Tài liệu Báo cáo khoa học: Identification of membrane-bound serine proteinase matriptase as processing enzyme of insulin-like growth factor binding protein-related protein-1 (IGFBP-rP1/angiomodulin/mac25) doc

Identification of membrane-bound serine proteinase
matriptase as processing enzyme of insulin-like growth
factor binding protein-related protein-1
(IGFBP-rP1/angiomodulin/mac25)
Sanjida Ahmed
1,2
, Xinlian Jin
1
, Motoki Yagi
1,2
, Chie Yasuda
1
, Yuichiro Sato
1,2
, Shouichi Higashi
1
,
Chen-Yong Lin
3
, Robert B. Dickson
3
and Kaoru Miyazaki
1,2
1 Division of Cell Biology, Kihara Institute for Biological Research, Yokohama City University, Japan
2 Graduate School of Integrated Sciences, Yokohama City University, Japan
3 Lombardi Comprehensive Cancer Center, Georgetown University Medical Center, Washington, DC, USA
Insulin-like growth factor (IGF) binding proteins
(IGFBPs) regulate cellular proliferation by modulating
the actions of insulin and IGFs [1,2]. Recent studies
have revealed a group of IGFBP-related proteins
(IGFBP-rPs), which have low affinity for IGFs ⁄

insulin and low structural homology to IGFBPs [3].
Keywords
angiomodulin; insulin-like growth factor;
insulin-like growth factor binding protein-
related protein-1; matriptase; proteolytic
processing
Correspondence
K. Miyazaki, Division of Cell Biology, Kihara
Institute for Biological Research, Yokohama
City University, 641–12 Maioka-cho,
Totsuka-ku, Yokohama 244–0813, Japan
Fax: +81 458201901
Tel: +81 458201905
E-mail:
(Received 25 August 2005, revised
15 November 2005, accepted 8 December
2005)
doi:10.1111/j.1742-4658.2005.05094.x
Insulin-like growth factor (IGF) binding protein-related protein-1
(IGFBP-rP1) modulates cellular adhesion and growth in an IGF ⁄ insulin-
dependent or independent manner. It also shows tumor-suppressive activity
in vivo. We recently found that a single-chain IGFB-rP1 is proteolytically
cleaved to a two-chain form by a trypsin-like, endogenous serine protein-
ase, changing its biological activities. In this study, we attempted to iden-
tify the IGFBP-rP1-processing enzyme. Of nine human cell lines tested,
seven cell lines secreted IGFBP-rP1 at high levels, and two of them, ovar-
ian clear cell adenocarcinoma (OVISE) and gastric carcinoma (MKN-45),
highly produced the cleaved IGFBP-rP1. Serine proteinase inhibitors effect-
ively blocked the IGFBP-rP1 cleavage in the OVISE cell culture. The con-
ditioned medium of OVISE cells did not cleave purified IGFBP-rP1, but

their membrane fraction had an IGFBP-rP1-cleaving activity. The mem-
brane fraction contained an 80-kDa gelatinolytic enzyme, which was
identified as the membrane-type serine proteinase matriptase (MT-SP1)
by immunoblotting. When the membrane fraction was separated by
SDS ⁄ PAGE, the IGFBP-rP1-cleaving activity comigrated with matriptase.
A soluble form of matriptase purified in an inhibitor-free form efficiently
cleaved IGFBP-rP1 at the same site as that found in a naturally cleaved
IGFBP-rP1. Furthermore, small interfering RNAs for matriptase efficiently
blocked both the matriptase expression and the cleavage of
IGBP-rP1 in OVISE cells. These results demonstrate that IGFBP-rP1 is
processed to the two-chain form by matriptase on the cell surface.
Abbreviations
AEBSF, 4-(2-aminoethyl)benzenesulfonyl fluoride hydrochloride; AGM, angiomodulin; FBS, fetal bovine serum; HAI-1, hepatocyte growth
factor activator inhibitor-1; HGF, hepatocyte growth factor; HLE, hepatocellular carcinoma; IGF, insulin-like growth factor; IGFBP, IGF-binding
protein; IGFPB-rP1, IGFBP-related protein-1; MMP, matrix metalloproteinase; MT-SP1, membrane-type serine proteinase matriptase; PSF,
prostacyclin-stimulating factor; TAF, tumor-derived cell adhesion factor; siRNA, small interfering RNA; tPA, tissue plasminogen activator uPA,
urokinase-type plasminogen activator.
FEBS Journal 273 (2006) 615–627 ª 2006 The Authors Journal compilation ª 2006 FEBS 615
Angiomodulin (AGM) was initially purified as a
tumor-derived cell adhesion factor (TAF) from human
bladder carcinoma cells [4,5]. Its cDNA was cloned
from human leptomeningial cells as mac25 [6] and from
human fibroblasts as prostacyclin-stimulating factor
(PSF) [7]. Since AGM has a relatively low structural
homology to IGFBPs, the name of IGFBP-related pro-
tein-1 (IGFBP-rP1) has recently been proposed for
AGM ⁄ mac25 ⁄ PSF. AGM ⁄ IGFBP-rP1 exerts a weak
cell adhesion activity through heparan sulfate proteog-
lycans on the cell surface [4,5,8] and stimulates cell
growth in culture medium containing insulin or IGFs

[9,10]. The IGFBP-rP1 mRNA is expressed in a wide
range of normal tissues including the heart, spleen,
ovary, small intestine and colon [11]. Immunohisto-
chemical analysis has shown that IGFBP-rP1 is highly
expressed in the blood vessels of various human cancer
tissues [5] and in invading tumor cells [12]. On the other
hand, other studies have shown that IGFBP-rP1 exhib-
its a tumor-suppressive activity when overexpressed in
cancer cells [13–15]. Thus, exact biological functions of
IGFBP-rP1 remain to be clarified.
Various extracellular proteinases regulate cellular
functions by degrading or processing protein sub-
strates including extracellular matrix proteins, growth
factors and cell surface proteins. For example, mat-
rix metalloproteinases (MMPs), such as membrane
type-1 matrix metalloproteinase (MT1-MMP), matri-
lysin (MMP-7) and gelatinases A ⁄ B (MMP-2 ⁄ 9), are
known to play important roles in the process of
tumor invasion and metastasis [16–18]. Serine pro-
teinases such as plasminogen activators, plasmin and
trypsin also contribute to expression of malignant
phenotypes in tumor cells [19,20]. Recently, consider-
able attention has been focused on the physiological
and pathological functions of a membrane-bound
serine proteinase, matriptase (MT-SP1) [21–23]. It is
well known that IGFBPs often undergo proteolytic
processing in various kinds of biological fluids such
as blood, synovial fluid and interstitial fluid, as well
as culture media [24,25]. Several types of proteinases,
such as pregnancy-associated plasma proteins [26,27],

prostate specific antigen [28] and MMP-3 [29], have
been reported to cleave IGFBPs. The proteolysis of
IGFBPs is thought to modulate the actions of IGFs
towards cells [29].
We recently found that IGFBP-rP1 is converted
from a single-chain form to a two-chain form by the
action of a trypsin-like serine proteinase [10]. The pro-
teolytic processing of IGFBP-rP1 greatly reduced its
insulin ⁄ IGF-dependent growth promoting activity but
enhanced its syndecan-1-mediated cell adhesion activity
[10]. We report here that the membrane-bound serine
proteinase matriptase is responsible for the processing
of IGFBP-rP1.
Results
Expression and processing of IGFBP-rP1
in various cell lines
We have reported that IGFBP-rP1 is proteolytically
converted to a two-chain form during purification [10].
The cleavage of IGFBP-rP1 leads to complete loss
of insulin ⁄ IGF-1-dependent cell growth-stimulatory
activity due to its loss of insulin ⁄ IGF-binding ability.
To examine whether the specific cleavage of IGFBP-
rP1 also occurs in cultured cell systems, we tested
expression and processing of IGFBP-rP1 in eight
human cancer cell lines and one immortalized epithe-
lial cell line (HEK293). When the conditioned media
were analyzed by immunoblotting after reducing
SDS ⁄ PAGE, seven of the nine cell lines tested were
found to secrete IGFBP-rP1 protein (Fig. 1). Among
them, OVISE ovarian adenocarcinoma cells and

MKN-45 gastric adenocarcinoma cells secreted signifi-
cant levels of the 25-kDa, cleaved form of IGFBP-rP1.
On nonreducing SDS ⁄ PAGE, IGFBP-rP1 in the two-
conditioned media was separated to a single band of
33 kDa (data not shown), indicating that the cleaved
IGFBP-rP1 was a two-chain form consisting of a
25 kDa chain and an 8 kDa chain [10]. These results
suggested that OVISE and MKN-45 cells expressed a
high level of proteinase(s) responsible for the process-
ing of IGFBP-rP1.
To determine the class of the IGFBP-rP1-cleaving
proteinase, OVISE cells were cultured in the presence
of various proteinase inhibitors and the processing of
IGFBP-rP1 was analyzed. As shown in Fig. 2, serine
proteinase inhibitors, 4-(2-aminoethyl)benzenesulfonyl
fluoride hydrochloride (AEBSF) and aprotinin, signifi-
cantly inhibited the production of the cleaved form of
IGFBP-rP1 (25 kDa), whereas a cysteine proteinase
inhibitor (leupeptin), an aspartic proteinase inhibitor
(pepstatin) and a metalloproteinase inhibitor (N-(R)-
(2-(hydroxyaminocarbonyl)methyl)-4-methylpentanoyl-l-
3-(2¢-naphthyl)alaninyl-l-alanine 2-aminoethyl amide;
TAPI-1) did not affect the processing. A commercially
available proteinase inhibitor mixture containing
AEBSF, aprotinin and some other inhibitors inhibited
the processing to the same extent as AEBSF alone.
As trypsin-type serine proteinases, but neither
chymotrypsin-type nor elastase-type, are susceptible to
the AEBSF inhibition, a trypsin-type serine proteinase
was thought to be responsible for the IGFBP-rP1

cleavage.
Processing of IGFBP-rP1 by matriptase S. Ahmed et al.
616 FEBS Journal 273 (2006) 615–627 ª 2006 The Authors Journal compilation ª 2006 FEBS
Analysis of serine proteinases secreted
by various cell lines
Trypsin-type serine proteinases present in the condi-
tioned media of the nine cell lines were analyzed by
gelatin zymography (Fig. 3A). To eliminate the activit-
ies of metalloproteinases, the renatured gel was first
treated with 1 mm EDTA and then incubated in the
presence of 10 mm CaCl
2
, as described in Experimental
procedures. Of the nine cell lines tested, OVISE,
Fig. 1. Analysis of noncleaved and cleaved
forms of IGFBP-rP1 in conditioned media of
eight cancer cell lines and one immortalized
cell line (HEK293). Concentrated conditioned
media were prepared from the cultures of
the nine indicated human cell lines, and
each sample containing the same amount of
protein (5 lg) was subjected to SDS ⁄ PAGE
under reducing conditions on a 14% gel,
followed by immunoblotting with the anti-
TAF ⁄ IGFBP-rP1 antibody (upper panel). Bars
indicate the noncleaved form (33 kDa) and
the cleaved form (25 kDa). The cleaved form
is detected highly in OVISE and MKN-45 cell
lines and slightly in HLE and HEK293 cell
lines. As a loading control, the same vol-

umes of conditioned media as those for the
immunoblotting were subjected to SDS ⁄
PAGE followed by protein staining with
Coomassie Brilliant Blue R-250 (lower
panel). Bars indicate the molecular size in
kDa. Other experimental conditions and the
types of the cell lines used are described in
‘Experimental procedures’.
Fig. 2. Effects of proteinase inhibitors on
processing of IGFBP-rP1 in culture of OVISE
cells. OVISE ovarian carcinoma cells were
incubated in a serum-free culture medium
supplemented without (None) or with one
of the following proteinase inhibitors: AE-
BSF (100 l
M), aprotinin (75 nM), leupeptin
(10 l
M), pepstatin A (1 lM), TAPI-1 (2 lM),
and a mixture (100 l
M AEBSF, 75 nM
aprotinin, 5 lM bestatin, 1.5 lM E-64, 2 lM
leupeptin, and 1 lM pepstatin A). After
incubation for 2 days, IGFBP-rP1 in each
conditioned medium was analyzed by
immunoblotting, as described in Fig. 1. The
upper panel shows immunoblots for
matriptase. Bars indicate the noncleaved
form (33 kDa) and the cleaved form
(25 kDa) of IGFBP-rP1. The lower panel
shows protein staining patterns of

conditioned media. Bars indicate the
molecular size in kDa.
S. Ahmed et al. Processing of IGFBP-rP1 by matriptase
FEBS Journal 273 (2006) 615–627 ª 2006 The Authors Journal compilation ª 2006 FEBS 617
MKN-45 and DLD-1 cells commonly expressed a
major gelatinolytic activity at approximately 75 kDa
and additional weak activities at 95 kDa, 105 kDa and
lower molecular weight positions. The 75 kDa activity
was faintly detected in HSC-4 and HEK293 cells.
To date, several families of type II transmembrane
serine proteinases have been identified [30]. Among
them, MT-SP1 is known to be expressed in many types
of epithelial cell lines and carcinoma cell lines, and to
be proteolytically released from the cell membranes
[21,31,32]. The molecular size of a major form of sol-
uble matriptase is approximately 75 kDa. To deter-
mine whether or not the 75 kDa proteinase in Fig. 3A
was matriptase, the conditioned media of the nine cell
lines were subjected to immunoblotting with an anti-
matriptase antibody (M32). As shown in Fig. 3B, the
conditioned media of OVISE, MKN-45 and DLD-1
cells clearly showed doublet bands at approximately
75 kDa and two minor bands at 95 and 110 kDa. The
conditioned media of HSC-4 and HEK293 cells also
showed weak immunoreactive bands at the same posi-
tions. These results attributed the gelatinolytic activit-
ies at 75, 95 and 105 kDa in Fig. 3A to matriptase.
The 75 kDa doublet has been reported to result from
glycosylation [33]. The glycosylation is thought to
increase the stability of this protein against trypsin and

other proteinases [34].
To further characterize the gelatinolytic activities in
the OVISE cell conditioned medium, we examined
effect of a serine proteinase inhibitor, AEBSF, on the
proteinase activities (Fig. 4). The sample was treated
with 2 mm AEBSF before and ⁄ or after SDS ⁄ PAGE
and subsequent renaturation of separated proteins.
The pretreatment of the sample with the inhibitor par-
tially blocked only the 75 kDa activity, whereas the
inhibitor treatment after the protein renaturation
strongly blocked the activities at 75, 95 and 110 kDa.
This indicated that only a part of the 75 kDa enzyme
existed in an inhibitor-free active form in the concen-
trated conditioned medium. A weak activity of 40 kDa
was not inhibited by the inhibitor, suggesting that it
might be a metalloproteinase. It has been reported that
110
95
75
EJ-1
ECV-304
HT1080
OVISE
MKN-45
HLE
HSC-4
DLD-1
HEK293
110
95

75
EJ-1
ECV-304
HT1080
OVISE
MKN-45
HLE
HSC-4
DLD-1
HEK293
A
B
Fig. 3. Analysis of serine proteinases secre-
ted by nine human cell lines. Conditioned
media were prepared from the cultures of
the indicated, nine human cell lines. The
concentrated conditioned media containing
the same amount of protein (5 lg) were
subjected to gelatin zymography (A) and
immunoblotting with the antimatriptase
antibody M32 (B) as described in ‘Experi-
mental procedures’. In the gelatin zymogra-
phy, metalloproteinase activities were
eliminated by incubating the renatured gel
with 1 m
M EDTA for 30 min. (A) Gelatin
zymograms of the conditioned media from
the nine cell lines. The conditioned media of
OVISE and MKN-45, which showed high
IGFBP-rP1 cleavage in Fig. 1, showed a

major activity at 75 kDa and minor activities
at 95 and 110 kDa, which were indicated by
bars. (B) Immunoblots for matriptase. The
75 kDa major band seems to correspond
mainly to a single-chain, latent matriptase,
while the 95 and 110 kDa bands seem to
correspond to two-chain, active forms
complexed with HAI-1 [18,32].
Processing of IGFBP-rP1 by matriptase S. Ahmed et al.
618 FEBS Journal 273 (2006) 615–627 ª 2006 The Authors Journal compilation ª 2006 FEBS
membrane-bound matriptase is shed into culture med-
ium in some different forms [21]. The soluble matrip-
tase includes a 70 kDa single-chain, latent form as a
major component, a two-chain, active enzyme com-
plexed with two different sizes of hepatocyte growth
factor (HGF) activator inhibitor (HAI-1) (total
molecular sizes of 95 and 110 kDa), and a trace of an
inhibitor-free active enzyme [21,31,32]. The 95 and
110 kDa matriptase ⁄ HAI-1 complexes are expected
not to react with AEBSF, because their active serine
residues are bound to HAI-1. Based on these facts,
the 95- and 110-kDa bands in zymography (Figs 3A
and 4) and immunoblotting (Fig. 3B) are most likely
to correspond to the matriptase ⁄ HAI-1 complexes.
The 95- and 110-kDa gelatinolytic activities seemed to
result from partial dissociation of an active matriptase
from its inhibitor HAI-1 after SDS ⁄ PAGE. Similarly,
the 75-kDa activity that was not inhibited by the pre-
treatment with AEBSF was thought to be an active
matriptase, which had been dissociated from HAI-1 by

the SDS treatment. In addition, immunoblotting under
reducing conditions suggested that the majority of the
75-kDa immunoreactive band was a single-chain latent
matriptase (data not shown) [32].
The correlation between the IGFBP-rP1-cleaving
activity (Fig. 1) and the expression of matriptase in
MKN-45 and OVISE cell lines suggested matriptase as
a candidate for the IGFBP-rP1-processing enzyme.
Moreover, it has been reported that matriptase prefers
basic P1, P3 and P4 residues to cleave substrates [33].
This consensus sequence is found in the cleavage site
sequence of IGFBP-rP1, where Lys97, Lys95 and
Arg94 residues are located at P4, P3 and P1 sites,
respectively [10]. These facts prompted us to examine
whether or not matriptase is the IGFBP-rP1-processing
enzyme. However, when purified IGFBP-rP1 was
incubated with the conditioned medium of OVISE or
DLD-1 cells, the latter of which contained the highest
amount of matriptase, its processing to a two-chain
form was not observed (data not shown).
IGFBP-rP1-cleaving activity of membrane fraction
from OVISE cells
As IGFBP-rP1 is cleaved in cultured OVISE cells, we
next examined the possibility that cell-associated serine
proteinase(s) might cleave IGFBP-rP1. A membrane
fraction of OVISE cells was prepared as described
under ‘Experimental procedures’ and incubated with
purified IGFBP-rP1. Figure 5A (lanes 1 and 2) shows
that the cleaved form of IGFBP-rP1 significantly
increased during the incubation. This cleavage was

inhibited effectively by aprotinin, suggesting that a
serine proteinase(s) catalyzed this processing (Fig. 5A,
lane 3). We further analyzed serine proteinase activit-
ies in the membrane fraction of OVISE cells by gelatin
zymography. This analysis revealed the presence of a
gelatinolytic enzyme, with an approximate molecular
size of 80 kDa, in the membrane fraction (Fig. 5B,
lane 1). When analyzed on the same gel, the apparent
molecular size of the gelatinolytic activity in the
membrane was slightly higher than the 75 kDa activity
found in the conditioned medium (data not shown).
Immunoblotting with the antimatriptase antibody
showed an immunoreactive band at the same position
as the gelatinolytic activity (Fig. 5B, lane 2), suggest-
ing that the 80-kDa serine proteinase was a mem-
brane-bound matriptase.
To verify that the 80 kDa proteinase has the
IGFBP-rP1-cleaving activity, we separated the proteins
in the membrane fraction of OVISE cells by
SDS ⁄ PAGE. After electrophoresis, the proteins on the
gel were renatured and the gel was horizontally divided
into eight equal parts. Each piece of the gel was then
incubated with purified IGFBP-rP1 and the cleavage
of the protein was analyzed as described under Experi-
mental procedures. As shown in Fig. 6A, gel fractions
Fig. 4. Effect of serine proteinase inhibitor AEBSF on gelatinolytic
activities of OVISE cell conditioned medium. AEBSF was added to
make a final concentration of 2 m
M into the OVISE cell conditioned
medium before SDS ⁄ PAGE (lanes 2 and 4) and ⁄ or into the reaction

buffer after the SDS ⁄ PAGE and subsequent renaturation (lanes 3
and 4). –, No addition, +, addition. Ordinate indicates the molecular
size in kDa. Arrowheads show the matriptase bands. Other experi-
mental conditions are described in ‘Experimental procedures’ and
in the legends to Fig. 3A.
S. Ahmed et al. Processing of IGFBP-rP1 by matriptase
FEBS Journal 273 (2006) 615–627 ª 2006 The Authors Journal compilation ª 2006 FEBS 619
4–6 had the IGFBP-rP1-cleaving activity and fraction
6 showed the highest activity. When the proteins in
the gel pieces were analyzed by immunoblotting, an
80 kDa matriptase was also detected in fraction 6
(Fig. 6B), indicating that this membrane-bound serine
proteinase is the prime candidate for the IGFBP-rP1-
cleaving proteinase in OVISE cells.
IGFBP-rP1-cleaving activity of purified matriptase
The results of the SDS ⁄ PAGE separation of the mem-
brane-bound enzyme suggested that the soluble form
of matriptase might cleave IGFBP-rP1 if separated
from inhibitors. To test this possibility, the condi-
tioned medium of DLD-1 cells, which contained the
highest activity of matriptase, was separated by
SDS ⁄ PAGE and assayed for the IGFBP-rP1-cleaving
80
B
80
1 2
A
123
33
25

Fig. 5. Analysis of IGFBP-rP1-cleaving enzyme present in mem-
brane fraction of OVISE cells. A membrane fraction was prepared
from OVISE cells as described in ‘Experimental procedures’ and
used for the following assays. (A) IGFBP-rP1-cleaving activity. Puri-
fied IGFBP-rP1 (500 ng) was incubated with the membrane fraction
(20 lg protein) in the presence (lane 3) or absence (lane 2) of
75 n
M aprotinin (serine proteinase inhibitor). The incubated samples
were subjected to immunoblotting under reducing conditions with
the anti-TAF ⁄ IGFBP-rP1 antibody. Lane 1, no incubation. Bars indi-
cate the noncleaved form (33 kDa) and the cleaved form (25 kDa).
(B) Gelatin zymography (lane 1) and immunoblotting with the anti-
matriptase antibody M32. The membrane fraction containing 15 lg
protein was run on a gelatin-containing gel under nonreducing con-
ditions for the zymography (lane 1), and the same sample contain-
ing 10 lg protein was subjected to immunoblotting with the
antimatriptase antibody M32. The bar indicates a gelatinolytic band
at approximately 80 kDa in lane 1 and a matriptase band at almost
the same position in lane 2. Other experimental conditions are des-
cribed in ‘Experimental procedures’.
Fig. 6. Fractionation of IGFBP-rP1-cleaving enzyme and matriptase
present in OVISE cell membrane by SDS ⁄ PAGE. The membrane
fraction (20 lg proteinÆlane
)1
) of OVISE cells was separated by non-
reducing SDS ⁄ PAGE on two lanes of a 7.5% gel. After the gel was
washed with 1% Triton X-100 for protein renaturation, each of the
two lanes was divided into eight fractions and used for one of the
two following assays. (A) IGFBP-rP1-cleaving activity. Each gel frac-
tion was incubated with IGFBP-rP1 (500 ng) at 4 °C for 24 h fol-

lowed by 37 °C for 24 h. The incubated samples were subjected to
immunoblotting under reducing conditions with the anti-TAF ⁄
IGFBP-rP1 antibody. (None), IGFBP-rP1 incubated without gel frac-
tion. Bars indicate the noncleaved form (33 kDa) and the cleaved
form (25 kDa). Fraction 6 shows the highest activity. (B) Fraction-
ation of matriptase. Gel fractions from another lane were used to
detect matriptase. Each gel fraction was extracted with the SDS
sample buffer and applied to SDS ⁄ PAGE under nonreducing condi-
tions on a 10% gel and immunoblotted with the antimatriptase anti-
body M32. An arrow indicates an immunoreactive band of
matriptase. MF, the membrane fraction before separation. Other
experimental conditions are described in ‘Experimental procedures’.
Processing of IGFBP-rP1 by matriptase S. Ahmed et al.
620 FEBS Journal 273 (2006) 615–627 ª 2006 The Authors Journal compilation ª 2006 FEBS
activity. The IGFBP-rP1-cleaving activity was detected
at fractions corresponding to approximately 75 and
95 kDa, which also showed immunoreactive bands to
the antimatriptase antibody in immunoblotting (data
not shown). A similar result was also obtained when
the conditioned medium of OVISE cells was separated
by SDS ⁄ PAGE (data not shown). These results
strongly suggested that soluble forms of matriptase are
able to cleave IGFBP-rP1 in the absence of HAI-1 or
other inhibitors.
To further confirm this possibility, we purified an
inhibitor-free, soluble form of matriptase from the
conditioned medium of DLD-1 cells, as described in
‘Experimental procedures’. The purified matriptase
preparation contained the 75-kDa matriptase and a
few minor proteins, as analyzed by SDS ⁄ PAGE

(Fig. 7A). Gelatin zymography under nonreducing
conditions confirmed that the 75 kDa matriptase had a
proteolytic activity (Fig. 7B). Immunoblotting analysis
under nonreducing conditions showed a single immuno-
reactive band for matriptase at 75 kDa, but under
reducing conditions it was split into a major 75-kDa
band of the single-chain, latent enzyme and a minor
50 kDa band of the two-chain, active enzyme (Fig. 7C)
[10]. Based on the relative band intensity, the percent-
age of the active enzyme to the total matriptase was
estimated to be approximately 30%.
We previously reported that trypsin cleaves IGFBP-
rP1 at the same site as an endogenous processing
enzyme [10]. The IGFBP-rP1-cleaving activities of
matriptase and trypsin were compared at varied con-
centrations (Fig. 8). Fifty nanograms of the 33 kDa
IGFBP-rP1 was almost completely cleaved to the
25 kDa form by 5 ng of the total matriptase, which
contained the active enzyme as a minor component.
On the other hand, 10 ng of trypsin converted a major
part of the 33 kDa IGFBP-rP1 to the 25 kDa form,
but an increased amount (50 ng) of trypsin nonspecifi-
cally degraded both forms of IGFBP-rP1. These results
indicated that the two-chain, active matriptase has a
much higher IGFBP-rP1-cleaving activity than trypsin.
Furthermore, we tried to identify the cleavage site of
IGFBP-rP1 by the purified matriptase. The 25 kDa
form of IGFBP-rP1 obtained by the treatment with
matriptase was applied to an automated protein
sequencer. The N-terminal amino acid sequence of the

25 kDa band of IGFBP-rP1 was determined to be
A
98
GAAAGGPG
106
, suggesting that this protein had
been cleaved between K(Lys)
97
and A(Ala)
98
. This
cleavage site was identical to that previously deter-
mined for a naturally cleaved IGFBP-rP1.
All these results demonstrate that the soluble form
of active matriptase cleaves IGFBP-rP1 but HAI-1
or some other inhibitors block its activity in culture
medium.
Effects of matriptase siRNAs on IGFBP-rP1
cleavage in OVISE cells
To show that matriptase is an endogenous IGFBP-
rP1-processing enzyme in OVISE cells, we designed
three siRNAs for matriptase and examined their effects
on the IGFBP-rP1 cleavage. As negative controls,
OVISE cells were treated with a scrambled RNA or
with the lipofectamine reagent alone. Although the
scrambled RNA had some cytotoxic effect, there was
no significant difference in the profiles of secreted pro-
teins among the control and siRNA-treated cultures
(Fig. 9A). Immunoblotting with the antimatriptase
antibody showed that all of the three siRNAs

decreased the amount of soluble matriptase to less
than 20% of the control levels (Fig. 9B). When the
IGFBP-rP1 secreted into culture medium by OVISE
cells was analyzed by immunoblotting, contrasting
patterns of the secreted IGFBP-rP1 were obtained
A
B
C
Fig. 7. Electrophoretic analyses of purified matriptase. Soluble mat-
riptase was purified from the conditioned medium of DLD-1 cells
as described in the text. The purified matriptase (approximately
200 ng of total proteins) was subjected to the following analyses.
(A) SDS ⁄ PAGE under nonreducing conditions followed by silver
staining. Arrowhead indicates the matriptase band. Bars indicate
the molecular size in kDa. (B) Gelatin zymography. Arrowhead indi-
cates the gelatinolytic activity by matriptase. (C) Immunoblotting
with the antimatriptase antibody M32 under nonreducing conditions
(lane 1) and reducing conditions (lane 2). In lane 2, the 75-kDa band
corresponds to the single-chain, latent enzyme, while the 50 kDa
band corresponds to the heavy chain of the two-chain, active
enzyme. Other experimental conditions are described in ‘Experi-
mental procedures’.
S. Ahmed et al. Processing of IGFBP-rP1 by matriptase
FEBS Journal 273 (2006) 615–627 ª 2006 The Authors Journal compilation ª 2006 FEBS 621
between the control and siRNA-treated cultures
(Fig. 9C). In the two control cultures, the 33-kDa,
uncleaved IGFBP-rP1 was faintly detected compared
with the 25-kDa, cleaved form, whereas in the siRNA-
treated cultures, the uncleaved IGFBP-rP1 was a
major component. This indicated that the cleavage of

IGFBP-rP1 was effectively reduced by the siRNA
treatment. These results confirmed that matriptase acts
as an IGFBP-rP1-processing enzyme in the culture of
OVISE cells.
Discussion
In this study, we first identified the type II membrane-
bound serine proteinase matriptase as a processing
enzyme of IGFBP-rP1 ⁄ AGM. Matriptase was initially
Fig. 8. Cleavage of IGFBP-rP1 by purified matriptase and trypsin. (A) Matriptase. IGFBP-rP1 (50 ng) was incubated with the indicated
amounts of purified matriptase at 37 °C for 5 h in 10 lL of a reaction mixture containing 20 m
M Tris ⁄ HCl (pH 7.5), 0.1 M NaCl and 10 mM
CaCl
2
. The original concentration of matriptase was determined for the total 75-kDa protein based on the band intensity relative to that of
bovine serum albumin as standard (Fig. 7A). (B) Trypsin. IGFBP-rP1 was incubated with the indicated amounts of TPCK-trypsin under the
same conditions except for the absence of CaCl
2
in the reaction mixture. The proteolytic cleavage of IGFBP-rP1 in (A) and (B) was analyzed
by immunoblotting as described in Fig. 1. Bars indicate the 33 kDa, uncleaved form and the 25 kDa, cleaved form of IGFBP-rP1.
AB C
Fig. 9. Effects of matriptase siRNAs on IGFBP-rP1 processing in culture of OVISE cells. OVISE cells at 50–60% confluence in 60 mm culture
dishes were transfected with 200 pmol of each of three siRNAs (si973, si1513 and si2578) using Lipofectamine 2000 reagent. As negative
controls, the cells were treated with a scrambled RNA (sc) or with the lipofectamine reagent alone (Cont.). These cultures were incubated in
serum-containing medium overnight, and then in serum-free medium for 2 days. The resultant conditioned media were collected and con-
centrated. The concentrated samples containing 10 lg protein were subjected to reducing SDS ⁄ PAGE followed by the Coomassie Brilliant
Blue staining (A), nonreducing immunoblotting to detect matriptase (B), and reducing immunoblotting to detect IGFBP-rP1 (C). Bars indicate
the molecular sizes of marker proteins in (A), the 110-, 95- and 75-kDa bands of matriptase in (B), and the 33-kDa, uncleaved form and the
25-kDa, cleaved form of IGFBP-rP1 in (C). The scrambled RNA-treated culture (2nd lane) was low in the total band intensity of matriptase (B)
and IGFBP-rP1 (C) as compared with the control or siRNA-treated cultures. This seemed due to the cytotoxic effect of the scrambled RNA.
Other experimental conditions are described in Experimental procedures.

Processing of IGFBP-rP1 by matriptase S. Ahmed et al.
622 FEBS Journal 273 (2006) 615–627 ª 2006 The Authors Journal compilation ª 2006 FEBS
found as a trypsin-like serine proteinase secreted by
human breast cancer cells [35] and later purified as a
complex with its natural inhibitor HAI-1 from human
milk [21]. Matriptase is expressed in normal epithelial
tissues such as the skin, stomach, colon, kidney, breast,
ovary and pancreas, but not in mesenchyma [36,37]. A
recent study with matriptase-deficient mice has shown
that matriptase plays critical roles in the epidermal
barrier function, hair follicle development and thymic
homeostasis [38]. Matriptase is also expressed in cancer
tissues of the breast, ovary, uterus and colon and also
by some mammary and ovarian carcinoma cell lines
in vitro [36]. Since matriptase is able to degrade extra-
cellular matrix proteins and to activate urokinase-
type plasminogen activator (uPA) [33,39], hepatocyte
growth factor (HGF) [39] and protease-activated
receptor-2 (PAR-2) [33], it is expected to play some
roles in the growth, invasion and metastasis of human
carcinoma cells.
We previously reported that IGFBP-rP1 is cleaved
to a two-chain form by a trypsin-like serine proteinase
[10]. In the present study, we found that OVISE ovar-
ian carcinoma cells cleaved endogenous IGFBP-rP1,
and their membrane fraction cleaved exogenous
IGFBP-rP1 in a cell-free solution. The IGFBP-rP1-
cleaving activity in the OVISE cell membrane comi-
grated with membrane-bound matriptase on SDS ⁄
PAGE. Furthermore, the treatment of OVISE cells

with matriptase siRNAs efficiently blocked both mat-
riptase expression and the cleavage of IGFBP-rP1. It
is also noted that the cleavage sequence of IGFBP-rP1
is consistent with the most preferable sequence in sub-
strate proteins of matriptase [33]. Indeed, a soluble
form of matriptase purified in an inhibitor-free form
efficiently cleaved IGFBP-rP1 at the same site as that
found in a naturally cleaved IGFBP-rP1. All these
facts indicate that the membrane-bound matriptase is
a natural processing enzyme of IGFBP-rP1. On the
other hand, soluble forms of matriptase, which are
released from cell membranes by proteolysis, were
detected in the conditioned media of at least 5 cell
lines out of 9 cell lines tested. The processing of endo-
genous IGFBP-rP1 was correlated with the amount of
soluble matriptase in the conditioned media. Although
the purified soluble matriptase could cleave IGFBP-
rP1, the conditioned media of OVISE and DLD-1 cells
did not show the IGFBP-rP1-processing activity unless
they were separated by SDS ⁄ PAGE. Our recent analy-
sis has detected soluble matriptase in 19 of 24 human
carcinoma cell lines tested, which included carcinomas
of the breast, lung, stomach and colon [32]. The sol-
uble matriptase mostly existed in a single-chain, latent
form as a major component and two-chain forms
complexed with its inhibitor HAI-1, in agreement with
the past reports [18,30]. Therefore, soluble matriptase
released from cell membrane is expected to have a very
low, if any, proteolytic activity. The IGFBP-rP1-clea-
ving activities of the SDS ⁄ PAGE fractions and the

matriptase purified from DLD-1 conditioned medium
indicate that the soluble form of activated matriptase
can cleave IGFBP-rP1, but its activity is masked by
HAI-1 in culture medium. It was recently reported that
the matriptase zymogen might be auto-activated by
interacting with HAI-1 on the cell surface [40]. The
proteolytic action of matriptase, including the process-
ing of IGFBP-rP1, seems to be restricted to the cell
surface and its close vicinity.
Although the present study demonstrates that mat-
riptase cleaves IGFBP-rP1, our data do not exclude
the possibility that other proteinases, especially serine
proteinases, also cleave IGFBP-rP1. We previously
reported that many human cancer cell lines secrete an
active or latent form of trypsin and a 75-kDa serine
proteinase [20], the latter of which was identified as
matriptase in a recent study [32]. We have also repor-
ted that OVISE cells secrete uPA, but neither trypsin
nor tissue plasminogen activator (tPA) [41]. In addi-
tion, it was previously found that trypsin cleaves
IGFBP-rP1 to the same two-chain form as that found
in conditioned media [10]. In the present study, we
examined whether or not uPA and tPA cleave IGFBP-
rP1 in test tubes, but the IGFBP-rP1 cleavage was seen
with neither uPA nor tPA (data not shown). There-
fore, matriptase seems to be a major IGFBP-rP1-pro-
cessing enzyme, at least in OVISE cells, and possibly
in MKN-45 cells. Recent studies have revealed the
presence of four families of type II transmembrane ser-
ine proteinase [30]. The matriptase subfamily consti-

tutes three members (matriptase, matriptase 2 and
matriptase 3). It is conceivable that some of these mem-
brane-bound serine proteinases or their soluble forms
are also involved in the processing of IGFBP-rP1.
It is well known that some IGFBPs undergo proteo-
lytic cleavage [24,25]. In biological fluids, IGFBPs bind
IGFs with high affinity, protecting the growth factors
from proteolytic degradation. The proteolytic cleavage
of IGFBPs is thought to contribute to the release of
IGFs from IGF ⁄ IGFBP complexes to interact with
IGF receptors on the cell surface. This mechanism
may not be directly applicable to the case of IGFBP-
rP1, because IGFBP-rP1 has a far lower affinity for
IGFs than IGFBPs. The high affinity binding of IGFs
to IGFBPs limits the interaction of the growth factors
with the cell surface receptors. Therefore, IGFBPs gen-
erally inhibit IGF-stimulated cell growth in vitro [25].
In contrast, IGFBP-rP1 stimulates cell growth in the
S. Ahmed et al. Processing of IGFBP-rP1 by matriptase
FEBS Journal 273 (2006) 615–627 ª 2006 The Authors Journal compilation ª 2006 FEBS 623
presence of insulin or IGF-1 in vitro, presumably due
to its low affinity for the factors [9,10]. On the other
hand, some studies have shown that IGFBP-rP1 exhib-
its a tumor-suppressive activity when overexpressed in
cancer cells [13–15]. The apparent discrepancy between
the in vitro and in vivo studies is not clearly explained.
IGFBP-rP1 also shows IGF ⁄ insulin-independent activ-
ities. It has affinity for heparin, type IV collagen and
syndecan-1 [5,8,10]. These activities seem to be respon-
sible for the cell adhesion activity of IGFBP-rP1

in vitro and its dense deposition on the basement mem-
brane of blood vessels in tumor tissues [5]. These
IGF ⁄ insulin-dependent and independent activities of
IGFBP-rP1 are notably altered by its proteolytic clea-
vage [10]. For example, IGFBP-rP1 loses its IGF ⁄
insulin-binding activity and IGF ⁄ insulin-dependent
growth-stimulating activity but acquires high cell adhe-
sion activity by proteolytic cleavage. It seems conceiv-
able that matriptase regulates tumor growth by
modulating biological activities of IGFBP-rP1, as well
as other growth-regulating proteins, in vivo.
In this study, we identified a new substrate of mat-
riptase. Matriptase may exert a broad range of func-
tions in regulating cellular growth, apoptosis and
differentiation by degrading or processing a variety of
extracellular proteins.
Experimental procedures
Materials
The sources of materials used are as follows: aprotinin, leu-
peptin, pepstatin A, AEBSF and proteinase inhibitor mixture
that contains AEBSF, aprotinin, bestatin, E-64, leupeptin
and pepstatin A from Wako Pure Chemical Industries
(Osaka, Japan); gelatin from Difco (Detroit, MI, USA);
N-(R)-(2-(hydroxyaminocarbonyl)methyl)-4-methylpentanoyl-
l-3-(2¢-naphthyl)alaninyl-l-alanine 2-aminoethyl amide
(TAPI-1) from Peptide Institute (Osaka, Japan). IGFBP-rP1
was purified from the conditioned medium of the human
bladder carcinoma cell line EJ-1, as described previously [10].
An anti-TAF ⁄ IGFBP-rP1 monoclonal antibody (#88) [5]
and an antimatriptase monoclonal antibody (M32) [21] were

raised against purified IGFBP-rP1 and purified matriptase,
respectively. All other chemicals were of analytical grade or
the highest quality commercially available.
Cell cultures and preparation of conditioned
medium
Types of human cancer cell lines used are as follows: HSC-
4, tongue squamous cell carcinoma; HT1080, fibrosarcoma;
EJ-1 and ECV-304, bladder carcinomas; DLD-1, colon
adenocarcinoma; OVISE, ovarian clear cell adenocarcino-
ma; MKN-45, adenosquamous carcinoma of the stomach;
HLE, hepatocellular carcinoma. The source and properties
of OVISE cells were described before [41]. The human
embryonic kidney cell line HEK293 (ATCC CRL-1573)
was purchased from American Type Culture Collection
(ATCC, Rockville, MD, USA). The other cell lines were
obtained from Japanese Cancer Resources Bank (JCRB) in
National Institute of Biomedical Innovation (Osaka,
Japan). To prepare conditioned medium, each cell line was
grown to semiconfluence in 90-mm culture dishes contain-
ing a 1 : 1 mixture of Dulbecco’s modified Eagles medium
and Ham’s F12 medium (Gibco; Grand Island, NY, USA),
DME ⁄ F12, supplemented with 10% fetal calf serum (FCS).
The cells were rinsed three times with serum-free
DME ⁄ F12, and the culture was further continued in the
presence or absence of various proteinase inhibitors in
serum-free DME ⁄ F12. After incubation for 2 days, the
resultant conditioned medium was collected, clarified by
centrifugation and dialyzed against distilled water at 4 °C.
The dialyzed sample was then lyophilized and dissolved in
a 100th volume of 10 mm Tris ⁄ HCl (pH 7.5) to the original

conditioned medium.
Preparation of membrane fractions
OVISE cells were grown to confluence in the serum-contain-
ing medium, rinsed three times with ice cold NaCl ⁄ P
i
and
then scraped in the presence of 20 mm Hepes (pH 7.5) con-
taining 250 m m sucrose at 4 °C. The cell suspension was then
homogenized with a Dounce homogenizer. The homogenate
was centrifuged at 1500 g for 7 min to remove the nuclei.
The postnuclear supernatant was further centrifuged at
50 000 g for 30 min. The resultant pellet was then dissolved
in 20 mm Tris ⁄ HCl (pH 7.5) containing 0.5 m KCl, 0.15 m
NaCl and 1% Triton X-100, and the insoluble substances
were removed by centrifugation at 14 000 g for 20 min. The
supernatant was used as a crude membrane fraction.
SDS–polyacrylamide gel electrophoresis
(SDS/PAGE) and immunoblotting
SDS ⁄ PAGE was performed on 7.5, 10, or 14% polyacryl-
amide gel slabs (85 mm wide, 1 mm thick, and 70 mm long)
under reducing or nonreducing conditions. Separated pro-
teins were stained with silver. Immunoblotting was per-
formed as described previously [42]. Briefly, proteins on
gels were transferred onto nitrocellulose membranes (Schlei-
cher & Schuell, Keene, NH, USA). The membrane was
blocked with skimmed milk and successively treated with
an anti-TAF ⁄ IGFBP-rP1 monoclonal antibody (#88) or an
antimatriptase monoclonal antibody (M32) as the first anti-
body, the second antibody (biotinylated antimouse IgG,
Vector Laboratory, Burlingham, CA, USA), and alkaline

Processing of IGFBP-rP1 by matriptase S. Ahmed et al.
624 FEBS Journal 273 (2006) 615–627 ª 2006 The Authors Journal compilation ª 2006 FEBS
phosphatase-coupled avidin (Vector Laboratory). Immuno-
reactive signals were visualized by the enhanced chemilumi-
nescence (ECL) detection method (Amersham Biosciences,
Piscataway, NJ, USA).
Gelatin zymography
Zymography was carried out on 10% polyacrylamide gels
containing 1 mgÆmL
)1
of gelatin, as described previously
[20]. Briefly, after SDS ⁄ PAGE, the gel was incubated in
50 mm Tris ⁄ HCl (pH 7.5) buffer containing 100 mm NaCl
and 2.5% Triton X-100 at room temperature for 1.5 h to
renature the proteins on the gel. After the incubation, the
gel was treated with 50 mm Tris ⁄ HCl (pH 7.5) containing
1mm EDTA at 37 °C for 30 min to inactivate metallopro-
teinases. After the treatment, the gel was further incubated
in 50 mm Tris ⁄ HCl (pH 7.5) containing 10 mm CaCl
2
at
37 °C for 18 h. Gelatinolytic bands were visualized by
staining the gel with Coomassie Brilliant Blue R-250.
Assay of IGFBP-rP1-cleaving activity of
membrane fraction and conditioned medium
Purified IGFBP-rP1 (500 ng) was incubated with a mem-
brane fraction (20 lg protein) or concentrated conditioned
medium (20 lg protein) of OVISE or DLD-1 cells in 40 lL
of 50 mm Tris ⁄ HCl (pH 7.5) containing 0.15 m NaCl and
10 mm CaCl

2
first at 4 °C for 24 h and then at 37 °C for
24 h. IGFBP-rP1 in the reaction mixture was precipitated
with cold 10% trichloroacetic acid, washed with cold eth-
anol, and then applied to SDS ⁄ PAGE on a 14% gel, fol-
lowed by immunoblotting with the anti-TAF ⁄ IGFBP-rP1
antibody #88.
Assay of IGFBP-rP1-cleaving activity after
SDS/PAGE separation of proteinases
IGFBP-rP1-cleaving proteinases present in the membrane
fraction of OVISE cells and conditioned medium of DLD-1
cells were separated by SDS ⁄ PAGE on 7.5% gels and
assayed as follows. Each sample containing 20 lg protein
was applied to two lanes on a gel. After electrophoresis, the
gel was incubated in the renaturation buffer described
above for 1.5 h and then washed with pure water for
15 min. Each lane (5 mm wide, 1 mm thick, and 70 mm
long) of the gel was divided horizontally into eight or 16
equal parts. For the assay of IGFBP-rP1 cleavage, each gel
piece from one lane was incubated with purified IGFBP-
rP1 (500 ng protein) in 40 lLof50mm Tris ⁄ HCl (pH 7.5)
containing 0.15 m NaCl and 10 mm CaCl
2
first at 4 °C for
24 h, and then at 37 °C for 24 h. The resultant IGFBP-rP1
fragments were precipitated by 10% trichloroacetic acid
and analyzed by immunoblotting with the anti-TAF ⁄
IGFBP-rP1 antibody #88 as described above. To detect
matriptase, gel pieces from another lane were individually
subjected to repeated freeze-thawing in 20 lL of three-fold

concentrated SDS sample buffer and then incubated at
room temperature for 6 h. Each extract was subjected to
nonreducing SDS ⁄ PAGE on a 10% gel followed by immu-
noblotting with the antimatriptase antibody M32.
Purification of soluble matriptase
Serum-free conditioned medium of DLD-1 cells was col-
lected, concentrated by ammonium sulfate precipitation,
and subjected to molecular-sieve chromatography on a
Cellullofine GCL-2000 m column (Seikagaku Kogyo,
Tokyo, Japan) pre-equilibrated with 20 mm Tris ⁄ HCl
(pH 7.5) buffer containing 0.5 m NaCl, 0.1% CHAPS
and 0.01% Brij35. Fractions containing matriptase were
pooled, dialyzed against 20 mm Tris ⁄ HCl (pH 7.5) con-
taining 0.01% Brij35 and applied to a Reactive Red
agarose column (Sigma-Aldrich, St. Louis, MO, USA).
Proteins bound to the column were sequentially eluted
with 0.2 m and 0.4 m NaCl in the buffer. The 0.4 m
NaCl fraction was dialyzed against 20 mm Tris ⁄ HCl
(pH 8.0) and applied to a Mono Q HR5 ⁄ 5 column
(Amersham Biosciences) pre-equibrated with the same
buffer. Proteins bound to the column were eluted with a
linear gradient of 0–0.5 m NaCl. Matriptase eluted from
the column was finally applied to reverse-phase HPLC on
a TSKgel TMS-250 column (Tosoh, Tokyo, Japan) in the
presence of 0.1% trifluoroacetic acid (TFA) and eluted
with a linear gradient of 0–80% acetonitrile in 0.1%
trifluoroacetic acid. Matriptase was dissociated and separ-
ated from HAI-1 in this HPLC. The matriptase-contain-
ing fractions were concentrated by centrifugation under
vacuum and then freeze-dried. The dried materials were

dissolved in 20 mm Tris ⁄ HCl (pH 7.5) and used for the
activity assay. The main fraction contained a 75 kDa
matriptase as a major component and a few contamin-
ating proteins as analyzed by SDS ⁄ PAGE.
RNAi experiments with OVISE cells
Matriptase siRNAs and a scrambled RNA as a control
were designed and synthesized at iGENE (Tsukuba, Japan).
The forward sequences of the siRNAs were: #973, sense
5¢-UCAUCACACUGAUAACCAACACUGA-AG-3¢;#2578,
sense 5¢-GGAUCAAAGAGAACACUGGGGUAUA-AG-3¢;
and #1513 sense, 5¢-AGUUCACGUGCAAGAACAAG
UUCUG-AG-3¢. The forward sequence of the scrambled
RNA was 5¢-GAUCCAAGUAAUACAGAGAUGGGAG
AG-3¢. OVISE cells were plated the day before transfection
at a cell density of 50–60% saturation in 60 mm culture
dishes. The cells were transfected with 200 pmol siRNA
using Lipofectamine 2000 reagent according to the manu-
facturer protocol (Invitrogen, Carlsbad, CA, USA). The
S. Ahmed et al. Processing of IGFBP-rP1 by matriptase
FEBS Journal 273 (2006) 615–627 ª 2006 The Authors Journal compilation ª 2006 FEBS 625
transfected cells were incubated in serum containing med-
ium overnight, and further incubated in serum-free medium
for 2 days. The resultant serum-free conditioned medium
was collected from each culture as described above, and
matriptase and IGFBP-rP1 in the conditioned medium were
analyzed by immunoblotting.
Acknowledgements
We thank K. Yamamoto and N. Akiyama for techni-
cal assistance, and Drs H. Yasumitsu, Y. Tsubota and
Y. Kariya for helpful suggestions and discussions. This

work was supported by Grants-in-Aid from the Minis-
try of Education, Culture, Sports, Science and Tech-
nology and from the Ministry of Welfare and Labor
of Japan.
References
1 Jones JI & Clemmons DR (1995) Insulin-like growth
factors and their binding proteins: biological actions.
Endocrinol Rev 16, 3–34.
2 Rechler MM (1993) Insulin-like growth factor binding
proteins. Vitam Horm Rev 47, 1–114.
3 Kim HS, Nagalla SR, Oh Y, Wilson E, Roberts CT Jr
& Rosenfeld RG (1997) Identification of a family of
low-affinity insulin-like growth factor binding proteins
(IGFBPs): characterization of connective tissue growth
factor as a member of the IGFBP superfamily. Proc
Natl Acad Sci USA 94, 12981–12986.
4 Akaogi K, Okabe Y, Funahashi K, Yoshitake Y, Nish-
ikawa K, Yasumitsu H, Umeda M & Miyazaki K
(1994) Cell adhesion activity of a 30 kDa major secreted
protein from human bladder carcinoma cells. Biochem
Biophys Res Commun 198, 1046–1053.
5 Akaogi K, Okabe Y, Sato J, Nagashima Y, Yasumitsu H,
Sugahara K & Miyazaki K (1996) Specific accumulation
of tumor-derived adhesion factor in tumor blood vessels
and in capillary tube-like structures of cultured vascular
endothelial cells. Proc Natl Acad Sci USA 93, 8384–8389.
6 Murphy M, Pykett MJ, Harnish P, Zang KD & George
DL (1993) Identification and characterization of genes
differentially expressed in meningiomas. Cell Growth
Differ 4, 715–722.

7 Yamauchi T, Umeda F, Masakado M, Isaji M, Mizu-
shima S & Nawata H (1994) Purification and molecular
cloning of prostacyclin-stimulating factor from serum-
free conditioned medium of human diploid fibroblast
cells. Biochem J 303, 591–598.
8 Sato J, Hasegawa S, Akaogi K, Yasumitsu H, Yamada
S, Sugahara K & Miyazaki K (1999) Identification of
cell-binding site of angiomodulin (AGM ⁄ TAF ⁄ Mac 25)
that interacts with heparan sulfates on cell surface.
J Cell Biochem 75, 187–195.
9 Akaogi K, Sato J, Okabe Y, Sakamoto Y, Yasumitsu H
& Miyazaki K (1996) Synergistic growth stimulation of
mouse fibroblasts by tumor-derived adhesion factor with
insulin-like growth factors and insulin. Cell Growth
Differ 7, 1671–1677.
10 Ahmed S, Yamamoto K, Sato Y, Ogawa T, Herrmann
A, Higashi S & Miyazaki K (2003) Proteolytic proces-
sing of IGFBP-related protein-1 (TAF ⁄ angiomodu-
lin ⁄ mac25) modulates its biological activity. Biochem
Biophys Res Commun 310, 612–618.
11 Oh Y, Nagalla SR, Yamanaka Y, Kim HS, Wilson E &
Rosenfeld RG (1996) Synthesis and characterization of
insulin-like growth factor-binding protein (IGFBP)-7.
J Biol Chem 271, 30322–30325.
12 Adachi Y, Itoh F, Yamamoto H, Arimura Y, Kikkawa-
Okabe Y, Miyazaki K, Carbone DP & Imai K (2001)
Expression of angiomodulin (tumor-derived adhesion
factor ⁄ Mac25) in invading tumor cells correlates with
poor prognosis in human colorectal cancer. Int J Cancer
95, 216–222.

13 Sprenger CC, Vail ME, Evans K, Simurdak J & Ply-
mate SR (2001) Over-expression of insulin-like growth
factor binding protein-related protein-1 (IGFBP-
rP1 ⁄ mac25) in the M12 prostate cancer cell line alters
tumor growth by a delay in G1 and cyclin A associated
apoptosis. Oncogene 21, 140–147.
14 Mutaguchi K, Yasumoto H, Mita K, Matsubara A,
Shiina H, Igawa M, Dahiya R & Usui T (2003)
Restoration of insulin-like growth factor binding pro-
tein-related protein 1 has a tumor-suppressive activity
through induction of apoptosis in human prostate can-
cer. Cancer Res 63, 7717–7723.
15 Wilson HM, Birnbaum RS, Poot M, Quinn LS & Swis-
shelm K (2002) Insulin-like growth factor binding pro-
tein-related protein 1 inhibits proliferation of MCF-7
breast cancer cells via a senescence-like mechanism. Cell
Growth Differ 13, 205–213.
16 Nelson AR, Fingleton B, Rothenberg ML & Matri-
sian LM (2000) Matrix metalloproteinases: biologic
activity and clinical implications. J Clin Oncol 18,
1135–1149.
17 Egeblad M & Werb Z (2002) New functions for the
matrix metalloproteinases in cancer progression. Nat
Rev Cancer 2, 161–174.
18 Seiki M & Yana I (2003) Roles of pericellular proteolysis
by membrane type-1 matrix metalloproteinase in cancer
invasion and angiogenesis. Cancer Sci 94, 569–574.
19 Blasi F & Verde P (1990) Urokinase-dependent cell sur-
face proteolysis and cancer. Semin Cancer Biol 1, 117–
126.

20 Miyata S, Koshikawa N, Higashi S, Miyagi Y,
Nagashima Y, Yanoma S, Kato Y, Yasumitsu H &
Miyazaki K (1999) Expression of trypsin in human
cancer cell lines and cancer tissues and its tight binding
Processing of IGFBP-rP1 by matriptase S. Ahmed et al.
626 FEBS Journal 273 (2006) 615–627 ª 2006 The Authors Journal compilation ª 2006 FEBS
to soluble form of Alzheimer amyloid precursor protein
in culture. J Biochem (Tokyo) 125, 1067–1076.
21 Lin CY, Anders J, Johnson M & Dickson RB (1999)
Purification and characterization of a complex con-
taining matriptase and a Kunitz-type serine protease
inhibitor from human milk. J Biol Chem 274, 18237–
18242.
22 Lin CY, Anders J, Johnson M, Sang QA & Dickson
RB (1999) Molecular cloning of cDNA for matriptase,
a matrix-degrading serine protease with trypsin-like
activity. J Biol Chem 274, 18231–18236.
23 Takeuchi T, Shuman MA & Craik CS (1999) Reverse bio-
chemistry: use of macromolecular protease inhibitors to
dissect complex biological processes and identify a mem-
brane-type serine protease in epithelial cancer and normal
tissue. Proc Natl Acad Sci USA 96, 11054–11061.
24 Maile LA & Holly JM (1999) Insulin-like growth binding
protein (IGFBP) proteolysis: occurrence, identification,
role and regulation. Growth Horm IGF Res 9, 85–95.
25 Collett-Solberg PF & Cohen P (1996) The role of the
insulin-like growth factor binding proteins and the
IGFBP proteases in modulating IGF action. Endocrinol
Metab Clin North Am 25, 591–614.
26 Lawrence JB, Oxvig C, Overgaard MT, Sottrup-Jensen L,

Gleich GJ, Hays LG, Yates JR & Conover CA (1999) The
insulin-like growth factor (IGF)-dependent IGF binding
protein-4 protease secreted by human fibroblasts is
pregnancy-associated plasma protein-A. Proc Natl Acad
Sci USA 96, 3149–3153.
27 Overgaard MT, Boldt HB, Laursen LS, Sottrup-Jensen L,
Conover CA & Oxvig C (2001) Pregnancy-associated
plasma protein-A2 (PAPP-A2), a novel insulin-like
growth factor-binding protein-5 proteinase. J Biol Chem
276, 21849–21853.
28 Cohen P, Peehl DM, Graves HC & Rosenfeld RG
(1994) Biological effects of prostate specific antigen as
an insulin-like growth factor binding protein-3 protease.
J Endocrinol 142, 407–415.
29 Fowlkes JL, Serra DM, Bunn RC, Thrailkill KM,
Enghild JJ & Nagase H (2003) Regulation of insulin-
like growth factor-1 (IGF-1) action by matrix metallo-
proteinase-3 (MMP-3) involves selective disruption of
IGF-1 ⁄ IGF-binding protein-3 (IGFBP-3) complexes.
Endocrinology 145, 620–626.
30 Szabo R, Wu Q, Dickson RB, Netzel-Avnett S, Antalis
TM & Bugge TH (2003) Type II transmembrane serine
proteases. Thromb Haemost 90, 185–193.
31 Benaud C, Dickson RB & Lin CY (2001) Regulation of
the activity of matriptase on epithelial cell surfaces by a
blood-derived factor. Eur J Biochem 268, 1439–1447.
32 Jin X, Hirosaki T, Lin C-Y, Dickson RB, Higashi S,
Kitamura H & Miyazaki K (2005) Production of
soluble matriptase by human cancer cell lines and cell
surface activation of its zymogen by trypsin. J Cell

Biochem 95, 632–647.
33 Takeuchi T, Harris JL, Huang W, Yan KW, Coughlin
SR & Craik CS (2000) Cellular localization of
membrane-type serine protease 1 and identification of
protease-activated receptor-2 and single-chain
urokinase-type plasminogen activator as substrates.
J Biol Chem 275, 26333–26342.
34 Ihara S, Miyoshi E, Nakahara S, Sakiyama H, Ihara H,
Akinaga A, Honke K, Dickson RB, Lin CY & Tanigu-
chi N (2004) Addition of beta1–6 GlcNAc branching to
the oligosaccharide attached to Asn 772 in the serine
protease domain of matriptase plays a pivotal role in its
stability and resistance against trypsin. Glycobiology 14,
139–146.
35 Shi YE, Torri J, Yieh L, Wellstein A, Lippman ME &
Dickson RB (1993) Identification and characterization
of a novel matrix-degrading protease from hormone-
dependent human breast cancer cells. Cancer Res 53,
1409–1415.
36 Oberst MD, Anders J, Xie B, Singh B, Ossandon M,
Johnson M, Dickson RB & Lin CY (2001) Matriptase
and HAI-1 are expressed by normal and malignant
epithelial cells in vitro and in vivo. Am J Pathol 158,
1301–1311.
37 Oberst MD, Singh B, Ozdemirli M, Dickson RB, John-
son MD & Lin CY (2003) Characterization of matrip-
tase expression in normal human tissues. J Histochem
Cytochem 51, 1017–1025.
38 List K, Haudenschild CC, Szabo R, Chen W, Wahl
SM, Swaim W, Engelholm LH, Behrendt N & Bugge

TH (2002) Matriptase ⁄ MT-SP1 is required for postnatal
survival, epidermal barrier function, hair follicle devel-
opment, and thymic homeostasis. Oncogene 21, 3765–
3779.
39 Lee SL, Dickson RB & Lin CY (2000) Activation of
hepatocyte growth factor and urokinase ⁄ plasminogen
activator by matriptase, an epithelial membrane serine
protease. J Biol Chem 275, 36720–36725.
40 Oberst MD, Williams CA, Dickson RB, Johnson MD
& Lin CY (2003) The activation of matriptase
requires its noncatalytic domains, serine protease
domain, and its cognate inhibitor. J Biol Chem 29,
26773–26779.
41 Miyagi E, Yasumitsu H, Hirahara F, Minaguchi H,
Koshikawa N, Miyazaki K & Umeda M (1995)
Characterization of matrix-degrading proteinases and
their inhibitors secreted by human gynecological
carcinoma cells. Jpn J Cancer Res 86, 568–576.
42 Higashi S & Miyazaki K (2003) Identification of a
region of b-amyloid precursor protein essential for its
gelatinase A-inhibitory activity. J Biol Chem 278,
14020–14028.
S. Ahmed et al. Processing of IGFBP-rP1 by matriptase
FEBS Journal 273 (2006) 615–627 ª 2006 The Authors Journal compilation ª 2006 FEBS 627