water was puried by reverse osmosis and passed through a Mil-
lipore Mili-Q unit. For enzymatic hydrolysis, Godo-YNL2 (Shusei
Co, Ltd., Tokyo, Japan) containing the b-galactosidase from K. lactis
was used.
Nicotinamide adenine dinucleotide phosphate (NADP) and
adenosine triphosphate (ATP) were obtained from Gerbu (Heidel-
berg Wieblingen, Germany). The hexokinase/glucose-6-phos-
phate-dehydrogenase solution was purchased from Roche
Diagnostics (Mannheim, Germany).
2.2. Enzymatic lactose hydrolysis in milk
The hydrolysis of the milk (Schwarzwlder H-Weidemilch,
3.8% obtained from Breisgaumilch GmbH, Germany) was per-
formed at 8 C in a table bioreactor (Multifors, Infors GmbH) cou-
pled with a cooling circulator (Julabo F12, JULABO Labortechnik
GmbH, Seelbach, Germany) at an operating volume of 0.5 L. The
lactose hydrolysis was started by adding 500 lkat Godo-YNL2 (as-
say, see Section 2.3).
2.3. b-Galactosidase activity assay
The hydrolytic activity of Godo-YNL2 was determined with lac-
tose as a substrate at 8 C in 1.2 mL volume (contained 0.6 mL of
400 mM lactose solution, 0.5 mL of 100 mM potassium phosphate
buffer, pH6.75, with 5 mMMgCl
2
). The assay was started by adding
0.1 mL enzyme solution. After 90 s, a 0.1 mL sample of the reaction
mixture was added to 0.2 mL of perchloric acid (1 M) to stop the
enzymatic reaction. Then, the sample was neutralised with 2 M
potassium hydroxide. The release of glucose was quantied enzy-
matically using the hexokinase/glucose-6-phosphate-dehydroge-
nase assay from Roche Diagnostics (Mannheim, Germany) at 37 C.
One katal was dened as the amount of enzyme which cata-
lysed the release of 1 mol glucose from lactose per second.
2.4. HPLC analysis
The chromatographic analyses were carried out on a Thermo
Finnigan high-performance liquid chromatograph equipped with
a Surveyor LC gradient pump and a Surveyor auto sampler with a
20 lL loop (injection volume 10 lL). For separation of the sugars,
an RP-C
18
column HyPurity (150 4 mm, 3 lm) from Thermo Sci-
entic (Schwerte, Germany) was used. A Finnigan Surveyor UV/VIS
detector with a detection wavelength of 303 nm and a Finnigan
Surveyor uorescence detector (Thermo Finnigan LLC, San Jose,
CA USA) with the variable wavelengths set at k
ex
313 nm and k
em
358 nm.
2.5. Milk sample preparation
After 0, 1, 2, 3, 4, 7 and 24 h of enzymatic lactose hydrolysis,
samples (0.5 mL) were taken and the reaction was stopped by add-
ing 1 mL of perchloric acid (1 M). After centrifugation for 10 min at
4 C and 8000 rpm (Eppendorf centrifuge 5417R), the samples
were neutralised with 2 M potassium hydroxide (approximately
0.3 mL). Reductive amination of the sugars was performed with
0.35 M p-aminobenzoic acid solved in DMSO/acetic acid solution
(70/30; v/v). A 0.2 mL aliquot of this sample was derivatised with
0.5 mL p-aminobenzoic acid reagent and 10 mg of sodium cyano-
borohydride for 15 min at 60 C in a 2-mL Eppendorf vial. Then,
the samples were cooled on ice and dissolved in an aliquot of the
HPLC eluent with the ion-pair reagent but without methanol (see
Section 2.5). If needed, samples were further diluted. Finally, the
samples were ltered through a 0.45 lm PTFE lter. Stock solu-
tions of the standard saccharides lactose, glucose and galactose
were prepared by weighing 20 mg of each saccharide in a 10 mL
volumetric ask and dissolution in water. All stock solutions were
stored at 4 C.
2.6. Chromatographic conditions
Separation was performed at 25 C on a RP-C
18
column HyPurity
(150 4 mm, 3 lm). Sodium phosphate buffer and methanol so-
dium phosphate buffer, containing the ion-pair reagent TBAHSO
4
,
were used as mobile phases for gradient separation. The eluents
A and B containing 20 mM THAHSO
4
were adjusted to different
pH values (pH 6.5, 5, 3.5 and 2), and the chromatographic separa-
tion of the sugars was investigated. Eluent A was 20 mM TBAHSO
4
in sodium phosphate buffer (0.1 M, pH 6.5) with methanol (50:50
(v/v) adjusted to each pH value by adding orthophosphoric acid.
Eluent B was 20 mM TBAHSO
4
in sodium phosphate buffer
(0.05 M, pH 6.5) adjusted to each pH value by adding orthophos-
phoric acid. Two different gradients were investigated: the gradi-
ent described in the literature of Meyer et al. and a gradient
started with 100% eluent B; after 20 min to 75% B/25% A and from
20 to 27 min re-equilibration to 100% eluent B. The ow rate was
0.5 mL min
1
, the injection volume was 10 lL and the total run-
time was 60 min with the gradient of Meyer et al. and 27 min with
our gradient for each sample (including re-equilibration time).
Peaks were identied by comparing the retention times with sugar
standards. Three ve-point calibration curves were prepared for
each sugar at concentrations between 1 and 100 mg L
1
, all dis-
solved in water. Each calibration solution was measured in
triplicate.
2.7. Validation of the method
For validation of the analytical method the linearity ranges, the
recovery rates, the reproducibility, the limits of detection and lim-
its of quantication were determined. The linear correlations be-
tween the concentrations of the sugars and the signals of UV- or
uorescence-detection were found in the ranges of 1100 and
0.110 mg L
1
, respectively. Each solution was analysed in
triplicate.
2394 S. Erich et al. / Food Chemistry 135 (2012) 23932396
The recovery rates were ascertained in quadruplicate by adding
standards of lactose, glucose and galactose in known concentrations
(1 and 10 g L
1
) into an original milk sample before its derivatisa-
tion. Each milk sample was measured by HPLC analysis with UV-
and uorescence detection in triplicate. The lactose content of the
original milk sample was 41.5 g L
1
2.9%. The milk samples con-
tained no glucose and galactose. Recovery of lactose, glucose and
galactose was expressed as the peak area of the saccharides in the
milk sample related to the peak area of the original milk sample.
To examine whether the reproducibility of the method is suf-
cient, the contents of lactose, glucose and galactose in milk sam-
ples were determined six times independently using UV-
detection. Therefore, six different sample aliquots of a hydrolysed
milk sample were derivatised and measured in triplicate.
To check the sensitivity of the method, the limit of detection
(LOD) and the limit of quantication (LOQ) were determined by
comparison to the signal-to-noise ratio (S/N). The LOD was dened
as three times the S/N-Ratio, and the LOQ was dened as ten times
the S/N-Ratio.
3. Results and discussion
3.1. Sample preparation and optimisation of the HPLC conditions
The enzymatic reaction in the milk samples was stopped by
adding perchloric acid (see Section 2.5). Meyer et al. reported an
ion-pair RP-HPLC method for the determination of sugars in soil
and landll samples, which was taken as a basis for further devel-
opment. The derivatisation of the sugars in the milk samples was
performed with p-aminobenzoic acid-reagent and sodium cyano-
borohydride. Different reaction times (5, 15, 30, 45 and 60 min)
and reaction temperatures (40, 50, 60 and 70 C) for derivatisation
were tested. At a temperature of 60 C and a reaction time of
15 min the highest peak areas could be achieved using UV detec-
tion with 303 nm (data not shown).
Meyer et al. suggested the application of a linear elution gradi-
ent of methanol (eluent A)/aqueous solution (eluent B; containing
20 mmol L
1
TBAHSO
4
, pH 2) starting with 0% A and ending with
50% A at a ow rate of 0.8 mL min
1
and approximately 60 min
run time (including re-equilibration). In our studies, mobile phases
with different pH values and two different elution conditions were
investigated (see Section 2.6). The best results were achieved with
the gradient described in 2.6. The eluents A and B were adjusted to
a pH of 2. Only with this pH value and this elution condition was
the separation of glucose and galactose sufcient. The overall run-
ning time took 27 min compared to approximately 60 min before
(see above). In contrast to our work, Meyer et al. separated more
analytes; therefore, they needed different chromatographic condi-
tions and a longer run time. Fig. 1 reveals a typical HPLC-chromato-
gramof a milk sample taken froma lactose hydrolysis reaction, and
prepared as described in Section 2.5. The retention times of lactose,
glucose and galactose were 8.5 0.1, 9.8 0.1 and 10.5 min 0.1 -
min, respectively. The retention times of the sugars showed a high
precision.
3.2. Lactose hydrolysis in milk
The enzymatic lactose hydrolysis in milk was performed in 0.5 L
operational volume at 8 C for 24 h (Fig. 2) using a common 1L-bio-
reactor. The b-galactosidase from K. lactis was used as a biocatalyst
and is commercially available as GODO-YNL2. The total amount of
enzyme was 500 lkat, estimated with lactose as substrate at 8 C
in buffer, pH 6.75. This is an important note because the product
specication of the supplier revealed the b-galactosidase activity
of GODO-YNL2 using o-nitrophenyl galactoside (oNPG) as sub-
strate at 37 C, but this is not relevant for the application of lactose
hydrolysis. However, the correlation between GODO-YNL2 lactose
activity at 8 C and oNPG activity at 37 C was approximately 1/4.5.
At various time intervals, milk samples (0.5 mL) were taken and
reconditioned for HPLC analysis. The lactose content of the original
milk samples was 41.5 g L
1
2.9%. After 24 h, the lactose content
of the milk was clearly below 100 mg L
1
(24 mg L
1
2.0%). Thus,
this milk was considered to be truly labelled as lactose free.
3.3. Validation of the ion-pair HPLC method
Table 1 summarises the validation parameters for the devel-
oped method with the HyPurity RP-C
18
column. The linear ranges
of UV- or uorescence-detection were between 1100 and 0.1
10 mg L
1
, respectively. The r
2
values were >0.99.
As displayed in Table 1, with UV-detection, the recovery rates of
the saccharides in milk ranged between 96% and 112%. The recov-
ery rates with uorescence detection varied between 98% and
104%. Thus, uorescence detection was not only one order of mag-
nitude more sensitive but also more accurate.
The reproducibility of the method was analysed as described
above (Section 2.7). The results showed a relative standard devia-
tion of 1.2%.
Fig. 1. Analysis of lactose (2), glucose (3) and galactose (4) in a milk sample after
7 h of lactose hydrolysis (see Fig. 2) using an ion-pair RP-HPLC with a UV-Detector
(see Section 2.6). (1) p-aminobenzoic acid.
Fig. 2. Hydrolysis of lactose in milk with the commercial enzyme preparation
Godo-YNL2 at 8 C along 24 h.
S. Erich et al. / Food Chemistry 135 (2012) 23932396 2395
The limit of detection for lactose with the UV detector was
2 mg L
1
, and the limit of quantication was 6 mg L
1
. With the
uorescence detector, the LOD for lactose was 0.13 mg L
1
and
the LOQ 0.45 mg L
1
. These data obtained for milk samples showed
a sensitivity similar to that recently reported for soil samples and
landll samples (0.03 mg L
1
) (Meyer et al.; Fischer et al.).
4. Conclusion
The method presented here has been applied for the sensitive
determination of lactose, glucose and galactose in original milk.
It is suitable for the investigation of lactose-free labelled food.
It showed good linearity, precision and an adequately low limit
of quantication and limit of detection. Furthermore, it was shown
that the commercial enzyme preparation Godo-YNL2 could hydro-
lyse lactose after 24 h at 8 C down to concentrations less than
100 mg L
1
, generating a lactose-free milk product.
Acknowledgements
We express our thanks to the German Federal Ministry of Eco-
nomics and Technology (AIF/FEI Project No. 15801 N) for nancial
support of this research.
References
Anumula, K. R. (1994). Quantitative determination of monosaccharides in
glycoproteins by high-performance liquid chromatography with highly
sensitive uorescence detection. Analytical Biochemistry, 220, 275283.
Binder, H. (1980). Separation of monosaccharides by high-performance liquid
chromatography: Comparison of ultraviolet and refractive index detection.
Journal of Chromatography, 189, 414420.
Cardelle-Cobas, A., Martnez-Villaluenga, C., Sanz, M. L., & Montilla, A. (2009). Gas
chromatographic-mass spectrometric analysis of galactosyl derivatives
obtained by the action of two different b-galactosidases. Food Chemistry, 114,
10991105.
Cataldi, T. R. I., Angelotti, M., & Bianco, G. (2003). Determination of mono- and
disaccharides in milk and milk productes by high-performance anion-exchange
chromatography with pulsed amperometric detection. Analytica Chimica Acta,
485, 4349.
Chvez-Servn, J., Castellote, A., & Lpez-Sabater, C. (2004). Analysis of mono- and
disaccharides in milk based formulae by high-performance liquid
chromatography with refractive index detection. Journal of Chromatography A,
1043, 211215.
Colinas, C., Barrera, I., & Blanco, C. A. (2006). A novel correlation for rapid lactose
determination in milk by a cryoscopic technique. Journal of AOAC International,
89, 15811584.
Ferreira, I. M. P. L. V. O., Gomes, A. M. P., & Ferreira, M. A. (1998). Determination of
sugars, and some other compounds in infant formulae, follow-up milks and
human milk by HPLC-UV/RI. Carbohydrate Polymers, 37, 225229.
Fischer, K., Wacht, M., & Meyer, A. (2003). Simultaneous and sensitive HPLC
determination of mono- and disaccharides, uronic acids, and amino sugars after
derivatization by reductive amination. Acta Hydrochimica et Hydrobiologica,
31(2), 134144.
Harju, M. (1987). Lactose hydrolysis. Bulletin International Dairy Federation, 212,
5055.
La Pera, L., Di Bella, G., Magnesi, R., Lo Turco, V., & Dugo, G. (2007). Analysis of
carbohydrates in sicilian DOC wines by high performance liquid
chromatography with evaporative light scattering detection. Italian Journal of
Food Science, 19, 319328.
Lucena, R., Gallego, M., Crdenas, S., & Valcrcel, M. (2003). Autoanalyzer for milk
Quality Control based on the lactose, fat, and total protein contents. Analytical
Chemistry, 75, 14251429.
Medeiros, P. M., & Simoneit, B. R. T. (2007). Analysis of sugars in environmental
samples by gas chromatographymass spectrometry. Journal of Chromatography
A, 1141, 271278.
Meyer, A., Raba, C., & Fischer, K. (2001). Ion-Pair RP-HPLC determination of sugars,
amino sugars, and uronic acids after derivatization with p-aminobenzoic acid.
Analytical Chemistry, 73, 23772382.
Positionspapier der Lebensmittelchemischen Gesellschaft zu den Angaben
,,laktosefrei und laktosearm (2005). Lebensmittelchemie, 59, 45.
Strydom, D. (1994). Chromatographic separation of 1-phenyl-3-methyl-5-
pyrazolone-derivatized neutral, acidic and basic aldoses. Journal of
Chromatography A, 678, 1723.
Valero, E., Villamiel, M., Sanz, J., & Martnez-Castro, I. (2000). Chemical and sensorial
changes in milk pasteurised by microwave and conventional systems during
cold storage. Food Chemistry, 70, 7781.
Wei, Y., & Ding, M.-Y. (2000). Analysis of carbohydrates in drinks by high-
performance liquid chromatography with a dynamically modied amino
column and evaporative light scattering detection. Journal of Chromatography
A, 904, 113117.
Xinmin, W., Ruili, Z., Zhihua, L., Yuanhong, W., & Tingfu, J. (2008). Determination of
glucosamine and lactose in milk-based formulae by high-performance liquid
chromatography. Journal of Food Composition and Analysis, 21, 255258.
Zhang, X., Cao, Y., & Ye, J. (2001). Determination of lactose in sugar-free powder by
capillary electrophoresis with electrochemical detection. Food Chemistry, 72,
385388.
Table 1
Validation data (recovery rates, limit of detection and limit of quantication, see Section 2.7) of the ion-pair HPLC method coupled with UV- and uorescence-detector.
Compound (retention time: min) Detection method
a
Recovery rates (%) (n = 4) LOD
b
(mg L
1
) LOQ
c
(mg L
1
)
Lactose (8.5 0.1) UV 106112 2.00 6.00
Fluorescence 99104 0.13 0.45
Glucose (9.8 0.1) UV 98104 1.80 5.80
Fluorescence 98102 0.12 0.35
Galactose (10.5 0.1) UV 96101 1.60 5.40
Fluorescence 97102 0.12 0.44
a
Linearity range: UV 1100 mg L
1
; Fluorescence 0.110 mg L
1
.
b
Limit of detection.
c
Limit of quantication.
2396 S. Erich et al. / Food Chemistry 135 (2012) 23932396