Geographic variations, cloning, and functional analyses of the venom acidic phospholipases A2 of Crotalus viridis viridis

(1)

Geographic variations, cloning, and functional analyses of the

venom acidic phospholipases A

2 of Crotalus viridis viridis

q

Inn-Ho Tsai,

a,*

Ying-Ming Wang,

a

Yi-Hsuan Chen,

a

and Anthony T. Tu

b a_{Institute of Biological Chemistry, Academia Sinica, and Institute of Biochemical Sciences, National Taiwan University, Taipei, Taiwan}

b

Department of Biochemistry, Colorado State University, Fort Collins, Colorado 80523, USA Received 6 December 2002, and in revised form 26 December 2002

Abstract

Geographic venom samples of Crotalus viridis viridis were obtained from South Dakota, Wyoming, Colorado, Oklahoma, Texas, New Mexico, and Arizona. From these samples, the phospholipases A2 (PLA2s) were puriﬁed and their N-terminal sequences,

precise masses, and in vitro enzymatic activities were determined. We puriﬁed two to four distinct acidic PLA2s from each sample;

some of them displayed diﬀerent inhibition speciﬁcities toward mammalian platelets. One of the acidic PLA2s induced edema, but

had no anti-platelet activity. There was also a common basic PLA2myotoxin in all the samples. We have cloned ﬁve acidic PLA2s

and several hybrid-like nonexpressing PLA2s. Molecular masses and N-terminal sequences of the puriﬁed PLA2s were matched with

those deduced from the cDNA sequences, and the complete amino acid sequences of ﬁve novel acidic PLA2s were thus solved. They

share 78% or greater sequence identity, and a cladogram based on the sequences of many venom acidic PLA2s of New World pit

vipers revealed at least two subtypes. The results contribute to a better understanding of the ecogenetic adaptation of rattlesnakes and the structure–activity relationships and evolution of the acidic PLA2s in pit viper venom.

Keywords: Crotalus v. viridis; Snake venom; Phospholipase A2; Cloning; Complete sequence; Geographic variation; Platelet aggregation; Edema

Phospholipases A2 (PLA2s)1 are present in most, if

not all, pit viper venoms. This secreted enzyme family is relatively stable, easy to be puriﬁed, and without post-translational modiﬁcations. The amino acid sequences of about 200 PLA2s from snake venom have been

de-termined and many of their 3-D structures were deter-mined [1,2]. The venom PLA2s have evolved into several

functional subtypes, which may play diﬀerent biological roles such as platelet aggregation inhibitor, neurotoxin, anticoagulant, or myotoxin [3,4]. The acidic PLA2s

usually show strong hydrolytic activity and may inhibit the aggregation of platelets [5,6].

Prairie rattlesnakes (Crotalus viridis viridis, abbrevi-ated Cvv) inhabit the region east of the Rocky Moun-tains in the United States, from southern Canada to northern Mexico [7]. Its microhabitats, including prairie, cropland, grassland, and desert, are diverse. Some of the venom components have been puriﬁed and studied, in-cluding a kallikrein-like protease [8], the 5-kDa myo-toxins [9], and a basic and myotoxic PLA2 [10], but not

the acidic PLA2s. The intraspecies variations of its

venoms collected in Texas and New Mexico have been studied by electrophoresis and immunochemistry [11]. Recently, the phylogenetic analyses based on mito-chondrial DNA sequences revealed the taxa variations of Cvv from diﬀerent localities [12].

In this study, the intraspecies diversity of venom composition and its signiﬁcance are investigated at the molecular level using PLA2s as the window. We

ana-lyzed the venom of Cvv using the samples from seven localities of its range. The acidic venom PLA2s were Archives of Biochemistry and Biophysics 411 (2003) 289–296

www.elsevier.com/locate/yabbi

ABB

q

Sequence data from this article have been deposited with the EMBL Data Library under Accession Nos.AF403134–AF403137for Cvv-E6a, f, g, and h and Accession Nos.AY120875–AY120877for Cvv-E6e, d, and b, respectively.

*

Corresponding author. Fax: +886-2-2363-5038. E-mail address:bc201@gate.sinica.edu.tw(I.-H. Tsai). 1

Abbreviations used: Cvv, Crotalus viridis viridis; dPPC, LL -dip-amitoylglycerophosphatidylcholine; PLA2, phospholipase A2; ESI-MS, electrospray ionization-mass spectrometry; PRP, platelet-rich plasma.

(2)

puriﬁed, cloned, and sequenced. Their functional activ-ities, including inhibition of the aggregation of platelets from several species, were compared. Moreover, we found a special acidic PLA2 with edema-inducing but

not anti-platelet activity. The amino acid sequences of these PLA2s were compared to discuss the structure–

activity relationships, and a phylogeny tree based on these sequences was built to decipher their structural or evolutionary relationships.

Materials and methods

Venoms and other materials. Lyophilized samples of Cvv venom from Wyoming and South Dakota were purchased from Miami Serpentarium Laboratory and Kentucky Reptile Zoo, respectively. Professor Steve P. Mackessy (University of Northern Colorado, Greeley, CO, USA) supplied the Cvv venom from Colorado as a gift, and Professor Eppie Rael (University of Texas, El Paso, TX, USA) generously donated the Cvv venoms from Texas, New Mexico, and western Oklahoma. A live specimen (Cvv No.2) of south Arizona origin was purchased from Glades Herps, Inc. (Fort Myers, FL, USA). Its venom was obtained 2 days before the venom glands were taken and the snake was sacriﬁced. The mRNA extraction and the cDNA synthesis kits were from Stratagene. Modiﬁcation and restriction enzymes were purchased from Promega. Synthetic dipalmitoyl phosphatidylcholine was purchased from Avanti Polar Lipids (Alabaster, AL, USA). Other chemicals were of reagent grade.

Puriﬁcation and assay of PLA2s. Dissolved Cvv venom

(3–15 mg) was fractionated on a FPLCsystem with a column of Superdex G-75 (HR10/30, Pharmacia) in 0.1 M ammonium acetate (pH 4.2) at room temperature. The fractions corresponding to 12 2 and 26 2 kDa were pooled separately and lyophilized. These PLA2

-containing fractions were further fractionated by re-versed-phase HPLCusing a column of silica gel (Vydac C8, 14 250 mm, 10 lm) equilibrated with 0.07%

aqueous triﬂuoroacetic acid (solvent A) and eluted with a 25–45% linear gradient of CH3CN containing 0.07%

triﬂuoroacetic acid (solvent B). The puriﬁed PLA2s were

dried in a vacuum-centrifuge device (Labconco, USA). The dye-staining method of Bradford was used for quantiﬁcation of the puriﬁed PLA2s. Bovine serum

al-bumin (1 mg/ml, A280¼ 0:56) was used to prepare the

standard curve. PLA2activity was measured by the

pH-stat titration method [13] using the micellar substrates of

L

L-dipalmitoylglycerophosphatidylcholine (dPPC; 3 mM)

with a pH-stat apparatus (RTS 822; Radiometer, Den-mark). The reaction rate was corrected for the nonen-zymatic spontaneous rate.

Amino acid sequences and molecular mass. The N-terminal sequences of puriﬁed PLA2s were determined

by an automated amino acid sequencer (Model 477A; PE–Applied Biosystems). Molecular weights of the PLA2s were determined by ESI-MS on a Sciex mass

analyzer (API100; Perkin–Elmer) using about 2 lg of the enzyme dissolved in 20 ll of 0.1% acetic acid with 50% (v/v) CH3CN.

Cloning and sequencing. The cDNA library of Cvv venom glands from a Texas specimen (Cvv No.1) had been previously constructed [14]. For Cvv No.2 from Arizona, the venom gland mRNA and the cDNA were prepared as before [13,15]. In order to amplify and clone PLA2s, PCR [16] was conducted using SuperTaq DNA

polymerase with a pair of mixed-base oligonucleotide primers (primer 1, TCTGGATTG/CAGGAGGATGA GG, and primer 2, GCCTGCAGA/GACTTAGCA), which were designed according to the highly conserved cDNA regions of the group II PLA2s from snake

ven-oms [13,17]. Fragments of 0.4 kb were speciﬁcally am-pliﬁed by PCR as shown by electrophoresis of the products on a 1% agarose gel.

After being treated with polynucleotide kinase, the ampliﬁed DNA was inserted into the pGEM-T-Easy vector (Promega Biotech, Madison, WI, USA). Then, it was transformed into Escherichia coli strain JM109. White transformants were picked up and cDNA clones were selected. The DNA sequencing system Model 373A and the Taq Dye-Deoxy terminator-cycle sequencing kit (PE–Applied Biosystems) were used to determine the cDNA sequences by the dideoxynucleotide method [18]. Platelet aggregation and edema. Blood was collected from a healthy human donor, rabbit, and guinea pig, in the presence of 3.8% sodium citrate (9:1, v/v). The blood was centrifuged at 130g for 15 min at room temperature to prepare platelet-rich plasma (PRP). Platelet aggre-gation was measured by an aggregometer (Payton, Model 600B, Canada) at 37°C. Aliquots (0.45 ml) of PRP were preincubated with the tested PLA2 for 5 min

in a siliconized glass cuvette under constant stirring. The aggregation was initiated by the addition of ADP to a ﬁnal concentration of 10 lM and followed for 5–10 min. The dose-dependent inhibition of the ADP-induced ag-gregation of PRP by the puriﬁed PLA2was investigated.

The edema-inducing activity of the puriﬁed PLA2on

the hind paw of Wistar rats (female, weight about 250 g) was measured as previously described [15]. A plethys-mometer (Type 7150; Ugo Basile) monitored the time course of the swelling.

Phylogenetic analysis of the acidic venom PLA2s. The

phylogenetic tree was constructed based on amino acid sequences of the venom acidic PLA2s so far available for

the New World pit vipers. Multiple alignments of the sequences were based on the program PILEUP. Neighbor-joining method was used in the program PHYLIP to build the tree [19]. Conﬁdence of the linkage at each node was estimated by bootstrap analyses of 1000 replicas [20].

(3)

Results

Puriﬁcation and characterization of venom PLA2s Separation of dimeric and monomeric PLA2s from

the Cvv venom was achieved by gel-ﬁltration (Fig. 1). From each of the 26-kDa fractions, two or three acidic PLA2 isoforms were puriﬁed by reversed-phase

HPLC(Fig. 2). Residue 6 in each of the acidic en-zymes is Glu (i.e., E6), thus the PLA2s are designated

Cvv-E6a–f. Additionally, a single Asn6-containing basic PLA2, Cvv-N6, was found in all the geographic

samples. It was puriﬁed by HPLCfrom the 13-kDa fraction of each sample (not shown). The molecular mass was determined by ESI-MS and then the enzy-matic activities of potential PLA2s were assayed. A

total of seven distinct PLA2s were identiﬁed from all

the geographic samples, and their N-terminal se-quences were determined as shown in Table 1. Nota-bly, the molecular weights for Cvv-E6b, Cvv-E6c, and Cvv-E6e are diﬀerent although their N-terminal 23 residues are identical.

Among the seven geographic samples, only those from Arizona, New Mexico, and Texas are from indi-vidual snakes, the others are pooled venom samples. Their relative yields were calculated from the areas un-der the HPLCabsorbance peaks (Table 2). Total con-tent of the venom PLA2s was usually 4–7% of the venom

mass, except a higher content of 9.5% was found in the South Dakota sample. The content of Cvv-N6 appar-ently increased as the snakesÕ habitat moved farther to

the South (Table 2). The N-terminal sequence of Cvv-N6 was consistent with that reported previously for the myotoxic PLA2 of this venom [10].

Cloning and sequence determination

The venom gland cDNAs of Cvv No.1 and No.2 were used separately as templates in the PCR ampliﬁcation of the PLA2s. After electrophoresis, the 0.4-kb products

were harvested from agarose gel. The cDNAs were cloned and sequenced by normal procedures [18] and their nucleotide sequences were checked twice. They all encoded a signal peptide of 16 amino acid residues fol-lowed by a PLA2domain of 122 residues. All the venom

PLA2 clones reported herein were independently

iden-tiﬁed more than once.

Of about 90 clones of PLA2selected from the cDNA

library of Cvv No.1, about 50 encoded Cvv-E6f, the rest encoded Cvv-E6a and Cvv-N6. The majority of the 40 clones from cDNA of Cvv No.2 encoded Cvv-N6 and E6e, 4 encoded E6d, and only 1 encoded Cvv-E6b. However, Cvv-E6b was not isolated from the ve-nom of Cvv No.2; it was isolated from the Colorado and Wyoming samples only (Table 2). We failed to clone Cvv-E6c, which was probably expressed only in the ve-nom of Cvv from the northern range of its distribution (e.g., from South Dakota). The deduced amino acid sequences are aligned in Fig. 3 according to the com-monly used numbering system [21]. The theoretical pI values of all 5 Cvv-E6 are between 4.9 and 5.6, while that of Cvv-N6 is 8.8.

Fig. 1. Gel filtration of the crude venom of C. v. viridis. C. v. viridis venom samples collected from different states were loaded onto a Superdex G75 (HR10/30) column. Elution was carried out with the equilibration buffer, 0.1 N ammonium acetate (pH 6.0), at a flow rate of 1.0 ml/min. Active fractions containing the dimeric (first bar) and the monomeric (second bar) PLA2s were pooled, respectively.

(4)

In addition, we obtained ﬁve unusual PLA2 clones,

which did not match any of the puriﬁed PLA2s (Fig. 4).

These appeared to be the hybrid types, i.e., with the N-terminal and the C-N-terminal sequences derived, respec-tively, from diﬀerent PLA variants found in the venom. The hybrid clones from individual specimens were dif-ferent; PLA-g and h were from Cvv No.1 and PLA-i, j, and k were from Cvv No.2.

Functional analyses of the acidic PLA2s

The inhibitions of the ADP-induced aggregation of the platelets from mammalian blood by the Cvv-E6

variants were analyzed. We found that Cvv-E6b, d, and f showed very low anti-platelet activity to-ward platelet-rich plasma of human, rabbit (Fig. 5), or guinea pig (not shown), whereas Cvv-E6a, c, and e showed signiﬁcant inhibition. Notably, Cvv-E6a was a more potent inhibitor than Cvv-E6e for the hu-man and guinea pig platelets, but the relative po-tency was reversed in the case of rabbit platelets (Fig. 5).

In the search for possible functions for the Cvv-E6 variants without obvious anti-platelet activity, their edema-inducing activities were also tested. We found that Cvv-E6f eﬀectively induced local swelling of the Fig. 2. Puriﬁcation of the PLA2s by reversed-phase HPLC. Lyophilized 26-kDa fractions from Fig. 1 were redissolved in solvent A and each was

fractionated on a C8-Vydac HPLCcolumn with a gradient of solvent B (dashed lines). All the puriﬁed PLA2s were conﬁrmed by ESI-MS and

enzymatic assay. The annotations are the same as those shown in Table 1. The peak for a serine protease (mass 24855) is marked with an asterisk. Table 1

The elution conditions, molecular masses, and enzymatic activities of the acidic PLA2s of Crotalus v. viridis venom

CVV-PLA HPLC, %B Mass Enzyme activities,a_dPPC _{N-terminal sequences}b

ESI-MS Theoretical +Deoxycholate +Triton X-100

E6a 39 13467 1 13467.2 351 16 208 8 SLVQFETLIMKIAGRSGLLWYSA E6b 36 13660 1 13659.5 1129 20 764 7 N L V K S E6c 35 13817 1 — 840 6 378 10 N L V K S E6d 37 13782 1 13782.5 680 5 345 10 M V K FS E6e 36 13633 1 13633.5 1306 42 678 29 N L V K S E6f 37 13876 1 13875.7 518 25 118 4 MM I V K F G N6 33 14200 1 14199.6 280 8 88 4 N L NKM KMMTKKNAFPF TS

a_{Duplicate determination of the initial rate toward 3 mM dPPC, in the presence of 10 mM Ca}2þ_{at 37}_{°C. Values shown are means error (lmol/}

min/lg enzyme).

(5)

hind paw while other acidic PLA2variants did not. The

myotoxic Cvv-N6 elicited moderate edematous eﬀect (Fig. 6).

Phylogenetic tree of the venom acidic PLA2s

A cladogram was constructed to study the structural relationships between the venom acidic PLA2s of

Cro-talus and those of other New World pit vipers. The protein sequence tree shown in Fig. 7 reveals an aﬃnity among the Glu6-containing PLA2s in the Crotalus

ven-oms, with the robustness at most major nodes supported by bootstrap analysis. The tree also indicates that the acidic PLA2s of Cvv venom may have evolved into two

subtypes, one with anti-platelet activity, e.g., Cvv-E6a and the acidic venom PLA2s from C. atrox and C.

ada-manteus, while the other may have other functions or speciﬁcities, e.g., Cvv-E6f becomes edema inducing (Fig. 6). The phylogenetic analysis using the cDNA sequences of these PLA2s gave almost the same tree topology (not

shown).

Discussion

The gel-filtration patterns of the Cvv venoms from different localities reveal more or less proportional differences (Fig. 1). These samples are rich in high-molecular-weight components (>32 kDa), and their geographic variations may be attributed to venom pro-teins other than PLA2s [11]. The South Dakota Cvv

venom has relatively high content of the 8- to 9-kDa polypeptides (Fig. 1), possibly disintegrins [22].

There are several PLA2 isoforms in the Cvv venom

samples (Table 2). We identiﬁed the cDNA match for each of the PLA2 puriﬁed except Cvv-E6c (Fig. 3). The

anti-platelet Cvv-E6a was present in all the Cvv venoms except the one collected from Arizona, in which another anti-platelet Cvv-E6e was found. In fact, a geographic North to South transition for the relative proportions of these PLA2s was displayed, namely, E6a and

E6c are decreasing, while E6f, E6e, and Cvv-N6 are increasing. Notably, the South Dakota sample

Fig. 3. Multiple alignments of the amino acid sequences of Cvv-E6 PLA2s. The amino acid sequences are deduced from the cDNA sequences and

single-letter codes are used. The numbering system follows that of Renetseder et al. [21]. Residues identical to those in the top line are denoted with dots, gaps are marked with hyphens. The sequences of the venom PLA2s of C. atrox and C. adamanteus are listed for comparison.

Table 2

Geographic variation in the venom PLA2s of C. v. viridis

Locality CVV-PLA2s Relative abundance

(%)a

South Dakota E6a 3.2

E6c 5.8 N6 1.0 Wyoming E6a 1.5 E6b 0.6 E6c 1.4 N6 1.1 Colorado E6a 0.7 E6b 3.5 E6c 0.4 N6 1.5

Western Oklahoma E6a 0.8

E6c 0.4 E6e 1.4 N6 2.1 Texas E6a 0.6 E6c 0.3 E6f 1.6 E6e 0.6 N6 3.5

New Mexico E6a 0.1

E6f 1.3

E6e 1.6

N6 2.2

Southeastern Arizona E6a 0

E6d 0.8

E6e 1.2

N6 1.0

a_{The relative abundance was calculated based on the UV} absor-bance area under the PLA2peak in HPLCand calibrated with the total

(6)

contains only Cvv-E6c, while both Cvv-E6c and E6b are present in the pooled venom from Wyoming, which possibly represents a mixture of the South Dakota and the Colorado venoms. Thus, Cvv-E6c and E6b appear to be the markers for the Cvv venoms from their northern habitat, while Cvv-E6e and Cvv-E6f appear to be associated with those from the southern range, and Cvv-E6d is limited only to those from the most south-western range (Arizona).

We also found that the contents of Cvv-E6e varied signiﬁcantly between the individual Cvv venoms col-lected from Texas and New Mexico (data not shown). This is consistent with the results from the venom studies using electrophoresis [11]. Moreover, previous phylogenetic analyses using the mitochondrial DNA sequences [12,23] showed that the Cvv samples from Nebraska and Montana were close to those from Wy-oming and Colorado, while a sample from Bernalillo (New Mexico) was branched out from the rest of Cvv [23]. This and our ﬁnding that only the venom of Ari-zona Cvv contains Cvv-E6d (Table 2) suggest the pres-ence of a distinct population of Cvv in its most southwestern habitat.

Based on the cDNA sequence data, we found that all the signal peptides of the Cvv-E6 isoforms and that of the acidic PLA2 of C. atrox (GenBank Accession No.

AF269131) were identical, but diﬀered from the signal peptide of Cvv-N6 PLA2 by three amino acid

substitu-tions (not shown). Moreover, the amino acid sequence Fig. 6. Edema-inducing activities of some selected Cvv PLA2s. Rat

paws were injected with 8 lg of the puriﬁed venom PLA2 in 100 ll

sterile saline (0.9% NaCl). Saline solution was injected as the control. The volumes of the hind paws were measured by a plethysmometer. Experiments were done in duplicate, and data points are the average values with errors indicated.

Fig. 4. Predicted amino acid sequences of the hybrid-like PLA2clones. The possible origins of their genes or structural features are denoted for each

clone.

Fig. 5. Dose-dependent inhibition of platelet aggregation by Cvv-E6 PLA2s. The ADP-induced aggregation was studied at various Cvv-E6

concentrations with platelet-rich plasma of human and rabbit. The data points are the averages from two or three experiments.

(7)

of Cvv-E6a is >93% similar to those of the acidic PLA2s

from C. atrox and C. adamanteus venoms, but only 78% identical to that of Cvv-E6f (Fig. 3).

Thus, the acid PLA2s showed more geographic

vari-ations than the basic PLA2 in the venom of Cvv. In

analogy, two or three acidic PLA2 isoforms have been

found to be responsible for the geographic variations of the venoms of C. r. ruber [24] and Calloselasma rho-dostoma [25]. In all the cases, the acidic PLA2s show

qualitative variations. This is probably attributable to more constrained or narrow functional speciﬁcities of the acidic PLA2s.

In addition to the present study, the cDNAs encoding hybrid-like PLA2s were previously cloned from venom

glands of Agkistrodon halys Pallas [26] and Ca. rhodos-toma [15]. The presence of hybrid-like mRNAs encoding the venom serine proteases has been found in an Asian viper (our unpublished results). Notably, these clones appear to be diﬀerent between individual snakes, i.e., those encoding PLAs-g and -h were derived from Cvv No.1 while the others (Fig. 4) were from Cvv No.2. The hybrid-like PLAs have never been expressed or puriﬁed in the venoms for all the cases so far studied. Therefore regulation mechanisms must exist to halt translation or secretion of the proteins in the venom glands.

To understand the mechanism behind the intraspe-ciﬁc venom variations, it is important to understand the structure–activity relationships and prey speciﬁcity of the venom proteins. Among the Cvv-E6 variants, the hot spots of mutations were found at residues 7–13, 20, 34, and 117–130 (Fig. 4), i.e., the surface residues lo-cated in the water–lipid interface of the enzyme [1,2]. Moreover, Cvv-E6a contains a unique ‘‘RGD motif’’ at 128–130, the motif for platelet binding through the platelet integrin GPIIb/IIIa [27,28]. In addition, all the Cvv-E6 PLA2s contain a proline bracket at the

C-ter-minal region 122–131, suggesting possible functional importance of this region [28]. Notably, since Cvv-E6e and Cvv-E6b diﬀer by only a substitution at position 34, H34 probably contributes to stronger anti-platelet ac-tivities of Cvv-E6e and Cvv-E6a, but not the Y34 in Cvv-E6b and the Q34 in Cvv-E6f (Fig. 5).

Crystallographic [5] and mutagenesis studies [29] of the acidic venom PLA2 of A. halys Pallas (i.e., Ahp)

illustrated that residues E6, F20, W21, Y113, D115, and W119 probably are important for the anti-platelet ac-tivities of the acidic venom PLA2. Among the

corre-sponding sites in the Cvv-E6b, c, d, and f, only E6 and Y113 are conserved, while L20, S21, and I119 substitute for the aromatic residues (F20, W21, and W119 in the Ahp PLA2) (Fig. 3). These substitutions possibly

de-crease the anti-platelet activities of Cvv-E6. On the other hand, the edema-inducing Lys49-PLA2s of pit

viper venoms usually contained distinct substitutions such as R38, K78, D86, and K119 or R120 [30]. Inter-estingly, these substitutions were also found in Cvv-E6f,

which showed lower enzymatic activity than the other Cvv-E6 isoforms, but signiﬁcant edema-inducing activity not usually observed with other acidic PLA2s

(Fig. 6).

It has been shown that the anti-platelet activities of venom PLA2s depend on their enzymatic activities. The

products of the enzymatic reaction, lysophospholipid and fatty acids, may inhibit signal pathways and alter the cytoskeleton of the platelet [31,32]. The enzymatic activities of Cvv-E6b, c, and d are relatively high but their anti-platelet activities are not potent. It remains to be studied whether some of the Cvv acidic PLA2s mainly

play a role in digestion of prey [33].

We have demonstrated [24] that an E6 PLA2 of Ca.

rhodostoma venom, like Cvv-E6e, is a stronger inhibitor toward rabbit and rat platelets than another E6 PLA2of

the same venom with stronger inhibition specificity for human platelets, like Cvv-E6a (Fig. 5). It is known that the platelets from different vertebrates may display ra-ther distinct specificities toward exogeneous agents [34]. However, a parallel comparison of the diet composition and the venom PLA2variants of each Cvv population is

not possible at this time.

The venom proteins have been under natural selec-tion for adaptaselec-tion to ecological diﬀerences and the shift in diet composition [35–37]. The phylogenetic tree of the acidic PLA2s (Fig. 7) also suggests the evolution of at

least two subtypes of the venom E6 PLA2s in the

pres-ent-day C. v. viridis. Apparently, the multiple genes en-coding diﬀerent PLA2 variants may have resulted from

gene duplication [17] and point mutations (e.g., among Cvv-E6b, E6c, and E6e). The quantitative diﬀerences in

Fig. 7. Phylogenetic relationship between acidic venom PLA2s from pit

vipers of the New World. The amino acid sequences of the acidic ve-nom PLA2s were aligned and used as the dataset for phylogenetic

analyses. The tree was constructed with the program PHYLIP with the E6-containing PLA2from Bothrops pictus venom as the outgroup. The

GenBank accession numbers for the PLA2s and the species are

Cvv-6a–f of C. v. viridis(the present report); Chh-E6 of C. h. horridus (Tsai et al., unpublished result); Cad (P00623) of C. adamanteus; C at (P00624) of C. atrox; App-D49 (P51972) and App-dimer (A53872) of A. p. piscivorus; Cdt-PreE6 (P08878) of C. d. terriﬁcus; Bpi (AF288754) of B. pictus.

(8)

the venom PLA2s may have resulted from diﬀerential

gene expression (e.g., among Cvv-N6, E6a, and E6f). In conclusion, we demonstrated that the acidic PLA2s

in the Cvv geographic samples are relatively diversified, which possibly resulted from the necessary adaptation to great variations in the diet environment [34–36] and as a strategy to cope with platelet properties of different preys (Fig. 5). It is found for the first time that in ad-dition to inhibiting platelet aggregation, the role of certain venom acidic PLA2s (e.g., Cvv-E6f) may become

edema induction, or even other less studied functions (e.g., digestion of the prey).

Acknowledgments

We heartily thank Professor E. Rael and Professor S.P. Mackessy for the generous gifts of C. v. viridis ve-nom. This study has been supported by research grants from the Administration of Education (Grant 89-B-FA01-1-4) and National Science Council of Taiwan, ROC.

References

[1] J.M. Danse, S. Gasparini, A. Menez, in: R.M. Kini (Ed.), Venom Phospholipase A2Enzyme: Structure, Function and Mechanism, Wiley, UK, 1997, pp. 29–71.

[2] B.Z. Yu, J. Rogers, M.D. Tsai, C. Pidgeon, M.K. Jain, Biochem-istry 38 (1999) 4875–4884.

[3] I.-H. Tsai, J. Toxicol. Toxin Rev. 16 (1997) 79–114.

[4] B.R. Francis, J. Meng, I.I. Kaiser, in: G.S. Bailey (Ed.), Enzymes from Snake Venom, Alaken Inc., Colorado, USA, 1998, pp. 503– 544.

[5] X.-Q. Wang, J. Yang, L.-L. Gui, Z.-J. Lin, Y.-C. Chen, Y.-C. Zhou, J. Mol. Biol. 255 (1996) 669–676.

[6] R.M. Kini, H.J. Evans, in: R.M. Kini (Ed.), Venom Phoshphol-ipase A2 Enzyme: Structure, Function and Mechanism, Wiley, UK, 1997, pp. 369–387.

[7] J.L. Glenn, R.C. Straight, in: A.T. Tu (Ed.), Rattlesnake Venoms: Their Actions and Treatments, Marcel Dekker, New York, 1982, pp. 3–120.

[8] Y. Komori, T. Nikai, H. Sugihara, Biochim. Biophys. Acta 967 (1988) 92–102.

[9] A.L. Bieber, D. Nedelkov, J. Toxicol. Toxin Rev. 16 (1997) 33–52. [10] C.L. Ownby, T.R. Colberg, S.P. White, Toxicon 35 (1997) 111–

124.

[11] M. Anaya, E.D. Rael, C.S. Lieb, J.C. Perez, R.J. Salo, J. Herpetol. 26 (1992) 473–482.

[12] C.E. Pook, W. Wuster, R.S. Thorpe, Mol. Phylogenet. Evol. 15 (2000) 269–282.

[13] I.H. Tsai, P.J. Lu, Y.M. Wang, C.L. Ho, L.L. Liaw, Biochem. J. 311 (1995) 895–900.

[14] J.W. Norris, R.M. Fry, A.T. Tu, Biochem. Biophys. Res. Commun. 230 (1997) 607–610.

[15] I.H. Tsai, Y.M. Wang, L.C. Au, T.P. Ko, Y.H. Chen, Y.F. Chu, Eur. J. Biochem. 267 (2000) 6684–6691.

[16] K.B. Mullis, F. Faloona, Methods Enzymol. 155 (1987) 335–350. [17] T. Ogawa, M. Kitajima, K.I. Nakashima, Y. Sakaki, M. Ohno, J.

Mol. Evol. 41 (1995) 867–877.

[18] T. Maniatis, E.F. Fritsch, J. Sambrook, Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1989.

[19] J. Felsenstein, PHYLIP: the PHYLogeny Inference Package, version 3.573. Computer program distributed by the University of Washington, Department of Genetics, Seattle, 1992.

[20] J. Felsenstein, Evolution 39 (1985) 783–791.

[21] R. Renetseder, B.W. Dijkstra, K. Huizinga, K.H. Kalk, J. Drenth, J. Mol. Biol. 200 (1988) 181–188.

[22] R.M. Scarborough, J.W. Rose, M.A. Naughton, D.R. Phillips, L. Nannizzi, A. Arfsten, A.M. Campbell, I.F. Charo, J. Biol. Chem. 268 (1993) 1058–1065.

[23] K.G. Ashton, A. de Queiroz, Mol. Phylogenet. Evol. 21 (2001) 176–189.

[24] R.C. Straight, J.L. Glenn, T.B. Wolt, M.C. Wolfe, Comp. Biochem. Physiol. 103B (1992) 635–639.

[25] I.H. Tsai, Y.H. Chen, Y.M. Wang, Arch. Biochem. Biophys. 387 (2001) 257–264.

[26] H. Pan, X.-L. Liu, L.-L. Ou-Yang, G.-Z. Yang, Y.-C. Zhou, Z.-P. Li, X.-F. Wu, Toxicon 36 (1998) 1155–1163.

[27] T.F. Huang, S. Niewiarowski, J. Toxicol. Toxin Rev. 13 (1994) 253–271.

[28] R.M. Kini, H.J. Evan, FEBS Lett. 375 (1995) 15–17.

[29] X.L. Liu, X.F. Wu, Y.C. Zhou, J. Nat. Toxins 10 (2001) 43–55. [30] I.H. Tsai, Y.H. Chen, Y.M. Wang, M.C. Tu, A.T. Tu, Arch.

Biochem. Biophys. 394 (2001) 236–244.

[31] Y. Yuan, S.P. Jackson, C.A. Mitchell, H.H. Salem, Thromb. Res. 70 (1993) 471–481.

[32] Y. Yuan, S.M. Schoenwaelder, H.H. Salem, S.P. Jackson, J. Biol. Chem. 271 (1996) 27090–27098.

[33] R.G. Thomas, F.H. Pough, Toxicon 17 (1979) 221–228. [34] Y.L. Chen, T.F. Huang, S.W. Chen, I.H. Tsai, Biochem. J. 305

(1995) 513–520.

[35] S. Creer, W.H. Chou, A. Malhotra, R.S. Thorpe, Zool. Sci. 19 (2002) 907–913.

[36] S. Creer, A. Malhotra, R.S. Thorpe, W.H. Chou, Mol. Ecol. 10 (2001) 1967–1981.

[37] J.C. Daltry, W. W€uuster, R.S. Thorpe, J. Herpetol. 32 (1998) 198– 205.