Báo cáo khoa học: Identification of the amniotic fluid insulin-like growth factor binding protein-1 phosphorylation sites and propensity to proteolysis of the isoforms docx

Identification of the amniotic fluid insulin-like growth
factor binding protein-1 phosphorylation sites and
propensity to proteolysis of the isoforms
Lorenzo Dolcini
1
, Alberto Sala
1
, Monica Campagnoli
1
, Sara Labo
`
1
, Maurizia Valli
1
, Livia Visai
1,2
,
Lorenzo Minchiotti
1
, Hugo L. Monaco
3
and Monica Galliano
1
1 Department of Biochemistry ‘A. Castellani’, University of Pavia, Italy
2 Center for Tissue Engineering (C. I. T), University of Pavia, Italy
3 Biocrystallography Laboratory, Department of Biotechnology, University of Verona, Italy
Introduction
The insulin-like growth factor binding proteins (IG-
FBP) are six homologous molecules (IGFBP-1–6) that
play critical roles in a wide variety of important physi-
ological processes. Upon binding they regulate the

availability of both insulin-like growth factors I and II
(IGF-I and -II) and their affinity for these ligands is
modulated by several mechanisms: attachment to the
extracellular matrix and post-translational modifica-
tions, such as proteolysis, phosphorylation and glyco-
sylation. Additionally, several studies have revealed
that IGFBPs, as well as their proteolytic fragments,
have IGF-independent biological activities in cell adhe-
sion and migration and in the regulation of the cell
cycle and apoptosis [1–4]. Furthermore, although these
Keywords
IGFBP; insulin-like growth factor binding
protein-1; mass spectrometry;
phosphorylation; proteolysis
Correspondence
M. Galliano, Department of Biochemistry ‘A.
Castellani’, University of Pavia, viale
Taramelli 3b, 27100 Pavia, Italy
Fax: +39 0382 423108
Tel: +39 0382 987724
E-mail:
(Received 16 March 2009, revised 27 July
2009, accepted 19 August 2009)
doi:10.1111/j.1742-4658.2009.07318.x
Insulin-like growth factor binding protein-1 (IGFBP-1) is the major
secreted protein of human decidual cells during gestation and, as a modula-
tor of insulin-like growth factors or by independent mechanisms, regulates
embryonic implantation and growth. The protein is phosphorylated and
this post-translational modification is regulated in pregnancy and repre-
sents an important determinant of its biological activity. We have isolated,

from human normal amniotic fluid collected in the weeks 16–18, the intact
nonphosphorylated IGFBP-1 and five electrophoretically distinct phospho-
isoforms and have determined their in vivo phosphorylation state. The
unmodified protein was the most abundant component and mono-, bi-, tri-
and tetraphosphorylated forms were present in decreasing amounts. The
phosphorylation sites of IGFBP-1 were identified by liquid chromatogra-
phy–tandem mass spectrometry analysis of the peptides generated with
trypsin, chymotrypsin and Staphylococcus aureus V8 protease. Five serines
were found to be phosphorylated and, of these, four are localized in the
central, weakly conserved, region, at positions 95, 98, 101 and 119, whereas
one, Ser169, is in the C-terminal domain. The post-translational modifica-
tion predominantly involves the hydrophilic stretch of amino acids repre-
senting a potential PEST sequence (proline, glutamic acid, serine,
threonine) and our results show that the phosphorylation state influences
the propensity of IGFBP-1 to proteolysis.
Abbreviations
IGF, insulin-like growth factor; IGFBP, insulin-like growth factor binding protein; LC-ESI-MS, liquid chromatography-electrospray ionization-
mass spectrometry.
FEBS Journal 276 (2009) 6033–6046 ª 2009 The Authors Journal compilation ª 2009 FEBS 6033
proteins are all secreted in the bloodstream, the func-
tional role of IGFBP-1 [5], IGFBP-3 [2] and IGFBP-5
[2] inside the cells has been well documented.
IGFBP-1, the subject of the present study, is predom-
inantly expressed in the liver and, in adult mammals,
the synthesis of the protein is upregulated in a number
of catabolic or stressful conditions, such as fasting and
diabetes [6]. Additionally, it has recently been shown
that a portion of intracellular IGFBP-1 localizes to
mitochondria, where it acts as a prosurvival factor and
protects the liver from apoptosis [5]. IGFBP-1 repre-

sents a minor IGF binding protein in the circulation of
nonpregnant adults, but it is regarded as the most
relevant member of the IGFBP family during gestation
[7–10]. Several studies indicate that IGFBP-1 regulates
embryonic growth as a local modulator of IGF bio-
availability and stimulates trophoblast migration
through binding of its C-terminal domain, which con-
tains an Arg–Gly–Asp sequence recognized by the
integrin family of cell surface receptors [11–14].
Furthermore, the protein has also been shown to play
crucial roles in ovarian, endometrial, trophoblast and
fetal–placental physiology and pathology and its level
in maternal as well as in fetal circulation and in the liver
increases under hypoxic conditions in the uterus, result-
ing in intrauterine growth retardation [15,16].
Its mature polypeptide chain consists of 234 amino
acids and, as all other IGFBPs, contains an N-terminal
and a C-terminal domain linked by a mid-region,
which has a less ordered structure [1,17]. This central
domain is more variable in the different members of
the family and contains most of the sites involved in
post-translational modifications. IGFBP-1 is subject to
phosphorylation on serine residues and backbone
cleavage [1,17].
In the serum of nonpregnant adults, the protein
occurs as a highly phosphorylated single species [18],
whereas differently phosphorylated IGFBP-1 isoforms
are present in the amniotic fluid [19]. Earlier reports
have indicated the presence of a variable number of
isomers [19], probably reflecting sampling at different

gestational periods. We have recently shown that nor-
mal amniotic fluid, collected in the weeks 16–18 of
gestation, contains the unmodified protein and five
electrophoretically distinct phosphoisoforms [20]. A
previous study has shown that IGFBP-1 expressed in
Chinese hamster ovary cells is phosphorylated at
Ser101, 119 and 169 [21,22] and a recent paper
described the modification of Ser98, which was
revealed in the highly phosphorylated protein isoforms
found in hypoxia-treated cells [23].
Post-translational modification influences the inter-
action of IGFBP-1 with IGF-I [21,22,24] and has been
associated with gestational and fetal abnormalities [24–
26]. However, neither the phosphorylation state of the
protein isolated from normal human amniotic fluid
nor the biological properties of the homogeneous
IGFBP-1 phosphoisoforms have been examined so far.
It has been shown that, in vitro, several metallo-
proteases recognize IGFBP-1 as a substrate [27] and a
modulating effect of phosphorylation on the backbone
cleavage of the protein has been recognized [8].
Recently, a specific IGFBP-1 protease activity has been
described in a patient with multiple myeloma and
identified as azurocidin [28].
Here we describe the purification on a preparative
scale of the six isoforms present in human normal amni-
otic fluid collected in the weeks 16–18. The availability
of sufficient amounts of the pure protein allowed a
detailed identification of the amino acid residues
involved in the IGFBP-1 post-translational modifica-

tions occurring in vivo. An important result is that three
of the five phosphorylation sites identified in the present
work are in a region enriched in proline, glutamic acid,
serine and threonine, as it is known that phosphate
addition represents a mechanism for activating a latent
PEST sequence [29]. Amniotic fluid contains a specific
IGFBP-1 protease activity, yielding a stable, functional
and well-structured C-terminal domain [30] and we have
examined the propensity of the different isoforms of
IGFBP-1 to backbone cleavage in the presence of this
partially purified specific protease.
Results
Purification of IGFBP-1 isoforms
The amniotic fluid proteins were separated by gel
filtration and the low molecular mass components,
eluting after the albumin peak, were pooled and fur-
ther purified by anion exchange chromatography. In a
previous study [30], we have shown that amniotic fluid
contains a metalloprotease that cleaves the majority of
IGFBP-1, yielding a stable C-terminal fragment. Thus,
to prevent the proteolytic process, 10 mm EDTA was
added during storage of the fluid and in all buffers
used throughout the gel filtration separation. The pro-
teins were then subjected to anion exchange chroma-
tography (Fig. 1) and the fractions obtained, resolved
by SDS gel electrophoresis, were transferred to
poly(vinylidene difluoride) membranes by electroblot-
ting and probed with anti-IGFBP-1 IgGs (data not
shown). Immunoblotting showed that IGFBP-1 eluted
in six peaks (I–VI) as a single positive band with

identical molecular mass of  30 kDa and that no
fragmentation of the protein had occurred during the
Phosphorylation of amniotic fluid human IGFBP-1 L. Dolcini et al.
6034 FEBS Journal 276 (2009) 6033–6046 ª 2009 The Authors Journal compilation ª 2009 FEBS
purification procedure. The peaks were pooled and
further purified by gel filtration chromatography. Six
homogeneous polypeptide chains with the same N-ter-
minal sequence, APWQCAPCSA, corresponding to
the first 10 residues of intact mature IGFBP-1, were
obtained. Nondenaturing gel electrophoresis (Fig. 2A)
and immunoblotting (Fig. 2B) showed that the electro-
phoretic mobilities of the IGFBP-1 isoforms increased,
but not progressively, as expected on the basis of the
anion exchange chromatography elution profile. Peak IV
migrated more anodically than the preceding chro-
matographic fraction and the last eluting fraction, VI,
displayed a very similar mobility to that of fraction V.
The six proteins were submitted to treatment with alka-
line phosphatase and, when analysed by isoelectric
focusing, all acquired the same isoelectric point of the
polypeptide chain eluting under peak I (data not shown).
This form thus represents the unmodified protein,
whereas the other fractions contain distinct phosphoiso-
forms. These results showed that the nonphosphorylated
protein, as well as the five distinct phosphoisoforms,
were obtained in homogeneous form by anion exchange
chromatography. Furthermore, the addition of the
chelating agent was effective in preventing proteolytic
cleavage, as only the intact proteins were obtained.
Mass determination of the IGFBP-1 isoforms

The molecular mass of the six IGFBP-1 isoforms was
probed by liquid chromatography-electrospray ioniza-
tion-mass spectrometry (LC-ESI-MS) (Table 1). The
mass value determined deconvoluting the multiple
charged ions of the protein eluting in peak I was
25 252 Da and, following reduction, 25 270 Da. These
values match the theoretical molecular mass of the
intact, unmodified IGFBP-1 and the difference
(+18 Da) measured for the reduced sample accounts
for the presence of nine disulfide bonds involving the
18 cysteine residues of the polypeptide chain. Analysis
of peak II yielded a molecular mass of 25 332 Da and
the increase of 80 Da over the previous component is
consistent with the presence of one phosphate group.
Peaks III and IV displayed the same molecular mass
increase of 160 Da over the unmodified protein, corre-
sponding to the presence of two phosphates. Peaks V
(+240 Da) and VI (+320 Da) contain the tri- and
tetraphosphorylated protein, respectively (Fig. 3).
The relative amount of each form was determined
by measuring the corresponding peak area in the chro-
matographic profile and were in good agreement with
the absolute amounts obtained from the absorbance at
280 nm of the pooled fractions (assuming an absor-
bance value of 1.42 for a 1 mgÆmL
)1
solution of
IGFBP-1 as calculated on the basis of the amino acid
Fig. 1. Purification of IGFBP-1 isoforms. The low molecular mass
fraction obtained from human amniotic fluid was resolved by

Q-Sepharose ion exchange chromatography. The elution was
carried out with a linear gradient from 0 to 100% using 6.25 m
M
Bistris-propane, pH 7.5 as buffer A and 6.25 mM Bistris-propane,
pH 9.5, 0.35
M NaCl as buffer B. The peaks I–VI, positive when
probed with anti-IGFBP-1 IgGs, were pooled and further purified by
gel filtration.
A
B
123456
123456
Fig. 2. Nondenaturing gel electrophoresis and western blot analysis
of IGFBP-1 isoforms. IGFBP-1 isoforms isolated by ion exchange
chromatography from peaks I–VI were (A) resolved on a 17% non-
denaturing polyacrylamide gel and stained with Coomassie Brilliant
Blue (lanes 1–6) and (B) following western blot were probed with
anti-IGFBP-1 IgGs (lanes 1–6).
Table 1. Molecular mass of the IGFBP-1 peaks isolated by anion
exchange chromatography.
IGFBP-1
isoforms
Measured
mass (Da)
Theoretical
mass (Da)
Dmass
(Da)
Peak I 25 251.2 25 252.4 +1.2
Peak II 25 331.1 25 332.4 +1.3

Peak III 25 411.0 25 412.4 +1.4
Peak IV 25 410.3 25 412.4 +2.1
Peak V 25 493.1 25 492.4 )0.7
Peak VI 25 572.2 25 572.3 +0.1
L. Dolcini et al. Phosphorylation of amniotic fluid human IGFBP-1
FEBS Journal 276 (2009) 6033–6046 ª 2009 The Authors Journal compilation ª 2009 FEBS 6035
composition). The most abundant nonphosphorylated
form, eluting as peak I, accounted for 69.5% of the total
protein; the monophosphorylated isoform, eluting in
peak II, was 14.9%; the doubly modified forms, eluting
as peaks III and IV, accounted for 6.5 and 5.5%, respec-
tively; the triphosphorylated protein, peak V, repre-
sented 2.7%, and peak VI, corresponding to the form
with four phosphate groups, was 0.9%. These results
show that amniotic fluid IGFBP-1 is modified by up to
four phosphate groups that produce five chromato-
graphically distinct isoforms. The different chromato-
graphic behaviour of the doubly modified protein is
probably due to the alternative positions of the charged
groups, resulting in different isoelectric points.
Identification of IGFBP-1 phosphorylation sites
The examination of the amino acidic sequence of
human IGFBP-1 with the protein pattern database
NetPhos [31] predicted 12 serines, four threonines and
one tyrosine, which could potentially serve as phosp-
hoacceptor sites. Thus, the assumption that the phos-
phate groups can have alternative distribution within
the polypeptide chain suggested that a low level of
modification at each site could be expected. Further-
more, it is known that the ionization efficiency of

phosphopeptides is generally slightly lower than that
of their unmodified counterparts and only relatively
large fragments, between six and 25 amino acids, are
suitable for mass spectral analysis [32]. Therefore, to
enhance sequence coverage and ensure that each
phosphorylation site was contained in at least one
peptide of suitable size and hydrophobicity, the
IGFBP-1 isoforms were cleaved with three different
proteolytic enzymes. Reduction of disulfide bonds was
carried out before each enzymatic digestion and the
resulting peptide mixtures were immediately submitted
to LC-ESI-MS ⁄ MS analysis. The HPLC separation
1000
25+
24+
23+
22+
21+
20+
19+
18+
17+
16+
15+
14+
13+
25+
24+
23+
22+

21+
20+
19+
18+
17+
16+
15+
14+
13+
25+
24+
23+
22+
21+
20+
19+
18+
17+
16+
15+
14+
13+
25+
26+
27+
24+
23+
22+
21+
20+

19+
18+
17+
16+
15+
14+
13+
25+
26+
27+
24+
23+
22+
21+
20+
19+
18+
17+
16+
1200 1400
m/z
1600 1800 2000
25251 Da
25331 Da
25410 Da
25572 Da
25493 Da
E
D
C

B
A
Fig. 3. Multicharged mass spectra of the IGFBP-1 phosphoisoform isolated from human amniotic fluid. The mass values reported on the
y-axis were obtained by deconvoluting the multiply charged pattern obtained for (A) nonphosphorylated protein eluting as peak I; (B) the
mono-1-phosphorylated form, peak II; (C) the biphosphorylated chains eluting in peaks III and IV; (D) the triphosphorylated protein, peak V
and (E) the tetraphosphorylated molecule eluting as peak VI.
Phosphorylation of amniotic fluid human IGFBP-1 L. Dolcini et al.
6036 FEBS Journal 276 (2009) 6033–6046 ª 2009 The Authors Journal compilation ª 2009 FEBS
was performed under acidic conditions in order to
maximize positive charge ionization and the elution
gradient was chosen in order to reduce the loss of
small and hydrophilic peptides during the desalting
step. Phosphoric acid was added to the sample solu-
tion because it significantly enhances the detection of
mono- and multiphosphorylated peptides following
RP-HPLC [33]. The identity of the fragments was
assessed by manual inspection and by comparing the
experimental fragmentation pattern with the theoretical
values obtained using the protein prospector
( )
and peaks studio software (Bioinformatic Solution
Inc., Waterloo, Canada). In addition, as phosphoryla-
tion can reduce the efficiency of cleavage [34], the data
were also examined for incompletely digested frag-
ments. The phosphorylated peptides were identified by
the presence of species with a theoretical peptide mass
increased by 80 Da (plus one phosphate group) or
multiples thereof, and in all cases, ions corresponding
to the loss of H
3

PO
4
()98 for single charged ions and
)49 for doubly charged ions) dominated the fragmen-
tation pattern. A complete sequence coverage of the
protein was obtained (data not shown) and Table 2
lists the phosphorylated peptides observed. The V8
fragments spanning residues 93–100 (AGSPESPE)
from the unmodified protein, at m⁄ z 773.3, and from
the monophosphorylated one, at m ⁄ z 853.2, displayed
a mass difference of 80 Da, which is consistent with
the presence of one phosphate group. Fragmentation
in the ion trap showed that this peptide was phosphor-
ylated at Ser95 and, to a much lower extent, also at
Ser98. The presence of the fragment ion at m ⁄ z 755.3
is due to the loss of phosphoric acid (98 Da), the ion
at m ⁄ z 296.1 (b3+P) is consistent with the modifica-
tion of Ser95 and the ions at m ⁄ z 638.2, (y5+P) and
at m ⁄ z 442.2 (b5) are indicative of the modification at
Ser98 (Fig. 4B). Figure 4A shows the MS ⁄ MS spec-
trum obtained for the unmodified peptide. The two
peptides coelute and the doubly modified fragment was
not observed. The MS ⁄ MS spectra of the V8 fragment
93–103 (AGSPESPESTE) from the monophosphory-
lated isoform, at m ⁄ z 1172.1 (+80 Da shift) is shown
in Fig. 5A. The loss of phosphoric acid (98 Da) from
the precursor yielded a charged species at m⁄ z 1072.1
and the presence of a product ion at m ⁄ z 642.0
(y5+P) is consistent with phosphorylation at Ser101.
The fragment ions at m ⁄ z 562.2 and at m ⁄ z 649.0

correspond to unmodified y5 and y6, respectively, and
indicate that Ser95 is phosphorylated as well. The
modification of Ser98 could not be distinguished in
this peptide. Analysis of the V8 digest of the trip-
hosphorylated protein showed the presence of peptide
93–103 carrying two phosphate groups. MS ⁄ MS spec-
tra (Fig. 5B) of the charged ion at m ⁄ z 1250.0 dis-
played fragment ions at m ⁄ z 1152.1 and at m ⁄ z 1054.1,
consistent with the loss of one and two phosphate
groups, respectively. The ion at m ⁄ z 642.0 (y5+P)
indicates the phosphorylation of Ser101 and the
ion at m ⁄ z 955.0 (y8+P) indicates that in the
Table 2. List of the phosphopeptides revealed by LC-ESI-MS ⁄ MS. Modified residues are underlined.
Residues Sequence
Calculated
mass (Da)
Measured
mass (Da)
Dmass
(Da)
Chymotryptic peptides
70-113 +P HALTRGQGACVQESDASAPHAAEAG
SPESPESTEITEEELLDNF 4660.1 4660.8 +0.7
70-113 +2P HALTRGQGACVQESDASAPHAAEAG
SPESPESTEITEEELLDNF 4741.0 4741.4 +0.4
114-127 +P HLMAP
SEEDHSILW 1743.7 1743.7 0
157-175 +P RVVESLAKAQET
SGEEISKF 2287.1 2287.4 +0.3
Glu C peptides

83-108 +2P SDASAPHAAEAG
SPESPESTEITEEE 2787.1 2787.8 +0.7
93-100 +P AG
SPESPE 853.4 853.2 )0.2
93-103 +P AG
SPESPESTE 1169.4 1169.3 )0.1
93-103 +2P AG
SPESPESTE 1249.4 1249.2 )0.2
93-108 +P AGSPE
SPESTEITEEE 1770.7 1771.4 )0.3
100-111 +P
STEITEEELLD 1357.6 1357.6 0
109-121 +P LLDNFHLMAP
SEE 1594.7 1595.2 )0.5
109-122 +P LLDNFHLMAP
SEED 1708.7 1708.6 )0.1
112-121 +P NFHLMAP
SEE 1253.5 1253.9 +0.4
113-121 +P FHLMAP
SEE 1139.4 1139.9 +0.5
161-172 +P SLAKAQET
SGEE 1330.0 1330.3 +0.3
Tryptic peptides
165-175 +P AQET
SGEEISK 1258.5 1258.6 +0.1
165-183 +P AQET
SGEEISKFYLPNCNK 2237.0 2239.0 +2.0
L. Dolcini et al. Phosphorylation of amniotic fluid human IGFBP-1
FEBS Journal 276 (2009) 6033–6046 ª 2009 The Authors Journal compilation ª 2009 FEBS 6037
biphosphorylated fragment, Ser95 and Ser101 are

modified. The triphosphorylated peptide was not
observed. These results show that the sequence span-
ning residues 93–103 contains three serine residues
whose adjacent amino acids form consensus sites for
phosphorylation. The V8 fragment spanning residues
109–121, at m ⁄ z 1515.3, was found in the unmodified
protein and the fragmentation pattern of its phosphor-
ylated counterpart, at m ⁄ z 1595.2, displayed ions
accounting for the modification of Ser119 (data not
shown). The modification of Ser169 was detected in
the V8 fragment spanning residues 161–172, at m ⁄ z
1330.3, which showed the expected fragmentation ions
(data not shown). The MS ⁄ MS spectrum of the V8
fragment TSMDGE, found in all the IGFBP-1 sam-
ples, and spanning residues 219–224 at m ⁄ z 654.9
(Fig. 6), showed that the molecular mass increase of
15.9 Da over the theoretical value was due to the
oxidation of methionine to sulfoxide, a common
post-translational modification to proteins occurring
in vivo. The nonoxidized sequence was also present at
m ⁄ z 639.1.
These results lead to the conclusion that the IG-
FBP-1 phosphoisomers isolated from normal amni-
otic fluid contain five phosphoacceptor sites. The
phosphorylation of the five serine residues was
detected in all the chromatographically distinct
isomers. A complete sequence coverage was obtained
for each polypeptide chain and Table 2 lists the
molecular mass of the peptides with phosphorylated
residues. One novel phosphorylation site on Ser101

was revealed, and the phosphoacceptor sites Ser95,
Ser98, Ser119 and Ser169, previously described in
cultured cells [22,23], were confirmed in our natural
samples of human amniotic fluid IGFBP-1. In agree-
ment with previous data [22,35], we did not find
any evidence of threonine and tyrosine phosphor-
ylation. For this reason, and because it is well
Fig. 4. LC-MS ⁄ MS ion spectra of IGFBP-1
peptide 93–100. (A) MS ⁄ MS spectrum of
the singly charged 773.3 Da ion correspond-
ing to the nonphosphorylated AGSPESPE
fragment. (B) MS ⁄ MS spectrum of the
singly charged 853.2 Da ion corresponding
to the monophosphorylated AGSPESPE
sequence.
Phosphorylation of amniotic fluid human IGFBP-1 L. Dolcini et al.
6038 FEBS Journal 276 (2009) 6033–6046 ª 2009 The Authors Journal compilation ª 2009 FEBS
known that the phosphopeptides are not eluted
with high efficiency from RP-HPLC, a quantitative
analysis of each phosphorylation site could not be
performed.
Chromatographic separation of the V8 peptides
In order to decide whether a major phosphoacceptor
serine is present in the monophosphorylated protein, it
A
B
Fig. 5. LC-MS ⁄ MS ion spectra of IGFBP-1
peptide 93–103. (A) MS ⁄ MS spectrum of
the singly charged 1172.1 Da ion corre-
sponding to the monophosphorylated AGSP-

ESPESTE fragment. (B) MS ⁄ MS spectrum
of the singly charged 1250.0 Da ion
corresponding to the triphosphorylated
AGSPESPESTE peptide.
Fig. 6. LC-MS ⁄ MS spectrum of the IGFBP-
1 peptide 219–224. The asterisk marks the
oxidized methionine in the sequence and in
the internal ions at m ⁄ z 234.5 and 262.9.
L. Dolcini et al. Phosphorylation of amniotic fluid human IGFBP-1
FEBS Journal 276 (2009) 6033–6046 ª 2009 The Authors Journal compilation ª 2009 FEBS 6039
was digested with endopeptidase V8 from Staphylococ-
cus aureus and compared with the unmodified form.
The peptide mixtures were resolved by RP-HPLC
chromatography and Fig. 6 shows the superimposition
of the UV traces monitored at 214 nm. The peaks were
manually collected and submitted to N-terminal
sequencing and LC-ESI-MS ⁄ MS analysis. The chro-
matographic profiles, as well as the peak area and
shape, did not differ in a significant manner (Fig. 7).
The only major discrepancy was the presence of a
peak, indicated as 1, in the trace of the unmodified
protein, which was absent in the profile of the
monophosphorylated molecule, where an additional
one, indicated as peak 2, was present. The N-terminal
sequence analysis showed that peak 1 contains the
sequence AGSPESPE (93–100) and that peak 2 corre-
sponds to the fragment AGSPESPESTE (93–103).
Mass analysis confirmed these data and showed that
the latter was phosphorylated. These results indicate
that the presence of the phosphate group prevents the

enzymatic cleavage. Furthermore, as no other evident
peak shift was appreciable, it can be argued that the
monophosphorylated IGFBP-1 form is predominantly
modified in the sequence spanning residues 93–103.
Proteolysis
In a previous paper [30], we described the isolation of
the C-terminal domain of IGFBP-1 and suggested that
the proteolytic cleavage could be ascribed to a specific
amniotic fluid metalloprotease. Therefore, prior to
examining the effect of phosphorylation on the proteo-
lytic process, we prefractionated the amniotic fluid
proteins by gel filtration and analysed the fractions
obtained for their IGFBP-1 protease activity. As shown
in Fig. 8A, unmodified IGFBP-1 incubated with the
fraction containing components with molecular mass
above 50 kDa, remained almost unchanged, whereas in
Fig. 7. HPLC elution profiles of the V8 peptides. Trace A was
obtained from the nonphosphorylated IGFBP-1 and trace B from
the monophosphorylated protein. The arrows indicate the major
differences between the two chromatograms (peaks 1 and 2).
A
B
C
Fig. 8. IGFBP-1 proteolysis. (A) Aliquots of the nonphosphorylated
protein (1 lgin20lLof20m
M Tris pH 7.5 containing 20 mM
CaCl
2
) were incubated overnight at 37 °C with 5 lL of the amniotic
fluid fractions containing proteins eluting ahead (lane 1) and after

(lane 2) the albumin peak in gel filtration chromatography; as a con-
trol, the isolated C-terminal fragment and the untreated protein are
shown in lanes 3 and 4, respectively. Following SDS gel electropho-
resis, the proteins were electroblotted and probed with
anti-IGFBP-1 IgGs. (B) Aliquots of the nonphosphorylated IGFBP-1
(1 lgin20m
M Tris pH 7.5 containing 20 mM CaCl
2
) were incubated
overnight with no addition (lane 1), and with the addition of increas-
ing amounts of the amniotic fluid fraction displaying proteolytic
activity: 2 lL (lane 2); 4 lL (lane 3); 8 lL (lane 4); and 16 lL (lane
5). Sample volumes were adjusted to 30 lL and incubations were
terminated after 48 h by the addition of 30 lL of sample buffer.
The samples were then submitted to SDS gel electrophoresis. The
proteins were electroblotted and probed with anti-IGFBP-1 IgGs.
(C) Substrate in gel electrophoresis of the amniotic fluid protease-
containing fraction. Gelatin (lane 1) and IGFBP-1 (lane 2) were
added to a 10% polyacrylamide solution before gel casting at a con-
centration of 1 mgÆmL
)1
. The asterisk indicates the lysis area
observed in IGFBP-1 zymography.
Phosphorylation of amniotic fluid human IGFBP-1 L. Dolcini et al.
6040 FEBS Journal 276 (2009) 6033–6046 ª 2009 The Authors Journal compilation ª 2009 FEBS
the presence of proteins eluting after the albumin peak,
IGFBP-1 was partially cleaved, yielding two major
bands, one corresponding to the C-terminal domain
and a second with an apparent molecular mass of
20 kDa. No fragmentation was observed when the incu-

bation was performed in the presence of 10 mm EDTA
(data not shown). The fraction containing the IGFBP-1
protease activity was then used to assess the dose
dependence of the proteolytic process. Figure 8B shows
that by increasing the amount added to the protein the
intensity of the 30 kDa band is progressively reduced.
The protease-containing fraction was then submitted to
zymography using gelatine and IGFBP-1 as substrates
in order to examine the specificity of the IGFBP-1
degrading activity. The gelatine substrate zymogram
showed the presence of several areas of lysis, represent-
ing gelatine-degrading proteinase activity in the sample.
On the contrary, only one area was evidenced in the
IGFBP-1 substrate zymogram (Fig. 8C). These results
indicate that a specific protease is involved in the prote-
olytic cleavage occurring in amniotic fluid. We then
examined the effect of phosphorylation on this process.
Each isoform was incubated with the proteinase-
containing fraction and the samples were then submit-
ted to western blot analysis, together with identical
aliquots of the untreated proteins (Fig. 9A). The
degradation process was evaluated on the basis of the
ratio between the integrated areas of the intact protein
in the treated and untreated samples. This value was
determined for each isoform and the results (Fig. 9B)
show that the monophosphorylated IGFBP-1 was
cleaved almost to the same extent as the unmodified
protein and that the susceptibility to proteolytic
degradation of the isoforms increased with the number
of phosphates linked to the polypeptide chains. The

different position of the groups bound to the biphosph-
orylated protein also played a role in the cleavage
process.
Discussion
The purification procedure used in this study permitted
the isolation of the unmodified as well as of the five
phosphoisoforms of IGFBP-1 previously detected by
proteomic analysis in human amniotic fluid collected
in weeks 16–18 of gestation [20]. To obtain a sufficient
amount of the biological fluid, several hundred individ-
ual samples were pooled and, therefore, the prepara-
tion was assumed to be normal as every individual
difference would be below the detectability score. The
molecular mass of the nonphosphorylated protein was
determined under both reducing and nonreducing con-
ditions. The values obtained matched the molecular
mass deduced from the amino acid sequence and were
consistent with the presence of nine disulfide bonds. In
addition, the molecular mass increments of the other
isoforms defined the exact number of phosphate
groups linked to the chromatographically distinct
IGFBP-1 phosphoisomers isolated. The characteriza-
tion of the phosphorylation sites of IGFBP-I isoforms
revealed that the post-translational modification
involves five serine residues (numbered as Ser95, Ser98,
Ser101, Ser119 and Ser169 in the mature polypeptide
chain). In a previous study, we examined the phos-
phorylation of IGFBP-1 isolated from amniotic fluid
following the selective extraction of phosphopeptides
on a titanium dioxide (TiO

2
) cartridge and only three
modified serines were located [36]. Enrichment meth-
ods are well suited when only minute amounts of the
phosphoprotein are available, but a detailed analysis
of phosphoptides demands relatively high quantities of
the homogeneous molecules. All the residues identified
in the present study were recognized by the protein
A
B
Fig. 9. Proteolysis of IGFBP-1 isoforms. (A) Aliquots of each
IGFBP-1 isoform (1 lgin20lLof20m
M Tris pH 7.5 containing
20 m
M CaCl
2
) were submitted to SDS electrophoresis and immuno-
blotting following incubation at 37 °C overnight with the addition of
a4lL aliquot of the amniotic fluid fraction containing IGFBP-1 pro-
tease activity. As a control, identical aliquots of the six isoforms
diluted to the same concentration with the buffer containing 20 m
M
CaCl
2
were incubated separately at 37 °C overnight and analysed.
For each isoform, indicated with the peak number, the left lane
contains the untreated protein and the right lane the sample treated
with the protease-containing fraction. (B) Histogram representation
of the susceptibility to proteolytic degradation of IGFBP-1 isoforms.
The y-axis shows the percentage of the ratio between the spot

area of the degraded and that of the untreated protein measured
using
IMAGE J 1.37V software.
L. Dolcini et al. Phosphorylation of amniotic fluid human IGFBP-1
FEBS Journal 276 (2009) 6033–6046 ª 2009 The Authors Journal compilation ª 2009 FEBS 6041
pattern database NetPhos [31] as phosphoacceptor
sites, with a score always higher than 0.99 and, in
agreement with previous data [22,35], no modification
occurring at tyrosine and threonine residues was
found. Although a previous study, using the protein
expressed in Chinese hamster ovary cells, identified
three serines (Ser101, Ser119 and Ser169) as the only
modified amino acids and excluded the involvement of
other residues [22], our results, showing the presence
of the tetraphosphorylated isoform, revealed a more
extensive post-translational modification of the amni-
otic fluid protein. Because the process can be the con-
sequence of a combination of biological variables,
including different activity levels of both the intracellu-
lar protein kinases and the extracellular phosphatases,
this discrepancy is probably due to the different source
of the examined molecules. Moreover, it has been
shown that the phosphorylation of IGFBP-1 is con-
trolled by placental steroid hormones [37] and by IGF-
II [38]. Both factors are probably regulated differently
in Chinese hamster ovary cells and in human decidual
cells. A recent paper revealed the modification of
Ser98 in the hyperphosphorylated IGFBP-1 expressed
in hypoxia exposed cells and the finding was suggested
to be associated with the condition [23]. However, our

data show that this phosphoacceptor site is also modi-
fied in protein purified from normal amniotic fluid.
The kinases responsible for in vivo phosphorylation of
IGFBP-1 are as yet unknown, but it has been shown
that IGFBP-1 is a substrate for several protein kinases
[31,35] and the residues adjacent to all the modified se-
rines form consensus sequences for post-translational
modification by casein kinase II.
A quantitative determination of the modification
occurring at each site could not be achieved because of
the poor ionization efficiency and fast degradation of
phosphopeptides. In particular, the detection of multi-
phosphorylated peptides is reduced owing to their high
propensity to adsorption to exposed surfaces [34]. Fur-
thermore, nonspecific and partial enzymatic cleavages
caused the presence of several fragments containing the
same phosphorylation site. However, the examination
of the peptides obtained following V8 proteolytic cleav-
age indicated that the monophosphorylated IGFBP-1
form is predominantly modified within the AGSPES-
PESTE sequence containing three of the five phosphor-
ylated residues identified in the present study. It is
interesting to note that this polypeptide sequence is
rich in proline, glutamic acid, serine and threonine, a
so-called PEST region, which is typical of rapidly
metabolized proteins. Briefly, the PEST hypothesis
suggests that protein regions containing a high local
concentration of the amino acids proline, glutamic acid,
serine, threonine, and, to a lesser extent, aspartic acid,
are suitable targets for proteolytic degradation [29].

Examination of the amino acid sequence of IGFBP-1
with the protein pattern database PESTfind (available
at a com-
putational tool designed to predict potential PEST
sequences, confirmed that the region spanning residues
89–114 (HAAEAG
SPESPESTEITEEELLDNFH) is
potentially one of those. This region is located 27
residues far from the proteolytic site that leads to the
formation of the C-terminal fragment previously
isolated and characterized from amniotic fluid [30].
Proteolysis of IGFBPs is an important mechanism
that controls IGF bioavailability [39,40] and previous
studies have focused on the cleavage of IGFBP-1 by
matrix metalloproteases present in amniotic fluid and
conditioned medium from decidualized endometrial
cells [8,25]. The effect of IGFBP-1 phosphorylation
on its susceptibility to enzymatic cleavage has been
examined in vitro and Gibson et al. [8] showed that the
highly phosphorylated protein is resistant to the prote-
ase activity in decidual conditioned medium and to
plasmin, whereas Kabir-Salmani et al. [37] suggested
that matrix metalloproteinase-9 selectively degrades
phosphorylated IGFBP-1. Recently, it has been
reported that the IGFBP-1 specific protease activity
identified as azurocidin cleaves both phosphorylated
and nonphosphorylated IGFBP-1 [28]. These discrep-
ancies probably reflect the specificity of any enzymatic
activity and, although the protein has been recognized
as a potential physiological substrate for several matrix

metalloproteases, little is known about the susceptibil-
ity of IGFBP-1 to enzymatic cleavage. Because the
only proteolytic degradation of IGFBP-1 occurring
without the addition of endogenous reactants so far
described under normal conditions is that observed in
the amniotic fluid [41], we performed the assays using
a partially purified fraction of amniotic fluid displaying
IGFBP-1 specific protease activity. Our data clearly
show that the post-translational modification increases
the susceptibility to cleavage by the amniotic fluid pro-
tease and suggest that not only the degree, but also the
position of the phosphate groups influence the process.
As for the other IGFBPs, the proteolytic cleavage of
IGFBP-1 occurs within the central domain and we
have shown that the process generates a C-terminal
domain (residues 141–234) carrying a phosphate group
linked to Ser169 [30]. It is possible that the charge
repulsion between the phosphate groups in the mid-
region and the one located in the C-terminal portion
of the protein may cause a conformational change that
makes the cleavage site more exposed to the protease.
Thus, the post-translational modification might act as
Phosphorylation of amniotic fluid human IGFBP-1 L. Dolcini et al.
6042 FEBS Journal 276 (2009) 6033–6046 ª 2009 The Authors Journal compilation ª 2009 FEBS
a mechanism for increasing the bioavailability of IGFs
by enhancing the proteolytic cleavage as well as for
producing the C-terminal domain.
We have examined the amino acid sequences of the
other protein family members using the PESTfind
database and found that a high score region as a

PEST sequence is present only in the IGFBP-5 poly-
peptide chain, spanning residues 130–146, and, with a
lower score, in IGFBP-6 between residues 150 and
175. The in vivo post-translational modification of
IGFBP-5 has recently been examined in detail and
mono- and biphosphorylated isoforms were shown to
be secreted by human breast carcinoma cells [42]. The
major phosphorylation site of the protein was found
to involve Ser96, in the central domain of IGFBP-5,
and a minor one was localized very close to the C-ter-
minus, at Ser248. Ser96 on IGFBP-5 shares the same
proximity to the N-terminal domain as Ser95, Ser98
and Ser101 on IGFBP-1 (our results), Ser111 and
Ser113 on IGFBP-3, and Ser106 on IGFBP-2 [42,43].
IGFBP-1 only seems to have phosphoacceptor sites
located in a PEST region, but the presence of phos-
phorylation at conserved serine residues points to a
pivotal role of this post-translational modification in
the regulation of biological activities of IGFBPs.
The present study identified five phosphorylated ser-
ine residues in the polypeptide chain of amniotic fluid
IGFBP-1 and showed that the modification does not
take place at the same extent at each of the potential
sites, leading to heterogeneity. Compared with previ-
ous studies, based on selective enrichment by affinity
techniques, the availability of relatively high amounts
of pure protein allowed us to carry out a more com-
plete and detailed characterization of the phosphoac-
ceptor sites. Four of the five phosphorylated residues
identified are located in the weakly conserved and

poorly structured sequence composed of  65 amino
acids connecting the N- and the C-terminal domains.
The prevalent phosphoacceptor sites of the monopho-
sphorylated protein are located in a PEST region pres-
ent in proteins that are a target for phosphorylated
mediated degradation and our data show that this
post-translational modification increases the propensity
of IGFBP-1 to proteolytic cleavage.
Experimental procedures
Chemicals
Immobilon-P poly(vinylidene) difluoride transfer mem-
branes were obtained from Millipore (Bedford, MA, USA).
Sephadex G-100, Supedex-75 and Q-Sepharose resins, as
well as the enhanced chemiluminescence detection system
were purchased from Amersham Biosciences (Piscataway,
NY, USA). Gelatin, type A from porcine skin, and alkaline
phosphatase were purchased from Sigma (St Louis, MO,
USA). A MilliQ system (Millipore) was used for water
purification. Laboratory chemicals were purchased from
Sigma, unless otherwise specified.
Isolation of IGFBP-1 isoforms from human
amniotic fluid
Human amniotic fluid was obtained from discarded
amniocentesis samples collected in weeks 16–18, pooled
and stored frozen until used. EDTA (10 mm) was added
to avoid unwanted proteolysis. The fluid (2 L) was satu-
rated to 90% with ammonium sulfate and after centrifu-
gation at 10 000 g for 1 h, the pellet was dissolved in
50 mm Tris ⁄ HCl, pH 8.0, 150 mm NaCl, 10 mm EDTA,
dialysed against the same buffer and fractioned twice by

gel filtration on a Sephadex G-100 column (4 · 100 cm).
Proteins eluting after the albumin peak were pooled,
equilibrated in 6.25 m m Bistris-propane, pH 7.5, and sep-
arated on a Q-Sepharose (75 mL) column equilibrated
with the same buffer (buffer A) and connected to an
A
¨
kta Prime system (Amersham Biosciences). The elution
was carried out with a linear gradient from 0% to 100%
of 6.25 mm Bistris-propane, pH 9.5, 350 mm NaCl (buffer
B) and monitored at 280 nm. The fractions containing
the IGFBP-1 isoforms were pooled, concentrated and
submitted to gel filtration on a Superdex G-75 column
(1.6 · 60 cm) equilibrated and eluted with 20 mm
Tris ⁄ HCl buffer, pH 8.0, 150 mm NaCl. The homogeneity
of the proteins was checked by SDS PAGE analysis and
by N-terminal sequencing performed in a Hewlett-Packard
model G 1000 A sequencer (Centro Grandi Strumenti,
University of Pavia).
Gel electrophoresis and western blot analysis
Isoelectric focusing was carried out on laboratory made
gels, cast on GelBond with a 4–10 nonlinear immobilized
pH gradient obtained with Acrylamido buffer solutions
(Fluka, Sigma-Aldrich Schweiz, Buchs, Switzerland). SDS
PAGE was carried out using a Mini PROTEAN II cell
(Bio-Rad, San Diego, CA, USA) and the gels were stained
with Coomassie Blue. Nondenaturing PAGE was carried
out on a discontinuous polyacrylamide gel system, using
pH 6.8 in the 4% stacking gel and pH 7.5 in the 10%
resolving gel. Polyclonal antibodies were produced as previ-

ously described [30] and antibody titres were assayed either
by ELISA or immunoblotting. The specific IgGs were puri-
fied by affinity chromatography on a Protein A-Sepharose
column according to the manufacturer’s recommendations
(Amersham Biosciences). For western blot analyses, after
electrophoresis, the proteins were transferred by electroblot-
ting to Immobilon-P poly(vinylidene difluoride) membranes
L. Dolcini et al. Phosphorylation of amniotic fluid human IGFBP-1
FEBS Journal 276 (2009) 6033–6046 ª 2009 The Authors Journal compilation ª 2009 FEBS 6043
using a Mini Protean II apparatus (Bio-Rad). The mem-
brane was blocked with 5% w ⁄ v skim milk and probed
with the mouse antiserum diluted 1 : 2000. Immunoreactive
spots were detected with horseradish peroxidase-conjugated
anti-mouse IgG and developed by the enhanced chemilumi-
nescence method.
LC-MS analysis of IGFBP-1 isoforms
The system used for LC-MS analyses was a Surveyor LC
coupled with an ion trap mass spectrometer LCQ (Thermo
Scientific, San Jose, CA, USA) equipped with an ESI
source and controlled by xcalibur software 1.3 (Thermo
Scientific). Experiments were carried out in positive ion
mode under constant instrumental conditions: source volt-
age 4.5 kV, capillary voltage )20 V, capillary temperature
210 °C, tube lens voltage )5 V. Mass spectra were acquired
in the mass range of 200–2000 m ⁄ z by scanning the
magnetic field in 200 ms. The purified isoforms, 20 lL out
of a 0.5 mgÆmL
)1
solution in 100 mm ammonium carbon-
ate, pH 8.5, were loaded onto a C18 column (Phenomenex

Jupiter 300, 5 lm, 2 · 50 mm) and eluted with a linear
gradient from solution A (5% acetonitrile) to solution B
(95% acetonitrile) both containing 0.1% formic acid over
15 min at a flow rate of 0.2 mLÆmin
)1
. An aliquot of the
unmodified protein was reduced by adding 10% 1,4-dith-
iothreitol to the 100 mm ammonium carbonate solution.
Deconvolution of multiple charged ions was achieved using
the mag tran 1.01 software [44].
Preparation of proteolytic digests for
LC-ESI-MS
⁄
MS
Three aliquots of each IGFBP-1 isoform, 50 lg in 100 lL
of 100 mm ammonium carbonate pH 8.5, were reduced for
30 min at 60 °C with 10 lL of a 10 mm freshly prepared
dithiothreitol solution in the same buffer. Proteolysis was
performed with either endopeptidase V8 from S. aureus,
trypsin or chymotrypsin. The enzymes, in 100 mm ammo-
nium bicarbonate at pH 8.5, were added at a ratio of 1 : 50
(w ⁄ w), allowed to react at 37 °C for 3 h and the digestions
were stopped by the addition of phosphoric acid (final con-
centration 0.8% v ⁄ v).
LC-ESI-MS
⁄
MS analysis of peptides
The peptide solutions (20 lL) were resolved in an analytical
C18 column (Phenomenex Jupiter 300, 5 lm, 250 · 2mm) at
a flow rate of 1 mLÆmin

)1
with a gradient 2–60% B in
120 min, 60–98% B in 30 min and 98% B for 10 min. Sol-
vents consisted of water (A) and 60% acetonitrile (B) both
containing 0.1% trifluoroacetic acid. The eluent was analysed
online with an LCQ ion trap mass spectrometer (Thermo Sci-
entific) with the ESI ion source controlled by the xcalibur
software 1.4 (Thermo Scientific). Mass spectra were gener-
ated in positive ion mode under constant instrumental condi-
tions: source voltage 4.0 kV, capillary voltage 46 V, sheath
gas flow 40 (arbitrary units), auxiliary gas flow 10 (arbitrary
units), sweep gas flow 1 (arbitrary units), capillary tempera-
ture 250 °C, tube lens voltage )105 V. MS ⁄ MS spectra,
obtained by collision-induced dissociation in the linear ion
trap, were performed with an isolation width of 3 Da (m ⁄ z),
the activation amplitude was 35% of ejection radio frequency
amplitude, which corresponds to 1.58 V. MS ⁄ MS spectra
were interpreted manually with the assistance of the predic-
tion algorithm for peptide fragmentation proteinprospec-
tor [45] and were automatically analysed using peaks
studio, version 4.2 (Bioinformatic Solution Inc., Waterloo,
Canada).
RP-HPLC separation of the V8 peptides
Nonphosphorylated and monophosphorylated fractions
(1 mg) were digested with the same protocol reported above
and were loaded on to a Vydac C18 reverse phase column
(4.6 · 250 mm). Elution was carried out with a linear gradi-
ent from 0% to 100% solvent B in 100 min at a constant
flow rate of 1 mLÆmin
)1

. The solvents consisted of water
(A) and acetonitrile (B), both containing 0.1% formic acid.
The elution profile was monitored, recording the UV trace
at 214 nm. Peaks were manually collected and subsequently
analysed via LC-ESI-MS ⁄ MS and with N-terminal sequenc-
ing as described above.
Degradation of IGFBP-1 isoforms by amniotic
fluid protease
Fresh amniotic fluid (100 mL), without the addition of
EDTA, was saturated to 90% with ammonium sulfate and
the precipitated proteins were redissolved and fractioned by
gel filtration chromatography as described above. The
components eluting ahead and after the albumin peak
were pooled and concentrated separately to  5mgÆmL
)1
.
Aliquots of the amniotic fluid fractions were added to 1 lgof
either unmodified IGFPB-1 or phosphoisoforms in 20 lLof
20 mm Tris ⁄ HCl pH 7.5 containing 20 mm CaCl
2
and incu-
bated at 37 °C overnight. The incubations were terminated
by the addition of sample buffer, without reducing agents,
and the degradation was analysed by western blot and quan-
tified using the image j 1.37v software, freely available online
at Each experiment was repeated
three times in order to check the reproducibility of both the
proteolytic process and the method of analysis.
For IGFBP substrate zymography, the unmodified pro-
tein was added to a 10% polyacrylamide solution before

gel casting (1 mgÆmL
)1
) and an aliquot of the amniotic fluid
fraction, diluted 1 : 1 in nonreducing sample buffer, was
electrophoresed. Only one lane, obtained with a spacer, was
made using this substrate, whereas the remaining part of
the cassette was filled with the resolving polyacrylamide gel
Phosphorylation of amniotic fluid human IGFBP-1 L. Dolcini et al.
6044 FEBS Journal 276 (2009) 6033–6046 ª 2009 The Authors Journal compilation ª 2009 FEBS
solution. Gelatine substrate zymography was carried out
using standard zymography methods, using gelatine at a
concentration of 1 mgÆmL
)1
[46].
Acknowledgements
We are grateful to Rossella Greco, Department of
General Biology and Medical Genetics, University of
Pavia for supplying the amniotic fluid; Patrizia Arcidi-
aco, Centro Grandi Strumenti, University of Pavia for
automated sequencing; and Angelo Gallanti, Depart-
ment of Biochemistry, University of Pavia for technical
assistance. This work was supported by a grant from
the Italian Ministry of Education and Scientific
Research (FIRB 2003).
References
1 Jones JI & Clemmons DR (1995) Insulin-like growth
factors and their binding proteins: biological actions.
Endocr Rev 16, 3–34.
2 Firth SM & Baxter RC (2002) Cellular actions of the
insulin growth factor binding proteins. Endocr Rev 23,

824–854.
3 Baxter RC (2000) Insulin-like growth factor (IGF-)
binding proteins: interactions with IGFs and intrinsic
bioactivities. Am J Physiol Endocrinol Metab 278,
967–976.
4 Kiepe D, Van der Paas A, Ciarmatori S, Sta
¨
ndker L,
Schu
¨
tt B, Hoeflich A, Hu
¨
gel U, Oh J & To
¨
nshoff B
(2008) Defined carboxy-terminal fragments of insulin-
like growth factor binding protein-2 exert similar
mitogenic activity on cultured rat growth plate
chondrocytes as IGF-I. Endocrinology 149, 4901–4911.
5 Leu JI & George DL (2007) Hepatic IGFBP1 is a
prosurvival factor that binds to BAK, protects the liver
from apoptosis, and antagonizes the proapoptotic actions
of p53 at mitochondria. Genes Dev 21, 3095–3109.
6 Lang CH, Vary TC & Frost RA (2003) Acute in vivo
elevation of insulin-like growth factor (IGF) binding
protein-1 decreases plasma free IGF-I and muscle
protein synthesis. Endocrinology 144, 3922–3933.
7 Westwood M (1999) Role of insulin-like growth factor
binding protein 1 in human pregnancy. Rev Reprod 4,
160–167.

8 Gibson JM, Aplin JD, White A & Westwood M (2001)
Regulation of IGF bioavailability in pregnancy. Mol
Hum Reprod 7, 79–87.
9 Fowler DJ, Nicolaides KH & Miell JP (2000) Insulin-
like growth factor binding protein-1 (IGFBP-1): a
multifunctional role in the human female reproductive
tract. Hum Reprod 6, 495–504.
10 Tisi DK, Liu XJ, Wykes LJ, Skinner CD & Koski KG
(2005) Insulin-like growth factor II and binding proteins
1 and 3 from second trimester human amniotic fluid are
associated with infant birth weight. J Nutr 135,
1667–1672.
11 Irving JA & Lala PK (1995) Functional role of cell
surface integrins on human trophoblast cell migration:
regulation by TGF-beta, IGF-II, and IGFBP-1. Exp
Cell Res 217, 419–427.
12 Jones JI, Gockerman A, Busby WH Jr, Wright G &
Clemmons DR (1993) Insulin-like growth factor binding
protein 1 stimulates cell migration and binds to the
alpha 5 beta 1 integrin by means of its Arg-Gly-Asp
sequence. Proc Natl Acad Sci USA 90, 10553–10557.
13 Gockerman A, Prevette T, Jones JI & Clemmons DR
(1995) Insulin-like growth factor (IGF)-binding proteins
inhibit the smooth muscle cell migration responses to
IGF-I and IGF-II. Endocrinology 136, 4168–4173.
14 Gleeson LM, Chakraborty C, McKinnon T & Lala PK
(2001) Insulin-like growth factor-binding protein 1
stimulates human trophoblast migration by signaling
through alpha 5 beta 1 integrin via mitogen-activated
protein kinase pathway. J Clin Endocrinol Metab 86,

2484–2493.
15 Huang ST, Vo KC, Lyell DJ, Faessen GH, Tulac S,
Tibshirani R, Giaccia AJ & Giudice LC (2004) Devel-
opmental response to hypoxia. FASEB J 18, 1348–1365.
16 Kajimura S, Aida K & Duan C (2005) Insulin-like
growth factor-binding protein-1 (IGFBP-1) mediates
hypoxia-induced embryonic growth and developmental
retardation. Proc Natl Acad Sci USA
102, 1240–1245.
17 Sitar T, Popowiicz GM, Siwanowicz I, Huber R &
Holak TA (2006) Structural basis for the inhibition of
insulin-like growth factors by insulin-like growth factor-
binding proteins. Proc Natl Acad Sci USA 103, 13028–
13033.
18 Westwood M, Gibson J, Davies AJ, Young RJ & White
A (1994) The phosphorylation pattern of insulin-like
growth factor-binding protein-1 in normal plasma is
different from that in amniotic fluid and changes during
pregnancy. J Clin Endocrinol Metab 79, 1735–1741.
19 Weber MM, Spo
¨
ttl G, Go
¨
ssl C & Engelhardt D (1999)
Characterization of human insulin-like growth factor-
binding proteins by two-dimensional polyacrylamide gel
electrophoresis and Western ligand blot analysis. J Clin
Endocrinol Metab 84, 1679–1684.
20 Gianazza E, Wait R, Begum S, Eberini I, Campagnoli
M, Labo

`
S & Galliano M (2007) Mapping the 5–50
fraction of human amniotic fluid by 2-DE and ESI-MS.
Proteomics Clin Appl 1, 167–175.
21 Jones JI, D’Ercole JA, Camacho-Hubner C & Clem-
mons DR (1991) Phosphorylation of insulin-like growth
factor (IGF)-binding protein 1 in cell culture and in
vivo: effects on affinity for IGF-I. Proc Natl Acad Sci
USA 88, 7481–7485.
22 Jones JI, Busby WH Jr, Wright G, Smith CE, Kimack
NM & Clemmons DR (1993) Identification of the sites
L. Dolcini et al. Phosphorylation of amniotic fluid human IGFBP-1
FEBS Journal 276 (2009) 6033–6046 ª 2009 The Authors Journal compilation ª 2009 FEBS 6045
of phosphorylation in insulin-like growth factor binding
protein-1. Regulation of its affinity by phosphorylation
of serine 101. J Biol Chem 268, 1125–1131.
23 Seferovic MD, Ali R, Kamei H, Liu S, Khosravi JM,
Nazarian S, Han VK, Duan C & Gupta MB (2009)
Hypoxia and leucine deprivation induce human insulin-
like growth factor binding protein-1 hyperphosphoryla-
tion and increase its biological activity. Endocrinology
159, 220–231.
24 Nissum M, Abuhehab M, Sukop U, Khosravi JM, Wild-
gruber R, Eckerskorn C, Han VKM & Gupta MB (2009)
Functional and complementary phosphorylation state
attributes of human insulin-like growth factor binding
protein-1 (IGFBP-1) isoforms resolved by free flow
electrophoresis. Mol Cell Proteomics 8, 1424–1435.
25 Bankowski E, Sobolewski K, Palka J & Jaworski S
(2004) Decreased expression of the insulin-like growth

factor-I-binding protein-1 (IGFBP-1) phosphoisoform
in pre-eclamptic Wharton’s jelly and its role in the
regulation of collagen biosynthesis. Clin Chem Lab Med
42, 175–181.
26 Loukovaara M, Leinonen P, Teramo K, Nurminen E,
Andersson S & Rutanen EM (2005) Effect of maternal
diabetes on phosphorylation of insulin-like growth
factor binding protein-1 in cord serum. Diabet Med 22,
434–439.
27 Coppock HA, White A, Aplin JD & Westwood M
(2004) Matrix metalloprotease-3 and -9 proteolyze
insulin-like growth factor-binding protein-1. Biol Reprod
71, 438–443.
28 Brandt K, Wang J, Lundell K, Sta
˚
hlberg M, Horn HV,
Ehrenborg E, Hall K, Jo
¨
rnvall H & Lewitt M (2009)
IGFBP-1 protease activity and IGFBP-1 fragments in a
patient with multiple myeloma. Growth Horm IGF Res
doi:10.1016/j.ghir.2009.05.002.
29 Rechsteiner M & Rogers SW (1996) PEST sequences
and regulation by proteolysis. Trends Biochem Sci 21,
267–271.
30 Sala A, Capaldi S, Campagnoli M, Faggion B, Labo
`
S,
Perduca M, Romano A, Carrizo ME, Valli M, Visai L
et al. (2005) Structure and properties of the C-terminal

domain of insulin-like growth factor-binding protein-1
isolated from human amniotic fluid. J Biol Chem 280,
29812–29819.
31 Blom N, Gammeltoft S & Brunak S (1999) Sequence-
and structure-based prediction of eukaryotic protein
phosphorylation sites. J Mol Biol 294, 1351–1362.
32 Gibson BW & Cohen P (1990) Liquid secondary
ion mass spectrometry of phosphorylated and
sulfated peptides and proteins. Methods Enzymol
193, 480–501.
33 Kim J, Camp DG II & Smith RD (2004) Improved
detection of multi-phosphorylated peptides in the
presence of phosphoric acid in liquid chromatography ⁄
mass spectrometry. J Mass Spectrom 39, 208–215.
34 Salih E (2005) Phosphoproteomics by mass spectro-
metry and classical protein chemistry approaches. Mass
Spectrom Rev 24, 828–846.
35 Frost RA & Tseng L (1991) Insulin-like growth factor-
binding protein-1 is phosphorylated by cultured human
endometrial stromal cells and multiple protein kinases
in vitro. J Biol Chem 266, 18082–18088.
36 Temporini C, Dolcini L, Abee A, Calleri E, Galliano
M, Caccialanza G & Massolini G (2008) Development
of an integrated chromatographic system for on-line
digestion and characterization of phosphorylated
proteins. J Chromatogr A 1183, 65–75.
37 Kabir-Salmani M, Shimizu Y, Sakai K & Iwashita M
(2005) Posttranslational modifications of decidual
IGFBP-1 by steroid hormones in vitro. Mol Hum
Reprod 11, 667–671.

38 Irwin JC, Fuentes LD, Dsupin BA & Giudice LC
(1993) Insulin-like growth factor regulation of human
endometrial stromal cell function: coordinate effects on
insulin-like growth factor binding protein-1, cell prolif-
eration and prolactin secretion. Regul Pept 48, 165–177.
39 Bunn CR & Fowlkes JR (2003) Insulin-like growth
factor binding protein proteolysis. Trends Endocrinol
Metab 14, 176–181.
40 Nakamura M, Miyamoto S, Maeda H, Ishii G, Hasebe
T, Chiba T, Asaka M & Ochiai A (2005) Matrix meta-
lloproteinase-7 degrades all insulin-like growth factor
binding proteins and facilitates insulin-like growth
factor bioavailability. Biochem Biophys Res Commun
333, 1011–1016.
41 Binoux M, Lalou C, Lassarre C & Segovia B (1994)
Regulation of IGF bioavailability by IGFBP proteases.
In The insulin-like growth factors and their regulatory
proteins (Baxter RC, Gluckman PD & Rosenfeld RG
eds), pp. 217–226. Elsevier Science, Amsterdam.
42 Graham ME, Kilby DM, Firth SM, Robinson PJ &
Baxter RC (2007) The in vivo phosphorylation and
glycosylation of human insulin-like growth factor
binding protein-5. Mol Cell Proteomics 6, 1392–1405.
43 Hoech WG & Mukku VR (1994) Identification of the
major sites of phosphorylation in IGF binding protein-
3. J Cell Biochem 56, 262–273.
44 Zhang Z & Marshall AG (1998) A universal algorithm
for fast and automated charge state deconvolution of
electrospray mass-to-charge ratio spectra. J Am Soc
Mass Spectrom 9, 225–233.

45 Clauser KR, Baker PR & Burlingame AL (1999) Role
of accurate mass measurement (±10 ppm) in protein
identification strategies employing MS or MS ⁄ MS and
database searching. Anal Chem 71, 2871–2882.
46 Morodomi T, Ogata Y, Sasaguri Y, Morimatsu M &
Nagase H (1992) Purification and characterization of
matrix metalloproteinase 9 from U937 monocytic
leukaemia and HT1080 fibrosarcoma cells. Biochem J
285, 603–611.
Phosphorylation of amniotic fluid human IGFBP-1 L. Dolcini et al.
6046 FEBS Journal 276 (2009) 6033–6046 ª 2009 The Authors Journal compilation ª 2009 FEBS