INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

HIGH-MOLECULAR-WEIGHT KININOGEN (HK) is a multidomain protein that has a number of important physiological activities. It is a precursor of bradykinin (BK) (24), it binds to platelets and endothelial cells through determinants on domains 3 and 5 (2, 7, 8, 14, 18, 23, 28, 31, 34), and it serves as a receptor for prekallikrein (PK) on endothelial cells (19). When PK binds to HK on endothelial cells, kallikrein is formed by a cell-associated, antipain-sensitive activator (16, 19). Kallikrein formation has the potential to liberate BK from its cell receptor HK (18, 19, 21). BK, a potent vasoactive agent, has the potential to induce endothelial cells to release nitric oxide (NO) (22). These data suggest that binding of HK to endothelial cells may influence their function as a result of kallikrein formation.

The interaction of kininogens with human umbilical vein endothelial cells (HUVEC) in culture has been described by binding studies using buffers that contain bovine serum albumin (BSA) (2, 10, 23, 28, 31). Kininogen binding to HUVEC in buffers containing BSA requires the addition of 25-50 µM Zn²⁺ (28, 31). Buffers that contain BSA bind substantial amounts of Zn²⁺ (26). However, recent studies indicate that only 4-10 µM Zn²⁺ is necessary for maximal PK activation in the absence of a carrier protein that binds Zn²⁺(26). Because PK binding and activation require the presence of HK, the actual Zn²⁺ requirement for HK binding to HUVEC in a buffer with a carrier protein that does not bind Zn²⁺ needs to be defined.

Zinc is known to participate in a number of activities of HK. The light chain of HK has a Zn²⁺ binding site on domain 5 that contributes to its binding to cells (3, 5). However, low-molecular-weight kininogen that does not contain a Zn²⁺ binding site also requires Zn²⁺ to bind to platelets and endothelial cells, suggesting that the Zn²⁺ requirement may also be related to the functional presentation of the kininogen receptor (3, 11, 17, 23, 34). We reexamined the requirements for Zn²⁺ in HK binding to HUVEC in buffer containing gelatin, a protein that does not bind Zn²⁺ (25, 26). These studies reveal that Zn²⁺ was required for HK binding, but at a concentration lower than that necessary to trigger maximal PK activation. When PK binds to HK on cells, kallikrein is formed that liberates BK from HK. Formed BK stimulates the BK B₂ receptor on endothelial cells to produce NO. These data indicate that the assembly and activation of the HK-PK complex on endothelial cells modulates their physiological activities.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Proteins, peptides, and antibodies. HK, PK, and plasma kallikrein were purchased from Enzyme Research Laboratory (South Bend, IN). Lisinopril and BK were purchased from Sigma (St. Louis, MO) and Dainippon Pharmaceutical (Osaka, Japan), respectively. Rabbit anti-BK B₂ receptor antibody was generously provided by Dr. Werner Muller-Esterl (Frankfurt, Germany). The B₂ BK receptor antagonist (B₂ inhibitor), D-Arg-Arg-Pro-Hyp-Gly-Thi-Ser-D-Tic-Oic-Arg, and a B₁/B₂ BK receptor antagonist (B₁/B₂ inhibitor), D-Arg-Arg-Pro-Hyp-Gly-Igl-Ser-D-Igl-Oic-Arg, were purchased from Phoenix Pharmaceuticals (Mountain View, CA). N^G-monomethyl-L-arginine salt (L-NMMA) and N^G-nitro-L-arginine methyl ester (L-NAME) were purchased from Calbiochem (San Diego, CA).

Biotinylation of HK. HK was biotinylated by using a previously reported method (9, 10). Briefly, 5 mg of single chain HK (120 kDa on reduced SDS-PAGE) in 4 mM sodium acetate-HCl and 0.15 M NaCl, pH 5.3, were dialyzed against 0.1 M sodium phosphate and 0.15 M NaCl, pH 7.2. The dialyzed HK was then incubated with a fivefold molar excess of EZ-Link Sulfo-NHS-LC-Biotin (Pierce Chemical; Rockford, IL) for 30 min at room temperature, followed by application onto a 10-ml Econo-Pac 10 DG column (Bio-Rad; Melville, NY). Biotinylated HK (biotin-HK) eluted from the column was monitored by absorbance at 280 nm by using an extinction coefficient of 7.0 and Bio-Rad protein assay. Biotin-HK had a specific activity of 17 U/mg. Biotin-HK with similar specific activity was also prepared by Enzyme Research Laboratory.

Preparation of BK-free HK. Liberation of BK from HK by plasma kallikrein was performed as previously described (18). Briefly, HK was dialyzed against 0.01 M Tris and 0.15 M NaCl, pH 7.8, overnight at 4°C and incubated with 200-fold molar excess kallikrein at 37°C for 16 h followed by the addition of 3 mM phenylmethylsulfonyl fluoride to stop the reaction. Identification of kinin-free HK was performed by electrophoresis of reduced protein on SDS-PAGE followed by staining with Coomassie blue.

Cell culture. HUVEC cultures were performed as previously described (9, 10, 28). HUVEC, which were purchased from Clonetics (San Diego, CA), were grown in endothelial cell growth medium (Clonetics), which contained 2% fetal bovine serum, 3 mg/ml bovine brain extract, and 10 ng/ml human recominant epidermal growth factor (Clonetics). Human microvascular endothelial cells (HMVEC), which were generously provided by Dr. Edwin W. Ades of the Center for Disease Control (Atlanta, GA) and obtained for us by Dr. Afshin Ameri, University of Michigan, were cultured in the same medium as HUVEC. Culture of human umbilical artery smooth muscle cells (UASMC) (Clonetics) was performed in smooth muscle cell basal medium-modified MCDB 131 (Clonetics) containing 5% fetal bovine serum, 2 ng/ml human fibroblast growth factor, 0.5 ng/ml human epidermal growth factor, and 5 µg/ml bovine insulin. Bovine pulmonary aortic endothelial cells (BPAEC) and normal human lung fibroblasts from the ATCC (Rockville, MD) were cultured in Dulbecco's modified Eagle medium containing 4.5 mg/ml D-glucose without sodium pyruvate, 1% fetal calf serum, and 1% penicillin-streptomycin-amphotericin mixture.

Binding of biotin-HK to HUVEC. Binding studies were performed on confluent HUVEC (4 × 10⁴ cells/well) in their second to fourth passage on fibronectin-coated, 96-well microtiter plates (Nunclon, Thomas Scientific; Swedesboro, NJ). Within 24 h of reaching confluence, the cells were washed three times in HEPES-Tyrode buffer (0.135 M NaCl, 2.7 mM KCl, 11.9 mM NaHCO₃, 0.36 mM NaH₂PO₄, 14.7 mM HEPES, pH 7.35) containing 100 mg/ml dextrose and 0.1% gelatin (HEPES-Tyrode gelatin buffer). When BSA was substituted for gelatin, it was added at 0.35%. In certain experiments the HEPES-Tyrode gelatin buffer was treated with Chelex-100 (Bio-Rad) according to the manufacturer instructions. Unless otherwise stated, all cells were incubated with 10 nM biotin-HK in HEPES-Tyrode gelatin buffer without any additions at 37°C for 20 min. Nonspecific binding was defined as the level of binding in the presence of 50-fold molar excess unlabeled HK. Specific binding was determined by substracting nonspecific binding from total binding. Cell-associated biotin-HK was measured using ImmunoPure streptavidin horseradish peroxidase conjugate (SA-HRP, Pierce; Rockford, IL) and peroxidase-specific fast-reacting substrate, turbo-3,3',5,5'-tetramethylbenzidine dihydrochloride (turbo-TMB, Pierce) (9, 21). The cells were washed three times and incubated with 100 µl of SA-HRP (1:500 dilution) in the gelatin buffer at room temperature for 1 h. After being washed three times in HEPES-Tyrode gelatin buffer, 100 µl of substrate turbo-TMB was added and developed for 5 min at room temperature. The reaction was stopped by adding 1 M phosphoric acid (100 µl), and the absorbance of the reaction mixture at OD₄₅₀ was measured using a Microplate autoreader EL 311 (Bio-Tek Instrument; Winooski, VT).

Prekallikrein activation on HUVEC and other cells. Investigations determined whether PK became activated to kallikrein when bound to HK on HUVEC as previously reported (16, 19, 26). Confluent HUVEC in fibronectin-coated 96-well microtiter plates were washed with HEPES-Tyrode gelatin buffer containing 2 mM Ca²⁺ and 1 mM Mg²⁺ followed by incubation with 20 nM HK or BK-free HK at 37°C for 60 min in the same buffer. After removal of unbound HK, the cells were incubated with 20 nM PK alone or in the presence of antipain (100 µM) or lisinopril (100 µM) in the same buffer containing 8 µM Zn²⁺ for 1 h at 37°C. Previous studies showed that 4-10 µM Zn²⁺ was optimal for PK activation (26). The addition of Mg²⁺ and Ca²⁺ was to maintain the integrity of the cultured cells through the incubation steps as previously performed (13, 19). After incubation, the supernatant was removed and frozen at 70°C for measurement of BK. The HUVEC were then washed followed by adding 100 µl of 0.4 mM HD-Phe-Pro-Arg-paranitroanilide (S2302) (DiaPharma; Franklin, OH) and incubated for 1 h at 37°C. The activity of kallikrein generated was measured at an optical density of 405 nm after 60 min hydrolysis of the substrate. The determination of kallikrein formation was also performed identically on UASMC, HMVEC, and BPAEC. When the activity of kallikrein generated was measured on UASMC and HMVEC, 8 µM Zn²⁺ was included in the buffer for each step. In certain experiments, PK activation of HUVEC was determined in HEPES-Tyrode buffer containing 0.35% BSA in the presence of 50 or 250 µM Zn²⁺. In other experiments, both the cell pellet and supernatant of HUVEC pretreated with 20 nM HK were collected from 10 to 60 min after they were incubated with 20 nM PK. Immunoblot studies were performed on the cell lysate and incubation buffer for HK by using a goat anti-human polyclonal antibody to HK as previously reported (29). Similarly, immunoblot studies for PK were also performed on the cell incubation buffer by using a polyclonal antibody to prekallikrein, which was obtained from Affinity Biologicals (Hamilton, Ontario, Canada). All immunoblots were detected by chemiluminescence (Amersham; Arlington Heights, IL). These immunoblots were scanned by densitometer scanning of the blot using a transmittance/reflectance scanner (model GS 300, Hoefer Scientific Instruments; San Francisco, CA) in the transmittance mode to determine the degree of cleavage and the relative amount of these proteins at each time point.

Bradykinin determination. Investigations were performed to determine whether BK was liberated from HK when kallikrein is formed on endothelial cells. In these experiments, HUVEC were incubated with HK and PK as described above, and the supernatants of these reactions after 1 h of incubation were collected and either frozen at 70°C or immediately deproteinized with trichlororacetic acid. BK in the samples was determined using a commercial kit (Markit BK, Dainippon Pharmaceutical; Osaka, Japan), performed according to the manufacturer instructions. In other experiments, lysates of BPAEC and human lung fibroblasts were examined for the presence of the BK B₂ receptor by immunoblot. These immunoblot studies were performed by the procedure of Muller-Esterl (20).

Measurement of NO formation. Confluent BPAEC grown in 96-well microtiter plates were washed with HEPES-Tyrode gelatin buffer, pH 7.4, containing 2 mM Ca²⁺ and 1 mM Mg²⁺. Confluent BPAEC were preincubated with 0.5 mM lisinopril for 30 min at 37°C. The cells were then incubated at 37°C for 30 min in the same buffer containing 100 nM HK or 100 nM PK alone or combined HK and PK in the presence of 0.5 mM lisinopril and 8 µM Zn²⁺ in the absence or presence of 100 µM antipain, 5 µM B₂ BK receptor antagonist, 5 µM B₁/B₂ BK receptor antagonist, 0.2 or 0.5 mM L-NAME, or 10 or 20 µM L-NMMA. Afterward, the supernatant was removed and frozen at 70°C for later measurement of NO formation indirectly by determining nitrite concentration (12). In these experiments, the supernatants from six wells, 600 µl total volume, were pooled, lyophilized, and resuspended in 200 µl of buffer. At the time of the assay, 70 µl of the sample in duplicate were incubated with Greiss reagent (Sigma), which consisted of equal volumes of 1% sulfanilamide in 2% phosphoric acid and 0.1% N-(1-naphthyl)ethylenediamine for 15 min at room temperature (12). The samples were then read in a spectrophotometer at OD₅₄₀ nm (12). The level of nitrite formed was compared with a standard curve that was prepared daily with 0.1 to 2.0 µM NaNO₂.

Atomic absorption spectroscopy. Atomic absorption spectroscopy was performed with a Perkin-Elmer 5100 atomic absorption spectrophotometer (Perkin-Elmer; Norwalk, CT) equipped with an HGA-600 graphite furnace with an L'vov platform and an AS-60 autosampler. A stock solution of zinc oxide in 5% nitric acid (Fisher Scientific; Fair Lawn, NJ) containing zinc at 1 mg/ml was used to prepare intermediate and working solutions. Water used for sample dilution and preparing of standards was deionized, millipore filtered, and did not contain any detectable zinc. The atomic absorption spectroscopy of zinc was monitored at 213.9 nm using 0.7-nm slits. The procedure was performed on HEPES-Tyrode gelatin buffer that had been concentrated 5 to 20 times or HEPES-Tyrode BSA buffer without added Zn²⁺.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Influence of Zn²+ and buffer proteins on HK binding to HUVEC. The role of Zn²⁺ on HK binding to HUVEC was examined. Initial studies determined the optimal Zn²⁺ concentration for HK binding to HUVEC in HEPES-Tyrode gelatin buffer. Atomic absorption spectroscopy showed that HEPES-Tyrode gelatin buffer had a Zn²⁺ concentration of 0.3 ± 0.1 µM. Optimal biotin-HK binding to HUVEC occurred with no added Zn²⁺ to HEPES-Tyrode gelatin buffer (Fig. 1). As the Zn²⁺ concentration increased from 1 to 4 µM, the total level of specific binding decreased and from 8 to 32 µM it remained constant (Fig. 1). These data indicated that HEPES-Tyrode gelatin buffer without added Zn²⁺ contained sufficient Zn²⁺ for maximal HK binding at 14 pmol/10⁶ HUVEC.

View larger version (18K): [in this window] [in a new window]   Fig. 1.   The Zn2+ requirement for biotin-high-molecular-weight kinogen (HK) binding to human umbilical vein endothelial cells (HUVEC). Zn2+ requirements for HK binding to HUVEC in buffer containing gelatin. HUVEC were incubated at 37°C for 20 min with 10 nM biotin-HK in HEPES-Tyrode buffer containing 0.35% BSA or 0.1% gelatin in the absence or presence of 1-64 µM added Zn2+. Specific binding is shown and is the level of binding when nonspecific binding, binding in the presence of 50-fold molar excess unlabeled HK, is subtracted from total binding. Data are means ± SE of three independent experiments.

Reversibility and saturability of HK binding to HUVEC in gelatin buffer. Because biotin-HK binding to HUVEC in HEPES-Tyrode gelatin buffer had a different requirement for added Zn²⁺ than in the presence of BSA (2, 10, 23, 26, 28, 31, 34), investigations next determined the characteristics of HK binding to HUVEC in gelatin buffer. The addition of 50-fold molar excess of unlabeled HK at 5 to 60 min after the initiation of binding resulted in 100 to 60% reversibility, respectively (Fig. 2). At 20 min biotin-HK binding to HUVEC was 90% reversible.

View larger version (11K): [in this window] [in a new window]   Fig. 2.   The reversibility of biotin-HK binding to HUVEC in gelatin buffer. Biotin-HK (10 nM) was incubated at 37°C with HUVEC in HEPES-Tyrode gelatin buffer followed by the addition of a 50-fold molar excess of unlabeled HK at the times indicated after the start of the incubation (5-60 min). The percent reversibility of biotin-HK binding at each time of addition of the unlabeled HK was determined. The percent reversibility was calculated by the ratio of the amount of biotin-HK bound in the presence of added unlabeled HK at each time point (5-60 min) over the level of binding of biotin HK that had been incubated simultaneously with 50-fold molar excess unlabeled HK. Data are means ± SE of three experiments.

View larger version (16K): [in this window] [in a new window]   Fig. 3.   Concentration-dependent binding of biotin-HK to HUVEC in gelatin buffer. A: increasing concentrations (1-60 nM) of biotin-HK were incubated at 37°C for 20 min with HUVEC in HEPES-Tyrode gelatin buffer not containing any added divalent cations. Specific binding was determined by subtraction of nonspecific binding from total binding. The mean ± SE pmoles bound biotin-HK per 106 cells from three experiments is shown. B: a Scatchard plot of the specific binding isotherm from A. Specific binding was plotted as Bound/Free vs. Bound using the data presented in A.

Activation of PK on cells. Investigations next proceeded to determine whether PK activation, which was characterized to occur on HUVEC (16, 19, 26), occurred on other cells when PK assembled on HK bound to these cells. HMVEC, UASMC, and BPAEC along with HUVEC were examined for their ability to support PK activation (Fig. 4). PK binding to HK on HUVEC did not require the addition of any Zn²⁺ (data not shown). However, when a complex of PK on HK on each of these cells was assembled in HEPES-Tyrode gelatin buffer containing 8 µM Zn²⁺ (26), PK was activated to kallikrein as determined by hydrolysis of a kallikrein chromogenic substrate (Fig. 4). The level of PK activation when bound to HK on plastic alone was <15% of that seen when bound to HUVEC (data not shown). Furthermore, the addition of 8 µM Zn²⁺ to optimize PK activation did not reduce the level of prebound HK (data not shown). The mechanism for PK activation on these various cells was similar to that seen on HUVEC because antipain inhibited PK activation as previously reported (19, 26). These data indicated that the PK activation mechanism that was originally described on HUVEC (16, 19) was present on a wide range of cells from both the intravascular and extravascular compartments.

View larger version (24K): [in this window] [in a new window]   Fig. 4.   Activation of prekallikrein (PK) bound to HK on different cells. HK (20 nM) in HEPES-Tyrode gelatin buffer containing 2 mM Ca2+ and 1 mM Mg2+ was incubated at 37°C for 60 min with human HUVEC, human microvascular endothelial cells (HMVEC), or bovine pulmonary artery endothelial cells (BPAEC). HK was bound to HMVEC and umbilical artery smooth muscle cells (UASMC) in the presence of 8 µM Zn2+. The cells then were incubated for another 60 min with 20 nM of PK in HEPES-Tyrode gelatin buffer containing 8 µM zinc, 2 mM calcium, and 1 mM magnesium. At the completion of the incubation, the cells were washed, 0.4 mM S2302 was added, and hydrolysis of the substrate was measured for 1 h. Data are means ± SE of three experiments. When antipain was added it was added with the PK in a concentration of 100 µM.

View larger version (66K): [in this window] [in a new window]   Fig. 5.   The change in structure in HUVEC-bound HK and soluble PK when PK is present in the incubation buffer over a monolayer of endothelial cells. Confluent monolayers of HUVEC were incubated with 20 nM HK in HEPES-Tyrode buffer containing 0.1% gelatin for 1 h at 37°C. After incubation, the cells were washed and incubated with 20 nM PK in the same buffer for 1 h. During the 1-h incubation period, the incubation buffer containing the PK was removed after 5-60 min, mixed with sample buffer for SDS-PAGE, and stored at 70°C for immunoblot studies. Similarly, the cells were washed three times and then solubilized in SDS-PAGE sample buffer. A: cleavage of HUVEC-bound HK when PK was present in the incubation buffer. The cell lysates were electrophoresed on a reduced 8% SDS-PAGE followed by electroblot onto nitrocellulose and immunoblot with an antibody to HK. The protein on the nitrocellulose was detected with a second antibody conjugated with horseradish peroxidase followed by chemiluminesence. The lane "HK" represents bound HK incubating with the cells for 1 h with buffer alone (i.e., no PK in the buffer). B: cleavage of PK in the incubation buffer over the monolayer of HUVEC. Aliquots of the incubation buffer over a monolayer of HUVEC containing PK were electrophoresed on a reduced 8% SDS-PAGE followed by electroblot onto nitrocellulose and immunoblot with an antibody to PK. The protein on the nitrocellulose was detected with a second antibody conjugated with horseradish peroxidase followed by chemiluminesence. The lane "PK" represents PK alone incubated for 1 h in buffer over cells that were not pretreated with HK. In both panels, the numbers at the bottom of the gels represent the incubation time in minutes. The numbers to the left of the panels represent molecular mass standards in kilodaltons.

View larger version (23K): [in this window] [in a new window]   Fig. 6.   Prekallikrein activation and bradykinin (BK) liberation on HUVEC. A: activation of PK on HUVEC. HUVEC were incubated with 20 nM HK alone, BK-free HK, or buffer in HEPES-Tyrode gelatin buffer containing 2 mM Ca2+ and 1 mM Mg2+ at 37°C for 60 min. Afterward, 20 nM PK without or with antipain (100 µM), lisinopril (100 µM) [angiotensin-converting enzyme inhibitor (ACEI)], or buffer was added and incubated with HUVEC at 37°C for 60 min in the same buffer containing 8µM zinc, 2 mM calcium, and 1 mM magnesium. At the end of the incubation, the wells were washed and 0.4 mM S2302 were added alone or in the presence of antipain (100 µM) in the same buffer and hydrolysis of the substrate was measured for 1 h. Data are means ± SE of five determinations. B: liberation of BK after PK activation. After PK activation on HUVEC as described above, the buffer from each of the wells was collected and deproteinized by treatment with trichloroacetic acid (see METHODS). A competitive ELISA for BK measured the peptide in deproteinized samples by the procedure described in the METHODS. The data from three experiments (means ± SE) are presented.

Bradykinin liberation on HUVEC. When PK assembled on HK on HUVEC became activated, BK was liberated (5.0 pmol/10⁶ HUVEC) (Fig. 6B). These data indicated that in the presence of PK most of the cell-bound HK became cleaved to liberate its BK. Furthermore, the measured BK was derived mostly from the HK that served as the receptor for PK because in the absence of added HK or PK, little BK was liberated. The extent of BK liberation from the assembly of PK on HK was abolished by the presence of antipain (Fig. 6B). No BK was liberated from BK-free HK (Fig. 6B). However, the extent of BK liberated (5.0 pmol/10⁶ HUVEC) was somewhat lower than the amount of HK bound (14 pmol/10⁶ HUVEC). This finding suggested that some of the liberated BK might have been metabolized by angiotensin-converting enzyme associated with the HUVEC. When the BK was released into the incubation buffer in the presence of lisinopril, there was 106% increase in the amount of the peptide (10.3 pmol/10⁶ HUVEC) measured in the sample (Fig. 6B).

NO formation in endothelial cells. Studies next determined whether BK liberated from HK by kallikrein formed on BPAEC membranes resulted in NO formation. BPAEC were chosen for these experiments because these cells have a constitutively higher expression of BK B₂ receptors and NO synthetase. On immunoblot using a rabbit antibody to the BK B₂ receptor, BPAEC along with normal human lung fibroblasts contained the 69-kDa epitope of the BK B₂ receptor (data not shown). When HK and PK complexes were assembled on BPAEC, NO was formed as measured by nitrite formation (0.6 ± 0.1 µM/2.4 × 10⁵ cells) (12) (Fig. 7). The presence of antipain, which inhibited PK activation, blocked NO formation. These data indicated that kallikrein formation was necessary for increased NO formation under these experimental conditions. The specificity that the NO formed was due to BK-mediated events was determined by including a B₂ or a BK B₁/B₂ receptor antagonist in the reaction. In the presence of each of these BK receptor antagonists, NO formation was inhibited indicating that NO formation is mediated by the BK B₂ receptor. Finally, two different NO synthetase inhibitors, L-NAME and L-NMMA, also reduced BK-mediated NO formation (Fig. 7). Thus HK and PK assembly and activation on endothelial cells directly influenced endothelial cell NO formation through a BK-mediated mechanism.

View larger version (29K): [in this window] [in a new window]   Fig. 7.   Role of BK in nitric oxide (NO) formation on BPAEC. BPAEC were preincubated with lisinopril (0.5 mM) follow by incubation with 100 nM HK and PK in HEPES-Tyrode gelatin buffer containing 2 mM Ca2+, 1 mM Mg2+, 8 µM Zn2+, and 0.5 mM lisinopril at 37°C for 30 min in the absence or presence of 100 µM antipain, 5 µM B2 BK receptor antagonist (B2 INHIBITOR), 5 µM B1/B2 BK receptor antagonist (B1/B2 INHIBITOR), 0.2 to 0.5 mM NG-nitro-L-arginine methyl ester (L-NAME), or 10 to 20 µM NG-monomethyl-L-arginine (L-NMMA). At the end of the incubation, the wells were washed and the amount of NO formation was measured as indicated in the METHODS. Data are means ± SE of five determinations.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The present investigations demonstrate two findings: 1) HK binding to endothelial cells in culture requires <1 µM Zn²⁺, and 2) when PK gets activated when bound to HK on HUVEC, BK is liberated from the bound HK to stimulate NO formation. Previous investigations determined that 25-50 µM added Zn²⁺ is required for HK binding to platelets and endothelial cells (7, 8, 28, 31). The requirement for this amount of added Zn²⁺ is due to the fact that the majority of the zinc becomes albumin bound and only a small fraction is free to support HK binding (26). With the use of a buffer protein that does not bind Zn²⁺ (25, 26), the concentration of Zn²⁺ necessary to support binding is 0.3 ± 0.1 µM. Recent studies using improved techniques to measure Zn²⁺ concentration indicate that the unbound Zn²⁺ concentration in blood is 0.15-0.5 µM in a total Zn²⁺ concentration of 10-25 µM (1, 6, 33). The present investigations indicate that this level of unbound Zn²⁺ is sufficient to support HK binding to HUVEC. Alternatively, in buffers containing BSA, 25 to 50 µM Zn²⁺ need to be added to elevate the free Zn²⁺ concentration to levels that will support HK binding (7, 8, 28, 31). Because PK binding to HK bound on HUVEC does not require the addition of Zn²⁺, the HK-PK complex must be constitutively assembled on the external membrane of endothelial cells. However, PK activation on HUVEC, both in buffers containing BSA and in plasma, requires supplemental Zn²⁺ to elevate the free Zn²⁺ level to a level (4-10 µM) that will support activation. This level of free Zn²⁺ is substantially higher than that needed to support binding as indicated in the present studies. Furthermore, the extent of PK activation was limited by the structure of HK, the PK receptor on HUVEC. When BK-free HK was substituted for HK, only 64% of PK is activated to kallikrein. Because BK-free HK has a different conformation than intact HK, PK must need to be oriented in a particular configuration on HK in order for it to be optimally activated (32). Similar findings have been noted when factor XI bound to BK-free HK on HUVEC is activated to factor XIa by factor XIIa (unpublished data).

Because the buffer protein and Zn²⁺ concentration were changed in the present study, investigations were performed to determine whether the kinetics of HK binding were altered. With the use of HEPES-Tyrode gelatin buffer, HK reversibly binds to endothelial cells with a similar affinity and number of binding sites as previously reported (10). It is still unclear what is the exact function of Zn²⁺ in kininogen binding. Because Zn²⁺ is required for low molecular weight kininogen binding to cells, it is not sufficient to justify its requirement for binding of HK through its domain 5, a zinc-binding region, because low-molecular-weight kininogen lacks this domain (3, 5). We have shown that both HK and low-molecular-weight kininogen require Zn²⁺ to support their binding to cytokeratin 1, a putative kininogen receptor (11). Similarly, Zn²⁺ is necessary for HK to bind to gC1qR and urokinase plasminogen activator receptor, two additional HK binding proteins, putative receptors, on endothelial cells (3, 15). It is possible that Zn²⁺ contributes to the presentation of the kininogen binding site similar to its role in the expression of an active human growth hormone receptor (4).

The second finding of these investigations is that when PK is activated on endothelial cells, some of the kallikrein diffuses off the cells and BK is liberated from HK. Over a 1-h observation period, the kallikrein detected within 10 min in the PK supernatant must have diffused off the cell membrane because it followed the cleavage of cell-bound HK and the detection of cleaved HK in the incubation buffer above the monolayer of cells. Furthermore, the kallikrein in the incubation media was not -kallikrein, a product of kallikrein autodigestion, because no cleaved kallikrein heavy chain was detected on immunoblot. Previous studies (19) had shown that when PK binds to bound HK on HUVEC, the formed kallikrein cleaves its receptor, HK, in a pattern consistent with that associated with liberated BK. Furthermore, Nishikawa et al. (21) showed that kallikrein incubated with endothelial cell-bound HK will liberate BK. The present investigations advance those investigations by showing that simultaneously with the generation of kallikrein activity on HUVEC, BK is liberated. The source of this BK has to be exogenously added HK that serves as the PK receptor, because in the absence of exogenous HK, there is little measurable BK. Also, the extent of measured BK was markedly influence by HUVEC-associated angiotensin-converting enzyme. In the presence of lisinopril, there was 106% increase in the amount of BK measured. Indirectly, these data also suggest that angiotensin-converting enzyme can only account for 50% of BK metabolism. Other peptidases as well as the BK receptor itself could be contributory. There is a correlation between the amount of HK bound to HUVEC (14 pmol/10⁶ HUVEC) and the amount of BK liberated from bound HK measured in the presence of lisinopril (10.3 pmol/10⁶ HUVEC). The liberated BK then activates its B₂ receptor to stimulate NO formation. Thus prekallikrein activation on endothelial cells modulates the activity of these cells through BK liberation that specifically induces NO production.

The potential importance of the above BK liberating pathway is underscored by the finding that the assembly mechanism for PK binding and activation, and thus BK liberation, is present on different vascular cells. The ability to activate PK to kallikrein is equal on HUVEC, HMVEC, UASMC, and BPAEC. Differences seen in the extent of formed kallikrein among these cell types may be the result of cell number, the number of HK binding sites on the cells, the concentration of the PK activator present, or variable Zn²⁺ requirements. Regardless, all these cells express the same mechanism for PK activation and BK liberation. Regulation of activation of PK complexed to HK on cell surfaces influences BK liberation and its associated physiological activities. Better understanding of the regulatory processes of PK activation on cells should increase our understanding of the physiological activities of the plasma kallikrein/kinin system.