Molecular cloning and expression of serum calcium-decreasing factor (caldecrin).

We previously reported on the purification of a serum calcium-decreasing factor, referred to as caldecrin, from porcine pancreas, that is thought to be a serine protease (Tomomura, A., Fukushige, T., Noda, T., Noikura, T., and Saheki, T. (1992) FEBS Lett. 301, 277-281). In the present study, we purified caldecrin from rat pancreas and determined its primary structure by cDNA cloning. The predicted caldecrin protein is presumed to be synthesized as a preproenzyme of 268 amino acids with a signal peptide of 16 amino acids and an activation peptide of 13 amino acids, and is, with the exception of a central region, almost identical to the reported rat pancreatic elastase IV sequence. The caldecrin gene is selectively expressed in the pancreas, as judged by Northern blot analysis. After expression in BMT-10 cells, immunoreactive caldecrin was found in the culture supernatant, and it inhibited the parathyroid hormone-stimulated 45Ca release from cultured fetal long bones. Catalytic site mutants were synthesized in a baculovirus system, and recombinant mutants also decreased the serum calcium level of mice. These data implicate caldecrin, a protease closely related to elastase IV, in the regulation of blood calcium levels.

It is well known that blood calcium is regulated by several calcitropic hormones, e.g. calcitonin, parathyroid hormone (PTH), 1 and 1,25-dihydroxy vitamin D 3 , synthesized by the thyroid, parathyroid gland, and kidney, respectively. However, the involvement of other organs in the regulation of blood calcium level is not clearly understood. It has been reported that some gastric factors reduce blood calcium levels and that vagally mediated hypocalcemia is induced by hypothalamic stimulation (1)(2)(3). Another etiologic organ is the pancreas. Acute pancreatitis causes hypocalcemia, which suggests that pancreas tissue contains hypocalcemic factors (4,5). Among pancreatic hormones, glucagon and amylin are reported to have hypocalcemic effects (6 -8). Earlier we purified a serum calcium-decreasing factor (caldecrin) from porcine pancreas (9). Caldecrin is a chymotrypsin-like serine protease that lowers blood calcium levels in mice and inhibits in vitro PTH-stimulated bone resorption in long bone cultures.
Most serine proteases are synthesized as preproenzymes and secreted as proenzymes (zymogens) that require enzymatic cleavage of an amino-terminal peptide for conversion to active proteases. Recently, a proform of caldecrin (procaldecrin) was successfully purified from porcine pancreas (10). Procaldecrin does not possess serum calcium-decreasing activity (SCDA), but acquires SCDA, as well as protease activity, upon trypsin treatment. Incubation of the activated caldecrin with the serine protease inhibitor phenylmethylsulfonyl fluoride (PMSF) destroys its protease activity, but has no effect on its SCDA (9,10).
The mammalian exocrine pancreas contains a predominant superfamily of serine proteases and isoforms of these enzymes. Two chymotrypsinogens, cationic chymotrypsinogen A and anionic chymotrypsinogen B, have been isolated from bovine and rat pancreas (11)(12)(13). Anionic chymotrypsinogen C was isolated from porcine pancreas (14). The nomenclature of the related elastase family is more complicated. Two elastase isozymes, elastases I and II, were found in pigs, humans, and rats (15)(16)(17)(18). Rat and porcine elastase I cDNAs (19 -20) and rat, porcine, and human elastase II cDNAs (19,21,22) were successfully isolated. Anionic protease E was purified from the human, porcine, and bovine pancreas (23)(24)(25), and it hydrolyzed synthetic elastase-specific substrates but failed to cleave elastin. Human protease E (26) and its isozyme (27) were found to be identical to elastase IIIB and IIIA (28), respectively. Kang et al. reported the sequence of elastase IV cDNA amplified by the polymerase chain reaction (PCR) from rat pancreas RNA (29).
In this paper, we report on the purification of a rat caldecrin that displays SCDA and the isolation of a rat caldecrin cDNA clone. On the basis of the predicted protein structure, caldecrin can be classified as an elastase-type protease. Recombinant rat caldecrin inhibited PTH-stimulated calcium release from cultured bones. Mutant caldecrins, which lacked protease activity, also decreased the serum calcium level of mice.

EXPERIMENTAL PROCEDURES
In Vivo Assay: SCDA-The SCDA of rat caldecrin was measured as described previously (9). Male BALB/c mice weighing 20 -25 g were used, and the caldecrin fraction (200 l/20 g body weight) was injected into a tail vein of mice previously starved for 18 -20 h. Blood was taken for determination of serum calcium levels 4 h after the injection. The calcium concentration was measured by the o-cresolphthalein complexon method (30). For inactivation of the protease activity of purified caldecrin, rat caldecrin at a concentration of 1.25 mg/ml was treated for 30 min at 4°C with 1 mM PMSF, diluted at least 100-fold with phosphate-buffered saline, and used for the in vivo assay. All assays were performed using at least 5 mice for each concentration of test material. Statistical significance was analyzed by Student's t test.
In Vitro Assay: Bone Resorption Measured in Mice Fetal Long Bone Cultures-Organ cultures of long bones from fetal mice were prepared * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
¶ To whom correspondence should be addressed: Department of Biochemistry, Faculty of Medicine, Kagoshima University, 8-35-1 Kagoshima 890, Japan. Tel: 81-992-75-5242; Fax: 81-992-64-6274. 1 The abbreviations used are: PTH, parathyroid hormone; SCDA, serum calcium-decreasing activity; PCR, polymerase chain reaction; PMSF, phenylmethylsulfonyl fluoride; PAGE, polyacrylamide gel electrophoresis; Tricine, N-[2-hydroxy-1,1-bis(hydroxymethyl)ethyl]glycine; bp, base pair(s); PX, porcine pancreatic extract; HPLC, high performance liquid chromatography. as described by Sato et al. (31). 45 Ca (370 GBq/g calcium, 370 kBq/ mouse; DuPont NEN) was injected into the abdomen of pregnant mice (ICR strain) at the 15th day of gestation, and the mice were killed the following day. The shafts of the radius and the ulna of the fetuses were dissected and cultured in serum free ␣-minimum essential medium (Life Technologies, Inc.). After 24 h of preculture, the bones were cultured for 72 h in ␣-minimum essential medium containing 0.2% bovine serum albumin (Seikagaku Kogyo, Tokyo, Japan), caldecrin fraction, and PTH (10 nM). All assays were performed in quadruplicate using bones from four fetuses. Bioactivity of caldecrin was measured by the inhibition of 45 Ca release into the medium from PTH-stimulated bones and expressed as percent inhibition of PTH-stimulated 45 Ca release minus basal release.
Purification of Rat Caldecrin from Rat Pancreas-Rat caldecrin was purified from acetone powder of rat pancreas according to the protocol described previously (9). The powder was stirred with 0.1 M Tris-HCl buffer (pH 7.5) in 2% NaCl for 1 h at 4°C, and the extract was then centrifuged at 10,000 ϫ g for 30 min. The supernatant was fractionated with acetone (30 -60%), dialyzed against water, and next fractionated with saturated ammonium sulfate (45-60%), followed by dialysis against 50 mM sodium acetate buffer (pH 5.5). The dialyzed material was applied to a Q Sepharose Fast Flow column (Pharmacia Biotech Inc.; 4.5 cm ϫ 16 cm) equilibrated with the same buffer. After having been washed, the absorbed material was eluted by a stepwise gradient of NaCl from 0.1 to 0.5 M in the starting buffer. The second peak, which eluted at 0.2 M NaCl, was collected, dialyzed against 0.2 M ammonium acetate buffer (pH 6.8), and chromatographed through a Superdex 75 FPLC column (Pharmacia) equilibrated with the above-mentioned buffer. Single main peaks that eluted at molecular masses between 22 and 15 kDa were combined. Caldecrin in each fraction was measured by the in vitro assay using fetal long bone cultures and Western blotting using anti-porcine caldecrin antibody.
Immunoscreening of a Pancreatic cDNA Library-Rat pancreas gt11 cDNA library (Clontech) was immunoscreened with rabbit polyclonal anti-porcine caldecrin antibody as the primary antibody. All other staining steps were performed with a picoBlue immunoscreening kit (Stratagene) according to the manufacturer's protocol.
gt11 recombinant lysogens were prepared from positive clones (rat pancreas caldecrin, pRPC) by the method of Huynh et al. (32). Briefly, BNN103 cells were infected with the gt11 clone. All clones that grew at 30°C but lysed the cells at 42°C were cultured at 30°C overnight, at 42°C for 2 min, and at 37°C for 2 h. After centrifugation, the precipitates were subjected to SDS-polyacrylamide gel electrophoresis (PAGE) and Western blotting.
The cDNA library was rescreened with an EcoRI/PstI fragment of pRPC18 used as a probe. Positive clones were subcloned into pUC19, and both strands of inserted DNA were sequenced by use of Sequenase (version 2) (U. S. Biochemical Corp.) and an automatic DNA sequencer (Applied Biosystems).
Northern Blot Analysis-Total RNA was prepared from rat brain, lung, liver, pancreas, kidney, and blood by the acid guanidinium thiocyanate-phenol-chloroform method (33). For Northern blot analysis, RNA (17 g) was denatured at 50°C for 60 min with glyoxal/dimethyl sulfoxide/sodium phosphate buffer, subjected to 1% agarose gel electrophoresis, and transferred onto a nitrocellulose filter (Schleicher & Schuell). After the RNAs had been linked to the filter by a UV linker (Funakoshi Co. Ltd., Tokyo), the filter was pre-hybridized for 18 h at 42°C in 50% formamide, 5 ϫ SSC(1 ϫ SSC: 0.15 M NaCl, 0.015 M sodium citrate, pH 7.0), 10 ϫ Denhardt's solution, 0.1% SDS, 0.2 mg/ml denatured salmon sperm DNA, and 50 mM sodium phosphate buffer (pH 6.5). Full-length rat caldecrin cDNA was labeled with 5Ј-[␣-32 P]dCTP (DuPont NEN) by use of a multiprimed DNA labeling system (Amersham), and employed as a probe. After hybridization overnight at 42°C, the filter was washed three times with 2 ϫ SSC/ 0.1% SDS at room temperature for 5 min and twice with 0.1 ϫ SSC, 0.1% SDS at 58°C for 30 min, and then exposed to x-ray film at Ϫ80°C for 24 h.
Protein Analysis-SDS-PAGE was performed with Tris-Tricine buffer (34), and the gels were stained with the reagents of a Quik-CBB gel staining kit (Wako Pure Chemical Industries, Ltd). The sample for SDS-PAGE was pretreated for 10 min with 5 mM PMSF. Then SDS and ␤-mercaptoethanol were added, and the sample was applied to 16.5% gels. Western blotting analysis with anti-porcine caldecrin antibody was performed in the presence of 5 mM PMSF.
For amino acid sequence analysis of proteolytic fragments, purified caldecrin from rat and pig were digested at 60°C for 1 h with metalloendopeptidase from Grifola frondosa (2.3 units/g) (Seikagaku Kogyo) at a substrate to enzyme ratio of 100:1 (w/w) in 50 mM ammonium acetate buffer (pH 10) containing 5 mM PMSF. The digests were immediately loaded onto a C18 reverse-phase HPLC column (Vyduc 0.46 ϫ 25 cm) and eluted with a linear gradient of 0 -80% acetonitrile in 0.1% trifluoroacetic acid. The amino acid sequences of the peptides were determined with an automated gas-phase protein sequencer (Applied Biosystems).

Construction of a Caldecrin Expression Vector and Heterologous Expression in Mammalian
Cells-To express rat caldecrin in mammalian cells, we inserted the caldecrin cDNA digested with EcoRI into a mammalian expression vector, pCAGGS, which contains a cytomegalovirus enhancer and a ␤-actin promoter (35). The orientation of the cDNA in the vector was confirmed by unique restriction enzyme digestion. Clones with sense and antisense strands were used for transfection experiments. The BMT-10 cells, established by Gerard and Gluzman (36) and kindly provided by Dr. J. Miyazaki, were cultured in Dulbecco's modified Eagle's medium (Life Technologies, Inc.) supplemented with 10% fetal calf serum in a 10% CO 2 atmosphere at 37°C. The cells were transfected with the cDNA by the DEAE-dextran method by use of a mammalian transfection kit (Stratagene) according to the manufacturer's manual. The plasmid containing the rat caldecrin gene (20 g) was mixed with DEAE-dextran (0.5 mg/ml) and used to transfect 1-day cultured subconfluent BMT-10 cells. After 30 min of incubation at room temperature, the cells were washed and incubated with 0.1 mM chloroquine for 5 h in Dulbecco's modified Eagle's medium supplemented with 2% fetal calf serum, and thereafter cultured for the appropriate time without serum. To detect caldecrin expression, the medium conditioned by the cells was dialyzed against distilled water, lyophilized, and analyzed by Western blotting or by the in vitro assay using the fetal long bone culture system.
Mutant Rat Caldecrin cDNA Construction-We introduced mutations H45A and S187A into the rat caldecrin cDNA using a transformer site-directed mutagenesis kit (Clontech). For introduction of the histidine to alanine or the serine to alanine mutation, the mutagenic primer 5Ј-TGTAACGGAGATGCTGGCGGCCC-3Ј or 5Ј-CTCACTGCCGCCGC-CTGCATCAAC-3Ј, respectively, and the transoligo SspI/EcoRV primer CTTCCTTTTTCGATATCATTGAAGCATTT were used. After DNA elongation, ligation, and selection by digestion with SspI, mutated DNAs were used to transform BMH71-18 mutS. The pooled cDNAs were then selected by digestion with SspI, and mutated cDNAs were excised with EcoRI and cloned into pUC19. Sequencing of relevant regions in the two mutants confirmed the presence of the mutation.

Production of Recombinant Wild-type and Mutant Caldecrin Virus and Expression in Insect Cells-
The recombinant wild-type and mutant caldecrins were expressed by the BacPAK baculovirus expression system (Clontech). The wild-type and mutated cDNAs were PCR-amplified with the following primer pairs: 5Ј-TTGAATTCATGTTGGGAATTACG-GTCCTCGCTG-3Ј and 5Ј-TTGAATTCTCACAGTTGTATTTTCTCGT-TGATCCA-3Ј. The products were cleaved with EcoRI, cloned into the transfer vector pBacPAK 9, and then used for co-transfection with Bsu36I-digested, AcMNPV C6 wild-type viral DNA into Sf9 cells.
Purification of Recombinant Wild-type and Mutant Rat Caldecrins-Cultures of Sf9 cells infected with the wild-type or mutant caldecrin baculovirus stock were harvested after 3 days of culture at 27°C in TNM-FH medium. The medium was concentrated 10 times by Amicon ultrafiltration (PM10), dialyzed against 10 mM sodium phosphate buffer (pH 6.8), and applied to a Mono Q FPLC column. After the column had been washed with 10 mM sodium phosphate buffer (pH 6.8) containing 0.15 M NaCl, caldecrin was eluted from it with 0.2 M NaCl. The mutant caldecrins were also purified by the same purification procedures.
In Vivo Assay of Recombinant Wild-type and Mutant Caldecrins-The purified recombinant wild-type and mutant caldecrins produced by the baculovirus expression system were activated at 24°C for 30 min by trypsin at a ratio of 1:50. Activated caldecrin was incubated with or without 1 mM PMSF for 10 min, and the buffer was then changed to phosphate-buffered saline. The mutant caldecrins were activated by the same procedure used for the wild-type caldecrin. Protease activity of recombinant caldecrin was measured with a synthetic substrate, Suc-Ala-Ala-Pro-Phe-pNA, as described previously (10).

Identification and Purification of Rat Caldecrin-To analyze
whether caldecrin is also present in rat pancreas, we examined the rat pancreas homogenate by Western blotting. Using antiporcine caldecrin antibody, we found the presence of an immunoreactive band (data not shown). On the basis of these data, we then purified caldecrin from rat pancreas acetone powder according to the purification protocol as described under "Ex-perimental Procedures," which had been originally developed for porcine caldecrin (9). Fig. 1 shows an SDS-PAGE gel and a Western blot of the purified rat caldecrin. A single band migrating at 30 kDa was revealed by both procedures.
Purified Rat Caldecrin Has SCDA-Purified rat caldecrin caused a dose-dependent decrease in serum calcium concentration, with a maximal decrease of 14% at a dose of 100 g/kg body weight (Fig. 2). These results are very similar to those obtained with porcine caldecrin (9). Treatment of rat caldecrin with PMSF had no effect on SCDA, strongly suggesting that the SCDA of the rat caldecrin does not depend on a functional catalytic site of its protease moiety.
Isolation of Rat Caldecrin cDNA Clones-To isolate rat caldecrin cDNA clones, we screened a rat pancreas gt11 cDNA library (8 ϫ 10 4 clones) with anti-porcine caldecrin antibody. After four rounds of immunoscreening, five immunoreactive clones were identified. Recombinant lysogens of these clones were then expressed and subjected to SDS-PAGE and Western blot analysis (data not shown). One of three positive clones, pRPC18, contained a cDNA insert of 634 base pairs (bp) in length. Sequencing analysis revealed that the insert cDNA of pRPC18 involved an open reading frame predicting a protein having high homology with the elastase family. Since the clone did not contain the full-length sequence, the cDNA library was rescreened with an EcoRI/PstI fragment of pRPC18 as a probe. One positive clone of longer size than pRPC18, containing a cDNA insert of 0.9 kilobase pairs, was then isolated and sequenced.
Tissue-specific Expression of Rat Caldecrin-To estimate the size of the caldecrin mRNA and to analyze its expression pattern in the rat, we analyzed total RNA from various tissues (brain, lung, liver, kidney, and blood) by Northern blotting, using the full-length (899 bp) rat caldecrin cDNA as a probe. With pancreatic RNA, a single band corresponding to an mRNA of approximately 1.1-1.2 kilobases hybridized with the probe; whereas no hybridization band was observed with any other tissue (Fig. 3). This indicates that the rat caldecrin gene is selectively expressed in the pancreas.
Primary Structure of Rat Caldecrin-As shown in Fig. 4, the rat caldecrin cDNA contained a 899-bp insert with a 807-bp open reading frame extending from an ATG codon at nucleotide 14 to a TGA stop codon at position 820. A polyadenylation signal, AATAAA, and a poly(A) tail were found in the 3Јuntranslated region of the cDNA. The deduced caldecrin polypeptide comprised 268 amino acids, including signal peptide, activation peptide, and mature form. The calculated molecular weight of full-length rat caldecrin was 29,374. The nucleotide sequence of rat caldecrin cDNA is very similar (99.3% homology) to that of rat elastase IV throughout the entire coding region (29), suggesting that both mRNAs are products of the same gene.
Amino Acid Sequencing of Rat and Porcine Caldecrin-To investigate whether the purified rat caldecrin corresponds to the protein predicted from the isolated cDNA, we sequenced the purified rat caldecrin. As given in Table I, part A, amino acid sequencing of rat caldecrin through cycles 1-15 showed that a valine appeared on the first cycle, and equivalent moles of glycine and valine on the second cycle, which was followed by pairs of amino acids for the next 7 cycles. The amino acid residue in the 1st cycle and those residues from cycles 10 to 15 corresponded to the amino acid sequences of the predicted mature form of rat caldecrin. Furthermore, pairs of amino acid residues from cycles 2 to 9 consisted of amino acid residues derived from the predicted activation peptide and mature form, but sequencing of the activation peptide was terminated at asparagine at position Ϫ5. Next, purified rat caldecrin was digested in the presence of PMSF with a metalloendopeptidase that cleaves specifically on the amino-terminal side of lysine residues. After separation of the resulting fragments by C18 reverse-phase HPLC, amino acid sequences of several major peptides were determined (Fig.  5A). All amino acid sequences of the six peptides examined  were identical to the corresponding parts of the amino acid sequences deduced from the rat caldecrin cDNA (Fig. 4). Then, to compare at the molecular level the caldecrins purified from rat and porcine pancreas, we also subjected porcine caldecrin to the same metalloendopeptidase cleavage protocol. Again, all sequences of the nine fragments obtained showed high homology with the deduced amino acid sequence of rat caldecrin (Figs. 4 and 5B). These results indicate that the caldecrins purified from rat and porcine pancreas are both members of the same elastase IV subfamily.
Expression of Rat Caldecrin in BMT-10 Cells-To examine the functional properties of recombinant rat caldecrin, we prepared a construct of the rat caldecrin cDNA in the pCAGGS expression vector and introduced the latter by the DEAE-dextran method into BMT-10 cells. Upon SDS-PAGE and Western blot analysis of cells harvested after 2 days in culture, a 30-kDa shown. An open reading frame extends from the translation initiation site (ATG) to the termination site (TGA). This is followed by a 3Ј-untranslated region that ends with a polyadenylation signal (bold underlined) and the poly(A) tail. The vertical arrowheads indicate proteolytic processing sites. The deduced amino acid sequence is numbered sequentially from the amino terminus of the predicted active enzyme. The amino acid residues of the charge-relay system are circled (37), and the residues determining the substrate binding specificity are boxed (15). The corresponding amino acid sequences of fragments obtained by proteolytic cleavage of rat caldecrin are underlined with a dashed line. The amino acid sequences of fragments analyzed after proteolytic cleavage of porcine caldecrin are shown below the rat caldecrin sequences (third row). The deduced amino acid sequence of elastase IV (NBRF-PDB accession number JQ1473) is given in the bottom row. Asterisks indicate residues conserved between rat caldecrin and rat elastase IV.
immunoreactive band was detected in lysates of cells bearing sense-strand cDNA but not in those containing the antisense cDNA or vector alone (Fig. 6). The culture medium of the transfected cells was also examined, since pancreas proteases are secretory proteins. As shown in Fig. 6, caldecrin immunoreactivity was high in the medium of cultured cells that had been transfected with the sense cDNA, whereas the medium of cultured cells transfected with antisense cDNA or vector alone did not contain detectable amounts of antigen.
Recombinant rat caldecrin was also examined for hypocalcemic activity. Medium conditioned by the cells transfected with the caldecrin cDNA was used in the in vitro assay utilizing fetal long bone cultures, which cultures have been shown to be a valid model system to study porcine caldecrin activity (9). As shown in Fig. 7, medium containing recombinant rat caldecrin inhibited dose-dependently the PTH-stimulated release of 45 Ca from the bone. The conditioned medium caused maximum inhibition at 20 l/ml of assay medium and maintained this level of inhibition at higher doses. The conditioned medium of cultured cells transfected with antisense cDNA did not affect the PTH-stimulated release of calcium from the bone.
Bioactivity of Mutant Caldecrins-To verify that the SCDA of caldecrin is not connected to its protease activity, we prepared recombinant wild-type and mutant caldecrins were synthesized by use of the baculovirus system, purified them by passage through a Mono Q column, and assayed them for SCDA. Sf9 cells transfected with recombinant wild-type or mutant caldecrin baculoviruses were synthesized and secreted polypeptides with molecular masses of about 60 and 30 kDa in the medium. The 30-kDa polypeptide was found to be caldecrin by Western blotting. A single step ion-exchange procedure was sufficient to obtain almost pure recombinant caldecrins with

Amino-terminal sequences of purified rat caldecrin (A) and comparison of the amino-terminal sequences deduced from rat caldecrin cDNA with other related enzymes (B)
A. Purified rat caldecrin cycles:

CGLPANLPQ-LPR VVGGEDARPN
Human proelastase IIA c  90 -95% purity as judged by SDS-PAGE. The purified wild-type caldecrin showed chymotrypsin activity with a synthetic substrate only after treatment with trypsin, indicating that it was the proform of caldecrin (10). Fig. 8 shows that serum calcium was decreased by the activated wild-type caldecrin but not by the proform of caldecrin. Again, treatment of the activated caldecrin with PMSF destroyed the chymotrypsin activity but permitted retention of the SCDA. Mutations of His-45 and Ser-187, which residues are required for serine protease activity, decreased the chymotrypsin activity to 0.6 and 0.11%, respectively, of that of wild-type calderin. However, the activated mutant caldecrins (H45A and S187A) also decreased serum calcium levels. These results suggest that the SCDA of caldecrin has no connection with the protease activity of the molecule.

DISCUSSION
In this paper, we reported the primary structure of rat caldecrin, a proteolytic enzyme from rat pancreas, and showed that it is highly related or identical to elastase IV, a previously analyzed protease of the chymotrypsin/elastase superfamily. This conclusion is based on the following findings. (i) The rat caldecrin cDNA encodes a serine protease zymogen consisting of a signal peptide, an activation peptide, and a mature enzyme. Eight cysteine residues present at positions 30,46,126,157,173,183,193, and 214 of the mature protein are conserved in all elastase family members, which suggests the formation of four disulfide bonds as observed in porcine elastase I (15). In addition, three cysteine residues are present at the positions Ϫ14, Ϫ13, and 112, indicating that a cysteine residue at position Ϫ14 or Ϫ13 may form a disulfide bond with Cys-112. (ii) The amino-terminal four residues of the mature form of pancreatic elastase, Val-Val-X-Gly, are highly conserved and these residues follow an arginine residue. Table I, part A, corroborates that rat caldecrin is synthesized as a zymogen with a signal peptide of 16 residues, and an activation peptide of 13 amino acids. Moreover, these data suggest that a cysteine (at position Ϫ13) of the activation peptide forms a disulfide link with an internal cysteine, i.e. 112, and that the activation peptide is associated with the enzyme after proenzyme activation, a situation analogous to that seen with both elastase II and chymotrypsin processing, as given in Table I, part B (11-13, 19, 22, 29, 37). (iii) The nucleotide sequences of rat caldecrin and rat elastase IV genes are 99.3% identical, but the deduced amino acid sequences display only 90.3% identity (Fig. 4). These results strongly suggest that rat caldecrin could be an isoform of elastase IV generated by some unusual splicing mechanism. Alternatively, and in our view more likely, however, the differences between our protein sequence and the elastase IV sequence published by Kang et al. (29) result from a frameshift reading error in the corresponding region of the cDNA by the latter group (see Fig. 4). Unfortunately, the elastase IV protein has so far not been investigated. Further studies will be required to resolve this problem. In any event, a stretch of the deduced amino acid sequence of rat caldecrin cDNA, which is quite different from that of elastase IV, was identical to the peptide sequence of a fragment obtained from the purified rat caldecrin and had a high homology with the corresponding peptide sequence of purified porcine caldecrin, indicating that the caldecrin gene is really translated. Moreover, the deduced amino acid sequence of the human caldecrin cDNA that we recently cloned displays, within the region of sequence divergence between caldecrin and elastase IV, a very high identity with the rat caldecrin analyzed in a study to be published elsewhere. 2 The amino acid residues characteristic of the serine protease catalytic triad (38) are retained at positions His-45, Asp-92, and Ser-187 of the deduced amino acid sequence. The other key amino acid residues, Gly-209 and Val-221, thought to contribute to the substrate specificity of such enzymes, may enlarge the substrate binding pocket to accommodate more bulky amino acid side chains (19). The presence of an activation peptide similar to that found in chymotrypsin and elastase II suggests a chymotrypsin-like substrate preference and indicates that caldecrin may be an evolutionary link between chymotrypsin and elastase I. Anionic elastase possessing chymotrypsin activity has not been found in rat or porcine pancreas. The partial amino acid sequence of anionic chymotrypsin C has been reported, but its cDNA sequence has not. The partial amino acid sequence around an essential histidine of porcine  8. SCDA of recombinant wild-type and mutant caldecrin is not required for the protease activity. Samples of recombinant wild-type (proform), trypsin-activated wild-type, and mutant caldecrins were assayed for chymotrypsin activities, and others were injected into mice at a dose of 100 g/kg body weight. Blood was taken 4 h after the injection, and the serum calcium concentration was then measured. chymotrypsin C (39) is not identical to that of porcine caldecrin.
So far, many proteases with elastolytic activity have been isolated from various tissues. Thus, careful consideration should be given to this nomenclature, in particular to elastase numbering. Rat and porcine caldecrins might be classified as elastase IV family members based on the current nomenclature rules, although the SCDA of these proteins apparently is not connected with protease activity. Yoneda et al. (40) have reported partial purification of a factor in a porcine pancreatic extract (PX) that decreases the blood calcium level, and they also showed that PX prevents progression of hypercalcemia and cachexia in mice inoculated with a carcinoma cell line (41). Amino-terminal sequencing of PX indicates that it belongs to the elastase IIIB family, and recombinant elastase IIIB displays hypocalcemic activity that is dependent on its protease activity (42). The reasons for this discrepancy between their results and ours are presently unclear. In any case, all presently available data show that same hypocalcemic factor(s) exists in the pancreas, and their structural analysis uncovers a novel regulatory function of these proteases. However, rat and porcine caldecrins treated with PMSF and even two kinds of the catalytic site mutants of rat caldecrin still displayed SCDA, suggesting that caldecrin may possess functional residues for SCDA that differ from those involved in the protease activity. The proform of caldecrin does not possess SCDA but acquires this property as well as protease activity after activation by trypsin treatment. These results suggest that the residues responsible for SCDA require a trypsin-induced conformational change to be come exposed at the outer surface of the molecule. The SCDA of caldecrin correlates with a decrease in serum hydroxyproline levels and with the inhibition of PTH-stimulated bone resorption (9), suggesting that caldecrin suppresses osteoclastic activity through an as yet unknown mechanism. All these data indicate that caldecrin, a protease closely related or identical to elastase IV, is a multifunctional protein that may be implicated in the regulation of blood calcium levels.