: H. Herwald, Department of Cell and Molecular Biology, Section for Molecular Pathogenesis, Lund University, Tornavägen 10, S-221 84 Lund, Sweden. Fax: + 46 46 157756, Tel.: + 46 46 2224522, E-mail: Heiko.Herwald@medkem.lu.se
Human high molecular mass kininogen (HK) participates as nonenzymatic cofactor in the contact system. Here, we show that recombinant domain D5 of HK (rD5) prolongs the clotting time of the intrinsic pathway of coagulation and attenuates the generation of bradykinin. Further studies indicate that a correct fold of domain D5 within HK is required for the activation of the contact system. The folding of rD5 seems to be modulated by the metal ions Zn2+, Ni2+, and Cu2+ as a specific antibody directed against the zinc-binding site in HK binds to HK and rD5 in a metal ion concentration dependent manner. The finding that these three metal ions specifically affect contact activation suggests that they regulate the accessibility of rD5 for negatively charged surfaces. Support for the assumption that the observed phenomena are due to conformational changes was obtained by fluorescence spectroscopy of rD5, demonstrating that its fluorescence spectrum was changed in the presence of ZnCl2. Moreover, negative staining electron microscopy experiments suggest that the zinc-induced changes in D5 also affect the conformation of the entire HK protein. The present data emphasize the role of zinc and other metal ions in the regulation of contact activation.
The human contact system comprises the three serine proteinases, factor XII (F XII), factor XI (F XI), and plasma kallikrein (PK) as well as the nonenzymatic cofactor high molecular mass kininogen (HK) (reviewed in ). Similar to other serine proteinase-triggered cascades, such as the complement system, the contact system is initiated by limited proteolysis . As indicated in Fig. 1, the contact system is activated on negatively charged surfaces . Initially, F XII is autocatalytically converted to its active form (F XIIa). This is followed by an activation of PK, anchored to the surface via HK, by F XIIa. PKa amplifies the activity of F XIIa and cleaves HK leading to release of the nonapeptide bradykinin. The peptide is considered to be a primary mediator of inflammatory processes involved in the production of pain, induction of vasodilatation and increase of vascular permeability . Finally, F XIIa triggers a conversion of F XI to F XIa leading to the initiation of the intrinsic pathway of coagulation. The contact activation on cellular surfaces is less clear. However, the mechanism of activation may be different from that on negatively charged surfaces . Originally, it was thought that the contact system has an important function in the initiation of blood coagulation, as deficiencies in any of the contact factors lead to abnormal high clotting times of the intrinsic pathway of coagulation in in vitro assays. However, no hemorrhagic disorders are observed in patients with deficiency in HK, F XII, or PK, indicating that the contact system plays a secondary role in hemostasis. In contrast, a growing body of evidence suggests that the contact system functions as an early pro-inflammatory effector system in pathophysiological processes .
The HK molecule is composed of two segments: a heavy chain containing four domains (D1–D4) and a light chain comprising the domains D5 and D6 . The first three domains of the heavy chain (D1–D3) have a cystatin-like structure. Domains D2 and D3 serve as specific inhibitors of a number of cysteine proteinases such as cathepsins and papain, whereas D1 has no proteinase-inhibitory activity . Domain D4 holds the kinin segment, which is released as bradykinin from the intact molecule following cleavage by PKa. The light chain of HK contains two domains. Domain D5 contains binding sites for zinc and for cellular and negatively charged surfaces (Table 1); whereas binding sites for PK and F XI are located in domain D6 [8–11].
a The first three amino acids followed by the total number of amino acids in the peptide are used to name each peptide. b The one-letter code for amino acids residues is used. c Zinc binding site of HK . d Cell binding site of HK .
The binding of HK to cellular surfaces is dependent on zinc [12,13]. It has been postulated that the presence of the metal ion is necessary for the correct folding of the surface-binding site in HK . The present study was therefore undertaken to investigate whether conformational zinc-induced changes in domain D5 may effect the functional activities of the contact system.
Materials and methods
Sources of proteins, antibodies, and peptides
HK was from Enzyme Research Laboratories Ltd. Swansea, UK. Peptides GCP28, GHG19, and HKH20 were synthesized in the Proteinchemisches Zentrallabor of the Johannes Gutenberg University (Mainz Germany) [9,15]. The monoclonal antibody HKL12 (in the following referred to as α-GHG19), directed against human HK, was produced in mice  and polyclonal antiserum AS88 to human HK in sheep . The polyclonal antiserum α-HKH20, directed against peptide HKH20, was produced in rabbit, following coupling of the cognate peptides to keyhole limpet hemocyanin (KLH) via the carbodiimide method . Peroxidase-conjugated goat anti-(rabbit Ig), goat anti-(mouse Ig) (Bio-Rad, Richmond, CA, USA), or donkey anti-(sheep Ig) (ICN, Aurora, OH, USA) were used as secondary antibodies.
Purification of the recombinant domain D5
For expression of rD5 a modified pET25b expression vector (Novagen, Inc., Madison, WI, USA) was used. The primers DK3 (5′-GCAGCAGTCATGACTGTAAGTCCACCCCACACTTCC-3′) and DK4 (5′-GCAGCAGGATCCACTGTCTTCAGAAGAGCTTGC-3′) were used to amplify a fragment encoding the D5 domain and part of the D6 domain. The fragment was cleaved with BspHI and BamHI and ligated to the vector, which was cleaved with NcoI and BamHI. The sequence of the construct was verified by DNA-sequencing using Sequenase version 2.0 DNA sequencing kit (Amersham, Arlington Heights, IL, USA). The protein was expressed in Escherichia coli strain BL21(DE3). Protein production was induced by addition of 1 mm isopropyl thio-β-d-galactoside to exponentially growing bacteria. After 3 h incubation bacteria were harvested by centrifugation. The pellet was resuspended in buffer A (50 mm phosphate 300 mm NaCl, pH 8.0). The bacteria were subsequently lysed by repeated cycles of freeze thawing. The lysate was then centrifuged at 29 000 g for 30 min. The supernatant was mixed with 2 mL Ni/nitrilotriacetic acid Sepharose (QIAGEN GmbH, Hilden, Germany) and incubated on rotation for 1 h. The Sepharose was loaded into a column and washed with 10 mL buffer A with 0.1% (v/v) Triton X-100, 10 mL buffer A, 5 mL buffer A with 1 m NaCl, 5 mL buffer A, 10 mL 20% ethanol, 10 mL buffer A containing 5 mm imidazole, and 10 mL buffer A containing 30 mm imidazole. The protein was eluted with buffer A containing 500 mm imidazole.
Proteins were separated by 12.5% (w/v) polyacrylamide gel electrophoresis in the presence of 1% (w/v) SDS . Standard molecular mass markers were from Sigma.
Western blotting and immunoprinting
Proteins were separated by SDS/PAGE and transferred onto nitrocellulose membranes for 30 min at 100 mA . The membranes were blocked with 50 mm KH2PO4, 0.2 m NaCl, pH 7.4, containing 5% (w/v) dry milk powder and 0.05% (w/v) Tween 20. Immunoprinting of the transferred proteins was performed according to Towbin et al. . The first antibody was diluted 1 : 6000 in the blocking buffer (see above). Bound antibody was detected by a peroxidase-conjugated secondary antibody to sheep Ig followed by the chemiluminescence detection method.
The molecular mass of rD5 was analyzed by a Bruker Reflex III MALDI-TOF mass spectrometer using the conventional dried droplet technique for sample preparation with ferulic acid as the MALDI matrix .
Amino-acid sequence analysis
N-terminal amino-acid sequences were determined using an Applied Biosystems model 477 A sequenator.
Indirect and competitive ELISA
Microtitreplates were coated by overnight incubation with 200 µL per well of a 1-µg·mL−1 antigen solution. Plates were washed five times using a Hepes buffer (50 mm Hepes, 0.15 m NaCl, 0.05% (v/v) Tween 20, pH 7.4). After blocking the plates with 200 µL per well of the washing buffer containing 2% (w/v) bovine serum albumin (incubation buffer) for 30 min at 37 °C, serial dilutions of the antisera (typical starting concentration 1 : 1000, v/v) or of the antibodies (1 µg·mL−1) in the incubation buffer were added to the plates for 1 h at 37 °C. Bound antibody was detected by a horseradish peroxidase labeled secondary antibody (1 : 3000, v/v; 1 h at 37 °C) and the chromogenic substrate solution 0.1% (w/v) 2,2′-azinobis-(3-ethylbenzo-6-thiazolinesulfonic acid (ABTS), 0.012% (v/v) H2O2 in 100 mm citric acid, 100 mm NaH2PO4, pH 4.5 for 30 min at 37 °C. Each incubation step was followed by a washing step. To study the binding of antibodies to HK and HK fragments in the presence of metal ions, plates were coated with the antigen (0.2–3 µg·mL−1) overnight followed by incubation of antisera (dilutions 1 : 8000 to 1 : 16 000) or antibodies (1–2 µg·mL−1) in the presence of a serial dilution of the different metal ions (starting concentration 1 mm).
The activated thromboplastin time (aPTT) was measured in a coagulometer (Amelung, Lemgo, Germany). 50 µL of a solution containing rD5 (2 mg·mL−1 in 12.9 mm sodium citrate pH 7.4), GHG19 (0.5 mg·mL−1), HKH20 (0.5 mg·mL−1), or GCP28 (0.5 mg·mL−1) were incubated with the aPTT reagent (Diagnostica Stago, Asnières, France) for 60 s at 37 °C followed by incubation with 50 µL of sodium citrate-treated normal human plasma for 200 s. To initiate clotting 100 µL of 25 mm CaCl2 was added. To monitor the thromboplastin time (PT), 200 µL of the PT reagent (Sigma Chemicals, St Louis, MO, USA) were incubated with proteins or peptides for 60 s at 37 °C followed by the initiation of clotting by addition of 50 µL plasma. When measuring the thrombin clotting time (TCT), 50 µL of the samples were added to 100 µL of the TCT reagent (Sigma) for 60 s at 37 °C and clotting was started by adding 150 µL of human plasma. The clotting times in the presence of metal ions were performed by preincubation of plasma (50 µL or 150 µL) with 50 µL of a solution containing 2 mm of each metal ion for 60 s at 37 °C prior to starting the coagulation by adding the different clotting reagents. Clotting assays were performed at least three times in duplicates.
Determination of kinin concentrations in plasma
The kinin concentrations in the reaction mixtures were quantitated by the Markit-A kit (Dainippon Pharmaceutical Co., Osaka, Japan) as described . Briefly, 2 µL of the plasma samples were added to 198 µL of a sodium citrate buffer (12.9 mm, pH 7.4) and mixed with 40 µL of the deproteinizing reagent. Samples were centrifuged at 1500 gfor 10 min. Aliquots of the supernatant (60 µL each) were mixed with 60 µL of the kit buffer, and applied to the wells (100 µL each) of microtitre plates coated with capture antibodies to rabbit Ig followed by specific antibradykinin Ig. After 1 h of incubation, the peroxidase-labeled bradykinin probe was applied and incubated for 1 h. The amount of bound peroxidase was visualized by the substrate solution, 0.1% (w/v) ABTS, 0.012% (v/v) H2O2 in 100 mm citric acid, 100 mm NaH2PO4, pH 4.5 for 30 min. The change of absorbance was read at 405 nm. The reference standards were prepared according to the manufacture's instructions. Measurements were performed at least three times, in duplicate each time.
A Perkin-Elmer LS50B luminescence spectrophotometer was used to measure fluorescence intensities of rD5. The excitation wavelength was set at 280 nm and the emission spectra were collected for 290–450 nm. Measurements were performed at 8.7 µm rD5 in 50 mm Hepes and 150 mm NaCl pH 7.4 with or without 1 mm ZnCl2.
Transmission electron microscopy
HK samples were incubated in 150 mm NaCl, 50 mm NaCl/Tris, pH 7.4 (TBS), containing 1 mm EDTA, MgCl2 or ZnCl2, respectively. 5 µL samples (typical concentrations 5–10 µg·mL−1) were adsorbed to 400 mesh carbon-coated copper grids for 1 minute, washed briefly with two drops of water, and stained with two drops 0.75% uranyl formate. The grids were rendered hydrophilic by glow discharge at low pressure in air. Specimens were observed in a Jeol 1200 EX transmission electron microscope operated at 60 kV accelerating voltage. Images were recorded on Kodak SO-163 plates without preirradiation at a dose of typically 2000 electrons·nm−2. Evaluation of the data from electron micrographs was performed as described previously .
Expression and characterization of rD5 in E. coli
To study the role of domain D5 of HK in contact activation, a recombinant polypeptide corresponding to this region (Table 1) was cloned in an E. coli expression vector. The domain was purified to homogeneity by affinity chromatography on a Ni2+- Sepharose column and its identity was verified by Western blot analysis (Fig. 2) and amino-terminal sequencing of the first five amino acids (Met-Thr-Val-Ser-Pro). Characterization of rD5 by MALDI-TOF mass spectrometry showed a molecular mass of 15.011 ± 0.002 kDa, which correlates well to the mass predicted from the deduced amino-acid sequence of the construct (15.009 kDa).
Immunological characterization of D5
Domain D5 contains binding sites for cellular and negatively charged surfaces as well as for zinc (Table 1). In a series of experiments, a number of antibodies directed against these segments in HK were analyzed in indirect ELISA in order to test whether they recognize the corresponding epitopes in rD5. As shown in Fig. 3A a polyclonal antiserum to HK (AS88) bound strongly to both HK and rD5 but only demonstrated weak reactivity with the D5-derived synthetic peptides GHG19, and HKH20. In contrast, peptide GCP28 from domain D3 was not recognized by AS88. Antibodies α-GHG19, directed against peptide GHG19 , and α-HKH20, directed against peptide HKH20 , were reactive with the corresponding peptides, but also cross-reacted with HK and rD5. The specificity of the antibodies was demonstrated by the failure of both antibodies to bind to control peptides (Fig. 3B,C). A control antiserum directed against GCP28, recognized HK and the peptide against which it had been raised, but failed to bind to rD5 or the two D5-derived peptides. (Fig. 3D). These data indicate that rD5 contains the information required for the recognition by specific antibodies to the surface and zinc-binding sites, respectively.
Influence of rD5 on the clotting time
Domain D5 plays an important role in the regulation of the intrinsic pathway of coagulation . Recent work with peptide HKH20 has demonstrated that the cell-surface binding site in domain D5 interferes with the procoagulant activity of HK . The functionality of rD5 was therefore analyzed by its ability to inhibit coagulation initiated by contact activation. As shown in Fig. 4A, similar to peptide HKH20, rD5 prolonged the intrinsic clotting time (aPTT, activated partial thromboplastin time) of normal human plasma. In contrast, the peptides GHG19 and GCP28 failed to influence the aPTT. Neither rD5 nor peptides HKH20, GHG19, and GCP28 attenuated the extrinsic pathway of coagulation (PT, prothrombin time) or thrombin induced fibrin polymerization (TCT; thrombin clotting time) (Fig. 4B,C). Taken together these experiments show that the surface binding site in rD5 is functionally active and can specifically block activation of the intrinsic pathway of coagulation.
Characterization of the zinc-binding site in rD5
The presence of zinc ions is required for binding of HK to cellular surfaces . The monoclonal antibody α-GHG19 directed to the heavy chain of HK binds to a region in domain D5 that contains the zinc-binding motive (HGLGHGHEQQHGLGHGH) . Figure 5A shows that antibody α-GHG19 lost its ability to bind to HK in the presence of high concentrations of ZnCl2. Similar to HK, rD5 and the peptide GHG19 were recognized by antibody α-GHG19 in a zinc-concentration dependent manner. In contrast, the binding of anti-GCP28 to HK or to GCP28 was unaffected by zinc ions. (Fig. 5B). As indicated by the finding that rD5 can be purified on Ni2+-sepharose (Fig. 2), Ni2+ also affected the binding of α-GHG19 to rD5 in a concentration dependent fashion, as did Cu2+. In contrast, other metal ions (calcium, cobalt, iron, magnesium, and manganese) had no influence on the α-GHG19–rD5 interaction (Fig. 5C). Interestingly, the same metal ions that interfered with the zinc-binding site in domain D5 (zinc, copper, and nickel) also prolonged the aPTT when added to plasma at high concentrations (Fig. 6A). In contrast, the PT and TCT were not significantly influenced by these ions; only the addition of ZnCl2 to plasma led to a small increase in the PT (Fig. 6A,B). These data suggest that the zinc-binding site remains intact in rD5. Furthermore, they also indicate that a zinc-dependent folding of domain D5 in the HK molecule is required for the activation of the intrinsic pathway of coagulation.
Attenuation of the generation of bradykinin by rD5
Kinins are generated as a consequence of the contact activation at the site where contact factors are assembled . We therefore analyzed whether rD5 can attenuate the formation of bradykinin in kaolin-treated plasma. Figure 7A shows that preincubation of the kaolin reagent with rD5 and peptide HKH20 prior to its addition to plasma led to a drastic reduction in the generation of bradykinin. In contrast, preincubation with control peptides was without any effect. The attenuation of bradykinin formation was time dependent and even five minutes after activation of plasma by a mixture of kaolin and rD5, a significant reduction was observed (data not shown). Thus, as was the case in the aPTT assay, addition of ZnCl2 to plasma led to disturbance of the contact activation resulting in diminished production of bradykinin (Fig. 7B). Also NiCl2, but not CuCl2, impaired bradykinin generation. As for the intrinsic driven clotting time, the formation of bradykinin seems to be dependent on a correct folding of domain D5. Together these data suggest that rD5 preserves the immunological and functional characteristics of domain D5, emphasizing the importance of this region both for the procoagulant and proinflammatory activity of HK.
Zinc induces conformational changes in D5 and HK
The inhibitory influence of high zinc-concentrations on contact activation together with the specific ability of this metal ion to block the binding of α-GHG19 to the zinc-binding sequence, suggested that zinc induces a conformational change in D5. To study if this is the case we used fluorescence spectroscopy. As the rD5 construct contains two tryptophan residues (at position 412 and 500, respectively) excitation at 280 nm results in fluorescence. The measurements demonstrated that the emission spectrum of rD5 was clearly shifted in the presence of 1 mm ZnCl2(Fig. 8). Thus the fluorescence emission maximum of rD5 alone appeared at 355 nm whereas upon addition of ZnCl2 the emission maximum was registered at 343 nm. As both the tryptophan residues are found outside the zinc-binding motif this indicates that zinc induces a conformational change of rD5 and that the change in emission spectrum is not due to direct effects of zinc.
To explore the influence of zinc on the overall structure of HK, the molecule was subjected to negative staining in the absence or presence of ZnCl2. Previous studies by Weisel et al. using glycerol spraying/rotary shadowing demonstrated that HK appears as a linear array of three linked globular regions with the two ends connected by another thin strand . After negative staining in the presence of EDTA large fields of uniform particles were visualized with a structure similar to that described by these authors (Fig. 9A). Note that noncleaved (circular, indicated by arrows) and cleaved (linear, indicated by arrowheads) forms of HK are clearly distinguishable by this technique. The noncleaved form accounted for about 70% of total HK. Conversion to the cleaved form occurred presumably during absorption to the carbon surface. In EDTA- or magnesium-containing buffers HK appeared as an egg-shaped circular array of five globular domains with a diameter of 9.5 ± 3 nm (width) and 15 ± 3 nm (length) (Fig. 9B,C). Two larger (5.5 ± 2 nm) and three smaller (3.5 ± 2 nm) domains were distinguished, probably representing the light and the heavy chains, respectively. In the presence of ZnCl2, HK appeared more spherical and compact with reduced diameter of 9 ± 3 nm (width) and 11 ± 3 nm (length) (Fig. 9D). Taken together, negative staining electron microscopy suggests that the binding of zinc to domain D5 induces a conformational change in the entire HK molecule. This modification can modulate the activity of HK in the activation of the contact system.
D5 plays a central role in contact activation by virtue of its capacity to mediate binding of HK to negatively charged cellular or subendothelial surfaces. The objective of the present study was to investigate the influence of zinc, known to interact with D5, upon contact activation, thereby establishing a functional link between the negative charge binding and the zinc binding properties. The recombinant D5 region used for these studies appeared to be immunologically intact, as it reacted with antibodies directed against the negative surface-binding sequence as well as the zinc-binding region. Interestingly, the binding of the antibodies raised against the zinc-binding region in D5 showed a concentration-dependent inhibition by zinc. The inhibition was highly specific and could only be mimicked by two other metal ions (nickel and copper). Three explanations for this phenomenon may be considered: (a) zinc binds directly to the region, thereby masking the epitope, (b) zinc induces conformational changes in the epitope, or (c) zinc binds to, or alters the conformation of the antigen-combining site in the antibody.
The findings in the functional assays measuring contact activation support the observations made in the antibody-binding assay: similar to rD5, zinc efficiently blocked activation of the contact system. Again inhibition was highly specific for certain divalent cations and, importantly, the set of ions that blocked contact activation was the same as the ones capable of inhibiting the binding of the antibody α-GHG19 to the zinc-binding site in D5. The similar effects of D5 and zinc suggest that the ions may indirectly influence the ability of the cell-binding site to interact with negative charges, lending support to the second alternative of the three, namely that zinc can induce structural changes in D5 influencing its biological properties.
We were able to obtain direct evidence for this hypothesis by fluorescence spectroscopy as rD5 contains two tryptophan residues both of which are located outside the zinc-binding site. Addition of zinc significantly shifted the emission spectrum of rD5 in a broad wavelength interval.
Moreover negative stain electron microscopy analysis suggested that the changes of the fold of D5 also effected the structure of the entire HK molecule. These studies demonstrated that addition of zinc evoked a conformational change in the protein. These findings are compatible with the observation that zinc inhibited the binding of α-GHG19 not only to rD5 but also to intact HK.
The observation made here are compatible with reports from Røjaker and Schousboe who showed that autoactivation of F XII on phosphatidylinositol phosphate or dextran sulfate occurs only in the presence of zinc . Similar to HK, F XII has affinity for zinc. The authors speculate that a zinc-induced conformational change is responsible for the autoactivation of F XII . Together, these data indicate that zinc plays an important role for the correct folding of the two surface binding proteins of the contact system.
In the present study, the contact system was triggered by the addition of kaolin. Apart from kaolin, other negatively charged surfaces, such as acidic phospholipids or dextran sulfates, have been demonstrated to activate the contact system in a zinc-dependent manner [26,27]. Recent investigations have revealed that the contact system can also be activated on bacterial surfaces as well as by the aggregated β amyloid protein [28–30]. As described for artificial surfaces the assembly of contact factors on microorganisms can lead to an activation of the system followed by the generation of bradykinin. Studies with the bacterial species Streptococcus pyogenes showed that the binding of HK to the streptococcal surface is mediated by the same binding-site that is involved in binding to cellular surfaces [9,31]. These data indicate that the cell-binding site in HK might also act as a recognition-site for invading microorganisms or for plaques followed by an activation of the contact system. Subsequent release of bradykinin at the site infection or deposition will lead to the formation of inflammatory reactions. Thus, in invasive bacterial infections an overwhelming activation of contact system can lead to serious complications, such as fever, pain, and hypotension. Whether or not the release of bradykinin from β amyloid protein plaques contributes to the extent of the disease is currently under investigation . It is possible that rD5 can be used as a substance for blocking the contact system activation on noxious surfaces. We are currently investigating whether rD5 is able to inhibit the generation of bradykinin triggered by bacteria. Future work will show whether rD5 or derivatives thereof can provide anti-inflammatory activity also in other systems.
We wish to thank Dr Müller-Esterl for generously providing us with antibodies to HK, Patrik Önnerfjord for analyzing rD5 by mass spectroscopy, and Monica Öberg for technical assistance. This work was supported in part by Active Biotech, Åke Wibergs Foundation, Alfred Österlunds Foundation, Anna-Greta Crafoords Foundation, Crafoordska Foundation, the Foundation for Strategic Research, Göran Gustafsson Foundation for Research in Natural Sciences and Medicine, Greta och Johan Kocks Foundation, King Gustaf V's 80th Anniversary Fooundation, The Royal Physiographic Society in Lund, Medical Faculty of Lund University, the Swedish Medical Research Council, and Tore Nilsons Foundation.