Tissue Kallikrein-binding Protein Is a Serpin I. PURIFICATION, CHARACTERIZATION, AND DISTRIBUTION IN NORMOTENSIVE AND SPONTANEOUSLY HYPERTENSIVE RATS*

Kallikrein-binding protein was purified to apparent homogeneity from rat serum by Affi-Gel Blue, DEAE-Sepharose CL-GB, Sephacryl S-200 chromatography, and preparative gel electrophoresis or high perform- ance liquid chromatography. The purified protein mi- grates as a single band of 60 kDa in a sodium dodecyl sulfate-polyacrylamide gel under reducing conditions. It is an acidic protein with isoelectric points ranging from 4.2 to 4.6. The amino terminus of the binding protein is an Asp residue as determined by sequence analysis. It forms a 92-kDa sodium dodecyl sulfate- stable complex with kallikrein with a t,,+ of 18 min. and showed distri- of the kallikrein-binding protein in and various tissues

Kallikrein-binding protein was purified to apparent homogeneity from rat serum by Affi-Gel Blue, DEAE-Sepharose CL-GB, Sephacryl S-200 chromatography, and preparative gel electrophoresis or high performance liquid chromatography.
The purified protein migrates as a single band of 60 kDa in a sodium dodecyl sulfate-polyacrylamide gel under reducing conditions. It is an acidic protein with isoelectric points ranging from 4.2 to 4.6. The amino terminus of the binding protein is an Asp residue as determined by sequence analysis.
It forms a 92-kDa sodium dodecyl sulfatestable complex with kallikrein with a t,,+ of 18 min. Western blot and radioimmunoassay showed a distribution of the kallikrein-binding protein in serum, urine, and various tissues with a 5-lo-fold lower amount in spontaneously hypertensive rats (SHR) than in Wistar-Kyoto rats (WKY). A full length cDNA clone encoding the kallikrein-binding protein was isolated from a rat liver cDNA library by immunoscreening and the translated amino acid sequence matches the amino-terminal 29-amino acid sequence of the binding protein.
The cDNA sequence shares 68.8% identity with human cY,-antichymotrypsin and is identical to that of a rat hepatic protein.
Dot blot analysis shows that kallikrein-binding protein is expressed at high levels in the liver and at low levels in the lung, salivary gland, and kidney. Its mRNA level in the liver decreases by a-fold after acute phase inflammation and is higher in male than in female rats. Genomic Southern blot analyses reveal restriction fragment length polymorphisms between SHR and WKY rats in the binding protein locus. The results indicate that rat kallikrein-binding protein belongs to the serpin superfamily and its level is significantly reduced in the spontaneously hypertensive rats.
Kallikrein-kinin system components are involved in many important pathophysiological processes such as hypertension, diabetes mellitus, allergy, and inflammation (Pisano, 1979;Schacter, 1980;MacDonald et al., 1988 (EC 3.4.21.35), one of the system's primary components, cleaves kininogens to produce vasoactive kinin peptides, which mediate a broad spectrum of biological effects including vasodilation, inflammation, pain, and smooth muscle contraction and relaxation.
Although the biochemistry of the kallikrein-kinin system and the regulation of kallikrein and kininogen gene expression have been extensively studied (Miiller-Ester1 et al., 1986;MacDonald et al., 1988), information regarding kallikrein inhibitors or other factors which modulate kallikrein functions at the post-translational level is very limited. Geiger et al. (1980) showed that human urinary kallikrein was progressively inhibited by normal human serum but not by homozygous ZZ a,-antitrypsin deficiency serum, and in addition, that human cY1-antitrypsin and human urinary kallikrein form a llO-kDa SDS-stable complex. It was also shown that human urinary kallikrein was inhibited progressively by human ocl-antitrypsin (Geiger et al., 1981;Hirano et al., 1984). We have recently identified circulating autoantibodies to tissue kallikrein (Chao et al., 1988a) and a kallikreinbinding protein in the systematic circulation of humans and rats, and in the secreted media of several transformed cell cultures (Chao et al., 1986a;Chao and Chao, 1988). Kallikreinbinding protein and kallikrein form a complex which is stable to SDS and heat treatment, suggesting a covalent binding. Our previous studies have shown that kallikrein-binding protein is distinct from other well characterized serine proteinase inhibitors (Chao et al., 1986a). Kallikrein-binding protein and al-antitrypsin interact differently with human tissue kallikrein; in addition, neither polyclonal nor monoclonal antibodies to human al-antitrypsin cross-react with kallikrein-binding protein . These findings indicate that there are two different potential tissue kallikrein modulators, kallikrein-binding protein and circulating autoantibodies, in addition to cul-antitrypsin in human serum. Previous studies have shown that kallikrein-binding protein only binds to active kallikrein but not to inactive (latent) kallikrein or to active-site blocked kallikrein (Chao et al., 1986a;Chao and Chao, 1988). Endogenous complexes of kallikrein and the binding protein have been identified in serum, urine, and kidney of normotensive rats (Chao et al., 1986b) but not from the spontaneously hypertensive rats.' These together with our recent studies showing a major difference in the amount of kallikrein-binding protein of the sponta- neously hypertensive rats (SHR) uersus normotensive control Wistar-Kyoto (WKY) rats (Chao and Chao, 1988) suggest that the kallikrein-binding protein may play an important physiological role in the clearance and catabolism of kallikrein. The present studies report the purification, characterization, and distribution of rat tissue kallikrein-binding protein and cloning of a cDNA encoding this binding protein.
Furthermore, the results suggest that the deficiency of kallikrein-binding protein found in spontaneously hypertensive rats may have a genetic basis.  (Chao and Margolius, 1979). Purified proteins were labeled with iz51 according to the lactoperoxidase method (Shimamoto et al., 1980 Gel Electrophoresis-A 7.5% non-SDS uolvacrvlamide slab gel was used for preparative electrophoresis as described previously (Xiong et al., 1990  . Briefly, the nitrocellulose membranes were blocked with BLOTTO (5% (w/v) nonfat dry milk in 0.01 M sodium phosphate, pH 7.4, 0.14 M NaCl, 1 PM parauhenvlmethvlsulfonvl fluoride, 1 ma/liter thimerosal, 200 mg/liter NaNi, and 0.01% antifoam A) ' (Johnson et al., 1984) for 1 h at30 "C and then incubated with rabbit anti-rat kallikrein-binding protein (RKBP) (1:200 in BLOTTO) or rabbit anti-cul-antitrypsin antiserum (I:500 in BLOTTO). After a 3 h incubation at 30 "C with gentle shaking, the nitrocellulose membranes were washed three times with BLOTTO and then incubated with iz51-RKBP or iz51-ai-antitrypsin (250,000 cpm/ml). The nitrocellulose membranes were then washed three times with BLOTTO and once with phosphate-buffered saline (0.01 M sodium phosphate, pH 7.4, 0.14 M NaCl), air-dried, and exposed to Kodak X-Omat film.

Tissue Extract
Preparation-Inflammation was induced in the rat by subcutaneous injection of 0.5 ml of turpentine per 100 g body weight at two sites in the dorsolumbar region (Chao et al., 1988b). The rats were fasted for 48 h following injection and then anesthetized with pentobarbital (50 mg/kg). The rats were perfused via cardiac puncture with normal saline until tissue appeared blood-free. Tissues were removed, minced, and homogenized in phosphate-buffered saline, pH 7.3, at 4 "C. Homogenates were centrifuged at 600 x g for 20 min, and the supernatant was treated with deoxycholate (0.5%, w/v) for 30 min at room temperature.
The supernatants were collected after centrifugation at 10,000 X g for 30 min, and protein concentrations were determined by the method of Lowry et al. (1951), using bovine serum albumin as the standard. Development of a Radioimmunoassay for Kallikrein-binding Protein-A specific radioimmunoassay for RKBP was developed according to the previously described procedures (Chao et al., 1989a). In the antibody titration curve, RKBP antisera dilutions in the assay buffer ranged from l:l,OOO to 1:640,000. One hundred ~1 of "'1-RKBP (10,000 cpm/lOO ~1) and 100 ~1 of antibody in assay buffer was added to 200 ~1 of assay buffer, bringing the final volume to 400 ~1. The assay mixtures were incubated at room temperature for 24 h. Antibody-bound RKBP was generated from the free protein through centrifugation in an optimum combination of 200 ~1 of polyethylene glycol(25%) and 100 ~1 of bovine y-globulin (1%). A l:lO,OOO antisera dilution was chosen for radioimmunoassay and the standard curve of the binding protein ranged from 640 pg to 80 ng. Immunological Screening of the Rut Liver cDNA Library-Immunoscreening procedures for the isolation of cDNAs encoding kallikrein-bindingprotein from a rat liver hgtll cDNA library were similar to those described by Chao et al. (1989b). Rabbit anti-RKBP antiserum (1:500 in BLOTTO) was used for the screening. The filters were then washed and incubated with lz51-RKBP (-250,000 cpm/ml). Purified RKBP at concentrations of 1.0, 0.1, and 0.02 pg was spotted on a piece of nitrocellulose and processed as above as positive controls. Nucleic Acid Sequencing-The positive clones were subcloned into the Ml3 mp19 sequencing vector according to the method described by Messing (1983). Nucleic acid sequencing was performed using the dideoxy chain termination method (Sanger et al., 1977).

RNA Preparation
and Dot Blot Hybridization-Sprague-Dawley rats (Charles River Laboratories) weighing 250 g were anesthetized with pentabarbital(50 mg/kg body weight) and perfused with normal saline via cardiac puncture.
The organs of interest were removed and RNAs were extracted as described ureviouslv  washing, the blots were exposed at -70 "C with intensifying screens to Kodak X-Omat film.

Purification of Kallikrein-binding
Protein-Fractions containing kallikrein-binding activity were monitored by the formation of a 92kDa complex identified on autoradiograms following SDS-PAGE. Pooled rat sera were first fractionated with 45-80% ammonium sulfate saturation. The precipitate was dissolved in 0.02 M sodium phosphate, pH 7.0, dialyzed against the same buffer, and then passed through an Aff-Gel Blue column equilibrated with the same buffer. Kallikreinbinding protein was in the flow-through fractions. The fractions containing the active material were combined and dialyzed against 0.05 M NaCl, 0.05 M Tris-HCl buffer, pH 8.8, and then passed through a DEAE-Sepharose CL-6B column equilibrated with the same buffer. The column was eluted with a 0.05-0.2 M NaCl gradient in 0.05 M Tris-HCl, pH 8.8. The binding protein was eluted from the DEAE-Sepharose column at 0.1 M NaCl, 0.05 M Tris-HCl buffer, pH 8.8. The fractions with binding activities were combined, precipitated with 80% ammonium sulfate, dissolved in 3-5 ml of 0.05 M ammonium formate, pH 8.0, and then passed through a Sephacryl S-200 column (2.5 x 90 cm) equilibrated with 0.05 M ammonium formate, pH 8.0. The binding protein was eluted from the column in the second peak between bovine serum albumin (68 kDa) and ovalbumin (43 kDa). Kallikrein-binding protein was subsequently electrophoresed on a 7.5% non-SDS polyacrylamide gel and eluted from the gel slices using an ISCO concentrator. Alternatively, the fractions with kallikrein-binding activity eluted from the Sephacryl S-200 column were separated by HPLC using a reverse phase Cq column. The kallikrein-binding protein was eluted at 45% acetonitrile at 44 min. Table I    panel, lane 2) and the corresponding cYr-antitrypsin in rat serum (right panel, lane 1) but does not bind to the purified kallikrein-binding protein (right panel, lane 3). The results indicated that kallikrein-binding protein is immunologically distinct from cri-antitrypsin.
Analytical isoelectric focusing revealed multiple bands of RKBP, with isoelectric points (pl) ranging from 4.2 to 4.6 (lane 2), rat cY1-antitrypsin with pl of 4.5-5.0 (lane 3), and human cY1-antitrypsin with pl of 5.0-5.5 (lane I) (Fig. 2). Sequential Edman degradation of the purified kallikreinbinding protein yielded a 29-amino acid sequence with Asp as the amino terminus. Tissue kallikrein (38 kDa) and the purified binding protein (60 kDa) forms an equimolar 92-kDa complex. Formation of the SDS-stable complex between kallikrein and the binding protein appears in 30 s and reaches half-maximal binding at 18 min (data not shown). When the complex formation was analyzed under non-SDS PAGE, the rate of binding is very rapid and reaches a maximal binding within 30 s (data not shown).

Tissue Distribution of Kallikrein-binding Protein in Normotensive and Spontaneously
Hypertensive Rats-Tissue distribution of RKBP was analyzed by both Western blot analysis and a specific radioimmunoassay. Fig. 3 shows a single protein band of 60 kDa corresponding to purified kallikreinbinding protein can be detected in serum and tissue extracts of heart, lung, kidney, salivary gland, uterus, testis, and pituitary by Western blot analysis. However, kallikrein-binding protein in the urine has a smaller molecular mass of 50 kDa, suggesting a possible cleavage of this protein before excretion. The binding protein was barely detectable in the pancreas, prostate, adrenal glands, and liver. Semiquantitation of immunoblot assays can detect rat cui-antitrypsin levels in serum  or the binding protein in serum or tissue extracts in a dose-dependent manner with a minimum detection of 20 ng of protein. Kallikrein-binding protein levels in the serum of spontaneously hypertensive rats (SHR) are significantly lower than those of the control normotensive Wistar-Kyoto (WKY) rats as shown by Western blot analysis (Fig. 4, left panel). Contrarily, there appears to be no difference of kallikrein-binding protein levels in the serum of Bio/ Breeder (BB(C)) control uersus Bio/Breeder spontaneously diabetic (BB(D)) rats (Fig. 4, left panel). When rat serum (1 ~1) was incubated with purified tissue kallikrein (1 pg), the amount of the 92-kDa kallikrein complex formation is significantly lower with the serum of SHR as compared to the control WKY rats or the BB diabetic and BB control rats (Fig. 4, right panel). The results clearly demonstrate a significant reduction of both kallikrein-binding protein and complex formation in the serum of SHR. The findings of a deficiency in RKBP levels in SHR using Western blotting were confirmed by quantitative determination based on a radioimmunoassay.
A highly specific radioimmunoassay was developed for measuring RKBP. Proteins were subjected to 7.5-15% linear gradient SDS-PAGE under reducing conditions and electrotransferred to nitrocellulose. Blots were developed with rabbit anti-RKBP antiserum followed by ""I-kallikrein-binding protein.
bovine y-globulin (1%) were found from exhaustive tests, yielding low background and high specific binding. The assays can detect kallikrein-binding protein levels ranging from 0.64 to 80 ng. Serial dilutions of sera or tissue extracts from normal and hypertensive rats showed complete parallelism with the kallikrein-binding protein standard (data not shown). Fig. 5 shows the tissue distribution of RKBP in SHR uersu.s WKY rats as determined by the radioimmunoassay. RKBP levels in the serum and all tissues of SHR are severalfold lower than those of the WKY rats.
Isolation and Characterization of cDNA Clones Encoding Kallikrein-binding Protein-By immunoscreening a rat liver cDNA library with the specific antiserum against the purified binding protein, a group of independent cDNA clones of various lengths were isolated and sequenced. Fig. 6 shows the amino-terminal 29-amino acid sequence of the purified kallikrein-binding protein matching with a region of the deduced amino acid sequence from the cDNA encoding kallikreinbinding protein with only two amino acid residues mismatched. The amino-terminal Asp residue of the purified RKBP can be aligned with residue 21 of the translated protein sequence from the RKBP cDNAs. The peptide bond between the Asp residue and the Cys residue amino-terminal to it, as shown by an arrow, is therefore considered to be the site of cleavage between the 20-amino acid signal peptide and the mature protein (Fig. 6). When the total cDNA sequence was searched against the GenBankTM nucleic acid sequence data base with the FASTN program by Lipman and Pearson (1985), it matches with the cDNA sequence of the rat Spi-2.3 (Yoon et al., 1987) with greater than 99% sequence identity, with the cDNA sequence of rat growth hormone-regulated protein (Le Cam et al., 1987), and with the cDNA sequence of rat thyroid hormone-regulated protein with 100% sequence identity (Tecce et al., 1986). The only mismatch found between the cDNA sequence of the kallikrein-binding protein and the rat Spi-2.3 is a single base replacement which does not alter the amino acid residue encoded, suggesting that the protein encoded by the Spi-2.3 cDNA, or by the rat growth hormone-regulated protein, rat thyroid hormone-regulated The calculated molecular weight of mature RKBP based on the translated amino acid sequence is 44,587, which is less than the molecular weight of 60,000 as determined by SDS-PAGE (Fig.  1). The difference in the molecular weights is ascribed to glycosylation of the mature protein at 5 potential residues, four of which are N-linked. These potential glycosylation sites were identified by two consensus sequences: Asn-X-Ser/Thr and Ser/Thr-X-X-Pro (Marshall, 1974). The rat kallikreinbinding protein contains 2 cysteine residues in its mature form which may be involved in a potential disulfide linkage. endothelial plasminogen activator inhibitor. The RKBP contains a similar reactive center sequence in its COOH-terminal region as the other serpin molecules. The sequence, starting from P1 to P,, residues, is Leu-Lys-Ser-Leu-Pro-Gln.
The cleavage site for target serine protease P1-P1, (boxed) or rat kallikrein-binding protein is the bond between Leu-Lys, different from those of other serpins.
Expression and Regulation of Rat Kallikrein-binding Protein-Tissue-specific and regulated expression of RKBP was identified by dot blot analysis using the RKBP cDNA probe, as shown in Fig. 8. Kallikrein-binding protein mRNA was found in the liver at high levels and also in the lung, salivary glands, and kidney but at low levels (Fig. 8A). Note that the amounts of RNA from liver loaded on the filter was IO-fold less than that from other tissues. The results indicate that the major site of synthesis of kallikrein-binding protein is in the liver with much less expression in other tissues. After acute phase inflammation, kallikrein-binding protein mRNA levels in the liver are reduced by one-half (Fig. 8B). The results indicate that in contrast to the acute-phase induced rat cY,-antitrypsin , the binding protein is a negative acute phase reactant. Similar to cYi-antitrypsin, a major sex difference in levels of kallikrein-binding protein mRNA is evident, with a severalfold higher level in male than in female rats (Fig. 8B).
Restriction Fragment Length Polymorphism Analysis--In genomic Southern blot analysis using the DNA from SHR and WKY rats, multiple restriction fragment length polymorphisms were detected. Using 11 different restriction endonucleases, we have identified restriction fragment length polymorphisms involving alterations in restriction fragment lengths with four restriction endonucleases: BclI, BglII, DruI, and NdeI as shown in Fig. 9. The results indicate potential mutation(s) associated with the kallikrein-binding protein gene or its flanking regions in SHR uersus WKY rats.

DISCUSSION
In the present report, we describe the purification, characterization, and cloning of a tissue kallikrein-binding protein from rat serum. Kallikrein-binding protein is an acidic protein with molecular mass of 60 kDa and isoelectric points ranging from 4.2 to 4.6. Different isoforms may be ascribed to varying carbohydrate content since the binding protein contains five potential glycosylation sites. The amino-terminal 29-amino acid sequences of the purified kallikrein-binding protein confirms the identity of the full length cDNA clones encoding kallikrein-binding protein which were isolated using antiserum against the binding protein. The cDNA encoding kallikrein-binding protein shares complete identity with the cDNA sequences of a rat hepatic protein, Spi-2.3 (Yoon et al., 1987), the rat GHRP (Le Cam et al., 1987), and the rat THRP (Tecce et al., 1986). Therefore, rat kallikrein-binding protein belongs to the serpin superfamily.
Previous studies by Geiger et al. (1981) have shown that human cui-antitrypsin binds to human urinary kallikrein and is a slow progressive inhibitor of human tissue kallikrein. We have confirmed the studies of Geiger et al. (1981), and in addition, we have also identified a novel human kallikreinbinding protein distinguishable from human cYi-antitrypsin  c&1990). Contrarily, we found that rat tissue kallikrein cannot form a SDS-stable complex with purified rat (piantitrypsin. It is known that human tissue kallikrein is able to cleave two peptide bonds between Met-Lys and Arg-Ala from human or bovine kininogens to produce lysyl-bradykinin, while rat tissue kallikrein lacks this ability but cleaves Arg-Arg and Arg-Ala bonds from rat kininogen to generate bradykinin (Kato et al., 1985). Since both human and rat CQantitrypsin have a Met-Ser bond as the reactive site, it is explainable why tissue kallikrein and cYi-antitrypsin of rat fail to complex. If cui-antitrypsin were indeed the sole regulator of kallikrein activity, in the case of rat, there would be a lack of such regulation, unless some other protein(s) function in substitution of cui-antitrypsin. The mechanisms by which kallikrein-binding protein acts is not understood at this moment. However, the generalized reaction mechanism for serpin-like inhibitors may apply to the interaction of kallikrein and the binding protein, which suggest a cleavage at the reactive center or the bait region of the inhibitor molecule (Carrel1 and Boswell, 1986). The purified binding protein (60 kDa) and tissue kallikrein (38 kDa) form a 1:l stoichiometric complex of 92 kDa which is resistant to SDS and heat treatment, suggesting a covalent linkage. The exact cleavage site at the reactive site of the binding protein by kallikrein has not been identified. Previous studies showed that active site blocked kallikrein or latent kallikrein cannot form complexes with the binding protein (Chao and Chao, 1988). The Pi-Pi, bond of the kallikrein-binding protein, Leu-Lys is not the favorite cleavage site for tissue kallikrein. Therefore, the next peptide bond, Pi,-P2,: Lys-Ser may serve as the cleavage site for tissue kallikrein since tissue kallikrein has a high affinity for the basic amino acid residues Arg and Lys. Such an alternative reactive site peptide bond has been observed recently in another serpin molecule, (YZantiplasmin, inactivating both plasmin and chymotrypsin using overlapping reactive sites (Potempa et al., 1988). Kallikrein-binding protein may interact with kallikrein by first