Peptidoglycan recognition proteins involved in 1,3-beta-D-glucan-dependent prophenoloxidase activation system of insect.

The prophenoloxidase (proPO) cascade is a major innate immune response in invertebrates, which is triggered into its active form by elicitors, such as lipopolysaccharide, peptidoglycan, and 1,3-beta-D-glucan. A key question of the proPO system is how pattern recognition proteins recognize pathogenic microbes and subsequently activate the system. To investigate the biological function of 1,3-beta-D-glucan pattern recognition protein in the proPO cascade system, we isolated eight different 1,3-beta-D-glucan-binding proteins from the hemolymph of large beetle (Holotrichia diomphalia) larvae by using 1,3-beta-D-glucan immobilized column. Among them, a 20- and 17-kDa protein (referred to as Hd-PGRP-1 and Hd-PGRP-2) show high sequence identity with the short forms of peptidoglycan recognition proteins (PGRPs-S) from human and Drosophila melanogaster. To be able to characterize the biochemical properties of these two proteins, we expressed them in Drosophila S2 cells. Hd-PGRP-1 and Hd-PGRP-2 were found to specifically bind both 1,3-beta-D-glucan and peptidoglycan. By BIAcore analysis, the minimal 1,3-beta-D-glucan structure required for binding to Hd-PGRP-1 was found to be laminaritetraose. Hd-PGRP-1 increased serine protease activity upon binding to 1,3-beta-D-glucan and subsequently induced the phenoloxidase activity in the presence of both 1,3-beta-D-glucan and Ca(2+), but no phenoloxidase activity was elicited under the same conditions in the presence of peptidoglycan and Ca(2+). These results demonstrate that Hd-PGRP-1 can serve as a receptor for 1,3-beta-D-glucan in the insect proPO activation system.

The innate immune system is a host defense mechanism that is evolutionarily conserved from plants to humans and is mainly involved in the recognition and control of the early stage of infection in all animals (1,2). It is activated by a group of germ line encoded receptors, conceptually termed pattern recognition receptors, that recognize microbial surface determinants that are conserved among microbes but absent in the host, such as lipopolysaccharide (LPS), 1 peptidoglycan (PGN), 1,3-␤-D-glucan, and mannan. Upon recognition, these receptors activate distinct signaling cascades that regulate specific immune-related proteins aimed at the aggressors. Recently, our knowledge of innate immunity in mammalian and insects has increased dramatically (1)(2)(3)(4)(5). The recruitment of similar receptors and pathways in both insects and mammals in the fight against infection suggests that they have developed similar mechanisms and molecular pathways to recognize and eliminate invaders (1,6).
Peptidoglycan recognition proteins (PGRPs) have been recognized as an important component of the innate immune system in a variety of organisms ranging from invertebrates to vertebrates including insects and mammals. Ashida and colleagues (7) were first to purify a soluble PGRP showing an affinity for PGN, and upon binding to PGN this complex induced activation of the prophenoloxidase (proPO) cascade. Subsequent identification and cloning of many other PGRPs demonstrated that PGRPs are conserved from insects to mammals (8 -10). However, the exact biological functions of short forms of PGRPs were not determined.
The proPO activation pathway, like the vertebrate complement system, is a proteolytic cascade comprising pattern recognition proteins, several serine proteases, their inhibitors, and terminates with the zymogen, proPO (11)(12)(13). It is known that microbial carbohydrates such as LPS, PGN, or 1,3-D-␤glucan are first recognized by pattern recognition proteins, which then induce activation of serine proteases within the proPO system (11)(12)(13). The proPO-activating enzyme or factor, which all are similar to Drosophila easter-type serine proteases, cleaves proPO to generate the active enzyme, phenoloxidase (PO) (14 -17). This enzyme produces toxic compounds to microorganisms by oxidizing phenols to melanin, and it also participates in the sclerotization of the cuticle, which is vital for the survival of insects (18). Many reports have been published about the proPO of the invertebrate and its activation mechanism (11)(12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24). A key question regarding the proPO activation system is how pattern recognition molecules can induce activation of the system in response to microbial infection.
We were able to reconstitute the downstream components of the proPO cascade using biochemically purified proteins from a large beetle, Holotrichia diomphalia (19). Using this in vitro system, we were able to show that the downstream part of the proPO system is regulated by two easter-type serine proteases and a masquerade-type serine protease homologue and involves a two-step proteolysis of proPOs before they exhibit any PO activity. However, the molecular mechanism of the upstream part of the proPO cascade is still poorly understood. Additionally, it remains to be elucidated how the connection between the upstream and downstream part of the proPO cascade is organized and activated. It was proposed that pattern recognition molecules, such as PGRPs, Gram-negative bacteria-binding proteins, LPS, or 1,3-␤-D-glucan-binding proteins (GRPs), make a complex with the proPO-activating enzyme(s) and microbial cell wall components, then activate proPO-activating enzyme(s), which then will convert proPO to active PO by a limited proteolysis (11-13, 19, 20, 24).
Recently, we demonstrated that a mixture of plasma and hemocyte lysate of the coleopteran insect Tenebrio molitor could induce proPO activation in response to 1,3-␤-D-glucan, but not to LPS and PGN (24). This result suggests that this mixture contains all the necessary components for 1,3-␤-Dglucan-dependent proPO activation, such as 1,3-␤-D-glucan pattern recognition protein(s), proPO-activating factors (PPAFs), and proPOs. In this article we report that Drosophila PGRP-SA-like proteins from the large beetle H. diomphalia recognize 1,3-␤-D-glucan and induce 1,3-␤-D-glucan-dependent proPO activation with the increase of serine protease activity upon binding to 1,3-␤-D-glucan in the presence of Ca 2ϩ . These proteins were also expressed in Drosophila S2 cells, and their biochemical properties were studied.

EXPERIMENTAL PROCEDURES
Animals and Collection of Hemolymph-Methods for rearing the insects and collection of hemolymph were as previously described (25). Hemocytes were collected from the hemolymph by centrifugation at 200 ϫ g for 10 min at 4°C, washed with anti-coagulation buffer (30 mM trisodium citrate, 26 mM citric acid, 20 mM EDTA, and 15 mM sodium chloride, pH 4.6, buffer A), and stored at Ϫ80°C. Approximately 3 ϫ 10 9 packed cells were obtained from 500 ml of hemolymph. The supernatant was taken as plasma, adjusted to pH 4.6 with 1 M citric acid solution, and stored at Ϫ80°C until use. The hemocyte lysate was prepared according to our previously published method (25). Briefly, 5 ml of packed cells were suspended in 10 ml of 50 mM Tris-HCl, 1 mM EDTA, pH 6.5, and then subjected to ultrasonification five times for 3 s at 4°C. The suspended solution was centrifuged at 22,000 ϫ g for 20 min at 4°C. The supernatant was used as hemocyte lysate in further experiments.
Preparation of 1,3-␤-D-Glucan-and PGN-immobilized Columns-A 1,3-␤-D-glucan immobilized column was prepared with a previously published method (26). Briefly, three ml of insoluble 1,3-␤-D-glucan (curdlan, Wako) stock solution (500 mg of curdlan was dissolved in 15 ml of 1 N NaOH solution) was added to 20 ml of 0. 2 M K 2 HPO 4 , and then adjusted to pH 8.4 with 6 M HCl. Two and a half grams of suction-dried AF-Amino Toyopearl 650M was suspended in the soluble curdlan solution prepared with 1 N NaOH. After addition of 1 g of NaCNBH 3 , the suspension was incubated at 60°C overnight. The gel was then acetylated to block remaining free amino groups by incubation with 8 ml of 0.2 M sodium acetate and 4 ml of acetic anhydride on ice for 30 min, followed by adding an additional 4 ml of acetic anhydride, and then incubating 30 min at room temperature. Finally, the resin was washed with 150 ml of 0.1 N NaOH, 150 ml of 1 M Tris-HCl, pH 8.0, and finally with distilled water. As a control, the resin was treated as described above but without curdlan. The amount of coupled 1,3-␤-D-glucans to resins were quantified by the sulfuric acid-phenol method (27). The resins that had coupled more than 3 g of glucose/mg of resin were used for further purification of Holotrichia 1,3-␤-D-glucan-binding protein.
We have made a PGN-immobilized resin by coupling soluble PGN fragments to CNBr-activated Sepharose. The soluble PGN fragments was prepared from Staphylococcus aureus-insoluble PGN (Fluka, catalog no. 77140) according to the published method (28).
Preparation of G-100 Solution and Glucan-specific Solution-G-100 solution from Holotrichia plasma was prepared by our previously published method (29). Briefly, the collected plasma solution (ϳ100 ml containing 900 mg of protein) was concentrated by ultrafiltration through a membrane filter (Amicon, YM1). Approximately 3 ml of the concentrated solution was applied to a Sephadex G-100 column (1.2 ϫ 42 cm) equilibrated with buffer B (50 mM Tris-HCl, containing 20 mM EDTA, pH 6.5) and then eluted with the same buffer at a flow rate of 12 ml/h. Fractions (ϳ12 ml containing 100 mg of protein) showing PO activity in the presence of 5 mM CaCl 2 were pooled and concentrated to 5 ml by ultrafiltration. This solution (referred as G-100 solution) was stored at Ϫ70°C until use. To examine the ␤-1,3-glucan-dependent PO activity, we prepared a solution (referred to as glucan-specific solution) by mixing G-100 solution (4 g of protein) from plasma and the crude hemocyte lysate (4 g of protein).
Assay of PO Activity-An assay of PO was carried out according to our previously published method (14). Briefly, to measure PO activity, 30 l of the glucan-specific solution (300 g of protein) or fractionated G-100 solution (150 g of protein) or hemocyte lysate (150 g of protein) was pre-incubated in 70 l of 20 mM Tris-HCl buffer, pH 8.0, containing 1 g of 1,3-␤-D-glucan for 10 min at 30°C, and then 400 l of substrate solution (1 mM 4-methylcatechol, 2 mM 4-hydroxyproline ethylester in 20 mM Tris-HCl buffer, pH 8.0, containing 5 mM CaCl 2 ) was added to the reaction mixture. After incubation at 30°C for 10 min, the increase in absorbance at 520 nm was measured using a Shimadzu spectrophotometer. One unit of phenoloxidase activity was defined as the amount of enzyme causing an increase in absorbance of 0.1 at 520 nm per 10 min of incubation (A 520 /10 min). Laminarioligosaccharides for examining 1,3-␤-D-glucan-dependent PO activity were purchased from Seikagaku Corp. (Tokyo, Japan).
To examine the effects of the eluate solution from 1,3-␤-D-glucanimmobilized column (referred to as ESGC) on PO activity, 1,3-␤-Dglucan-binding proteins depleted solution was obtained by passing the glucan-specific solution through the 1,3-␤-D-glucan-immobilized column. To determine the effects of the purified recombinant Hd-PGRP-1 and Hd-PGRP-2 on the Holotrichia proPO system, the purified Hd-PGRPs (1 g) was incubated with the glucan-specific solution (150 g of protein), or the eluate solution or the pass-through solution (150 g of protein) from the 1,3-␤-D-glucan-immobilized column at 30°C for 20 min in the presence of Ca 2ϩ and 1,3-␤-D-glucan as described above.
Measurement of the Amidase Activity during ProPO Activation-To determine amidase activity in the samples, commercially available trypsin substrate (t-butyloxycarbonyl-benzyl-L-phenylalanyl-L-seryl-L-arginine 4-methylcoumaryl-7-amide (Boc-Phe-Ser-Arg-MCA)) was used. In our previous studies, we observed that this substrate was mostly hydrolyzed during Holotrichia proPO activation in the presence of 1,3-␤-Dglucan and calcium ion (14,30). This substrate was dissolved in dimethylformamide according to the instruction from the manufacturer. One hundred l of reaction mixture for measuring PO activity was incubated with 490 l of substrate solution, which contains 40 M substrate in 20 mM Tris-HCl buffer, pH 8.0. After incubation of the mixture at 30°C for 1 h, 500 l of 17% (v/v) acetic acid was added to terminate enzyme reaction. The specific amidase activity of the eluate solution can be detected by a fluorescence spectrophotometer at ex ϭ 380 nm and em ϭ 460 nm. As a control, 100 l of buffer A was added to check amidase activity as above. One unit of the amidase activity was defined as the amount that liberated 1 nmol of 7-amino-4-methylcoumarin/min.

Purification of the Denatured Holotrichia 1,3-␤-D-Glucan-binding
Proteins-To purify proteins that can recognize 1,3-␤-D-glucan from the hemolymph of H. diomphalia larvae, the glucan-specific solution was applied to a 1,3-␤-D-glucan-immobilized Toyopearl column (2.5 ϫ 5 cm) equilibrated with buffer B at 0.6 ml/min. After the washing column with buffer B until no absorbance at 280 nm was evident, bound proteins were eluted with buffer B containing 8 M urea at 0.5 ml/min. Then they were analyzed by SDS-PAGE under reducing and non-reducing conditions. The enriched two bands with molecular sizes of 20 and 17 kDa were cut out from the gel and then extracted by electroelution with 250 mA for 3 h at 4°C according to the instructions from the manufacturer. Finally the purity of the two bands was checked by SDS-PAGE under reducing conditions.
To determine the partial amino acid sequences of the purified 20-and 17-kDa proteins, the purified proteins (25 g each) were reduced, alkylated, and digested with 2 g of lysylendopeptidase at 37°C for 13 h. The digested peptides were separated by HPLC on a C 18 reverse phase column (Gilson) with a linear gradient between 0.05% trifluoroacetic acid in water and 0.052% trifluoroacetic acid in 80% acetonitrile (30). The amino-terminal amino acid sequence of the purified proteins and the internal peptides from HPLC were determined on an Applied Biosystem Procise automated gas-phase amino acid sequencer (31).
cDNA Cloning and Nucleotide Sequencing of the 20-and 17-kDa Proteins-A cDNA library from H. diomphalia larvae was constructed as previously described (14) by using a ZAP-cDNA synthesis kit (Stratagene). Three partial amino acid sequences of the purified 20-kDa protein were determined. Among them, an oligonucleotide corresponding to IQNWEDFT was synthesized as follows: 5Ј-AT(T/C/A)CA(A/ G)AAT(T/C)TGGGA(A/G)GGITT(T/C)AC-3Ј, and it was labeled with [␥-32 P]ATP using a previously described method (32). Additionally, three partial amino acid sequences of the purified 17-kDa protein were determined. Among them, an oligonucleotide corresponding to NRWG-GQQA was synthesized as follows: 5Ј-AA(T/C)CG(A/I)TGGGGIGGI-CA(A/G)CA(A/G)GC-3Ј, and it was also labeled with [␥-32 P]ATP. We screened 5 ϫ 10 4 colonies and obtained six hybridization-positive clones. We analyzed two plasmids containing two chemically determined amino acid sequences and amino-terminal sequences. The deduced amino acid sequences of the 20-kDa (Hd-PGRP-1) and 17-kDa (Hd-PGRP-2) were compared with the protein sequence data base of the National Center for Biotechnology Information using the Genetyx system (Software Development Co., Ltd., Tokyo, Japan). 1,3-␤-D-Glucan and PGN Binding Assay-The cultured medium (Invitrogen) of S2 stable cells expressing Hd-PGRP-1 (360 l) or Hd-PGRP-2 (720 l) or the mixture of Hd-PGRP-1 (360 l) and Hd-PGRP-2 (720 l) was mixed with 40 l of 50% (v/v) suspension of the 1,3-␤-Dglucan-or PGN-immobilized resins in 50 mM Tris-HCl, pH 7.0, at 4°C overnight with rocking. After centrifugation, the resins were washed three times with the same buffer. Proteins bound to the 1,3-␤-D-glucanor PGN-immobilized resins were dissolved in 4ϫ SDS-PAGE sample buffer and subjected to SDS-PAGE. For immunoblotting, the proteins were transferred to nitrocellulose membranes overnight at 20 V, 4°C using an electroblot apparatus. The membranes were then treated with the anti-V5 antibody (Invitrogen) and incubated with horseradish peroxidase-conjugated goat anti-mouse IgG and visualized using an ECL kit, according to the protocol provided by the manufacturer (Amersham Biosciences).

Expression and Purification of Recombinant Proteins in Drosophila
BIAcore Analysis-For calculation of association constants (K D ) for Hd-PGRPs, biotinylated laminarioligosaccharide were immobilized on to a streptavidin-coated sensor chip SA (SA chip, BIAcore AB). The biotination reactions of laminarioligosaccharide were performed as described (26). Each solution of laminarioligosaccharides in water was incubated at 90°C for 2 h with Biotin-LC-Hydrazide (Pierce). The reaction mixture was directly injected onto the surface of the SA chip at a flow rate of 2 l/min with BIAcore 1000 (BIAcore AB). Samples in 10 mM HEPES buffer, pH 7.0, containing 150 mM NaCl at a concentration of 100 nM to 1 M were passed over the surface of the sensor chip at a flow rate of 10 l/min. The interaction was monitored as the changes of surface plasmon resonance response at 25°C. After 3 min of monitoring, the same buffer was introduced onto the sensor chip in place of the protein solution to start the dissociation. At the end of each cycle, regeneration of the chip was accomplished by washing away the surface-bound protein with 5 l of 10 mM H 3 PO 4 . Both the association rate constant (k a ) and the dissociation rate constant (k d ) were obtained from the surface plasmon resonance signal binding data and calculated using a program named 1:1 (Langmuir) binding model to obtain constants with the BIA-Evaluation software (version 3.2, BIAcore AB). The association constant (K D ) was subsequently determined by k d /k a .
Antibody and Immunoblotting-Antibodies against the purified Hd-PGRP-1 and Hd-PGRP-2 were raised by injecting 10 g of the purified proteins into a male albino rabbit with complete Freund's adjuvant and twice giving booster injections with the same amount of protein 7 and 14 days later (33). The resulting antibodies were affinity-purified as previously described (14). For immunoblotting, the proteins separated on SDS-PAGE were transferred electrophoretically to a polyvinylidene difluoride filter, and the filter was blocked by immersion in 5% skim milk solution containing 1% horse serum for 12 h. The membrane was then transferred to a rinse solution I (20 mM Tris-HCl, pH 7.5, containing 150 mM NaCl, 0.1% Tween 20, and 2.5% skim milk) containing the affinity-purified antibodies against Hd-PGRP-1 and Hd-PGRP-2 (50 ng/ml) and then incubated at 4°C for 2 h. The bound antibody was identified using an ECL Western blotting reagent kit.
Determination of Hd-PGRP Localization-Hemolymph and fat body were obtained from 5 larvae. The hemolymph was centrifuged separately at 3,000 rpm at 4°C for 10 min. The supernatant was used as plasma. The hemocyte lysate was prepared as described above. The soluble proteins of plasma, hemocyte lysate, and fat body were precipitated with trichloroacetic acid and subjected to SDS-PAGE and then immunoblotting with affinity-purified Hd-PGRP-1 and Hd-PGRP-2 antibodies, respectively.

RESULTS
Relationship between PO Activity and Amidase Activity of 1,3-␤-D-Glucan-dependent ProPO Activation-It is well known that invertebrate proPO system can be activated by LPS, PGN, and 1,3-␤-D-glucan (11)(12)(13). To purify 1,3-␤-D-glucan recognition protein(s) from Holotrichia proPO system, it is necessary to make a solution showing PO activity by 1,3-␤-D-glucan, not by PGN or LPS. As shown in Fig. 1A, we prepared a glucanspecific solution showing 1,3-␤-D-glucan-dependent PO activities using a mixture of G-100 solution from plasma and hemocyte lysate. This solution specifically showed PO activity in the presence of both 1,3-␤-D-glucan and Ca 2ϩ (column 10), but not in the presence of Ca 2ϩ and PGN or LPS (columns 11 and 12). This result suggests that the glucan-specific solution contains all necessary proPO-activating enzymes, proPOs, and unidentified 1,3-␤-D-glucan recognition protein(s).
Previously it was reported that insect proPO system induced the activation of serine protease zymogen to active serine protease during 1,3-␤-D-glucan-dependent proPO activation (11)(12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)30). To explore the existence of serine protease activity during 1,3-␤-D-glucan-specific proPO activation, we examined the amidase activity in the presence of 1,3-␤-D-glucan and Ca 2ϩ by using commercially available trypsin substrate, Boc-Phe-Ser-Arg-MCA. As shown in Fig. 1B, the hemocyte lysate did not show any amidase activities even in the presence of Ca 2ϩ and 1,3-␤-D-glucan (column 8). When Ca 2ϩ ion was added to G-100 solution, amidase activity increased compared with the G-100 alone (column 3). By addition of hemocyte lysate to G-100 solution in the presence of Ca 2ϩ , more amidase activity was increased (column 6). However, when 1,3-␤-D-glucan was added to the mixture of hemocyte lysate and G-100 solution in the presence of Ca 2ϩ , the amidase increased ϳ2-fold compared with that of Ca 2ϩ only (column 10). These results suggest that 1,3-␤-D-glucan might increase the amidase activity by activation of the unidentified serine protease zymogen to active serine protease or that some hemocyte lysate protein(s) recognizing 1,3-␤-D-glucan will enhance the activation of zymogen serine protease. Interestingly, even though the endogenous amidase activities were observed in the presence of PGN or LPS (columns 11 and 12), PGN-or LPS-dependent PO activity was not shown, suggesting that PGN or LPS pattern recognition protein(s) or cofactor (s) will be necessary to show PGN-or LPS-specific PO activity. These factors might be excluded during the preparation of G-100 solution or hemocyte lysate from the crude hemolymph.  Fig. 2A). When the 8 M urea elution solution was analyzed by SDS-PAGE under non-reducing conditions, bands 7 and 8 were found to be mostly enriched (lane 4 in Fig.  2, A and B).
To characterize the biochemical properties of these proteins, the eight eluted proteins were blotted to polyvinylidene difluoride membrane and their amino-terminal sequences or partial amino acid sequences were determined. The amino-terminal sequences of bands 1, 2, 4, and 6 were blocked. The partial amino acid sequences of bands 2 and 4 did not show any homology with known proteins. However, the partial amino acid sequence of band 5 showed an identity with that of Holotrichia proPO-activating factor II that was previously shown to be involved in Holotrichia proPO activation (Ref. 25, data not shown). Interestingly, the amino-terminal sequences of bands 7 and 8 showed high identities with those of Drosophila peptidoglycan recognition protein-SA (Fig. 2C). Bands 7 and 8 had molecular masses of 20 and 17 kDa under reducing conditions, respectively (lane 4 in Fig. 2A), whereas under non-reducing conditions the mass was ϳ16 kDa for both (lane 4 in Fig. 2B). We have named bands 7 and 8 as H. diomphalia peptidoglycan The cDNA sequences of Hd-PGRP-1, -2, and -3 have been submitted to DDBJ as accession numbers AB115774, AB115775, and AB115776, respectively. The two amino-terminal sequences and four partial amino acid sequences of the Hd-PGBP-1 and Hd-PGBP-2 perfectly matched the deduced amino acid sequences in the open reading frames (data no shown). Therefore, we conclude that these are cDNAs for Hd-PGRP-1 and Hd-PGRP-2.
Hd-PGRP-1 and Hd-PGRP-2 Are Localized in Plasma-To examine the localization of Hd-PGRP-1 and Hd-PGRP-2, we prepared fat body, plasma, and hemocyte lysate. As shown in Fig. 4, Hd-PGRP-1 and Hd-PGRP-2 were not detected in the hemocyte lysate or fat body (lanes 2 and 3), whereas a significant amount of protein was detected in the plasma (lane 1), indicating that Hd-PGRPs are localized in the plasma. Silkworm PGRP, Drosophila PGRP-SA and PGRP-SD were constitutively expressed in hemocytes, whereas inducible short PGRPs and PGRP-LB family expressed in the fat body (9,35).

Recombinant Hd-PGRP-1 and Hd-PGRP-2 Proteins and Their Binding Specificity for PGN and 1,3-␤-D-Glucan-To
gain insight into the mechanism of each Hd-PGRP during proPO activation, the recombinant Hd-PGRP-1 and Hd-PGRP-2 were synthesized using the Schneider expression system and S2 insect cells. For easier purification, His 6 tag was added to the carboxyl-terminal end of the proteins. When proteins were purified by nickel particles, 3 mg of each pure protein was obtained from 1 liter of culture medium (Fig. 5A). To confirm the sugar recognition specificity of the recombinant Hd-PGRP-1 and Hd-PGRP-2 for 1,3-␤-D-glucan and PGN, we prepared soluble PGN-coupled Sepharose CL-6B, 1,3-␤-D-glucan-coupled Toyopearl, a control Sepharose CL-6B resin, and a control Toyopearl resin. Hd-PGRPs were detected both in the eluate solution of 1,3-␤-D-glucan-resin and PGN-resin, but not in the control resins (Fig. 5B). Although the binding ability for Hd-PGRPs against glucan is apparently much weaker than for PGN, Hd-PGRPs specifically recognized both PGN and 1,3-␤-D-glucan.
For quantitative analysis of Hd-PGRPs as 1,3-␤-D-glucan pattern recognition proteins, we examined interactions between laminarioligosaccharides and Hd-PGRPs using a BIAcore system. The water-soluble laminarioligosaccharides were first derivatized with biotin and fixed onto the surface of a streptavidin-immobilized sensor chip. When the purified recombinant Hd-PGRP-1 was injected onto the laminaritetraoseand laminaripentaose-immobilized chip, it bound to those chips time-and dose-dependently. However, no specific binding could be detected using laminaribiose and laminaritriose chips (Fig.  6A). Under the same conditions, Hd-PGRP-2 did not bind to the laminaripentaose-immobilized chip (Fig. 6B). The binding parameters were obtained from the sensorgrams with different kinds of laminarioligosaccharides (Table I). The association constant (K D ) of the laminaripentaose with PGRP-I was even higher that that of laminaritetraose. However, K D values of laminaritriose and laminaribiose cannot be determined because their binding to Hd-PGRP-1 is too low. These results support that Hd-PGRP-1 has a specific binding ability against FIG. 4. Localization of Hd-PGRP-1 (A) and -2 (B). The plasma, fat body, and hemocyte lysate were obtained as described under "Experimental Procedures." The proteins were precipitated with trichloroacetic acid and subjected to SDS-PAGE under reducing conditions. Hd-PGRP-1 and Hd-PGRP-2 were detected by immunoblotting with affinity-purified antibodies against Hd-PGRP-1 and Hd-PGRP-2.  (37). The closed circles and closed reverse triangles indicate the conserved cysteine and tryptophan residues, respectively. The closed squares indicate the cysteine residues are present in vertebrate PGRPs. Residues corresponding to the catalytic sites of bacteriophage T7 lysozyme are marked with boxes. Gaps were introduced to obtain maximal sequence similarity.
To obtain further information about minimum structural requirement of pattern recognition molecules to induce proPO activation, we tested the effects of laminarioligosaccharides on 1,3-␤-D-glucan-dependent proPO activation. Laminaribiose and laminaritriose did not activate Holotrichia proPO system, whereas laminaritetraose with 4 glucose residues activated the proPO system (columns 4 -6 in Fig. 7). When Hd-PGRP-1 or the mixture of Hd-PGRP-1 and Hd-PGRP-2 was added to the glucan-specific solution containing laminariheptaose and Ca 2ϩ , PO activity was increased ϳ1.5and ϳ2-fold, respectively (lanes 9 and 11 in Fig. 7). However, Hd-PGRP-2 alone did not affect PO activity under the same conditions (lane 10). These results support that laminaritetraose is enough for 1,3-␤-Dglucan-dependent proPO activation.
To further prove that Hd-PGRPs are 1,3-␤-D-glucan pattern recognition proteins involved in the proPO system, we checked PO activity by using the purified recombinant Hd-PGRPs. When Hd-PGRP-1 (1 g) was added to the glucan-specific solution in the presence of Ca 2ϩ and 1,3-␤-D-glucan, PO activity was increased ϳ1.5 fold after 20 min incubation compared with the control (column 8 in Fig. 8). But, addition of Hd-PGRP-2 did not result in an increase in PO activity under the same conditions (column 9), whereas a mixture of Hd-PGRP-1 and Hd-PGRP-2 increased the PO activity ϳ2.0-fold (column 11). Additionally, without 1,3-␤-D-glucan, the increase of the PO activity was not induced in the presence of Hd-PGRP-1, or with Hd-PGRP-2 (columns 5 and 7). Under the same conditions, the PGN-dependent PO activity in the glucan-specific solution was not influenced by Hd-PGRP-1 or Hd-PGRP-2 (data not shown). The slightly increased PO activity by addition of Hd-PGRPs and soluble PGN might be attributed to a contamination of PGN with lipoteichoic acid. These results taken together suggest that Hd-PGRP-1 recognizes 1,3-␤-D-glucan and then induce Holotrichia 1,3-␤-D-glucan-dependent proPO activation. When both Hd-PGRP-1 and Hd-PGRP-2 were present during the proPO activation, the PO activity was more enhanced compared with Hd-PGRP-1 alone. The main difference between Hd-PGRP-1 and Hd-PGRP-2 is that Hd-PGRP-1 has six additional amino acid residues in the amino-terminal region (Fig.  3), which then suggests that this region in the amino terminus may be involved in correct binding of 1,3-␤-D-glucan to this PGRP and that Hd-PGRP-2 lacking those 6 amino acids has a  low affinity for binding 1,3-␤-D-glucan. The exact mechanism of this difference in PO activity remains to be determined.
To further explore whether Hd-PGRP-1 and Hd-PGRP-2 increase the amidase activity of the glucan-specific solution dur-ing 1,3-␤-D-glucan-dependent proPO activation, we examined the change of amidase activity with trypsin substrate as previously described in Fig. 1B. As shown in Fig. 9, the amidase activity was increased by addition of Hd-PGRP-1, but not Hd- These findings indicate that the recognition signal of Hd-PGRP-1 for 1,3-␤-D-glucan might be transferred to the activation of serine protease zymogen, but Hd-PGRP-2 might involve as a cofactor for enhancing PO activity. The isolation and characterization of this serine protease zymogen remains to be performed. DISCUSSION We report here another function for the short form of PGRPs, which is that it can function as a 1,3-␤-D-glucan pattern recognition protein and induce proPO activation in addition to binding PGN. The analysis of BIAcore experiments clearly supports that the short form Hd-PGRP-1 can bind laminaritetraose. By performing in vitro reconstitution experiments, it was found that Hd-PGRP-1 could induce 1,3-␤-D-glucan-dependent PO activity. Söderhä ll and colleagues (38) reported that the minimum structure for crayfish proPO activation system by laminarin was a laminaripentaose. Although the silkworm PGRP was the first identified protein that bound PGN and activated the proPO cascade in the hemolymph, Hd-PGRP-1,which has high sequence homology with silkworm PGRP, recognizes and binds 1,3-␤-D-glucan and, after binding, elicits activation of the proPO activation system.
Regarding the function of Hd-PGRP-2, there is one possible explanation that a heterodimer of Hd-PGRP-1 and Hd-PGRP-2 induces more efficient 1,3-␤-D-glucan-dependent proPO activation rather than Hd-PGRP-1 alone. Recently, the functional diversity of PGRP-LC family was reported by Werner et al. (39). Alternative splicing of Drosophila PGRP-LC results in that these PGRPs have different affinity to LPS and PGN. The simultaneous requirement of two splice forms (PGRP-LCa and -LCx) for the response to LPS suggests that the PGRPs may act as heterodimer or as higher multimers.
Recent data show that the recognition ligands of the PGRP family such as microbe cell wall components are quite diverse. For example, the long form PGRP-LC and PGRP-LE are absolutely required for the induction of antibacterial peptide genes through Imd/Relish pathway in response to Gram-negative bacteria infection in Drosophila (40,41) or in Drosophila larvae (42). The activation of the Imd/Relish pathway by LPS as well as by Gram-negative bacteria raises the possibility that PGRP-LC recognizes LPS rather than PGN. Further support for this is the fact that PGRP-LC is involved in phagocytosis of Gram-negative bacteria but not Gram-positive bacteria (43). Also, it was reported that Drosophila Toll was activated by Gram-negative bacteria through a circulating PGRP-SA molecule (44). The crystal structure of Drosophila PGRP-LB demonstrates that poor conservation of surface residues at the active sites predicts a widely varying individual specificity of PGRPs for molecular patterns on microbial cell walls (45). Further these data suggest that the PGRP family has a principal role in sensing pathogens and that distinct PGRP molecules recognize different classes of microorganisms.
We isolated Hd-PGRPs by a soluble 1,3-␤-D-glucan-immobilized Toyopearl 650M, a hydrophilic vinylpolymer-based resin. Hd-PGRPs bound tightly to 1,3-␤-D-glucan column resin, but it could be eluted by 8 M urea. Interestingly, Hd-PGRPs were co-purified with six other proteins ( Fig. 2A). The isolation and structural determination of these proteins will give us useful information for elucidation of proPO activation mechanism by 1,3-␤-D-glucan. We reported the molecular relationship between three proPO-activating factors and two proPOs of the large beetle, H. diomphalia larvae. The two proPOs are activated by a two-step proteolysis reaction by two serine proteases and one serine protease homologue in the downstream part of proPO system (19). It is possible that pattern recognition proteins, such as GRP or PGRP, can make a complex with microbial cell components and proPO-activating serine protease (s), and then activated proPO-activating serine protease will convert proPO to active PO by a limited proteolysis, in a manner similar to Drosophila Toll, which is activated through a cleaved form of the cytokine Spatzle by an easter serine protease (46).
Recently, it was reported that Drosophila PGRP-SC1B has a N-acetylmuramyl-L-alanine amidase activity and that it might have a scavenger function (47). Also, Drosophila PGRP-LB has been shown to have bacterial cell wall lytic activity (45). The sequence homology comparison based on these reports suggests that Zn 2ϩ -coordinating amino acid residues of PGRP-LB, such as His-42 and His-152, and Cys-160 and Glu-182, are important for the molecule to exhibit amidase activity (45). Hd-PGRPs and some other short form PGRPs did not show conservation of Cys-160 and His-152 of PGRP-LB. Interestingly, it was reported that invertebrate PGRP and 1,3-␤-D-glucan pattern recognition proteins, which are involved in activation of the proPO cascade ,are non-enzymatic homologues of bacteriophage T7 lysozyme and bacterial 1,3-␤-glucanase, respectively (9, 24, 48 -50). These invertebrate proteins are catalytically inactive, owing to the mutation of critical catalytic residues in the active site of these enzymes. Whether these differences between enzymatic active PGRPs and non-enzymatic homologues have any importance for the recognition of pattern molecules, i.e. affinity constant differences or how to cross-talk with the proPO-activating enzyme(s) remains to be shown.
In the arthropod proPO system, a threshold concentration of calcium is required for proPO activation (51)(52)(53), suggesting that calcium plays an important role in this process. However, the role(s) of calcium is still unclear. In the present study, we confirmed that calcium was required for 1,3-␤-D-glucandependent proPO activation and for amidase activity during proPO activation. Further studies are necessary to determine the exact function of calcium in the proPO system.