J Mol Evol (2003) 57:546–554 DOI: 10.1007/s00239-003-2508-4

Interisland Mutation of a Novel Phospholipase A2 from Trimeresurus flavoviridis Venom and Evolution of Crotalinae Group II Phospholipases A2 Takahito Chijiwa,1 Sachiko Hamai,2 Shoji Tsubouchi,3 Tomohisa Ogawa,3 Masanobu Deshimaru,3 Naoko Oda-Ueda,1 Shosaku Hattori,4 Hiroshi Kihara,2 Susumu Tsunasawa,2 Motonori Ohno1 1 2 3 4

Department of Applied Life Science, Faculty of Engineering, Sojo University, 4-22-1 Ikeda, Kumamoto 860-0082, Japan Protein Analysis Center, Takara Shuzo, Kusatsu, Shiga 525-0055, Japan Department of Chemistry, Faculty of Science, Kyushu University, Higashi-ku, Fukuoka 812-8581, Japan Institute of Medical Science, University of Tokyo, Oshima-gun, Kagoshima 894-1531, Japan

Received: 19 February 2003 / Accepted: 20 May 2003

Abstract. Trimeresurus flavoviridis (Crotalinae) snakes inhabit the southwestern islands of Japan: Amami-Oshima, Tokunoshima, and Okinawa. Affinity and conventional chromatographies of AmamiOshima T. flavoviridis venom led to isolation of a novel phospholipase A2 (PLA2). This protein was highly homologous (91%) in sequence to trimucrotoxin, a neurotoxic PLA2, which had been isolated from T. mucrosquamatus (Taiwan) venom, and exhibited weak neurotoxicity. This protein was named PLA-N. Its LD50 for mice was 1.34 lg/g, which is comparable to that of trimucrotoxin. The cDNA encoding PLA-N was isolated from both the AmamiOshima and the Tokunoshima T. flavoviridis venomgland cDNA libraries. Screening of the Okinawa T. flavoviridis venom-gland cDNA library with PLA-N cDNA led to isolation of the cDNA encoding one amino acid-substituted PLA-N homologue, named PLA-N(O), suggesting that interisland mutation occurred and that Okinawa island was separated from a former island prior to dissociation of Amami-Oshima and Tokunoshima islands. Construction of a phylogenetic tree of Crotalinae venom group II PLA2’s

based on the amino acid sequences revealed that neurotoxic PLA2’s including PLA-N and PLA-N(O) form an independent cluster which is distant from other PLA2 groups such as PLA2 type, basic [Asp49]PLA2 type, and [Lys49]PLA2 type. Comparison of the nucleotide sequence of PLA-N cDNA with those of the cDNAs encoding other T. flavoviridis venom PLA2’s showed that they have evolved in an accelerated manner. However, when comparison was made within the cDNAs encoding Crotalinae venom neurotoxic PLA2’s, their evolutionary rates appear to be reduced to a level between accelerated evolution and neutral evolution. It is likely that ancestral genes of neurotoxic PLA2’s evolved in an accelerated manner until they had acquired neurotoxic function and since then they have evolved with less frequent mutation, possibly for functional conservation. Key words: Interisland mutation — Trimeresurus flavoviridis — Phospholipase A2 — Neurotoxicity — Amino acid sequence — cDNA — Accelerated evolution — Phylogeny

Introduction The nucleotide sequences reported in this paper are available from the GenBank/EMBL/DDBJ databases under accession numbers AB102728 and AB102729. Correspondence to: Motonori Ohno; email: [email protected]. ac.jp

Phospholipase A2 (PLA2; EC 3.1.1.4) catalyzes the hydrolysis of the 2-acyl ester linkage of 3-sn-phosphoglycerides with the requirement of Ca2+ to produce 3-sn-lysophosphoglycerides and fatty acids

547

(Dijkstra et al. 1981, 1983). Snake venoms contain PLA2 isoforms as major toxic components. Snake venom PLA2’s are classified into groups I and II based on the mode of disulfide pairings (Dufton and Hider 1983). Group I PLA2’s are found in Elapidae (Elapinae and Hydrophiinae) venoms, whereas group II PLA2’s are found in Viperidae (Viperinae and Crotalinae) venoms. Group II PLA2’s are divided into two subgroups, [Asp49]PLA2 forms and [Lys49]PLA2 forms (Maraganore et al. 1984; Maraganore and Heinrikson 1986). They share the same scaffold (Brunie et al. 1985; Renetseder et al. 1985; Holland et al. 1990; Suzuki et al. 1995). Trimeresurus flavoviridis (Crotalinae) snakes inhabit the southwestern islands of Japan: AmamiOshima, Tokunoshima, and Okinawa. AmamiOshima island is the northernmost and Tokunoshima island is 30 km south of Amami-Oshima island. Okinawa island is located a further 120 km south of Tokunoshima island. Four PLA2 isozymes were isolated from Tokunoshima T. flavoviridis venom. They all consist of 122 amino acid residues and have been well studied: [Asp49]PLA2, called PLA2 (pI 7.9, highly lipolytic and myolytic) (Oda et al. 1990; Kihara et al. 1992); more basic [Asp49]PLA2, called PLA-B (pI 8.6, edema-inducing) (Yamaguchi et al. 2001); and two [Lys49]PLA2’s called BPI and BPII (pI’s 10.2 and 10.3, both extremely weakly lipolytic and strongly myolytic) (Yoshizumi et al. 1990; Liu et al. 1990; Kihara et al. 1992). The cDNAs and genes encoding PLA2, BPI, and BPII have been cloned (Oda et al. 1990; Ogawa et al. 1992; Nakashima et al. 1993, 1995). Two cDNAs encoding PL-X0 and [Thr38]PLA2, both [Asp49] PLA2’s, have also been cloned (Ogawa et al. 1992), although their proteins have not been isolated. Comparison of their nucleotide sequences led to a novel discovery that Darwinian-type accelerated evolution has occurred to acquire diverse physiological activities (Ogawa et al. 1992; Nakashima et al. 1993, 1995; Ohno et al. 1998, 2002). Thus, the isolation of novel PLA2 isozymes from the venoms of T. flavoviridis snakes from the three islands is of particular interest in terms of inquiry into their molecular evolution. In the work reported here, we isolated a novel neurotoxic PLA2, named PLA-N, from AmamiOshima T. flavoviridis venom by affinity chromatography on a T. flavoviridis serum undecapeptide (DVEDHPWTISL)-immobilized column as well as by conventional chromatography. Its amino acid sequence was determined. The cDNA encoding PLAN was isolated from both Amami-Oshima and Tokunoshima T. flavoviridis venom-gland cDNA libraries. Screening of Okinawa T. flavoviridis venomgland cDNA library with PLA-N cDNA led to isolation of the cDNA coding for a PLA-N homologue, named PLA-N(O). Interisland mutation of T. flavoviridis venom neurotoxic PLA2’s and molec-

ular evolution of Crotalinae venom group II PLA2’s are discussed. Materials and Methods Materials The venom and blood were collected from various individuals of Amami-Oshima T. flavoviridis. The venom glands were excised from a specimen of Amami-Oshima T. flavoviridis. Achromobacter protease I, endoprotease Asp-N, and pronase were obtained from Wako Pure Chemical Industries (Osaka, Japan), Hoffman-LaRoche (Basel, Switzerland), and Nacalai Tesque (Kyoto, Japan), respectively. Restriction endonucleases and Taq DNA polymerase were from Takara Shuzo (Kyoto) and Pharmacia Biotech (Uppsala, Sweden), respectively. [a-32P]dCTP (3000 Ci/mmol) and [a-35S]dATP (1000 Ci/mmol) were from Amersham (Buckinghamshire, UK).

Fractionation of an Undecapeptide from T. flavoviridis Serum Amami-Oshima T. flavoviridis serum diluted three fold with 1.0 M acetic acid–20 mM HCl was ultrafiltrated with an MWCO 500 membrane (Spectrum Laboratories, Rancho Dominguez, CA) against water and an inner solution was then ultrafiltrated with an MWCO 15,000 membrane against water. An outer solution was concentrated under reduced pressure to a small volume and solid sodium bicarbonate was added to 0.1%. The solution was subjected to high-performance liquid chromatography (HPLC) on a Cosmosil 5C18-AR column (0.46 · 15 cm) (Nacalai Tesque) before and after pronase digestion (0.01% w/w, pH 8.0, 37C, 15 h). Four major peaks (A, B, C, and D) disappeared by pronase digestion (data not shown), indicating that they are polypeptides. Peptide A (molecular weight 1311.7 measured by mass spectroscopy) was sequenced to be DVEDHPWTISL.

Purification of PLA-N from Amami-Oshima T. flavoviridis Venom Affinity Chromatography on a Peptide A-Immobilized Column. The serum peptide A (30 mg) synthesized by the solid-phase method was immobilized on POROS-EP (PerSeptive Biosystems, Framingham, MA; 1.0 g). Otherwise, Amami-Oshima T. flavoviridis venom was fractionated on a Sephadex G-100 column to give three peaks (Ishimaru et al. 1980). The proteins (1.5 mg) from the second peak which exhibited PLA2 activity was applied onto a peptide A-immobilized column (0.4 · 10 cm), washed with 0.05 M ammonium acetate (pH 6.8) containing 0.5 mM NaCl, and eluted with 0.1% trifluoroacetic acid (TFA) at a flow rate of 5.0 ml/min (data not shown). The eluate was immediately neutralized with 2 M Tris to pH 7.4.

Conventional Chromatography. This was conducted as described in the legend to Fig. 1. Sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) was carried out according to the method of Laemmli (1970). Polyacrylamide gel isoelectric focusing was conducted with Biolyte 3/10 (Bio-Rad Japan, Tokyo). Enzymatic and Physiological Activities PLA2 activity was routinely assayed at pH 8.0 with egg-yolk emulsion as described previously (Ishimaru et al. 1980). One unit of

548

Amino Acid Sequence Analysis of PLA-N The N-terminal sequences of proteins and peptides were determined on a pulse liquid-phase sequencer (Applied Biosystems 470A, Foster City, CA). PLA-N obtained by conventional chromatography was reduced and S-carboxymethylated. The C-terminal three residues were analyzed on a C-terminal sequencer (Hewlett-Packard HPG1009A, Palo Alto, CA). S-Carboxymethyl (Cm)-PLA-N was digested with Achromobacter protease I or endoprotease Asp-N. The peptides were fractionated by HPLC on a Wakosil 5C18-200 column (0.21 · 15 cm) (Wako Pure Chemical Industries) and sequenced.

Cloning and Sequencing of the cDNAs Encoding PLA-N and PLA-N(O)

Fig. 1. The second peak proteins (0.96 g) from Sephadex G-100 of Amami-Oshima T. flavoviridis venom were loaded onto a CMSepharose CL6B column (2.5 · 14 cm) equilibrated with 0.025 M ammonium acetate (pH 6.8) and eluted with linear concentration gradient of ammonium acetate (pH 6.8) up to 0.8 M over 750 ml (a). Fractions of 4 ml were collected. The fractions indicated by a bar were collected, dialyzed against water, and lyophilized (94 mg). The proteins were applied onto a POROS 20S column (0.46 · 10 cm) equilibrated with 0.025 M Tris–HCl (pH 8.5) containing 0.1 mM CaCl2 and eluted with a linear concentration gradient of NaCl up to 1.0 M at a flow rate of 5.0 ml/min (b). The proteins from the fractions indicated by a bar were rechromatographed on a POROS 20S column under the same condition (data not shown). The protein finally obtained (6.3 mg) gave a sharp band on HPLC (l-Bondasphere 5l, C4, 300 A˚, 0.39 · 15 cm, Waters, Millford, MA) when eluted with a linear concentration gradient of CH3CN-2-propanol (7:3, v/v) in 0.05% TFA at a flow rate of 0.4 ml/min (c). This protein was named PLA-N.

enzyme activity is defined as the uptake of NaOH in micromoles per minute. Neurotoxic activity was assayed with chicken biventer cervicis muscle (Ginsborg and Warriner, 1960). The LD50 was assayed in mice (ICR strain) by intravenous injection.

Total RNAs were extracted from Amami-Oshima T. flavoviridis venom gland and mRNAs were purified with Dynabeads mRNA purification kit (Dynal, Oslo, Norway). The venom-gland cDNA library was constructed with kgt10 by the standard method (Sambrook et al. 1989). Sense primer, 50 -ATCAAGATAATGACGAAGAAAAAC-30 , and antisense primer, 50 -AAGAAAGTCCGGGTAAAAACATAT-30 , which correspond to Ile19– Asn16 and Tyr108–Leu115, respectively, of PLA-N, were employed for reverse transcription-polymerase chain reaction against the cDNA library. A 0.32-kbp DNA fragment amplified was subcloned into pTV118N plasmid vector at the HincII site. Its nucleotide sequence was confirmed to encode the partial sequence of PLA-N. The plasmid was cleaved by PstI. The resulting 0.18kbp DNA fragment corresponding to Asp56–Leu116 was labeled with [a-32P]dCTP by the random priming method (Feinberg and Vogelstein 1983) and employed for screening of Amami-Oshima T. flavoviridis venom-gland cDNA library by the plaque hybridization method (Sambrook et al. 1989). The cloned DNA in phage vector was digested with KpnI and subcloned into pTV118N. The nucleotide sequence of this cDNA determined by the dideoxy chain termination method (Sanger et al. 1977) completely matched with the amino acid sequence of PLA-N. This cDNA was named cPLA-N. The same screening conducted for the Tokunoshima T. flavoviridis venom-gland cDNA library (Oda et al. 1990) gave a cDNA with the same sequence as cPLA-N. Screening of the Okinawa T. flavoviridis venom-gland cDNA library (Chijiwa et al. 2000) with PLA-N cDNA led to isolation of four cDNA clones. All the clones had the same nucleotide sequence, which is only one base (nonsynonymous) substituted from that of cPLA-N. This clone was named cPLA-N(O).

Data Analysis for Molecular Evolution A phylogenetic tree of PLA2’s from Amami-Oshima, Tokunoshima, and Okinawa T. flavoviridis venoms together with PLA2’s from other Crotalinae snake venoms was constructed based on their amino acid sequences and by the neighbor-joining algorithm (Saitou and Nei 1987). Alignment was made by the CLUSTAL W program (Thompson et al. 1994). The degrees of confidence for internal lineages in a phylogenetic tree were determined by the bootstrap confidence (Felsenstein 1985) using Kimura’s (1969) method to compute a distance matrix with 1000 replicates. The DNASIS package developed by Hitachi Software Engineering was employed for analysis and alignment of DNA sequences. The numbers of nucleotide substitutions per site (KN) in the 50 - and 30 -untranslated regions (UTRs) and the numbers of

549

Fig. 2. The nucleotide sequences of cPLA-N (A, T) and cPLA-N(O) (O) and their amino acid sequences: 10 for PLA-N (A, T) and 30 for PLA-N(O) (O). A, Amami-Oshima; T, Tokunoshima; O, Okinawa. An asterisk indicates the position of nonsynonymous nucleotide substitution between cPLA-N and cPLA-N(O).

nucleotide substitutions per synonymous site (KS) and per nonsynonymous site (KA) in the protein-coding region were computed for pairs of T. flavoviridis venom PLA2 cDNAs and for pairs of Crotalinae venom neurotoxic PLA2 cDNAs according to the method of Nei and Gojobori (1986).

Results and Discussion Purification and Amino Acid Sequence of PLA-N The protein fractionated from Amami-Oshima T. flavoviridis venom by affinity chromatography on its serum peptide A-immobilized column gave a major band (about 90% purity) on SDS-PAGE with an apparent molecular weight of 14,000 (data not shown). Its N-terminal sequence was determined to be NLXQFNGMIKIMTKKNGFPFYTSYGXYXG (X = unidentified), which was highly similar to that of trimucrotoxin, a neurotoxic PLA2, from T. mucrosquamatus (Taiwan) venom (Tsai et al. 1995). This protein was then purified from the same venom by conventional chromatographic procedures (Fig. 1). Its pI was estimated to be 10.3 (data not shown). The N-terminal 51 residues were first determined. The alignment of the peptides obtained from Cm-PLA-N by digesting with Achromobacter protease I or endoprotease Asp-N together with the C-terminal three residues determined by the C-terminal sequencing established the complete amino acid sequence of PLA-N (Fig. 2; the sequence excluding the signal peptide). PLA-N had the structural features of group II PLA2 category.

Chemical and Biological Properties of PLA-N The specific activity of PLA-N against egg-yolk emulsion at pH 8.0 was 12% that of PLA2 (1300 units/ mg). PLA-N was only weakly inhibited by peptide A so that a 3000-fold molar excess of peptide A was required for 50% inhibition. Thus, the reason why PLA-N was specifically separated from T. flavoviridis venom by passing through a peptide A-immobilized column remains to be clarified. The sequence homologous to peptide A was not found in known T. flavoviridis serum PLA2 inhibitors (PLIs), PLI-IV (or PLI-A) and PLI-V (or PLI-B), which inhibit PLA2 (Inoue et al. 1991; Nobuhisa et al. 1997a), and PLI-I, which inhibits BPI and BPII (Nobuhisa et al. 1997b). It is conceivable that since the sequences of PLA2 and BPI or BPII are fairly different from that of PLA-N, neither PLI-IV and PLI-V nor PLI-I has affinity to PLA-N. In addition, no homologous sequence was found in any other proteins hitherto sequenced. PLA-N caused a dose-dependent inhibition of contraction of chicken biventer cervicis muscle without affecting acetylcholine response, indicating that PLA-N is a presynaptic neurotoxin. However, its activity was one order of magnitude lower than that of trimucrotoxin from T. mucrosquamatus venom. For example, the time necessary to inhibit 90% of contraction was 109.1 + 6.5 min at 10 lg/ml for PLA-N, whereas it was reported to be 115 + 15 min at 0.25 lg/ml for trimucrotoxin (Tsai et al. 1995). The LD50 for mice was 1.34 lg/g, which is not significantly different from that of trimucrotoxin (1.2 lg/g)

550

Fig. 3. The aligned amino acid sequences of T. flavoviridis (Tf) and T. gramineus (Tg) venom PLA2’s and other Crotalinae venom neurotoxic PLA2’s. A, Amami-Oshima: T, Tokunoshima; O, Okinawa. Tg PLA2’s are PLA-I (Oda et al. 1991), PLA-II and PLA-III (Fukagawa et al. 1993), and PLA-V (Nakai et al. 1995). Crotalinae

neurotoxic PLA2’s are trimucrotoxin from T. mucrosquamatus (Tm), agkistrotoxin from Agkistrodon halys pallas (Ag), and crotoxin B from Crotalus durissus terrificus (Cd) (Bouchier et al. 1991). Bovine pancreatic (Bp) [Asp49]PLA2 (Fleer et al. 1978) is included for alignment.

(Tsai et al. 1995). Recently we found that PLA-N exhibits strong cell toxicity in cancer cells such as HL-60 cells (Oda-Ueda et al. unpublished work).

Molecular Evolution

The Nucleotide Sequences of cPLA-N (Amami-Oshima and Tokunoshima) and cPLA-N(O) (Okinawa) These are shown in Fig. 2. The predicted amino acid sequence of PLA-N(O) is also included. cPLA-N and cPLA-N(O) are identical in the protein-coding region except for nonsynonymous substitution at position 499. The codon AAA (Lys) of cPLA-N was changed to AAC (Asn) for cPLA-N(O).

The amino acid sequences of PLA-N (Amami-Oshima and Tokunoshima) and PLA-N(O) (Okinawa) were aligned with other T. flavoviridis venom PLA2’s and T. gramineus (Taiwan) venom PLA2’s and with some neurotoxic PLA2’s from other Crotalinae snake venoms (Fig. 3). The homologies of PLA-N and PLA-N(O) to other T. flavoviridis PLA2 isozymes such as PLA2, PLA-B, and BPII are in the range of 52 to 67%. However, they are most homologous (91%) to neurotoxic trimucrotoxin from T. mucrosquamatus (Taiwan) venom. Their homologies to other neurotoxic PLA2’s such as agkistrotoxin

551

Fig. 4. Phylogenetic tree constructed for PLA2’s from Crotalinae venoms based on their amino acid sequences. The numerals at the nodes represent the bootstrap confidence. The branch lengths are drawn to scale and represent the numbers of amino acid substitu-

tion per site. The PLA2’s appeared for the first time are Crotalus atrox (Ca) PLA2 (Randolph and Heinrikson 1982) and Deinagkistrodon acutus (Da) [Lys49]PLA2 (Fan et al. 1999) and neurotoxic [Asp49]PLA2 (Wang et al. 1996).

(Kondo et al. 1989) from Agkistrodon halys pallas venom and crotoxin B (Bouchier et al. 1991) from Crotalus durissus terrificus venom are moderately high (both 71%). Thus, PLA-N and PLA-N(O) are more homologous to neurotoxic PLA2’s from the different origins than to other T. flavoviridis PLA2’s. PLA-N(O) is different from PLA-N at position 115 (Asn/Lys) (Fig. 3). Thus, interisland mutation occurred between them. This may be ascribed to the subtle difference in the environments between Amami-Oshima and Tokunoshima islands and Okinawa island. It is unknown to what degree such substitution affects the physiological activities of PLA-N in terms of neurotoxicity and cell toxicity against cancer cells. A phylogenetic tree was constructed for Crotalinae venom group II PLA2’s based on their amino acid sequences (Fig. 4). It reveals that Crotalinae venom group II PLA2’s are separated into four groups. This branching pattern should reflect the differences in the particular structures and functions of PLA2’s of each branch. We denote them PLA2 type (acidic to neutral [Asp49]PLA2’s), basic [Asp49]PLA2 type, [Lys49]PLA2 type, and neuroPLA2 type. The [Lys49]PLA2-type isozymes are basic proteins and known to have ex-

tremely low lipolytic activity against ordinary PLA2 substrates such as egg-yolk emulsion but have strong myolytic activity (Yoshizumi et al. 1990; Liu et al. 1990; Kihara et al. 1992). The other three groups are in the form of [Asp49]PLA2. The neuroPLA2-type isozymes uniquely form a cluster. The PLA2-type isozymes exhibit high lipolytic activity against ordinary PLA2 substrates and myolytic activity (Kihara et al. 1992). Evidence was recently presented that the mechanism of myolytic action of [Lys49]PLA2-type isozymes is different from that of PLA2-type isozymes (Nu´n˜ez et al. 2001). Basic [Asp49]PLA2-type isozymes in Fig. 4 are all originated from T. flavoviridis venom. Their lipolytic activities toward egg-yolk emulsion are 10–30% that of PLA2 (Yamaguchi et al. 2001; Chijiwa et al. 2003). Their particular physiological functions remain to be clarified except that PLA-B exhibits edema-inducing potency (Yamaguchi et al. 2001) but PL-Y does not (Chijiwa et al. 2003). The evolutionary relationship of snake venom PLA2 isozymes was studied for the nucleotide sequences of their cDNAs by computing the KN, KS, and KA values. When PLA-N cDNA was compared with other T. flavoviridis PLA2 cDNAs, the 50 - and 30 -

552 The KA/KS and KN/KS values for pairs of the cDNAs encoding T. flavoviridis venom PLA2’s

Table 1.

Mature protein-coding region cDNA pair PLA-N PLA-N PLA-N PLA-N

Table 2.

vs vs vs vs

PL-X0 PLA2 BPI PL-Y

50 -UTR (KN)

KS

KA

KA/KS

KN/KS

0.060 0.060 0.060 0.060

0.144 0.229 0.237 0.144

0.189 0.237 0.278 0.193

1.36 1.04 1.22 1.40

0.417 0.262 0.253 0.417

The KA/KS values for pairs of the cDNAs encoding Crotalinae venom neurotoxic PLA2’s Mature protein-coding region

cDNA pair

KS

KA

KA/KS

PLA-N vs trimucrotoxin PLA-N vs agkistrotoxin PLA-N vs crotoxin B Trimucrotoxin vs agkistrotoxin Trimucrotoxin vs crotoxin B Agkistrotoxin vs crotoxin B

0.0760 0.190 0.190 0.228 0.217 0.249

0.0387 0.0775 0.0775 0.0916 0.161 0.166

0.497 0.375 0.375 0.360 0.707 0.622

UTRs and the signal peptide-coding domains are highly conserved but the mature protein-coding regions are quite variable. Accelerated evolution was observed for the mature protein-coding region because the KA/KS values are greater than one (Table 1) (Nakashima et al. 1993, 1995; Nobuhisa et al. 1997). However, comparison was made within neuroPLA2type PLA2’s, the KA/KS values are in the range of 0.36–0.71 (Table 2). These values are lower than those for the protein-coding regions of the genes under accelerated evolution but higher than those for the genes under neutral evolution, which have been estimated at below 0.2 or so (Miyata et al. 1980; Wu and Li 1985; Nei 1987). It may be said that the genes encoding neurotoxic PLA2’s evolved in an accelerated manner until they had acquired neurotoxic function and after that they have evolved under less frequent mutation, possibly for functional conservation. Amami-Oshima, Tokunoshima, and Okinawa islands were separated from a former island 1 to 2 million years ago by eustacy (changes in sea levels) in the orogenic stage (Hoshino 1975). Thus, T. flavoviridis snakes in these islands have been kept isolated in the different environments over a long period. Our recent works indicated that some of the PLA2 isozymes of T. flavoviridis venom have evolved independently in these isolated islands. The most striking phenomenon is the lack of strongly myolytic BPI and BPII in Okinawa T. flavoviridis venom (Chijiwa et al. 2000), which are abundantly expressed in AmamiOshima and Tokunoshima T. flavoviridis venoms (Chijiwa et al. 2000, 2003). BPI and BPII genes in Okinawa T. flavoviridis had been converted to a pseudogene by their fusion. Interisland sequence diversities were also noted among PLA-B0 (Amami-

Oshima), PLA-B (Tokunoshima), and PL-Y (Okinawa), which belong to the basic [Asp49]PLA2 type and have one to four amino acid substitutions in the b-sheet and its vicinity (Chijiwa et al. 2003). In line with these facts, the present study evidenced that interisland mutation occurred between PLA-N (Amami-Oshima and Tokunoshima) and PLA-N(O) (Okinawa). Since T. flavoviridis venom PLA2 isozymes have evolved in an accelerated manner (Nakashima et al. 1993, 1995; Ohno et al. 1998, 2002), it seems that they are sensitive to the surrounding environments and tend to undergo mutation in an adaptive manner, possibly for prey species (Chijiwa et al. 2000). On the other hand, PLA2 is ubiquitously involved in the venoms of T. flavoviridis snakes from the three islands, although there are one to three synonymous nucleotide substitutions in their cDNAs (Chijiwa et al. 2003). The rigidity of the PLA2 gene in T. flavoviridis snakes even in different environments can be assumed to be based on its essential and principal roles in venom toxicity. From the observations that BPI and BPII are missing in only Okinawa T. flavoviridis venom and PLA-N is involved in Amami-Oshima and Tokunoshima T. flavoviridis venoms but its homologue, PLA-N(O), is contained in only Okinawa T. flavoviridis venom, it is inferred that Okinawa island was first separated from a former island and the rest was later disunited into Amami-Oshima and Tokunoshima islands.

Acknowledgments. The authors are indebted to Dr. Chewn-Lang Ho, Institute of Biological Chemistry, Academia Sinica, Taiwan, for the neurotoxicity assay of the PLA2 samples.

553

References Bouchier C, Boulain JC, Bon C, Me´nez A (1991) Analysis of cDNAs encoding the two subunits of crotoxin, a phospholipase A2 neurotoxin from rattlesnake venom: the acidic nonenzymatic subunit derives from a phospholipase A2-like precursor. Biochim Biophys Acta 1088:401–408 Brunie S, Bolin J, Gewirth D, Sigler PB (1985) The refined crystal structure of dimeric phospholipase A2 at 2.5 A˚. Access to a shielded catalytic center. J Biol Chem 260:9742–9749 Chijiwa T, Deshimaru M, Nobuhisa I, Nakai M, Ogawa T, Oda N, Nakashima K, Fukumaki Y, Shimohigashi Y, Hattori S, Ohno M (2000) Regional evolution of venom-gland phospholipase A2 isoenzymes of Trimeresurus flavoviridis snakes in the southwestern islands of Japan. Biochem J 347:491–499 Chijiwa T, Yamaguchi Y, Ogawa T, Deshimaru M, Nobuhisa I, Nakashima K, Oda-Ueda N, Fukumaki Y, Hattori S, Ohno M (2003) Interisland evolution of Trimeresurus flavoviridis venom phospholipase A2 isozymes. J Mol Evol 56:286–293 Dijkstra BW, Kalk KH, Hol WGJ, Drenth J (1981) Structure of bovine pancreatic phospholipase A2 at 1.7 A˚ resolution. J Mol Biol 147:97–123 Dijkstra BW, Renetseder R, Kalk KH, Hol WGJ, Drenth J (1983) Structure of porcine pancreatic phospholipase A2 at 2.6 A˚ resolution and comparison with bovine phospholipase A2. J Mol Biol 168:163–179 Dufton MJ, Hider RC (1983) Classification of phospholipases A2 according to sequence. Evolutionary and pharmacological implications. Eur J Biochem 137:545–551 Fan CY, Qian YC, Yang SL, Gong Y (1999) cDNA cloning and sequence analysis of Lys-49 phospholipase A2 from Agkistrodon acutus. Genet Anal 15:15–18 Feinberg AP, Vogelstein B (1983) A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity. Anal Biochem 132:6–13 Felsenstein J (1985) Confidence limits of phylogenetics: An approach using the bootstrap. Evolution 39:783–791 Fleer EAM, Verheij HM, de Haas GH (1978) The primary structure of bovine pancreatic phospholipase A2. Eur J Biochem 82:261–269 Fukagawa T, Matsumoto H, Shimohigashi Y, Ogawa T, Oda N, Chang CC, Ohno M (1992) Sequence determination and characterization of a phospholipase A2 isozyme from Trimeresurus gramineus (greeen habu snake) venom. Toxicon 30:1331–1341 Fukagawa T, Nose T, Shimohigashi Y, Ogawa T, Oda N, Nakashima K, Chang CC, Ohno M (1993) Purification, sequencing and characterization of single amino acid-substituted phospholipase A2 isozymes from Trimeresurus gramineus (green habu snake) venom. Toxicon 31:957–967 Ginsborg BL, Warriner J (1960) The isolated chick biventer cervicis nerve-muscle preparation. Br J Pharmacol 15:410–411 Holland DR, Clancy LL, Muchmore SW, Ryde TJ, Einspahr HM, Finzel BC, Heinrikson RL, Watenpaugh KD (1990) The crystal structure of a lysine49 phospholipase A2 from the venom of the cottonmouth snake at 2.0-A˚ resolution. J Biol Chem 265: 17649–17656 Hoshino M (1975) Eustacy in relation to orogenic stage. Tokai University Press, Tokyo Inoue S, Kogaki H, Ikeda K, Samejima Y, Omori-Satoh T (1991) Amino acid sequences of the two subunits of a phospholipase A2 inhibitor from the blood plasma of Trimeresurus flavoviridis. Sequence homologies with pulmonary surfactant apoprotein and animal lectins. J Biol Chem 266:1001–1007 Ishimaru K, Kihara H, Ohno M (1980) Purification and properties of phospholipase A2 from venom of Trimeresurus flavoviridis (habu snake). J Biochem (Tokyo) 88:443–451

Kihara H, Uchikawa R, Hattori S, Ohno M (1992) Myotoxicity and physiological effects of three Trimeresurus flavoviridis phospholipases A2. Biochem Int 28:895–903 Kimura M (1969) The rate of molecular evolution considered from the stand point of population genetics. Proc Natl Acad Sci USA 63:1181–1188 Kondo K, Zhang J, Xu K, Kagamiyama H (1989) Amino acid sequence of a presynaptic neurotoxin, agkistrodotoxin, from the venom of Agkistrodon halys pallas. J Biochem (Tokyo) 105:196–203 Laemmli UK (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680–685 Liu SY, Yoshizumi K, Oda N, Ohno M, Tokunaga F, Iwanaga S, Kihara H (1990) Purification and amino acid sequence of basic protein II, a lysine-49-phospholipase A2 with low activity, from Trimeresurus flavoviridis venom. J Biochem (Tokyo) 107:400– 408 Maraganore JM, Heinrikson RL (1986) The lysine-49 phospholipase A2 from the venom of Agkistrodon piscivorus piscivorus. Relation of structure and function to other phospholipases A2. J Biol Chem 261:4797–4804 Maraganore JM, Merutka G, Cho W, Welches W, Ke´zdy FJ, Heinrikson RL (1984) A new class of phospholipases A2 with lysine in place of aspartate 49. Functional consequences for calcium and substrate binding. J Biol Chem 259:13839–13843 Miyata T, Yasunaga T, Nishida T (1980) Nucleotide sequence divergence and functional constraint in mRNA evolution. Proc Natl Acad Sci USA 77:7328–7332 Nakai M, Nakashima K, Ogawa T, Shimohigashi Y, Hattori S, Chang CC, Ohno M (2001) Purification and primary structure of a myolytic lysine-49 phospholipase A2 with low lipolytic activity from Trimeresurus gramineus venom. Toxicon 33:1469– 1478 Nakashima K, Ogawa T, Oda N, Hattori M, Sakaki Y, Kihara H, Ohno M (1993) Accelerated evolution of Trimeresurus flavoviridis venom gland phospholipase A2 isozymes. Proc Natl Acad Sci USA 90:5964–5968 Nakashima K, Nobuhisa I, Deshimaru M, Nakai M, Ogawa T, Shimohigashi Y, Fukumaki Y, Hattori M, Sakaki Y, Hattori S, Ohno M (1995) Accelerated evolution in the protein-coding regions is universal in crotalinae snake venom gland phospholipase A2 isozyme genes. Proc Natl Acad Sci USA 92:5605– 5609 Nei M (1987) Molecular evolutionary genetics, Columbia University Press, New York Nei M, Gojobori T (1986) Simple methods for estimating the numbers of synonymous and nonsynonymous nucleotide substitutions. Mol Biol Evol 3:418–426 Nobuhisa I, Nakashima K, Deshimaru M, Ogawa T, Shimohigashi Y, Fukumaki Y, Sakaki Y, Hattori S, Kihara H, Ohno M (1996) Accelerated evolution of Trimeresurus okinavensis venom gland phospholipase A2 isozyme-encoding genes. Gene 172:267–272 Nobuhisa I, Deshimaru M, Chijiwa T, Nakashima K, Ogawa T, Shimohigashi Y, Fukumaki Y, Hattori S, Kihara H, Ohno M (1997a) Structures of genes encoding phospholipase A2 inhibitors from the serum of Trimeresurus flavoviridis snake. Gene 191:31–37 Nobuhisa I, Inamasu S, Nakai M, Tatsui A, Mimori T, Ogawa T, Shimohigashi Y, Fukumaki Y, Hattori S, Kihara H, Ohno M (1997b) Characterization and evolution of a gene encoding a Trimeresurus flavoviridis serum protein that inhibits basic phospholipase A2 isozymes in the snake’s venom. Eur J Biochem 249:838–845 Nu´n˜ez CE, Angulo Y, Lomonte B (2001) Identification of the myotoxic site of the Lys49 phospholipase A2 from Agkistrodon piscivorus piscivorus snake venom: Synthetic C-terminal pep-

554 tides from Lys49, but not from Asp49 myotoxins, exert membrane-damaging activities. Toxicon 39:1587–1594 Oda N, Ogawa T, Ohno M, Sasaki H, Sakaki Y, Kihara H (1990) Cloning and sequence analysis of cDNA for Trimeresurus flavoviridis phospholipase A2, and consequent revision of the amino acid sequence. J Biochem (Tokyo) 108:816–821 Oda N, Nakamura H, Sakamoto S, Liu SY, Kihara H, Chang CC, Ohno M (1991) Amino acid sequence of a phospholipase A2 from the venom of Trimeresurus gramineus (green habu snake). Toxicon 29:157–166 Ogawa T, Oda N, Nakashima K, Sasaki H, Hattori M, Sakaki Y, Kihara H, Ohno M (2002) Unusually high conservation of untranslated sequences in cDNAs for Trimeresurus flavoviridis phospholipase A2 isozymes. Proc Natl Acad Sci USA 89:8557– 8561 Ohno M, Me´nez R, Ogawa T, Danse JM, Shimohigshi Y, Fromen C, Ducancel F, Zinn-Justin S, Le Du MH, Boulain JC, Me´nez A (1998) Molecular evolution of snake toxins: Is the functional diversity of snake toxins associated with a mechanism of accelerated evolution? In: Moldave K (ed) Progress in nucleic acid research and molecular biology. Academic Press, New York, Vol 59, pp 307–364 Ohno M, Ogawa T, Oda-Ueda N, Chijiwa T, Hattori S (2002) Accelerated and regional evolution of snake venom gland isozymes. In: Me´nez A (ed) Perspectives in molecular toxinology, John Wiley & Sons, New York, pp 387–400 Randolph A, Heinrikson RL (1982) Crotalus atrox phospholipase A2: Amino acid sequence and studies on the functions of the NH2-terminal region. J Biol Chem 257:2155–2161 Renetseder R, Brunie S, Dijkstra BW, Drenth J, Sigler PB (1985) A comparison of the crystal structures of phospholipase A2 from bovine pancreas and Crotalus atrox venom. J Biol Chem 260:11627–11634 Saitou N, Nei M (1987) The neighbor-joining method: A new method for reconstructing phylogenetic trees. Mol Biol Evol 4:406–425

Sambrook J, Fritsch EF, Maniatis T (1989) Molecular cloning: A laboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY Sanger F, Nicklen S, Coulson AR (1977) DNA sequencing with chain-terminating inhibitors. Proc Natl Acad Sci USA 74:5463– 5467 Suzuki A, Matsueda E, Yamane T, Ashida T, Kihara H, Ohno M (1995) Crystal structure analysis of phospholipase A2 from Trimeresurus flavoviridis (Habu snake) venom at 1.5 A˚ resolution. J Biochem (Tokyo) 117:730–740 Thompson JD, Higgins DJ, Gibson TJ (1994) CLUSTAL W: Improving the sensitibility of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res 22:4673– 4680 Tsai IH, Lu PJ, Wang YM, Ho CL, Liaw LL (1995) Molecular cloning and characterization of a neurotoxic phospholipase A2 from the venom of Taiwan habu (Trimeresurus mucrosquamatus). Biochem J 311:895–900 Wang YM, Wang JH, Tsai IH (1996) Molecular cloning and deduced primary structures of acidic and basic phospholipases A2 from the venom of Deinagkistrodon acutus. Toxicon 34:1191– 1196 Wu CI, Li WH (1985) Evidence for higher rates of nucleotide substitution in rodents than in man. Proc Natl Acad Sci USA 82:1741–1745 Yamaguchi Y, Shimohigashi Y, Chijiwa T, Nakai M, Ogawa T, Hattori S, Ohno M (2001) Characterization, amino acid sequence and evolution of edema-inducing, basic phospholipase A2 from Trimeresurus flavoviridis venom. Toxicon 39:1069– 1076 Yoshizumi K, Liu SY, Miyata T, Saita, S, Ohno M, Iwanaga S, Kihara H (1990) Purification and amino acid sequence of basic protein I, a lysine-49-phospholipase A2 with low activity, from the venom of Trimeresurus flavoviridis (habu snake). Toxicon 28:43–54