Prophenoloxidase-activating Enzyme of the Silkworm, Bombyx mori

Prophenoloxidase-activating enzyme (PPAE) was purified to homogeneity as judged by SDS-polyacrylamide gel electrophoresis from larval cuticles of the silkworm, Bombyx mori. The purified PPAE preparation was shown to be a mixture of the isozymes of PPAE (PPAE-I and PPAE-II), which were eluted at different retention times in reversed-phase high performance liquid chromatography. PPAE-I and PPAE-II seemed to be post translationally modified isozymes and/or allelic variants. Both PPAE isozymes were proteins composed of two polypeptides (heavy and light chains) that are linked by disulfide linkage(s) and glycosylated serine proteases. The results of cDNA cloning, peptide mapping, and amino acid sequencing of PPAE revealed that PPAE is synthesized as prepro-PPAE with 441 amino acid residues and is activated from pro-PPAE by cleavage of a peptide bond between Lys152 and Ile153. The homology search showed 36.9% identity of PPAE to easter, which is a serine protease involved in dorso-ventral pattern formation in theDrosophila embryo, and indicated the presence of two consecutive clip-like domains in the light chain. A single copy of the PPAE gene was suggested to be present in the silkworm genome. In the fifth instar larvae, PPAE transcripts were detected in the integument, hemocytes, and salivary glands but not in the fat body or mid gut. A polypeptide cross-reactive to mono-specific anti-PPAE/IgG was transiently detected in the extract of eggs between 1 and 3 h after they were laid.

and the synthesis of immune proteins with antimicrobial activity effectively engage in defending the insect from the invasion of microorganisms (1)(2)(3). Unlike the well established interrelationships among defense mechanisms and other mechanisms maintaining homeostasis in higher vertebrates, there is a paucity of information concerning such interrelationships in insects.
One of the immune mechanisms in insects is the pro-PO cascade, a serine proteinase cascade involved in the generation of superoxide, melanin synthesis, and subsequent sequestration of foreign matter entering the hemocoel of the insect (3)(4)(5). The cascade is composed of recognition proteins that specifically bind microbial cell wall components such as peptidoglycan or ␤-1,3-glucan, serine protease zymogens, and pro-PO (6 -8). It is present in the hemolymph (9) and in the cuticle of the body wall (10) and is triggered by injury or minute amounts of peptidoglycan and ␤-1,3-glucan (3). Activation of the cascade results in limited proteolysis of pro-PO to the active phenoloxidase (PO). The active enzyme catalyzes two kinds of reactions: the oxygenation of monophenols to o-diphenols and oxidation of o-diphenols to the corresponding quinones necessary for melanin synthesis (11). During melanin synthesis, the various quinones formed are cytotoxic to invading microorganisms and also facilitate the formation of covalently linked coagulum around microorganisms invading the insect and at the injured part (4). As the insect pro-PO has been shown to be a protein homologous to the arthropod hemocyanins, we have previously speculated that the pro-PO may be involved in physiological processes other than defense mechanisms, such as O 2 transport (10).
Another serine proteinase cascade in insects is involved in the establishment of the dorso-ventral pattern in the Drosophila embryo (12). Very recently, Lemaitre et al. (13) have provided evidence demonstrating the links between the dorsoventral pattern formation and immune protein induction. They reported the involvement of components of the dorso-ventral pattern formation pathway, named spä tzle and Toll, in the induction of an antifungal peptide (drosomycin) synthesis in the fruit fly, Drosophila melanogaster, following fungal infection. At the blastoderm stage of the Drosophila embryo, spä tzle is a ligand of the cell surface receptor, Toll, and is produced from pro-spä tzle following a limited proteolysis by the serine protease easter in the perivitelline space. The binding of spä tzle to Toll causes activation of the intracellular signaling pathway for the dorso-ventral pattern formation. These previously reported observations raise the possibility that easter is involved in the extracellular signaling pathways for drosomycin synthesis and the dorso-ventral pattern formation in Drosophila. The parallels between intracellular signaling path-ways for the dorso-ventral pattern formation of the Drosophila embryo and activation of the Drosophila immune protein gene have been known for several years (14). Thus, the links between the dorso-ventral pattern formation and immune protein induction seem to be much closer than they had previously been thought to be.
Before the implication of the Toll pathway, the pro-PO cascade was the only protease cascade known to be present in larval and adult stages of insects. At present, two serine proteases of the insect pro-PO cascade have been purified. One of them, which is tentatively referred to as BAEEase, was purified from the silkworm hemolymph and has been shown to be activated from its zymogen (pro-BAEEase) by limited proteolysis (15). The function of pro-BAEEase remains unknown. The other enzyme, pro-PO-activating enzyme (PPAE), was purified by Dohke 26 years ago (16) from larval cuticles of silkworms. We previously reported that PPAE is also present as an inactive zymogen (pro-PPAE) in the hemolymph and cuticles of silkworm larvae (6,10). Since the first purification by Dohke, PPAE has been purified from several insect species (17)(18)(19)(20). However, characterization of PPAE at the molecular level and elucidation of the activation mechanism of PPAE zymogen remain to be carried out.
Here, we report the purification, characterization, and cDNA cloning of PPAE. PPAE is shown to be a homologous protein to easter and to be synthesized in the epidermis of the integument, the hemocytes, and the salivary glands in the larval stage of the silkworm, Bombyx mori. Preliminary immunochemical study suggests the presence of pro-PPAE in silkworm eggs.

EXPERIMENTAL PROCEDURES
Silkworm (B. mori)-For obtaining cuticles, silkworms at the beginning of the fifth instar were purchased from local sericulture farms. They were reared on mulberry leaves at room temperature (22-28°C) to day 5 or 6 of the fifth instar. Cuticles were collected according to the method of Dohke (16) and stored at Ϫ20°C until use. For experiments other than the purification of PPAE, silkworms were reared on an artificial diet as described previously (6). Nondiapausing eggs laid within 2 h were collected, and all of them were designated to be at 1 h from the oviposition.
Preparation of CH-Sepharose 4B Coupled to m-Aminobenzamidine-4 g of m-aminobenzamidine dihydrochloride (Sigma) was coupled to 15 g of CH-Sepharose 4B by using 5 g of 1-ethyl-3-(3Ј-dimethylaminopropyl) carbodiimide hydrochloride (Sigma) at pH 4.75 according to the manufacturer's instructions. The resulting CH-Sepharose 4B coupled to m-aminobenzamidine (Sepharose-mABA) was subjected to eight alternate washings with 10 mM HCl containing 0.5 M NaCl and with 10 mM NaOH containing 0.5 M NaCl, followed by final washing with 1 liter of distilled water.
Pro-PO and the Assay of PO Activity-Pro-PO was purified from hemolymph of the silkworm, B. mori, as described previously (21), and the amount of one unit of PO was defined according to Ashida (22).
Assay of PPAE-PPAE was assayed using the following three different substrates: (a) A mixture of equal volumes of pro-PO (0.180 mg/ml of 10 mM potassium phosphate buffer, pH 7.5) and PPAE was incubated on ice, and an aliquot of the mixture was assayed for PO activity at 5 and 15 min from the beginning of the incubation using L-DOPA as a substrate, according to the method of Ashida (22). During assay of PO activity, the activation of pro-PO was insignificant because pro-PO and PPAE were diluted 50-fold in the assay mixture, and pH was lowered to 6.0. One unit of PPAE was defined as the amount that produced 200 units of PO for 10 min from 5 to 15 min of the incubation. In the activation reaction, the concentrations of the buffer solution and the PPAE were 0.01 M and below 2.5 units/ml, respectively, and pH was maintained at 7.5. This assay method was used in the course of purification of PPAE. (b) The standard assay mixture for determining esterase activity was composed of 1 mM BAEE, 50 mM Tris-HCl buffer, pH 8.0, at 25°C, 0.5 M KCl and enzyme in a final volume of 1 ml. The reaction was initiated by the addition of enzyme at 25°C and monitored at 254 nm using a Shimadzu spectrophotometer (model 240) with a constant temperature cuvette holder and an accessory unit (OPI-4) for memorizing the monitored data and processing them to output in a form of the differential. In the experiments to determine K m and V max at varied KCl concentrations, the concentrations of BAEE at certain times were read on the records of the time course of the reactions, and the values of the differential of the time courses were read at the corresponding times. Thus, the velocities of the hydrolysis at many different concentrations of BAEE were determined by the analysis of single time course of the reaction with high accuracy. The activity was expressed as mol substrate hydrolyzed/min/mg protein under the above conditions. The concentration of BAEE hydrolyzed was calculated from the molar absorbancy index adopted from Schwert and Takenaka (23). (c) Amidase activity of PPAE was assayed by using various fluorogenic substrates, peptidyl-7-amino-4-methyl coumarins (NH-Mecs, Peptide Institute Inc.). The reaction mixture was composed of 20 mM Tris-HCl buffer, pH 7.5, at 25°C, 0.2 M NaCl, 50 M NH-Mecs, and PPAE solution in a total volume of 0.5 ml. After incubation of the reaction mixture at 25°C for 60 min, the enzyme reaction was terminated by adding 0.5 ml of 50% (by volume) acetic acid. The amount of liberated 7-amino-4 methyl coumarin was determined by the method of Kojima et al. (24).
Preparation of Ammonium Sulfate Fractions of PPAE from Cuticles of Silkworm Larvae-Ammonium sulfate fractions of PPAE were prepared by two methods as described below. All subsequent procedures were performed at 0 -4°C, and centrifugation was carried out at 12,000 ϫ g for 20 min unless otherwise specified. pH values of buffers were those at room temperature (22-26°C). (a) 4 kg of frozen cuticles obtained from about 13,000 larvae were extracted with distilled water, fractionated with ammonium sulfate, and dialyzed according to the method of Dohke (16). The resultant fraction was designated PPAE (AS-dw). The PPAE (AS-dw) was 436 ml in volume and contained 181,314 units of PPAE and 13.42 g of protein. Its specific activity was 13.46 units/mg protein. (b) The residue of cuticles extracted with distilled water in a above was homogenized in a Warning blender (model 34BL21) for 30 s at the maximum speed in 4 liters of 10 mM Tris-HCl buffer, pH 7.1, containing 20 mM CaCl 2 and 0.5 M NaCl. The homogenate was stirred for 2 h, followed by centrifugation. The supernatant was subjected to ammonium sulfate fractionation as in a, and the precipitate that appeared between 0.2 and 0.4 saturation was dissolved in 660 ml of Tris-HCl buffer, pH 8.1, containing 10 mM CaCl 2 and dialyzed against 10 liters of the same buffer for 24 h with two changes of the buffer. The dialyzed solution was centrifuged at 77,000 ϫ g for 30 min, and the supernatant was named PPAE (AS-salt). The PPAE (ASsalt) was 1,145 ml in volume and contained 2,999,900 units of PPAE and 11.7 g of protein. Its specific activity was 256.9 units/mg protein.
Purification of PPAE from PPAE (AS-salt)-Column chromatography of PPAE on DEAE-cellulose (DE23, Whatmann), hydroxyapatite (prepared according to Tiselius et al. (25)), and CM-cellulose (CM52, Whatmann) was carried out by adopting essentially the same methods as those described by Dohke (16). Briefly, PPAE (AS-salt) was applied to a DEAE-cellulose column (7.0 ϫ 55 cm) equilibrated with 10 mM Tris-HCl buffer, pH 8.5. Unabsorbed materials were washed with the same buffer at a flow rate of 200 ml/h. 20-ml fractions were collected from the beginning of the sample application. Fractions containing PPAE activity (numbers 39 -135, DEAE fraction) were pooled. After pH was adjusted to 7.5 with 0.01 N HCl, the pooled fractions were applied to a hydroxyapatite column (3 ϫ 30 cm) equilibrated with 10 mM Tris-HCl buffer, pH 7.5, followed by washing of the column with 300 ml of the buffer. Adsorbed proteins were eluted at a flow rate of 50 ml/h with a linear potassium phosphate gradient from 0.01 M to 0.5 M at pH 7.5 in a total volume of 1 liter. 10-ml fractions were collected from the beginning of the gradient elution. Active fractions (numbers 37-48, first HA fraction) were dialyzed against 5 liters of 10 mM potassium phosphate buffer, pH 7.5, for 12 h. The dialyzed fractions were loaded onto a CM-cellulose column (1.8 ϫ 30 cm) previously equilibrated with the same buffer. After washing the column with 100 ml of the buffer, a linear gradient of KCl (0 -0.4 M) in 600 ml of the buffer was applied to the column at a flow rate of 25 ml/h, and 6-ml fractions were collected from the beginning of the gradient elution. Active fractions (numbers 1-20, CM fraction) were pooled and subjected to further purification. The CM fraction was loaded onto a column (1.5 ϫ 11.5 cm) of Sepharose-mABA after adjustments of the NaCl concentration to 0.5 M and pH to 8.0 by adding 0.66 M NaCl and 1 M Tris solution, respectively. The column was eluted sequentially with solutions of 10 mM Tris-HCl buffer, pH 8.0, containing NaCl at the following concentrations (volume of the solution is indicated in parentheses): 0.5 M (500 ml), 0.1 M (100 ml), 0.05 M (300 ml), and 0 M (150 ml). 6-ml fractions were collected from the beginning of the elution with the buffer containing 0.05 M KCl, and fractions (numbers 4 -52) were pooled (Sepharose-mABA fraction). After pH of the Sepharose-mABA fraction was adjusted to 6.5 by the addition of 0.3 ml of 1 M potassium phosphate buffer and about 33 ml of 0.01 N HCl, the fraction was applied to a hydroxyapatite column (1.7 ϫ 26 cm) equilibrated with 0.01 M potassium phosphate buffer, pH 6.5. Adsorbed proteins were eluted at a flow rate of 20 ml/h with a linear gradient of potassium phosphate buffer, pH 6.5, (0.01-0.3 M) in a total volume of 240 ml. 3-ml fractions were collected from the beginning of the gradient elution. Fractions (numbers [45][46][47][48][49][50][51][52][53][54][55][56][57][58][59][60] corresponding to those eluted between phosphate concentrations of 0.078 and 0.116 M were pooled and used as a purified PPAE preparation (second HA fraction).
Examination of pH Stability of PPAE-One volume of PPAE (1.647 mg protein/ml of 10 mM potassium phosphate buffer, pH 6.37, containing 0.1 M NaCl) was added to nine volumes of buffer solutions with various pH values. In the buffer solutions being used for the dilution of PPAE, the concentration of the buffers was 71.4 mM and the KCl concentration was 714.3 mM. The diluted PPAE mixtures were incubated on ice for 24 h, followed by the addition of 15 volumes of 1 M Tris-HCl buffer, pH 8.0. After the dilution, an aliquot (15 l) of each PPAE solution was assayed for activity by using BAEE.
SDS-PAGE, PAGE under Nondenaturing Conditions, and Isoelectric Focusing PAGE-SDS-PAGE was carried out in a 1-mm-thick slab gel according to the method of Laemmli (26), with 12% acrylamide in the separation gel unless otherwise specified. PAGE under nondenaturing conditions was performed in 7.5% separating gel at pH 9.2 and 6.5 according to the method of Davis (27).
Samples Used in SDS-PAGE and Immunoblotting-The plasma fraction of hemolymph was obtained as described previously (6). Extracts of cuticles were prepared by using 5% SDS and 5 M urea (10) in the presence or absence of 3% 2-mercaptoethanol. Eggs (100 mg) and ovaries including oviducts (300 mg) were homogenized in 300 l of 10 mM potassium phosphate buffer, pH 7.5, containing 1% SDS, followed by centrifugation at 15,000 ϫ g for 30 s. After the supernatants were subjected to centrifugation again under the same conditions, the resulting supernatants were used as the extracts.
Gel Permeation Chromatography of PPAE on Sephacryl S-200 -A Sephacryl S-200 (Amersham Pharmacia Biotech) column (1.08 ϫ 100 cm) was equilibrated with 10 mM Tris-HCl buffer, pH 7.5, containing 0.1 M NaCl. The column was calibrated by chromatography of a mixture of molecular mass standards (Combithek Calibration Proteins II, Roche Diagnostics): bovine ␣-chymotrypsin (25 kDa), hen egg albumin (45 kDa), bovine serum albumin (68 kDa), and aldolase (158 kDa). PPAE (5.87 mg of protein in 4.6 ml of 0.5 M potassium phosphate buffer, pH 7.5) was separately chromatographed on the calibrated column. The proteins were eluted with the same buffer as that used for the equilibration at a flow rate of 20 ml/h, and their elution volumes were measured. Elution profiles of PPAE examined by measuring protein and activity were superimposable and symmetrical. The molecular mass of PPAE was deduced graphically from the observed elution volumes.
Mass Number Analysis by Matrix-assisted Laser Desorption and Ionization (MALDI) Spectrometry-Mass number analyses were carried out according to the method of Peter et al. (29) by using a Kompact MALDI 4 (Shimadzu Corp.) operated in linear time of flight mode. Ionization was accomplished by nitrogen laser at a wavelength of 337 nm for 3 ns. Sinapinic acid was used as a matrix. A saturated matrix solution was prepared in 0.1% trifluoroacetic acid containing 50% acetonitrile (AcCN). After the saturated matrix solution was dried on a sample slide, 1 l of the sample solution was applied on matrix crystals and dried. Ten spectra/sample were examined, and the mean of the observed mass/H ϩ values was calculated. Horse heart myoglobin (M r 16, 950.9) and bovine serum albumin (M r 66, 525) were used as mass number standards.
Examination of Reactivity of PPAE to Lectins-The reactivity was examined using peroxidase-conjugated lectins (8). Briefly, PPAE-I, PPAE-II (5 g protein/lane), and molecular mass markers (Bio-Rad) were run on SDS-PAGE under reducing conditions and transferred to a polyvinylidine difluoride membrane (Immobilon, Millipore). The mem-brane was cut into small pieces with PPAE-I and PPAE-II, and each of the pieces was blotted with one of the peroxidase-conjugated lectins (Peroxidase-conjugated Lectin kit-B, Seikagaku Kogyo) at a concentration of 10 g conjugated lectin/ml. The blots were visualized by incubation with 4-chloro-1-naphtol and H 2 O 2 . After the transfer of PPAE from polyacrylamide gel to the membrane, the gel with residual PPAE was stained with CBB.
S-Pyridylethylation and Separation of the Light and Heavy Chains of PPAE-PPAE-I and PPAE-II were obtained from purified PPAE by reversed-phase octyl (C 8 ) column (Bakerbond, J. T. Baker Inc.) chromatography (pore size, 300 Å; column size, 4.6 ϫ 100 mm) (see Fig. 2 for the chromatography). Each of them (ϳ250 g of protein) were subjected to S-pyridylethylation of cysteines as described previously (21). The S-pyridylethylated PPAE-I and PPAE-II were applied separately onto the C 8  For fluorography, each of the labeled PPAE-I and PPAE-II was dissolved in 53 l of 8 M urea. A half volume of each dissolved solution was run on SDS-PAGE (26) under reducing conditions and the other half under nonreducing conditions. 2-mm-thick and 17% polyacrylamide-separating gel was employed in the SDS-PAGE. After the electrophoresis, the gel was processed for fluorography by using Enlightning (NEN Life Science Products) according to the manufacturer's instructions and exposed to x-ray film (X-OMAT, Kodak) at Ϫ80°C for 1 week before development.
Polyclonal Rabbit Anti-PPAE/IgG-PPAE (320 g) in phosphatebuffered saline was emulsified with an equal volume of Freund's complete adjuvant and injected subcutaneously into a rabbit. Two additional injections were administered at 10-day intervals. Blood was collected at 10 days after the last injection.
Anti-PPAE/IgG from the immunized rabbit serum and nonimmunized IgG from nonimmunized rabbit serum were obtained by protein A-Sepharose 4B (Amersham Pharmacia Biotech) column chromatography according to the manufacturer's instructions.
Immunoblotting-After SDS-PAGE, proteins were transferred to a polyvinylidine difluoride membrane (31). Following the transfer, polypeptides cross-reactive to anti-PPAE/IgG were blotted by the standard method (32). Anti-PPAE/IgG and anti-rabbit goat IgG conjugated to horseradish peroxidase (Santa Cruz Biotechnology) were used at concentrations of 9.6 g/ml and 1 g/ml, respectively. The blots were visualized by incubation with H 2 O 2 and 4-chloro-1-naphtol.
Extraction of RNA-On day 3 of the fifth instar, the silkworms were dissected in a dissecting buffer containing 0.11 M KCl, 4 mM NaCl, 5 mM KH 2 PO 4 , and 4 mM EDTA that had been adjusted to pH 6.5 by KOH. The fat body, mid gut, integument, and silk gland were obtained from a larva, and salivary glands were obtained from 60 larvae. For collection of hemocytes, 60 larvae were bled by cutting one of their abdominal legs, and the discharged hemolymph was collected into 40 ml of the dissecting buffer that was gently being stirred. The collected hemolymph was centrifuged at 1,000 ϫ g for 5 min at 4°C, and the resulting hemocytes were suspended in 40 ml of the dissecting buffer. The suspension-and-centrifugation was repeated twice. The final sediment was used as hemocytes. Total RNA of the tissues was extracted using Isogen (Nippongene) according to the manufacturer's instructions.
Construction of cDNA Library of Larval Integument-The mRNA in the total RNA obtained from larval integument was purified by using the QuickPrep Micro mRNA purification kit (Amersham Pharmacia Biotech). The integument cDNA library was constructed using the Lambda Zap II cDNA Synthesis kit (Stratagene) and Gigapack III Goldpacking Extract (Stratagene). Purification of the mRNA and construction of the cDNA library were performed according to the manufacturer's instructions. ) for 30 cycles, each of which consisted of denaturation at 95°C for 1 min, annealing at 55°C for 2 min, and extension at 72°C for 3 min. The PCR product was digested with EcoRI, purified by agarose gel electrophoresis, and subcloned into EcoRI site of pBluescript II SK(ϩ) (Stratagene). The cloned cDNA (0.5 kbp) was confirmed to be a part of PPAE cDNA by DNA sequencing. The 0.5-kbp cDNA in pBluescript II SK(ϩ) was used as a probe to screen the integument cDNA library.
Screening of the Integument cDNA Library-The cloned 0.5-kbp cDNA was labeled with [␣-32 P]dCTP by using the Ready-To-Go-DNA Labeling kit (Amersham Pharmacia Biotech). About 2.5 ϫ 10 6 clones of the cDNA library were screened at high stringency by using the labeled 0.5-kbp cDNA as a probe (32). Out of several positive plaques, six clones that had different sizes were excised in vivo with helper phage and recirculated to generate subclones in the pBluescript SK. Although all of the clones were confirmed to contain PPAE cDNA, the clone containing the longest cDNA was used as PPAE cDNA in the present study.
Northern Blot Analysis-60-g aliquots of RNA from several tissues were run on electrophoresis in 1% agarose gel and transferred to Hybond N ϩ (Amersham Pharmacia Biotech). The membrane was hybridized with ␣-32 P-random-primed-labeled EcoRI/XhoI digest of PPAE cDNA, which corresponded to 1-629 nt of the clone. Another Hybond N ϩ membrane with the same separated RNAs as above was prepared and hybridized with ␣-32 P-labeled and random-primed ␣-tubulin cDNA as an internal standard for assessing the amount of RNA subjected to the electrophoresis. An ␣-tubulin (1.3 kbp) probe was obtained as a PCR product, and the identity of its sequence to that having been reported was confirmed by referring to the GenBank™ (accession number X83429). Hybridization, washing, and exposure of the blotted membrane to x-ray film were carried out by using standard procedures (32).
Southern Blot Analysis of Genomic DNA-The epidermis was collected from nine silkworms on day 5 of the fifth instar (33). DNA in the epidermis was prepared according to the method of Sambrook et al. (32) with minor modifications. Briefly, the extract of the epidermis was treated with phenol saturated with 10 mM Tris-HCl buffer, pH 8.0, containing 1 mM EDTA after digestion of the extract with RNase and proteinase K and then centrifuged at 1,500 ϫ g for 10 min at 4°C. DNA in the water phase was treated with 80% formamide (v/v) containing 20 mM Tris-HCl buffer, pH 8.0, and 0.8 M NaCl and was dialyzed first against 20 mM Tris-HCl buffer, pH 8.0, containing 10 mM EDTA and 0.1 M NaCl and then against 10 mM Tris-HCl buffer, pH 8.0, containing 10 mM EDTA and 0.01 M NaCl. The dialyzed solution was concentrated by ultrafiltration to 25 ml and used as the genome DNA solution of the epidermis. The genome DNA was treated again with RNase and purified by ethanol precipitation. 10-g aliquots of the purified genome DNA were digested separately with BamHI, EcoRI, HindIII, KpnI, PstI, SacI, SalI, and XbaI (Roche Diagnostics). DNA fragments in the digests were separated by electrophoresis on a 0.8% agarose gel, transferred onto a Hybond N ϩ membrane, and hybridized with the SmaI/HindIII digested ␣-32 P-labeled PPAE cDNA clone, which corresponded to 439 -1070 nt of PPAE cDNA. Hybridization and washing were carried out by the standard method.
Rapid Amplification of cDNA Ends-5Ј-and 3Ј-RACE analyses were performed using RACE systems for rapid amplification of cDNA ends (Life Technologies, Inc.). Based on the integument PPAE cDNA sequence, oligonucleotide primers were designed. The sequences and the names given to the primers are listed in Table I. For 5Ј-RACE, 0.63 g of total RNA from hemocytes was reverse-transcribed with Superscript II (Life Technologies, Inc.) using the primer SP1r. After the cDNA synthesis, the cDNA was incubated with terminal deoxynucleotidyl transferase to add a homopolymeric tail of deoxycytidine residues. The cDNA was then diluted 100 times and used directly for further PCR. The first round of PCR was performed using the Abridged Anchor Primer, SP1r, and Taq polymerase for 35 cycles under the following conditions: denaturation for 1 min at 94°C, annealing for 1 min at 55°C, and extension for 2 min at 72°C. The first PCR product was used for the secondary "nested" PCR using the Universal Amplification Primer and SP2r. The secondary PCR was performed under the same conditions as above except for annealing at 60°C for 1 min. The 3Ј-RACE was performed by employing the same procedure as that for 5Ј-RACE, except that the oligo(dT)-containing adoptor primer, universal adopter primer, and SP6f were used.
RT-PCR-Total RNA (0.63 g) from hemocytes was used for RT-PCR using the Superscript One-step RT-PCR System (Life Technologies, Inc.). The cDNA synthesis was carried out at 45°C for 30 min. The subsequent PCR was performed according to the manufacturer's instructions. Cycling conditions were: 30 s at 94°C, 30 s at 60°C, and 2 min at 72°C. The cycle was repeated 40 times. The oligonucleotide primers for the RT-PCR and PCR are given in Table I. The pairs of primers used were as follows: SP3f and SP1r, SP3f and SP6r, SP4f and SP6r, and Sp5f and SP7r. The PCR products obtained were amplified again by using the same pair of primers by PCR before the analyses of the DNA sequences.
Direct Sequencing of PCR Products-PCR products were electrophoresed and stained with ethidium bromide. PCR products with the expected size were excised from agarose gel, purified by using Genepure (Nippongene), and subjected to DNA sequencing.
DNA Sequencing-DNA sequencing was performed with an auto-

5Ј-AAGAATTCATCGTCCCCAGCCAGCCACG-3Ј
1009-1028 a f, forward; r, reverse. b Location is indicated by nucleotide numbers from the nucleotide at the 5Ј end of integument PPAE cDNA in Fig. 8.
Homology Search-All sequence data were analyzed by GENETYX System, version 8.0 (Software Development Co., LTD Tokyo). A homology search was carried out by using SWISS-PROT (Release 36.0) and PIR (Release 57.0).
Determination of Protein-Protein was determined by the method of Lowry et al. (34) with bovine serum albumin as a standard. Dohke (16) extracted PPAE from the cuticles of silkworms with deionized water and reported purification of the enzyme. We obtained the ammonium sulfate fraction of PPAE (PPAEdw) by adopting his procedures for the extraction and ammonium sulfate fractionation. We re-extracted the residue of the water extraction with buffer containing 0.5 M NaCl and subjected the extract to ammonium sulfate fractionation between 0.2 and 0.4 saturation to obtain a PPAE fraction (PPAE-salt). We found that PPAE-salt contained much more (16.5 times more) PPAE than PPAE-dw and that the specific activity of PPAE-salt was 19 times higher than that of PPAE-dw. PPAE in PPAE-salt was purified by column chromatography on DEAEcellulose, hydroxyapatite, and CM-cellulose according to the method of Dohke. In the chromatography, PPAE appeared to be eluted at the essentially same salt and buffer concentrations as those reported by Dohke (data not shown). PPAE in the CM fraction obtained after CM-cellulose column chromatography was further purified by column chromatography on Sepharose-mABA and hydroxyapatite as is described under "Experimental Procedures." The column chromatography on Sepharose-mABA was very effective for removing contaminated proteins (Table II). In the second column chromatography on hydroxyapatite, elution profiles of protein and PPAE activity were symmetrical and superimposable (Fig. 1), indicating that the purity of PPAE eluted from the column was high. About 20 mg of homogeneous PPAE was obtained, and the preparation was named the second HA fraction (Table II).

Reversed-phase C 8 Column Chromatography of PPAE
Purified PPAE in the second HA fraction was eluted in two peaks in reversed-phase C 8 column chromatography (Fig. 2). Proteins with a shorter retention time and with a longer retention time were named PPAE-I and PPAE-II, respectively. PPAE-I and PPAE-II could be separated in reversed-phase chromatography on columns such as a C 8 column and ODS column but not by any other fractionation methods examined so far, including polyacrylamide gel electrophoresis at pH 9.2 and 6.5 under nondenaturing conditions, iso-electric polyacrylamide gel electrophoresis (data not shown), and ConA-Sepharose column chromatography. 2

Molecular Mass of PPAE
The molecular mass of PPAE under nondenatured conditions was determined to be 45,000 Da in gel permeation chromatography on a Sephacryl S 200 column (data not shown). PPAE (second HA fraction and a mixture of PPAE-I and PPAE-II) migrated to the position corresponding to that of 45,000-Da protein in SDS-PAGE under nonreducing conditions. Under reducing conditions, PPAE gave two bands (light chain and heavy chain) in SDS-PAGE (Fig. 3a). The light chain and heavy chain of PPAE were determined to be 22.0 kDa and 38.5 kDa, respectively, from their mobilities. In a separate experiment, both PPAE-I and PPAE-II were shown to give the same electrophoretic pattern, as is seen in Fig. 3a. These results indicate that both PPAE-I and PPAE-II are composed from light and heavy chains. In the present report, we refer to the light chain and the heavy chain of PPAE-I as PPAE-I-L and PPAE-I-H, respectively, and those of PPAE-II as PPAE-II-L and PPAE-II-H. In MALDI mass spectrometry, mass numbers of PPAE-I and PPAE-II were determined to be 48,811 and 48,821, respectively (Fig. 4).

Iso-electric Point of PPAE
The iso-electric point of PPAE seemed to be higher than 10, because PPAE could not stay in the gel where a pH gradient from 4 to 10 had been established during electrofocusing. Furthermore, PPAE could not enter into the separating gel of pH 9.2 in PAGE under nondenaturing conditions. The isoelectric point of PPAE predicted from its deduced amino acid sequence was 7.82 (35). The reason for such a large discrepancy between the observed and the predicted isoelectric points is not known.

Detection of Carbohydrates of PPAE-I and PPAE-II by SDS-PAGE and Lectin Blotting
None of the lectins used in the present study bound to PPAE-I-L and PPAE-II-L under the experimental conditions (Fig. 5). PPAE-I-H and PPAE-II-H were blotted by concanavalin A, lentil seed agglutinin A, and Dolichos biflorus agglutinin but 2 D. Satoh and M. Ashida, unpublished observations.

N-terminal Amino Acid Sequences of PPAE-I and PPAE-II
PPAE-I and PPAE-II were reduced and pyridylethylated. The light chain and heavy chain of each of S-pyridylethylated PPAE-I and PPAE-II were separated by reversed-phase chromatography on a C 8 column. The S-pyridylethylated PPAE-I-H and PPAE-II-H were analyzed for N-terminal amino acid sequences. The results showed that both of the polypeptides have the same amino acid sequences up to 16th residue from the N termini. They were NH 2 -IVGGAPASIDSYPWLV.
Neither PPAE-I-L nor PPAE-II-L gave N-terminal amino acid after Edman degradation. However, both of the S-pyridylethylated light chains gave the same N-terminal sequences after deblocking with pyroglutamate aminopeptidase. We analyzed the sequences of the deblocked PPAE-I-L and PPAE-II-L up to the 10th residue. Their N-terminal sequences appeared to be pyroglutamyl-SCRTPNGLNG.

Some Enzymatic Properties of PPAE
Because we could not separate PPAE-I and PPAE-II by any means other than reversed-phase chromatography, which caused inactivation of enzymes, we used a mixture of the isoforms to examine the enzymatic properties.
Thermostability-PPAE was stable at 25°C, but above that temperature, inactivation occurred. At 50°C, PPAE lost its activity completely in 5 min (Fig. 6a). pH stability at 0°C for 24 h: PPAE was shown to be stable in a pH range of 6.0 to 8.5.
Below pH 6.0, inactivation occurred, and PPAE lost its activity completely during incubation at pH 3.5. In alkaline pH, PPAE was relatively stable, and 80% of the activity remained even at pH 10.5 (Fig. 6b).
Effect of the Concentration of Salt on Esterase Activity-The esterase activity of PPAE is greatly affected by the salt concentration. The ability of PPAE to hydrolyze BAEE at 4 mM KCl was less than 10% of that at 500 mM KCl (Fig. 7a). KCl was shown to have an effect on K m and V max of PPAE when its activity was assayed by using BAEE as a substrate. At low concentrations (4 -40 mM) of KCl, K m for BAEE was about 1 mM, and at high concentrations, K m decreased to 0.35 mM at 250 mM KCl. The value of V max obtained at 250 mM KCl was about 4.5 times higher than that obtained at 4 mM KCl (Fig. 7b).
Substrate Specificity-Substrate specificity of PPAE was examined by using various fluorogenic substrates, NH-Mecs (Table III). NH-Mecs were required to have Arg at the P1 site to be hydrolyzed efficiently by PPAE. However, Arg at the P1 site was not sufficient for the substrates to be hydrolyzed. The amino acids at P2 and P3 sites also affected the ability of the substrates to be hydrolyzed. Substrate specificity of BAEEase (15), which has been shown to be activated in the silkworm plasma from pro-BAEEase by microbial cell wall components, was different from that of PPAE.

Molecular Cloning of the Integument PPAE cDNA
The six partially overlapping clones were screened from the Lambda ZAPII integument cDNA library. These clones were subcloned into pBluescript SK and subjected to DNA sequence analyses. Of these six clones, the one with the longest insert contained 1447 bp of PPAE cDNA (Fig. 8). The insert included 43 bp of 5Ј-untranslated region, 1323 bp of an open reading frame, a TGA termination codon, 60 bp of 3Ј-untranslated region, and 18 bp of a poly(A) ϩ tail. A poly adenylation signal, AATAAA, was found at 28 bases upstream of the poly(A) ϩ tail. The open reading frame, nt 43-1366, encoded a polypeptide consisted of 441 amino acid residues and terminated in front of the TGA termination codon. The N-terminal sequences of PPAE-I-L and PPAE-II-L were pyroglutamyl-SCRTPNGLNG, as described above. Thus, the mature protein seemed to be from the 22nd glutamic acid to the 420th proline in the deduced amino acid sequence. The molecular mass of the deduced mature protein was calculated to be 45,637. Two potential glycosylation sites for the N-linked carbohydrate chain (Asn-Xaa-(Ser/Thr)) were observed at Asn 219 and Asn 314 in the deduced amino acid sequence. In fact, Asn 219 and Asn 314 could not be detected in the Edman degradation of the peptides containing these residues, indicating that these residues are really glycosylated.

Activation Mechanism of pro-PPAE
From the lysyl endopeptidase digest of the S-pyridylethylated PPAE-I-L, we could isolate a peptide containing the amino acid sequence NH 2 -CCGRDIAVGDK-COOH. This sequence is followed by an N-terminal sequence of PPAE-I-H in the deduced amino acid sequence. We could also isolate a peptide containing the C-terminal sequence of the deduced amino acid sequence in Fig. 8 from the lysyl endopeptidase digests of the S-pyridylethylated PPAE-I-H. These results indicate that pro-PPAE is activated by being hydrolyzed at a peptide bond between Lys 152 and Ieu 153 and that no other peptide bond of pro-PPAE is hydrolyzed in the process of activation of the zymogen.

Tissue-specific Expression of PPAE mRNA
Transcripts of 1.5 kilobases were detected in RNA preparations from the integument, hemocytes and salivary glands of silkworms on day 3 of the fifth instar by Northern blot analysis using a PPAE cDNA probe. The most intense hybridization signal was observed with the RNA preparation from the integument, and the intensity decreased with hemocytes and with salivary glands in this order. PPAE mRNA was not detected in RNA preparations from the fat body, mid gut, and silk gland (Fig. 9).

Southern Blot Analysis
The number of PPAE genes in the genome was estimated by Southern blot analysis (Fig. 10). The probe fragment was a SalI/HindIII digest of PPAE cDNA, corresponding to 439 -1070 nt, and did not have any site for the restriction enzymes used in the present analysis. We found one to three hybridization bands with different mobilities in the digests obtained by treating the genomic DNA with eight restriction enzymes. More than one hybridization band was observed after EcoRI, Hin-dIII, and PstI digestion of the genomic DNA. However, only one hybridization band was detected in the other digests. The three hybridization bands from HindIII digest and the two hybridization bands from EcoRI and PstI digests suggest the presence of more than two introns and more than one intron, respectively, in the region of the genome containing the sequence of the probe fragment.

Hemocyte PPAE cDNA Sequence
The hemocyte PPAE cDNA sequence was determined by RT-PCR and 5Ј-and 3Ј-RACE analyses. The hemocyte cDNA was shown to have a sequence different from that of integument PPAE cDNA at six bases, as stated in the legend to Fig.  8. At the protein level, however, the difference results in only a substitution from Gly 202 to Ala 202 .

Detection of a Polypeptide Cross-reactive to Anti-PPAE/IgG
A polypeptide was detected after SDS-PAGE and immunoblotting of SDS/urea extract of cuticles of silkworm larvae on day 6 of the fifth instar. Under nonreducing conditions, the mobility of the blotted polypeptide was the same as that of purified PPAE. Under reducing conditions, however, the polypeptide in the SDS/urea extract cross-reactive to anti-PPAE/IgG migrated to a position corresponding to the mobility of a 53-kDa protein (Fig. 11). The 53-kDa protein seems to be a zymogen form of PPAE. In plasma, a polypeptide cross-reactive to anti-PPAE/IgG was not detected. Among extracts from ovaries including oviducts and eggs at different developmental stages, the extract prepared from eggs after 2 h from oviposition was shown to contain a polypeptide cross-reactive to anti-PPAE/IgG and with the same mobility in SDS-PAGE as that of cuticular PPAE. A polypeptide cross-reactive to the antibody was not detected in extracts prepared from other developmental stages.

DISCUSSION
Purification of PPAE from larval cuticles of silkworm, B. mori, and some properties of the enzyme were reported by Dohke about 26 years ago (16). Since then, no further characterization of the enzyme has been reported. Since Dohke's work, PPAE has been demonstrated to be a member of a protease cascade (pro-PO cascade), which is now recognized as one of the major defense mechanisms of insects and is speculated to interplay with other defense mechanisms. To understand the mechanism of the interplay, thorough investigation of the properties of each component of the pro-PO cascade is necessary. In the present study, we developed an improved purification procedure to obtain homogeneous PPAE in high yield from larval cuticles of the silkworm, B. mori, and we examined the properties of the enzyme by biochemical, immunochemical, and molecular biological techniques.
The introduction of an extraction method using a buffer with a high salt concentration (0.5 M NaCl) and column chromatography on Sepharose-mABA were extremely effective for obtaining PPAE in high yield and high purity in the present purification from cuticles of the silkworm, B. mori. One of the present After incubation, the remaining PPAE activity was assayed. Activity after incubation on ice is taken as 100%. b, pH stability. PPAE was incubated at various pHs for 24 h on ice, and the activity was assayed. Activity that recovered after incubation at pH 6.75 in phosphate buffer was taken as 100%. Buffers used were: ƒ, glycine-HCl buffer; Ⅺ, acetate buffer; q, citrate buffer; ‚, potassium phosphate buffer; E, Tris-HCl buffer; f, carbonate buffer. Details are described under "Experimental Procedures." authors has previously reported that pro-PPAE in the extract of silkworm cuticles is activated to PPAE if CaCl 2 is present in the extract (10). Because we employed a buffer containing CaCl 2 for extracting PPAE, pro-PPAE in cuticles is likely to be activated during the extraction. We could obtain about 20 mg of PPAE from the cuticles of 13,000 larvae on day 5 or 6 of the fifth instar. In SDS-PAGE, the purified PPAE gave a single band at 45.0 kDa under nonreducing conditions and two bands (heavy chain and light chain) at 38.5 and 22.0 kDa under reducing conditions.
The mass number of PPAE-I and PPAE-II observed in MALDI spectrometry (Fig. 4), and those of PPAE, its light and heavy chains observed in SDS-PAGE (Fig. 3a) were not consistent. Considering the accuracy of MALDI mass spectrometry, PPAE, its light chain, and its heavy chains seem to indicate that they behave anomalously in SDS-PAGE. The reason why the sum of observed mass number of light chain and heavy chains was larger than the one observed for PPAE under nonreducing conditions in SDS-PAGE is not clear. However, a similar phenomenon has been observed on the proclotting enzyme of horseshoe crab (36).
Our results are not consistent with those in Dohke's report (16) in which PPAE was estimated to be a 33-kDa protein in SDS-PAGE under reducing conditions and to be composed of a single polypeptide. As Dohke extracted cuticles with deionized and distilled water, we examined whether our AS fraction-ddw (prepared from the water extract of cuticles) contained the protein cross-reactive to monospecific anti-PPAE/IgG prepared in the present study. By SDS-PAGE and immunoblotting of AS fraction-ddw, we could detect a polypeptide of which the mo-bility was the same as that of the PPAE purified in the present study (data not shown). This result indicates that PPAE extracted from cuticles with deionized and distilled water is the same molecule as PPAE extracted under a high salt concentration. The reason for the discrepancies in molecular mass and subunit structure between Dohke's PPAE and ours remains to be clarified. Both PPAE preparations by Dohke and by us in the present study could activate silkworm pro-PO without the participation of any other factor (16), 3 contrasting the silkworm PPAE with coleopteran PPAE (20), which has been reported to require another proteinaceous factor to convert the respective pro-PO into the active PO. Chosa et al. (19) have also observed that Drosophila PPAE could activate pro-PO by itself by using homogeneous preparations of pro-PO and PPAE.
We examined the following enzymatic properties of purified PPAE: substrate specificity (Table III), effects of KCl on K m and V max in hydrolyzing BAEE, and pH stability and thermostability. In the principal points, the results confirmed and extended what Dohke reported 26 years ago. Among NH-Mecs used as substrates, those with Arg at the P1 site were hydrolyzed. This result is consistent with Dohke's observation that BAEE and tosyl-L-arginine methyl ester were hydrolyzed but tosyl-L-lysine methyl ester was not. PPAEs of the fruit fly (D. melanogaster) and beetle (Holotrichia diomphalia) have been examined for their substrate specificity using NH-Mecs (19,20). PPAEs of the silkworm, fruit fly, and beetle seem to have the common feature of exhibiting a preference for the substrate with Arg at the P1 site. The fruit fly PPAE and silkworm PPAE hydrolyze Boc-Val-Pro-Arg-Mec best among the substrates so far examined, but Boc-Phe-Ser-Arg-Mec has been reported to be the better substrate for the beetle PPAE than Boc-Val-Pro-Arg-Mec. We previously reported that another serine protease of the pro-PO cascade, BAEEase, also hydrolyzes NH-Mecs with Arg at the P1 site (15). However, the preferences of PPAE and BAEEase for NH-Mecs with Arg at P1 site as substrates were different (Table III). We recently showed that PPAE hydrolyzes only the peptide bond between R and F in N 49 R 50 F 51 G 52 sequences of pro-PO polypeptides I and II (37). 3 This observation seems to indicate that PPAE is an enzyme demanding strict requirements to its proteinaceous substrates. K m and V max of PPAE decreased and increased, respectively, in response to the increase in KCl concentration in the reaction mixture when BAEE was used as a substrate. K m was 1.2 mM and 0.35 mM at 4 mM and 500 mM KCl, respectively. V max at 4 mM KCl was about one-fourth of that at 500 mM KCl. Contrary to the observation with BAEE, the ability of PPAE to activate pro-PO was not enhanced under high salt concentrations (16). We have observed that the molecular size of pro-PO does not change after incubation with PPAE in the presence of 0.5 M KCl, 2 indicating that the peptide bond in N 49 R 50 F 51 G 52 sequences of pro-PO is not hydrolyzed under these conditions. It is possible that conformational change of pro-PO due to high salt concentration caused the steric hindrance for PPAE to form an enzyme-substrate complex.
Although the purified PPAE was judged to be homogeneous by SDS-PAGE (Fig. 3a) and elution profiles in the column chromatography on hydroxyapatite (Fig. 1) and Sephacryl S-200, PPAE was resolved into two fractions in HPLC on a C 8 column. PPAE eluted at shorter and longer retention times was named PPAE-I and PPAE-II, respectively. Both PPAEs are composed of a light chain and heavy chain, which are linked by disulfide linkage(s). The difference between the light chains and between the heavy chains could not be detected in such properties as the mobility in SDS-PAGE, N-terminal amino 3 Y. Yasuhara, M. Ochiai, and M. Ashida, manuscript in preparation.

TABLE III
Substrate specificity of PPAE Purified PPAE (0.851 mg/ml of 0.1 M potassium phosphate buffer, pH 6.5) was diluted to 1:200 with 20 mM Tris-HCl buffer, pH 7.5, containing 0.2 M KCl at 25°C. The diluted PPAE solution was used for examining substrate specificity by using various NH-Mec substrates at concentration of 25 M. The amount of 7-amino-4-methyl-coumarin liberated from Boc-Val-Pro-Arg-Mec is taken as 100%. About 1.5% of the substrate was hydrolyzed under the assay conditions employed. See "Experimental Procedures" for details of the assay. Z, benzyloxycarbonyl; Obzl, benzyloxy; Glt, glutaryl; Suc, succinyl; Boc, t-butylcarbonyl. acid sequence, reactivities to lectins, and tritiated DFP. Nterminal amino acids of light chains of PPAE-I and PPAE-II were determined to be pyroglutamic acid. Furthermore, practically the same mass numbers, 48,811 and 48,821 for PPAE-I and PPAE-II, respectively, were observed in MALDI mass spectrometry (Fig. 4). The structural basis for causing the difference in the retention times and mass numbers of PPAEs I and II will be discussed later in this section with the results on cDNA sequencing. Because we could not find any means to The active triad of the serine protease domain is indicated by closed circles. The nucleotide sequence of hemocyte PPAE cDNA was determined by RT-PCR 5Ј-RACE and 3Ј-RACE as described under "Experimental Procedures." The cDNA was found to be longer by the sequence 5Ј-ATTTCGCTCAGAACAGCAACAAAG-TCCGAACCT-3Ј than integument cDNA in the noncoding 5Ј sequence. The sequence was found to be attached to the 5Ј end of integument cDNA. The sequence in the open reading frame of the hemocyte PPAE cDNA was different from that of integument PPAE cDNA at five nucleotide residues. The nucleotide residues A 222 , C 250 , T 344 , G 711 , A 712 , and G 928 of integument PPAE cDNA are replaced by G, T, C, C, G, and A, respectively, in hemocyte PPAE cDNA. At the amino acid level, the deduced hemocyte pro-PPAE amino acid sequence was the same as the integument pro-PPAE sequence except that Gly 202 of cuticle pro-PPAE was replaced by Ala.  (1)(2)(3)(4)(5)(6) linkage to N-linked N-acetylglucosamine. The mass numbers of S-pyridylethylated PPAE-I-H and PPAE-II-H observed in MALDI mass spectrometry were larger by 2,554 and 2,540, respectively, than the mass number calculated from the deduced amino sequence of the heavy chain, of which cysteine residues were assumed to be S-pyridylethylated. The difference between the observed and calculated mass numbers could account for two N-linked carbohydrate chains and is thus consistent with the fact that the residues of Asn 218 and Asn 313 seemed to be linked to carbohydrate chains.
The longest cDNA of PPAE screened from the cDNA library of body wall epidermis of a larva on day 3 of the fifth instar, was 1,447 bp. It encoded a single open reading frame consisting of a 21-amino acid prepropeptide followed by a 420-amino acid mature protein. We obtained results indicating that the mature protein (pro-PPAE) with pyroglutamyl N-terminal amino acid residue is activated to PPAE by cleavage of the peptide bond between Lys 152 and Ile 153 (Fig. 8). By the cleavage of the peptide bond, pro-PPAE seemed to be converted to an active serine protease with a light chain and heavy chain consisting of 152 and 268 amino acid residues, respectively, that are linked by disulfide linkage(s). The 268-amino acid heavy chain contains a typical catalytic domain of trypsin-type serine protease with active site triads, His 187 , Asp 264 , and Ser 370 . Labeling of the heavy chain with tritiated DFP (Fig. 3b) supports the notion that both triads in PPAE-I and PPAE-II function as an active site.
The hemocyte PPAE cDNA sequence was determined by RT-PCR and 5Ј-and 3Ј-RACE analyses. Although the cDNA was longer by 34 bp than the integument PPAE cDNA at the 5Ј upstream region, sequences of an open reading frame and 3Ј-untranslated region were shown to be the same as the integument PPAE cDNA, except that substitutions of a base at six nucleotide residues were observed in the open reading frame. Among six substitutions, only the substitution of guanine 710 to cytosine 710 caused the substitution of an amino acid from Gly 202 to Ala 202 . Considering the technique used for the determination of the hemocyte PPAE cDNA sequence, we could not rule out the possibility that the observed sequence includes artifacts. Therefore, it is not clear whether the gene coding for the PPAE variant with Ala 202 really exists in the silkworm strain used in the present study. Because only a single copy of the PPAE gene was indicated to be present in the silkworm genome (Fig. 10), PPAE variants, even if they exist, must be allelic variants. It is possible to explain the presence of PPAE-I and PPAE-II as the result of the expression of allelic genes that code for the PPAEs with Ala 202 and Gly 202 . In Northern blot analyses, PPAE transcripts were detected in total RNA preparations from the epidermis, hemocytes, and salivary glands but not in those from the fat body and mid gut (Fig. 9). Because PPAE is synthesized as prepro-PPAE, it is likely that the molecule is secreted from the cell where it is synthesized. It has not been elucidated, however, whether prepro-PPAE synthesized in the epidermis is secreted in both directions, to the cuticle and to hemocoel, or only to the cuticle. Furthermore, the type of hemocyte engaging in the synthesis of the prepro-PPAE remains to be studied. The implication of the presence of PPAE transcripts in the salivary gland is not clear at present. By Western blotting, we examined whether the same molecule in terms of mobility in SDS-PAGE is present in both the cuticle and plasma (Fig. 11). In the extract of cuticle with SDS and urea, only protein with mobility corresponding to 53 kDa under reducing conditions and 45 kDa under nonreducing conditions was shown to be cross-reactive to anti-PPAE/ IgG. The mass number, 45 kDa, was also observed for the purified PPAE in SDS-PAGE under nonreducing conditions. The above results seem to show that the present anti-PPAE/ IgG is mono-specific to PPAE and the zymogen form of PPAE. We observed that the purified PPAE zymogen is a protein with mobilities corresponding to 45 and 53 kDa under nonreducing conditions and reducing conditions, respectively. 2 When plasma was examined for the presence of pro-PPAE by Western blot analysis, no cross-reactive band was detectable (Fig. 11, b  and c).
Previously, we could detect PPAE activity in plasma treated with a microbial cell wall component, and we also obtained results suggesting that PPAE in plasma cleaved pro-PO subunits at the same peptide bonds as cuticular PPAE (38). Furthermore, we detected the same PPAE transcripts in hemocytes as those in the body wall epidermis in the present study. Therefore, the reason why we could not detect PPAE (or pro-PPAE) in plasma seems to be the low concentration of the zymogen, which was under the limit of detection by the present method of Western blot analysis.
We searched for proteins with similarity to the PPAE sequence. The proteins with similarity (the degree of their similarity is indicated in parentheses) were as follows: Drosophila serine protease zymogens, easter (35.4%) (39) and snake (25.3%) (40), and horseshoe crab (Tachypleus tridentus) serine protease zymogens, proclotting enzyme (27.4%), and Factor B (26.8%). These zymogens are known as the components of invertebrate protease cascades such as the dorso-ventral pathway in the perivitelline space of the developing Drosophila embryo and the blood coagulation cascade in the hemolymph of the horseshoe crab. The above proteins were demonstrated, or speculated, to have a so-called "clip-like domain" or "disulfide knot domain" (41,43) in which six cysteine residues form three disulfide linkages. As is shown in Fig. 12, the light chain of PPAE seemed to contain two consecutive clip-like domains. To our knowledge, a protease zymogen with two clip-like domains has not been reported before in the literature, although protease-like molecules, masquarde (44) and Anopheles gambiae ISLP5 (45), have been shown to have two consecutive clip-like domains. The former molecule is required for the morphogenesis of Drosophila muscle but has not been shown to have protease activity. The function of latter molecule is still not known. As the clip-like domain is also present in ␤-defensins, an antimicrobial peptide in mammals (46) and in horseshoe crabs (47), we speculated that pro-PPAE or PPAE might have antimicrobial activity. We failed, however, in demonstrating the anti-microbial activity of the silkworm PPAE in the preliminary experiments, where Escherichia coli, Bacillus licheniformis, and Micrococcus luteus were incubated with 1.76 M PPAE. 2 The function of the clip-like domain remains to be studied for PPAE as well as for other serine proteases.
The light chain of PPAE has a sequence with three consecutive threonine residues, Thr 116 -Thr 117 -Thr 118 . The same sequence has been reported in the light chains of T. tridentus proclotting enzyme (41) and Factor B (42). All of the residues have been indicated to be O-glycosylated. MALDI mass spectrometry of S-pyridylethylated light chains of PPAE-I and PPAE-II showed that their mass numbers are larger by 247 and 419, respectively, than those calculated from the deduced amino acid sequence in which cysteine residues were assumed to be S-pyridylethylated. This difference in mass numbers seems significant if we consider the accuracy of the present MALDI mass spectrometry. Thus, the present results do not exclude the possibility of the presence of O-glycosylated threonine residue(s) in the PPAE light chain. We noticed that the PPAE light chain showed anomalous behavior: lower mobility in SDS-PAGE than expected and less stainability with CBB. The same anomalous properties have been observed for the light chain of proclotting enzyme (47).
The pro-PO cascade (pro-PO activating system) of the silkworm hemolymph has at least one reaction step where Ca 2ϩ ions are required (48). The ␥-carboxyglutamic acid domain is known to be present in some proteases and to bind Ca 2ϩ ions for the activation of proteases in the mammalian blood coagulation system (49). Therefore, one can expect to find a ␥-carboxyglutamic acid domain in one of the proteases in the pro-PO cascade. However, such a domain was not detected in the deduced amino acid sequence of pro-PPAE. The absence is consistent with our previous observation that PPAE does not require Ca 2ϩ ions to activate pro-PO. In the cascade of the dorso-ventral pathway of Drosophila, a protease, snake, which converts pro-easter to easter, was shown to have a ␥-carboxyglutamic acid domain (40). Future studies may show that protease(s) upstream of PPAE in the pro-PO cascade have such a domain.
Catalytic subunits of serine proteases have conserved sequences especially around amino acids participating in the catalytic action and the formation of disulfide linkages (50). The sequence of the heavy chain of PPAE was found to be similar to the catalytic subunits of the following proteases (the degree of similarity in parentheses): prothrombin (30.8%) (51), factor IX (32.3%) (52), factor XI (33.7%) (53), human plasma kallikrein (35.2%) (54), and Limulus proclotting enzyme (29.7%) (41). From the alignment of these sequences, we learned that disulfide linkages are likely to be formed between cysteine residues of PPAE as follows: Cys 183 -Cys 199 , Cys 355 -Cys 366 , and Cys 366 -Cys 395 . Furthermore, in the light chain of PPAE, one of cysteine residues, except those being engaged in the formation of two clip-like domains, seemed to be linked to Cys 284 of the heavy chain by a disulfide linkage.
Before the present study, insect proteases whose functions are known and that have a clip-like domain have been reported only from the protease cascade in the dorso-ventral pathway, which operates in perivitelline space at the early embryonic The amino acid sequences of protease zymogens previously known to have a clip-like domain were extracted from the PIR data base, and multiple alignment was performed with the PLASTP (62). Numbers in parentheses indicate the amino acid residues from N termini of each protease. Identical residues found in at least four aligned sequences are boxed. Cysteines are in shaded boxes. Disulfide bonds indicated by solid lines are those found in the horseshoe crab proclotting enzyme (41). Gaps are indicated by hyphens. CLD, clip-like domain. stage of D. melanogaster. Easter and snake are such proteases. Snake activates easter, which, in turn, converts prospä tzle to spä tzle. As was mentioned in the Introduction, the protease cascade with easter or a protease with a substrate specificity similar to that of easter have been shown to be involved in the extracellular signaling pathway for the ventralization of blastdermal cells and for the induction of antifungal peptide synthesis of the adult Drosophila fat body. To date, the pro-PO cascade is the only protease cascade that could be activated by one of fungal cell wall components, ␤-1,3-glucan, and is present in both larval and adult stages of insects. In the present study, pro-PPAE of the pro-PO cascade was shown to be homologous to easter. Furthermore, preliminary immunochemical study showed that silkworm eggs at 2 h after being laid have a polypeptide cross-reactive to the mono-specific anti-PPAE/IgG and that the mobility of the polypeptide is about the same as that of purified PPAE (Fig. 11). Our present observations that have been reported so far about the dorso-ventral pathway, the extracellular signaling pathway for activation of the drosomycin gene, and the indication of the presence of pro-PO cascade in silkworm eggs (55) imply that the pro-PO cascade may operate not only for the synthesis of melanin from phenolic compounds but also in developmental events in the early embryonic stage and in the signaling pathway for activation of immune protein gene(s).
After submission and during the review of the present paper, two papers dealing with prophenoloxidase-activating enzyme appeared (56,57). The enzymes from lepidopteran (Manduca sexta) and coleopteran (H. diomphalia) were reported to be serine proteases as ours but require another proteinaceous factor to make them competent to activate pro-PO and have one clip-like domain. From the reported sequences, the similarity of the silkworm PPAE to the Manduca PPAE and Holotrichia PPAE is calculated to be 34.7 and 35.0%, respectively. The Holotrichia PPAE has been claimed to hydrolyze peptide bond at two sites in each of pro-PO subunits when it works with auxiliary proteinaceous component (20). As described in the present paper and as reported before (6, 10, 16, 21, 37, 58 -60), the silkworm PPAE is different in these properties. Thus, there seem to be two kinds of serine-type proteinase among insects that are directly involved in the activation of pro-PO. It remains to be studied whether the two kinds of PPAE are present in each of insect species and, if so, what the physiological significance of their presence is. Apparently, more thorough investigations of the pro-PO cascade components are needed to answer the questions surrounding the pro-PO cascade.