Phospholipases A
2from
Callosellasma rhodostoma
venom gland
Cloning and sequencing of 10 of the cDNAs, three-dimensional modelling and chemical
modification of the major isozyme
Inn-Ho Tsai1,2, Ying-Ming Wang1, Lo-Chun Au3, Tzu-Ping Ko1, Yi-Hsuan Chen1and Yi-Fang Chu2
1Institute of Biological Chemistry, Academia Sinica, Taipei, Taiwan;2Institute of Biochemical Science, National Taiwan University,
Taipei, Taiwan;3Department of Medical Research, Veterans General Hospital-Taipei, Taiwan,
Callosellasma rhodostoma (Malayan pitviper) is a monotypic Asian pitviper of medical importance. Three acidic phospholipases A2 (PLA2s) and one basic PLA2-homolog were purified from its venom while 10 cDNAs
encoding distinct PLA2s were cloned from venom glands of a Thailand specimen of this species. Complete
amino-acid sequences of the purified PLA2s were successfully deduced from their cDNA sequences. Among
the six un-translated PLA2cDNAs, two apparently result from recombination of its Lys49-PLA2gene with its
Asp49-PLA2genes. The acidic PLA2s inhibit platelet-aggregation, while the noncatalytic PLA2-homolog induces
local edema. This basic PLA2-homolog contains both Asp49 and other, unusual substitutions unique for the
venom Lys49-PLA2 subtype (e.g. Leu5, Trp6, Asn28 and Arg34). Three-dimensional modelling of the basic
protein revealed a heparin-binding region, and an abnormal calcium-binding pocket, which may explain its low catalytic activity. Oxidation of up to six of its Met residues or coinjection with heparin reduced its edema-inducing activity but methylation of its active site His48 did not. The distinct Arg/Lys-rich and Met-rich region at positions 10±36 of the PLA2homolog presumably are involved in its heparin-binding and the cell
membrane-interference leading to edema and myotoxicity.
Keywords: phospholipase A2; cDNA sequences; modelling; edema; Malayan pitviper.
Calloselasma rhodostoma (Malayan pitviper) is the commonest cause of snakebites in Malaysia and Thailand. The venom causes local effects and systemic bleeding. Its venom has been reported to contain moderate to low levels of phospholipases A2
(PLA2, E.C.3.1.1.4) [1]. However, the structure and activity of
the venom PLA2s of this primitive monotypic pitviper were
not known [2]. As one of the commonest viperid venom components, the PLA2s usually exist in multiple isoforms,
which show distinct pharmacological effects such as neuro-toxicity, myoneuro-toxicity, edema induction, platelet or anti-coagulating activities [3,4]. These isoforms are the results of gene duplication and accelerated evolution and have acquired functional diversity [5]. Ca21 ion plays an essential role in
the catalytic mechanism of the enzyme, however, some of the venom PLA2s have substitution such as Lys49 at the Ca21
binding site (Asp49) and evolved into toxins independent of phospholipid-hydrolysis to exert their functions [4].
Asia is known as the original and the major habitat for present day pitvipers. C. rhodostoma is a primitive monotypic species which contain venom proteins distinct from those of other Asian pitvipers, e.g. its zinc-metalloprotease has been found to have special structure [6]. Thus, characterization of the PLA2s of this venom species at gene and protein
levels would presumably significantly enhance our knowledge about the evolution of the venom PLA2s and the Asian
pitvipers.
In the present study, we cloned at least 10 different PLA2s
from the venom gland cDNA library of C. rhodostoma. Complete sequences of the purified venom PLA2s were inferred
by matching their partial amino-acid sequences and masses against the cDNA-deduced sequences. The major PLA2of this
venom, designated as CRV-W6D49 because of its Trp6 and Asp49 substitutions, was a novel type of basic PLA2-homolog
without detectable hydrolytic activities. Heparin may bind this protein and inhibit its edema-inducing activity. Its 3D structure was predicted by homologous modelling and the abnormal active site was examined. Possible involvement of its exposed methionine residues in the edema-inducing effect was studied by the modification using N-chlorosuccinimide. M A T E R I A L S A N D M E T H O D S
Materials
Venom glands from a single specimen and pooled venom of C. rhodostoma were from the Thai Red-Cross Society (Bangkok). Dipalmitoyl-glycerophosphocholine (diC16PCho) powder was
purchased from Avanti Polar labs. Methyl p-nitrobenzene-sulfonate was a product of Aldrich. Bovine serum albumin, heparin, N-acetyl-methionine, N-chlorosuccinimide, sodium
Correspondence to I.-H. Tsai, PO Box 23-106, Taipei, Taiwan 10798. Fax: 1 886 2 23635038, E-mail: bc201@gate.sinica.edu.tw Abbreviations: CRV, Callosellasma rhodostoma venom; ESI-MS, electrospray ionization mass; NCS, N-chlorosuccinimide; NOB,
4-nitro-3-(octanoyloxy)benzoic acid; PLA2, phospholipase A2; diC16PCho,
dipalmitoyl-glycerophosphocholine . Enzyme: phospholipase A2(E.C.3.1.1.4).
Note: the novel nucleotide sequences for the C. rhodostoma PLA2s have
been submitted to the EMBL databank and are available under accession nos: AF104066 (for K49), AF104073 (for W6hybrid), AF104065 (for W6D49), AF10406870 (for S1E6a-c), AF104067 (for H1E6), AF104071-2 (for R6a-b) and AF104074 (for R6K49), respectively.
(Received 19 June 2000, revised 12 September 2000, accepted 14 September 2000)
deoxycholate, and Triton X-100 were purchased from Sigma Chemical. 4-Nitro-3-(octanoyloxy)benzoic acid (NOB) is a gift from F. J. Ketzdy (The Upjohn Company, Kalamazoo, MI, USA) [7] and is now available from Sigma. Other reagents and buffers are of reagent grade.
Purification of PLA2
The crude venom of C. rhodostoma (10±15 mg) was frac-tionated on a Superdex G-75 column (HR 10/30, Pharmacia, Sweden) in 0.1 m ammonium acetate (pH 4.2), at room temperature. Fractions containing proteins of 14 and 26 ^ 2 kDa were pooled separately and lyophilized. PLA2s
of the venom were further purified by RP-HPLC using a column of C8-silica gel (Vydac, 14 250 mm, 10m) equili-brated with 0.07% aqueous trifluoroacetic acid. The enzymes were eluted with a 28±50% linear gradient of CH3CN
containing 0.07% trifluoroacetic acid. The HPLC conditions usually do not have an adverse effect on the PLA2 activities.
The PLA2 peaks were identified and dried in a
vacuum-centrifuge device (Labconco).
Mass spectrometry and partial sequencing of purified PLA2s
The molecular masses of PLA2s were determined by
electro-spray-ionization mass spectrometry (ESI-MS) using a PE-Sciex API100 mass analyzer (PerkinElmer). Samples were dissolved in 20 mL of 50% CH3CN containing 0.1% acetic acid
and injected into the analyzer for positive-mode analysis. The N-terminal amino-acid sequences of proteins were determined by a gas-phase sequencer (model 477A, Applied Biosystems) with an on-line phenylthiohydantoin amino-acid analyzer using the normal program [8].
Protein quantitation and enzyme assay
The dye-staining method of Bradford was used for quantifi-cation of the purified PLA2 in solution. BSA (1 mg´mL21,
A280 0.56) was used to establish the standard curve [9]. PLA2
activity was measured by pH-stat titration at pH 7.4 and 37 8C. Fresh egg yolk (2.5 g) was dissolved in 100 mL 0.10 m NaCl and mixed thoroughly. Alternatively, diC16PCho (3 mm) was
mixed with 3 mm sodium deoxycholate or 6 mm Triton X-100 and 100 mm NaCl in a glass-Teflon tissue homogenizer to form micelles. With constant stirring, 10 mm CaCl2was added just
before addition of the enzyme to the thermostated substrate [10]. The liberated fatty acid was titrated with 6 mm NaOH using a pH-stat apparatus (Radiometer RTS822, Denmark). The reaction rate was corrected for the nonenzymatic rate. Indirect and direct hemolysis of red blood cells by CRV-W6D49 were performed as described [11] to detect the hydrolase activity towards natural membrane bilayer. The esterase activity towards NOB in Tris buffer (pH 7.4) with 10 mm CaCl2 was
followed spectrophotometrically at 400 nm [7]. Cloning and nucleotide sequencing
The cDNA library of C. rhodostoma venom gland was con-structed by Au et al. [12]. In order to amplify the PLA2cDNA,
PCR was carried out using the cDNA library as the template. A pair of degenerate oligonucleotide primers, with 21 and 18 mixed-base, respectively, were designed based on the highly conserved 50and 30regions in cDNA of other known group-II
snake venom PLA2s [10]. The PCR procedure was performed
with SuperTaq DNA polymerase. A 0.4-kb fragment was
specifically amplified as shown by 1% agarose gel electro-phoresis. After treatment with polynucleotide kinase, the amplified DNA was inserted into the pGEM-T vector (Promega Biotech) and then transformed into Escherichia coli strain JM109. White transformants were picked-up and their DNA extracted and selected by examination of its restriction enzyme products under agarose gel electrophoresis.
The DNA Sequencing System model 373A and the Taq DyeDeoxy terminator cycle sequencing kit (PE Applied Biosystems) were used to determine the cDNA sequences of more than 80 clones obtained in this study [13].
Modification of CRV-W6D49
CRV-W6D49 was oxidized with N-chlorosuccinimide (NCS) which selectively converts methionine residues into methionine sulfoxide derivatives [14,15]. The modification of CRV-W6D49 was in 0.1 m sodium borate (pH 8.5) at room temperature (18 8C). The protein concentration was 75 mm throughout and the NCS concentration was 10-fold that of the protein. After reacted for 1 h, excess N-acetyl methionine was added to halt the reaction. The modified proteins were then purified by RP-HPLC and dried in vacuum. Molecular masses of the oxidized proteins were analyzed by ES1-MS to find out the extent of oxidation. Methylation of the imidazole group or the only His residue (His48) in CRV-W6D49 with methyl p-nitrobenzenesulfonate was carried out by the methods of Verheij et al. [16].
Circular dichroism
CD measurements were carried out on a J715 spectropolari-meter (Jasco, Japan) under constant flushing of nitrogen at 25 8C. Each sample was scanned from 200 to 250 nm for two or three times and superimposable spectra were obtained. The mean residue elipticity [u] was calculated from the mean residue weight.
Binding of PLA2to heparin and its effect on edema-induction
Relative affinity of the basic PLA2for binding to heparin was
studied using a HiTrap heparin column (5 mL, Pharmacia Co.) on an FPLC apparatus [17]. The hind-paw edema of a rat (female Wistar, 200±250 g) was measured with a plethys-mometer (Ugo Basile, model 7150). Extent of hind-paw swelling was calculated based on the paw volume after the injection with native or chemically modified PLA2in contrast
to buffer injected in the control experiment [18]. Platelet inhibition by the acidic PLA2s
Fresh blood was obtained by bleeding the white rabbits through the ear-vein, and the platelet-rich plasma was prepared. The effect of the venom PLA2 on the ADP-induced platelet
aggregation was studied on an aggregometer (Payton, module 600B, Canada) [18].
Computer modelling of the basic PLA2
The atomic coordinates of K49-PLA2 from Bothrops venom
were obtained from the Protein Data Bank (PDB accession no. 1GOD). The amino-acid sequence differs from CRV-W6D49 by 46 residues. These were replaced using the program o [19] and conformations of the substituted side-chains were also adjusted. To assess proper side-chain orientation, another K49-PLA2
no. 1PPA) was superimposed on the working model for reference. There are 19 other crystal structures of PLA2 that
have been refined to higher than 2.0 AÊ resolution, although 15 of these are the PLA2s from bovine pancreas. These 3D
structures were also superimposed on and compared with the working model to provide additional reference for adjustments. The model was subjected to molecular dynamic (MD) and energy minimization (EMin) using xplor [21]. The tempera-ture started at 500 K and increased to 800 K, then decreased in 10- to 25-K steps to 300 K. The model coordinates were restrained by harmonic coefficients to the original in the course of MD but were allowed to shift unrestrained in EMin. The best models were selected using stereochemical criteria as evaluated using procheck [22], and were used for next cycles of MD and EMin.
R E S U L T S A N D D I S C U S S I O N cDNA cloning and nucleotide sequencing
By sequencing more than 80 clones from the C. rhodostoma venom gland cDNA library we identified the clones encoding 10 distinct PLA2s. More than 50% of the clones obtained were
found to encode CRV-W6D49. Only the PLA2s that have been
cloned repetitively and their cDNA nucleotide sequences confirmed at least twice were reported herein. The signal peptides of 16 residues are almost identical or differed by less
than three semiconserved residues in these PLA2 precursors
(data not shown). Their protein sequences were deduced from the nucleotide sequences and aligned in Fig. 1. The isoelectric points were predicted from their sequences, CRV-H1E6, S1E6 (a±c) and R6 are acidic PLA2s, W6D49 and K49 are basic
PLA2homologs, while the basic W6-hybrid and R6K49 are of
hybrid types as their names imply.
Purification and characterization of PLA2s
Gel filtration of the crude venom of C. rhodostoma using a Superdex G75 FPLC column separated the monomeric PLA2s
from the dimeric PLA2s [2]. Using the 26-kDa fraction
obtained from the gel-filtration three dimeric PLA2 showing
high hydrolytic activities toward lecithin substrates were purified by HPLC. They were designated as CRV-H1E6, S1E6a and b, respectively (Fig. 2A). Content for each of these acidic PLA2s was estimated to be about 1±2% (mass/
mass) of the total venom proteins. Their masses (Fig. 2B) are in consistent with those calculated from the cDNA-deduced sequences (Fig. 1) and they inhibited the aggregation of rabbit platelets induced by ADP.
On the other hand, a basic PLA2 homolog (designated as
CRV-W6D49) was purified from the 14-kDa fraction of gel filtration with a yield of about 4% (mass/mass) (Fig. 2A). This basic protein with realtive molecular mass of 13 674 shows no detectable enzyme activities toward egg yolk or the micellar
Fig. 1. Amino-acid sequences and molecular masses of venom PLA2s deduced from the nucleotide sequences of 10 cDNA clones of C. rhodostoma.
The protein sequences were deduced from the nucleotide sequences of PLA2-cDNAs from the venom gland. Single-letter codes of amino acids were used, the
numbering system follows that of Renetseder et al. [24]. Residues identical to those in the top line is denoted with a dot; gaps are marked with hyphens. Abbreviations are: TMV, T. mucrosquamatus venom and DAV, D. acutus venom [29].
substrates containing synthetic diC16PCho [2]. To test whether
it may hydrolyze the phospholipids in model cell membranes, red cell ghosts were prepared from the washed human erythrocytes by hypotonic lysis [23] and incubated with 20 mg´mL21of CRV-W6D49 in Tris buffer (pH 7.4) at 37 8C
for 6 h. Then the ghosts and the venom protein were added to intact erythrocytes and the rate of hemolysis was followed. We found hardly any direct and indirect hemolysis [11] by CRV-W6D49 in the presence of Ca21, while a control
experi-ment using CRV-S1E6a show prominent indirect hemolysis. The results suggested that CRV-W6D49 could not generate lysophospholipid from leaky cell membrane or the unsealed ghosts unlike normal PLA2s.
It is unusual that a D49-PLA2 from pitviper venom does
not show enzymatic activities in vitro using either egg yolk, the micellar dipalmitoyl lecithin, or the ghosts as the substrates (see above). As some of these assays contain phos-pholipids with various head-groups and different fatty-acyl chains, CRV-W6D49 apparently does not behave like a phos-pholipase enzyme by definition. Although the protein appears to have a very low activity towards a pseudo-substrate, NOB, in the presence of Ca21 [7], this may merely reflect an
insufficient catalytic apparatus in CRV-W6D49. Furthermore, when active site His48 of CRV-W6D49 was methylated with methyl p-nitrobenzenesulfonate [16], the relative molecular mass was increased to 13 686 as expected, and the derivative was found to retain the edema-inducing potency of the native form (data not shown). Thus, the results suggest that His48 or an intact catalytic site is not required for the in vivo edematous effect of CRV-W6D49.
Results of ESI-MS of the purified venom proteins show the presence of minor forms of CRV-W6D49 and CRV-S1E6a (Fig. 2B), each with a molecular mass higher than the
respective major form by 16. As both proteins contain methionine residues, the minor forms may be oxidation products occurring during venom storage or purification rather than products of limited hydrolysis as only one N-terminus was detected during the protein sequencing.
Distinctness of venom PLA2s ofC. rhodostoma
It has been hypothesized that Glu6, Asp114, and four aromatic residues at positions 20, 21, 113 and 119 at the surface of Crotaline venom acidic PLA2s are associated with their
anti-platelet effects [24]. These sites are also conserved in the purified acidic PLA2s from C. rhodostoma venom (Fig. 1).
However, we noticed that some common residues in other acidic PLA2s of Crotaline venom, namely Ala40, Gly60, Ser76,
Gly80, Ile104 and Asn125, have been substituted in the acidic C. rhodostoma PLA2s.
The deduced amino-acid sequence of CRV-W6D49 with theoretical pI of 8.3 is 75% similar to those of the K49-PLA2s
(usually with pI of 8.8±9.5) from various pitviper venoms but only , 60% identical to the other D49-PLA2s from the
same venom (Fig. 1). Notably, CRV-W6D49 contains unusual substitutions, e.g. Met2, Asn4, Val11, Pro31, Gly32M, Gly33K and Gly53K and Trp6, Leu5, Lys7, Asn28, Pro31, Val102, Lys53, Glu86 which are uniquely conserved in the venom K49 PLA2s [25±29]. These substitutions together may be
respon-sible for its low catalytic activity and Ca21-independent
myo-toxicity and edema-inducing activities. Notably, CRV-W6D49 (Fig. 1) shares with K49 PLA2s the proposed myotoxic sites
involving Thr13 and Lys residues 7, 16, 78, 79 or 80, 116 and 117 [26,27]. Thus, CRV-W6D49 and K49 PLA2s share
structural and functional similarities.
Fig. 2. Purification and ESI-MS analyses of venom PLA2s. (A) Fractions I and II (Fig. 1)
from gel filtration were further purified by RP-HPLC on a C8 silica gel column. The peaks corresponding to a monomeric basic PLA2
(CRV-W6D49, i.e. W6) and three dimeric acidic PLA2s (CRV-H1E6, S1E6a and S1E6b) were
denoted with arrows, respectively. (B) Molecular masses of the purified PLA2s as determined by
Effect of heparin on the function of basic PLA2homolog
CRV-W6D49 also contains a potential heparin-binding motif of multiple positive charges [17] at residues 33±36, which is flanked by Pro31 and Pro37 forming a proline-bracket [30] (Fig. 1). In K49 PLA2s, heparin-binding motifs are located in
the C-terminal region 115±119 [27,28], which is in fact not far from region 33±36 in the 3D structures or models of these basic PLA2homologs (Fig. 3). The heparin-binding property of
CRV-W6D49 was experimentally shown by their strong adsorption to a heparin-column and elution off the column at a salt concentration above 0.4 m NaCl (data not shown) [17]. By incubating CRV-W6D49 with various concentration of heparin before injection into the paws, the edema was greatly reduced by heparin in a dose-dependent manner (data not shown).
Untranslated PLA2mRNAs
Polymorphic PLA2 genes and mRNAs are usually found in
snake venom glands [25]. Six of the 10 C. rhodostoma PLA2
clones are apparently not translated into venom proteins. Some of them possibly represent spare or regressed genes with a potential to be expressed through the snake natural history. The retention of low-expressing mRNA of venom proteins may be beneficial for snake survival under ever-changing environment. Interestingly, two of the untranslated PLA2s, CRV-K49 and
CRV-R6 (Fig. 1), are structurally almost identical to the abundantly expressed K49-PLA2 [18] and PLA2 III in
Trimeresurus mucrosquamatus venom, respectively (GenBank accession no. X77088). This finding also supports the func-tional significance of both PLA2genes.
Two of the untranslated clones obtained in this study apparently encode hybrid types of PLA2s, which we designated
as CRV-W6-hybrid and CRV-R6K49, respectively (Fig. 1). Both PLA2 hybrids have a C-terminal sequence highly
similar to that of CRV-K49 PLA2 while the N-terminal part
of CRV-W6-hybrid is almost identical to CRV-W6D49 and that of CRV-R6K49 is similar to CRV-R6. Hybrid types of PLA2
mRNAs were also recently found in the venom glands of Agkistrodon halys Pallas [31,32] and Deinagkistrodon acutus [33]. However, hybrid proteins were not isolated in any of the crude venoms, suggesting that they probably are pseudo-genes. Notably, most of the recombination or crossing-over apparently occurred at a highly conserved region near an intron±exon junction encoding amino-acid residues 42±48 of the PLA2s [34].
Computer modelling and structure-activity relationship The 3D model of CRV-W6D49 PLA2 is shown in Fig. 3, and
contains 943 nonhydrogen atoms in 122 amino-acid residues, showing excellent geometry with rmsd of 0.003 AÊ in bond lengths and 0.4468 in bond angles. The distribution of peptide dihedral angles was also improved for this model in which 85.0% of nonglycine and nonproline residues are in the most favored region compared with only 76.6% for the initial model of the K49 PLA2structure. All of the remaining residues were
also found in the allowed regions. The overall coordinate difference between the initial and final models is low, with rmsd of 0.35 AÊ in coordinates for backbone atoms and 1.16 AÊ for the side-chains. Beside the altered regions near residues Met32 and Glu86, maximal deviation in backbone atoms occurred in regions near the C-terminus. Specifically, the peptides Thr 121-Gly 122 and Cys126-Ser127 became more similar to those of bovine pancreatic PLA2 and Agkistrodon
K49-PLA2[20], respectively.
Presumably the binding sites for sulfated aminoglycan contains clusters of positively charges, as shown in the surface-potential map of the model (Fig. 3). The distances between the positively charged residues range between 9.8 and 13.8 AÊ, which probably play essential roles in the binding of CRV-W6D49 to heparin or heparan sulfate [35].
The N-terminal region of CRV-W6D49 is exceptionally rich in Met-residues (at positions 2, 8, 10, 12, 24 and 32). It has been shown that surface Met cluster may serve as a flexible and
Fig. 3. Backbone structure and molecular surface of the CRV-W6D49 model. (Right) Backbone tracing of the CRV-W6D49 3D model in which the N- and C-termini are indicated. Also shown are the side-chains of all positively charged residues, arginines in blue and lysines in cyan. (Left) Electrostatic potential with range of 2 10 to 110 kBT is
shown (blue for positive and red for negative charge as calculated by grasp) [38]. The view is in the direction of putative heparin-binding site and residues speculated to interact with the heparin are labelled. Numbers in italic are the distances in AÊ between the nearest positively charged side-chains.
Fig. 4. Accessible surface area of the N-terminal region in CRV-W6D49 model. The exposed areas for residues 1±35 of the CRV-W6D49 model were calculated using program areaimol of the CCP4 Suite (Collaboration Computational Project [37]) and plotted. Single-letter codes of amino acids are used.
efficient hydrophobic protein±protein interacting site [36]. The local solvent accessibilities of the Met residues were thus analyzed and Met2, Met10, Met24 and Met32 are found to be exposed (Fig. 4).
Figure 5 shows the active site of the CRV-W6D49 model in comparison with that of an ordinary D49-PLA2 from cobra venom. In the normal active site the catalytic calcium is coordinated by seven ligands, two from the carboxyl oxygens of Asp49 side-chain, three from the backbone carbonyl oxygens of residues 28, 30 and 32, and two water molecules [20,24]. Our model shows that only the CO groups of Asn28 and Gly30 are maintained in the calcium binding position of CRV-W6D49. The distances are 2.36 and 2.68 AÊ from the metal ion, respectively, while the latter is not short enough to make proper bonding interaction. The third CO group of our model assumes a completely different arrangement from the corres-ponding residue Gly32 in the active enzyme. Due to the flipping over of peptide Pro31-Met32, the carbonyl oxygen of Met32 was moved 4.98 AÊ farther away from the calcium. The distance is about 7.2 AÊ, unable to interact with the metal ion. The Ca21-binding loop seems no longer effective
in W6D49. Notably, Arg34 in the active site of CRV-W6D49 probably form a stable salt bridge with Asp 49 whose acidic side-chain is rotated 120 AÊ (Fig. 5). This is possibly the reason why CRV-W6D49 does not bind calcium.
Effect of oxidation on the edematous potency of CRV-W6D49 Methionine residues of CRV-W6D49 were modified by NCS in a mild alkaline solution, resultant derivatives were purified by RP-HPLC (Fig. 6). Because the higher degree of Met oxidation in the protein, the less hydrophobic the protein became, the modified CRV-W6D49 was eluted from the column followed the order of derivatives of (MetO)6, (MetO)5and (MetO)4, and
the native protein, as verified by the ESI-MS results (Fig. 6).
The conformation of CRV-W6D49 was apparently altered by the modification as revealed by the CD spectra (Fig. 7). As 51 among its total 122 residues are located in helices according to the 3D model (Fig. 3), the helical content for native CRV-W6D49 is postulated to be 42%. Based on the CD spectra, the helical content for the native, the (MetO)4 and the (MetO)6
derivatives of CRV-W6D49 were calculated to be 41, 31 and 25%, respectively. The b-structures were increased by the modification, while portions of random structures were kept relatively constant (35 ^ 1%). Met2, Met8 and Met12 are present in a-helix 1±12 of the native PLA2and Met24 is at the
end of a-helix 18±22. Oxidation of these Met residues probably loosened the N-terminal helix of the protein. Oxidation of
Fig. 5. Structure of Ca21-binding region of the CRV-W6D49 model.
The model of the active site of CRV-W6D49 is shown in pink and that of an active PLA2from Taiwan cobra venom (PDB accession no. 1POA) is in
blue. Selected residues in our model are labeled with its position number. The calcium ion is shown in green and the three back-bone oxygens that bind the Ca21ion in the active PLA
2are also labeled. The carbonyl oxygen
of Met 32, labeled in red, was moved 5 AÊ further away from the position for calcium binding. This figure was produced using molscript [39] and raster3d [40].
Fig. 6. HPLC purification and ESI-MS of NCS-modified CRV-W6D49. (A) After oxidative modification by NCS, CRV-W6D49 was fractionated by RP-HPLC on a C8 silica gel column. Peaks 1±3 were the differentially oxidized CRV-W6D49, peak 4 was the unmodified protein. (B) Trans-formed ESI-mass spectra of peaks 1 and 3, showing molecular masses which are consistent with those calculated for the (MetO)6 and (MetO)4
four of the Met residues only slightly decreased, but oxidation of all the six Met residues of CRV-W6D49 greatly reduced its edema-inducing activity (Fig. 8). Thus, these Met residues are probably involved in the interactions between CRV-W6D49 and the target cells such as myocytes, neutrophils or polymorphonuclear leukocytes [18,28].
To conclude, this study has elucidated the structures and the activities of four PLA2s present in the venom of a medically
important pitviper C. rhodostoma. Acidic PLA2 variants of
this venom inhibit platelet aggregation and contribute to systemic bleeding, while the basic PLA2 homolog causes
local myonecrosis and edema following the snakebites. This basic and Met-rich PLA2homolog is more similar to the Lys
49-PLA2 homolog than to other Asp 49-PLA2s, and therefore
appears to be a novel type of venom PLA2. The cDNAs
encoding six untranslated C. rhodostoma PLA2s were also
cloned, two of them are almost identical to those for the PLA2s
abundantly present in T. mucrosquamatus venom. Special features of the venom PLA2s molecules of this monotypic
pitviper were highlighted to testify the biodiversity of the pitviper venom PLA2s and shed light on the evolutionary
relationships between the Asian pitvipers. A C K N O W L E D G E M E N T S
We thank our colleague Professor S. H. Wu for providing 3D modelling facilities, and Mr W. Z. Yang and Mr Y. F. Liew for analyses and cloning of some of the PLA2, and Dr Kay-Hooi Khoo for critical reading of the
manuscript. This study is supported by research grants from the National Science Council and Academia Sinica, Taiwan.
R E F E R E N C E S
1. Tan, N.-H. & Ponnudurai, G. (1996) The toxicology of Callose-lasma rhodostoma (Malayan pit viper) venom. J. Toxicol.-Toxin Rev. 15, 1±17.
2. Wang, Y.-M., Liew, Y.-F., Chang, K.-Y. & Tsai, I.-H. (1999) Purification and characterization of the venom phospholipase A2
from Asian monotypic Crotalinae snakes. J. Nat. Toxins 8, 331±340. 3. Danse, J.M., Gasparini, S. & Menez, A. (1997) Molecular biology of snake venom phospholipases A2. In Venom Phospholipase A2
Enzyme: Structure, Function and Mechanism. (Kini, R.M., ed.), pp. 29±71. J. Wiley & Sons, UK.
4. Tsai, I.-H. (1997) Phospholipases A2of Asian snake venoms. J.
Toxicol.-Toxin Rev. 16, 79±114.
5. Ogawa, T., Wakashima, K.I., Nobuhisa, I., Deshimaru, M., Shimo-higashi, Y., Fukumaki, Y., Sakaki, Y., Hattori, S. & Ohno, M. (1996) Accelerated evolution of snake veom phospholipase A2 isozymes
for acquisition of diverse physiological functions. Toxicon 34, 1229±1236.
6. Tsai, I.-H., Wang, Y.-M., Chiang, T.-Y., Chen, Y.-L. & Huang, R.-J. (2000) Purification, cloning and sequence analyses for pro-metallo-protease-disintegrin varients from Deinagistrodon acutus venom and subclassification of the small venom metalloproteases. Eur. J. Biochem. 267, 1359±1367.
7. Cho, W.H., Markowitz, M.A. & Ketzdy, F.J. (1988) A new class of phospholipase A2 substrates: kinetics of the phospholipase A2
catalyzed hydrolysis of 3-(acyloxy)-4-nitrobenzoic acids. J. Am., Chem. Soc. 110, 5166±5171.
8. Hunkapiller, M.W. & Hood, L.E. (1983) Analysis of phenylthio-hydantoins by ultrasensitive gradient high-performance liquid chromatography. Methods Enzymol. 91, 486±493.
9. Bradford, M.M. (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72, 248±254.
10. Tsai, I.-H., Lu, P.-J., Wang, Y.-M., Ho, C.-L. & Liaw, L.-L. (1995) Molecular cloning and characterization of a neurotoxic phospho-lipase A2 from the venom of Taiwan habu (Trimeresurus
mucrosquamatus). Biochem. J. 311, 895±900.
11. Jeng, T.W., Hendon, R.A. & Fraenkel-Conrat, H. (1978) Search for relationships among the hemolytic, phospholipolytic and neurotoxic activities of snake venom. Proc. Natl Acad. Sci. USA 75, 600±604.
12. Au, L.-C., Chou, J.-S., Chang, K.-J., Teh, G.-W. & Lin, S.-B. (1993) Nucleotide sequence of a full-length cDNA encoding a common precursor of platelet aggregation inhibitor and hemorrhagic protein from Calloselasma rhodostoma venom. Biochim. Biophys. Acta 1173, 243±245.
13. Maniatis, T., Fritsch, E.F. & Sambrook, J. (1989) Molecular Cloning: a Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York.
14. Shechter, Y., Burstein, Y. & Gertler, A. (1977) Effect of oxidation of Fig. 7. CD spectra of the native and NCS-modified CRV-W6D49
derivatives. The spectra of the native (Ð), the NCS-oxidized CRV-W6D49 (MetO)6(-´-´-) and (MetO)4(---) in 12 mm Tris/HCl buffer, pH 7.4, with
2 mm CaCl2at 25 8C.
Fig. 8. Effect of chemical modification on the edema-inducing activity of CRV-W6D49. Time course of the edema induced by injecting the native (X), the NCS-oxidized CRV-W6D49 derivatives (MetO)6(A) and (MetO)4
(W), and methylated-CRV-W6D49 to rat paws, and that of a control experiment (---), with injection of the same volume (100 mL) of sterile NaCl/Pi. The experiments were repeated twice and similar results were
methionine residues in chicken ovomucoid on its inhibitory activities against trypsin, chymotrypsin, and elastase. Biochemistry 16, 992±997.
15. Boudier, C. & Bieth, J.G. (1994) Oxidized mucus proteinase inhibitor: a fairly potent neutrophil elastase inhibitor. Biochem. J. 303, 61±68.
16. Verheij, H.M., Volwerk, J.J., Jansen, E.H., Puyk, W.C., Dijkstra, B.W., Drenth, J. & de Haas, G.H. (1980) Methylation of histidine-48 in pancreatic phospholipase A2. Role of histidine and calcium ion in
the catalytic mechanism. Biochemistry 19, 743±750.
17. Vyas, A.A., Pan, J.-J., Patel, H.V., Vyas, K.A., Chiang, C.-M., Sheu, Y.-C., Hwang, J.-K. & Wu, W.-G. (1997) Analysis of binding of cobra cardiotoxins to heparin reveals a new b-sheet heparin-binding structural motif. J. Biol. Chem. 272, 9661±9670.
18. Liu, C.-S., Chen, J.-M., Chang, C.-H., Chen, S.-W., Teng, C.-M. & Tsai, I.-H. (1991) The amino acid sequence and properties of an edema-inducing Lys-49 phospholipase A2homolog from the venom
of Trimeresurus mucrosquamatus. Biochim. Biophys. Acta 1077, 362±370.
19. Jones, T.A., Zou, J.Y., Cowan, S.W. & Kjeldgaard, M. (1991) Improved methods for building protein models in electron density maps and the location of errors in these models. Acta Crystallogr. A47, 110±119. 20. Scott, D.L., Achari, A., Vidal, J.C. & Sigler, P.B. (1992) Crystallo-graphic and biochemical studies of the (inactive) Lys-49 phos-pholipase A2from the venom of Agkistridon piscivorus piscirorus.
J. Biol. Chem. 267, 22645±22657.
21. Brunger, A.T. (1992) X-PLOR: a System for X-Ray Crystallography and NMR. Yale University Press, New Haven, USA.
22. Laskowski, R.A., MacArthur, M.W., Moss, D.S. & Thronton, J.M. (1993) PROCHECK: a program to check the stereochemical quality of protein structures. J. Appl. Cryst. 26, 283±291.
23. Tsai, I.H., Murthy, S.N.P. & Steck, T.L. (1982) Effect of red cell membrane binding on the catalytic activity of glyceraldehyde-3-phosphate dehydrogenase. J. Biol. Chem. 257, 1438±1442. 24. Wang, X.-Q., Yang, J., Gui, L.-L., Lin, Z.-J., Chen, Y.-C. & Zhou, Y.-C.
(1996) Crystal structure of an acidic phospholipase A2 from the
venom of Agkistrodon halys Pallas at 2.0 AÊ resolution. J. Mol. Biol. 5, 669±676.
25. Ohno, M., Menez, R., Ogawa, T., Danse, J.M., Shimohigashi, Y., Fromen, C., Ducancel, F., Zinn-justin, S., Le Do, M.H., Boulain, J.-C., Tamiya, T. & Menez, A. (1998) Molecular evolution of snake toxins: is the functional diversity of snake toxins associated with a mechanism of accelerated evolution? Prog. Nucleic Acid Res. Mol. Biol. 59, 307±364.
26. Selistre de Araujo, H.S., White, S.P. & Ownby, C.L. (1996) Sequence analysis of Lys 49 phospholipase A2myotoxins: a highly conserved
class of proteins. Toxicon 34, 1237±1242.
27. Ward, R.J., Alves, A.R., Neto, J.K., Arni, R.K. & Casari, G. (1998) A sequence space analysis of Lys-49 phospholipases A2: clues
towards identification of residues involved in a novel mechanism of membrane damage and in myotoxicity. Prot. Engineer. 11, 285±294.
28. Lomonte, B., Moreno, E., Tarkowski, A., Hanson, L.A. & Maccarana, M. (1994) Neutralizing interaction between heparins and myotoxin II, a lysine 49 phospholipase A2 from Bothrops asper snake venom.
J. Biol. Chem. 269, 29867±29873.
29. Wang, Y.-M., Wang, J.-H., Pan, F.-M. & Tsai, I.-H. (1996) Lys-49 phospholipase A2homologs from venoms of Deinagkistrodon acutus
and Trimeresurus mucrosquamatus have identical protein sequence. Toxicon 34, 485±489.
30. Kini, R.M. & Evan, H.J. (1995) A novel approach to the design of potent bioactive peptides by incorporation of proline brackets: antiplatelet effects of Arg-Gly-Asp peptides. FEBS Lett. 375, 15±17. 31. Pan, H., Liu, X.-L., Ou-Yang, L.-L., Yang, G.-Z., Zhou, Y.-C., Li, Z.-P. & Wu, X.-F. (1998) Diversity of cDNAs encoding phospholipase A2
from Agkistrodon halys Pallas venom, and its expression in E. coli. Toxicon 36, 1155±1163.
32. Liu, X.-L., Pan, H., Yang, G.-Z., Wu, S.-F. & Zhou, Y.-C. (1999) Cloning, expression and biochemical characterization of a basic-acidic hybrid phospholipase A2-II from Agkistrodon halys Pallas.
Biochim. Biophys. Acta 1431, 157±165.
33. Liu, X.-L., Pan, H., Yang, G.-Z., Wu, S.-F. & Zhou, Y.-C. (1999) Cloning and sequencing of genes encoding phospholipase A2
from Agkistrodon acutus. Acta Biochim. Biophys. Sinica (China) 31, 41±45.
34. John, T.R., Smith, L.A. & Kaiser, I.I. (1994) Genomic sequences encoding the acidic and basic subunits of Mojave toxin: unusually high sequence identity of non-coding regions. Gene 139, 229. 35. Lindahl, U., Marion, K.G. & Kjellen, L. (1998) Regulated diversity of
heparan sulfate. J. Biol. Chem. 273, 24979±24982.
36. Gellman, S.H. (1991) On the role of methionine residues in the sequence-independent recognition of nonpolar protein surfaces. Biochemistry 30, 6633±6636.
37. Collaborative Computational Project, Number, 4. (1994) The CCP4 Suite: programs for protein crystallography. Acta Crystallogr. D50, 760±763.
38. Nicholls, A., Sharp, K.A. & Honig, B. (1991) Protein folding and association: insights from the interfacial and thermodynamic properties of hydrocarbons. Proteins 11, 281±296.
39. Kraulis, P.J. (1991) MOLSCRIPT: a program to produce both detailed and schematic plots of protein structures. J. Appl. Cryst. 24, 946±950.
40. Merrit, E.A. & Murphy, M.E.P. (1994) Raster3D, Version 2.0: a program for photorealistic molecular graphics. Acta Crystallogr. D50, 869±873.
