ARTICLE IN PRESS

International Dairy Journal 17 (2007) 606616

www.elsevier.com/locate/idairyj

The molecular characterisation and antimicrobial activity of amidated

bovine lactoferrin
Y. Pana, B. Shiellc, J. Wanb, M.J. Coventryb, H. Roginskia, A. Leeb, W.P. Michalskic,
a

School of Agriculture and Food Systems, The University of Melbourne, Sneydes Rd., Werribee, Vic 3030, Australia
b
Microbiology and Biotechnology Group, Food Science Australia, Private Bag 16, Werribee, Vic 3030, Australia
c
Protein Biochemistry and Proteomics Group, Australian Animal Health Laboratory, CSIRO Livestock Industries, Private Bag 24,
East Geelong, Vic 3220, Australia
Received 19 January 2006; accepted 9 August 2006

Abstract
Chemical modication of bovine lactoferrin (LF) by amidation increased the net positive charge of the protein and markedly enhanced
its activity against most Gram-positive and Gram-negative bacteria tested, but not against Saccharomyces cerevisiae or Penicillium
candidum. Bovine lactoferrin was amidated with 1-ethyl-3-[3-(dimethylamino) propyl] carbodiimide in the presence of ammonium ions.
Mass spectrometric analysis of the modied protein conrmed that amidation converted aspartyl or glutamyl residues to asparaginyl or
glutaminyl residues. The changes in net charge and electrophoretic mobility of amidated LF were conrmed by electrophoretic and
chromatographic methods. Isoionic point titration was used to determine changes to net charges and ion exchange chromatography
results conrmed that the increased positive charge played an important role in the antimicrobial action of the amidated protein. Thus,
chemical amidation of LF introduced substantial structural changes in the LF molecule, reected in amino acid composition and
increased net positive charge, and enhanced biological activity.
r 2006 Elsevier Ltd. All rights reserved.
Keywords: Lactoferrin; Amidation; Antimicrobial protein

1. Introduction
Lactoferrin (LF), an iron-binding glycoprotein (approximately 80 kDa) of the transferrin family, is present mainly
in milk of some mammals (Lonnerdal, 2003) and other
external secretions such as tears, saliva and synovial uid
(Masson, Heremans, & Dive, 1966). Bovine LF (bLF) is a
single polypeptide chain protein (689 amino acids) and has
a broad antimicrobial spectrum against many Grampositive and Gram-negative bacteria. It was also reported
that charge modication of plasma and milk proteins
resulted in antiviral compounds (Swart et al., 1999). The
biochemical properties of LF appear to determine its
bacteriostatic and bactericidal activities through its powerful iron-binding capacity, net positive charge and the
strong interaction with prokaryotic or eukaryotic memCorresponding author. Tel.: +61 3 5227 5772; fax: +61 3 5227 5555.

E-mail address: wojtek.michalski@csiro.au (W.P. Michalski).

0958-6946/$ - see front matter r 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.idairyj.2006.08.005

branes. Positively-charged LF has been shown to bind to

the lipopolysaccharide (LPS) on the cell wall of Gramnegative bacteria, resulting in increased permeability of the
bacterial outer membrane (Ellison, Giehl, & Laforce,
1988). In Gram-positive bacteria, LF binds to the anionic
lipoteichoic acid (LTA), and this has been shown to allow
better accessibility of lysozyme to the peptidoglycan in the
bacterial membrane, again resulting in cell damage (Leitch
& Willcox, 1999).
The N-terminal domain of bLF, which carries a net
positive charge, has very high afnity for bacterial
membranes (Appelmelk et al., 1994; Elass-Rochard et al.,
1995). Lactoferricin B (LFcin B), a basic peptide (residues
1741) resulting from peptic digestion of the N-terminal
domain, has much greater antimicrobial activity than
native LF (Bellamy, Takase, Wakabayashi, Kawase, &
Tomita, 1992; Bellamy et al., 1992). Another positively
charged peptide (corresponding to residues 268284 of bLF
and designated as lactoferrampin) also has substantially

ARTICLE IN PRESS
Y. Pan et al. / International Dairy Journal 17 (2007) 606616

higher antimicrobial activity than LF (van der Kraan et al.,

2004). These observations indicate that increased net
positive charges appear to markedly enhance the antimicrobial activity of the protein.
The amidation of proteins has been shown to increase
their net positive charges and has been widely used in
studies of enzymic catalysis (Banks, Blossey, & Shafer,
1969; Johjima, Wariishi, & Tanaka, 2002) and for
producing or immobilizing biomaterials (Gratzer & Lee,
2001; Spagna, Barbagallo, Casarini, & Pifferi, 2001; Stile,
Barber, Castner, & Healy, 2002). Amidation also has been
used to chemically modify the milk protein b-lactoglobulin
to improve its physicochemical and functional properties
(Mattarella & Richardson, 1982; Mattarella & Richardson,
1983; Mattarella, Creamer, & Richardson, 1983). Several
studies have shown that bLF possesses antibacterial and
antifungal activities, and one of the mechanisms for these
activities appears to be due to the binding ability of LF to
cell membranes through its positive charges (Ellison et al.,
1988; Leitch & Willcox, 1999).
Pan et al. (2005) rst reported the enhancing effect of
amidation on bactericidal activity of bLF against the dairy
spoilage psychrotroph Pseudomonas fluorescens. In the
current study, bLF was chemically amidated to increase its
net positive charge and the resultant proteins molecular
properties and spectrum of antimicrobial activity were
quantitatively evaluated and compared with those of native
bLF.
2. Materials and methods
2.1. Chemical modification of lactoferrin
Bovine LF (Biopole, Les Isnes, Belgium), partially
saturated with iron (1520%) was chemically modied by
amidation. All reagents used for chemical modication and
analysis were of analytical or electrophoresis grade. Protein
concentrations were determined by the Bradford method
(Bradford, 1976), using bovine serum albumin (BSA) as
standard, and each sample was measured in duplicate.
Amidation reactions were based on the method of
Mattarella et al. (1983). Bovine LF (salmon pink in colour)
was dissolved in 5.5 mol L
1 NH4Cl (pH 4.75) to a nal
concentration
of
20 mg mL
1
(approximately

1
250 nmol mL ). Solid 1-ethyl-3-[3-(dimethylamino) propyl] carbodiimide (EDC; 273 mg) was slowly added to the
protein solution (10 mL, nal EDC concentration of
100 mmol mL
1) with continuous stirring and the pH of
the reaction mixture adjusted to 4.75 with 0.1 mol L
1 HCl
or 0.1 mol L
1 NaOH. Following 60 min incubation at
room temperature, the reaction mixture was dialysed
overnight against distiled water (pH 5.0, adjusted with
HCl). The nal sample (referred to as amiLF) was claried
by centrifugation at 8000  g for 5 min, passed through a
0.22 mm lter and kept at
80 1C. This sample was
colourless, indicating that the amidated protein did not
contain iron and may have lost its ability to bind iron due

607

to structural changes. In addition, LF was dissolved in

distiled water and mixed with EDC in the absence of
NH4Cl (incomplete amidation reaction), extensively dialysed against distiled water (pH 5.0) and the resultant
sample was referred to as EDC-LF.
2.2. Mass spectrometry
The native and amidated LF were characterised by liquid
chromatographymass spectrometry (LCMS) using the
method described by Shiell, Beddome, and Michalski
(2002). Protein samples (500 mg mL
1) were dissolved in
100 mmol L
1 NH4HCO3, pH 8.5, and digested with
trypsin (proteomics sequencing grade, Sigma Inc., St Louis,
MO, USA) in a ratio of 50:1 (w/w) at 37 1C for 20 h.
Reversed-phase chromatography (RPC) of digests was
performed on a ProteCol C18 column (3 mm,
0.3 mm  150 mm, SGE International Pty Ltd, Ringwood,
Victoria, Australia) using a Surveyor-MS HPLC system
(Thermo, San Jose, CA, USA). Mobile-phase buffers were
0.2% (v/v) formic acid in water (A) and 0.2% (v/v) formic
acid in 80% acetonitrile (B) and were run at a ow rate of
4 mL min
1 with a linear gradient of 5100% B over 20 min,
followed by holding at 100% B for 10 min and nally
returning to 5% B over 2 min. The efuent from the
column was connected directly to the nanospray ion source
of a LCQ Classic quadrupole ion-trap mass spectrometer
(Thermo). Mass spectral data were acquired in a datadependent mode known as the triple play, where the most
intense ion in each full scan (m/z 4002000) was
automatically selected and subjected to a high-resolution
zoom scan by a tandem mass spectrometry (MS/MS)
product ion scan. The zoom scan allowed determination of
the mass/charge state of the selected ion and hence an
accurate mass measurement of the selected ion. The MS/
MS product ion scan provided detailed conrmatory
information on the amino acid sequence of the selected
ion. The MS/MS scans were performed with normalised
collision energy of 35%. Nanospray ion source settings
were as following: nanospray tip voltage 1.8 kV, capillary
temperature 150 1C, capillary voltage 24 V and tube lens
offset 5 V. Raw mass spectral data were analysed using the
QualBrowser tool from Xcalibur Core Data System software version 1.3 (Thermo) to compile a list of monoisotopic peptide masses for each trypsin-digested protein.
The peptide masses obtained were compared with theoretical peptide masses expected from tryptic cleavage of the
target proteins, which were determined using the PeptideMass program from ExPASy tools (http://kr.expasy.org/
tools/peptide-mass.html) (Wilkins et al., 1997). To further
verify the identity of amidated peptides, in particular, the
MS/MS data obtained were compared with the theoretical
MS/MS pattern derived using the MS-Product program
(http://prospector.ucsf.edu/ucsfhtml4.0/msprod.html) from
the Protein Prospector site of MS analysis tools (http://
prospector.ucsf.edu/). The SwissProt primary accession
number for 708 amino acid bovine lactoferrin precursor

ARTICLE IN PRESS
608

Y. Pan et al. / International Dairy Journal 17 (2007) 606616

(TRFL_BOVIN) is P24627 (http://au.expasy.org/uniprot/

P24627). The sequence used in the paper was that of the
mature protein after removal of the signal peptide residues
119; hence, in the paper the protein sequence is shown as
from residues 1689.
2.3. Electrophoretic analysis
For denaturing polyacrylamide gel electrophoresis in the
presence of ionic detergent (referred to as SDS-PAGE), all
samples were run on NuPAGE (Invitrogen, Carlsbad, CA,
USA) 412% Bis-Tris gels used with 3-(N-morpholino)
propane sulphonic acid (MOPS) running buffer at pH 7.7.
Protein samples (1 mg mL
1) were mixed with NuPAGE
LDS sample buffer and dithiothreitol (50 mM nal concentration), heated at 90 1C for 5 min, and centrifuged at
6000  g for 5 s prior to loading on gels. Following
electrophoresis, gels were stained with 0.25% (w/v)
Coomassie brilliant blue R250 in 40% (v/v) methanol
and 10% (v/v) acetic acid, and destained in 15% (v/v)
methanol and 5% (v/v) acetic acid. In some cases, proteins
were visualised in gels by staining with silver nitrate
according to the procedure of Heukeshoven and Dernick
(1985).
Polyacrylamide gel electrophoresis in the presence of
acetic acid and urea (AU-PAGE) was performed as
described previously (Waterborg, 1996). Protein samples
for AU-PAGE analysis were desalted using Vivaspin 6
(Sartorius AG, Goettingen, Germany) centrifugal devices
with an 8 kDa molecular mass cut-off membrane.
In capillary zone electrophoresis (CE), the samples and
standards were separated on a P/ACE system 5510
(Beckman, Fullerton, CA, USA) equipped with proprietary
software Version 1.2 and a 27 cm  25 mm i.d. uncoated
CElect fused-silica capillary (Supelco, Bellefonte, PA,
USA). The temperature of the capillary was 35 1C, the
voltage and current across the capillary were maintained at
20 kV and 20 mA, respectively, and proteins were detected
at 214 nm. Protein samples were dissolved in freshly
prepared sample buffer (consisting of 5 mmol L
1 trisodium citrate dihydrate, 10 mmol L
1 DTT and 6 mol L
1
urea, pH 8.0). Samples were analysed in running buffer
(pH 3.0) comprising of 20 mmol L
1 trisodium citrate
dihydrate, 0.05% (w/v) methyl hydroxyethyl cellulose
(MHEC, 400600 cps) and 6 mol L
1 urea. An internal
standard glycyl-L-tyrosine (10 mg mL
1) was used in all
analyses.
Isoelectric focusing (IEF) Ready gels (pH 3-10) were
used to compare the pI values of native and amidated LF
according to the manufacturers (Bio-Rad, Hercules, CA,
USA) protocol. The separation was performed at constant
voltage: 100 V for 1 h, 250 V for 1 h and 500 V for 30 min.
Separated proteins were visualised with IEF staining
solution (27%, v/v, isopropanol, 10%, v/v, acetic acid,
0.04%, w/v, Coomassie Brilliant blue R250 and 0.05%, w/
v Crocein scarlet), and destained in a solution comprising
40% (v/v) methanol and 10% (v/v) acetic acid.

Native and amidated proteins were fractionated by

preparative IEF in a Rotofor cell (Bio-Rad) according to
the manufacturers protocol. Protein samples (20 mg) were
dissolved in 50 mL of sample buffer containing 1% (v/v)
glycerol, 1% (v/v) pH 310 ampholyte Bio-Lyte and 0.5%
(w/v) n-octyl glucoside (both Bio-Rad). The separation was
conducted at a constant power of 12 W at 10 1C. Twenty
fractions, containing proteins separated at different pI
values, were collected and analysed by SDS-PAGE.
Isoionic point determinations of proteins in solution
were performed by pH titration in the presence of an ionexchange resin (AG501-X8D, Bio-Rad) as described by
Mattarella et al. (1983).
2.4. Chromatographic analysis
Properties of native and amidated LF were analysed by
various chromatographic methods using HPLC-grade
solvents or freshly made buffers that were always ltered
through 0.22 mm lter prior to use.
Ion-exchange chromatography (IEC) of native and
modied LF was performed on a cation-exchange column
(Mono S HR 5/5 column, 50 mm  5 mm i.d., Pharmacia,
Uppsala, Sweden) connected to a Hewlett Packard 1050
Series HPLC system. The elution buffers consisted of:
50 mmol L
1 sodium acetate buffer (pH 5.0) (buffer A) and
2 mol L
1 sodium chloride in 50 mmol L
1 sodium acetate
buffer (pH 5.0) (buffer B). Samples (450 mL) were
fractionated at a ow rate of 1 mL min
1, and eluted
proteins were detected at 254 and 280 nm. Following the
sample injection the column was equilibrated with 100%
buffer A for 5 min, followed by elution with a linear
gradient from 0 to 100% of buffer B for 20 min and with
100% buffer B for 5 min (total run time 30 min). After each
run, the column was washed with a gradient from 100% to
0% buffer B over 5 min. Eleven injections of amidated LF
at 15 mg mL
1 were applied and the chromatogram for
each injection was compared. Following injections, 10
fractions containing amidated LF were collected every
minute, corresponding to retention times between 14 and
23 min; fractions with the same retention time were pooled
from each run. The protein concentration of each fraction
and pool was determined using the Bradford method. The
pooled fractions were also evaluated with the bactericidal
assay against Pseudomonas fluorescens (ATCC 31732),
described below.
LF and modied LF (5 mg in 500 mL injection volume)
were fractionated by size-exclusion chromatography (SEC)
on Superdex 200HR 10/30 column (Pharmacia) connected
to a HRLC 800 series system (Bio-Rad). The column was
calibrated with a mixture of proteins with known
molecular mass (molecular mass standard kit 12 to
200 kDa range, Sigma-Aldrich Inc, St Louis, MO, USA).
The column was equilibrated and protein samples were
eluted with 0.05 mol L
1 sodium phosphate buffer (pH 6.0)
containing 0.15 mol L
1 NaCl at a ow rate of
0.5 mL min
1, with protein detection at 280 nm. Selected

ARTICLE IN PRESS
Y. Pan et al. / International Dairy Journal 17 (2007) 606616

fractions were manually collected by hand from the start to

end point of each peak. All fractions were analysed by
SDS-PAGE.

609

cultural Trust Rothamsted Experimental Station, Harpenden, Hertfordshire, UK) and least-square differences
(LSD) of means were determined at the 5% signicance
level.

2.5. Microorganisms and media

3. Results
Cultures were obtained from the Food Science Australia,
Werribee (FSAW) culture collection. Growth conditions
were: Ps. fluorescens (ATCC 948, ATCC 31732, ATCC
49838, 3.0%, w/v, tryptone soy broth and 0.6%, w/v, yeast
extract, pH 7.0, TSYE, 16 h at 30 1C), Ps. fragi (FSAW
3121, TSYE, 16 h at 30 1C), Bacillus subtilis (ATCC 6633,
brain heart Infusion, BHI, 16 h at 30 1C), Listeria monocytogenes (NTCC 11994, TSYE, 24 h at 30 1C), Salmonella
Typhimurium (ATCC 14028, TSYE, 24 h at 37 1C),
Escherichia coli (ATCC 11775, nutrient broth, NB, 16 h
at 37 1C), Enterococcus faecalis (ATCC 51299, BHI or NB,
16 h at 37 1C), Saccharomyces cerevisiae (FSAW 3304,
TSYE, 48 h at 25 1C) and Penicillium candidum (VB 10D,
LACTOLABOs, Texel, Dange St Romain, France, malt
extract, ME, 48 h at 2022 1C). All microbiological media
and media supplements were obtained from Oxoid
Australia Ltd. (Heidelberg, Victoria, Australia) with the
exception of the ME, obtained from Amyl Media Pty. Ltd.
(Dandenong, Victoria, Australia).
2.6. Bactericidal assay against resting cells of bacteria
Cultures were grown for 16 h at their optimum growth
temperatures on the media listed above. Cell suspensions
were diluted to 106 cfu mL
1 with 0.1% (w/v) peptone
(1 mL) and centrifuged at 12,000  g for 10 min. Cells were
resuspended in 20 mmol L
1 phosphate (Na2HPO4 and
KH2PO4) buffer (pH 7.0) and either native or amidated LF
was added to make up 1 mL volumes, and an equal volume
of sterile distiled water was used in control samples. The
samples were incubated for 60 min at 30 1C and centrifuged
at 12,000  g for 10 min. The cells pellets were resuspended
in phosphate buffer (1 mL), serially diluted in 0.1% (w/v)
peptone and plated (1 mL pour plates or spirally using a
Whitley Automatic Spiral Platter, Don Whitley Scientic
Limited, Shipley, West Yorkshire, England) on TSYE. The
viable count (cfu mL
1) was enumerated after the plates
were incubated for periods up to 48 h at the appropriate
temperatures indicated above. The effect of different
concentrations of LF and amidated LF (0.11.0 mg mL
1)
were tested for bactericidal activity against Ps. fluorescens
(ATCC 31732) and B. subtilis. To evaluate the possible
contribution to any observed antimicrobial activity, the
effect of the amidation catalyst, EDC (0.55 mmol L
1) on
Ps. fluorescens (ATCC 31732) was evaluated. EDC-LF
(at 1 mg mL
1) was also tested against Ps. fluorescens.
2.7. Statistical analysis
Analysis of variance on microbiological data was
performed using Genstat version 8.0 (2005, Lawes Agri-

3.1. Mass spectrometric analysis of native and amidated LF

The amidation reaction was performed with an excess of
reagents and the efcacy of chemical amidation of bLF was
determined by mass spectrometric analysis; tryptic digests
of native and amidated LF were compared by LCMS.
Total ion chromatograms of the tryptic digests of amidated
LF and native LF are shown in Fig. 1 and the list of
detected peptides is shown in Table 1. Amidated LF had
peptide MS spectra different from those of the native LF,
indicating chemical modication of LF (Fig. 1). Twentythree peptides were identied in the trypsin-digested LF,
with 31% coverage of the whole sequence. Twenty-seven
peptides were identied in digests of amidated LF with
39% sequence coverage, among which 17 fragments
contained one to three amino acids that were converted
from aspartic acid (D) or glutamic acid (E) to the respective
amidated amino acid. The amidation appeared incomplete,
as peptides derived from amidated LF contained both
unmodied aspartic acid (D) and glutamic acid (E) and
modied asparagine (N) and glutamine (Q). Fragments
containing residues 5469 and 302309 encompassed both
unmodied (D) and modied (N) residues (Table 1).
3.2. Electrophoretic mobility of amidated LF
Native and amidated LF were analysed by SDS-PAGE
and resolved proteins were stained with both Coomassie
blue and silver nitrate. Amidated LF (Fig. 2a and b, lane 3)
demonstrated marked differences in electrophoretic mobility and migrated faster than native LF (Fig. 2a and b, lane
2). Several protein bands with mass4160 kDa were also
observed in the amidated LF sample and the band for
amidated LF (at a position of 70 kDa) appeared broader
than that of native LF, suggesting that LF was not
homogenously modied. These patterns were consistent
with modied proteins being most likely polymerised
(cross-linked) or aggregated and this was also supported
by results of size exclusion chromatography, where
amidated LF separated in a broad range of molecular
masses from 80 to 4200 kDa (data not shown). In
addition, a mixture of native and amidated LF (lane 4) was
separated into the two forms by comparison with native
LF (lane 2) and amidated LF (lane 3), suggesting that
amidation generated entirely different molecules from the
native protein. In addition, two closely migrating protein
bands were observed for native LF (lane 2), indicating
that native LF contained two forms of protein with
similar molecular masses possibly differing in degree of
glycosylation.

ARTICLE IN PRESS
Y. Pan et al. / International Dairy Journal 17 (2007) 606616

610

amidated LF were analysed by AU-PAGE under nondenaturing conditions designed for basic proteins; it was
thereby conrmed that amidated LF was composed of
chemically modied monomers and high molecular mass
oligomers or aggregates (data not shown).

RT:
21.50

100
90

18.39
16.03

80

3.3. Isoelectric properties of amidated LF

Relative Abundance (%)

70
60
50
40
15.09

30
20
14.13
12.13

10
0

3.20

22.72
34.14
23.12
24.15 28.25 33.54 35.14
38.32

6.08 6.77 11.77

10

15

(a)

20

25

30

35

39.05 45.67

40

45

Retention time (min)

RT:
100

3.4. Chromatographic characterisation of amidated LF

17.51

Amidated LF was eluted from the cation-exchange

column more slowly than native LF (Fig. 3a), indicating
that the modied LF carried more positive charges than its
native form. Moreover, the peak of amidated LF was much
wider than and overlapped with that of native LF,
suggesting that amidated LF may still contain amounts
of unmodied protein, as well as modied protein with
different degrees of amidation. Ten fractions were collected
from IEC runs at the retention times from 14 to 23 min,
and tested for antimicrobial activity (Fig. 3b). The
antimicrobial activity peak was associated with more
positively charged fractions: fraction 4 contained the
highest concentration of protein, but fraction 5 and 6,
which contained much less protein, had the highest activity
against Ps. fluorescens. Fractions 5 and 6 contained more
positively charged proteins that contributed to a longer
retention time.

90
80
70

18.35

21.50

60

Relative Abundance (%)

50
15.07

40
30

14.80
14.14

20

22.45

13.23

10
0

23.49
6.89 7.36 10.30
4.23 4.94

(b)

It was anticipated that amidation of LF would change its

isoelectric point (pI) as a reection of change in the
proteins net charge. Indeed amidated LF did not focus on
IEF gel electrophoresis with an effective separation range
of pI 38 and precipitated at pI 10.3 in liquid-phase
preparative IEF (Rotofor, Bio-Rad) (effective range of
pI 310) (data not shown). In contrast, native LF was
focused at pI 9 in the Rotofor separations, indicating its
true pI value. The isoionic points of native and amidated
LF were determined to be 8.7 and 9.7, respectively
(Table 2). The isoionic point of a reference protein, BSA,
was determined as 4.8, which is within its reported pI value
range of 4.74.9 (Malamud & Drysdale, 1978). This study
conrmed that amidation resulted in a signicant change in
the net charge of LF.

10

24.97

15

20

25

27.25

34.79

30

35

38.08
39.72 42.65 46.13

40

3.5. Antimicrobial activity of amidated LF

45

Retention time (min)

Fig. 1. Mass spectrometric analysis of lactoferrins: total ion chromatograms of the tryptic digests of (a) amidated and (b) native lactoferrins.

The effect of addition of positive charges to the protein

on its electrophoretic mobility was also veried by capillary
zone electrophoresis (CE), on which amidated LF (retention time 5.4 min) eluted faster than native LF (retention
time 5.9 min, data not shown). Moreover, native and

Amidated LF showed different inhibitory activities

against Gram-positive and Gram-negative bacteria
(Fig. 4). An increased activity (46 log viable count
reduction) against B. subtilis, L. monocytogenes,
Ps. fluorescens (3 strains) and Ps. fragi was observed as a
result of amidation, compared to native LF and the
control, containing no protein. Under the same test
conditions, a moderate inhibition was observed against
Salm. Typhimurium (3-log reduction), Ent. faecalis
(1-log reduction) and E. coli (2-log reduction). Amidated

ARTICLE IN PRESS
Y. Pan et al. / International Dairy Journal 17 (2007) 606616

611

Table 1
Peptides identied in tryptic digests of native and amidated lactoferrins by LCMS analysisa
Native LF

Amidated LF

Peptides detected

Amino acid sequence

Peptides detected

Amino acid sequence

f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f

WQWR
AIAEK
DPYK
LRPVAAEIYGTK
ESPQTHYYAVAVVK
KGSNFQLDQLQGR
GSNFQLDQLQGR
EPYFGYSGAFK
ETTVFENLPEK
DQYELLCLNNSR
APVDAFK
APVDAFKECHLAQVPSHAVVAR
SFQLFGSPPGQR
DLLFK
DSALGFLR
VDSALYLGSR
YLTTLK
ETAEEVK
HSSLDCVLRPTEGYLAVAVVK
CLAEDVGDVAFVK
EDFR
QVLLHQQALFGK
NGKNCPDK

f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f
f

RWQWR
WQWR
AD(or N)AVTLDGGMVFEAGR*
LRPVAAEIYGTK
QSPQTHYYAVAVVK
QSPQTHYYAVAVVKK
GSNFQLNQLQGR
QAYPNLCQLCK
ETTVFQNLPQK
APVNAFK
APVNAFKQCHLAQVPSHAVVAR
SVNGKQNLIWK
SFQLFGSPPGQR
DLLFK
D(or N)SALGFLR*
VNSALYLGSR
YLTTLK
QTAQQVKAR
GQADALNLDGGYIYTAGK
CVPNSKQK
YYGYTGAFR
CLAENVGDVAFVK
LLCLNGTR
QVLLHQQALFGK
NCPNKFCLFK
LGGRPTYQQYLGTEYVTAIANLK
CSTSPLLQACAFLTR

(2225)
(4852)
(7073)
(7485)
(8699)
(100112)
(101112)
(187197)
(211221)
(225236)
(237243)
(237258)
(285296)
(297301)
(302309)
(314323)
(324329)
(333339)
(420440)
(532544)
(567570)
(609620)
(621628)

(2125)
(2225)
(5469)
(7485)
(8699)
(86100)
(101112)
(164174)
(211221)
(237243)
(237258)
(259269)
(285296)
(297301)
(302309)
(314323)
(324329)
(333341)
(387404)
(515522)
(523531)
(532544)
(571578)
(609620)
(624633)
(651673)
(675689)

a
Listed peptides represented 30.9% and 39.4% of sequence for native and amidated lactoferrin, respectively. Amidated amino acid residues: Q
(glutamine) and N (asparagine) are indicated in peptides from amidated protein. Peptides found in the tryptic digest of amidated lactoferrin, in amidated
and unamidated form, are indicated with asterisk.

LF had little or no effect on growth of S. cerevisiae and P.

candidum (o1-log reduction).
The effect of the concentration of amidated LF on its
bactericidal activity against Ps. fluorescens and B. subtilis
was also investigated. Native LF at 1.0 mg mL
1 did not
show antibacterial activity against Ps. fluorescens compared with the control (Fig. 5a). In contrast, incubation
with amidated LF at the same concentration for 60 min
reduced the bacterial viable count by 6-log cycles. At a
10-fold lower concentration (0.1 mg mL
1), the amidated
LF still reduced viable count by 3 logs compared with the
control and native LF (1 mg mL
1). Amidated LF at
1.0 mg mL
1 resulted in a 7-log reduction in the viable
count of B. subtilis, whereas native LF (1 mg mL
1)
reduced the viable count by 2-log units when compared
with the control (Fig. 5b). At concentrations of 0.5 and
0.1 mg mL
1, amidated LF resulted in a 6 and 5 log
reduction in viable count of B. subtilis, respectively,
notably greater than that of native LF.
To conrm that the antimicrobial activity of amidated
LF was not associated with the use of EDC, concentrations
of EDC ranging from 0.5 to 5.0 mmol L
1 were tested
against resting cells of Ps. fluorescens. At the concentration
used in the amidation reaction (2.5 mmol L
1), the addition

of EDC resulted in only one log reduction in viable count

compared to the control, containing only phosphate buffer
(data not shown). The nal concentration of water-soluble
EDC in amidated LF preparations, following dialysis most
likely would be negligible (Gratzer & Lee 2001, Spagna
et al., 2001) and not contribute to any antimicrobial
activity of the preparation. EDC functions as a crosslinking reagent in the amidation reaction and can generate
random polymer molecules. However, there was no
difference in antimicrobial activity against Ps. fluorescens
between the polymerised LF in the EDC-LF sample and
native LF (data not shown). These results suggest that the
increase of antimicrobial activity of amidated LF was most
likely attributed to its charge and possibly to conformational change due to amidation.
4. Discussion
Chemical amidation of LF as a means of enhancing
bactericidal activity against Ps. fluorescens was rst
reported by Pan et al. (2005). In the current study, bLF
was amidated to increase its net positive charge and the
results demonstrated that amidated LF had markedly
enhanced antimicrobial activity against a wide range of

ARTICLE IN PRESS
Y. Pan et al. / International Dairy Journal 17 (2007) 606616

612

Fig. 2. SDS-PAGE analysis of native and amidated lactoferrin (LF). (a) Gel stained with Coomassie Brilliant blue: lane 1, molecular mass markers; lane 2,
native LF (0.5 mg); lane 3, amidated LF (2.5 mg); lane 4, native LF (0.25 mg)+amidated LF (1.25 mg). (b) Gel stained with silver nitrate: lane 1, molecular
mass markers (diluted 10-fold); lane 2, native LF (0.05 mg); lane 3, amidated LF (0.25 mg); lane 4, native LF (0.025 mg)+amidated LF (0.125 mg). ( v )
indicates native LF and (b) indicates amidated LF.

Table 2
Determination of isoionic points of native and amidated lactoferrins, and bovine serum albumin, and comparison with theoretical pI values
Protein

LF
Amidated LF
BSA

Isoelectric point
(measured)
8.09.0b

4.9 or 4.7c

Isoelectric
pointa(theoretical)

8.67
9.93

Isoionic point
(measured)

5.0c

Isoionic point (determined in this study)

Assay 1

Assay 2

Average

8.67
9.40
4.90

8.66
9.49
4.70

8.67
9.45
4.80

The theoretical values were obtained from http://kr.expasy.org/tools/pi_tool.html.

Malamud & Drysdale (1978).
c
Shimazaki (2000).
b

Gram-positive and Gram-negative bacteria. Like LF, the

positively charged amidated LF would be expected to bind
to the negatively charged LPS and or LTA on the bacterial
membrane, resulting in an alteration of membrane permeability and leading to membrane damage. Some bacterial
strains, however, seem to have stronger resistance to the
amidated LF, possibly due to the differences in membrane
structure and length and complexity of the hydrophilic
polysaccharide in its outer layer (Papo & Shai, 2003).
Based on the same explanation, LFcin B had very low
activity against Ps. fluorescens, Ent. faecalis and Bifido-

bacterium bifidum strains (Bellamy, Takase, Wakabayashi,

et al., 1992).
The molecular characterisation of amidated LF was
conducted using several techniques. MS proved to be a
powerful method for direct determination of the chemical
changes of modied proteins, and indeed the MS analysis
of native and amidated LF clearly demonstrated that
amidation converted aspartyl (D) or glutamyl (E) residues
to asparaginyl (N) or glutaminyl (Q) residues, respectively,
also resulting in the noticeable changes in molecular mass.
In addition, the modication did not occur uniformly in

ARTICLE IN PRESS
Y. Pan et al. / International Dairy Journal 17 (2007) 606616

613

Fraction number
1 2 3 4 5 6 7 8 9 10

3900
3400

100%

Buffer B

Native LF
2900
Absorbance (mAU)

0%
2400
1900
1400

Amidated LF

900
400
-100
0

10

15

20

25

30

Retention time (min)

10.0

10.0

9.0

9.0

8.0

8.0

7.0

7.0

6.0

6.0

5.0

5.0

4.0

4.0

3.0

3.0

2.0

2.0

1.0

1.0

0.0

0.0
1

(b)

Protein concentration (mg mL-1)

Log (cfu mL-1)

Viable count reduction of Ps. fluorescens

(a)

5
6
Fraction Number

10

Fig. 3. (a) Separation of native and amidated lactoferrin by ion-exchange chromatography; the light line stands for native lactoferrin and dark line for
amidated lactoferrin (offset by +100 mAU to avoid overlap). Ten fractions (numbered 110) were collected during separation of amidated lactoferrin
(every minute from retention time 1423 min) for testing of antimicrobial activity against Ps. fluorescens (ATCC 31732). (b) Antimicrobial activity of
amidated LF fractions; bars represent reduction in viable count log (cfu mL
1, LSD 0.6 log units at po0.05 for three replicates), (E) protein
concentration (mg mL
1) and (m) absorbance at 280 nm recorded from chromatogram. The error bars are the standard deviation based on three replicates
in each test.

every molecule and in every possible amino acid residue,

with some peptides existing in amidated and non-amidated
forms. MS analysis yielded limited sequence coverage,
approximately 30% and 39% of native and amidated LF
sequences, most likely due to the presence of glycan
moieties of LF, which may have interfered with both
tryptic digestion and peptide identication (Berkel, Geerts,
Veen, Kooiman, & Piepper, 1995; Shimazaki, 2000). It is
also possible that the protein did not unfold completely
upon modication due to structure stabilising effect of
glycan moieties (Moore, Anderson, Groom, Haridas, &
Baker, 1997). There are ve potential N-glycosylation sites
in bLF sequence and four glycan chains have been

chemically identied as actually attached (Spik et al.,

1994).
Amidation of LF led to obvious changes in the net
charge of the protein and visibly affected its mobility on
SDS-PAGE. This phenomenon has been reported for
extremely basic (positively charged) proteins such as
histones, where the charges introduced by SDS molecules
were balanced to a signicant degree by the proteins own
charges and resulted in anomalous mobility (Laas, 1989;
Panyim, Chalkley, Spiker, & Oliver, 1970). These observations plausibly explain the accelerated mobility of amidated
LF on SDS-PAGE gels. According to theoretical calculation, amidation would not signicantly change molecular

ARTICLE IN PRESS
Y. Pan et al. / International Dairy Journal 17 (2007) 606616

614

Viable count Log (CFU mL-1)

8.0
7.0
6.0
5.0
4.0
3.0
2.0
1.0
S. cerevisiae

P. candidum

Ent. faecalis

E. coli

L. monocytogenes

Salm.Typhimurium

B. subtilis

Ps. fragi

Ps. fluorescens (ATCC49838)

Ps. fluorescens (ATCC31732)

Ps. fluorescens (ATCC948)

0.0

Fig. 4. Viable count of resting cells of bacteria incubated in 20 mM phosphate buffer for 60 min at 30 1C in the absence ( ), and presence of native
lactoferrin (1 mg mL
1, ) or amidated lactoferrin (1 mg mL
1, ) (LSD 0.6 log units at po0.05 for three replicates). The error bars are the standard
deviation based on three replicates in each test.

7.0
6.0
Log (cfu mL-1)

Viable count of Ps. fluorescens

8.0

5.0
4.0
3.0
2.0
1.0
0.0
Control

(a)

LF (1.0)

Ami LF (0.1)

Ami LF (0.5)

Ami LF (1.0)

-1

Concentration of LF and Ami LF (mg mL )

8.0

6.0
Log (cfu mL -1)

Viable count of B. subtilis

7.0

5.0
4.0
3.0
2.0
1.0
0.0
Control

(b)

LF (1.0)

Ami LF (0.1)

Ami LF (0.5)

Ami LF (1.0)

Concentration of LF and Ami LF (mg mL -1)

Fig. 5. Effect of different concentrations of amidated lactoferrin on resting cells of (a) Ps. fluorescens (ATCC 31732), (LSD 0.7 log units at po0.05 for
three replicates) and (b) B. subtilis (ATCC 6633), (LSD 0.6 units at po0.05 for three replicates) incubated in 20 mM phosphate buffer for 60 min at 30 1C
compared to control and native lactoferrin. The error bars are the standard deviation based on three replicates in each test.

weight due to the minor mass difference between aspartyl

and asparaginyl residues, and glutamyl and glutaminyl
residues. However, mobility change equivalent to about
10 kDa difference was observed between native and
amidated LF. This increase of net positive charge of
amidated LF was also conrmed by CE, IEF and IEC
analyses.

Preparative IEF in an aqueous environment (effective pI

range from 3 to 10), and analytical in a vertical IEF gel
(effective pI range from 3 to 8), proved to be unsuitable for
analysis, as theoretical calculation based on the possible
amino acid sequence of amidated LF indicates that
modied protein has pI approximately 9.9. Moreover,
amidation might have increased the hydrophobicity of the

ARTICLE IN PRESS
Y. Pan et al. / International Dairy Journal 17 (2007) 606616

protein (Mattarella et al., 1983) and led to aggregation and

precipitation of the protein at pH X 10. The isoionic point
titration of proteins in solution, another method often used
for the characterisation of modied proteins (Malamud &
Drysdale, 1978; Mattarella et al., 1983), conrmed that
isoionic points and pIs of the proteins studied here were
generally very close to those reported previously (Mattarella et al., 1983), again demonstrating that protein
amidation resulted in an increase in net positive charge.
Amidation not only led to the addition of positive
charges but also caused the polymerisation of the protein.
Amidated LF contained a high molecular mass component
of 4160 kDa, which has been conrmed by SDS-PAGE,
AU-PAGE and SEC, suggesting inter-molecular crosslinking between the carboxyl and amino groups of
polypeptide chains (Mattarella et al., 1983; Tropini, Lens,
Mulder, & Silvestre, 2000). This reaction has been reported
already for amidated BSA even in the absence of the
nucleophile (ammonium chloride) (Murphy & Howell,
1991) and was also conrmed by the polymerisation of the
EDC-LF sample in this study. In addition, native LF
separated on SDS-PAGE into two forms most likely
related to LF-a and -b (84 and 81 kDa), which differ in the
composition of glycan moieties (Wei, Nishimura, &
Yoshida, 2001).
Native LF, according to manufacturer, was 1520%
saturated with iron and was salmon pink in colour.
Following amidation, LF was colourless, indicating that
the amidated protein did not contain iron and may have
lost its ability to bind iron due to structural changes. The
3-D structures of bovine and human lactoferrin are very
similar but not identical, with a sequence identity of 69%.
Each lactoferrin molecule binds two ferric ions with high
afnity in the two specic metal-binding sites of the N- and
C-terminal lobes. Aspartic acid, a histidine, two tyrosines
and a bicarbonate ion contribute to the binding of the
ferric ions in the two specic clefts located in the N- and
C-terminal lobes (Moore et al., 1997). Although, in this
study, the mass spectrometric analysis did not identify
modication of two aspartic acid residues (Asp 60 and 395,
Moore et al., 1997) specically involved in iron binding, a
number of aspartic and glutamic acid residues in the
N- and C-terminal lobes were found to be converted
(Table 1). Iron binding by bLF depends on intrinsic
geometry of the binding site and may also be inuenced by
other factors like interdomain interactions and glycosylation (Moore et al., 1997). Introduction of residue
modications through amidation (although incomplete) is
therefore likely to dramatically reduce, if not abolish, the
iron-binding capacity of amidated protein.
5. Conclusions
Amidation reactions caused dramatic changes in LF
molecules, and these were reected in amino acid
composition, net charge, conformation, hydrophobicity
and solubility. More importantly, amidation enhanced the

615

antimicrobial properties of LF against most Gram-positive

and Gram-negative bacteria tested, but not against
Saccharomyces cerevisiae or Penicillium candidum. These
observations may provide a basis for further development
of LF formulations with enhanced antimicrobial activity
for use in food process hygiene, veterinary and health-care
applications.
Acknowledgements
Yu Pan was the recipient of a Dairy Australia
Postgraduate Research Scholarship. The authors would
like to thank Dr. Jean-Paul Perraudin from Biopole S.A.,
Belgium for kindly providing lactoferrin samples.
References
Appelmelk, B. J., An, Y. Q., Geerts, M., Thijs, B. G., de Boer, H. A., Mac
Laren, D. M., et al. (1994). Lactoferrin is a lipid A-binding protein.
Infection and Immunity, 62, 26282632.
Banks, T. E., Blossey, B. K., & Shafer, J. A. (1969). Inactivation of
chymotrypsin by a water-soluble carbodiimide. The Journal of
Biological Chemistry, 244, 63236333.
Bellamy, W., Takase, H., Wakabayashi, K., Kawase, K., & Tomita, M.
(1992). Antibacterial spectrum of lactoferricin B, a potent bactericidal
peptide derived from the N-terminal region of bovine lactoferrin.
Journal of Applied Microbiology, 73, 472479.
Bellamy, W., Takase, M., Yamauchi, K., Wakabayashi, H., Kawase, K.,
& Tomita, M. (1992). Identication of the bactericidal domain of
lactoferrin. Biochimica et Biophysica Acta, 1121, 130136.
Berkel, P. H. C., Geerts, M. E. J., Veen, H. V., Kooiman, P. M., &
Piepper, F. R. (1995). Glycosylated and unglycosylated human
lactoferrins both bind iron and show identical afnities towards
human lysozyme and bacterial lipopolysaccharide, but differ in their
susceptibilities towards tryptic proteolysis. Biochemical Journal, 312,
107114.
Bradford, M. M. (1976). A rapid and sensitive method for the quantitation
of microgram quantities of protein utilizing the principle of proteindye binding. Analytical Biochemistry, 72, 248254.
Elass-Rochard, E., Roscanu, A., Legrand, D., Trif, M., Salmon, V.,
Motas, C., et al. (1995). Lactoferrinlipopolysaccharide interaction:
involvement of the 2834 loop region of human lactoferrin in the highafnity binding to E. coli 055B5 lipopolysaccharide. Biochemical
Journal, 312, 839845.
Ellison III, R. T., Giehl, T. J., & Laforce, F. M. (1988). Damage of the
outer membrane of enteric Gram-negative bacteria by lactoferrin and
transferrin. Infection and Immunity, 56, 27742781.
Gratzer, P. F., & Lee, J. M. (2001). Control of pH alters the type of crosslinking produced by 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide
(EDC) treatment of acellular matrix vascular grafts. Journal of
Biomedical Materials Research (Applied Biomaterials), 58, 172179.
Heukeshoven, J., & Dernick, R. (1985). Simplied method for silver
staining of proteins in polyacrylamide gels and the mechanism of silver
staining. Electrophoresis, 6, 103112.
Johjima, T., Wariishi, H., & Tanaka, H. (2002). Veratryl alcohol binding
sites of lignin peroxidase from Phanerochaete chrysosporium. Journal of
Molecular Catalysis B: Enzymatic, 17, 4957.
Laas, T. (1989). Isoelectric focusing in gels. In J. C. Janson, & L. Ryden
(Eds.), Protein purification: Principles, high resolution methods, and
applications (pp. 376403). New York, NY: VCH Publishers, Inc.
Leitch, E. C., & Willcox, M. D. (1999). Elucidation of the antistaphylococcal action of lactoferrin and lysozyme. Journal of Medical
Microbiology, 48, 867871.
Lonnerdal, B. (2003). Lactoferrin. In: P.F. Fox, & P.L.H. McSweeney
(Eds.), Advanced dairy chemistry, vol.1: Proteins (3rd ed.,

ARTICLE IN PRESS
616

Y. Pan et al. / International Dairy Journal 17 (2007) 606616

pp. 449466). New York, NY, USA: Kluwer Academic/Plenum

Publishers.
Malamud, D., & Drysdale, J. W. (1978). Isoelectric points of proteins: a
table. Analytical Biochemistry, 86, 620647.
Masson, P. L., Heremans, J. F., & Dive, C. (1966). An iron-binding
protein common to many external secretions. Clinical Chimica Acta,
14, 735.
Mattarella, N. L., & Richardson, T. (1982). Adsorption of positivelycharged b-lactoglobulin derivatives to casein micelles. Journal of Dairy
Science, 65, 22532258.
Mattarella, N. L., & Richardson, T. (1983). Physicochemical and
functional properties of positively charged derivatives of bovine
b-lactoglobulin. Journal of Agricultural and Food Chemistry, 31,
972978.
Mattarella, N. L., Creamer, L. K., & Richardson, T. (1983). Amidation or
esterication or bovine b lactoglobulin to form positively charged
proteins. Journal of Agricultural and Food Chemistry, 31, 968972.
Moore, S. A., Anderson, B. F., Groom, C. R., Haridas, M., & Baker, E.
N. (1997). Three-dimensional structure of diferric bovine lactoferrin at
2-8 A resolution. Journal of Molecular Biology, 274, 222236.
Murphy, M. C., & Howell, N. K. (1991). Effects of the attachment of
small molecules by carbodiimide-mediated condensation on the
physicochemical and functional properties of bovine serum albumin.
Journal of Science, Food and Agriculture, 55, 117140.
Pan, Y., Wan, J., Roginski, H., Lee, A., Shiell, B., Michalski, W. P., et al.
(2005). Effect of chemical modication on anti-microbial and anti-viral
properties of milk proteins. Australian Journal of Dairy Technology, 60,
149151.
Panyim, S., Chalkley, R., Spiker, S., & Oliver, D. (1970). Constant
electrophoretic mobility of the cysteine-containing histone in plants
and animals. Biochimica et Biophysica Acta, 214, 216221.
Papo, N., & Shai, Y. (2003). Can we predict biological activity of
antimicrobial peptides from their interactions with model phospholipid
membranes. Peptides, 24, 16931703.
Shiell, B. J., Beddome, G., & Michalski, W. P. (2002). Mass spectrometric
identication and characterisation of the nucleocapsid protein of
Menangle virus. Journal of Virology Methods, 102, 2735.

Shimazaki, K. (2000). Lactoferrin: A marvelous protein in milk. Animal

Science Journal, 71, 329347.
Spagna, G., Barbagallo, R. N., Casarini, D., & Pifferi, P. G. (2001). A
novel chitosan derivative to immobilize-L-rhamnopyranosidase from
Aspergillus niger for application in beverage technologies. Enzyme and
Microbial Technology, 28, 427438.
Spik, G., Coddeville, B., Mazurier, J., Bourne, Y., Cambillau, C., &
Montreuil, J. (1994). Primary and three-dimensional structure
lactotransferrin (lactoferrin) glycans. Advances in Experimental Medicine and Biology, 357, 2132.
Stile, R. A., Barber, T. A., Castner, D. G., & Healy, K. E. (2002).
Sequential robust design methodology and X-ray photoelectron
spectroscopy to analyze the grafting of hyaluronic acid to glass
substrates. Journal of Biomedical Materials Research (Applied Biomaterials), 61, 391398.
Swart, P. J., Harmsen, M. C., Kuipers, M. E., Van Dijk, A. A.,
Van der Strate, B. W. A., Van Berkel, P. H. C., et al. (1999).
Charge modication of plasma and milk proteins results
in antiviral active compounds. Journal of Peptide Science, 5,
563576.
Tropini, V., Lens, J. P., Mulder, W. J., & Silvestre, F. (2000). Crosslinking of wheat gluten using a water-soluble carbodiimide. Cereal
Chemistry, 77, 333338.
van der Kraan, M. I. A., Groenink, J., Nazmi, K., Veerman, E. C. I.,
Bolscher, J. G. M., & Nieuw Amerongen, A. V. (2004). Lactoferrampin: a novel antimicrobial peptide in the N1-domain of bovine
lactoferrin. Peptides, 25, 177183.
Waterborg, J. H. (1996). Acetic acid-urea polyacrylamide gel electrophoresis of basic proteins. In J. M. Walker (Ed.), The protein protocols
handbook (pp. 8391). Totowa, NJ, USA: Humana Press.
Wei, Z., Nishimura, T., & Yoshida, S. (2001). Characterization of glycans
in a lactoferrin isoform, lactoferrin-a. Journal of Dairy Science, 84,
25842590.
Wilkins, M. R., Lindskog, I., Gasteiger, E., Bairoch, A., Sanchez, J.-C.,
Hochstrasser, D. F., et al. (1997). Detailed peptide characterisation
using PEPTIDEMASSa World-Wide Web accessible tool. Electrophoresis, 18, 403408.