Comparative proteomics and subtyping of venom phospholipases
A
2
and disintegrins of Protobothrops pit vipers
Inn-Ho Tsai*, Yi-Hsuan Chen, Ying-Ming Wang
Institute of Biological Chemistry, Academia Sinica, and Institute of Biochemical Sciences, National Taiwan University, Taipei, Taiwan Received 28 May 2004; received in revised form 23 July 2004; accepted 16 August 2004
Availble online 26 August 2004
Abstract
To explore the venom diversity and systematics of pit vipers under the genus Protobothrops, the venom phospholipases A2(PLA2s) of
P. mangshanensis, P. elegans and P. tokarensis were purified and characterized for the first time. The results were compared with the corresponding venom data of other co-generic species including P. mucrosquamatus, P. flavoviridis and P. jerdonii. Based on sequence features at the N-terminal regions, we identified five PLA2subtypes, i.e., the Asp49-PLA2s with N6, E6 or R6 substitution and the
Lys49-PLA2. However, not all subtypes were expressed in each of the species. Venom N6-PLA2s from P. mangshanensis and P. tokarensis
venom were weakly neurotoxic toward chick biventer cervicis tissue preparations. The venoms of P. tokarensis and P. flavoviridis contained identical PLA2 isoforms. In most Protobothrop disintegrins, sequences flanking the RGD-motif are conserved. Phylogenetic
analyses based on amino acid sequences of both families of the acidic PLA2s and the disintegrins clarify that these species could belong to
a monophyletic group.
D 2004 Elsevier B.V. All rights reserved.
Keywords: Pit viper venom; Phospholipase A2; Disintegrin; Anticoagulant; Neurotoxin; Protobothrops; Trimeresurus
1. Introduction
Phospholipase A2 (PLA2, EC 3.1.1.4) catalyzes the
hydrolysis of the 2-acyl ester bonds of phosphoglycerides to produce lysophosphoglycerides and fatty acids. By gene duplication and fast evolution [1,2], PLA2s of pit viper
venoms have been diversified into several distinct structural subtypes[3,4]and thus serve different functional roles. They all consist of 122 amino acid residues and share the same structural scaffold. On the other hand, venom disintegrins are short (less than 9 kDa) polypeptides with RGD or KGD exposed-loop for specific binding to a variety of glycoprotein
receptors on cell membrane. Both PLA2s and disintegrins are
common components of pit viper venoms.
Protobothrops is a genus of Asian pit vipers, comprised of about 12 terrestrial species. The distribution covers mainly China, Taiwan, Ryukyu (southwestern Japan), and they are responsible for significant portions of snakebites in these areas. P. mucrosquamatus inhabits northern India through southern China to Taiwan. P. jerdonii inhabits northern Indochina and southern China, and P. mangshanensis is an endemic species of Hunan province of China. Among the species of East Asian Islands, P. flavoviridis inhabits several islands of central Ryukyu, while P. tokarensis inhabits only the Tokaren islands of central Ryukyu, and P. elegans inhabits Yaeyama islands of southern Ryukyu[5].
A phylogenetic tree constructed from snake mitochondria DNA sequences suggested that Protobothrops is a mono-phyletic genus distinctive from the arboreal Trimeresurus (sensu stricto) and Tropidolaemus, and the terrestrial Ovophis
[6–9]. These four genera were formerly grouped under Trimeresurus (sensu lateral). The evolution of pit vipers on
1570-9639/$ - see front matterD 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.bbapap.2004.08.006
Abbreviations: HPLC, high performance liquid chromatography; dPPC, dipalmitoyl glycerophosphocholine; PLA2, phospholipase A2; Pel,
P. elegans; Pfl, P. flavoviridis; Pma, P. mangshanensis; Pmu, P. mucrosquamatus; Pto, P. tokarensis; Ook, Ovophis okinavensis
* Corresponding author. Fax: +886 2 23635038. E-mail address: bc201@gate.sinica.edu.tw (I.-H. Tsai).
East Asian Islands has been studied to address the question of whether the degree of genetic divergence reflects the history of isolation on the Ryukyu Archipelago; however, some controversy still remains[5]. Intrageneric venom variations of Protobothrops species also remain to be studied.
Among venom PLA2s of various Protobothrops species,
four from P. mucrosquamatus [10–12], four from P. flavoviridis [13,14] and at least one from P. jerdonii [15]
have been fully sequenced. Amino acid sequences of venom disintegrins of several Protobothrops are also available
[16,17]. In the present study, novel PLA2s from several less
studied Protobothrops were purified and characterized, their N-terminal sequences and masses were compared with the data of other co-generic species. Systematics of these species among the Old World and the New World pit vipers were examined based on the subtypes and molecular phylogeny of the venom PLA2s and disintegrins.
2. Materials and methods 2.1. Venoms and other materials
Venoms of P. mucrosquamatus, P. flavoviridis, P. tokarensis and P. elegans were purchased from Latoxan Co. (Valence, France). Venom of P. mangshanensis is a gift from Prof. Shu, Yu-Yen (Kuangxi Medical University, China).
2.2. Purification and characterization of venom PLA2s
Venom powder (5–15 mg) was dissolved in 0.1 or 0.2 ml of reagent-grade water. After repeated centrifugations at
12,000g for 5 min each time, aliquots of 100 Al were injected into a pre-equilibrated gel-filtration column (Super-dex 75, HR10/30) on a Pharmacia FPLC system (Amersham Biosciences). The column was eluted at 1.0 ml/min with 0.1 M ammonium acetate (pH 6.2) at room temperature of 23– 26 8C. Fractions with PLA2 activities were collected
separately and freeze-dried. They were further purified by reversed-phase HPLC with a C8 column (4.5250 mm, Vydac Co., USA). Purified PLA2s were dried in a
vacuum-centrifuge device (Labconco, USA). Their molecular masses were determined by electrospray ionization mass spectrom-etry on a mass spectrometer (API-100; Perkin Elmer, Foster City, USA). Protein sequences were determined by a gas-phase amino acid sequencer coupled with a phenylthiohy-dantoin amino acid analyzer (model 120A; Perkin Elmer)
[18].
2.3. Enzymatic activities and other functional assays Concentration of PLA2 in solution was determined by
reading the absorbance at 280 nm and assuming a molar absorption coefficient of 1.5 at 1.0 mg/ml of the protein. The hydrolytic activities of PLA2s towards mixed micelles
of l-dipalmitoyl phosphatidylcholine (dPPC, Avanti polar lipid, USA) and sodium deoxycholate or Triton X-100 (Sigma) were assayed at pH 7.4 and 37 8C on a pH-stat apparatus (Radiometer, Copenhagen) [3]. Neurotoxicity was assayed with chicken biventer cervicis neuromuscular tissue [19].
The effect of venom PLA2s on blood coagulation was
studied by activated partial thromboplastin time (APTT) with a Hemostasis Analyzer (model KC1, Sigma Diagnos-tics, USA) as described previously [20].
Fig. 1. Separation of PLA2s from crude venom by gel filtration. Dissolved venom was eluted at a flow rate of 1.0 ml/min at room temperature (25 8C) on a
FPLC system with a Superdex G75 (HR 10/30) column in equilibration with 0.1 M ammonium acetate (pH 6.2). Fractions containing PLA2s (indicated by
2.4. Phylogenetic analysis
Sequences closely related to the major acidic PLA2s
from P. mucrosquamatus and P. flavoviridis were selected by BLASTp search [21]. Partial sequences of three novel acidic PLA2s (Table 2) were also included in the data set
for the phylogenetic analysis. The disintegrin sequences
of related species were also collected [16,17,22]. Amino acid sequence alignment was made using PILEUP program. Cladograms were constructed based on these sequences by neighbor-joining algorithm using PHYLIP program [23], and the degree of confidence for the internal branch was determined by bootstrap methods
[24].
Fig. 2. Purification of PLA2s by reversed-phase HPLC. Lyophilized fractions fromFig. 1were redissolved and fractionated on a Vydac C8-HPLC column with
a gradient of B solvent (acetonitrile, dashed lines). Venom PLA2s were purified and confirmed by ESI-MS and pH-stat enzyme assay. Annotations of the PLA2s
were the same as those inTables 1 and 2, except MaTx denoted mangshantoxin.
Table 1
Molecular weights and partial sequences of the basic PLA2of Protobothrops venoms
PLA2 a
Venom species Mol. wt.F1 N-terminal sequencesb K49-PLA2
*Pmu-K49 P. mucrosquamatus 13667 SLIELGKMIFQETG –KNPVKNYGLYGCNCG
Pma-K49-like P. mangshanensis 13640 d Vd d d d d d Vd d d d d –d d d d d d d d d d d d d d d Pel-K49-like P. elegans 13701 d d d d d Wd d Vd d d d d –d EAd d d d d d d d d d d d Pto-K49-like P. tokarensis 13752 d d VQd Wd d d d d d d d –d EAAd d d d d d d d d d d *Tf-BP-I/II P. flavoviridis 13753 d d VQd Wd d d d d d d d –d EAAd d d d d d d d d d d N6D49-PLA2
*Trimucrotoxin P. mucrosquamatus 13902 NLLQFNKMIKIMTK–KNAIPFYSSYGCYCG
Mangshantoxin P. mangshanensis 13902 d d d d d d d d d d d d d d –d d d d d d d d d d d d d d d d Pma-N6 P. mangshanensis 14055 d d d d d d d d d d d d d d –d d d d d d d LFd d d d d d d Pto-N6 P. tokarensis 14033 d d d d d d d d d d d d d d –d d GFd d d Td d d d d d d *Tf-PLA-N P. flavoviridis 14033 d d d d d d d d d d d d d d –d d GFd d d Td d d d d d d
a
Asterisk denotes the PLA2whose full amino acid sequences has been solved. b
One letter amino acid codes are used, residues identical to the first line are marked with a dot. Gap at residue 15 is introduced to follow the common numbering system for the enzymes[25].
3. Results
3.1. Purification, characterization and subtyping of PLA2s
Proteins from the venoms of P. mucrosquamatus, P. tokarensis, P. elegans and P. mangshanensis were separated by gel-filtration column (Fig. 1). About 10% of the protein in the largest venom peak of the P. mangshanensis elution-profile contained acidic E6-PLA2 (possibly homodimers),
most of the PLA2s were in the peak (corresponding to 12–
14 kDa proteins) that follows. Fractions containing PLA2s
were further purified by HPLC, and they were eluted from the column in the following order: basic R6-PLA2,
W6/G6-PLA2(probably K49 PLA2), N6-PLA2, acidic R6-PLA2and
finally the E6-PLA2(Fig. 2). After the N-terminal sequence
was determined, each of the PLA2was annotated based on
the abbreviated species name (e.g., Pma for P. mangsha-nensis) and its apparent structural subtype judged by their N-terminal sequences. Orthologous PLA2s of these four
venom species and those of P. flavoviridis [13] and P. jerdonii [15] were grouped together (Tables 1 and 2). Protobothrops PLA2s could apparently be classified into
five subtypes, i.e., K49, N6, acidic and basic R6, and E6-PLA2s.
Each subtype could be easily identified based on the common structural features at residues 1–29, e.g., the substitutions L5, G6 or W6, Q11 and N28 were characteristic for K49-PLA2s (or R49-PLA2s [10]), while
the substitutions N6, I11, Y28 and D49 were character-istic for N6-PLA2s (Table 1). Acidic R6-PLA2s usually
contained E7 and E11 while basic R6-PLA2s contained
K7 and/or K11; and E6-PLA2s were acidic enzymes with
E6 and K11 substitutions (Table 2). Notably, although the residue at position 49 of some of the PLA2s is not
known, whether they belong to Asp49 or not is predicted from homology to others with full sequence solved (Tables 1 and 2). Members of the same subtypes apparently shared more than 76% sequence identity, while the identities across the subtypes are less than 60% [3].
Table 2
Molecular data of PLA2s with E6 or R6 substitution from Protobothrops venoms
PLA2a Venom species Mol. wt.F1 N-terminal sequencesb
Acidic R6
*Pmu-PLA-III P. mucrosquamatus 13973 NLWQFREMIKEATG–KEPLTTYLFYACYCG
Pma-R6-III P. mangshanensis 13922 d d d d d d d d d d d d d d –d d d d d d d d d d d d d d d *Jerdoxin P. jerdonii 13855 Hd d d d d d d d d d d d d –d d d d d d d d d d d d d d d Basic R6
Pma-R6-I P. mangshanensis 13873 NLLQFRKMIKKMTG –KEPILSYATYGCNCG
Pma-R6-II P. mangshanensis 13845 d d d d d d Dd d d d d d d –d d d d Vd d d Fd d d Yd d Pto-R6 P. tokarensis 14020 Hd d d d d d d d d d d d d –d d d d Vd d d Fd d d Yd d *Tf-PLA-B P. flavoviridis 14039 Hd d d d d d d d d d d d d –d d d d Vd d d Fd d d Yd d *Tf-PLA-BV/XV/Y P. flavoviridis 13949 Hd d d d d d d d d d d d d –d d d d Vd d d Fd d d Yd d E6
*Pmu-PL-I P. mucrosquamatus 13601 NLWQFENMIMKVAK–KSGILSYSAYGCYCG
Pma-E6 P. mangshanensis 13561 Gd d d d d d d d d d d d d –d d d d d d d d d d d d d d d Pel-E6a P. elegans 13588 Gd d d d d d d d d d d d d –d d d d d d d d d d d d d d d Pel-E6b P. elegans 13571 Gd d d d d d d d d d d d d –d d d d d d d d d d d d d d d Pto-E6 P. tokarensis 13764 Gd d d d d d d Id d Vd d –d d d d d d d d d d d d d d d *Tfl-PLA1a P. flavoviridis 13764 Gd d d d d d d Id d Vd d –d d d d d d d d d d d d d d d *Tfl-PLA1b P. flavoviridis 13925 Hd Md d d d d d Kd d TG–Rd d d WWd GSd d d d d d *Ts-A6 T. stejnegeri 13939 Hd Md d d d d d Kd d TG–Rd d d WWd GSd d d d d d *Ook-E6 O. okinavensis 13786 Hd Md d d TLd d Id d G–Rd d VWWd GSd d d d d d a
Asterisk denotes the PLA2whose full amino acid sequences has been solved. b
One letter amino acid codes are used, residues identical to the first line are marked with a dot. Gap at residue 15 is introduced to follow the common numbering system for the enzymes[25].
Table 3
Enzymatic and anticoagulating activities of the Protobothrops D49-PLA2s
PLA2a Specific activity (Amol/mg/min),
toward dPPC micelles with
Anticoagulating activity Deoxycholate Triton X-100 ED (Ag/ml)b
Pma-R6-I (K7) 8F1 n.d. 0.38 Pma-R6-II (D7) 376F27 113F12 0.73 Pto-R6 (K7) 238F11 436F35 0.36 Mangshantoxin (S24) 521F15 155F4 0.70 Pto-N6 (S24) 698F38 417F24 0.64 Pma-N6 (F24) 71F3 33F1 5.1 Pma-R6-III (E7) 659F25 368F19 5.2 Pmu-PL-III (R6E7) 690F26 228F17 2.4 Pmu-PLA-I (E6) 739F32 130F10 N40 Pma-E6 563F12 63F2 6.9 Pel-E6a 761F24 82F2 21 Pel-E6b 827F10 153F10 N40 Pto-E6 1428F12 941F31 37 a
Special amino acid substitution is shown in parentheses.
b
Effective dose of the PLA2to prolong coagulation time from 30 s (of
3.2. Assay and functional study
The in vitro enzymatic activity of purified PLA2 was
determined at 37 8C by pH-stat using micellar lecithin substrates (Table 3). Enzymatic activities of all the K49-PLA2s (not shown) and some basic R6-PLA2s (e.g.,
Pma-R6-II) were hardly detectable. The hydrolytic activities of the PLA2s for dPPC in the negatively charged
(deoxycho-late) micelles were between two- and threefold higher than those in the neutral (Triton X-100) micelles, except for the basic R6-PLA2s (e.g., Pto-R6) which had higher
specific-ities toward Triton X-100 micelles.
In agreement with previous data[20], the K49-PLA2s are
relatively poor anticoagulants (not shown). Anticoagulating activities of the D49-PLA2s are shown in the last column of
Table 3. We found that basic R6-PLA2s and most
N6-PLA2s, e.g., Pma-R6-II, trimucrotoxin, mangshantoxin and
Pto-N6, are strong anticoagulants, but acidic R6-PLA2s and
E6-PLA2s are not.
All the N6-PLA2s are basic enzymes[25]. Among them,
trimucrotoxin has been shown to be neurotoxic [10] and myotoxic[26], and Tf-PLA-N is weakly neurotoxic[13]. We
further discovered that the neurotoxicity of Pto-N6 toward chick tissue was as weak as that of Tf-PLA-N and Pma-N6 was hardly neurotoxic (Table 4). In addition, the toxicity of Pto-N6 was not significantly increased by the addition of crotoxin A (the acidic subunit of crotoxin[25]) (Table 4). 3.3. Molecular phylogeny of venom proteins
Based on the protein sequences determined or deduced from gene sequences, phylogenetic trees were constructed for medium-sized disintegrins [22] (Fig. 3) and acidic E6-PLA2s (Fig. 4). Kistrin, the disintegrin of Calloslasma
rhodostoma venom was assigned as an out-group for the disintegrin tree. The trees revealed the relationship between the venom proteins of Protobothrops and those of other related species with high bootstrap values or confidence.
4. Discussion
Like Bothrops venoms, most of the Protobothrops venoms contain K49-PLA2s. The gel-filtration patterns of
Protobothrops venoms are usually similar to one another (Fig. 1) but distinct from those of other genera[20,25]. We have previously suggested that four venom PLA2s subtypes
(E6, N6, R6, and K49) have been diverged and evolved in parallel in the present-days pit vipers [3,11]. Recently, another report including a tree/cladogram of full sequences of the Asian crotalid venom PLA2s also revealed the
presence of the same four subtypes [13,27]. The acidic R6-PLA2s are further separated from the basic R6-PLA2s in
our investigation (Table 2). Each PLA2 subtype possibly
plays special functional roles.
The acidic R6-PLA2s have been found so far only in
Protobothrops venoms, while the E6- and K49-PLA2s are
Fig. 3. Phylogenetic tree showing structural relationships of disintegrins from Protobothrops venoms. The sequence data set includes: trimucrin (X77089, GenBank accession number), jerdonin[16], elegantin E1 and E2[38], trimestatin[41], and other medium-sized disintegrins of related species[22]. Kistrin is assigned as an out-group, and values on branching points are calculated bootstrap values. Specific sequence flanking the RGD motif (represented by a rectangle) is also shown.
Table 4
Neurotoxicity of N6-PLA2toward the chick biventer cervicis tissue
N6-PLA2toxin Dose, Ag/ml Time for 90% inhibition
of the twitch, min
Trimucrotoxin 2.0 38F2 (n=4)a 1.0 76F13 (n=5) 0.3 106F2 (n=3) Mangshantoxin 1.0 77F3 (n=2) Pma-N6 2.0 N240 (n=2) Pto-N6 3.0 205F5 (n=2) Tf-PLA-Nb 3.0 230 (n=3)
a Numbers of experiments are shown in parentheses. b Data taken from Ref.[13].
rather common pit viper venom components[28]and basic R6- and N6-PLA2 are also present in other genera of pit
vipers [20,25]. Judging from the phylogeography of pit vipers and the content of their venom, it appears that the major PLA2s subtypes diverged possibly before the
branch-ing of most if not all the Protobothrops species. Then, minor diversity of the venom also occurred during separation, speciation, and adaptation of the species.
The K49-PLA2s of pit viper venom are basic proteins
with extremely low lipolytic activity but strong myotoxic and edema-inducing activities. Their highly basic C-terminal regions have been known to be related to the toxicity[29,30]. The N6-PLA2s exhibit neuro-myotoxicities
and rather high lipolytic activities, and their structure– activity relationships have been studied extensively[10,25]. They are lethal for mice at 1.2–2.0 mg/g[10,13]. Tf-PLA-N also exhibited strong cytotoxicity to cancer cells such as HL-60[27]. Moreover, most of the N6-PLA2s and basic
R6-PLA2s of Protobothrops venoms caused prolonged blood
coagulation time (Table 3).
Many of the acidic venom PLA2s are conceived as
platelet aggregation inhibitors [20]. The Glu6 and some aromatic amino acids at regions 20, 21, and 113–119 have been shown to be important for the anti-platelet activity of the E6-PLA2s[31]. In addition, the venom acidic R6-PLA2s
of Tmu-PLA-III [11] and P. jerdonii venom [32] were shown to inhibit platelet aggregation. It is interesting that all the acidic R6-PLA2s contain acidic residues at positions 7
and 11, while both positions are usually basic in the basic R6-PLA2s (Table 1). Residues responsible for inhibitory
activities of the acidic R6-PLA2s are not clear but Lys7, 10,
11 and 16 at the N-terminal helix are probably involved in the anticoagulating effect of basic R6-PLA2s since previous
mutations of basic residues at all positions 7, 10 and 16 to
Glu7 reduced the anticoagulating activities of human group II PLA2 [33]; these surface exposed sites has been
implicated in the anticoagulating action of another viperid basic R6-PLA2[34].
Notably, the data presented inTable 3revealed that basic R6-PLA2s (e.g., Pma-R6-I and Pto-R6) had much lower
enzymatic activities than the acidic R6-PLA2s (e.g.,
Pma-III). Previous report also showed that the basic R6-PLA2s of P. flavoviridis had weak lipolytic activities toward
egg-yolk emulsion [14]. In corroboration, the anticoagulat-ing mechanism of some basic group II PLA2s is not related
with their enzymatic activities [33].
Although two extra PLA2pseudogenes have been cloned
[27], Ovophis okinavensis venom contains only one acidic PLA2(Ook-E6) [35], which is structurally rather different
from those of Protobothrops venoms (Table 2). Moreover, we found that both Ovophis monticola and Ovophis gracilis venoms express only E6-PLA2(Tsai, et al., unpublished).
Thus, our results demonstrate that co-generic viperid species usually express the same conserved set of venom PLA2
subtypes. Previous species trees support the monophyletic nature of Protobothrops, which is at relatively root position among the crotalid snakes [6,9]. Unlike Protobothrops, other genera of pit vipers usually contain less or only an incomplete set of the venom PLA2s subtypes[3,20]. Thus, it
seems likely that certain venom PLA2subtypes could be lost
before or during evolution of the viperid genus.
Notably, P. mangshanensis expresses a rather complete set of venom PLA2subtypes although each is not in large
quantity, and it is the only Protobothrops venom species found to contain both the acidic and the basic R6-PLA2s and
the N6-PLA2s of either S24 or F24 substitution (Tables 1
and 2). Previously, we reported the cloning of both S24 and F24 N6-PLA2s from the venom glands of a Sistrurus
Fig. 4. Phylogenetic tree of venom E6-PLA2s from Protobothrops and related species. Full amino acid sequences of the acidic PLA2s form P.
mucrosquamatus (X77088, GenBank accession number) and P. flavoviridis (D10724, D10725 GenBank accession numbers) were included in addition to those previously used[20]. Partial sequences (N-terminal residues 1–30) were underlined and the sequence of Pma-E6 is identical to that of Pel-E6. Ook-E6a (P00625, GenBank accession number) from O. okinavesis venom is the out-group. Values are calculated bootstrap values, and the amino acid residue-31 characteristic for each E6-PLA2cluster is also shown on the node.
species while other rattlers’ venoms express only one of them or none [25]. The close evolutionary relationships between Protobothrops and the New World rattlesnakes are thus further supported.
In spite of the genus-specific expression of venom subtypes, venom contents may be complicated by geo-graphic variations and differential expression as a result of adaptation to their feeding habits or ecology [20,36]. For examples, P. jerdonii venom appears to lack E6- and N6-PLA2s [31], some populations of P. flavoviridis do not
express N6-PLA2s[13], and some Taiwanese populations of
P. mucrosquamatus do not express acidic R6-PLA2s[11],
which are expressed in the venoms of continental Proto-bothrops species but absent from those of the three Ryukyu species (Table 2). Inter-island sequence diversities of the basic R6-PLA2s have been reported for P. flavoviridis
venoms [14], Tf-PLA-N also revealed some geographic variations [13,27]. An extraordinary diversity of the disintegrins of P. flavoviridis (Fig. 3) also suggests its rich inter-island biodiversities.
Judging from the N-terminal sequences, the venom PLA2s of three continental Protobothrops species (P.
mucrosquamatus, P. mangshanensis and P. jerdonii) are close to each other than to those from the three Ryukyu species (P. flavoviridis, P. tokarensis and P. elegans) (Tables 1 and 2). However, the E6-PLA2s sequences (Table 2) of
southern Ryukyu P. elegans are rather similar to those of P. mucrosquamatus and P. mangshanensis. It is conceived that central Ryukyu was insulated in the late Pliocene (1.8–2 my)[5]. Its fauna probably diverged from the ancestral or allied species from China and Taiwan at relatively early stages of evolution. In contrast, the identical PLA2s in P.
flavoviridis and P. tokarensis venoms suggests that the separation of their inhabiting islands and subsequent speciation are relatively recent events[5].
Because of the small size and stability, disintegrins are as good as venom PLA2s for studying the biosystematics of
venom species. Cladogram for the medium-size disintegrins commonly found in Protobothrops venom has been con-structed based on their amino acid sequences (Fig. 3). The tree shows that southern Ryukyu P. elegans is similar to P. mucrosquamatus, while the central Ryukyu P. flavoviridis is not. Previous species trees based on mtDNA sequences[5]
or snake morphologies[37]also showed the sisterhood of P. elegans with P. mucrosquamatus, and bflavoviridis and tokarensis cladeQ with P. jerdonii. Three variants of trigramin (the disintegrins from the bamboo tree viper T. stejnegeri) were also included in the phylogenetic tree (Fig. 3), and trigraminh is identical to albolabrin of T. albolabris
[22]. Protobothrops disintegrins are separated from Trimer-esurus disintegrins in this cladogram.
Moreover, amino acid residues flanking the RGD motif are rather conserved among the Protobothrops disintegrins (Fig. 3). Most of them conserved the sequences RARGDNP, which is the binding site for the integrins aIIbh3, avh3and
a5h1receptors [38,39]. This is different from those of the
arboreal Trimeresurus containing IARGDDL. However, P. elegans venom contains additional disintegrin variant with RARGDDL while P. flavoviridis contains additional variant with the sequences IARGDFP. Previous finding showed that large hydrophobic side chains at the position X of the RGDX motif are favorable for high-affinity interactions with human platelet aIIbh3receptor[38]. The position right
after X was also found to be important for the disintegrin binding and specificity [40,41]. Notably, the sequences of the RGD-imbedding loop in Protobothrops disintegrins (Fig. 3) and those in the rattlesnake disintegrins (e.g., molossin, viridian, basilicin and cereberin [38]) are highly similar.
In the cladogram for acidic E6-PLA2(Fig. 4), Tf-PL-Ib
(a non-expressing E6-PLA2 from P. flavoviridis venom
gland) is looped out. The topology of the venom tree agrees, in general, with snake taxonomy except that T. stejnegeri has exceptionally high diversity of the E6-PLA2 variants
[20]. Interestingly, each of the E6-PLA2 cluster shares a
distinct residue 31 (W31, K31 or A31, respectively) (Fig. 4). This residue at the entrance of substrate of the enzyme is one of the important interface recognition sites regulating its catalytic specificity [42].
In conclusion, by comparing HPLC profiles, and N-terminal sequences, we found that the venom PLA2s of
Protobothrops are genus-specific, and evolved with five paralogous subtypes (Tables 1 and 2). Both venom trees of the acidic PLA2s and the disintegrins support the
mono-phyletic nature of Protobothrops and its distinctness from other genera including Trimeresurus (sensu stricto) and Ovophis. We found that P. mangshanensis retains many venom PLA2s typical of Protobothrops; whether it belongs
to a separated genus is questionable [5]. The venom of P. tokarensis is almost identical to those of P. flavoviridis. The fact that P. flavoviridis differs from P. mucrosquamatus and P. jerdonii possibly resulted from either its especially fast evolution [27] or the hybridization of P. flavoviridis with other species in ancient time as revealed by its extra E6-PLA2and disintegrin variants in the venom (Figs. 3 and 4).
Acknowledgments
This study was supported by research grants from the Administration of Education (grant 89-B-FA01-1-4) and National Science Council (grant NSC91-2311-B001-114) of Taiwan, R. O. C. We thank Prof. Yu-Yan Su for the gift of P. mangshanensis venom.
References
[1] F. Gubensek, D. Kordis, Venom phospholipases A2genes and their
molecular evolution, in: R.M. Kini (Ed.), Venom Phospholipase A2
Enzyme: Structure, Function and Mechanism, J. Wiley and Sons, London, 1997, pp. 73 – 95.
[2] K. Nakashima, T. Ogawa, N. Oda, M. Hattori, Y. Sakaki, H. Kihara, M. Ohno, Accelerated evolution of Trimeresurus flavoviridis venom gland phospholipase A2 isozymes, Proc. Natl. Acad. Sci. U. S. A. 90 (1995) 5964 – 5968.
[3] I.H. Tsai, Phospholipases A2 from Asian snake venom, J. Toxicol.,
Toxin Rev. 16 (1997) 79 – 113.
[4] B.R. Francis, J. Meng, I.I. Kaiser, Classification of snake venom group II phospholipases A2according to amino acid sequences, in:
G.S. Bailey (Ed.), Enzymes from Snake Venom, Alaken, Colorado, USA, 1998, pp. 503 – 544.
[5] M.C. Tu, H.Y. Wang, M.P. Tsai, M. Toda, W.J. Lee, F.J. Zhang, H. Ota, Phylogeny, taxonomy, and biogeography of the oriental pitvipers of the genus Trimeresurus (Reptilia: Viperidae: Crotalinae): a molecular perspective, Zool. Sci. 17 (2000) 1147 – 1157.
[6] C.L. Parkinson, Molecular systematics and biogeographic history of pitviper as determined by mitochondrial and ribosomal DNA sequences, Copeia 1999 (1999) 576 – 586.
[7] S. Creer, A. Malhotra, R.S. Thorpe, Assessing the phylogenetic utility of four mitochondrial genes and a nuclear intron in the Asian pit viper genus, Trimeresurus: separate, simultaneous and conditional data combination analyses, Mol. Biol. Evol. 20 (2003) 1240 – 1251. [8] A. Malhotra, R.S. Thorpe, A phylogeny of the Trimeresurus group of
pit vipers: new evidence from a mitochondrial gene tree, Mol. Phylogenet. Evol. 16 (2000) 199 – 211.
[9] F. Klaus, D.G. Mink, W.M. Brown, Crotalin intergeneric relationships based on mitochondrial DNA sequence data, Copeia 1996 (1996) 763 – 773.
[10] I.H. Tsai, P.J. Lu, Y.M. Wang, C.L. Ho, L.L. Liaw, Molecular cloning and characterization of a neurotoxic phospholipase A2 from the
venom of Taiwan habu (T. mucrosquamatus), Biochem. J. 311 (1995) 895 – 900.
[11] I.H. Tsai, Y.H. Chen, Y.M. Wang, M.Y. Liau, P.J. Lu, Differential expression and geographic variation of the venom phospholipases A2
of Calloselasma rhodostoma and Trimeresurus mucrosquamatus, Arch. Biochem. Biophys. 387 (2001) 257 – 264.
[12] C.S. Liu, J.M. Chen, C.H. Chang, S.W. Chen, C.M. Teng, I.H. Tsai, The amino acid sequence and properties of an edema-inducing Lys-49 phospholipase A2 homolog from the venom of Trimeresurus
mucrosquamatus, Biochim. Biophys. Acta 1077 (1991) 362 – 370. [13] T. Chijiwa, S. Hamai, S. Tsubouchi, T. Ogawa, M. Deshimaru, N.
Oda-Ueda, S. Hattori, H. Kihara, S. Tsunasawa, M. Ohno, Interis-land mutation of a novel phospholipase A2 from Trimeresurus
flavoviridis venom and evolution of Crotalinae group II phospho-lipases A2, J. Mol. Evol. 57 (2003) 546 – 554.
[14] T. Chijiwa, Y. Yamaguchi, T. Ogawa, M. Deshimaru, I. Nobuhisa, K. Nakashima, N. Oda-Ueda, Y. Fukumaki, S. Hattori, M. Ohno, Interisland evolution of Trimeresurus flavoviridis venom phospholi-pase A2isozymes, J. Mol. Evol. 56 (2003) 286 – 293.
[15] Q.M. Lu, Y. Jin, J.F. Wei, W.Y. Wang, Y.L. Xiong, Biochemical and biological properties of Trimeresurus jerdonii venom and character-ization of a platelet aggregation-inhibiting acidic phospholipase A2,
J. Nat. Toxins 11 (2002) 25 – 33.
[16] X.D. Zhou, Y. Jin, R.Q. Chen, Q.M. Lu, J.B. Wu, W.Y. Wang, Y.L. Xiong, Purification, cloning and biological characterization of a novel disintegrin from Trimeresurus jerdonii venom, Toxicon 43 (2004) 69 – 75.
[17] I.H. Tsai, Y.M. Wang, Y.H. Lee, Characterization of a cDNA encoding the precursor of platelet aggregation inhibition and metalloproteinase from Trimeresurus mucrosquamatus venom, Biochim. Biophys. Acta 1200 (1994) 337 – 340.
[18] M.W. Hunkapiller, L.E. Hood, Analysis of phenylthiohydantoins by ultrasensitive gradient high-performance liquid chromatography, Methods Enzymol. 91 (1983) 486 – 493.
[19] B.L. Ginsborg, J. Warriner, The isolated chick biventer cervicis nerve-muscle preparation, Br. J. Pharmacol. 15 (1960) 410 – 441. [20] I.H. Tsai, Y.M. Wang, Y.H. Chen, T.S. Tsai, M.C. Tu, Venom
phospholipases A2 of bamboo viper (Trimeresurus stejnegeri):
molecular characterization, geographic variations and evidence of multiple ancestries, Biochem. J. 377 (2004) 215 – 223.
[21] S. McGinnis, T.L. Madden, BLAST: at the core of a powerful and diverse set of sequence analysis tools, Nucleic Acids Res. 32 (2004) W20 – W25.
[22] J.J. Calvete, M.P. Moreno-Murciano, R.D.G. Theakston, D.G. Kisiel, C. Marcinkiewicz, Snake venom disintegrins: novel dimeric disinte-grins and structural diversification by disulfide bond engineering, Biochem. J. 372 (2003) 725 – 734.
[23] J. Felsenstein (1992) PHYLIP: the PHYLogeny Inference Package, version 3.573. Computer program distributed by the U. of Wash-ington, Dept. of Genetics, Seattle, U.S.A.
[24] J. Felsenstein, Confidence limits on phylogenies: an approach using the bootstrap, Evolution 39 (1985) 783 – 791.
[25] Y.H. Chen, Y.M. Wang, M.J. Hseu, I.H. Tsai, Molecular evolution and structure–activity relationships of crotoxin-B like asparagin 6 phospholipases A2 in pit viper venoms, Biochem. J. 381 (2004)
25 – 34.
[26] L.N. Chen, C.S. Liu, C.C. Chang, Isolation and characterization of a toxic phospholipase A2 from the venom of the Taiwan habu (Trimeresurus mucrosquamatus), Biotechnol. Appl. Biochem. 19 (1994) 61 – 73.
[27] M. Ohno, T. Chijiwa, N. Oda-Ueda, T. Ogawa, S. Hattori, Molecular evolution of myotoxic phospholipases A2from snake venom, Toxicon
42 (2003) 841 – 854.
[28] M. Wang, Y.F. Liew, K.Y. Chang, I.H. Tsai, Purification and characterization of the venom phospholipases A2 from Asian
monotypic Crotalinae snakes, J. Nat. Toxins 8 (1999) 331 – 340. [29] L. Chioato, A.H.C. de Oiveira, R. Ruller, J.M. Sa’, R.J. Ward, Distinct
sites for myotoxic and membrane-damaging activities in the C-terminal region of a Lys49-phospholipase A2, Biochem. J. 366 (2002)
971 – 976.
[30] B. Lomonte, E. Moreno, A. Tarkowski, L.A. Hanson, M. Maccarana, Neutralizing interaction between heparins and myotoxin II, a lysine 49 phospholipase A2 from Bothrops asper snake venom. Identi-fication of a heparin-binding and cytolytic toxin region by the use of synthetic peptides and molecular modeling, J. Biol. Chem. 269 (1994) 29867 – 29873.
[31] X.L. Liu, X.F. Wu, Y.C. Chou, Identification of key residues responsible for enzymatic and platelet-aggregation-inhibiting activ-ities of acidic phospholipase A2s from Agkistrodon halys Pallas,
J. Nat. Toxins 10 (2001) 43 – 55.
[32] Q.M. Lu, Y. Jin, J.F. Wei, D.S. Li, S.W. Zhu, W.Y. Wang, Y.L. Xiong, Characterization and cloning of a novel phospholipase A2
from the venom of Trimeresurus jerdonii snake, Toxicon 40 (2002) 1313 – 1319.
[33] C.M. Mounier, P. Luchetta, C. Lecut, R.S Koduri, G. Faure, G. Lambeau, E. Valentin, A. Singer, F. Ghomashchi, S. Be´guin, M.H. Gelb, C. Bon, Basic residues of human group II phospholipase A2are
important for binding to factor Xa and prothrombinase inhibition comparison with other mammalian secreted phospholipases A2, Eur. J.
Biochem. 267 (2000) 4960 – 4969.
[34] K.H. Zhao, Z.J. Lin, S.Y. Song, Y.C. Zhou, Structure of a basic phospholipase A2from Agkistrodon halys Pallas at 2.13 2 resolution,
Acta Crystallogr., D 54 (1998) 510 – 521.
[35] F.J. Joubert, T. Haylett, Snake venoms. Purification, some properties and amino acid sequence of a phospholipase A2 (DE-I) from Trimeresurus okinavensis (Hime-habu) venom, Hoppe Seyler Z. Physiol. Chem. 362 (1981) 997 – 1006.
[36] J.C. Daltry, W. Wuster, R.S. Thorpe, The role of ecology in determining venom variation in Malayan pitviper, C. rhodostoma, in: R.S. Thorpe, W. Wuster, A. Malhotra (Eds.), Venomous Snakes, Clarendon Press, Oxford, U.K, 1997, pp. 155 – 172.
[37] B.H. Brattstrom, Evolution of the pit vipers, Trans. S. Diego Soc. Nat. Hist. 13 (1964) 185 – 268.
[38] A. Scaloni, E.D. Matino, N. Miraglia, A. Pelagalli, R.D. Morte, N. Staiano, P. Pucci, Amino acid sequence and molecular modelling of
glycoprotein IIb–IIIa and fibronectin receptor iso-antagonists from Trimeresurus elegans venom, Biochem. J. 319 (1996) 775 – 782. [39] L.W. Lee, H.C. Peng, W.C. Ko, W.C. Hung, C.H. Su, C.H. Lin, T.F.
Huang, M.H. Yen, J.R. Sheu, Triflavin potentiates the antiplatelet activity of platelet activating factor receptor antagonist on activated neutrophil-induced platelet aggregation, Eur. J. Pharmacol. 364 (1999) 239 – 246.
[40] C.P. Chang, J.C. Chang, H.H. Chang, W.J. Tsai, S.J. Lo, Positional importance of Pro53 adjacent to the Arg49–Gly50–Asp51sequence of
rhodostomin in binding to integrin aIIbh3, Biochem. J. 357 (2001) 57 – 64.
[41] Y. Fujii, D. Okuda, Z. Fujimoto, K. Horii, T. Morita, H. Mizuno, Crystal Structure of trimestatin, a disintegrin containing a cell adhesion recognition motif RGD, J. Mol. Biol. 332 (2003) 1115 – 1122. [42] O.P. Kuipers, J. Kerver, J. van Meersbergen, R. Vis, R. Dijkman, H.M.
Verheij, G.H. de Haas, Influence of size and polarity of residue 31 in porcine pancreatic phospholipase A2on catalytic properties, Protein
