A Novel Phospholipase A2 Inhibitor with Leucine-rich Repeats from the Blood Plasma of Agkistrodon blomhoffii siniticus

The phospholipase A2(PLA2) inhibitor PLIβ, purified from the blood plasma of Chinese mamushi snake (Agkistrodon blomhoffii siniticus), is a 160-kDa trimer with three 50-kDa subunits; and it inhibits specifically the enzymatic activity of the basic PLA2 from its own venom (Ohkura, N., Okuhara, H., Inoue, S., Ikeda, K., and Hayashi, K. (1997) Biochem. J. 325, 527–531). In the present study, the 50-kDa subunit was found to be glycosylated withN-linked carbohydrate, and enzymatic deglycosylation decreased the molecular mass of the 50-kDa subunit to 39-kDa. One 160-kDa trimer of PLIβ was found to form a stable complex with three basic PLA2 molecules, indicating that one basic PLA2 molecule would bind stoichiometrically to one subunit of PLIβ. A cDNA encoding PLIβ was isolated from a Chinese mamushi liver cDNA library by use of a probe prepared by a polymerase chain reaction on the basis of the partially determined amino acid sequence of the subunit. The cDNA contained an open reading frame encoding a 23-residue signal sequence followed by a 308-residue protein, which contained the sequences of all the peptides derived by lysyl endopeptidase digestion of the subunit. The molecular mass of the mature protein was calculated to be 34,594 Da, and the deduced amino acid sequence contained four potentialN-glycosylation sites. The sequence of PLIβ showed no significant homology with that of the known PLA2inhibitors. But, interestingly, it exhibited 33% identity with that of human leucine-rich α2-glycoprotein, a serum protein of unknown function. The most striking feature of the sequence is that it contained nine leucine-rich repeats (LRRs), each of 24 amino acid residues and thus encompassing over two-thirds of the molecule. LRRs in PLIβ might be responsible for the specific binding to basic PLA2, since LRRs are considered as the motifs involved in protein-protein interactions.

Phospholipases A 2 (PLA 2 s, 1 EC 3.1.1.4) catalyze the hydrol-ysis of the acyl-ester bond at the sn-2 position of glycerophospholipids to yield fatty acids and lysophospholipids. Snake venom is one of the most abundant sources of secretory PLA 2 s. Secretory PLA 2 s have been classified into three groups, I, II, and III, according to their primary structures (1). Elapidae venom contains group-I PLA 2 s; and Viperidae venom, group-II ones. These snake venom PLA 2 s exhibit a wide variety of toxicity, such as neurotoxicity and myotoxicity (2). Venomous snakes have PLA 2 inhibitory proteins (PLIs) in their blood sera in order to protect them from the leakage of their own venom PLA 2 s into the circulatory system. We have purified three distinct types of PLA 2 inhibitors (PLI␣, PLI␤, and PLI␥) from the blood plasma of the Chinese mamushi, Agkistrodon blomhoffii siniticus (3). PLI␣ was found to be a 75-kDa glycoprotein and a trimer of 20-kDa subunits having sequence homology to the carbohydrate-recognition domain (CRD) of C-type lectins (4). We have already purified this type of inhibitor from the blood plasma of the Habu, Trimeresurus flavoviridis (5,6). These CRD-like inhibitors inhibited specifically the group II acidic PLA 2 s, and thus they seemed to recognize a unique aromatic patch structure on the surface of these acidic PLA 2 molecules (7). On the other hand, PLI␥ was found to be the same type of inhibitor as those purified from the plasma of Naja kaouthia (8) and Crotalus durissus terrificus (9,10). It is a 100-kDa glycoprotein containing 25-and 20-kDa subunits, and their amino acid sequences were characterized by two tandem patterns of cysteine residues found in Ly-6 related proteins (11). This type of inhibitor exhibited a broad inhibition spectrum against PLA 2 s, and it is thought to recognize the Ca 2ϩ -binding loop of PLA 2 s, which is conserved among all the secretory PLA 2 s. PLI␤ was shown to be a 160-kDa glycoprotein having a trimeric structure composed of 50-kDa subunits (3). It inhibited only some group II basic PLA 2 s and did not inhibit other kinds of PLA 2 s. The N-terminal 30-amino acid sequence of the 50-kDa subunit was determined, but no significant homology was found to the known PLA 2 inhibitors.
In the present study, to elucidate the complete primary structure of PLI␤, we cloned and sequenced the PLI␤ cDNA and found that PLI␤ is a protein with leucine-rich repeats (LRR) and that its sequence is 33% identical to that of leucinerich ␣ 2 -glycoprotein (LRG) from human serum.

EXPERIMENTAL PROCEDURES
Materials-The blood plasma, venom, and liver of Chinese mamushi (A. blomhoffii siniticus) were obtained from Ueda Trading Co. (Gifu, Japan). A. blomhoffii siniticus basic PLA 2 was purified from the venom as described previously (4). Chinese mamushi PLI␤ was purified to homogeneity from blood plasma as described previously (3). * This work was supported in part by Grant 07672400 (to S. I.) from the program Grants-in-Aid for Scientific Research (C) of the Ministry of Education, Science, and Culture of Japan. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AB007198.
Deglycosylation and SDS-PAGE of PLI␤-PLI␤ was deglycosylated with an enzymatic deglycosylation kit (Bio-Rad). O-Linked oligosaccharides were cleaved by the combination of O-glycanase and N-acetylneuraminidase at 37°C for 1 h in 50 mM sodium phosphate buffer, pH 6.0. N-Linked oligosaccharides were cleaved by peptide-N-glycosidase F (PNGase F) at 37°C for 3 h in 150 mM sodium phosphate buffer, pH 7.6. PLI␤ used for the PNGase F digestion was denatured by treatment with 0.1% SDS and 50 mM ␤-mercaptoethanol at 100°C for 5 min prior to the digestion.
SDS-PAGE on 11% gels was carried out by the method of Laemmli (12). After the electrophoresis, protein bands were visualized with Coomassie Blue. For the detection of sugars in glycoproteins, proteins on SDS-polyacrylamide gel were transferred electrophoretically to a nitrocellulose membrane. The glycoproteins on the membrane were detected with a DIG glycan detection kit (Boehringer Mannheim). They were oxidized, labeled with digoxigenin, and then detected by an enzyme immunoassay using a digoxigenin-specific antibody conjugated with alkaline phosphatase.
Mass Spectrometry Analysis of Deglycosylated PLI␤-Mass analysis was performed on a Voyager DE STR (PerSeptive Biosystems) matrixassisted laser desorption time of flight (MALDI-TOF) mass spectrometer. The instrument was operated in the linear positive ion mode with an accelerating potential in the source of 20 kV. Ionization was accomplished with the 337-nm beam from a nitrogen laser with the laser energy precisely tuned for an optimal signal to noise ratio. Calibration was done externally with a mixture of two standard compounds (bovine serum albumin and horse heart myoglobin).
An aliquot (2 l) of the HPLC-purified, deglycosylated PLI␤ in 40% acetonitrile was mixed with an equal volume of matrix solution (10 mg/ml of sinapinic acid in 30% acetonitrile containing 0.1% TFA). This mixture was deposited on the metal target, allowed to air-dry, and introduced into the MALDI source.
Chemical Cross-linking-Aliquots (5 l) of the purified PLI␤ (140 g/ml) in 50 mM Hepes buffer (pH 7.5, ionic strength 0.2) were treated with bis(sulfosuccinimidyl)-suberate (Pierce) for 3 h at room temperature. The samples were diluted with the double-strength gel sample buffer containing 0.2 M dithiothreitol, heated to 100°C for 5 min, and analyzed by SDS-PAGE, followed by Coomassie Blue staining.
Direct Binding of Basic PLA 2 to PLI␤-A 0.23-nmol amount of PLI␤ was incubated for 1 h at room temperature with 0, 0.7, and 2.1 nmol of basic PLA 2 in 100 l of 50 mM Hepes buffer (pH 7.5, ionic strength 0.2) containing 0.05% (w/v) Tween 20. The mixture was then applied to a Superose 12 HR10/30 column (Amersham Pharmacia Biotech) equilibrated with the same buffer. The fractions containing the PLI␤⅐PLA 2 complex were further analyzed by reversed-phase HPLC on a Vydac C4 column (The Separations Group).
Amino Acid Sequencing of Peptides Derived from PLI␤-PLI␤ was pyridylethylated in the presence of 6 M guanidine hydrochloride by the method described previously (13). The pyridylethylated PLI␤ was digested with lysyl endopeptidase (Wako Chemicals), and the resultant fragments were separated by HPLC on a Vydac C4 column with 0.1% trifluoroacetic acid containing a linear concentration gradient of acetonitrile from 0 to 48%. The fragments thus obtained were sequenced with an Applied Biosystems model 477A protein sequencer equipped with a 120A PTH analyzer.
Titration of Sulfhydryl Groups-The free thiol group in PLI␤ was quantified in 50 mM Hepes buffer, pH 7.5, containing 0.6 mM 5,5Јdithiobis(2-nitrobenzoic acid) (DTNB) in the absence or presence of 6 M guanidine HCl according to Riddles et al. (14).
Construction of Chinese Mamushi Liver cDNA Library-Total RNA was isolated from fresh liver by use of a total RNA extraction kit (Amersham Pharmacia Biotech) according to the manufacturer's instructions. Poly(A) ϩ -enriched RNA was obtained using an oligo(dT)cellulose column (Amersham Pharmacia Biotech). Oligo(dT)-primed cDNA synthesis was carried out by means of a Superscript choice system (Life Technologies, Inc.). cDNAs were ligated to the MOSElox arms (Amersham Pharmacia Biotech), and packaged via a -in vitro packaging module (Amersham Pharmacia Biotech).
Screening and Isolation of Clones Encoding PLI␤-To obtain the probe for cDNA screening, we performed the reverse transcriptase polymerase chain reaction (RT-PCR) using a Chinese mamushi liver total RNA as template. On the basis of the determined N-terminal and internal peptide sequences of PLI␤, two degenerated oligonucleotide primers, ABSPR5 and ABSPR6, were designed for amplification of a cDNA internal fragment by PCR. The primer ABSPR5 (5Ј-GARAAY-GTNACNGARTTYGT-3Ј, where N, R, and Y denote G/A/T/C, G/A, and T/C, respectively) was oriented in the sense direction and corresponded to the N-terminal residues 11-17 of PLI␤. The primer ABSPR6 (5Ј-ACDATRCAYTGDATNGGRTT-3Ј, where D denotes G/A/T), was oriented in the antisense direction and corresponded to residues 186 -192. The total RNA was subjected to first-strand cDNA synthesis with Superscript reverse transcriptase (Life Technologies, Inc.) using an oligo(dT) primer. The obtained single-strand cDNA was used as a template for PCR amplification with Taq DNA polymerase (Takara) under the following thermal cycling conditions: 5 cycles of 1 min at 94°C, 2 min at 60°C, and 3 min at 72°C, followed by 50 cycles of 1 min at 94°C, 2 min at 55°C, and 3 min at 72°C. The amplified 550-bp product was cloned into pMOSBlue T-vector (Amersham Pharmacia Biotech), cut out, and labeled with horseradish peroxidase by means of an ECL direct nucleic acid labeling system (Amersham Pharmacia Biotech).
Plaques were formed using the Chinese mamushi liver cDNA library, then transferred to a Hybond Nϩ membrane (Amersham Pharmacia Biotech) and screened with the peroxidase-labeled 550-bp fragment as a probe by hybridizing at 42°C overnight in an ECL gold hybridization buffer (Amersham Pharmacia Biotech) containing 5% blocking agent. The filters were washed twice for 20 min with 0.5ϫ SSC (1ϫ SSC ϭ 0.15 M NaCl in 0.015 M sodium citrate) containing 0.4% SDS and 6 M urea at 42°C, and then twice for 5 min with 2ϫ SSC at room temperature. The positive phages were located with ECL detection reagents (Amersham Pharmacia Biotech) according to the manufacturer's instructions. Plasmids were obtained from the isolated phages by in vivo excision.
DNA Sequencing and Analysis-Both strands of DNA were sequenced with a Thermo Sequenase core sequencing kit with 7-deaza-dGTP (Amersham Pharmacia Biotech) on an SQ-3000 DNA sequencer (Hitachi) after subcloning into pBluescript (Stratagene). Analysis of DNA sequencing data and alignments of protein sequences were performed by use of DNASIS (Hitachi Software Engineering Co. Ltd.) and BLAST (15) software on EMBL (16) and GenBank TM (17) data bases.
RNA and DNA Blotting Analyses-A 0.23-g amount of poly(A) ϩ RNA prepared from Chinese mamushi liver was electrophoresed on a 1.2% agarose gel containing 2.2 M formaldehyde and transferred to a Hybond-Nϩ membrane (Amersham Pharmacia Biotech). The biotinylated probes for RNA blotting analysis were prepared from the amplified 550-bp product with a BioPrime DNA labeling kit (Life Technologies, Inc.). After prehybridization, the membrane was incubated with the probes at 42°C overnight in SuperSignal NA hybridization buffer (Pierce), washed twice for 5 min with 2ϫ SSC containing 0.1% SDS at room temperature, twice for 5 min with 0.2ϫ SSC containing 0.1% SDS at room temperature, and then twice for 15 min with 0.2ϫ SSC containing 0.1% SDS at 42°C. After the final wash, the PLI␤ transcript was chemiluminescently detected with a SuperSignal NA complete blotting kit for detection of biotinylated probes (Pierce) according to the manufacturer's instructions.
Genomic DNA was prepared from Chinese mamushi liver by the standard procedure (18). For DNA blotting, DNA was completely digested with various restriction endonucleases, electrophoresed on a 0.7% agarose gel, and transferred to a nylon membrane (Hybond-Nϩ, Amersham Pharmacia Biotech). The membrane was hybridized with the peroxidase-labeled 550-bp fragment at 42°C overnight in an ECL gold hybridization buffer (Amersham Pharmacia Biotech) containing 5% blocking agent, washed twice for 20 min with 0.5ϫ SSC containing 0.4% SDS at 42°C, and then twice for 5 min with 2ϫ SSC at room temperature. The hybridized fragments were chemiluminescently detected with an ECL direct nucleic acid detection system (Amersham Pharmacia Biotech) according to the manufacturer's instructions.

RESULTS
Chemical Characterization of PLI␤-PLI␤ was purified from the blood plasma of Chinese mamushi, A. blomhoffii siniticus, by sequential chromatography on Sephadex G-200, Q-Sepharose, Hi-Trap Blue, Hi-Trap Phenyl HP, and Mono-Q HR5/5 columns, as described previously (3). SDS-PAGE of the purified PLI␤ showed one protein band with an apparent molecular mass of 50 kDa (Fig. 1). When the PLI␤ was treated with N-glycosidase F and a combination of N-acetylneuraminidase and O-glycanase, the apparent molecular mass was reduced to 39 and 47 kDa, respectively. Since PLI␤ treated with N-glycosidase F after the treatment with the combination of N-acetylneuraminidase and O-glycanase gave the same molecular mass of 39 kDa as that treated only with N-glycosidase F, PLI␤ was suggested to be a glycoprotein having N-linked, but not Olinked, carbohydrate. Actually, no sugars in PLI␤ were detected after the N-glycosidase F treatment (data not shown).
Since the molecular mass of 39 kDa, estimated from the migration of the deglycosylated PLI␤ on SDS-PAGE gel, seemed to be different from that of 34,594 Da, which was calculated from the sequence determined in the present study, the molecular mass of the deglycosylated PLI␤ was confirmed by means of mass spectrometry. The observed molecular mass was 34,674 Ϯ 61.8 Da, which was comparable to the calculated value, indicating that the leucine-rich structure of PLI␤ may affect its migration on SDS-PAGE. The molecular mass of 50 kDa obtained from SDS-PAGE of the intact PLI␤ subunit was also determined to be 43,857.2 Ϯ 81.9 Da by the mass spectrometry.
As described previously (3), the molecular mass of PLI␤ was determined to be 160 kDa by gel filtration. Thus the native PLI␤ was considered to be a trimer of 50-kDa subunits. In order to confirm the trimeric structure of PLI␤, chemical crosslinking experiments were performed. As shown in Fig. 2, the subunits of PLI␤ could be cross-linked to form dimer and trimer, but no other higher multimer were formed. Therefore, native PLI␤ was found to be composed of three homogeneous 50-kDa subunits. Table I shows the quantification of free thiol groups of PLI␤. In the presence of denaturating agents, 1.7 mol of free thiol group was detected in one mol of the PLI␤ subunit, while only 0.2 mol was detected in the absence of the denaturating agents. Therefore, two half-cysteine residues of the PLI␤ subunit was found to occur as free thiol groups accessible to the modifying reagent only in the denatured conformation.
Direct Interaction of PLI␤ with Basic PLA 2 -As shown in Fig. 3, complex formation of PLI␤ with A. blomhoffii siniticus basic PLA 2 was investigated by gel filtration on a Superose 12 column. When 0.23 nmol of PLI␤ and 0.7 nmol of PLA 2 were separately applied to the column, they were eluted with a single peak at 28 min and 42 min, corresponding to their respective apparent molecular masses of 160 and 8 kDa. However, when their mixture was applied to the column, the latter peak disappeared, and the height of the former peak increased, suggesting the formation of a stable complex between PLI␤ and basic PLA 2 . The fraction containing the complex of PLI␤ and basic PLA 2 was obtained by the gel filtration of the mixture containing an excess amount of the PLA 2 over PLI␤. Then, PLI␤ and basic PLA 2 in this fraction were separately quantified by reverse-phase HPLC. The molar ratio of PLA 2 to PLI␤ was found to be 3.17, suggesting that each subunit of PLI␤ may bind one molecule of basic PLA 2 .
Cloning of PLI␤ cDNA-Previously, we determined the Nterminal amino acid sequence of PLI␤ (3). In order to elucidate the internal sequences of PLI␤, we digested reduced and Spyridylethylated PLI␤ with lysyl endopeptidase. Of the resulting peptides, fourteen peptides were isolated by reverse-phase HPLC, and their complete or partial amino acid sequences were determined. On the basis of the determined partial amino acid sequences of PLI␤ and its peptides, we synthesized degenerate primers, ABSPR5 and ABSPR6, corresponding to residues 11-17 (ENVTEFV) and 186 -192 (NPIQCIV), respectively. Using these degenerate primers for RT-PCR with mamushi liver total RNA as template, we obtained a cDNA fragment of 550 bp. When this fragment was subcloned and sequenced, the deduced amino acid sequence coincided with that of some lysyl endopeptidase peptides, suggesting that the fragment was a part of PLI␤ cDNA. Thus this fragment was used as a probe to screen the mamushi liver MOSElox cDNA library. From the 2 ϫ 10 6 plaques screened, five positive clones were obtained. Clone pABS205, which contained the largest insert (2.   encoded a protein of 331 amino acids, and its deduced amino acid sequence included the amino acid sequences determined from the lysyl endopeptidase peptides of the purified PLI␤ (Fig.  4, underlined). The first 23 amino acids following the initiator methionine have the features of a signal sequence (19). Valine at the residue 1 was regarded as the N-terminal residue of the mature protein on the basis of the N-terminal sequence determined from the purified PLI␤. Thus the mature PLI␤ was predicted to be composed of 308 amino acid residues with a calculated molecular mass of 34,594 Da, which was consistent with the corresponding value of the deglycosylated PLI␤ determined by mass spectrometry. There were four potential Nlinked glycosylation sites (Asn-X-Ser/Thr) in the deduced sequence (Fig. 4, boldface). One of them, asparagine at residue 12, would actually be glycosylated, since no PTH derivatives of asparagine were detected at the corresponding position on the amino acid sequence study of the purified PLI␤ (3). Two potential polyadenylation signals (AATAAA) occurred at 1503 and 2226 (Fig. 4). Since the latter signal was present 17 bp up-stream of the poly(A) tail, it was used as the actual signal by the mRNA from which this cDNA clone was derived.
RNA and Genomic DNA Blotting Analysis for PLI␤-Northern blot analysis revealed that a single hybridizing band of approximately 2.7 kb was observed on the blot of the poly(A) ϩ RNA isolated from Chinese mamushi liver, indicating that the size of the obtained cDNA is close to that of the full-length copy of mRNA (Fig. 5). Since no other shorter transcripts could be detected, no alternative polyadenylation would occur.
Genomic Southern analysis by use of the fragment corresponding to the coding region (nucleotide 112-656) as a probe showed about four bands of different intensity in each lane of genomic DNA digested with different restriction endonucleases (Fig. 6A). A nearly identical pattern of hybridizing bands was observed, when the fragment corresponding to the 3Ј-noncoding region (nucleotide 1132-2256) was used as a hybridization probe (Fig. 6B). The restriction digest with EcoRI revealed a major hybridizing band of 2.5 kb on probing with the coding fragment as well as on probing with the noncoding fragment; although the recognition sites of EcoRI were present at positions 1102 and 1132 of PLI␤ cDNA, suggesting that the PLI␤ gene was not a single copy and that PLI␤-related sequences formed a multigene family in the snake genome. DISCUSSION Three distinct types of PLA 2 inhibitors (PLI␣, PLI␤, and PLI␥) are present in the blood plasma of the Chinese mamushi, A. blomhoffii siniticus (3). These inhibitors inhibit three group-II PLA 2 s (acidic, neutral, and basic PLA 2 s) from the venom with different specificity. PLI␣, specifically inhibits the acidic PLA 2 by binding of 1 mol of the PLA 2 to 1 mol of the PLI␣ which is a trimer of 20-kDa subunits (4). PLI␥ shows a broad inhibition spectrum; i.e. it inhibited all three venom PLA 2 s almost equally and, furthermore, it inhibited other groups of PLA 2 s, including Elapidae venom PLA 2 s (group I) and honeybee PLA 2 (group III) (3). PLI␤ is a specific inhibitor of the venom basic PLA 2 of A. blomhoffii siniticus, and it does not inhibit the other two PLA 2 s (3). As shown in Fig. 3, 1 mol of the PLI␤, which was composed of three identical subunits of 50-kDa, was found to bind 3 mol of the basic PLA 2 s, suggesting that each subunit would bind one PLA 2 molecule.
In the present study, the complete nucleotide sequence of cDNA encoding PLI␤ was determined. The mature PLI␤ was composed of 308 amino acid residues with a calculated molecular weight of 34,594. This value is consistent with that obtained by mass spectrometry of the deglycosylated PLI␤. PLI␤ was found to be a heavily glycosylated protein having N-linked glycosidic chains. Four potential N-linked glycosylation sites were identified in the deduced sequence.
The amino acid sequence of PLI␤ showed no significant structural similarities to those of PLI␣ and PLI␥. PLI␤ contained neither the CRD sequences, which were found in the sequences of PLI␣ (6) and PLA 2 receptors (20,21), nor the two tandem patterns of cysteine residues found in the sequences of the two homologous subunits of PLI␥ (11) or in Ly-6-related proteins. Searching for homologies by use of the BLAST program revealed that PLI␤ shared 33% identity over the whole molecule with human leucine-rich ␣ 2 -glycoprotein (LRG) (Fig.  7). LRG is a human serum protein in which leucine-rich repeats (LRRs) were first discovered (22). Like LRG, PLI␤ contained 9 tandem LRRs of 24 residues each, which encompassed over two-thirds of the molecule (Fig. 8). An alignment of LRRs of PLI␤ showed that they contained the consensus sequence of X-L-X-X-L-D-L-S-X-N-X-L-X-X-L-X-X-X-X-F-X-X-L-X (X denotes any amino acid). Similar LRR consensus sequences have been found in the primary structure of a large number of proteins, including proteins that participate in biologically important processes, such as receptors for hormone, enzymes, enzyme inhibitors, proteins for cell adhesion, and ribosomebinding proteins (reviewed in Refs. 23 and 24). Although the functions of these proteins are different, all proteins containing LRRs are thought to be involved in protein-protein interactions. The functions of LRG remain unknown, and its physiological ligand remains unidentified. But, both LRG and PLI␤ are serum proteins, and there is 33% identity between their sequences. This may raise the possibility that PLI␤ corresponds to the snake LRG and, furthermore, that human LRG functions as a PLA 2 inhibitor, although evidence on this hypothesis has yet to be demonstrated.
Many LRR-containing proteins contain homologous regions flanking the LRR domains. These regions are characterized by the patterns of cysteine residues; the consensus sequences can be described as CP(ϳ2X)CXC(ϳ6X)C for the amino-flanking and PXXCXC(ϳ20X)C(ϳ20X)C for the carboxy-flanking re-gions (23). The carboxy-flanking region of PLI␤ was nearly identical to the consensus sequence. The amino-flanking region of PLI␤ did not conform to the consensus sequence, but contained a related sequence with two cysteine residues. In addition, there was a proline-rich cluster in the amino-flanking region of PLI␤. This proline-rich cluster is not present in any other LRR-containing proteins, and thus might play an important structural or functional role in PLI␤.
Two cysteine residues, Cys-147 and Cys-190, were found in the LRR domain of PLI␤, while no cysteine residues were found in that of LRG. One of the LRR-containing proteins, ribonuclease inhibitor (RI), contained most of cysteine residues in its LRR domain, and all of the cysteine residues occur as free thiol groups (25). Of the 10 cysteine residues present in the PLI␤ subunit, two were found to occur as free thiols as shown in Table I. Therefore, the 2 cysteine residues with free thiol groups might be assigned to be Cys-147 and Cys-190 found in the LRR domain, because the intrachain disulfide bonding of these cysteine residues could be expected to be unfavorable geometry for the conformation of LRR.
The crystal structure of porcine liver RI has been determined (26). RI was found to be a nonglobular, horseshoe-shaped molecule with a curved parallel ␤-sheet lining the inner circumference of the horseshoe and the helices flanking its outer circumference. In this structure, the individual LRR units have essentially the same conformation of ␤/␣ structural units, consisting of a short ␤-strand and an ␣-helix approximately parallel to each other. The crystal structure of the complex of RI and bovine ribonuclease A (RI-RNase A) has also been determined (27). RNase A binds to a concave surface of the inhibitor, mainly comprising its parallel ␤-sheet and the loops C-terminal to the ␤-strands. Therefore, the structure of RI-RNase A could serve as a model for the interaction between PLI␤ and PLA 2 . If the LRRs in PLI␤ constitute ␤/␣ structural units that occur in RI and the whole molecule reveals a horseshoe structure just like the molecule of RI, the PLA 2 molecule can be expected to bind to the concave surface of the horseshoe structure of PLI␤. That is, the concave surface would contain a large excess of negatively charged residues, especially aspartate at the sixth residue of the consensus sequence of LRR, and these negative charges could electrostatically interact with the positively charged residues on the basic PLA 2 molecule. PLI␤ is a selective inhibitor against the group II basic PLA 2 s from Crotalidae venom, and it does not inhibit other kind of PLA 2 s. The residues His-1, Arg-6, Glu-17, Trp-70, Lys-111, and Ile-124 are conserved among these group II basic PLA 2 s, but they are replaced by other residues in all the PLA 2 s not inhibited by PLI␤. Some of these residues in the PLA 2 molecule are therefore thought to be involved in the interaction with the PLI␤ molecule.
The very extensive binding surface of the RI-RNase A complex was suggested to be in part responsible for the extremely low inhibitory constant (K i ) of 5.9 ϫ 10 Ϫ14 M (28). The inhibitory constant of PLI␤ for A. blomhoffii siniticus basic PLA 2 was 7.5 ϫ 10 Ϫ10 M (3). That this interaction is much weaker than that of RI and RNase suggests that the number of amino acid residues of PLI␤ closely contacting with the PLA 2 molecule is much smaller than that for the RI-RNase complex and that they were located scatteringly and specifically on the extensive concave surface of the horseshoe. The narrow inhibition spectrum of PLI␤ may be accounted for by the local distribution of the contact residues despite the extensive contact area.