(Received for publication, December 11, 1996)
From We reported previously the cloning of a novel
human serine protease inhibitor containing two Kunitz-like domains,
designated as placental bikunin, and the subsequent purification of a
natural counterpart from human placental tissue (Marlor, C. W.,
Delaria, K. A., Davis, G., Muller, D. K., Greve, J. M., and
Tamburini, P. P. (1997) J. Biol. Chem. 272, 12202-12208). In this report, the 170 residue extracellular domain of
placental bikunin (placental bikunin(1-170)) was expressed
in baculovirus-infected Sf9 cells using its putative signal peptide.
The resulting 21.3-kDa protein accumulated in the medium with the
signal peptide removed and could be highly purified by sequential
kallikrein-Sepharose and C18 reverse-phase chromatography.
To provide insights as to the potential in vivo functions
of this protein, we performed an extensive investigation of the
inhibitory properties of recombinant placental
bikunin(1-170) and both of its synthetically prepared Kunitz domains. All three proteins inhibited a number of serine proteases involved in the intrinsic pathway of blood coagulation and
fibrinolysis. Placental bikunin(1-170) formed
inhibitor-protease complexes with a 1:2 stoichiometry and strongly
inhibited human plasmin (Ki = 0.1 nM),
human tissue kallikrein (Ki = 0.1 nM),
human plasma kallikrein (Ki = 0.3 nM)
and human factor XIa (Ki = 6 nM).
Conversely, this protein was a weaker inhibitor of factor VIIa-tissue
factor (Ki = 1.6 µM), factor IXa
(Ki = 206 nM), factor Xa
(Ki = 364 nM), and factor XIIa
(Ki = 430 nM). This specificity profile
was to a large extent mimicked, albeit with reduced potency, by the
individual Kunitz domains. As predicted from this in vitro specificity profile, recombinant placental bikunin(1-170) prolonged the clotting time in an activated partial thromboplastin time
assay.
Blood clotting, resulting either from the extrinsic pathway
following tissue injury or the intrinsic pathway following contact activation, involves tightly regulated proteolytic cascades (1). The
intrinsic pathway is initiated by activation of factor XII either
through proteolysis or contact with negatively charged surfaces.
Activated factor XIIa, in turn, converts plasma prekallikrein to
kallikrein, which can then activate additional factor XII. Factor XIIa
activates factor XI, which, in turn, activates factor IX. Activated
factor IX forms a complex with factor VIIIa, phospholipid, and calcium,
which converts factor X to factor Xa. Factor X is also activated by the
factor VIIa-tissue factor complex operating within the extrinsic
pathway. Thrombin generation by factor Xa in complex with factor Va
leads ultimately to the formation of the fibrin clot. Thrombus
formation is also regulated by the fibrinolytic system whereby plasmin,
formed from plasminogen by the action of kallikrein, tissue plasminogen
activator (tPA),1 or urokinase, breaks down
both fibrinogen and fibrin (2).
Protease inhibitors play critical roles in the regulation of the
coagulation and fibrinolytic systems. Tissue factor pathway inhibitor
(TFPI), a multivalent Kunitz-type inhibitor, is a major regulator of
the extrinsic pathway through factor Xa-dependent inhibition of factor VIIa-tissue factor (3). C1 inhibitor is thought to
be the major physiological inhibitor of the intrinsic pathway enzymes
plasma kallikrein and factor XII (4). More recently, TFPI-2, a serine
protease inhibitor structurally related to TFPI, has been proposed to
be an important physiological inhibitor of enzymes involved in both
coagulation and fibrinolysis (5).
In a previous study, we identified and cloned a human cDNA encoding
a novel protein containing two Kunitz-like domains which we termed
placental bikunin (6). In this study a soluble recombinant fragment of
this protein, placental
bikunin(1-170),2 was expressed
and found to be a potent serine protease inhibitor. Using the
recombinant protein and both of its synthetic NH2- and COOH-terminal Kunitz domains, we dissect the specificity of the protein
more comprehensively. In doing so, we demonstrate that placental
bikunin is a potent inhibitor of plasma and tissue kallikreins, plasmin
and factor XIa.
Bovine chymotrypsin, bovine trypsin, and tPA
(single chain form from human melanoma cell culture), urokinase,
Suc-A-A-P-F-AMC, Boc-L-G-R-AMC, Suc-A-A-A-AMC, Boc-Q-G-R-AMC,
Bz-P-F-R-pNa, and P-F-R-AMC were from Sigma. Neutrophil
elastase was from Athens Research and Technology, Inc. (Athens, GA).
Human plasmin, factor VIIa, tissue factor (lipidated), and
(CH3SO2-D-cyclohexylpyrosyl)-G-R-pNa (Spectrozyme tPA substrate) were from American Diagnostica, Inc. (Greenwich, CT). Human plasma kallikrein and human factors IXa, X, Xa,
XIa, and XIIa were from Enzyme Research Laboratories (South Lafayette,
IN). Tos-G-P-K-AMC, Suc-A-A-P-R-pNa, and
Boc-E(OBzl)-A-R-AMC, were from Bachem Bioscience (King of Prussia, PA).
H-D-I-P-R-pNa was from Chromogenix AB. Bovine pancreatic
kallikrein, human tissue kallikrein, and recombinant aprotinin were
kindly provided by Bayer AG (Wuppertal, Germany).
A cDNA fragment encoding an
NH2-terminal 170-amino acid fragment of placental bikunin
was expressed in Sf9 cells (7) as follows. Placental bikunin cDNA
obtained by polymerase chain reaction and contained within the pCRII
vector as described previously (6) was liberated by digestion with
HindIII and XbaI. This fragment was gel purified and then
cloned into the M13mp19 vector (New England Biolabs, Beverly, MA).
In vitro mutagenesis (8) was used to generate a
PstI site between the upstream XbaI site and the
sequence encoding the ATG start site at the 5 Peptides corresponding to amino acids 7-64 and
102-159 of mature placental bikunin (6) were synthesized on an Applied
Biosystems model 433A peptide synthesizer using NMP-HBTU Fmoc
chemistry. Synthesis was on a preloaded Gln resin (Nova Biochem, La
Jolla, CA) with an 8-fold excess of Fmoc-amino acid/coupling. The side chain protecting groups used were as follows: for cysteine, histidine, tryptophan, and asparagine, trityl; for glutamic acid, serine, tyrosine, aspartic acid, and threonine, t-butyl; for lysine,
t-Boc; and for arginine,
2,2,5,7,8-pentamethylchroman-6-sulfonyl. In addition, glycine was
protected with
N- The purified synthetic peptides corresponding to placental
bikunin(7-64) and placental
bikunin(102-159) were refolded to yield the respective
functional NH2-terminal(7-64) and COOH-terminal(102-159) Kunitz domains using an adaptation of the method of Tam et al. (9) as follows. A solution
containing 23% (v/v) dimethyl sulfoxide in 0.1 M Tris-HCl
(pH 6.0) was added dropwise to a solution of reduced peptide (2.1 and 2.9 mg/ml for the NH2-terminal(7-64) and
COOH-terminal(102-159) domain, respectively) in 0.1 M Tris-HCl (pH 6.0), containing 8 M urea to
obtain a final concentration of 0.3 mg/ml peptide in 20% dimethyl
sulfoxide, 0.1 M Tris-HCl (pH 6.0), and 1 M
urea. The solutions containing either domain were each stirred at
25 °C for 24 h and then diluted 1:10 with 50 mM
Tris-HCl (pH 8.0) containing 0.1 M NaCl. Refolded material
was isolated by affinity chromatography over immobilized bovine
pancreatic kallikrein under conditions described above for the
purification of recombinant placental bikunin(1-170).
Fractions containing trypsin inhibitory activity were pooled, adjusted
to pH 2.5 with trifluoroacetic acid, then applied directly to a Vydac
C18 reverse-phase column (5 µm, 0.46 × 25 cm). The
NH2-terminal(7-64) Kunitz domain was purified
using a 40-min linear gradient of 22.5-50% acetonitrile in 0.1%
trifluoroacetic acid at a flow rate of 1.0 ml/min. The COOH-terminal(102-159) Kunitz domain was purified using a
40-min linear gradient of 20-40% acetonitrile in 0.1%
trifluoroacetic acid. Active fractions were pooled, lyophilized,
dissolved in 0.1% trifluoroacetic acid, and stored at To determine the stoichiometry with trypsin, the protease
(2.2 nM) was incubated with recombinant placental bikunin
(0-2 nM) or aprotinin (0-4 nM) in 1.0 ml of
50 mM Hepes (pH 7.5), 0.1 M NaCl, 2.0 mM CaCl2, 0.01% Triton X-100 (buffer A) for 30 min at room temperature, followed by the addition of
Suc-A-A-P-R-pNa (90 µM, final concentration),
after which residual enzymatic formation of pNa was
monitored at 410 nm. To determine the stoichiometry with plasma
kallikrein, 40 nM protease was incubated with recombinant placental bikunin (0-40 nM) in 1.0 ml of 50 mM
Tris-HCl (pH 8.0) containing 0.1 M NaCl and 0.01% Triton
X-100 (buffer B). After 30 min at room temperature,
Bz-P-F-R-pNa was added (300 µM, final concentration) and residual enzymatic activity monitored at 410 nm. Fractional protease activity was calculated as
V(+I)/V( Apparent equilibrium dissociation constants
(Ki*), were determined as described previously (6)
using methods for tight binding inhibitors (10) assuming
enzyme:inhibitor stoichiometries of 1:1 and 2:1 for binding to single
Kunitz domains and placental bikunin(1-170), (1)
respectively. Active site concentrations of trypsin, plasma and tissue
kallikreins, and plasmin were determined by titration with
p-nitrophenyl-p Preclinical Research,
Institute for Research Technologies,
ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES
Materials
end of the cDNA
insert. The oligonucleotide used for the mutagenesis had the sequence
CGC GTC TCG GCT GAC CTG GCC CTG CAG ATG GCG CAC GTG TGC GGG. A stop
codon (TAG) and BglII/XmaI site was similarly engineered at the 3
end of the cDNA using the oligonucleotide CTG
CCC CTT GGC TCA AAG TAG GAA GAT CTT CCC CCC GGG GGG GTG GTT CTG GCG GGG
CTG. The stop codon was in-frame with the sequence encoding placental
bikunin and caused termination immediately following the lysine at
amino acid residue 170 of the mature protein sequence. The mutated
cDNA thus encoded a truncated placental bikunin protein containing
both Kunitz domains but which was devoid of the putative transmembrane
domain. The product from digestion with PstI and
BglII was isolated and cloned into the BacPac8 vector (CLONTECH) for expression of placental
bikunin(1-170). The expression of placental
bikunin(1-170) by Sf9 insect cells was optimal at a
multiplicity of infection of 1 when the medium was harvested at 72 h post-infection. Sf9 cell culture supernatant (2-3 liters) was
harvested by centrifugation (1,500 × g for 30 min),
adjusted to pH 8.0 by the addition of 1 M Tris-HCl (pH 8.0) to a final concentration of 50 mM, then applied at 2.0-ml
min
1 to a 5-ml column comprised of bovine pancreatic
kallikrein (70 mg) which had been immobilized onto CNBr-activated
Sepharose 4B (Pharmacia Biotech Inc.) according to the manufacturer's
instructions. After loading, the column was washed with 0.1 M Tris-HCl (pH 8.0) containing 0.1 M NaCl until
the A280 nm of the wash could no longer be
detected. The column was washed further with 0.1 M Tris-HCl
(pH 8.0) containing 0.5 M NaCl and then eluted with 0.2 M acetic acid (pH 2.0). Fractions (2.0 ml) were collected in tubes containing 0.25 ml of 1 M Tris-HCl (pH 7.5) and
assayed for the ability to inhibit bovine trypsin as described (6). Active fractions were pooled, adjusted to pH 2.5 with trifluoroacetic acid and subjected to chromatography on a C18 reverse-phase
column (1.0 × 25 cm) equilibrated with 0.1% trifluoroacetic acid
containing 22.5% acetonitrile at a flow rate of 1 ml/min. The
placental bikunin(1-170) was eluted with a linear gradient
of 22.5-50% acetonitrile in 0.1% trifluoroacetic acid over 50 min.
Fractions containing trypsin inhibitory activity were pooled,
lyophilized, redissolved in 5 mM sodium acetate (pH
5.0), 0.1 M NaCl, and stored at
20 °C until needed.
The highly purified protein exhibited an
280 nm of
4.52 × 104 liters M
1
cm
1 based upon composition analysis.
-Fmoc-N-
-(2-Fmoc-oxy-4-methoxybenzyl) when it occurred after aspartic acid. Cleavage and deprotection were
performed in 84.5% trifluoroacetic acid, 4.4% thioanisole, 2.2%
ethanedithiol, 4.4% liquefied phenol, and 4.4% H2O for
2 h at room temperature. The crude peptide was precipitated with t-butyl methyl ether, centrifuged at 3,000 rpm in a Sorval
RT 6000D model table top centrifuge for 5 min, and washed twice in t-butyl methyl ether. Reduced peptides were purified by HPLC
(10 ml/min) on a Dynamax 60A C18 column (2.1 × 30 cm)
using a 30-min linear gradient of either 23-34% acetonitrile in 0.1%
trifluoroacetic acid (placental bikunin(7-64)) or a 30-min
linear gradient of 20-30% acetonitrile in 0.1% trifluoroacetic acid
(placental bikunin(102-159)).
20 °C until
needed.
I) and
stoichiometries obtained from plots of
V(+I)/V(
I)
versus ratio of inhibitor to enzyme
concentrations.
-guanidinobenzoate as described (11). The concentration of bovine chymotrypsin was determined by
titration with N-trans-cinnamoylimidazole as described (12). The amount of the NH2-terminal(7-64) and
COOH-terminal (102-159) Kunitz domain, and aprotinin were
determined by titration with active site-titrated trypsin and by amino
acid analysis. The concentration of placental
bikunin(1-170) was quantified by amino acid analysis. The
concentrations of factors VIIa, IXa, X, Xa, XIa, and XIIa, tissue
factor, elastase, urokinase, and tPA were based on the manufacturer's
specifications. Following preincubation of protease with inhibitor for
5 min at 37 °C in 990 µl of the appropriate buffer (see below),
reactions were initiated with substrate to achieve the following
initial component concentrations: bovine pancreatic kallikrein (buffer
B) with [E0] = 92 pM and 100 µM P-F-R-AMC (Km = 82 ± 8.8 µM); human tissue kallikrein, 50 mM Tris-HCl
(pH 9.0), 50 mM NaCl, and 0.01% Triton X-100 with [E0] = 0.35 nM and 10 µM P-F-R-AMC (Km = 5.7 ± 0.4 µM); human neutrophil elastase, 0.1 M
Tris-HCl (pH 8.0), and 0.05% Triton X-100 with
[E0] = 19 nM and 0.6 mM Suc-A-A-A-AMC (Km = 1,300 µM); human factor Xa, 20 mM Tris-HCl (pH
8.0), 0.1 M NaCl, 0.1% bovine serum albumin with
[E0] = 0.87 µM and 0.6 mM Boc-L-G-R-AMC (Km = 750 ± 98 µM); and human factor XIa (buffer A with 1 mg/ml bovine
serum albumin) with [E0] = 0.1 nM
and 400 µM Boc-E(OBzl)-A-R-AMC (Km = 450 ± 50 µM). The following assays were performed following preincubation of protease with inhibitor for 30 min at
37 °C: trypsin (buffer A) with [E0] = 50 pM and 30 µM Tos-G-P-K-AMC (Km = 22 ± 2 µM); chymotrypsin
(buffer A) with [E0] = 50 pM and
80 µM Suc-A-A-P-F-AMC (Km = 70 ± 12 µM); human plasmin, 50 mM Tris-HCl (pH
7.5), 0.1 M NaCl, and 0.02% Triton X-100 with
[E0] = 50 pM and 500 µM Tos-G-P-K-AMC (Km = 726 ± 70 µM); human plasma kallikrein (buffer B) with
[E0] = 0.2 nM and 100 µM P-F-R-AMC (Km = 457 ± 28 µM); and factor XIIa (buffer B) with
[E0] = 4.0 nM, and 200 µM Boc-Q-G-R-AMC (Km = 200 ± 37 µM). Factor VIIa-tissue factor inhibition was measured by
incubating enzyme (2.8 nM factor VIIa, 2.8 nM
tissue factor) with inhibitor in 100 µl of buffer A for 30 min at
25 °C. Reactions were initiated with 1 mM final
(CH3SO2-D-cyclohexylpyrosyl)-G-R-pNa, and the A405 nm was monitored. Factor IXa
inhibition was measured as described (5). Briefly, factor IXa (10 nM) was preincubated for 15 min at 37 °C with inhibitor
in 50 mM Tris-HCl (pH 7.5) containing 0.1 M
NaCl, 250 µM CaCl2, 60 nM
poly-D-lysine (Mr = 209,200) and
0.1% polyethylene glycol 8000. Factor X (200 nM) was then
added and the incubation continued for 30 min at 37 °C. Substrate
(600 µM Boc-L-G-R-AMC) was then added and factor Xa
activity measured. To determine the inhibition of tPA, tPA was
preincubated with inhibitor for 2 h at room temperature in 20 mM Tris-HCl (pH 7.2) containing 150 mM NaCl and
0.02% sodium azide. Reactions were initiated with substrate to achieve
initial component concentrations of: tPA (16.7 nM),
inhibitor (0-6.6 µM), I-P-R-pNa (1 mM) in 28 mM Tris-HCl (pH 8.5) containing
0.004% (v/v) Triton X-100 and 0.005% (v/v) sodium azide. The amount
of pNa formed after incubation for 2 h at 37 °C was
determined from the increase in the A405 nm.
Hydrolysis of AMC-conjugated peptides was monitored on a Perkin-Elmer
model LS50B fluorometer (excitation = 370 nm, emission = 432 nm) over the first 2 min of the reaction, while hydrolysis of
pNa conjugates was monitored at 405 nm on a Hewlett-Packard
model HP8452 spectrophotometer. Ki* values were
determined from plots of fractional rate versus inhibitor
concentration, which were fit by nonlinear regression analysis
(Enzfitter by Biosoft, Cambridge, U. K.) using the following equation
(Eq. 1)
where Vi and V0 are
the enzyme activities in the presence or absence of a total inhibitor
concentration of [I]0, and [E]0
is the total concentration of enzyme. Ki values were
obtained by correction for the effect of substrate using the following
equation: Ki = Ki*/(1 + [S]0/Km). Ki values
for inhibition by placental bikunin(1-170) were determined
assuming an inhibitor:protease stoichiometry of 1:2 and that the two
Kunitz inhibitory domains are equivalent and act independently.
The effect of recombinant placental bikunin(1-170) and aprotinin on the APTT was determined as follows. Inhibitor in 20 mM Tris-HCl (pH 7.2) containing 150 mM NaCl and 0.02% sodium azide was added (0.1 ml) to a cuvette within a MLA ElectraR 800 Automatic Coagulation Timer coagulometer (Medical Laboratory Automation, Inc., Pleasantville, NY). The instrument was set to APTT mode with a 300-s activation time, and samples were run in duplicate. Following the addition of 0.1 ml of plasma (Specialty Assayed Reference Plasma 1-6-5185, Helena Laboratories, Beaumont, TX), the APTT reagent (Automated APTT-lot 102345, Organon Teknika Corp., Durham, NC) and CaCl2 (final concentration of 25 mM) were automatically added to initiate clotting, which was monitored automatically.
Analytical ProceduresCNBr digestion was performed as follows. Purified placental bikunin(1-170) in water (165 pmol) was taken to dryness on a Speed-Vac concentrator (Savant Instruments, Farmingdale, NY), resuspended in 0.05 ml of 250 mM Tris-HCl containing 6 M guanidine HCl, 0.5% (v/v) 2-mercaptoethanol, and 1 mM EDTA, and incubated for 2 h in the dark at room temperature before the addition of 4% (v/v) final 4-vinylpyridine. Following an additional 2-h incubation, the sample was applied to a biphasic reaction cartridge containing C18 (Hewlett-Packard), eluted in methanol, then taken to dryness by Speed-Vac concentration. The sample was incubated with 0.05 ml of 70% (v/v) formic acid saturated with CNBr for 12 h, taken to dryness, reconstituted in 0.1 ml of 0.1% (v/v) trifluoroacetic acid, then subjected to NH2-terminal sequencing. NH2-terminal sequence analysis was performed essentially as described (13), but with minor modification (6). Amino acid analysis was performed as described previously (14). Reducing SDS-polyacrylamide gel electrophoresis was performed with 10-20% Tricine-buffered polyacrylamide gels (Novex, San Diego, CA) according to the manufacturer's instructions, and protein was visualized with Coomassie Brilliant Blue R-250 using standard protocols. Electrospray ionization mass spectrometry was performed on a Finnigan TSQ7000 mass spectrophotometer capable of unit resolution, with the following operational parameters: scan rate, 200-2,000 over 3 s; spray voltage, 5 kV; capillary temperature, 220 °C; nitrogen pressure, 50 p.s.i.; auxillary nitrogen, 30 ml/min. Samples were loaded on a PLRP cartridge (Microm Bioresources, Auburn, CA), washed with water, and eluted at 0.2 ml/min with 90% (v/v) acetonitrile, 2% (v/v) acetic acid.
To explore the protease inhibitory capacity of placental bikunin, a cDNA encoding the entire extracellular domain of 170 amino acids of placental bikunin plus its natural signal peptide (6) was prepared by site-directed mutagenesis and expressed in the baculovirus/Sf9 system. The protein was purified from the cell culture supernatant by sequential kallikrein-Sepharose affinity chromatography and C18 reverse-phase HPLC (Table I). An overall enrichment of 240-fold was achieved based on the increase in specific activity; however, the actual fold enrichment of the protein was likely to be much higher as only a fraction of the inhibitory activity of the culture supernatant was due to placental bikunin(1-170). This latter conclusion is based on the fact that the majority (>99%) of the trypsin inhibitory activity present in the starting supernatant did not bind to the kallikrein-Sepharose column, whereas all the placental bikunin(1-170) bound the column based on immunoblot analysis of the column starting material and flow-through (not shown). Fractionation of 2 liters of medium typically yielded between 0.7 and 1 mg of placental bikunin(1-170) based on amino acid analysis.
|
The final preparation was highly pure as judged by SDS-polyacrylamide
gel electrophoresis (Fig. 1) and exhibited a molecular mass of 21.3 kDa, consistent with the expected size of placental bikunin(1-170). NH2-terminal sequence analysis
of the protein (26 cycles) yielded the same NH2 terminus
for as obtained for the natural protein (6). The sequence started at
residue +1 and continued with the sequence ADRER-. Furthermore, the
amino acid composition was > 95% accurate relative to the
theoretical composition of placental bikunin(1-170).
Purified placental bikunin(1-170) (100 pmol) was then
pyridylethyl alkylated, CNBr digested, and then sequenced without
purification of the resulting peptide fragments. Sequencing for 20 cycles yielded the sequence for each of the NH2 termini
expected from the digestion of placental bikunin(1-170)
(Table II).
|
Titration of fixed concentrations of either bovine trypsin or human
plasma kallikrein with purified placental bikunin(1-170) showed that each molecule of inhibitor can interact with two molecules of protease simultaneously (Fig. 2). On the other hand,
titration of trypsin with aprotinin yielded a stoichiometry of 1:1 in
the enzyme inhibitor complex (Fig. 2).
To attempt to dissect the specificities of the individual domains within this bikunin, synthetic peptides were made corresponding to the NH2-terminal(7-64) and COOH-terminal(102-159 Kunitz domains of placental bikunin. Purified, reduced NH2-terminal(7-64) ([M-H]+ = 6,567.5), and COOH-terminal(102-159) domain ([M-H]+ = 6,835.6) were prepared, each of which had the expected amino acid composition. Refolding of these peptides using dimethyl sulfoxide as the oxidizing agent, followed by purification on a C18 column, yielded purified refolded NH2-terminal(7-64) (2% of the starting reduced peptide) and COOH-terminal(102-150) domains (1.4% of the starting reduced peptide) which exhibited [M-H]+ values of 6,561.2 and 6,829.3, respectively. The mass reduction upon refolding of each peptide (6 ± 1 mass units) suggests that refolding had in each case resulted in the formation of three intrachain disulfide bonds from the six cysteines present within each of the reduced peptides.
A panel of serine proteases was used to determine the relative specificity and potency of recombinant placental bikunin(1-170) and its individual Kunitz domains (Table III). Aprotinin, a Kunitz inhibitor of defined specificity, was included in these studies for comparison. Recombinant placental bikunin(1-170) was a potent inhibitor of several trypsin-like proteases and also of chymotrypsin. Of potential relevance to its putative physiological role, this protein was a potent inhibitor of human plasma and tissue kallikreins, human plasmin and factor XIa. The potency of recombinant placental bikunin(1-170) against plasmin, plasma kallikrein, and factor XIa were 2-, 45-, and 47-fold greater, respectively, than observed with aprotinin. Recombinant placental bikunin(1-170) was also a more potent inhibitor than aprotinin against several other enzymes of the intrinsic pathway of coagulation, including factors IXa, Xa, and XIIa, although the Ki values were in the hundred nanomolar range.
|
Collectively, these data suggest that recombinant placental
bikunin(1-170) should be a more effective inhibitor of the intrinsic coagulation pathway than aprotinin. To evaluate this hypothesis further, placental bikunin(1-170) and aprotinin were compared for their capacity to prolong the APTT, a measure of the
intrinsic pathway of coagulation. As predicted from the in
vitro specificity data (Table III), recombinant placental
bikunin(1-170) was significantly more potent than
aprotinin in prolonging the APTT (Fig. 3). To achieve a
doubling of the clotting time, at least 28 µM aprotinin
but only 0.7 µM recombinant placental
bikunin(1-170) was required.
Recombinant placental bikunin(1-170) was a poor inhibitor of the factor VIIa-tissue factor complex and factor Xa, suggesting that it may be a poor inhibitor of the extrinsic pathway of coagulation. This prediction was confirmed by the finding that placental bikunin(1-170) had no significant effect on the prothrombin time when tested at concentrations up to 2 µM (data not shown). Recombinant placental bikunin(1-170) did not inhibit urokinase, tPA, or elastase at concentrations up to 1 µM (data not shown).
The individual Kunitz domains within placental bikunin were each functional protease inhibitors and exhibited roughly similar specificities, although the NH2-terminal(7-64) Kunitz domain was significantly less potent against tissue kallikrein and factor XIa. Only versus factor XIIa was the NH2-terminal domain more potent than the COOH-terminal domain. Overall, the specificity of the COOH-terminal Kunitz(102-159) domain more closely resembled that of recombinant placental bikunin(1-170) except that on average the potency of the COOH-terminal Kunitz(102-159) for each protease was approximately 5-fold lower.
In this report we describe the expression and functional characterization of placental bikunin. Because the natural protein was available only in low amounts (6) the functional studies were performed instead with a recombinant form of the protein. A cDNA sequence encoding a 170-residue placental bikunin fragment containing both Kunitz domains but truncated immediately prior to the putative transmembrane segment (6) was expressed in Sf9 cells. This was done to alleviate concerns that expression of the full-length protein would complicate its recovery as a soluble protein. The putative signal peptide encoded within the cloned cDNA (6) was recognized as such by the Sf9 system as judged by its absence from the NH2-terminal sequence of the purified protein. For comparison with the recombinant protein, the individual Kunitz domains were prepared synthetically and effectively folded as judged by reductions in mass consistent with formation of the expected number of intrachain disulfide bonds. Recombinant placental bikunin(1-170) and the synthetic NH2-terminal(7-64) and COOH-terminal(102-159) Kunitz domains were all potent serine protease inhibitors. Each protein was a potent inhibitor of trypsin, chymotrypsin, plasmin, factor XIa, and tissue and plasma kallikrein, yet a relatively weak inhibitor of factors IXa, Xa, and XIIa. Stoichiometry studies with placental bikunin(1-170) verified that both Kunitz domains were functional and could direct the formation of ternary complexes with selected target proteases. The fact that placental bikunin(1-170) was functional lends credence to the notion that the hydrophobic segment encoded by the full-length cDNA acts as a membrane anchor and does not participate in the globular structure necessary for protease inhibition.
Although placental bikunin(1-170) potently inhibited two key enzymes, kallikrein and factor XIa, of the contact activation/intrinsic pathway, it was an extremely weak inhibitor of factor VIIa-tissue factor complex. This suggests that it is unlikely involved in the regulation of the extrinsic coagulation cascade. Consistent with these observations, placental bikunin (1-170) was extremely potent in extending the APTT time, a measure of the rate of coagulation driven by the intrinsic pathway, but did not extend the prothrombin time, a measure of the rate of coagulation driven by the extrinsic pathway. Based on these findings, placental bikunin exhibits an in vitro activity profile that might be expected of a physiological regulator of both contact activation and the intrinsic pathway of coagulation. On the other hand, the kallikreins play a key upstream role in initiating fibrinolysis through activation of plasmin, which in turn participates directly in fibrinolysis. The high potency of placental bikunin(1-170) against the kallikreins and plasmin suggests that this inhibitor could function to stabilize fibrin clots. The possibility that the native full-length protein might associate with biological membrane is consistent with the notion that it may serve to regulate the processes of contact activation and fibrinolysis since elements of these processes take place at damaged tissue surfaces.
In general, placental bikunin(1-170) exhibited Ki values that were significantly lower than those observed with the individual domains. Similar results have been obtained with TFPI wherein the recombinant protein containing all three of the Kunitz domain had anywhere from a 2-fold to 10-fold greater affinity for the factor VII-tissue factor complex than any of the individual domains (15). Furthermore, cooperatively has been demonstrated for the interaction of TFPI with its target proteases. Specifically, Girard and Broze demonstrated that TFPI inhibits the tissue factor-factor VIIa complex 50-fold more potently when first allowed to complex with factor Xa (16). From a biological perspective this is consistent with a feedback inhibition mechanism allowing some active factor Xa to form before the entire extrinsic pathway terminates. The possibility that placental bikunin may regulate the intrinsic pathway through a similar cooperative mechanism warrants further investigation. We have demonstrated that although the NH2- and COOH-terminal Kunitz domains have similar potencies versus a number of proteases, the COOH-terminal Kunitz domain is 10-fold more potent against factor XIa. Perhaps the COOH-terminal domain of placental bikunin inhibits factor XIa in vivo, which in turn triggers the subsequent formation of a ternary complex through interaction of the NH2-terminal domain with factor XIIa or kallikrein. This would serve as a feedback inhibition mechanism for the extrinsic pathway where some factor XIa is allowed to form before the entire pathway is shut down. On the other hand, the differences in potency against factor XIa between the synthetic NH2- and COOH-terminal Kunitz domains could reflect the fact that the putative N-linked carbohydrate residue within the NH2-terminal domain of native placental bikunin (6) plays a role in determining the specificity of this domain.
Blockage of fibrinolysis and extracorporeal coagulation have been observed in open heart surgery patients receiving aprotinin (17, 18), and these effects are believed to arise from the inhibition of kallikrein, plasmin, and the intrinsic pathway of coagulation (19). Because placental bikunin(1-170) is significantly more potent than aprotinin as an inhibitor of plasma kallikrein, and the intrinsic pathway of coagulation, further exploration of the use of this novel human bikunin in this clinical setting is warranted. In addition, placental bikunin should also be considered for other indications for which aprotinin has been shown to be beneficial preclinically. One potential application is the blockage of the plasmin-dependent invasion of cancer cells into surrounding tissue (20). In addition, placental bikunin may be useful for the reduction of severity of brain edema secondary to brain or spinal chord injury (21, 22), a process that may result from the local release of the vasodilator bradykinin (21-23) through the actions of serine proteases such as kallikrein (24).
We acknowledge the following members of the research staff at the Bayer Research Center for important contributions: Carla Pellegrino (NH2-terminal sequencing), Thomas Buckholz (peptide synthesis), John Kupcho (amino acid analysis), Susan Barkowski-Clark and Bradley Rose (Sf9 expression), and Tony Paiva (mass spectrometry).