Effects of chemical modifications of Pa-11, a phospholipase A2 from the venom of Australian king brown snake (Pseudechis australis), on its biological activities

Tasicarr Vol . 28, No . 1, pp . 107-117, 1990. Printed in Great Britain .

0041-0101/90 $3.00 + .00

m 1989 Pergamon Press pic

EFFECTS OF CHEMICAL MODIFICATIONS OF PA-11, A PHOSPHOLIPASE A2 FROM THE VENOM OF AUSTRALIAN KING BROWN SNAKE (PSEUDECHIS A USTRALIS), ON ITS BIOLOGICAL ACTIVITIES C. TAKASAKI, I

A.

SUGAMA, I

A.

YANAGITA, I N. TAhfFYA,' E . G . ROWAN' HARVEY2*

and A. L.

'Department of Chemistry, Faculty of Science, Tohoku University, Aobayama, Sendai 980, Japan and 'Department of Physiology and Pharmacology, University of Strathclyde, Glasgow GI 1XW, Scotland, U .K . (Acceptedfor publication 6 July

1989)

C . TAKAsAKI, A. SUGAMA, A. YANAGITA, N. TAmIYA, E. G . ROWAN and A. L. HARVEY. Effects of chemical modifications of Pa-11, a phospholipase A2 from

the venom of Australian king brown snake (Pseudechis australis), on its biological activities . Toxicon 28, 107-117, 1990.-Pa-11, a phospholipase A2 isolated from the venom of an Australian elapid snake Pseudechis australis, was chemically modified and its enzymic, neuromuscular and lethal activities were studied. Carboxymethylation of Met-8 gave a derivative with 2% of the enzymic activity and less than 3% of the lethal activity of native Pa-11 ; it had about 5% of the original ability to block directly and indirectly stimulated mouse phrenic nerve-hemidiaphragm preparations . Nitrophenylsulfenylation of tryptophanyl residues at positions 31 and 69 caused loss of all activities . Amidination of all 141ysy1 residues gave a derivative with 41 % and 16% of the enzymic and lethal activities, respectively, but with less than 5% of the original neuromuscular blocking activity. Mono-carbamoylation of lysyl residues at positions 58, 63, 81 and 85 was achieved . The most abundant derivative, 58carbamoyl-lysine Pa-l l was enzymically 130% and lethally 100% as active as native Pa-11, but it had only about 20% of the native's neuromuscular activity in vitro . 63-Carbamoyl-lysine Pa-11 had 10% of the enzymic and 20% of the lethal activities, respectively; however, it retained at least 50% of its ability to block neuromuscular transmission in vitro, while losing most of its activity to block directly stimulated muscle contractions . 81- and 85-Carbamoyl derivatives have the same enzymic and lethal activities as the original protein, but the 85 derivative had less than 10% of the native neuromuscular activity . Hence, modifications of lysine residues at positions 58, 63 and 85 seem to be particularly significant in altering the neuromuscular, but not enzymic, activity of Pa-11, perhaps by altering the ability of the toxin to bind to its target on nerve and muscle membranes. Modification at position 63 appeared to lead to a dissociation of effects on neuromuscular transmission and directly on muscle cells.

*Author to whom correspondence should be addressed . 107

108

C . TAKASAKI et al. INTRODUCT1ON

of phospholipases A2 (phosphatidate 2-acylhydrolase, EC 3.1 .1 .4) have been purified from the venoms of various snakes, and many of them show presynaptic toxicity at the neuromuscular junction and others produce damage to skeletal muscle fibres . There are large variations in the enzymic and lethal activities among them, and not all phospholipases AZ from snake venoms are neurotoxic or myotoxic (see reviews by ERKER, 1978 ; CHANG, 1985). Studies on the enzymic and lethal activities of the snake venom phospholipase AZ show no simple relationship between the two activities . Although there is considerable evidence that phospholipase Az activity is indispensable for the toxicity (STRONG et al., 1976; HALPERT et al., 1976 ; HOWARD and TRUOG, 1977), CoNDREA et al. (1981a) demonstrated that extensive modification of lysine residues of Naja nigricollis phospholipase AZ drastically reduced its lethal activity with little effect on the enzymic activity . More recently, inactivation of the enzyme activity of a toxic phospholipase AZ from Bothrops alternatus venom was reported to have little effect on its lethal potency (NISENBom et al., 1988). The controversy regarding the role of phospholipase AZ activity and biological effects of venom components has been reviewed by ROSENBERG (1986) . A basic phospholipase AZ with lethal activity, Pa-11, was purified (NISHIDA et al., 1985a) from the venom of an Australian elapid snake, Pseudechis australis and sequenced (NISHIDA et al., 19856) as shown in Fig. 1 . Pa-11 has both presynaptic and myotoxic activities on isolated nerve-muscle preparations (ROWAN et al., 1989a) . In the present paper, Pa-11 was subjected to chemical modifications on methionyl, tryptophanyl or lysyl residues and the enzymic, neuromuscular, and lethal activities of the derivatives were studied in an attempt to explore the role of phospholipase activity in the pharmacological action of the toxin . A NUMBER

MATERIALS AND METHODS Materials Pa-11 was purified from the venom of the king brown snake (Pseudechis attstralis), as described previously (NIStnnA et al., 1985a) . 1,2-Dipalmitoylglycero-phosphocholine was purchased from Calbiochem-Behring, Hoechst Corp ., La Jolla, CA, U.S .A. Methylacetimidate hydrochloride was purchased from Sigma Chemical Co ., St Louis, MO, U .S .A . ; K'4CNO (55 Ci/mole) was from Àmersham International plc, Buckinghamshire, U .K . Lysylendopeptidase (EC 3 .4.14.50) from Achromohacter lyticus was from Wako Pure Chemical Industries, Osaka,Japan . Assay of phospholipase A 7 activity The enzymic activity was measured by a titration method using an emulsion of 1,2-dipahnitoylglycerophosphocholine (4 mM), Triton X-100 (8 mM) and CaCl, (20 mM) as the substrate at pH 8 .0 and 37°C as described previously (TAKASAKI and TAMIYA, 1982) . One unit of enzyme activity was defined as the amount of the enzyme that released 1 umole of fatty acid/min under the above conditions . Usually a triplicate run was carried out in each experiment . If the standard deviation was greater than 5%, a duplicate run was performed. Neuromuscular activity Hemidiaphragms and attached phrenic nerves were removed from male mice (20-25 g, Bantin and Kingman A strain), as described previously (ROWAN et al., 1989a). The preparations were mounted with a resting tension of approximately 1 g in 5 ml tissue baths containing a physiological salt solution of the following composition (mM) : NaCl, 118 .4; KCI, 4.7; MgSO 1 .2; KH,PO 1 .2; CaCl,, 2.5 ; NaHCO 25 ; and glucose, 11 .1 . The solution was maintained at 32°C and pH 7 .3, and bubbled with oxygen containing 5% CO, . For indirect stimulation, the phrenic nerve was stimulated by pulses of 0.2 msec of sufficient strength to produce a maximal twitch. For direct muscle stimulation, two silver electrodes were hooked on to either side of the rib ; pulses of 2 msec were applied with sufficient strength to produce a maximal twitch . Preparations were stimulated alternately indirectly via the phrenic nerve and directly via the hook electrodes at a frequency of 0 .05 Hz. Prior to the addition of the test

Chemical Modifications of Snake Venom PLA, 10

20

109 30

NLIQFGNMIQCANKGSRPSLDYADYGCYCG WGGSGTPVDELDRCCQVHDNCYEQAGKKGC

FPKLTLYSWKCTGNVPTCNSKPGCKSFVCA CDAAAAKCFAKAPYKKENYNIDTKKRCK Fia.

1 . AIHIO ACID SEQUENCE of

Pa-11 .

compounds, tubocurarine (15 pM) was always added to ensure that activation of nerve endings by the direct stimulation did not contribute to the overall tension recorded in response to the direct stimulation. With most of the derivatives, there was sufficient material for testing at one concentration; 10 pg/ml was chosen as a standard . Twitches were monitored continuously until complete blockade or for 300min, whichever was shorter. Responses of control preparations stimulated in the absence of toxin were reduced by about 20% after 300 min. Lethal activity Male mice (19-21 g) were injected via the tail vein with protein contained in 0.1-0 .4 ml of saline and observed for 24 hr . Four mice were used at each dose, with the doses being increased by a factor of 1.1 or 1.2 . With all the lethal derivatives, mice passed red-brown urine and showed signs of limb paralysis before death. Carboxymethylation of Met-8 Pa-11 (12 mg) was dissolved in 4 .0 ml of 0.2 M sodium formate buffer, pH 3.0 containing 6 M guanidine hydrochloride and 223 mg of iodoacetic acid . The reaction mixture was left at 40°C for 16 hr in the dark, then applied to a column (2 .2 cm x 29 cm) of Sephadex G-25 equilibrated with 50 mM sodium borate buffer, pH 8.1 . The protein-containing fraction was pooled and applied to a column (1 .2 cm x 9.5 cm) of CM-cellulose CM-52 equilibrated with the same buffer . The fraction containing the main component was pooled and freeze-dried, and the residue was desalted and freeze-dried again. In order to control for effects of the denaturation process in 6 M guanidine HCl, 1 mg Pa-I 1 was dissolved in 0.4 ml of 0.2 M sodium formate buffer, pH 3.0, containing 6 M guanidine HCI. The solution was left in the dark at 40°C for 16 hr and then applied to a column (1 .2 x 30 cm) of Sephadex G-25 equilibrated with 0.1 M acetic acid. The protein-containing fraction was pooled and freeze-dried and the residue was dissolved in 3 ml of 0.14 M NaCl-0.015 M sodium/potassium buffer pH 7.4. The concentration was determined from A,, and the phospholipaseA, and lethal activities were determined as described above. Renaturation is essentially complete during the gel filtration step . Bromophenacylation of His-48 The modification with p-bromophenacylbromide was carried out as described previously 1985a). The derivatives completely lost enzymic and lethal activities .

(NISHIDA

et al.,

Nitrophenylswlfenylation of tryptophanyl residues Pa-11 (9 mg) was dissolved in 4ml of 50% (v/v) acetic acid containing 1 mg of 2-nitrophenylsulfenyl chloride . The reaction mixture was left for 1 hr at room temperature and applied to a column (1 .5 cm x 4 cm) of Sephadex G-25 equilibrated with 10% (v/v) acetic acid. The fraction containing protein was pooled, freeze-dried, dissolved in 4 ml of 0.2 M ammonium acetate, buffer, pH 7.5 and applied to a column (1 .2 cm x 18 cm) of Bio Rex 70 (minus 400 mesh) equilibrated with te same buffer . The fraction containing the main component was pooled, freeze-dried, desalted and freeze-dried again. Anddination of lysyl residues Pa- I 1 (10 mg) was dissolved in 2 ml of distilled water and the solution adjusted to pH 9.5 with 0.05 M NaOH . To this solution, 1 I mg of methyl acetimidate hydrochloride in 50 pl of 2 M NaOH was added, and the reaction mixture was kept at pH 9.5 with 0.5 M HCl for 40 min. This procedure was repeated three times, and then the

C. TAKASAKI et al.

110 0.8

a

ro3

~2 ro 1 0 0 E

0.4

O

0 .3

E È

0_

ô

T

0 .2

> U V

ro?

0 .1

0 ro

0.5 .m 0 ro

Elution volume/l FIG . 2. SEPARATION OF CARBAMOYLATED DERIVATIVES OF Pa-11 . (a) Chromatography on CM-cellulose CM-52. A mixture of carbamoylated Pa-11 (61.5 mg) was chromatographed on a column (2 .2 x 36 cm) equilibrated with 0.01 M sodium-potassium phosphate buffer, pH 6.5. A linear concentration gradient elution of NaCI was carried out from 0.1 to 0.25 M over 1.2 litre in the same buffer . The protein-containing fractions (I-III) indicated by the bars were collected separately. ( ), A280 nm; (-------), NaCl concentration; (. . . . .), radioactivity. Inset shows the molar ratios of '4C-labelled to toxin, and indicates the presence of tri-, di- and mono-labelled derivatives, and of native Pa-11. (b) Separation of the monocarbamoylated derivatives (fraction II from (a)) on Bio Rex 70 . Twenty-four milligrams of freezedried fraction II was dissolved in 3 ml of 0.1 M ammonium acetate buffer (pH 8.0), and applied to a column (2 .2 x 72 cm) of Bio Rex 70 (minus 400 mesh) equilibrated with the same buffer . The elution was carried out with a linear concentration gradient of ammonium acetate buffer from 0.1 to 0.6 M over 1 .6 litre. The fractions (II-1-II-7) indicated by the bars were collected separately . ( ), A280 nm ; (. . . . .), radioactivity.

reaction mixture was applied to a column (1 .5 cm x 35 cm) of Sephadex G-25 equilibrated with 0.1 M ammonium acetate buffer, pH 8.6 . The protein-containing fraction was pooled and applied to a column (l .8 cm x 33 cm) of Bio Rex 70 (minus 400 mesh) equilibrated with the same buffer. The fraction containing the main component was pooled and freeze-dried, and the residue desalted and freeze-dried again.

Carbamoylation of lysyl residues

Pa-11 (61 .5 mg) was dissolved in 3 ml of 0.1 M sodium borate buffer, pH 8.6 containing 0.05 M K'CNO (I50 uCi) and kept at 25°C for 5 hr. The reaction mixture was added with 3 ml of 2 M hydroxylamine hydrochloride, left for 16 hr and then applied to a column (1 .2 cm x 50 cm) of Sephadex G-25 equilibrated with 0.01 M sodium-potassium phosphate buffer, pH 6.5 . The protein-containing fraction was pooled and then applied to a column (2.2 cm x 36 cm) of CM-cellulose CM-52 equilibrated with the same buffer (Fig. 2a). Fraction II (a mixture of monocarbamoylated derivatives, 24 mg) was pooled, desalted by passing through a column (1 .5 cm x 50 cm) of Sephadex G-25 equilibrated with 0.1 M acetic acid, and freeze-dried. The residue was dissolved in 3 ml of 0.1 M ammonium acetate buffer, pH 8.0, and applied to a column (2 .2 cm x 72 cm) of Bio Rex 70 (minus 400 mesh) equilibrated with the same buffer (Fig. 2b). Fractions II-1 to 11-7 were separately pooled and freeze-dried, and the residues desalted and freeze-dried again.

Chemical Modifications of Snake Venom PLA,

111

Determination of carbamoylated lysyl residues

About 30 nmoles of mono-carbamoylated Pa-11 from each fraction (II-1-II-7) was separately reduced, Scarboxymethylated and digested with lysylendopeptidase in 0.05 M Tris-HCl buffer, pH 9.1 at an enzyme/ substrate ratio of 1 :200 (w/w) for 5 hr at 37°C . The digests were subjected to a fast protein liquid chromatography system equipped with a column of PepRPC HR 5/5 (Pharmacia Biotechnology, Uppsala, Sweden). Peptides were eluted by a linear concentration gradient of acetonitrile in 10 mM ammonium acetate buffer, pH 5.3 . The peptides with radioactivity were freeze-dried, hydrolyzed with 5.7 M HCl at 110°C for 22 hr in vacuo and analyzed with an automated amino acid analyzer (JLC 200, JEOL, Japan). The amino acid compositions of the radioactive peptides from the lysylendopeptidase digests are given in Table I. RESULTS Carboxymethylation of Met-8

Pa-11 has a single methionine residue at position 8. On CM-52 column chromatography, the derivative was eluted at a concentration of 65 mM NaCl in 50 mM sodium borate buffer, pH 8.1, while the native Pa-11 was eluted at 80 mM NaCl . Amino acid analysis of the performic acid-oxidized derivative by the method of GUNDLACH et al . (1959) showed that the methionyl residue was carboxymethylated completely and that no other residues were modified (data not shown) . The enzymic activity of the derivative was 2.4% ofnative Pa-11 . The derivative was not toxic to mice at a dose level of 7.5 lug/g body wt. The lethal activity is reduced, therefore, to less than 3% of native Pa-11 (Table 2). At 10 ug/ml, the methionine-modified derivative produced about 40% block of indirectly and directly stimulated mouse hemidiaphragm preparations over a 3 hr exposure period (Table 3). Therefore, it could be estimated that it had less than 5% of the activity of native Pa-11 . TABLE 1 . AMINO ACID COMPOSITIONS OF MONOCARBAMOYLATED PEPTIDEs FROM DERIVATIsm

Amino acid

Peptides obtained from BioRex fraction II II-1

CMC Asp Thr Ser Glu Pro Gly Ala* Hct* Val Met Ile Leu Tyr Phe His Lys Arg Tip Total Location Position of modified Lys Yield %

Pa-11

2.54 2.00 1 .75 0.90

II-2a

II-2b

H-4

II-7

(5) (3) (2) (2)

0.59 (1)

0.57 (1)

0.88 (1) 1 .00 (1) 0.92 (0) (l)

0.92 (1) 1.10 (1) 0.78 (0) (1)

(3) (2) (2) (1)

2.04 (3) 1 .00 (1) 0.86 (1)

2.85 3.00 1 .91 2.15

2.07 (2) 3.32 (2) 0 .42 (0) (1) 0.93 (1)

0.92 (1) 0.95 (1) 5.23 (5) (I) 0.84 (1)

2.21 (2) 2.13 (2) 6.12 (5) (1) 1 .91 (2)

0.95 (1) 0.95 (1)

0.91 (1) 0.13 1 .49 (2)

0.80 (1)

1 .31 (1)

0.73 (1) 0.11 1.14 (1)

2.00 (2) 1.03 (1) 1.15 (1)

1 .19 (1)

2.21 (2)

15 71-85 81 93

16 82-97 85 38

27 71-97 85 44

6 58-63 58 90

(1) 13 58-70 63 97

*Alanine and homocitrulline residues were not separated.

112

C. TAKASAKI et at . TABLE

2.

ENZYMIC AND LE rHAL ACTMTIFS OF MODIFIED

Pa- I I derivative

PLA2 activity* units/mg %

Native Met-8 Guanidine treated NPS-Trp Am-Lys Lys-58 Lys-63 Lys-81 Lys-85

2630±91 63±3 2580±47 0 1090± 51 3344±64 282± 12 2092±80 2560±78

100 2 98 0 41 127 11 80 97

Pa-I I

Lethality (LD P) ug/g body wt 0.23 >7 .5 -<0 .25t > 12 1 .47 0.22 1 .10 0.25 0.27

100 <3 <2 16 100 21 90 90

*Means ±S .E.M . of at least 3 determinations . Met-8=carboxymethylated at position 8; Guanidine treated=denatured by 6M guanidine HCI without carboxymethylation and renatured by removing top salt by gel formation; NPS-Trp=nitrophenylsulfenylated at Trp-31 and Trp-69; Am-Lys=amidinated at all 14 Lys; Lys-58 etc=monocarbamoylated at the specified Lys residues. ±The one mouse injected with a dose of 0.25 pg/g died .

Bromophenacylation of His-48 The modification of His-48 with p-bromophenacylbromide produced derivatives that had completely lost enzymic and lethal activities, as previously reported (NISHIDA et al., 1985a) . At 10,ug/ml, His-48-modified Pa-11 had no significant effects on the twitch responses of hemidiaphragm preparations to direct or indirect stimulation (Table 3). Modification of tryptophan residues Two tryptophan residues are present in Pa-11 at positions 31 and 69 (Fig. 1). The molar absorption coefficient at 365 nm (Ew5) of the main component obtained by Bio Rex 70 chromatography was 7,440 M - 'cm - ' . The value indicates that both Trp-31 and Trp-69 were modified . (The value of E, 5 of nitrophenylsulfenylated tryptophan is 4000 M- 'cm - ', according to SCHoFFoNE et al., 1968.) The derivative lost all enzymic activity, and was not lethal at 12,ug/g (Table 2). At 1 0,ug/ml, the derivative only produced a slight depression of indirectly and directly elicited twitches of mouse hemidiaphragm preparations (Table 3). TABLE

3.

NEUROMUSCULAR EFFECTS OF MODIFIm

Pa- I 1 derivative (Pg/ml) Native (1 jug/ml) Met-8 (10 jug/ml) NPS-Trp (10 jug/ml) Am-Lys (10 jug/ml) His-48 (10 pg/ml) Lys-58 (I0 pg/ml)* Lys-63 (10 pg/ml)* Lys-85 (10 pg/ml)*

Pa-11

Twitch height after 300 min exposure (% control) Indirect stimulation Direct stimulation 20±16 57±17 88± 4 85± 9 108± 8 22 ± 15 0± 0 50±21

34± 8 64±I0 76+ 6 78+ 7 93± 14 41±21 63±28 58± 10

Abbreviations are given in the footnote to Table 2. Values are means ±S .E .M .s from four experiments, except in case of Lys-63 and Lys-85 derivatives, where n=2. *Twitch tension was measured after 200 min exposure.

113 Chemical Modifications of Snake Venom PLA2 --

120

a

100

L . O) Lm L

60 40 20 0

100 200 Time in toxin (min)

300

0

c 8 t M m L L .

3 H

0

100 200 Time in toxin (min)

300

c 8 L .Q !D L L . 3 H

FIG.

3.

0

100 200 Time in toxin (min)

EFFECTS OF MONO-CARRAMOYLATED LYSYL DERrvATIvEs OF PREPARATIONS.

300 Pa-11

ON MOUSE HEMMIAPHILAGM

Effects of Lys-58 modified Pa-11 ([1, Q) compared with native Pa-11 (/, 0), both at 10 jig/ml, on indirect (/, p) and direct (0, p) stimulation of mouse hemidiaphragm preparations. Points are means of 3-4 preparations ; S.E .M .s are not shown but were less than 10% . A and Q indicate the average twitch heights of two control preparations to indirect and direct stimulation, respectively . (b) Lys-63 modified Pa-11, and (c) Lys-85 modified Pa-11 ; details as in (a) . (a)

Amidination of lysyl residues The main component obtained by Bio Rex 70 chromatography was hydrolyzed with 5.7 MHCl at 110°C for 20 hr or 48 hr and subjected to amino acid analysis. Lysine content was 1 .89 or 4.22 residues/mole protein, respectively ; it was concluded that all 14 lysyl residues were modified because the amidinated lysine residues were recovered as lysine residues with the pseudo first order rate constant of 5 x 10 - ` hr- I under the above conditions (GARNER and GuRD, 1975). The derivative shows 41 % and 16% of the enzymic

114

C . TAKASAKI e al.

and lethal activities of original protein, respectively (Table 2), but it had virtually no activity when tested at 10 mg/ml on mouse hemidiaphragm preparations (Table 3). Carhamoylation of lysyl residues A mixture of carbamoylated Pa-11 was separated into three fractions (I-III) by CMcellulose column chromatography at pH 6.5 (Fig. 2a). The content of carbamoyl groups was calculated from radioactivity/protein molecule, and fractions 1, II and III contained di-, mono-carbamoylated derivatives, and intact Pa- 11, respectively . Separation of the monocarbamoylated derivatives from each other was carried out by Bio Rex 70 column chromatography at higher pH, at which the pKa differences between individual lysyl residues are larger (Fig. 2b). Fraction II, which contained mono-carbamoylated derivatives, was separated into seven fractions (II-1-II-7). Lysylendopeptidase digests gave a single peptide with radioactivity by reversed phase chromatography from each of fractions I1-1, 11-4 and 1I-7 . The amino acid compositions of these peptides showed that II-1, II-4 and 11-7 were 81-, 58- and 63-carbamoyl-lysine Pa-11, respectively (Table 1). Two radioactive peptides were obtained from II-2 (II-2a and II-2b), but both of them came from the same derivative, 85-carbamoyl-lysine Pa-11 . 58-Carbamoyl-lysine Pa-11 has 30% higher enzymic activity and the same lethal activity as compared to native Pa-11 . 63-Carbamoyl-lysine Pa-11 has 10% of the enzymic, and 20% of the native lethal activities, respectively . 81- and 85-Carbamoyl derivatives have the same enzymic and lethal activities as the original protein (Table 2). The changes to the neuromuscular activities induced by monocarbamoylation were complex . 58-Carbamoyl-lysine Pa-11 had reduced activity on both directly and indirectly stimulated mouse hemidiaphragm preparations (Fig. 3a); it was equivalent to about 20% of the activity of native Pa-11 . With 63-carbamoyl-lysine Pa-11, there was a great loss of activity against direct muscle stimulation, but relatively little change of effect on indirect stimulation (Fig. 3b). 85-Carbamoyl-lysine Pa-11 still blocked both directly and indirectly stimulated twitch responses, but there was a very pronounced delay of about 2 hr before twitch block was noticeable (Fig. 3c). There was insufficient 81-carbamoyl-lysine derivative for neuromuscular testing.

DISCUSSION

A basic phospholipase Az with presynaptic and myotoxic activities, Pa-11 (RowAN et al., 1989a), was chemically modified on methionyl, tryptophanyl or lysyl residues, and enzymic, lethal and neuromuscular activities of the modified derivatives were determined . Carboxymethylation of Met-8 diminished enzymic, lethal and neuromuscular activities to less than 5%, of native Pa-11 . Pa-11 treated with 6 M guanidine hydrochloride but without iodoacetic acid fully retains enzymic and lethal activities . Therefore, the loss of activities in the derivative was caused not by denaturation but by modification . The 8th position in mammalian pancreatic or snake venom phospholipases AZ is occupied by hydrophobic amino acid residues (Met, Leu or Val) and is involved in the interface recognition site (DAM-MIERAS et al., 1975). WEZEL et al. (1976) reported that porcine pancreatic isophospholipase Az completely lost its activity by carboxymethylation of Met8, and they speculated that the introduction of the negatively charged carboxymethyl group impeded the formation of the interface recognition site . In this study, Pa-11 lost its

Chemical Modifications of Snake Venom PLA,

11 5

enzymic activity drastically but not completely . The amino acid sequences of the aminoterminal region of Pa-11 and porcine isophospholipase AZ are Asn'-Leu-Ile-Gln-Phe-GlyAsn-Met-Ile-Gln'°- and Ala'-L,éu-Trp-Gln-Phe-Arg-Ser-Met-Ile-Lys` °-, respectively . Thus, the porcine enzyme has two basic amino acid residues at positions 6 and 10, while Pa-11 lacks both of them. It seems that the introduction of a negative charge at position 8 in Pa-11 does not impede the formation of the interface recognition site as seriously as in the porcine enzyme. Nitrophenylsulfenylation of both Trp-31 and Trp-69 in Pa-11 caused complete loss of activities . YOSHIDA et al. (1979) reported that the single tryptophan residue located at position 64 in phospholipase AZ I of Laticauda semifasciata was essential for its high enzyme activity, as it decreased to 5% on modification of the residue with N-bromosuccinimide ; the modified phospholipase also lost the weak neuromuscular blocking activity found in the native protein (HARVEY and TANIIYA, 1980). Broadly similar results were found after alkylation of tryptophan residues in phospholipases from cobra venoms (CoNDREA et al ., 1985). ViLjoEN et al . (1976) reported that one of the two tryptophan residues (Trp-28) in phospholipase AZ of Bitis gabonica was essential for its enzymic activity, while the modification of the other (Trp-59) did not alter the activity . The sequence comparison suggests that Trp-31 of Pa- I 1 corresponds to Trp-28 of B. gabonica phospholipase AZ. The loss of enzymic activity of nitrophenylsulfenylated Pa-I 1 is probably due to the modification of Trp-31 . Trp-69 is present only in phospholipases Az from Australian elapid snakes, and its role in the enzymic activity is not clear . Pa-11 amidinated on all of the 141ysyl residues shows 41 % and 16% of the enzymic and lethal activities, respectively, but probably had less than 5% neuromuscular activity in vitro. A similar dissociation of enzymic and pharmacological properties had been reported by CoNDREA et al. (1981b) with phospholipases from Naja naja atra and Naja nigricollis . Amidinated porcine pancreatic phospholipase AZ also retains about 60% of the activity of the native enzyme . Both Pa-11 and the porcine enzyme have the same pH optimum and a similar affinity for Caz+ (SLoTBoom and HAAS, 1975). Amidination of lysyl residue in phospholipases AZ does not affect the enzymic activity greatly, but the larger drop in pharmacological activity implies that some of the lysyi residues are important in the binding to membrane-bound targets. Previously, RAPUANo et al. (1986) demonstrated that radiolabelled phospholipases from Naja nigricollis and Naja naja atra lost their ability to bind to synaptosomal membranes after carbamylation of Lys residues . Although in this study no single lysyl residue was found to be essential for lethal and enzymic activities, several modifications altered the pattern of neuromuscular activity . Four mono-carbamoyl derivatives on Lys-58, -63, -81 or -85 were isolated by sequential chromatography of CM-cellulose, CM-52 and Bio Rex 70 . The yields of the derivatives modified on Lys-58 (27%) and Lys-85 (17%) were larger than those of the others, suggesting that these residues are located on the surface of the protein molecule . KARLssoN and PONGSAWASDI (1980) proposed that basic regions of the neurotoxic phospholipase AZ molecule facilitate their binding to negatively charged muscle and nerve membranes. Based on a comparison of the amino acid sequences of some toxic phospholipases A2, KoNno et al . (1982) thought that Lys-58 and Lys-63 are essential for neurotoxicity. From our results, Lys-58 seems not to be essential to toxic activity, although it may play a role in binding of the toxin to membranes. Similarly, modification of Lys-85 did not greatly affect the enzymic or lethal activities, but profoundly slowed the onset of muscular blockade in vitro; Lys-85 may also, therefore, be important for binding. The role of Lys-63 seems more critical for some aspects of the pharmacological activity .

116

C . TAKASAKI et al.

Recently, a highly toxic, but weakly enzymic phospholipase AZ (LD 50 = 0.048 yg/g) with Phe-63 instead of Lys-63 was isolated from Laticauda colubrina (TAKASAKI et al., 1988), suggesting that Lys-63 is not essential for the lethal activity. This colubrina toxin blocks acetylcholine release, while having no myotoxic actions (ROWAN et al., 19896) . When Pall was modified at Lys-63, there was relatively little change in its ability to block nerveevoked contractions, but there was a marked loss in effects on directly stimulated muscle contractions . This implies that Lys-63 might be important for binding of Pa-11 to muscle, but not neuronal membranes. The 63rd position is located at the entrance of the active site pocket of bovine pancreatic phospholipase AZ (DuKsTRA et al., 1981a,b) . Therefore, the lower enzymic activity of 63-carbamoyl-lysine Pa-11 may be due to the steric effect. In Naja phospholipase A2, the 63rd residue is tyrosine ; this may be the reason that carbamoylation does not affect the enzymic activity . Modification of Tyr-63 does, in fact, greatly reduce lethal activity, with less effect on enzymic activity (SOONs et al ., 1986) . Acknowledgements-We are grateful to Professor K . OGURA of Tohoku University, Sendai, Japan for the use of the Pharmacia fast protein liquid chromatography system. We thank Mr H . ABE for the amino acid analysis .

REFERENCES CHANG, C. C. (1985) Neurotoxins with phospholipase A, activity in snake venoms . Proc . natn . Sci. Council ROC,

B, 9, 126-142 .

CONDREA, E., FLETCHER, J., RAPUANO, B. R., YANG, C. C. and ROSENBERG, P. (1981a) Effect of modification of

one histidine residue on the enzymatic and pharmacological properties of a toxic phospholipase A, from Naja nigricollis snake venom and less toxic phospholipase A, from Hemachatus haemachatus and Naja naja atra snake venoms . Toxicon 19, 61-71 . CONDREA, E., FLETCHER, J. E., RAPUANO, B. E., YANG, C. C. and ROSENBERG, P. (19816) Dissociation of enzymatic activity from lethality and pharmacological properties by carbamylation of lysines in Naja nigricollis and Naja naja atra snake venom phospholipases A,. Toxicon 19, 705-720 . CONDREA, E., SOONS, K. R., BARRINGTON, P. L., YANG, C. C. and ROSENBERG, P. (1985) Effect of alkylation of tryptophan residues on the enzymatic and pharmacological properties of snake venom phospholipase A, . Can . J . Physiol. Pharmac . 63, 331-339 . VAN DAM-MIERAS, M. C. E., SLoTBoom, A . J ., PIE[ERSON, W. A. and DE HAAS, G. H . (1975) The interaction of phospholipase A, with micellar interfaces, the role of the N-terminal region . Biochemistry 14, 5387-5394. DUKSTRA, B. W., DRENTH, J. and KALK, K. H . (1981a) Active site and catalytic mechanism of phospholipaseA, . Nature, Lond. 289, 604-606 . DIJKSTRA, B. W., KALK, K. H., HOL, W. G. J. and DRENTH, J. (19816) Structure of bovine pancreatic phospholipase A at 1 .7 A resolution . J. molec . Biol. 147, 97-123 . ERKER, D. (1978) Studies on presynaptically neurotoxic and myotoxic phospholipasesA,. In : Versatility of Proteins ; Proc. Int . Symp. on Proteins, Taipei, Taiwan, pp . 413-431 (Li, C . H ., Ed .) New York : Academic Press . GARNER, W . H . and GURD, F. R. N. (1975) Semisynthesis of a specific NH ; terminal [1-' 3C1 glycine adduct to sperm whale myoglobin: Intermediate protection of e-amino groups with methy acetimidate . Biochem. biophys . Res. Commun . 63, 262-268. GUNDLACH, H . G ., MOORE, S . and STErN, W . H . (1959) The reaction of iodoacetate with methionine . J. biol . Chem . 234, 1761-1764 . HALPERT, J., ERKER, D . and KARLSSON, E . (1976) The role of phospholipase activity in action of a presynaptic neurotoxin from the venom of Notechis scutatus scutatus . FEES Lett . 61, 72-76 . HARVEY, A . L. and TAMIYA, N . (1980) Role of phospholipase A activity in the neuromuscular paralysis produced by some components isolated from the venom of the sea snake, Laticauda semifasciata . Toxicon 18, 65-69 . HOWARD, B. D. and TRUOG, R. (1977) Relationship between the neurotoxicity and phospholipase A activity of ß-bungarotoxin . Biochemistry 16, 122-125. KARLSSON, E. and PONGSAWASDI, P. (1980) Purification of two phospholipase A isoenzymes with anticoagulant activity from the venom of the cobra Naja naja siamensis . Toxicon 18, 409419 . KoNDo, K., TODA, H., NARITA, K. and LEE, C . Y . (1982) Amino acid sequences of three ß-bungarotoxins from Bungarus multicinctus venom. Amino acid substitutions in the A chains . J. Biochem . (Tokyo) 91, 1531-1548.

Chemical Modifications of Snake Venom PLA,

11 7

NIsENBOm, H. E., PERAzzo, J. C., MoNSeutAT, A. J. and ViDAL, J. C. (1988) Effect of chemical modification with p-bromophenacyl bromide on the enzymatic and lethal properties of phospholipase A2 from Bothrops alternates (Vibora de la Cruz) venom. Toxicon 26, 1137-1144. NISIMA, S., TERAsHiMA, M., SHniAzu, T., TAKAsAKi, C. and TAMIYA, N. (1985a) Isolation and properties of two phospholipases A, from the venom of an elapid snake (Pseudechis australis) . Toxicon 23, 73-85. NISHIDA, S., TERAsHnmA, M. AND TAMIYA, N. (19856) Amino acid sequences of phospholipases A2 from the venom of an Australian elapid snake (king brown snake, Pseudechis australis). Toxicon 23, 87-104. RAPuANO, B. E., YANG, C.-C. and RosENBERG, P. (1986) The relationship between high-affinity noncatalytic binding of snake venom phospholipases A, to brain synaptic plasma membranes and their central lethal potencies. Biochim. biophys. Acta 856, 457-470. ROWAN, E. G., HARvEY, A. L., TAKAsAKI, C. and TAMIYA, N. (1989a) Neuromuscular effects of three phospholipases A, from the venom of the Australian king brown snake Pseudechis australis. Toxicon 27, 551-560. ROWAN, E. G., HARvEY, A. L., TAKASAKi , C. and TAhUYA, N. (19896). Neuromuscular effects of a toxic phospholipase A2 and its nontoxic homologue from the venom of the sea snake, Laticauda colubrina . Toxicon 27, 587-591. RosENBERG, P. (1986) The relationship between the enzymatic activity and pharmacological properties of phospholipases in natural poisons. In: Natural Toxins. Animal, Plant, and Microbial, pp. 129-174 (HARRts, J. B., Ed .). Oxford: Clarendon Press. ScoFFoNE, E., FONTANA, A. and Rocclu, R. (1968) Sulfenyl halides as modifying reagents for polypeptides and proteins . I. Modification of tryptophan residues . Biochemistry 7, 971-979. SLOTBoom, A. J. and DE HAAs, G. H. (1975) Specific transformations at N-terminal region of phospholipase A2. Biochemistry 14, 5394-5399 . SooNs, K. R., CONDREA, E., YANG, C.-C. and Ros;NBERG, P. (1986) Effects of modification of tyrosines 3 and 62 (63) on enzymatic and toxicological properties of phospholipases A, from Naja nigricollis and Naja naja atra snake venoms . Toxicon 24, 679-693. STRONG, P. N., GoEaxE, J., OBERG, S. G. and KELLY, R. B. (1976) ß-Bungarotoxin, a presynaptic toxin with enzymic activity . Proc. natn . Acad. Sci. U.S.A . 73, 178-182. TA KAsAKi , C. and TAhUYA, N. (1982) Isolation and properties of lysophospholipases from the venom of an Australian elapid snake, Pseudechis australis. Biochem. J. 203, 269-276 . TAKAsAKI, C., KIMuRA, S., KOKuBuN, Y. and TAnmrA, N. (1988) Isolation, properties and amino acid sequences of a phospholipase A2 and its homologue without activity from the venom of a sea snake, Laticauda colubrina, from the Solomon Islands. Biochem. J. 254, 869-875. VnroEN, C. C., VIsmR, L. and BoTEs, D. P. (1976) An essential tryptophan in the active site of phospholipase A, from the venom of Bitis gabonica . Biochim. biophys. Acta 438, 424-436. vAN WEzEL, F. M., SLOTBoom, A. J. and DE HAAs, G. H. (1976) Studies on the role of methionine in porcine pancreatic phospholipase A2. Biochim. biophys. Acta 452, 101-111 . YOSIMA, H., KuDo, T., SHINKAI, W. and TAMIYA, N. (1979) Phospholipase A of sea snake Laticauda semifasciata . J. Biochem. (Tokyo) 85, 379-388.