Primary sequence determination of the most basic myonecrotic phospholipase A2 from the venom of Vipera russelli

~1-~I01(~~)~-.S

Toxlmn, Vol . 32. No. 6, pp. é65-673, 1994 ELevier Science Ltd Rimed in Gnat B~tain

PRIMARY SEQUENCE DETERMINATION OF THE MOST BASIC MYONECROTIC PHOSPHOLIPASE AZ FROM THE VENOM OF YIPERA RUSSELLI VEBRABASAPPA T. Gownw, Jwl~s SCFIII~T and JoxN L. MII~DLSBROOK * Toxinology Division, United States Army Medical Research Institute of Infectious Diseases, Frederick, MD 21701-5011, U.S .A. (Recei°ed 10 September 1993 ; accepted S January 1994)

V. T. GOWDw, J. $CFII1~T and J . L. M)DDLEBROOK . Primary sequence determination of the most basic myonecrotic phospholipase AZ from the venom of Vipera russelli. Toxicon 32, 665-673, 1994.-The most basic phospholipase AZ (PLAZ), VRV-PL-VIIIa, was purified from (Sri Lankan) Vipera russelli venom. It is a major component of the venom, contributing over 40% to the whole venom PLAZ activity . The purity of VRV-PL-VIIIa was ascertained by electrophoresis and by reverse phase high-pressure liquid-chromatography (RP-HPLC). VRV-PL-VIIIa had an apparent mol. wt of 13,000 and was a single polypeptide. The protein was reduced, pyridylethylated and subjected to sequence analysis. The N-terminal amino acid sequence was established up to the 39th residue. Pyridylethylated VRV-PL-VIIIa was digested with endoprotease Glu-C, and several peptides were purified by RP-HPLC; six purified peptides were sequenced. The sequence of the C-terminal was established by sequencing a CNBr-produced peptide purified by RP-HPLC. Several peptides were also generated by digestion with endoprotease Asp-N . Two peptides were sequenced to obtain overlapping regions. The complete structure was deduced from sequences of overlapping peptides and through homology with other group II PLAZ sequences. Sequence homology was greatest with ammodytozin A: 99 amino acid residues out of 121 occurred in identical positions. Myotozin III of Bothrops aspen showed 73% homology, 89 out of 121 residues. In agreement with the sequence data, polyclonal antiserum against VRV-PL-VIIIa cross-reacted in ELISA with ammodytoxin A and, to a lesser extent, with caudoxin .

INTRODUCTION

Mwxx phospholipases AZ (phosphatidyl choline 2-acylhydrolase, EC 3.1 .1 .4) (PLAZ) found in snake venons induce more than one pharmacological effect in animals (IWANAGA and $U2:URI, 1979; Ros$xs>~G, 1979, 1986), while their counterparts from mammals are without observable effects. PLAZ are divided into two structural groups : class I, similar to pancreatic PLAZ , are typically found in elapid venoms, whereas class II PLAZ are usually " Author to whom correspondence should be addressed. 665

666

V . T . GOWDA et al.

found in crotalid and viperid venoms . Over 100 PLA Z analogs have been sequenced to date. Because of the striking sequence homology around the catalytic site and relatively small differences in the sequences occwring near the C-terminus, it has been difficult to explain the structure-function relationship of these molecules . Three PLAZ have been pwified and characterized from Russell's viper venom: VRV-PLVI (Ln~ of 3 .5 mg/kg) by ViswnxnTH et al. (1988) ; VRV-PL-VIIIa (Ln~ of 5 .3 mg/kg) by KASTURi and Gowns (1989); and VRV-PL-V (Ln~ of 1 .8 mg/kg) by .TAYANTHI et al. (1989) . These isoforms of PLA Z induce newotoxicity, edema and anticoagulant effects in experimental models. VRV-PL-V and VRV-PL-VIIIa are myonecrotic, while all three enzymes induce hemorrhage in lung and liver. Rabbit antiserum prepared against VRV-PL-V/VRV-PL-VIIIa neutralized the lethal toxicity and neurotoxicity of all three toxic enzymes, but the PLAZ activity was not affected by the antiserum when phosphatidyl choline was used as substrate . Similarly, three PLA Z toxins from the venom of Vipera ammodytes ammodytes have been characterized and sequenced (GUBENSEK et al., 1980; RITONJA and GUBENSEK, 1985; RITONJA et al., 1986; KRIZAJ et al., 1989) . These toxins have extensive sequence homology among themselves, with mutations occurring in only five positions localized between residues 115 and 128 . Three amino acid changes from ammodytoxin A to ammodytoxin B at 115 (Y-H), 118 (R-M) and 119 (N-Y) led to a 30-fold increase in the Ln~ . Similarly, two amino acid changes from ammodytoxin A to ammodytoxin C at 124 (F-I) and 128 (K-E) increased the t-n~ by 18-fold . Because of their biological properties, it would be interesting to know the primary sequences of the three PLA Z isoforms from Russell's viper venom . In this paper, we describe the primary sequence of VRV-PL-VIIIa, the most basic PLAZ found in Russell's viper venom, and compare the sequence homology with other snake venom PLAZ . MATERIALS AND METHODS Chemicals and reagents Russell's viper venom (Sri Lankan) was a kind gift from Professor Dtsrxtctt Mees (Frankfurt, Germany) . CM-Sephadex C-25, and Sephadex-G-50 were obtained from Pharmacia Inc . (Piscataway, NJ, U.S .A .) . Endoproteinases, Glu-C (protease V8 from Staphylococcws aureus) and Asp-N (Pseudomorurs Jragi) were of sequencing grade from 13oehringer Mannheim (Germany) . Trifluoroacetic acid (TFA), sequence grade, was purchased from Pierce Chemical Company (Rockford, IL, U .S.A.). Other chemicals used in the study were of analytical grade from commercial sources. VRV-PLrVIIIa was purified from Vipera russelli venom as described by K~sivw and Gowne (1989). Purity of the isolated PLA Z was checked by polyacrylamide gel electrophoresis (with and without sodium dodecylsulfate) and reverse phase high-pressure liquid~hromatography (RP-HPLC) . Phospholipase assay Assays were carried out by pH slat-titralions . Phosphatidyl choline was the substrate with a 2 : I molar ratio of Triton X-100 :phospholipid in the reaction mixture . Released fatty acids were titrated at pH 8 .0 with NaOH at room temperature with a Radiometer apparatus (Aran and Ketsete, 1985). Preparation of antiserum Rabbit antiserum against VRV-PL-VIIIa was prepared by immunizing a New Zealand white rabbit, essentially as described previously (Mtnnt.Eexootc and K~tsEtt, 1989) . Antibodies in the serum reacted with toxin positively as determined by standard ELISA (Mtnnt.eettooic and K~a, 1989) . Chemical modifications VRV-PL-VIIIa was dissolved in 6 .0 M guanidinium hydrochloride, 0 .25 M Tris, pH 8.5, 5 mM EDTA . Reduction with 0 .085 M ß-mercaptcethanol was carried out under nitrogen at room temperature for 6 hr . Subsequently, a 1 .5-fold molar excess (over sulfhydryl groups) of 4vinyl-pyridine was added . ARer 30 min, the sample was dialyzed against 1 % acetic acid at 4°C.

Russell's Viper Myonecrotic PLAZ Sequence

667

Protease and chemical digestions

Digestion with endoprotease Glu-C was performed at room temperature for 6 hr in 0.2 M ammonium acetate, pH 4.0, with reduced pyridylethylated VRV-PL-VIIIa in a molar ratio of 25 :1 (substrate :protease). The reaction was stopped by acidification with TFA to pH 2.0-2 .5 . Digestion with cndoproteinase Asp-N was carried out after adjusting the pH of the pyridylethylated protein to 8.0 with Tris (free base). Reaction at 37°C for 20 hr wascarried out at a molar ratio of substrate to protease of 100 :1 . Insolublepolypeptide was present throughout the digestion. The mixture was centrifuged, and the supernatant was acidified to pH 2.0-2 .5 and fractionated by RP-HPLC. Reaction with CNBr

Pyridylethylated VRV-PL-VIIIa (0.80 mg) was dissolved in 300 pl of CNBr (0 .37 M in 70% vJv formic acid). The solution was incubated at room temperature for 24 hr, then dried with a stream of dry nitrogen . The dried sample was dissolved in distilled water and lyophylized three times. Sequence analyses

Automated degradation was done in a model 470A amino acid sequencer from Applied Biosystems (Foster City, CA, U.S .A.) . Residues were identified with a model 120A liquid chromatograph from the same manufacturer. Purification of peptides

RP-HPLC employed various gradients of acetonitrile in 0.09% TFA (see figure legends for details). The column was a Hi-Pore RP-318 (C-18) from Biorad Laboratories (Richmond, CA, U.S .A.) . All other equipment was from Waters Division of Millipore Corp . (Milford, MA, U.S.A.). RESULTS

Pyridylethylated VRV-PL-VIIIa (RPE-VRV-PL-VIIIa) gave a sharp single peak on RP HPLC (data not shown). When this material was subjected to automated sequencing, the N-terminal sequence of 39 residues was established : SLLEFGKMILEETGKLAIPSYSSYGCYCGWGGKGTPKDA . The remaining structure of VRV-PL-VIIIa (Fig. 1) was assembled from sequences of peptides produced by digestion with protease V8, cyanogen bromide or protease Asp-N . Fractionation of the V8 protease digest of VRV-PL-VIIIa on reverse phase chromatography yielded 19 peaks; many contained more than one peptide (Fig. 2a) . Several peptides were purified by rechromatography of isolated peaks and subjected to sequencing (designated with circles in Fig . 2). Six peptides were found to provide support for the overlapping sequence and confirmed the N-terminal sequence . These were V8-II, V8-IV, V8-VI, V8-VII (2 peaks upon rechromatography) and V8-XVII, as shown in Fig . 2. From the above, it is evident that V8 protease also cleaved a bond between glycine and pyridylethylated cysteine . Since we had already determined the sequence of the first peptide, the sequence of the second peptide could be deduced . The sequence of the last three residues in the peptide in peak XVII (C-terminal peptide) was not certain, because of small peaks on the detector chart . To firmly establish this part of the sequence, RPE-VRV-PL-VIIIa was subjected to degradation with cyanogen bromide and the peptides purified by RP-HPLC (Fig. 2b) . Peak `b' revealed the sequence shown in Fig . 1, confirming the carboxy-terminal sequence of the toxin . When the peptides were arranged by comparison with the sequence of ammodytoxin A, 80% homology was observed, although there was a sequence gap of five residues missing in VRV-PL-VIIIa . In consequence, it was necessary to obtain a peptide which included this region . Therefore, the protein was digested with Asp-N, which gave 11 well-separated

668

V. T. GOWDA et al.

peaks on RP-HPLC (Fig. 2c). Several peptides were sequenced, two of which helped complete the missing sequence of VRV-PL-VIIIa (Fig. 1). Though described as specific for aspartic acid, in this case, the protease Asp-N cleaved on the amino-side of glutamic acid. With this information, the complete sequence of the toxin was deduced and is shown in Fig. 1 along with the placement of cleavage sites . The sequence is not completely unambiguous in that we did not generate a peptide spanning the peptides abuting amino acids 52-53 . Nevertheless, the high degree of homology in this region with other PLAZ supports the sequence as shown. The amino acid composition ofVRV-PL-VIIIa calculated from the sequence is: Asp,, Asn b , Thr s , Ser,, Glue , Pro s , Gly  , Alab, Val,, Met Z , Ilex, Lew,, Tyr9, Phe4 , Lys,3 , His,, Args, Cys,4 , Gln,, and Trp, . Finally, using the consensus numbering system for class II phospholipases Az (RENETSF.DER et al., 1985), the sequence for VRV-PL-VIIIa was compared with ammodytoxin A, as shown in Fig. 3. Antiserum raised against purified VRV-PL-VIIIa showed strong cross-reactions by ELISA with ammodytoxin A and caudoxin and to a slightly lesser degree with crotoxin (Fig. 4). The same antiserum did not react in ELISA with the elapid toxins ß-bungarotoxin or notexin (data not shown).

1 2 3 9 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 S L L E F G K M I L E E T G K L A I P S ______________________direet sequence-_____________________ ____________V8(VII) -____ __V8(IV)-> 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 90 Y S S Y G C Y C G W G G K G T P K D A T

91 42 43 44 95 46 47 98 49 50 51 52 53 54 55 56 57 58 59 60 D R C C F V H D C C Y G N L P D C N P K ____ V8(VI)-_______________________> ______________________ 61 62 63 69 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 S D R Y K Y K R V N G A I V C E K G T S ___________V8(VI)_____________________________~ _______________________________________p,sP(III)_________ 81 82 83 89 85 86 87 88 89 90 91 92 93 99 95 96 97 98 99 100 C E N R I C E C D K A A A I C F R Q N L ___V8(II)-__~ _____________________________________ _> ___p,sp (I) ______> 10 11 12 13 14 15 16 17 18 19 20 21 1 2 3 9 5 6 7 8 9 N T Y S K K Y M L Y P D F L C K G E L K C __________________________V8(XVII)___________________________> ______________CnBr digest(b)-_________> FIG. I. DFDUCPD AdI1N0 ACm s~uexcE of VRV-PLrVIIIa . The sequence of VRV-PL-VIIIa is shown along with the cleavage sites for the protease or chemical reactions employed to produce peptides .

Russell's Viper Myonecrotic PLA Z Sequence

669

DISCUSSION

Over 100 phospholipases and analogs have been sequenced, more than 60 class I and 36 class II from snake venoms with the remainder from mammalian pancreas. A majority of the class II PLAZ have serine as their N-terminal amino acid. BABU and %OWDw (1991) reported that VRV-PL-VIIIa has serine as its N-terminal amino acid and its chemical modification with 4-NN-dimethyl amino azo benzene-4'-isothiocyanate results in the loss of both enzymatic activity and pharmacological effects. Most of class II PLAZ have 120-123 amino acids in their primary sequences, although the consensus numbering system is based on 138 residues. VRV-PL-VIIIa has 121 amino acid residues, with a glutamic acid missing, third from the C-terminal end, as compared to ammodytoxin A. Ninety-nine out of 121 amino acid residues occur in identical positions in ammodytoxin A and VRV-PLVIIIa (depicted in bold in Fig. 3). As in ammodytoxin, residue 4 of VRV-PL-VIIIa is glutamic acid, instead of the class II invariant glutamine. There are 19 basic amino acid residues, unevenly distributed in the sequence of the molecule. Nearly two-thirds of them occur towards the C-terminal and one-third towards the N-terminal. In contrast, the 14 acidic amino acids are more evenly spaced throughout the sequence, except between residues 88 and 102 where there are four . There are 21 hydroxyl containing amino acids which are also evenly spaced. Bulky hydrophobic side-chain containing amino acids such as leucine, isoleucine and phenylalanine are concentrated in the first 20 residues (N-tenminus) and the last 30 residues (C-terminus) . Therefore, the C-terminal sequence of VRV-PL-VIIIa is rich in both basic and hydrophobic amino acids. The positions of all cysteines are conserved as found in other class II PLAZ . A striking sequence, rich in basic and hydroxy amino acids, is found between residues 68 and 78 in VRV-PL-VIIIa and ammodytoxin A, B or C. In addition, there is a partly similar sequence (69-77) in Naja nigricollis basic PLAZ , nigexine (N. nigricollis), and Naja mossambica mossambica (CMIII) . VRV-PIrVIIIa : Ammadytoxins (A, B, C) : Naja nigricollis (basic) : Nigexine (N. nlgrtcollis) : CM-III (N. moss. moss.) :

68 78 PKSDRYKYKR PKTDRYKYHR 69 77 PYLTLYKYK PYFTLYKYK PYFTLYKYK

Polylysines form random structures at neutral pH, because the closely spaced, positively charged side-chains repel each other strongly and disrupt the tendency to form interchain hydrogen bonds. If the side-chains are bulky or have like charges, pleated sheets cannot exist because of unfavorable interactions . Hence, a flexible structure probably results from the above sequences, and this region may be important for protei~protein interactions . VRV-PL-VIIIa, N. nigricollis basic PLAZ and CM-III of N. moss . moss . each have anticoagulant activity (KASTURI and GOWDw, 1989 ; Lip et al., 1977; RowwN et al., 1991), while VRV-PL-VIIIa and N. nigricollis PLAZ are both myonecrotic. Knvi and Evwxs (1987) predicted that sequence 54-77 is important in the anticoagulant properties of snake venom PLAZ . It is possible that this homologous sequence might represent a domain important for one or both of the toxic effects resulting from protein-protein interactions . The C-terminal region between residues 80 and 110 is predicted to be important for the expression of presynaptic neurotoxic activity in PLAZ (Kixt and IWANAGA, 1986). This region is rich in bulky hydrophobic residues in neurotoxic PLAZ like N. nigricollis basic PLAZ , textilotoxin (Pseudonaja textilis) subunits and ammodytoxin A. The high neurotoxrox ulb-B

670

V . T. GOWDA et al.

icity of ammodytoxin A is attributed to residues at 119-Y, 122-R, 123-N, 127-F and 131-K (Kxizw et al., 1989). While VRV-PL-VIIIa possesses a significant homology to ammodytoxin A in this region, there are notable substitutions. Chemically unrelated amino acids are substituted at positions 119 (Y-K), 122 (R-M), 123, (N-L) and 131 (K-G), while phenylalanine was retained at position 127. Note that two substitutions in ammodytoxin

o.e~

O.B ~

0 .2

i 10

i 15

i 20

25

~ Fletention Time (min) FtG . 2(a),(b)

30

85

40

46

Russell's Viper Myonecrotic PLA= Sequence

67l

Ratendon Time (min)

FIa. 2(c) FIa. 2. RP-HPLC FRACTIONA170N OF PEPTn)E4 B~ F~cr:n BY DIGE4ITON OF RPE-VRV-PIrVIIIa wlrx Pleo~"~ VS (a), CNBr (b) or protease Asp-N (c). For each separation, solvent A was 0.09% TFA, solvent B was 70% acetonitsile in 0.09% TFA, flow was 1 .0 ml/min, temperature was maintained at 30°C and efliuent was monitored at 210 nm. Initially, the column was equilibrated with 12% solvent H, then the sample was applied and run for 2.5 min. Next, the following gradients were used: protease Glu-C, a linear gradient to 50% B at 41 min; CNBr, stepwise increase of solvent B to 28%,followed by a linear gradient to 42% at 45 min; protease Asp-N, a linear gradient to 55% solvent B at 47 min.

C vis-à-vis ammodytoxin A increased its Lasso by 18-fold compared to ammodytoxin A and, similarly, three substitutions in ammodytoxin B increased its LDSp by 30-fold. A further substitution (four out of five residues) in VRV-PL-VIIIa led to an increase of LD Sp by

Ammodytoxin A : VRV-PL-Vllls :

20 30 40 1 10 SLL)cFGN1!!dâGKISIL-EBTGICLAIPSYSSYGCYCGWGGICGTPICDA

70 80 90 50 60 TDRCCI'VHDCCYGNLP ---DC ---- -SPICTDRYICYHR--ENG7II-VCGKG TDRCC1rVHDCCYGäLP - --DC --- --NPICSDRYICYKR--VNG7II-VCEKG 110 120 130 138 100 T$-C>sNRICâCDR7171AIC~RKâLKTY-NYIYRNYPDr-LC1CiCESElCC T$-CENRIC8CDK711171IC1'RQNLNTY-SKKYMLYPDr-LCICGZL 1CC FIa. 3. Coelp~IttsoN OF THE ~uewcrs of VRV-PL-VIIIa ertD wr~oDYroxnv A. The sequences of VRV-PIrVIIIA and ammodytoxin A are aligned and spaced by the consensus numbering system for PLA2 developed by R>~Dr:B et a! . (1985) .

Ammydoxin Caudodn Crolo~dn VRV-PL-VIII

Fta. 4. ELISA cRass-xEACrm~rr oF xAeHrr ANr13Eteu~r rtereAxeD AaArNST VRV-PIrVIIIa wtrx AMMODYTOXIN A, CAUDOXIN AND CROi'OXIN.

270-fold . Therefore, the low neurotoxicity and high LDsp (5.4 mg/kg) for VRV-PL-VIIIa compared to high neurotoxicity and low LD sp (0.02 mg/kg) for atnmodytoxin A are probably due to the above cited amino acid changes. However, VRV-PL-VIIIa has three more substitutions and one deletion located between residues 118 and 136, and the effects of these changes on neurotoxicity cannot be ruled out at this time . Nonetheless, our data support the concept that the toxicity of snake venom PLAZ is determined not by any single amino acid residue, but by a specific sequence of residues forming a domain . The immunological cross-reactivity of antiserum prepared against VRV-PL-VIIIa with ammodytoxin A, crotoxin and caudoxin almost certainly results from the extensive sequence homology between the toxins. Thus, there should be common conformational and/or linear epitopes recognized by the polyclonal VRV-PL-VIIIa specific antiserum. It has been shown that PLAZ toxins can be grouped into three serogroups (MmDL>~BROOx and KAISeR, 1989), and following those criteria, VRV-PL-VIIIa falls into the viperid serotype subgroup, which is hardly unexpected . In contrast, none of 22 monoclonal antibodies against VRV-PL-VIIIa appeared to recognize a linear epitope. These results may mean that most of the cross-reacting antibodies in our rabbit antiserum are directed to conformational epitopes . Acknowledgement-T. VrmQ~-g~-gwpA GOWDA acknowled~s the award of a Senior Research Associateship by the National Research Council. The views of the authors do not purport to reflect the positions of the Department of the Army or the Department of Defense (pare. 4-3, AR 360-5). REFERENCES AIxD, S. D. and ICAL9QR, I. I. (1985) Comparative studies on three rattlesnake toxins . Toxicon 23, 361-374. HAeu, A. S. and GOwDA, T. V. (1991) Effects of chemical modification on enzymatic and toxicological properties of phoapholipese A~ from Naja raja aaja and Vipers russelli snake venoms . Toxicon 29, 1251-1262. Gr~.~arr- F., RrrONZA, A., ZUPAN, J. and Ttrntc, V. (1980) Basic proteins of Vipers anvnodytes venom: studies of structure and function . Period. Biol. 82, 44347. IWANAaA, S. and Suzuru, T. (1979) Enzymes in snake venoms . In: Handbook of Experimental Pharmacology, Vol. 52, Snake Venovnr, pp . 61-144 (L~, C. Y., Ed .). Berlin : Springer. JAYAN1éü, G. P., ICASruAr, S. and GOWDA, T. V. (1989) Dissociation of catalytic activity and neurotoxicity of a basic phoapholipase A from Russell's viper (Vipers rurselli) venom. Toxicon 27, 875-885. KASruxt, S. and GowDA, T. V. (1989) Purification and characterization of a major phosphoGpase A= from Russell's viper (Vipers russelli ) venom. Toxicon 27, 229-237.

Russell's Viper Myonecrotic PLA2 Sequence

673

Ktxt, R . M . and Evexs, H. J . (1987) Structure-function relationship of phospholipases : the anticoagulant region of phospholipase A z . J. biol. Chem. 262, 14,402-14,407 . Kttvt, R . M . and Iwetvea~, S . (1986) Structure-function relationships of phospholipasea I : Prediction of presynaptic neurotoxicity . Toxicon 24, 527-541 . Kxtzet, L, Tuts¢, D ., Rtrotvt~, A . and Gtmrnw+é- F. (1989) Primary structure of ammodytoxin C further reveals the toxic site of ammodytoxin. Biochem. Biophys. Acta 999, 198-202 . Lt~, C . Y ., Hoo, C. L. and E~, D . (1977) Cardiotoxin like action of a basic phospholipase A z from Naja nigricollis venom . Toxicon 15, 355-356 . Mtnut.t:enootc, J . L. and Kets~, I . I . (1989) Immunological relationships of phospholipase AZ neurotoxins from snake venoms . Toxicon 27, 965-977 . Rav»t=tt, R., BRUME, S ., Dttttsrtu, B . W ., Dttmvrx, J . and Stat.> :a, P . H. (1985) A comparison of the crystal structures of phoapholipase A2 from bovine pancreas and Crotales atrox venom. J. blot. Chem. 260, 11,627-11,634 . Rtrota~, A. and GUSExstac, F . (1985) Ammodytoxin A, a highly lethal phospholipase A= from Vipera ammodytes ammodytes venom . Biochem. Biophys. Acta 828, 306-312 . Rtrorue, A ., Mectu.Ero, T. W ., TvRtc, D . and GueetvsEtr, F . (1986) Amino acid sequence of ammodytozin B partially reveals the location of the site of toxicity of ammodytoxins . Biol. Chem. Hoppe-Seyler 367, 919-923 . Rosentta ;tta, P. (1979) Enzymes in snake venoms . In : Handbook ojExperimental Pharmacology, Vol. 52, Snake Venomr, pp . 404-434 (L~, C . Y., Ed.) . Herlin : Springer . RostareERC, P. (1986) The relationship betwcen enzymatic activity and pharmacological properties of phospholipeses in natural poisons . In : Natural Toxins, Animal, Plant and Microbial, pp . 128-166 (HARRIS, J . B., Ed .). Oxford: Oxford Science Publication . RowAx, E. G., HARVEV, A . L . and M1nvEZ, A . (1991) Neuromuscular effects of nigexine, a basic phospholipase Az from Naja nigricollis venom. Toxicon 29, 371-374. VISHWANATH, B . S., Ktxl, R. M . and GownA, T. V . (1988) Purification and partial biochemical characterization of an edema inducing phospholipase AZ from Vipera russelli (Russell's viper) venom. Toxicon 26, 713-720.