Cloning, characterization, and tissue distribution of porcine SPAI, a protein with a transglutaminase substrate domain and the WAP motif.

The primary and gene structures and tissue distribution of porcine SPAI-2, a protein that belongs to the WAP protein superfamily and has a sodium-potassium ATPase inhibitory activity, were determined by molecular cloning and Northern analysis. A full-length cDNA clone was isolated from a porcine duodenum cDNA library. The cDNA insert encoded a polypeptide of 187 amino acids, which is composed of three domains: a hydrophobic presequence of 21 amino acids, a prosegment of 105 amino acids ending with Asp, and the mature SPAI-2 sequence of 61 amino acids beginning with Pro. The prosegment contained 16 repeats of a hexapeptide that is highly homologous to the repetitive sequence found in the transglutaminase domain of the human elafin, an elastase-specific inhibitor that also belongs to the WAP superfamily. The repetitive sequence was demonstrated to be a good substrate of transglutaminase using a recombinant preparation produced in Escherichia coli. A porcine genomic library was then screened for the SPAI gene. Characterization and sequencing of positive clones indicated that the gene is similar to the elafin gene, having 3 exons encoding the 5′-untranslated region and signal sequence, proSPAI, and 3′-untranslated region, respectively. Northern blot analysis revealed intestine-specific expression of SPAI mRNA; the message was especially abundant in the small intestine. ProSPAI was also found in the circulation. The similarity of proSPAI to elafin in the domain structure, the acid-labile nature of the cleavage site (Asp-Pro), and the fact that the major form of SPAI in the plasma is proSPAI strongly suggest that proSPAI is not the precursor but rather it is the native form of SPAI. Like elafin, therefore, SPAI appears to be a new type of biologically active substance with a transglutaminase substrate domain that acts as an anchoring sequence.

SPAI-2 1 is a 61-amino-acid peptide isolated from porcine duodenum as a Na ϩ ,K ϩ -ATPase inhibitor (1) and is now considered as an inhibitor specific for monovalent cation-transporting ATPases (2); however, whether this inhibitory action observed in vitro is its true physiological function remains to be established since the IC 50 is relatively high (ϳ10 M). It belongs to the WAP protein superfamily that has the "four-disulfide core" (3) or "WAP" motif (4); the members include, among others, mucous proteinase inhibitor MPI (5-7) and elafin, an elastase inhibitor that was isolated by Wiedow et al. (8) from the horny layers of patients with psoriasis. MPI and elafin are well characterized at the molecular level; for example, their domain structures and gene structures have been determined (9,10). Less is known about the molecular biology of SPAI. We therefore cloned cDNA from a porcine duodenum cDNA library, determined its nucleotide sequence to deduce the primary structure of the precursor, and examined the exon-intron organization of the corresponding gene. The results indicated that SPAI is much more similar to elafin than to MPI in its domain and gene structures, suggesting that SPAI and elafin are, despite their difference in biological activities, evolutionarily very close among the members of the WAP superfamily.
At present, the physiological significance of SPAI is not clear. Localization often offers valuable clues to the physiological roles. We therefore also examined 1) tissue distribution of the SPAI message by Northern analysis and 2) whether SPAI is present in plasma by immunoaffinity chromatography.

EXPERIMENTAL PROCEDURES
Materials-Fresh porcine tissues were obtained from the Shibaura abattoir sanitation inspection station, Tokyo, Japan. Restriction enzymes were obtained from Toyobo, Osaka, Japan; random primer DNA labeling kit version 2 was from Takara, Kyoto, Japan; Sequenase version 2.0 sequencing kit was from U. S. Biochemical Corp; 32 P-labeled nucleotides were from DuPont NEN. Rabbit polyclonal antiserum to porcine SPAI was kindly donated by Peptide Institute Inc., Osaka, Japan. Monoclonal antibodies against porcine SPAI were obtained by standard hybridoma techniques. The hybridoma producing the monoclonal antibodies 1F4 and 1H12 were amplified in the ascites fluid of mice. The IgG produced by the ascites induction was then purified on Protein A-Sepharose 4 FF (Pharmacia Biotech Inc.) according to the instructions of the manufacturer. Monospecificities of the antibodies were established by Western blotting; they stained only the SPAI band among various protein bands when proSPAI, prepared as described below, was mixed with pig kidney extracts and analyzed by SDS-PAGE and immunostaining (11).
Preparation of the Oligonucleotide Probe SP-P5A-The oligonucleotide used for cDNA library screening was designed according to the nucleotide sequence of short PCR products. The following PCR also served as a test for the presence of SPAI message. Poly(A) ϩ RNA (170 g) was isolated from porcine duodenum (2 g) using a FastTrack mRNA isolation kit. A mixture of oligo(dT)-primed cDNA was synthesized from the duodenal mRNA using reverse transcriptase and used as templates for PCR. The primers used for PCR were made based on the consensus amino acid sequence of the three SPAIs reported previously (1); the consensus regions selected were those corresponding to amino acid residues 34 -40 (5Ј primers) and amino acid residues 52-57 (3Ј primers) of SPAI-2. The sequences of the 5Ј primers were: CAAGCTTGAA (T/  C . The 5Ј ends of the 5Ј and 3Ј primers contained the HindIII and EcoRI restriction sequences, respectively, for the subsequent cleavage of cloned cDNA fragments. All six possible combinations of the primers (three 5Ј primers versus two 3Ј primers) were examined using a GeneAmp DNA amplification kit (Perkin-Elmer) according to the following schedule: 94°C for 1 min, 44°C for 2 min, and 72°C for 3 min for 30 cycles, followed by further incubation for 7 min at 72°C. Only two combinations (primer 53 versus primer 31 and primer 53 versus primer 32) yielded amplified products of about 80 bp. After electrophoretic purification on an agarose gel, the products were treated with T4 polymerase and T4 kinase, subcloned into M13 mp18 at the SmaI site and sequenced. The sequence flanked by the primers were 5Ј-TGACTGTCCAGGGGTCAAGAAGTGCTGT-GAAGGC-3Ј (SP-P5) and corresponded, as expected, to amino acid residues 40 -51 of SPAI-2; therefore, the antisense sequence (SP-P5A) complementary to SP-P5 was chemically synthesized and used as an oligonucleotide probe for screening.
Construction and Screening of cDNA Library-A cDNA library was constructed, using an Amersham cDNA cloning system, in gt10 from porcine duodenum poly(A) ϩ RNA isolated with a FastTrack mRNA isolation kit (Invitrogen). Plaques were screened using the Escherichia coli strain NM514 by hybridization of 32 P-labeled SP-P5A, a synthetic 36-base antisense probe, to nitrocellulose replicates in 50 mM sodium phosphate buffer, pH 7.0, containing 60 mM sodium citrate, 0.6 M NaCl, 1 ϫ Denhardt's (1 ϫ Denhardt's: 0.1% each of bovine serum albumin, polyvinylpyrrolidone, and Ficoll) at 65°C. Filters were washed in a solution containing 0.6 M NaCl, 60 mM sodium citrate, and 50 mM sodium phosphate (pH 7.0) at 65°C. Positive clones were plaque-purified, and the cDNA inserts were subcloned into pBluescript II (Stratagene) for sequencing.
Southern Blot Analysis of Porcine Genomic DNA-Porcine genomic DNA was isolated from liver. Ten g of the DNA were digested with EcoRI, HindIII, and XbaI. The digests were electrophoresed in a 0.7% agarose gel at 0.6 V/cm and transferred to Magnagraph nylon membrane (Micron Separations Inc.) using an LKB VacuGene vacuum blotting unit. The membrane was prehybridized at 42°C for 2 h and hybridized with the 32 P-labeled XbaI-PvuII fragment (561 bp) derived from SPAI-2 cDNA at 42°C in 50% formamide, 5 ϫ SSPE, 5 ϫ Denhardt's, and 0.1% SDS and followed by washes and autoradiography as described above.
Enzyme-linked Immunosorbent Assay-To determine plasma SPAI levels and to monitor purification of SPAI-LI, the following sandwich ELISA was devised, consisting of solid-phase monoclonal antibody 1H12 as the capture reagent and biotinylated monoclonal antibody 1F4 as the second antibody. Microtiter plates were coated with affinitypurified monoclonal antibody 1H12 (10 g/ml) in 20 mM phosphatebuffered saline, pH 7.4, at 4°C overnight. After blocking with Block Ace (Dainippon Seiyaku; diluted 4-fold in H 2 O) for 1 h at room temperature, the wells of the plates were loaded with 100 l of samples and allowed to stand for 1 h at room temperature followed by washing with 0.9% NaCl containing 0.05% Tween (saline-Tween). Subsequently the wells were incubated at room temperature for 1 h with a 10-fold diluted Block Ace solution containing biotinylated monoclonal antibody 1F4 (4.5 g/ ml), which had been prepared by coupling the purified 1F4 IgG with N-hydroxysuccinimide biotin (Pierce). Following washing with saline-Tween, to quantitate the sandwiched antigen, the wells were then incubated for 2 h at room temperature with 100 l of horseradish peroxidase avidin D (Vector Laboratories) (0.2 g/ml of Block Ace diluted 1:10 with water), and developed with o-phenylenediamine (Sigma) and hydrogen peroxide as substrates.
Purification of SPAI-like Immunoreactivity from Plasma-Porcine serum (1 liter) was acidified to pH 2.0 with 2 N HCl and filtered through no. 2 qualitative filter papers (Advantec) to remove precipitates. The filtrate was loaded onto a silica column (Sep-Pak Vac, Waters), which had been equilibrated with 0.1% trifluoroacetic acid. After washing with 0.1% trifluoroacetic acid, the column was eluted with 15% CH 3 CN containing 0.1% trifluoroacetic acid. The eluate was lyophilized, redissolved in 75 ml of 20 mM Tris-HCl buffer, pH 7.4, containing 0.5 M NaCl, and applied to an immunoaffinity column (1.5 ϫ 3 cm), which had been prepared by cross-linking 5 mg of the anti-SPAI monoclonal antibody 1H12 to 5 ml of CNBr-activated Sepharose 4B (Pharmacia) according to the manufacturer's instruction, at a flow rate of 5 ml/h and equilibrated with 20 mM Tris-HCl, pH 7.4, containing 0.5 M NaCl. After washing the column with the same buffer, the absorbed proteins were eluted with 0.1 M Gly-HCl, pH 2.5, lyophilized, and dissolved in 1 ml of 10% CH 3 CN containing 0.1% trifluoroacetic acid. The affinity-purified sample was further purified by reverse-phase HPLC on a YMC-GEL C 18 column (4.6 ϫ 250 mm). The HPLC column was eluted first by an isocratic mode with 25% CH 3 CN containing 0.1% trifluoroacetic acid for 3 min and then by a linear gradient from 25% to 30% CH 3 CN containing 0.1% trifluoroacetic acid for 40 min at a flow rate of 1 ml/min monitoring absorbance at 225 nm. Every peak was collected and assayed for porcine SPAI-LI.
Sequencing of Plasma SPAI-LI-The SPAI-LI purified by immunoaffinity chromatography and HPLC was sequenced after carboxymethylation, fragmentation with chymotrypsin, and, in the case of the N-terminal fragment, pyroglutamyl peptidase treatment. SPAI-LI 1 (2 nmol), the major peak obtained by the C 18 HPLC, was dissolved in 200 l of 1 N Tris-HCl, pH 8.6, containing 3.5 M guanidine and 0.25% EDTA and reduced with 20 l of 0.6 M dithiothreitol under N 2 gas. Following a 2.5-h incubation at 37°C, 30 l of 0.6 M carboxymethyliodide and 30 l of 0.5 N NaOH were added to the mixture. Carboxymethylated SPAI-LI 1 was purified by HPLC on an ODS column (4.6 ϫ 250 mm, YMC) using a 30-min acetonitrile gradient from 26% to 34% containing 0.1% trifluoroacetic acid at a flow rate of 1 ml/min. The carboxymethylated SPAI-LI 1 preparation was then dissolved in 200 l of 0.1 M Tris-HCl, pH 8.0, and digested with 1 g of N ␣ -p-tosyl-L-lysine chloromethyl ketone-treated chymotrypsin (Sigma) at 37°C for 2.5 h. The digest was chromatographed on the 4.6 ϫ 250-mm ODS column, and five peaks were obtained using a 60-min 1-32% gradient of acetonitrile containing 0.1% trifluoroacetic acid. The retention times of the first four peaks coincided with those of the chymotryptic fragments of the carboxymethylated form of the mature SPAI previously isolated from porcine duodenum (1), and their identities were confirmed by sequencing. The fifth peak, which was considered to be the N-terminal fragment, was lyophilized, redissolved in 100 l of 0.1 M Tris-HCl, pH 8.0, and treated with 755 milliunits of pyroglutamate aminopeptidase (Nakarai) at 37°C for 6 h. The reaction was monitored by the reverse phase HPLC described above using a 40-min acetonitrile gradient from 25 to 30%. The cleaved peptide was isolated with the same HPLC condition and sequenced in an Applied Biosystems 473A pulsed liquid sequenator.
Construction of MBP-proSPAI Expression Vector-A cDNA fragment coding for porcine proSPAI was prepared by PCR using SPAI cDNA as template. The following oligonucleotides were synthesized to amplify a segment of the porcine SPAI cDNA sequence corresponding to amino acid residues 21-167 (Fig. 2): The sense primer was a 17-mer (5Ј-GCGCAGAGACTTGACAG-3Ј), and the antisense primer was a 24-mer (5Ј-GAAGCTTTCACTTGGGATACAAAC-3Ј). The PCR reactions were carried out with Taq DNA polymerase for 30 cycles (1 min at 94°C, 0.4 min at 47°C, and 2.5 min at 75°C). The resulting PCR product was purified, blunt-ended, digested with HindIII, and ligated into the StuI/ HindIII site of the pMAL-p vector (New England Biolabs). The ligations were used to transform E. coli XL1-Blue, and the desired recombinants were selected by sequencing their plasmid DNA (pMAL-p-proSPAI).
Expression and Purification of ProSPAI-Production of MBP-proS-PAI and cleavage of the fusion protein with Factor Xa were carried out using a New England Biolabs expression kit exactly as described by the manufacturer. Briefly, XL1-Blue cells transformed with pMAL-p-proS-PAI were grown at 37°C for 4 h and the expression of the MBP-proSPAI fusion protein was induced by adding isopropylthiogalactoside. Following a 2-h incubation, the cells were pelleted, washed with 20% sucrose, and treated with 5 mM MgCl 2 on ice, and the periplasmic fraction (cold osmotic shock fluid) was recovered by centrifugation. Cold osmotic shock fluid (1.5 l) was applied to a column containing amylose resin (100 ml), and the MBP-proSPAI fusion protein was eluted with 10 mM maltose. The eluate, containing 70 mg of protein in 40 ml, was dialyzed against 50 mM Tris-HCl, pH 8.0, containing 100 mM NaCl and 1 mM CaCl 2 , cleaved with Factor Xa (10 g; Danex Biotek, Mundelstrup, Denmark) at 20°C for 10 h, and dialyzed against 10 mM phosphate, pH 7.2. The final purification of proSPAI was achieved by removing the MBP fragment by HPLC on a hydroxyapatite column (1.2 ϫ 5 cm) equilibrated with 10 mM phosphate buffer, pH 7.2. When stained with Coomassie Brilliant Blue after SDS-PAGE, proSPAI gave a reddish band that is quite different from those of ordinary proteins that stain blue.

RESULTS
Isolation and Sequence Analysis of Porcine SPAI cDNA-A gt10 cDNA library was prepared from porcine duodenum and screened with an oligonucleotide probe, SP-P5A, designed from the amino acid sequence of SPAI-2 (1). This screening (of about 3 ϫ 10 6 recombinants) yielded one positive clone. The positive phage clone was named 57G and characterized. The insert was excised with EcoRI and subcloned into the EcoRI site of pBluescript II for sequencing. A restriction map and sequencing strategy for the cDNA clone are shown in Fig. 1. Fig. 2 shows the nucleotide and deduced amino acid sequence of clone 57G. The cDNA insert consisted of a total of 741 bases, with 6 bases upstream of the ATG codon and a poly(A) tail. An open reading frame of 561 bases encoded a protein of 187 amino acids. The protein sequence contained a hydrophobic signal sequence at its N terminus, a prosequence, and the SPAI sequence at the C terminus (Fig. 4C). One of the most striking features is the presence of 16 repeats of a hexapeptide with the following consensus sequence: GQDPVK, which is repeated 4 times intact and 12 times in part (Fig. 3). Since the Asp 126 -Pro 127 bond that is assigned as the prosequence cleavage site is known to be susceptible to strong acid hydrolysis and such acidic conditions were used in our previous isolation of SPAI, it is highly likely that the cleavage of Asp 126 -Pro 127 bond might be an experimental artifact; however, for simplicity, we use, in this paper, the term "proSPAI" for the higher molecular weight native form of SPAI to distinguish it from the short form of 61 Signal sequence and prosequence cleavage sites are indicated by arrowhead and upward arrow, respectively. Note that the "prosequence" cleavage site (Asp 126 -Pro 127 ) is acid-labile and is readily cleaved when exposed to strong acid (15,16). Eight Cys residues involved in the formation of four-disulfide core are shaded. amino acids that was isolated from acidified porcine duodenum (1).
Isolation and Characterization of the SPAI Gene-A porcine genomic DNA library in EMBL3 SP6/T7 was screened with 32 P-labeled full-length cDNA, and numerous phage clones were obtained. One set of the positive clones contained a 5.4-kb EcoRI fragment that hybridized strongly to the probe. The EcoRI fragment was subcloned into the plasmid vector pBluescript II for further characterization. After its mapping using various restriction enzymes and Southern blot analysis of the resulting fragments, SacI digestion was found to produce three fragments appropriate for sequencing the exon-intron boundaries: a 1.3-kb fragment with two XbaI sites one of which corresponds to the XbaI site in the cDNA sequence, a 1.2-kb fragment with a unique SpeI site corresponding to that in the cDNA sequence, and a 2.9-kb fragment. Sequence analysis of these fragments according to the strategy shown in Fig. 4A indicated that the porcine SPAI gene spans approximately 1.8 kb and is divided into 3 exons and 2 introns (Fig. 4B); detailed positions of the introns are shown in the Fig. 2 sequence. Fig. 5 shows the result of Southern blot analysis performed to determine the copy number of the porcine SPAI gene. Genomic DNA was digested with three restriction enzymes, transferred to a filter, and hybridized to the XbaI-PvuII fragment of SPAI cDNA. In all digests, single dense bands and two or three faint bands were detected under moderate stringency conditions (Fig. 5A); the faint bands disappeared under stringent conditions (Fig. 5B). The sizes of the dense bands are in good agreement with those predicted from the genomic restriction map illustrated in Fig. 4A, indicating that the SPAI gene is present as a single copy in the porcine genome. Three forms of SPAI have been isolated (1): SPAI-1, SPAI-2, and SPAI-3. SPAI-1 is a N-terminally truncated form of SPAI-2. SPAI-2 and SPAI-3 are of the same length and differ only at two positions (Arg 148 3 Gly, Ser 156 3 Gly); each substitution can be produced by a single nucleotide change in the corresponding codon. The substitutions are therefore probably allelic variants in the SPAI-2 gene. The additional bands with relatively weak hybridization signals suggest the existence of at least two other closely related genes.
Tissue Distribution of mRNA-To define tissue distribution patterns of SPAI, expression of the SPAI mRNA in various porcine tissues was studied by Northern blot analysis. Total RNA preparations from the brain, liver, lung, adrenal, stomach, small intestine, large intestine, and kidney were electrophoresed on an agarose gel, transferred to nylon membrane, and probed with 32 P-labeled cDNA. A transcript of about 900 nucleotides was detected in the intestine but in none of the other tissues examined (Fig. 6). Within the intestine, SPAI mRNA was much more abundant in small intestine than in large intestine.
A High Molecular Weight Form of SPAI in the Circulation-We further examined SPAI levels in plasma by immunoaffinity chromatography. A monoclonal antibody to SPAI, 1H12, was produced, purified with Protein A, and coupled to Sepharose 4B. Immunoreactive SPAI in porcine plasma was adsorbed to the immunoaffinity gel, eluted with glycine-HCl, and further purified by reverse-phase HPLC on a C 18 column (Fig. 7A). Chemical sequencing of the major peak indicated that porcine plasma contains a SPAI species much larger than the SPAI originally isolated from duodenum by us (1). As shown in Fig. 7B, the N terminus of the circulating form corresponded to Gln at position 22 in the nascent biosynthetic precursor (Fig.  2). This fact indicates that 1) the signal sequence of porcine SPAI consists of 21 amino acids and 2) the plasma species is an unprocessed form of the precursor proSPAI and contains the complete sequence of the repetitive domain (Fig. 3) and SPAI.
Transglutaminase-mediated Cross-linking of ProSPAI-To demonstrate that the N-terminal repetitive sequence of proS-PAI serves as a substrate for transglutaminase, we expressed and purified proSPAI using the MBP fusion protein system (New England Biolabs), and examined whether it is crosslinked by transglutaminase. Highly purified proSPAI migrated as a closely spaced doublet on SDS-PAGE (Fig. 8, lane 2); the upper band could be converted to the lower band by increasing the incubation time with Factor Xa, which was used to cleave proSPAI from the MBP-proSPAI fusion protein. Furthermore, both bands reacted with anti-SPAI antibodies, indicating that the doublet represents two forms of the same protein probably generated by limited proteolysis during purification. When treated with transglutaminase, proSPAI was readily crosslinked and yielded higher molecular weight species (Fig. 8,  lane 3); under the same experimental conditions, MBP and mature SPAI, which lacks the repetitive sequence, were not cross-linked.  5. Southern blot analysis of porcine genomic DNA probed with SPAI cDNA. High molecular weight genomic DNA was isolated from porcine liver, digested to completion with the indicated restriction enzymes, electrophoresed on an agarose gel, transferred to nylon membrane, and hybridized to the 32 P-labeled cDNA clone 57G. The filter was washed in 1 ϫ SSC containing 0.1% SDS at 55°C for 2 h (A), followed by a 2-h wash at 65°C (B), before autoradiography. DISCUSSION We have isolated and characterized a full-length cDNA and genomic DNA clones for porcine SPAI-2. The amino acid sequence deduced from the nucleotide sequence coincided with that previously determined by direct sequencing of purified SPAI (1), validating our previous work. The presence of a hydrophobic presequence indicates that SPAI-2 is a secreted protein. The presence of SPAI in the plasma matches this secretory nature. Northern blot analysis indicated that SPAI-2 mRNA is essentially confined to small intestine (Fig. 6). Taken together, these results indicate that although the major site of action of SPAI-2 is probably the intestine, it can be delivered to many other tissues through the circulation.
The sequence comparison between the "mature" SPAI and its precursor revealed that the precursor is cleaved at the Asp 126 -Pro 127 bond (Figs. 2 and 4C). The Asp-Pro bonds in proteins, however, have been demonstrated to be unstable under strongly acidic conditions (15,16). The fact that the mature SPAI was isolated from duodenum extracts exposed to such acidic conditions (1), therefore, strongly suggests that the 61amino acid SPAI might occur as an artifact during purification. In support of this possibility, we could detect only proSPAI in plasma, suggesting that the proSPAI is not a precursor of SPAI, but rather it is the native form of SPAI.
A wide variety of proteins have been shown to have the four-disulfide core structure and constitute the WAP protein superfamily. They include neurophysins (17), plant agglutinin (3), adhesion molecules (18), whey proteins (19 -22), proteinase inhibitors (6,23,24), scorpion toxins (25), bactericidal peptides (26), pollen proteins (27), and SPAI (1). So far the gene structures of the following three members have been determined: neurophysins (28 -30), mucous proteinase inhibitor MPI (9), and elafin (10). The structure of the SPAI-2 gene determined here is very similar to that of elafin (10) including the intron positions, indicating that the two genes have arisen by a gene duplication. The genes for other members might also have arisen from a common ancestor gene by exon and gene duplications followed by gene conversion (28,29) and intron insertion (9).
It is interesting that SPAI has a repetitive sequence in its N-terminal region (Fig. 3) that is very similar to that found previously in the human elafin precursor (10), whose sequence characteristics (i.e. rich in Gln and Lys) led us to propose that the repetitive sequence serves as a substrate for transglutaminase, an enzyme that catalyzes the formation of N ⑀ -(␥-glutamyl)lysine isopeptide cross-links between proteins (31, 32). This was soon confirmed to be the case by Molhuizen et al. (33) and by ourselves (11), establishing that elafin consists of two domains: the transglutaminase substrate domain and the elastase inhibitor domain. The transglutaminase substrate domain serves as an anchor to localize elafin covalently to specific sites on extracellular matrix proteins; therefore, we termed the transglutaminase substrate domain of elafin the cementoin moiety (11). The striking similarity between the repetitive sequences of elafin and SPAI suggests a similar role for the SPAI N-terminal extension; indeed, cross-linking experiments indicated that the SPAI repetitive sequence is a good substrate for transglutaminase (Fig. 8). The chances of the interaction between the anchored SPAI and Na ϩ ,K ϩ -ATPase seem, however, very small since the enzyme is a membrane protein and its movement is restricted. The possibility, therefore, remains that SPAI, especially the anchored SPAI, might have other as-yetunidentified major targets of soluble nature. The structural FIG. 6. Intestine-specific expression of SPAI mRNA revealed by Northern blot analysis. About 20 g of total cellular RNA isolated from the indicated tissues was fractionated on a denaturing gel and transferred to a nylon membrane filter. The filter was probed with a radiolabeled cDNA probe and autoradiographed. Equal loading of RNA is demonstrated by the ethidium bromide staining of the 28 and 18 S rRNA bands (lower panel).

FIG. 7. Purification of plasma SPAI-like immunoreactivity (SPAI-LI) and its identification as proSPAI by N-terminal amino acid sequencing.
A, a typical elution profile of SPAI-LI from a reversephase C 18 HPLC column. SPAI-LI was first extracted with a silica cartridge from porcine plasma (purity, Ͻ 0.1%; yield, 80%) and partially purified with an antibody column (purity, 25%; yield, 55%). The affinity-purified SPAI-LI was then loaded onto the HPLC column and eluted by a CH 3 CN gradient. One major and one minor peaks of SPAI-LI were emerged and termed SPAI-LI 1 and SPAI-LI 2 , respectively, as indicated by shaded boxes. About 2 nmol or 37 g of SPAI-LI 1 was obtained from 1 liter of plasma with a yield of 43% (estimated by ELISA); normal plasma levels of porcine SPAI were 1-5 nM. Although plasma samples did not contain low molecular weight SPAI species (i.e. SPAI-1, SPAI-2, and SPAI-3), their elution positions are also indicated by arrows. B, the N-terminal amino acid sequence of SPAI-LI 1 determined by direct sequencing. Initial sequencing attempts failed probably because of the modification of the N terminus. We therefore determined the sequence after removal of the N-terminal pyroglutamate with pyroglutamyl peptidase. The sequence analysis and alignment with the deduced amino acid sequence (Fig. 2) indicated that the SPAI-LI 1 purified from porcine plasma begins at Gln 22 as indicated by numbers in parentheses. The amount of SPAI-LI 2 was not sufficient for sequencing. similarity between SPAI and elafin suggested that SPAI is also a proteinase inhibitor; we therefore examined its effects on the following proteinases by the standard assays using chromogenic or fluorogenic synthetic substrates, but no inhibitory activity was observed even at 5 M: leucocyte elastase, leucocyte chymotrypsin-like proteinase (cathepsin G), pancreatic elastase, trypsin, ␣-chmotrypsin, and plasmin. This result indicates that, at least, SPAI is not the porcine homolog of human elafin. In support of this conclusion, we have recently identified a SPAI analog in porcine trachea that is much similar to human elafin than SPAI, and chemically synthesized it. The chemically synthesized porcine SPAI analog (probably porcine elafin) inhibited pancreatic elastase effectively (IC 50 ϭ 50 nM) but exhibited no Na ϩ ,K ϩ -ATPase inhibitor activity. 2 In the present study, we determined the structure, properties, and tissue distribution of porcine SPAI by cDNA and gene cloning and Northern blot analysis. The interesting features revealed are the presence of the transglutaminase substrate domain in the SPAI sequence, the preferential localization of the SPAI message in the intestine, and the presence of higher molecular weight form of SPAI, which is considered to be the native form, in the circulation. The information may form a valuable basis for elucidating, for example by gene targeting, the physiological significance of SPAI.