Báo cáo khoa học: Identification and characterization of plasma kallikrein–kinin system inhibitors from salivary glands of the blood-sucking insect Triatoma infestans pptx

Identification and characterization of plasma
kallikrein–kinin system inhibitors from salivary glands
of the blood-sucking insect Triatoma infestans
Haruhiko Isawa
1,2
, Yuki Orito
2
, Naruhiro Jingushi
2
, Siroh Iwanaga
3
, Akihiro Morita
2
, Yasuo Chinzei
2
and Masao Yuda
2
1 Department of Medical Entomology, National Institute of Infectious Diseases, Tokyo, Japan
2 Department of Medical Zoology, School of Medicine, Mie University, Japan
3 Laboratory of Chemistry and Utilization of Animal Resources, Faculty of Agriculture, Kobe University, Japan
The plasma kallikrein (EC 3.4.21.34)–kinin system
plays an important role in the initiation and amplifi-
cation of surface-mediated, acute inflammatory
responses following tissue injury [1–4]. This system is
composed of three serine protease zymogens [prekal-
likrein (PK), factor XII (FXII) (EC 3.4.21.38) and
factor XI] and the nonenzymatic procofactor, high
molecular weight kininogen (HK). Kallikrein–kinin
system activation is initiated by binding of FXII
and a PK–HK complex to a biological activating
surface, such as an endothelial cell surface, and is

then accelerated by the reciprocal activation of FXII
and PK on the surface. Zn
2+
is essential for binding
of FXII and HK to a biological activating surface,
and induces their conformational changes [5–11].
Activation of the kallikrein–kinin system results in
the release of bradykinin, a primary mediator of
acute inflammatory responses [3,4,12]. Bradykinin
causes vasodilation, increases microvascular perme-
ability, and enhances pain sensitivity, resulting in
inflammatory symptoms such as redness, edema and
pain around the injured site. Activated FXII (FXIIa)
Keywords
factor XII; high molecular weight kininogen;
kallikrein–kinin system; salivary gland;
Triatoma infestans
Correspondence
H. Isawa, Department of Medical
Entomology, National Institute of Infectious
Diseases, Toyama 1-23-1, Sinjuku-ku,
Tokyo 162-8640, Japan
Fax: +81 3 5285 1147
Tel: +81 3 5285 1111
E-mail:
(Received 20 February 2007, revised
18 June 2007, accepted 27 June 2007)
doi:10.1111/j.1742-4658.2007.05958.x
Two plasma kallikrein–kinin system inhibitors in the salivary glands of the
kissing bug Triatoma infestans, designated triafestin-1 and triafestin-2, have

been identified and characterized. Reconstitution experiments showed that
triafestin-1 and triafestin-2 inhibit the activation of the kallikrein–kinin sys-
tem by inhibiting the reciprocal activation of factor XII and prekallikrein,
and subsequent release of bradykinin. Binding analyses showed that triafes-
tin-1 and triafestin-2 specifically interact with factor XII and high mole-
cular weight kininogen in a Zn
2+
-dependent manner, suggesting that they
specifically recognize Zn
2+
-induced conformational changes in factor XII
and high molecular weight kininogen. Triafestin-1 and triafestin-2 also inhi-
bit factor XII and high molecular weight kininogen binding to negatively
charged surfaces. Furthermore, they interact with both the N-terminus of
factor XII and domain D5 of high molecular weight kininogen, which are
the binding domains for biological activating surfaces. These results suggest
that triafestin-1 and triafestin-2 inhibit activation of the kallikrein–kinin
system by interfering with the association of factor XII and high molecular
weight kininogen with biological activating surfaces, resulting in the inhibi-
tion of bradykinin release in an animal host during insect blood-feeding.
Abbreviations
APTT, activated partial thromboplastin time; DS 500, dextran sulfate of M
r
500 000; FXII, factor XII; FXIIa, activated FXII; HK, high molecular
weight kininogen; HKa, two-chain HK; PK, prekallikrein; PT, prothrombin time; SPR, surface plasmon resonance; RU, resonance unit;
PtdInsP, phosphatidyl inositol phosphate; S-2302, H-
D-Pro-L-Phe-L-Arg-p-nitroanilide.
FEBS Journal 274 (2007) 4271–4286 ª 2007 The Authors Journal compilation ª 2007 FEBS 4271
converts factor XI to factor XIa, which in turn causes
activation of the intrinsic coagulation pathway. Glass,

kaolin, dextran sulfate, sulfatide and acidic phospho-
lipids are negatively charged surfaces that can activate
the kallikrein–kinin system in vitro. Recent studies
indicate that there is a multiprotein receptor complex
on endothelial cells that functions as the physiologic
surface receptor for FXII and HK activation [13–16]:
this complex consists of gC1qR, urokinase plasmino-
gen activator receptor, and cytokeratin 1 [17–19]. It
has also been shown that in mice, FXII contributes
to pathologic thrombus formation through the intrin-
sic pathway [20,21].
Blood-sucking arthropods have several physiologi-
cally active molecules in their saliva, such as an anti-
coagulant, inhibitor of platelet aggregation, and
vasodilator [22,23]. These molecules are injected into
host animals and act to assist arthropod blood-feed-
ing. We have identified and characterized the first
reported kallikrein–kinin system inhibitor, hamadarin,
from salivary glands of the malaria vector mosquito
Anopheles stephensi [24]. Hamadarin belongs to the
D7 family of proteins, which are widely present in
the saliva of mosquitoes [25,26]. We suggested that
hamadarin may attenuate the host’s acute inflamma-
tory responses to the bite by inhibition of bradykinin
release, thereby enabling a mosquito to take a blood
meal efficiently and safely. However, it remains
unclear whether kallikrein–kinin system inhibitors are
widely distributed in saliva of other blood-sucking
arthropods and whether they prevent the release of
bradykinin during blood-feeding by a mechanism sim-

ilar to that of hamadarin.
In this study, we report two potent inhibitors of
the plasma kallikrein–kinin system in salivary glands
of the kissing bug Triatoma infestans, which is
known to be a vector of Chagas’ disease in South
America. These inhibitors, designated triafestin-1 and
triafestin-2, are major components of saliva, and
were classified in the lipocalin family on the basis of
amino acid sequence similarities. We showed that
triafestin-1 and triafestin-2 inhibit the reciprocal acti-
vation of FXII and PK and subsequent generation
of bradykinin. These results suggested that the inhib-
itory activities are produced through interference
with the binding of FXII and HK to activating sur-
faces. Furthermore, triafestin-1 and triafestin-2 specif-
ically interacted with the N-terminal region of FXII
and domain D5 of HK, which are the interactive
sites of FXII and HK for activating surfaces.
The biological significance of kallikrein–kinin system
inhibition for the blood-sucking arthropods is
discussed.
Results
cDNA cloning and expression of recombinant
triafestin-1 and triafestin-2
A cDNA library was constructed from the salivary
glands of unfed fifth instar nymphs of T. infestans.
cDNA fragments from 550 clones were randomly
selected from this library and sequenced as described
previously [27]. Sequence similarity searches of all
cDNA clones were performed using the blast pro-

gram. To screen for saliva proteins, signal peptide
prediction was carried out with the encoded proteins
using the signalp program. In total, 173 cDNA frag-
ments encoded the predicted secreted proteins. Of
these, 127 fragments (73.4%) had sequence similarities
to salivary gland lipocalins [27], such as triabin
(antithrombin) [28] and pallidipin (antiplatelet) [29]
identified from T. pallidipennis (Fig. 1A). Of these
lipocalin-like T. infestans fragments, four had identical
sequences and were designated clone Ti369, and six
had identical sequences and were designated clone
Ti263. Clones Ti369 and Ti263 have about 90%
amino acid sequence identity and are predicted to
have the same 18 amino acid signal peptide, suggest-
ing that they belong to the same protein group, desig-
nated ‘triafestin’. Therefore, the gene products of
clones Ti369 and Ti263 were designated triafestin-1
and triafestin-2, respectively. To study the phylogeny
of these triafestins, the amino acid sequences of tria-
festins and five salivary lipocalins of triatomine bugs
were aligned by the clustalw protocol. The align-
ment data were tested by the bootstrap method, and
a neighbor-joining tree was constructed (Fig. 1B).
Triafestins were most closely related to pallidipin and
then to Rhodnius platelet aggregation inhibitor-1 [30].
Nitrophorin 1 (NO-carrying protein) [31] and prolix-
in-S (antifactor IX ⁄ IXa) [32] were distantly related to
triafestins.
To investigate the bioactivities of these triafestins,
recombinant proteins were produced in a baculo-

virus–insect cell system. Secreted recombinant proteins
formed a major fraction of the proteins in the cell
culture medium, and were purified by a combination
of ion exchange and gel filtration chromatography.
Analysis by reducing SDS ⁄ PAGE showed that the
apparent molecular masses of recombinant triafestin-1
and triafestin-2 were approximately 25–26 kDa, which
is slightly larger than calculated from the sequence
data (21.8 kDa) (Fig. 2). Antiserum to either triafes-
tin-1 or triafestin-2 cross-reacted with the other pro-
tein, probably due to their high amino acid sequence
similarity. In western blot analysis, these antisera
Plasma kallikrein–kinin system inhibitors H. Isawa et al.
4272 FEBS Journal 274 (2007) 4271–4286 ª 2007 The Authors Journal compilation ª 2007 FEBS
reacted with two major salivary gland proteins, which
migrated with the same mobilities as purified recom-
binant triafestins on two-dimensional gel electrophore-
sis (data not shown). These results indicate that
triafestin-1 and triafestin-2 are abundant in the saliva
of T. infestans.
Triafestin inhibits the plasma kallikrein–kinin
system
Kissing bugs in the subfamily Triatominae have anti-
coagulant(s) in their saliva [33,34]. Therefore, we
examined the activated partial thromboplastin time
(APTT) and prothrombin time (PT) prolongation
activities of triafestin-1 and triafestin-2 using human
plasma. As shown in Fig. 3, both triafestin-1 and triaf-
estin-2 increased APTT clotting time in a dose-depen-
dent manner, but showed no effect on PT clotting

time. Intrinsic tenase [factor VIIIa–factor IXa–phospo-
lipid–Ca
2+
complex] is involved in both intrinsic and
extrinsic coagulation pathways in vivo. However, some
intrinsic tenase inhibitors do not prolong PT clotting
time [35]. Therefore, we next investigated the inhibi-
tory activity of triafestin-1 and triafestin-2 towards the
Fig. 1. (A) Amino acid sequence alignment of triafestin-1 and triafestin-2 with five other salivary lipocalins of triatomine bugs. The underlines,
asterisks and periods indicate signal sequences, strongly conserved residues, and weakly conserved residues, respectively. The conserved
cysteine residues are indicated in bold type. (B) Dendrogram showing the phylogenetic relationships of triafestin-1, triafestin-2 and five other
salivary lipocalins of triatomine bugs based on the amino acid sequence similarities. The tree was constructed using the neighbor-joining
method, and bootstrap values correspond to 500 replications. The bar shows the amino acid sequence distance.
H. Isawa et al. Plasma kallikrein–kinin system inhibitors
FEBS Journal 274 (2007) 4271–4286 ª 2007 The Authors Journal compilation ª 2007 FEBS 4273
intrinsic tenase. Assays with triafestins showed that
both triafestin-1 and triafestin-2, even when added in
excess, could not inhibit the intrinsic tenase (data not
shown), suggesting that they are specific inhibitors of
the intrinsic coagulation pathway. We then examined
the anticoagulation activity of triafestin-1 and triafes-
tin-2 using human plasma pretreated with APTT
reagent in the absence of Ca
2+
. In these experiments,
the kallikrein–kinin system had been activated before
the addition of triafestin-1 or triafestin-2. In the
absence of Ca
2+
, factor XIa was generated, but fac-

tor IX was not activated, thereby preventing the down-
stream reactions. The addition of Ca
2+
allowed the
cascade to proceed, but neither triafestin-1 nor triafes-
tin-2 prolonged APTT, showing that they could not
inhibit the reactions following the kallikrein–kinin sys-
tem in plasma (data not shown). Therefore, we suggest
that both triafestin-1 and triafestin-2 inhibit activation
of the plasma kallikrein–kinin system, leading to inhi-
bition of the intrinsic coagulation pathway.
Triafestin inhibits FXII and PK reciprocal
activation and bradykinin generation
To investigate inhibition of the kallikrein–kinin system
by triafestin-1 and triafestin-2, the effects of triafestin-1
and triafestin-2 on the amidolytic activities of plasma
serine proteases in the kallikrein–kinin system (FXIIa,
kallikrein, and factor XIa) were examined using chro-
mogenic substrates. Neither triafestin-1 nor triafestin-2
inhibited the amidolytic activities of any of these
plasma proteins (data not shown). Therefore, we exam-
ined the effects of triafestin-1 and triafestin-2 on
activation of the kallikrein–kinin system in several
different reconstitution assays [36,37]. In the first
assay, the effects of triafestin-1 and triafestin-2 on the
reciprocal activation of FXII and PK were studied. As
shown in Fig. 4A, both triafestin-1 and triafestin-2
inhibited FXII and PK reciprocal activation in a dose-
dependent manner. We then examined the effects of
triafestin-1 and triafestin-2 on kallikrein-catalyzed acti-

vation of FXII and FXIIa-catalyzed activation of PK.
As shown in Fig. 4B,C, triafestins inhibited both reac-
tions in a dose-dependent manner. These results sug-
gested that both triafestin-1 and triafestin-2 inhibited
kallikrein–kinin system activation by inhibiting the
reciprocal activation of FXII and PK without affecting
their amidolytic activities. Activation of the kallikrein–
kinin system results in the generation of bradykinin, a
potent proinflammatory and pain-producing nonapep-
tide that is released from HK after cleavage by kallik-
rein. Therefore, we examined whether triafestin-1 and
triafestin-2 could attenuate the generation of bradyki-
nin in a reconstitution system. FXII was preincubated
with triafestin-1 or triafestin-2, reciprocal activation
was started by addition of PK, HK and a negatively
ABC
Fig. 2. SDS ⁄ PAGE and western blot analysis of a T. infestans sali-
vary gland extract and purified recombinant triafestin-1 and triafes-
tin-2. A crude salivary gland extract (lane 1), 2 lg of recombinant
triafestin-1 (lane 2) and 2 lg of recombinant triafestin-2 (lane 3)
were separated on a 15% gel under reducing conditions. Proteins
were stained with Coomassie Brilliant Blue (A) or detected with
antibody against recombinant triafestin-1 (B) or antibody against
recombinant triafestin-2 (C), respectively. M, molecular mass stan-
dards.
Fig. 3. Inhibition of intrinsic coagulation by recombinant triafestin-1
and triafestin-2. The inhibitory activities of triafestin-1 or triafestin-2
on the intrinsic and extrinsic coagulation pathways were estimated
by APTT and PT assays, respectively. Citrated human plasma was
incubated with increasing concentrations of triafestin-1 or triafestin-

2, and the mixture was activated with diluted APTT assay reagent
or PT assay reagent. Plasma clotting times were recorded with a
coagulometer. Results are presented as the mean ± SD of triplicate
determinations.
Plasma kallikrein–kinin system inhibitors H. Isawa et al.
4274 FEBS Journal 274 (2007) 4271–4286 ª 2007 The Authors Journal compilation ª 2007 FEBS
charged surface, and bradykinin generation was
assayed by ELISA. As shown in Fig. 4D, both triafes-
tin-1 and triafestin-2 reduced bradykinin generation in
a dose-dependent manner. This result indicated that
not only could triafestins inhibit the intrinsic coagula-
tion pathway, but they could also inhibit bradykinin
generation by interfering with activation of the kallik-
rein–kinin system.
Fig. 4. Inhibitory effects of triafestin-1 (filled bars) and triafestin-2 (unfilled bars) on surface-mediated reactions involving FXII and on bradyki-
nin release from HK. (A) Effect of triafestin-1 and triafestin-2 on reciprocal activation of FXII and PK. FXII was preincubated with various con-
centrations of triafestin-1 or triafestin-2 for 10 min. The activation reaction was started by addition of PK and DS 500. The total activity of
generated FXIIa and kallikrein was measured using chromogenic substrate S-2302. (B) Effect of triafestin-1 and triafestin-2 on kallikrein-cata-
lyzed activation of FXII. FXII was preincubated with various concentrations of triafestin-1 or triafestin-2 for 10 min. The activation reaction
was started by addition of kallikrein and DS 500. The generated FXIIa activity was measured using S-2302 with soybean trypsin inhibitor, a
potent inhibitor of kallikrein, to minimize kallikrein activity. (C) Effect of triafestin-1 and triafestin-2 on FXIIa-catalyzed activation of PK. FXIIa
was preincubated with various concentrations of triafestin-1 or triafestin-2 for 10 min. The activation reaction was started by addition of PK
and DS 500. The generated kallikrein activity was measured using S-2302 with corn trypsin inhibitor, a potent inhibitor of FXIIa, to minimize
FXIIa activity. (D) Effect of triafestin-1 and triafestin-2 on bradykinin release. FXII was preincubated with various concentrations of triafestin-1
or triafestin-2 for 10 min. Activation of the reconstituted kallikrein–kinin system was started by addition of PK, HK and DS 500, and bradyki-
nin generation was assayed by competitive ELISA. A control containing no DS 500 (NC) was included for comparison. Results are presented
as the mean ± SD of triplicate determinations.
H. Isawa et al. Plasma kallikrein–kinin system inhibitors
FEBS Journal 274 (2007) 4271–4286 ª 2007 The Authors Journal compilation ª 2007 FEBS 4275
Triafestin binds to both FXII and HK

To identify the target molecule(s) of triafestin-1 and
triafestin-2, we studied the interactions between both
triafestins and plasma proteins using a surface plas-
mon resonance (SPR) biosensor. As shown in
Fig. 5A,B, injection of FXII ⁄ FXIIa onto immobilized
triafestin-1 gave a significant response in a dose-
dependent manner, indicating that triafestin-1 binds
both the zymogen and enzyme forms of FXII. Exper-
iments using the same buffer condition demonstrated
clear interactions between both HK and two-chain
HK (HKa) and triafestin-1, which were also dose-
dependent (Fig. 5C,D). Similar results were obtained
with triafestin-2 (data not shown). PK and kallikrein
did not interact with both triafestins (data not
shown).
To evaluate the binding kinetics, interactions
between triafestins and target molecules were exam-
ined in the presence of 100 lm ZnCl
2
(Fig. 5). In
this analysis, various concentrations of target mole-
cules were used. In the kinetic analysis, a two-state
(biphasic) binding model fitted the data and was
used. Other binding models (e.g. the simple 1 : 1
Langmuir model) did not fit the SPR sensorgram
data. The two-state binding model is represented as
follows:
A þ B ()
ka1
kd1

AB ()
ka2
kd2
ðABÞ
Ã
where AB is the initial binding complex and (AB)* is
the final tight binding complex. The calculated kinetic
constants are summarized in Table 1. These kinetic
constants suggest that triafestin-1 and triafestin-2 rap-
idly associate with their target molecules, form tran-
sient initial complexes, and finally form tight
complexes.
Fig. 5. Sensorgrams for the binding of FXII (A), FXIIa (B), HK (C) and HKa (D) to immobilized triafestin-1 measured by SPR. Triafestin-1 was
coupled onto a B1 sensor chip at 2136 RU in binding assays for FXII, FXIIa and HK, and at 1876 RU in the binding assay for HKa. Different
concentrations of FXII, FXIIa, HK and HKa were injected at a flow rate of 20 lLÆmin
)1
in buffer containing 100 lM ZnCl
2
, and association
was monitored for 2 min. After a return to buffer flow, dissociation was monitored for 2 min. The sensor chip surface was regenerated with
25 m
M EDTA after each injection.
Plasma kallikrein–kinin system inhibitors H. Isawa et al.
4276 FEBS Journal 274 (2007) 4271–4286 ª 2007 The Authors Journal compilation ª 2007 FEBS
Zinc ions modulate FXII and HK binding
to triafestin-1 and triafestin-2
Zinc ions markedly affect FXII and HK structure and
function. Zn
2+
binding to FXII and HK induces their

conformational changes, which regulate the accessibil-
ity of these proteins to activating surfaces [5–11].
Therefore, we investigated the effects of Zn
2+
concen-
trations on the interaction of these plasma proteins
with triafestins. As shown in Fig. 6A,B, FXII and
FXIIa binding to triafestin-1 increased with increasing
Zn
2+
concentrations up to 150 lm, but decreased at
200 lm. For FXIIa, unlike FXII, weak binding to
triafestin-1 and triafestin-2 was observed even in the
absence of Zn
2+
. Similar results were obtained with
triafestin-2 (data not shown). HK binding to triafestin-
1 (Fig. 6C) and triafestin-2 (data not shown) peaked
at 25 lm and 100 lm Zn
2+
, respectively, and then
decreased at higher Zn
2+
concentrations. In the
absence of Zn
2+
, triafestin-1 weakly bound to HK,
but triafestin-2 did not. For HKa, unlike HK, binding
to both triafestins increased with increasing Zn
2+

con-
centration (Fig. 6D). At a low Zn
2+
concentration
($ 25 lm), the shape of SPR sensorgrams for HK and
HKa binding to triafestins were different from those at
higher Zn
2+
concentrations. These results suggest that
triafestins bind to FXII and FXIIa, and HK and
HKa, as a function of their Zn
2+
-induced conforma-
tional changes.
Triafestin inhibits FXII and HK binding to
negatively charged surfaces
Binding of FXII and HK to negatively charged sur-
faces can initiate and accelerate activation of the
plasma kallikrein–kinin system in vitro [38,39]. There-
fore, we examined the inhibitory effects of both triafes-
tins on adhesion of FXII and HK to a negatively
charged surface using an SPR biosensor. In these
assays, the sensor chip surface was coated with an
acidic phospholipid monolayer containing phosphati-
dylinositol phosphate (PtdInsP) as a negatively
charged surface. FXII or HK preincubated with triaf-
estin-1 or triafestin-2 was injected onto the sensor chip
surface, and association of FXII and HK with the
lipid monolayer was then measured. As shown in
Fig. 7, FXII and HK binding to the negatively charged

surface decreased as the molar ratio of triafestin-1 to
FXII or HK increased. Similar results were obtained
with triafestin-2 (data not shown), showing that both
triafestins interfered with FXII and HK binding to
the acidic phospholipid surface by a dose-dependent
mechanism. These results suggest that triafestin-1 and
Table 1. Kinetic constants for triafestin-1 and triafestin-2 interactions with FXII, FXIIa, HK, HKa, FXII
1)77
and HKD5. Kinetic constants were calculated from sensorgram curves using kinetic
evaluation software for a two-state binding model.
Triafestin-1 Triafestin-2
k
a1
(M
)1
Æs
)1
)
(· 10
5
)
k
d1
(s
)1
)
(· 10
)2
)
k

a2
(s
)1
)
(· 10
)2
)
k
d2
(s
)1
)
(· 10
)4
)
K (
M
)1
)
(· 10
8
) Chi
2
k
a1
(M
)1
Æs
)1
)

(· 10
5
)
k
d1
(s
)1
)
(· 10
)2
)
k
a2
(s
)1
)
(· 10
)2
)
k
d2
(s
)1
)
(· 10
)4
)
K (
M
)1

)
(· 10
8
) Chi
2
FXII 2.44 ± 0.06 6.27 ± 0.06 1.70 ± 0.12 3.45 ± 0.16 1.92 ± 0.16 49.77 ± 5.69 4.00 ± 0.20 6.00 ± 0.53 1.74 ± 0.10 3.37 ± 0.41 3.49 ± 0.51 11.04 ± 1.18
FXIIa 3.67 ± 0.04 1.61 ± 0.06 1.83 ± 0.05 15.60 ± 0.80 2.68 ± 0.14 36.53 ± 2.11 9.30 ± 0.90 6.30 ± 0.86 1.84 ± 0.08 7.46 ± 0.31 3.67 ± 0.17 3.68 ± 0.67
HK 1.94 ± 0.10 3.56 ± 0.20 1.90 ± 0.14 8.24 ± 0.10 1.25 ± 0.11 10.99 ± 1.26 1.09 ± 0.07 3.49 ± 0.07 1.99 ± 0.08 3.02 ± 0.80 2.18 ± 0.71 1.06 ± 0.17
HKa 2.46 ± 0.06 1.75 ± 0.24 2.84 ± 0.28 12.43 ± 1.1 3.23 ± 0.33 15.60 ± 0.17 2.70 ± 0.03 2.18 ± 0.10 3.36 ± 0.10 13.47 ± 0.42 3.09 ± 0.12 8.12 ± 0.33
FXII
1)77
0.091 ± 0.002 4.03 ± 0.30 1.89 ± 0.05 0.12 ± 0.02 3.51 ± 0.50 3.60 ± 0.10 0.071 ± 0.002 5.26 ± 0.20 1.37 ± 1.02 0.14 ± 0.04 2.10 ± 0.71 2.38 ± 0.16
HKD5 0.22 ± 0.01 0.33 ± 0.19 2.14 ± 0.20 0.43 ± 0.11 38.57 ± 10.07 33.77 ± 1.63 0.341 ± 0.023 0.47 ± 0.31 2.34 ± 0.73 0.24 ± 0.10 93.07 ± 55.52 26.1 ± 4.88
H. Isawa et al. Plasma kallikrein–kinin system inhibitors
FEBS Journal 274 (2007) 4271–4286 ª 2007 The Authors Journal compilation ª 2007 FEBS 4277
Fig. 7. Effect of triafestin-1 on the binding of FXII (A) and HK (B) to an acidic phospholipid monolayer measured by SPR analysis. HPA sensor
chips with hydrophobic surfaces were coated with an acidic phospholipid monolayer (phosphatidylcholine ⁄ PtdInsP ¼ 6 : 4) at 1027 RU
and 1088 RU for FXII and HK, respectively. FXII and HK were preincubated with various concentrations of triafestin-1 in buffer containing
200 l
M ZnCl
2
for FXII binding and 25 lM ZnCl
2
for HK binding. Sensorgrams were obtained from injections of these mixtures at a flow rate
of 20 lLÆmin
)1
, and association was monitored for 2 min. After a return to buffer flow, dissociation was monitored for 2 min. The sensor
chip surface was regenerated with 10 m
M NaOH after each injection.
Fig. 6. Effect of Zn

2+
concentration on the binding of FXII (A), FXIIa (B), HK (C) and HKa (D) to immobilized triafestin-1 measured by SPR.
The same sensor chips as in Fig. 5 were used in these assays. Interactions between FXII and triafestin-1 (A), FXIIa and triafestin-1 (B), HK
and triafestin-1 (C) and HKa and triafestin-1 (D) were investigated at Zn
2+
concentrations ranging from 0 to 200 lM. Sensorgrams were
obtained from injection at a flow rate of 20 lLÆmin
)1
, and association was monitored for 2 min. After a return to buffer flow, dissociation
was monitored for 2 min. The sensor chip surface was regenerated with 25 m
M EDTA after each injection.
Plasma kallikrein–kinin system inhibitors H. Isawa et al.
4278 FEBS Journal 274 (2007) 4271–4286 ª 2007 The Authors Journal compilation ª 2007 FEBS
triafestin-2 interact with the binding regions of FXII
and HK for the negatively charged surface.
Triafestin interacts with both the N-terminal
region of FXII and domain D5 of HK
An FXII putative binding region for activating sur-
faces has been mapped in its N-terminal region, which
contains a fibronectin type II domain [40]. In addition,
domain D5 of HK has been identified as a major bind-
ing region for negatively charged surfaces as well as
cellular surfaces [41–43]. These functional regions are
also known to contain several binding sites for Zn
2+
.
Therefore, we investigated whether triafestins could
interact with these binding regions and whether this
interaction is affected by Zn
2+

concentration. For
this analysis, N-terminus of FXII (FXII
1)77
) and
domain D5 of HK (HKD5) recombinant proteins were
prepared as described in Experimental procedures.
Interactions between triafestins and these recombinant
peptides were examined using an SPR sensor. As
expected, triafestin-1 bound to both FXII
1)77
and
HKD5 in a Zn
2+
concentration-dependent manner
(Figs 8 and 9). Similar results were obtained with triaf-
estin-2 (data not shown). Interestingly, hamadarin also
exhibited Zn
2+
-dependent binding to these functional
regions (Figs 8 and 9; Table 2). Therefore, we con-
cluded that triafestin-1 and triafestin-2, as well as
hamadarin, block FXII and HK interactions with acti-
vating surfaces by binding to their N-terminal region
and domain D5, respectively.
Discussion
It has been estimated that at least 24 putative secreted
proteins are present in the saliva of T. infestans [34].
However, most of these proteins have no known func-
tion(s) [27]. In this work, we have identified and char-
acterized two closely related inhibitors of the

kallikrein–kinin system, triafestin-1 and triafestin-2,
from T. infestans saliva.
Binding of the PK–HK complex to a biological acti-
vating surface such as the surface of endothelial cells
triggers the kallikrein–kinin system by promoting PK
activation by cell matrix-associated PK activator(s)
[44,45]. Furthermore, binding of both FXII and the
PK–HK complex to the surface accelerates their reci-
procal activation. Thus, it is possible that triafestins
inhibit initiation of kallikrein–kinin system activation
and the subsequent reciprocal FXII and PK activation.
Fig. 8. Sensorgrams for the binding of
FXII
1)77
(A) and HKD5 (B) to immobilized
triafestin-1 (left) and hamadarin (right) mea-
sured by SPR. Triafestin-1 was coupled onto
a sensor chip at 1876 RU and 1672 RU in
binding assays for FXII
1)77
and HKD5,
respectively. Hamadarin was coupled onto a
sensor chip at 2005 RU and 1917 RU in
binding assays for FXII
1)77
and HKD5,
respectively. Different concentrations of
FXII
1)77
and HKD5 were injected at flow

rate of 20 lLÆmin
)1
in buffer containing
100 l
M ZnCl
2
, and association was moni-
tored for 2 min. After a return to buffer
flow, dissociation was monitored for 2 min.
The sensor chip surface was regenerated
with 100 m
M EDTA after each injection.
H. Isawa et al. Plasma kallikrein–kinin system inhibitors
FEBS Journal 274 (2007) 4271–4286 ª 2007 The Authors Journal compilation ª 2007 FEBS 4279
FXII and HK compete for anionic and endothelial cell
surfaces in the presence of Zn
2+
. Therefore, there may
be common receptors for FXII and HK [18,19,46].
Triafestins might associate with both FXII and HK by
mimicking such common receptors and interfering with
their binding to activating surfaces.
Bradykinin induces vascular hypotension and blood
flow retardation by dilating blood vessels. It also
induces plasma diapedesis into tissue by increasing vas-
cular permeability, leading to blood condensation and
blood flow retardation in blood vessels. These effects
would not only reduce blood flow into an insect feed-
ing site but would also enhance host hemostatic
responses, such as blood coagulation and platelet

aggregation, started by vascular injury. Therefore,
early acute inflammation induced by bradykinin at an
injured site would be a serious disadvantage for blood-
feeding arthropods. Indeed, kissing bugs often feed on
blood successfully, without being noticed by the host
animal. Triafestins may function as anti-inflammatory
molecules to inhibit pain generation, as triafestins
strongly inhibit release of bradykinin through inhibi-
tion of kallikrein–kinin system activation.
The kallikrein–kinin system has been suggested to
have little influence on physiologic hemostasis, because
hereditary deficiencies in FXII are not associated with
spontaneous or excessive bleeding [47]. However,
recent studies have shown that FXII can induce patho-
logic thrombosis via both the intrinsic and extrinsic
coagulation pathways [48]. FXII-deficient and FXII
inhibitor-treated mice are protected against arterial
thrombosis and stroke, indicating that FXII plays a
Fig. 9. Effect of Zn
2+
concentration on the
binding of FXII
1)77
(A) and HKD5 (B) to
immobilized triafestin-1 (left) and hamadarin
(right) measured by SPR. The same sensor
chips as in Fig. 8 were used in these
assays. Interactions were investigated at
Zn
2+

concentrations ranging from 0 to
200 l
M. Sensorgrams were obtained with
an injection flow rate of 20 lLÆmin
)1
, and
association was monitored for 2 min. After
a return to buffer flow, dissociation was
monitored for 2 min. The sensor chip sur-
face was regenerated by 100 m
M EDTA
after each injection.
Table 2. Kinetic constants for hamadarin interactions with FXII
1)77
and HKD5. Kinetic constants were calculated from sensorgram curves
using kinetic evaluation software for a two-state binding model.
Hamadarin
k
a1
(M
)1
Æs
)1
)(· 10
3
) k
d1
(s
)1
)(· 10

)2
) k
a2
(s
)1
)(· 10
)2
) k
d2
(s
)1
)(· 10
)6
) K (M
)1
)(· 10
8
) Chi
2
FXII
1)77
5.91 ± 0.10 1.84 ± 0.09 2.17 ± 0.02 2.95 ± 1.68 29.83 ± 17.24 0.90 ± 0.12
HKD5 20.67 ± 0.55 1.99 ± 0.15 1.68 ± 0.12 2.22 ± 0.67 68.33 ± 13.92 67.37 ± 1.33
Plasma kallikrein–kinin system inhibitors H. Isawa et al.
4280 FEBS Journal 274 (2007) 4271–4286 ª 2007 The Authors Journal compilation ª 2007 FEBS
central role in pathologic thrombus formation in vivo
[20,21]. These observations, together with the data
reported here, suggest that triafestins may also func-
tion in vivo as an anticoagulant to counteract throm-
bus formation promoted by FXII activation. It may be

that the bite of blood-feeding insects is one of the
inducible factors in pathologic thrombus formation in
blood vessels.
The free Zn
2+
concentration in plasma is usually
kept at a significantly low level (< 1 lm) [49]. Because
triafestin interacts with FXII and HK at high Zn
2+
concentrations (Fig. 6), this implies that triafestin may
not act under normal conditions. However, once insect
blood-sucking injures blood vessels, localized increase
of free Zn
2+
may occur; Zn
2+
is released from the
cytoplasm of the injured cells into the plasma, and
platelets secrete granules storing high concentrations of
Zn
2+
by contact with exposed extracellular matrix
[50–52]. Triafestins may counteract the kallikrein–kinin
system in such a microenvironmental milieu.
On the basis of amino acid sequence similarities,
triafestin-1 and triafestin-2 apparently belong to the li-
pocalin family (Fig. 1B), whereas their inhibitory prop-
erties closely resemble those of hamadarin and
haemaphysalin [24,53], salivary kallikrein–kinin system
inhibitors that do not belong to this family. Hamada-

rin is a kallikrein–kinin system inhibitor identified in
the mosquito salivary gland, and belongs to the D7
family proteins, which have significant sequence simi-
larity to the pheromone ⁄ odorant-binding protein
superfamily [24,26]. Haemaphysalin was identified
from a hard tick, Haemaphysalis longicornis, and con-
sists of two Kunitz-type serine protease inhibitor
domains [53]. These three kallikrein–kinin system
inhibitors specifically bind both FXII and HK,
depending on Zn
2+
concentration, but do not share
any significant sequence similarity that may serve as a
common binding site for FXII and HK. The fact that
these three unrelated blood-sucking arthropods share
such similar bioactive substances as major components
of the saliva emphasizes the biological importance of
inhibition of the kallikrein–kinin system for blood-
feeding behavior. These observations may be an exam-
ple of the convergent evolution of salivary proteins in
blood-sucking arthropods.
In conclusion, the studies reported here show that
triafestin-1 and triafestin-2 inhibit kallikrein–kinin sys-
tem activation, and inhibit the reciprocal activation of
FXII and PK and subsequent bradykinin generation.
The triafestins exert this activity by binding to both
FXII and HK in the presence of Zn
2+
and inhibiting
the association of FXII and HK with an activating

surface. These results provide new insights into the
biochemical and pharmacologic complexity of blood-
sucking arthropods. On the basis of the unique biolog-
ical activities of triafestins, FXII and HK should be
promising new targets for antithrombotic therapies
that would present a low risk or no risk of excessive
bleeding.
Experimental procedures
Materials
FXII, FXIIa, PK, kallikrein, single-chain HK, HKa and
corn trypsin inhibitor were purchased from Enzyme
Research Laboratories (South Bend, IN). To determine
protein concentrations, the following extinction coefficients
(E
0:1%
280
) and relative molecular masses were used: FXII, 1.41,
80 000; FXIIa, 1.41, 80 000; PK, 1.17, 86 000; kallikrein,
1.17, 86 000; HK, 0.701, 120 000; and HKa, 0.701, 110 000.
Dextran sulfate of M
r
500 000 (DS 500) and soybean tryp-
sin inhibitor were purchased from Wako Pure Chemical
Industries Ltd (Osaka, Japan). Recombinant hamadarin
was produced as described previously [24]. The chromo-
genic substrate S-2302 (H-d-Pro-l-Phe-l-Arg-p-nitroanilide)
was purchased from Chromogenix AB (Mo
¨
lndal, Sweden).
Phosphatidylcholine and PtdInsP were purchased from

Sigma-Aldrich (St Louis, MO).
Isolation and sequencing of cDNA clones
Paired salivary glands were dissected from 30 fifth instar
nymphs of T. infestans, and polyA (+) RNA was isolated
using a MicroPrep mRNA isolation kit (Amersham Bio-
sciences, Piscataway, NJ). A salivary gland cDNA library
was constructed from the isolated mRNA using the Super-
Script plasmid system (Gibco BRL, Gaitherburg, MD). In
total, 550 cDNA clones were picked randomly from this
library, and plasmids were purified from overnight cultures
using the Qiaprep Spin Miniprep kit (Qiagen, Hilden, Ger-
many). The nucleotide sequences of cloned cDNA frag-
ments were determined using the ABI PRISM BigDye
Terminator cycle sequencing kit (Applied Biosystems, Fos-
ter City, CA) and an automated DNA sequencer (ABI 310
genetic analyzer; Applied Biosystems). The cDNA
sequences of triafestin-1 and triafestin-2 were deposited in
the DNA Data Bank of Japan (DDBJ) (accession numbers
AB292809 and AB292810). Analyses of the sequence data
were performed using the genetyx version 8.5 program
(GENETYX Co., Tokyo, Japan) and the blast program
( The amino acid sequences
of triafestin-1 and triafestin-2 were compared with those of
five salivary lipocalins from triatomine bugs: triabin (Gen-
Bank accession no. Q27049); pallidipin (GenBank accession
no. AAA30329); nitrophorin 1 (GenBank accession no.
Q26239); prolixin-S (GenBank accession no. AAB41587);
H. Isawa et al. Plasma kallikrein–kinin system inhibitors
FEBS Journal 274 (2007) 4271–4286 ª 2007 The Authors Journal compilation ª 2007 FEBS 4281
and Rhodnius platelet aggregation inhibitor-1 (GenBank

accession no. AAB09090). The amino acid sequences of
triafestin-1, triafestin-2 and the other lipocalins listed above
were aligned using the clustalw method (http://www.
clustal.genome.ad.jp/) and then analyzed by the mega pro-
gram version 3.1 [54]. A phylogenetic dendrogram for the
alignment was constructed using the neighbor-joining
method. The secretory signal sequence and its cleavage site
were predicted by the signalp program (.
dtu.dk/services/SignalP/).
Production and purification of the recombinant
proteins
Triafestin-1 and triafestin-2 recombinant proteins were pro-
duced in a baculovirus–insect cell system. Triafestin-1 and
triafestin-2 full-length cDNAs were cloned into the BamHI
site of the baculovirus transfer vector pAcYM1. Sf9 cells
were cotransfected with the constructed plasmid and linear-
ized baculovirus DNA (Linearized BD BaculoGold Baculo-
virus DNA; BD Biosciences, San Jose, CA). Tn5 cells
infected by recombinant baculovirus secreted recombinant
protein into the culture medium due to the original peptide
signal sequence. Recombinant proteins were purified from
the cell culture medium as follows. Culture medium
containing secreted recombinant protein was applied to a
PD-10 column (Amersham Biosciences) in 20 mm sodium
acetate buffer (pH 5.2). The eluted sample was applied to a
RESOURCE S column (Amersham Biosciences) equili-
brated with the same buffer and eluted with a 0–1 m NaCl
gradient at a flow rate of 4 mLÆmin
)1
. Fractions containing

the recombinant protein were pooled, concentrated using
Centricon 10 (Millipore Co., Bedford, MA), and applied to
a TSK G2000 SW column (Tosoh Co., Tokyo, Japan)
equilibrated with Tris-buffered saline (10 mm Tris ⁄ HCl,
pH 7.0, 150 mm NaCl). Peak fractions were analyzed by
15% SDS ⁄ PAGE under reducing conditions. The purified
protein concentration was determined with a Coomassie
protein assay kit (Pierce Biotechnology, Inc., Rockford, IL)
using bovine c-globulin as a standard.
Antibody preparation and western blot analysis
Recombinant triafestin-1 and triafestin-2 were also pro-
duced in Escherichia coli BL21 (Amersham Biosciences)
and used to immunize rats. Briefly, cDNA fragments
encoding predicted mature triafestin-1 and triafestin-2 were
amplified by PCR, and subcloned into the pMAL-c2G
expression plasmid (New England Biolabs, Beverly, MA) at
the unique restriction sites BamHI–XhoI. Recombinant
triafestin-1 and triafestin-2 were produced as maltose-bind-
ing protein-fused proteins and affinity-purified on amylose
resin. Fused maltose-binding protein was cleaved by Genen-
ase I (New England Biolabs) according to the manufac-
turer’s instructions. For antibody production, rats were
immunized with maltose-binding protein-free triafestin-1 or
triafestin-2. Western blot analysis was performed essentially
as described previously [55].
Expression and purification of the recombinant
N-terminal region of FXII and domain D5 of HK
Expression vectors for the recombinant N-terminal region of
FXII and domain D5 of HK were constructed using a
pET22b vector (Novagen, Darmstadt, Germany) as follows.

For expression of the FXII N-terminal region, a cDNA frag-
ment coding for amino acid residues Ile1 to Lys77 of FXII
(FXII
1)77
), which contains a fibronectin type II domain, was
amplified by PCR using a human liver cDNA library (First-
Choise PCR-Ready Human Liver cDNA; Ambion Inc.,
Austin, TX) with primers FN2f (5¢-GCTCATGATTCCAC
CTTGGGAAGCCCCCA-3¢) and FN2r (5¢-CGGGATCCT
CATTTCACTTTCTTGGGCTCC-3¢). For expression of
domain D5 of HK, a cDNA fragment encoding domain D5
and part of domain D6 (amino acid residues Val383 to
Tyr514, HKD5) was amplified by PCR from the same human
liver cDNA library with primers DK3 (5¢-GCAGCAGTCA
TGACTGTAAGTCCACCCCACACTTCC-3¢) and modi-
fied DK4 (5¢-GCAGCAGGATCCTCAACTGTCTTCAGA
AGAGCTTGC-3¢), as described by Herwald et al. [11]. Each
resultant fragment was digested with BspHI and BamHI, and
then inserted into the NcoI–BamHI sites of the pET22b vec-
tor. Recombinant proteins were expressed in E. coli BL21-
codon plus (DE3)-RP strain (Stratagene, La Jolla, CA)
according to the manufacturer’s instructions. The soluble
form of FXII
1)77
was extracted from bacterial cell lysates
and purified to homogeneity by a combination of cation
exchange HPLC and size exclusion HPLC as described
above. Purification of recombinant HKD5 was performed as
described by Herwald et al. [11].
Assay for the effect of triafestin-1 and triafestin-2

on plasma coagulation and tenase activity
The effects of triafestin-1 and triafestin-2 on APTT and
PT were assayed as follows. Twenty microliters of citrat-
ed normal human plasma (Caliplasma Index 100; Bio-
merieux, Marcy l’Etoile, France) and 20 lL of triafestin-1
or triafestin-2 were preincubated for 5 min. The mixture
then was activated for 2 min at 37 °C with 35 lLof
diluted APTT reagent (actin; Dade Behring, Liederbach,
Germany) for the APTT assay, or with 35 lL of diluted
PT reagent (rabbit brain thromboplastin; Ortho-Clinical
Diagnostic, Inc., Raritan, NJ) for the PT assay. The clot-
ting reaction was initiated by addition of 25 lLof
25 mm CaCl
2
, and the clotting time was measured using
a KC-10 coagulometer (Heinrich Amelung, Lemgo, Ger-
many). The effect of triafestin on the intrinsic tenase
activity was examined by a reconstitution system as
described previously [35].
Plasma kallikrein–kinin system inhibitors H. Isawa et al.
4282 FEBS Journal 274 (2007) 4271–4286 ª 2007 The Authors Journal compilation ª 2007 FEBS
Assay for the effect of triafestin-1 and triafestin-2
on activation of the kallikrein–kinin system
To assay for the effects of triafestin-1 and triafestin-2 on
reciprocal activation of FXII and PK, FXII (0.2 nm, final
concentration) was preincubated with triafestin-1 or triafes-
tin-2 in an assay buffer (50 mm Tris ⁄ HCl, pH 7.4, 50 mm
NaCl, 1% BSA, 0.1% polyethylene glycol 8000) for 10 min
at room temperature. Autoactivation of FXII and reciprocal
activation of FXII and PK were initiated by addition of PK

(10 nm, final concentration) and DS 500 (0.2 lgÆmL
)1
, final
concentration). After 10 min of incubation, S-2302 (170 l m,
final concentration) was added to the reaction mixture, and
absorbance changes at 405 nm were recorded at 2 min inter-
vals. To assay for the effects of triafestin-1 and triafestin-2
on FXII activation by kallikrein, FXII (20 nm, final concen-
tration) was preincubated with triafestin-1 or triafestin-2 in
the same buffer for 10 min. Activation was initiated by
addition of kallikrein (0.2 nm, final concentration) and
DS 500 (0.3 lgÆmL
)1
, final concentration). After 10 min of
incubation, soybean trypsin inhibitor (0.6 lm, final concen-
tration) and S-2302 (340 lm, final concentration) were
added, and the increase in absorbance at 405 nm was
recorded at 2 min intervals. To assay for the effects of triaf-
estin-1 and triafestin-2 on PK activation by FXIIa, FXIIa
(50 pm, final concentration) was preincubated with triafes-
tin-1 or triafestin-2 in the same buffer for 10 min. Activa-
tion of PK by FXIIa was started by addition of PK (10 nm,
final concentration) and DS 500 (0.1 lgÆmL
)1
, final concen-
tration). After 5 min of incubation, corn trypsin inhibitor
(100 nm, final concentration) and S-2302 (170 lm, final con-
centration) were added, and absorbance changes at 405 nm
were monitored at 2 min intervals.
Assay for the effect of triafestin-1 and triafestin-2

on bradykinin generation
The effects of triafestin-1 and triafestin-2 on bradykinin
generation were investigated in a reconstitution assay. FXII
(0.2 nm, final concentration) was preincubated with triafes-
tin-1 or triafestin-2 in buffer (50 mm Tris ⁄ HCl, pH 7.4,
50 mm NaCl, 1% BSA, 0.1% polyethylene glycol 8000) for
10 min at room temperature. Autoactivation of FXII and
reciprocal activation of FXII and plasma proteins were ini-
tiated by addition of PK (10 nm, final concentration), HK
(10 nm, final concentration) and DS 500 (0.4 lgÆmL
)1
, final
concentration). After 10 min of incubation, released brady-
kinin was quantified by ELISA using a Markit-M Brady-
kinin kit (Dainippon Pharmaceutical Co., Osaka, Japan).
Binding analysis with SPR
SPR measurements were performed using a BIAcore X bio-
sensor system (Biacore AB, Uppsala, Sweden). Triafestin-1
or triafestin-2 was immobilized onto the surface of a B1
sensor chip in 10 mm sodium acetate buffer (pH 4.5) by the
amine coupling procedure. An empty flow cell was prepared
by the same immobilizing procedure, but without ligand
protein, and used for the control assay. Analyses of binding
of triafestin-1 and triafestin-2 to plasma proteins and their
derivatives (FXII, FXIIa, HK, HKa, PK, kallikrein,
FXII
1)77
or HKD5) were performed at 25 °C in running
buffer (10 mm Hepes, pH 7.4, 150 mm NaCl, 0.005%
Tween-20) containing various concentrations of ZnCl

2
.
Forty microliters of each protein mixture was injected at a
flow rate of 20 lLÆmin
)1
, and association with triafestin-1
or triafestin-2 was monitored for 120 s. After a return to
buffer flow, dissociation was monitored for 120 s. The sen-
sor chip surface was regenerated by a pulse injection of 25
or 100 mm EDTA after each experiment. These regenera-
tion conditions reproducibly brought resonance unit (RU)
values back to those observed before each new injection.
Kinetic binding constants were determined using biaevalu-
ation 3.0 software (Biacore AB). To assay for Zn
2+
depen-
dency in triafestin binding to plasma proteins and their
derivatives, running buffer was passed through a column of
Chelex 100 (BioRad, Hercules, CA) to remove divalent
metal ion contaminants before addition of ZnCl
2
. Each
protein solution (250 nm) was dialyzed in a metal-chelating
running buffer containing 0.1% Chelex 100, and was
injected at different Zn
2+
concentrations.
Triafestin-1 and triafestin-2 interference with the binding
of FXII and HK to a negatively charged surface was also
monitored using a BIAcore X instrument. Phospholipid

vesicles (phosphatidylcholine and PtdInsP in a 6 : 4 (w ⁄ w)
ratio) were prepared as described previously [24], and then
immobilized on the hydrophobic alkanethiol-coated surface
of an HPA sensor chip according to the manufacturer’s
instruction. FXII (20 nm) or HK (20 nm) was preincubated
with various concentrations of triafestin-1 or triafestin-2 for
5 min in Tris-buffered saline (50 mm Tris ⁄ HCl, pH 7.4,
150 mm NaCl) containing 100 lm or 25 lm ZnCl
2
, respec-
tively. Each mixture was injected onto the sensor chip sur-
face, and association and dissociation were monitored for
120 s each. To regenerate the phospholipid surface, com-
plete dissociation of bound proteins was done by the addi-
tion of 100 m m NaOH for 30 s.
Acknowledgements
This study was supported by a grant-in-aid for
Research on Health Sciences focusing on Drug Inno-
vation (KH23306) and Young Scientists Fellowship B
(16790249) to H. Isawa from the Japan Health Sci-
ences Foundation and the Japan Society for the Pro-
motion of Science (JSPS), respectively, and a grant
from the Research for the Future Program from JSPS
to Y. Chinzei. It was also supported by grants from
H. Isawa et al. Plasma kallikrein–kinin system inhibitors
FEBS Journal 274 (2007) 4271–4286 ª 2007 The Authors Journal compilation ª 2007 FEBS 4283
Health Sciences Research Grant in Research on
Emerging and Reemerging Infectious Diseases from
the Japanese Ministry of Health, Labor and Welfare.
References

1 Colman RW (1984) Surface-mediated defense reactions.
The plasma contact activation system. J Clin Invest 73,
1249–1253.
2 Colman RW & Schmaier AH (1997) Contact system: a
vascular biology modulator with anticoagulant, profi-
brinolytic, antiadhesive, and proinflammatory attributes.
Blood 90, 3819–3843.
3 Schmaier AH (2003) The kallikrein–kinin and the
renin–angiotensin systems have a multilayered interac-
tion. Am J Physiol Regul Integr Comp Physiol 285,
R1–R13.
4 Joseph K & Kaplan AP (2005) Formation of bradyki-
nin: a major contributor to the innate inflammatory
response. Adv Immunol 86 , 159–208.
5 Shimada T, Kato H & Iwanaga S (1987) Accelerating
effect of zinc ions on the surface-mediated activation of
factor XII and prekallikrein. J Biochem (Tokyo) 102,
913–921.
6 Schousboe I & Halkier T (1991) Zinc ions promote the
binding of factor XII ⁄ factor XIIA to acidic phospholip-
ids but have no effect on the binding of high-Mr kinino-
gen. Eur J Biochem 197, 309–314.
7 Lin Y, Pixley RA & Colman RW (2000) Kinetic analy-
sis of the role of zinc in the interaction of domain 5 of
high-molecular weight kininogen (HK) with heparin.
Biochemistry 39, 5104–5110.
8 Røjkjær R & Schousboe I (1997) The surface-dependent
autoactivation mechanism of factor XII. Eur J Biochem
243, 160–166.
9 Bernardo MM, Day DE, Halvorson HR, Olson ST &

Shore JD (1993) Surface-independent acceleration of
factor XII activation by zinc ions. II. Direct binding
and fluorescence studies. J Biol Chem 268, 12477–12483.
10 Bernardo MM, Day DE, Olson ST & Shore JD (1993)
Surface-independent acceleration of factor XII activa-
tion by zinc ions. I. Kinetic characterization of the
metal ion rate enhancement. J Biol Chem 268, 12468–
12476.
11 Herwald H, Mo
¨
rgelin M, Svensson HG & Sjo
¨
bring U
(2001) Zinc-dependent conformational changes in
domain D5 of high molecular mass kininogen modulate
contact activation. Eur J Biochem 268, 396–404.
12 Scott CF, Silver LD, Schapira M & Colman RW (1984)
Cleavage of human high molecular weight kininogen
markedly enhances its coagulant activity. Evidence that
this molecule exists as a procofactor. J Clin Invest 73,
954–962.
13 Schmaier AH, Røjkjær R & Shariat-Madar Z (1999)
Activation of the plasma kallikrein ⁄ kinin system on
cells: a revised hypothesis. Thromb Haemost 82, 226–
233.
14 Røjkjær R, Hasan AAK, Motta G, Schousboe I &
Schmaier AH (1998) Factor XII does not initiate pre-
kallikrein activation on endothelial cells. Thromb
Haemost 80, 74–81.
15 Joseph K, Shibayama Y, Ghebrehiwet B & Kaplan AP

(2001) Factor XII-dependent contact activation on
endothelial cells and binding proteins gC1qR and cyto-
keratin 1. Thromb Haemost 85, 119–124.
16 Motta G, Røjkjær R, Hasan AAK, Cines DB & Schma-
ier AH (1998) High molecular weight kininogen regu-
lates prekallikrein assembly and activation on
endothelial cells: a novel mechanism for contact activa-
tion. Blood 91, 516–528.
17 Hasan AAK, Zisman T & Schmaier AH (1998) Identifi-
cation of cytokeratin 1 as a binding protein and presen-
tation receptor for kininogens on endothelial cells.
Proc
Natl Acad Sci USA 95, 3615–3620.
18 Joseph K, Ghebrehiwet B, Peerschke EIB, Reid KBM
& Kaplan AP (1996) Identification of the zinc-depen-
dent endothelial cell binding protein for high molecular
weight kininogen and factor XII: identity with the
receptor that binds to the globular ‘heads’ of C1q
(gC1q-R). Proc Natl Acad Sci USA 93, 8552–8557.
19 Herwald H, Dedio J, Kellner R, Loos M & Mu
¨
ller-
Esterl W (1996) Isolation and characterization of the
kininogen-binding protein p33 from endothelial cells.
Identity with the gC1q receptor. J Biol Chem 271,
13040–13047.
20 Renne T, Pozgajova M, Gruner S, Schuh K, Pauer HU,
Burfeind P, Gailani D & Nieswandt B (2005) Defective
thrombus formation in mice lacking coagulation factor
XII. J Exp Med 202, 271–281.

21 Kleinschnitz C, Stoll G, Bendszus M, Schuh K, Pauer HU,
Burfeind P, Renne C, Gailani D, Nieswandt B & Renne
T (2006) Targeting coagulation factor XII provides pro-
tection from pathological thrombosis in cerebral ische-
mia without interfering with hemostasis. J Exp Med
203, 513–518.
22 Law JH, Ribeiro JMC & Wells MA (1992) Biochemical
insights derived from insect diversity. Annu Rev Biochem
61, 87–111.
23 Ribeiro JMC (1995) Blood-feeding arthropods: live
syringes or invertebrate pharmacologists? Infect Agents
Dis 4, 143–152.
24 Isawa H, Yuda M, Orito Y & Chinzei Y (2002) A mos-
quito salivary protein inhibits activation of the plasma
contact system by binding to factor XII and high molec-
ular weight kininogen. J Biol Chem 277, 27651–27658.
25 Arca
´
B, Lombardo F, de Lara Capurro M, della Torre A,
Dimopoulos G, James AA & Coluzzi M (1999) Trapping
cDNAs encoding secreted proteins from the salivary
glands of the malaria vector Anopheles gambiae. Proc
Natl Acad Sci USA 96, 1516–1521.
Plasma kallikrein–kinin system inhibitors H. Isawa et al.
4284 FEBS Journal 274 (2007) 4271–4286 ª 2007 The Authors Journal compilation ª 2007 FEBS
26 Calvo E, Mans BJ, Andersen JF & Ribeiro JM (2006)
Function and evolution of a mosquito salivary protein
family. J Biol Chem 281, 1935–1942.
27 Morita A, Isawa H, Orito Y, Iwanaga S, Chinzei Y &
Yuda M (2006) Identification and characterization of a

collagen-induced platelet aggregation inhibitor, triplatin,
from salivary glands of the assassin bug, Triatoma infe-
stans. FEBS J 273, 2955–2962.
28 Noeske-Jungblut C, Haendler B, Donner P, Alagon A,
Possani L & Schleuning WD (1995) Triabin, a highly
potent exosite inhibitor of thrombin. J Biol Chem 270,
28629–28634.
29 Noeske-Jungblut C, Kratzschmar J, Haendler B,
Alagon A, Possani L, Verhallen P, Donner P &
Schleuning WD (1994) An inhibitor of collagen-induced
platelet aggregation from the saliva of Triatoma pallidi-
pennis. J Biol Chem 269, 5050–5053.
30 Francischetti IM, Ribeiro JM, Champagne D &
Andersen J (2000) Purification, cloning, expression, and
mechanism of action of a novel platelet aggregation
inhibitor from the salivary gland of the blood-sucking
bug, Rhodnius prolixus. J Biol Chem 275, 12639–12650.
31 Champagne DE, Nussenzveig RH & Ribeiro JM (1995)
Purification, partial characterization, and cloning of
nitric oxide-carrying heme proteins (nitrophorins) from
salivary glands of the blood-sucking insect Rhodnius
prolixus. J Biol Chem 270, 8691–8695.
32 Sun J, Yamaguchi M, Yuda M, Miura K, Takeya H,
Hirai M, Matsuoka H, Ando K, Watanabe T, Suzuki K
et al. (1996) Purification, characterization and cDNA
cloning of a novel anticoagulant of the intrinsic path-
way, (prolixin-S) from salivary glands of the blood
sucking bug, Rhodnius prolixus. Thromb Haemost 75,
573–577.
33 Stark KR & James AA (1996) Anticoagulants in vector

arthropods. Parasitol Today 12, 430–436.
34 Pereira MH, Souza ME, Vargas AP, Martins MS,
Penido CM & Diotaiuti L (1996) Anticoagulant activity
of Triatoma infestans and Panstrongylus megistus saliva
(Hemiptera ⁄ Triatominae). Acta Trop 61, 255–261.
35 Isawa H, Yuda M, Yoneda K & Chinzei Y (2000)
The insect salivary protein, prolixin-S, inhibits factor
IXa generation and Xase complex formation in
the blood coagulation pathway. J Biol Chem 275,
6636–6641.
36 Samuel M, Pixley RA, Villanueva MA, Colman RW &
Villanueva GB (1992) Human factor XII (Hageman
factor) autoactivation by dextran sulfate. Circular
dichroism, fluorescence, and ultraviolet difference
spectroscopic studies. J Biol Chem 267, 19691–19697.
37 Citarella F, Wuillemin WA, Lubbers YTP & Hack CE
(1997) Initiation of contact system activation in plasma
is dependent on factor XII autoactivation and not on
enhanced susceptibility of factor XII for kallikrein
cleavage. Br J Haematol 99, 197–205.
38 Schousboe I (1990) The inositol-phospholipid-acceler-
ated activation of prekallikrein by activated factor
XII at physiological ionic strength requires zinc
ions and high-Mr kininogen. Eur J Biochem 193,
495–499.
39 Schousboe I & Halkier T (1991) Zinc ions promote the
binding of factor XII ⁄ factor XIIA to acidic phospholip-
ids but have no effect on the binding of high-Mr kinino-
gen. Eur J Biochem 197
, 309–314.

40 Røjkjær R & Schousboe I (1997) Partial identification
of the Zn
2+
-binding sites in factor XII and its activa-
tion derivatives. Eur J Biochem 247 , 491–496.
41 DeLa Cadena RA & Colman RW (1992) The sequence
HGLGHGHEQQHGLGHGH in the light chain of high
molecular weight kininogen serves as a primary struc-
tural feature for zinc-dependent binding to an anionic
surface. Protein Sci 1, 151–160.
42 Hasan AA, Cines DB, Herwald H, Schmaier AH &
Muller-Esterl W (1995) Mapping the cell binding site on
high molecular weight kininogen domain 5. J Biol Chem
270, 19256–19261.
43 Kunapuli SP, DeLa Cadena RA & Colman RW (1993)
Deletion mutagenesis of high molecular weight
kininogen light chain. Identification of two anionic
surface binding subdomains. J Biol Chem 268, 2486–
2492.
44 Shariat-Madar Z, Mahdi F & Schmaier AH (2002)
Identification and characterization of prolylcarboxypep-
tidase as an endothelial cell prekallikrein activator.
J Biol Chem 277, 17962–17969.
45 Joseph K, Tholanikunnel BG & Kaplan AP (2002) Heat
shock protein 90 catalyzes activation of the prekallik-
rein–kininogen complex in the absence of factor XII.
Proc Natl Acad Sci USA 99, 896–900.
46 Reddigari SR, Shibayama Y & Brunne
´
e Kaplan AP

(1993) Human Hageman factor (factor XII) and high
molecular weight kininogen compete for the same bind-
ing site on human umbilical vein endothelial cells. J Biol
Chem 268, 11982–11987.
47 Saito H (1987) Contact factors in health and disease.
Semin Thromb Hemost 13, 36–49.
48 Colman RW (2006) Are hemostasis and thrombosis two
sides of the same coin? J Exp Med 203, 493–495.
49 Cunningham BC, Bass S, Fuh G & Wells JA (1990)
Zinc mediation of the binding of human growth hor-
mone to the human prolactin receptor. Science 250,
1709–1712.
50 Whitehouse RC, Prasad AS, Rabbani PI & Cossack ZT
(1982) Zinc in plasma, neutrophils, lymphocytes,
and erythrocytes as determined by flameless
atomic absorption spectrophotometry. Clin Chem 28,
475–480.
51 Foley B, Johnson SA, Hackley B, Smith JC Jr &
Halsted JA (1968) Zinc content of human platelets. Proc
Soc Exp Biol Med 128, 265–269.
H. Isawa et al. Plasma kallikrein–kinin system inhibitors
FEBS Journal 274 (2007) 4271–4286 ª 2007 The Authors Journal compilation ª 2007 FEBS 4285
52 Baker RJ, McNeil JJ & Lander H (1978) Platelet metal
levels in normal subjects determined by atomic absorp-
tion spectrophotometry. Thromb Haemost 39, 360–365.
53 Kato N, Iwanaga S, Okayama T, Isawa H, Yuda M &
Chinzei Y (2005) Identification and characterization of
the plasma kallikrein–kinin system inhibitor, haemaphy-
salin, from hard tick, Haemaphysalis longicornis.
Thromb Haemost 93, 359–367.

54 Kumar S, Tamura K & Nei M (2004) MEGA3: inte-
grated software for molecular evolutionary genetics
analysis and sequence alignment. Brief Bioinform 5,
150–163.
55 Yuda M, Sawai T & Chinzei Y (1999) Structure and
expression of an adhesive protein-like molecule of mos-
quito invasive-stage malarial parasite. J Exp Med 189,
1947–1952.
Plasma kallikrein–kinin system inhibitors H. Isawa et al.
4286 FEBS Journal 274 (2007) 4271–4286 ª 2007 The Authors Journal compilation ª 2007 FEBS