Lethal toxins and cross-neutralization of venoms from the African water cobras, Boulengerina annulata annulata and Boulengerina christyi

Tosico~n Vol . 29 No . l l, pp . 1311327, 1991 . Printed in crest Britain .

Pergamon Prae pk

LETHAL TOXINS AND CROSS-NEUTRALIZATION OF VENOMS FROM THE AFRICAN WATER COBRAS, BOULENGERINA ANNULATA ANNULATA AND BOULENGERINA CHRISTYI Scorn A. W1:nvsTmv, JA~s J. ScfnKmT and LEONARD A. SMITH*

Pathophysiology Division, U.S . Army Medical Research Institute of Infectious Diseases, Fort Detrick, Frederick, MD 21702-5011, U.S .A . (Received 30 March 1991 ; accepted

12

.tune 1991)

S. A. WEINSTE>rr, J. J. Sl.FII1IIDT and L. A. SMITH. Lethal toxins and crossneutralization of venoms from the African water cobras, Boulengerina annulata annulata and Boulengerina christyi. Toxicon 29, 1315-1327, 1991 .-Venoms of the water cobras, Boulengerina, were assayed for lethality, proteolytic activity and protein content. Boulengerina annulata annulata and B. christyi venoms averaged 89% protein and lacked proteolytic activity . The murine i.p. LD S° Of B. a. annulata and B. christyi venoms were 0.143 and 0.120 mg/kg, respectively . Polyvalent antivenom produced by the South African Institute of Medical Research neutralized 575 and 200 LD Sp Of B. a. annulata and B. christyi venoms/ml antivenom, respectively . Cation exchange chromatography resolved four lethal peaks from B. a. annulata venom and six lethal peaks from B. christyi venom. The major lethal peaks (about 12% of total venom protein) were purified further with molecular sieve chromatography and were characterized as 61 (B. a. annulata toxin) and 62 (B . christyi toxin) residue polypeptides with four half-cystines. Elucidation of the complete amino acid sequences indicated that these toxins belonged to the short-chain class of postsynaptic neurotoxins. Short-chain neurotoxins 1 from B. a. annulata and B. christyi had murine i.p. tD~ of 0 .052 and 0.083 mgfkg, respectively, and showed over 80% homology with N. nigricollis alpha toxin. Reverse-phase analysis of another peak present in both venoms resolved a toxin that had an N-terminus identical to B. christyi short-chain neurotoxin 1 . These fractions also contained toxins readily separable from the short-chain isotoxin by preparative reverse-phase chromatography . Amino acid sequencing of the first 28 residues indicated that both toxins were long-chain neurotoxins with identical N-termini. The LDso of long-chain neurotoxins 2 from B. a. annulata and B. christyi venoms were 0 .086 and 0.090 mg/kg, respectively . The venoms of these little-known elapids have the lowest LDP of any African proteroglyph studied thus far and have high concentrations of potent postsynaptic neurotoxins. INTRODUCTION

elapid genus, Boulengerina (Dor.c.o, 1886) contains two species of distinctive, poorly known snakes . Boulengerina annulata annulata (BUCHHOLZ and PETERS, 1876) THE AFRICAN

'Author to whom correspondence should be addressed. 1315

1316

S. A. WEINSTEIN et al.

is found along streams, river tributaries and lake margins in Cameroon, Zaire and Gabon. It is light brown or reddish tan ground with 21-24 black bands, and reaches an average adult length of 2 m (maximum length about 2.8-3 m). The eastern race, B. a. stormsi (DOLLO, 1886), is restricted to Lake Tanganyika and the immediate surrounding areas. It is morphologically the same as B. a. annulata but has fewer (3-5) bands in the neck region . Boulengerina christyi (BOULENGER, 1904) ranges throughout Zaire and the Congo in habitats similar to that of B. annulata . Although B. christyi is sometimes considered a subspecies of B. annulata, and is probably sympatric with it in some localities, differences in morphology and pattern (black ground color with yellow bands) have afforded it full species status . These differences have caused some investigators to consider it a monotypic species of the genus Limnonaja, rather than a congeneric (SCHblmT, 1923). It is similar in size to B. annulata . Apart from their oviparity, aquatic habits and tendency towards favoring low-lying trees near aquatic environments, their natural history and reproductive biology are almost wholly unknown. They are nocturnal foragers and feed primarily upon fish and frogs. Except for a single LD P report (CHRISTENSEN, 1971), the lethality of their venoms has not been reported . There are no data describing the lethal components of these venoms . We report here on the isolation and characterization of four lethal toxins from the venoms of B. a. annulata and B. christyi. We describe also the neutralization of these venoms by a commercial polyvalent antivenom. MATERIALS AND METHODS Venoms

Lyophilized venoms of B. annulata annulata and B . christyi were purchased from Latoxan (Rosans, France). Lyophilized venoms of Naja haje haje and Hemachatus haemachatus were purchased from Sigma Chemical Co . (St Louis, MO, U.S.A.) . Venoms of Naja nivea, N. melanoleuca, N. nigricollis, Dendroaspis polylepis, and D . angusticeps were extracted from long-term captive specimens. Freshly extracted venoms were immediately frozen and lyophilized. All lyophilized venoms were stored at 4°C over desiccant in the dark . Chemicals

Trifluoroacetic acid fTFA), sequenal grade, was purchased from Pierce Chemical Co . (Rockford, IL, U.S .A.) . Acetonitrile and water (HPLC grade) were from Thomas Scientific (Swedesboro, NJ, U.S.A.), and 4-vinylpyridine was from Sigma Chemical Co . (St Louis, MO, U.S .A.). Ouchterlony double irnrrptnodijjusion

Immunodiffusion was performed as described previously (Wetxsretx et al., 198 .

Determirratèon of protein content and proteolytic activity

Protein content of crude venoms was determined by using the BCA assay (Pierce Chemicals, Rockford, IL, U.S.A .) (Sra'nt et al., 1985). Protein concentrations of purified toxins were calculated from the molar extinction coefficients. Venom proteasè activity was measured by using a commercial casein substrate (Bio-Rad Laboratories, Richmond, CA, U.S .A .). The casein was incorporated in agarose, wells were punched in the agarose, and an appropriate venom dilution (from 0.125 to 2 mg/ml) in phosphate-buffered saline (PBS, 0.05 M, pH 7.2) was applied. A trypsin solution (232 units/mg, 5 Kg/assay, Millipore Corp ., Freehold, NJ, U.S.A.) served as a positive control and reference, and a well filled with PBS served as a negative control. Lethal potency determination

The i.p . [.nw values of the crude venoms were determined by injecting male Swiss-Webster mice (18-20 g) in four groups of four mice per group. All injections were administered in the lower quadrants of the abdomen. Dosage was derived from a 1 mg/ml solution of venom in PBS (0.05 M, pH 7.2). Animals were observed after injection and mortality was recorded after 24 hr. The ~n,~ of lethal fractions from Superose-12 (molecular sieve)

Lethal Toxins and Neutralization of Boulengerina Venoms

1317

chromatography was determined by injecting male Swiss-Webster mice (18-20 g) in either fow groups (B . annulata annerlata, short~hain neurotoxin 1 and long-chain neurotoxins 1 from both species) or three groups (B. christyi, short-chain neurotoxin I) of fow mice per group. Toxin diluent was Tris-HCl (0.05 M, pH 7 .2) containing 0 .7 M NaCI. Lethal potency of HPLC fractions was determined by using five groups of four mice per group . In the first two dose levels involving venom fractions, equivalent volumes of buffer were injected into control groups of two mice per group . Animals succumbing to either crude venom or venom fractions were necropsied and any gross tissue pathology was examined and noted . The t.n, ° was calculated by the Spearman-Karber method (Woxr n Ha+r.rtt ORGANIZATION, 1981) . The 95% fiducial limits for the r.u~ were determined .

Venom fractionation Venom samples were prepared for fractionation by dissolving 30 mg in 650 tel of 0 .05 M 2{N-morpholino)ethanesulfonic acid (MES) (pH 6 .5) . The sample was filtered through a 0 .22-tun membrane and chromatographed on a Mono-S 10/10 (ration exchange, 1 an x 10 cm) column equilibrated with 0 .05 M MES (pH 6.5 ; Buffer A) . The column was eluted with a 0-1 .0 M NaCI gradient in Buffer A . Fractions of 2 .5 ml were wllected, with a flow rate of 2.75 ml/min . The column elutee was monitored by absorbance at 280 nm, and corresponding peak fractions were pooled for further study . Fractions corresponding to lethal activity were pooled, lyophilized, filtered through a Millipore membrane (0.22 tem) and applied in G00-tel aliquots onto a Superose-12 10/30 (molecular sieve) whunn . The column was developed with 0 .05 M Tris-HCl containing 0.7 M NaCI . Each run was performed at a flow rate of 0.5 ml/min and 1-ml fractions were collected . The elutee was monitored at 280 nm and corresponding fractions of high absorbance were pooled and examined for lethal activity. Selected fractions were acidified with TFA and subjected to reverse-phase chromatography (Waters Associates, Milford, MA, U .S.A .) using a Hi-Pore RP-318 column (Bio-Red Laboratories, Richmond, CA, U .S .A .) . Fractions obtained from reverse-phase chromatography were neutralized with N~thylmorpholine (1 .0-1 .5 pl/ml of fraction) and most of the acetonitrile was removed by evaporation with a stream of dry nitrogen at room temperature . Fractions were then screened in mice for lethality. Control animals were injected with neutralized TFA only.

Sequence analysis Before sequencing, samples were reduced and pyridylethyhrted (CAVUVS and FxmntKAx, 1970 ; SCHMIDT and Miuni .eaxoox, 1989) . Automated Edman degradation was then performed by a model 470A amino acid segrtencer from Applied Hiosystems (Foster City, CA, U.S .A.) . Residues were identified with an on-line highpressure liquid chromatograph, model 120A (Applied Biosystems) . Digestions with staphylococcal protease V8 were performed as described previously (SCHA4DT and MIDDLEBROOK, 1989).

Venom neutralization assay Polyvalent antivenom [South African Institute of Medical Research (SAIMR)] and venorns of interest were prepared in volumes sufficient for i .p. injection of one group of mice . Each group consisted of fow male Swiss-Webster mice (18-20 g) . Typically, 200 pl of antivenom was mixed with C>-28 Ln~ of each venom or toxin of interest, incubated at 37°C for 45 min, and then injected into mice. The number of survivors after 24 hr was recorded .

Aretyleholine receptor assay The binding assays were performed using the non-radioactive method of STU.es (1991). Immulon II plates were coated with 100 tel of toxin (50 ug,/ml) in carbonate buf%r at 4°C overnight . Wells were blocked with PHS containing 1 % gelatin (PBSG) and 100 td of acetylcholine receptor (AchR) (50 trg/ml PHSG) was added to each well . After incubation, wells were washed with PBS containing 0.1 % Tween 20, and mouse AchR antiserum was added . Anti-mouse antibody alkaline phosphatase conjugate (Sigma) was used as the secondary antibody, and paranitrophenyl phosphate (Kirkegaard and Perry Laboratories, Gaithersbwg, MD, U.S .A.) was the substrate . Absorbance of samples was read at 405 nm, 30 min after substrate was added. Competitive ELISA was performed by mixing 1 mM carbamylcholine chloride with an equal volume of AchR (100pg/ml) and immediately applying the mixture to toxin-containing wells . Values in the text represent means of triplicate assays .

131 8

S. A . WEINSTEIN et al. RESULTS

Protein content, lethal potencies and proteolytic activity of B. a. annulata and B. christyi venoms The crude venoms of B. a. annulata and B. christyi were white, suggesting a lack or low

content of L-amino acid oxidase. Assay for protein content of crude venoms indicated that protein content averaged 89% (range = 8Cr92%). Neither venom in concentrations up to 100 pg had any detectable proteolytic activity using casein as substrate. The marine i.p. LDP and 95% confidence limits of B. a. annulata and B. christyi venoms were 0.143 (0.131-0 .156) and 0.120 (0.109-0 .130) mg/kg, respectively . Animals that succumbed to lethal injections of either venom showed sudden collapse accompanied by myoclonus and uresis . Death was preceded by a period of tachypnea, the duration of which was dependent on the amount of venom injected. Tachypnea was a feature more prominent with administration of B. christyi venom. Gross necropsy was unremarkable . Neutralization of B. a. annulata and B. christyi venoms by SAIMR polyvalent antivenom

Neutralization tests with SAIMR polyvalent antivenom, which contains antibodies against venoms from eight species of elapids and two species of viperids, showed that this antivenom afforded good protection against B. annulata venom and about 65% less protection against B. christyi venom (Table 1). Neutralization of Dendroaspis polylepis venom, one of the homologous venoms of this antivenom, was more than twice as effective as neutralization of B. a. annulata venom. It is noteworthy that although the antivenom neutralized N. nivea venom (another venom homologous for SAIMR polyvalent) about twice as effectively as B. christyi venom, it neutralized only one-third the number of mouse LD S° compared with D. polylepis venom (Table 1). These results suggest there is significant antigenic distance between venoms of Boulengerina and those of other African elapid taxa. Fractionation of B. a. annulata and B. christyi venoms

Figures 1 and 2 illustrate the chromatographic profiles obtained from subjecting to Mono-S (ration exchange) chromatography 30 mg of crude venoms of B. a. annulata (lA) or B. christyi (2A). Analysis of B. a. annulata venom resolved at least 10 peaks, of which four were lethal to mice (Fig. lA). Boulengerina chrisryi venom was resolved into at least 11 peaks, of which six were lethal to mice (2A). Animals succumbed within 15 min after TABLE

1.

NEUTRALIZATION of Boutengerina, Naja nivea Arm Derrdroaspis potylepis VENOMS HY SAIMR POLYVALENT ANTIVENOM

Venom

Locality

B. a. annutata B. christyi N. nivea

Unknown Unknown SW Cape South Africa Western Kenya

D. potytepis '

i .p . LDSp pgl20 g mouse (95% confidence limits)

Marine i .p . LDS° doses neutralized by 1 ml antivenom'

2 .86 (2.62-3.13) 2 .40 (2.18-2.61) 6 .68 (6.11-7.31)

575 f 22 200 f 15 375 f 18

5 .12 (4.52-5 .79)

1200 f 30

Values represent means tS.E .M . (n = 3) .

Lethal Toxins and Neutralization of Boulengerina Venoms

1319

1ao so eo

o.1s

A eo

r0 0.03

90

0.00

10 0.12

_

B

FRACTION NUMBER

a.oE

E

0.0E

0.0~

0 .00

10

~ 16

_T

20

- r

23

~r 30 93

FRACTION NUAABER

Frc. 1. CHROMAiOORAPrßC PROFIIE OF B. a. anrwlara cxtme veivord . venom subjected to FPLC with cation exchange. Crude venom (30 mg) was dissolved in 2-(N-morpholino}ethanesulfonic acid (MFS) 0.05 M (pH 6.5) and filtered through a 0.2 um Millipore filter . The filtrate was then subjected to cation exchange by using a Mono-S IO/10 column attached to a FPLC unit. The column was equilibrated with the MES buffer and 625-p1 aliquots of sample were injected onto the column . Chromatography was carried out with a flow rate of 3 ml/min, and elution was accomplished with a linear gradient of Buffer A to 1 M NaCI in Buffer A (% Buffer B) . Fractions of 2.5 ml were collected, monitored for absorbance at 280 nm and tested in mice for lethality. Lethal peaks are indicated by letter designation. (B) Molecular sieve chromatography of lethal peak D (Mono-S fractions 30-31) . Samples were chromatographed at room temperature by using a Superose 12 H 10/30 column, and eluted with 0.05 M Tris-HCl (pH 7.2) containing 0.7 M NaCI . A flow rate of 0.5 ml/min was maintained and fractions of 1 .0 m1 were collected. All fractions went monitored for absorbance at 280 nm and lethal activity. (A) Crude

injection with 0.4 mg/kg of any of the lethal peaks from either venom. Mice injected with B. a. annulata peaks A, B, or D (fractions 20-22, 25, and 30-31, respectively) exhibited myoclonus and rapid prostration accompanied by uresis and death. Injecting B. christyi peaks A1, B1, C1 or D1 (fractions 15-16, 21, 23, 24-25, respectively) resulted in symptomology resembling closely the efïects observed with B. a. annulata peaks A, B, and D. Boulengerina a. annulata peak C (fractions 27-28) and B. christyi peaks E1, and F1

1320

S, A. WEINSTEIN et al. 100 0.20

e0 e0

0.15

70

E Q

a0 00 0.10

30 10

O.Os

]0

20 0.00

10

O.tx r

E c

gN

O.e

0.15

o.e

o.to

0.9

C

O.OS

o .o ~~T e to 1s

xo

xs

FRACTION NUIiABER

9e

as

~

T1

ITTTs 1a ts J

xo

xs

FRACTION NUMBER

F1G . 2. CHROMATOGRAPHIC PROFILE OF B. christyi CRUDE VENOM . Crude venom subjected to FPLC with cation exchange. Crude venom (30 mg) was dissolved in 0.05 M 2{N-morpholino}ethaneaulfonic acid (MES) (pH 6.5 ; Buffer A) and filtered through a 0.2-tml Millipore filter. The filtrate was then subjected to cation exchange by using a Mono-S 10/10 column attached to a FPLC unit. The column was equilibrated with Buffer A and 625-p1 aliquots of sample were injected onto the column . Samples were chromatographed with a flow rate of 3 ml/min, and elution was accomplished with a linear gradient of 1 M NaCI in Buffer A (% Buffer B) . Fractions of 2.5 ml were collected, monitored for absorbance at 280 nat and tested in mice for lethality. Lethal peaks are indicated by letter designation. (B) Molecular sieve chromatography of lethal Peak Dl (Mono-S fraction 24-25) . Samples were chromatographed at room temperature by using a Superose 12 H 10/30 column, and eluted with 0.05 M Tris-HCI (pH 7.2) containing 0.7 M NaCI . A flow rate of 0.5 ml/min was maintained and fractions of 1 .0 ml were collected. All fractions wtre monitored for absorbance at 280 nm and lethal activity . (C) Molecular sieve chromatography of lethal Peak El (Mono S fractions 28-29) . Performed under experimental conditions detailed for Panel B. (A)

(fractions 28-29 and 33, respectively) caused symptoms distinguishable from those caused by either peaks A, B, D or A1, B1 Cl or D1 . Injecting 0.35 mg/kg of these fractions led

Lethal Toxins and Neutralization of Boutetrgerina Venoms 1 B. a. annulata: K B. christyi : M 1

2 I E 2

3

4

5

C H N 3 4 5

6

7

8

Q O S 6 7 8

9 10 1 1 12 13 14 1 5 1 8 1 7 1 B 1 9

132 1 20 21 22 23 24

S Q P P T T T H C S G© E T N C Y 9 1 0 1 1 1 2 1 3 14 15 18 1 7 18 19 20 21 22 23 24 25

25282728293031323334353837383940414243444548474849 B. a. annulata : R K ~O W ~S D H R G T I I E R G C G C P T V K P G V B. christyi : E K R W H D H R G T I I E R G C ß C P T V K P G V 28 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 48 47 48 49 50 .~ . " ~ ©®~® ~Til.Y~.1~l.Yl.7.~.Z~.YZ.Y :1.YZa~IaZ:Ya FIG .

3.

An71NO ACID SEQUENCES OF PEAK D FROM B. Q. Otirlttiata VENOM AND PEAK DI

B. christyi vENOM.

FROM

rapidly to protracted tachypnea, prostration and death. To purify further the major lethal fractions, Mono-S fractions 30-31 (comprising about 11.6% of total venom protein) and 24-25 (comprising about 12.6% of total venom protein) from B. a. annulata and B. christyi venoms, respectively, were subjected to Superose-12 (molecular sieve) chromatography . Figures 1B and 2B show the profiles obtained from gel filtration analysis of B. a. annulata Mono-S peak D and B. christyi Mono-S peak D1 . Both lethal toxins appeared as a single peak, which eluted in fraction 18 (Figs 1 B and 2B). About 2.70 mg of B. a. annulata toxin or 2.85 mg of B. christyi toxin could be obtained from 30 mg of either crude venom (about a 75% yield from the Mono-S pools) . Injection of either of these peaks caused rapid collapse, myoclonus, uresis, prostration and death. The mucine i.p. LDS° of Superose-12 lethal pools obtained from Mono-S major lethal fractions of B. a. annulata and B. christyi venoms were 52 ug/kg (95% confidence limits 45 .9-60.0) and 83 hg~g (95% confidence limits 70-97.5), respectively . Superose-12 analysis of Mono-S fractions 27-28 from B. a. annulata venom and fractions 28-29 from B. christyi venom resulted in a single peak with a trailing shoulder. Figure 2C illustrates the Superose-12 profile obtained from B. christyi Mono-S fractions 28-29 (consisting of 7.2% total venom protein). Boulengerina a. annulata Mono-S fractions 27-28 (comprising about 9% total venom protein) appeared identical (not shown), and these peaks both eluted in fraction 18. Injecting 250 hg/kg of either peak elicited an insidious onset of tachypnea, which led rapidly to prostration and death. Reversephase chromatography and amino acid sequences of major lethal toxins from venoms of B. a. annulata and B. christyi

A single homogeneous peak was found by reverse-phase analysis (data not shown) of lethal toxins obtained from molecular sieve chromatography of Mono-S fractions 30-31 and 24-25 from venoms of B. a. annulata and B. christyi, respectively . Each toxin was pyridylethylated and placed in the automatic sequencer. 1'he first 41 residues of the B. a. annulata peak and the first 38 of the B. christyi peak were identified . The remainder of each sequence was obtained after cleaving of the derivatiz;ed toxins with the glutamic acid-specific, staphylococcal protease V8. Resulting fragments were purified by reversephase HPLC (not shown) . A peptide was isolated from the B. a. annulata toxin which

min this column 4 residues column Aresidues Rsvea concentration BAbsorbance and waswas Pox Bwas 23 0illustrated (C) 3equilibrated A38-61 B a39-62 F456 C Bio-Rad was SHQOSS (rancnoxs TFA/70% for 17-19 claeosUroax~rxv 8 monitored 2min, P77891011121314151617181920 in and R8both These (B) with Hi-Pore V9 aacetonitrile 25-26) then 10 Analysis St]panels peptide 10% 11Patfragments raised AR-318, 12 QP210 solvent ~xn of 13 of WEINSTEIN ATof Anm this (C) 14 Str~eose-12 to Twas Flow B01520% BPTclvistyi christyi The figure 18After Hobtained rate cmpartial were (C)x(A) B17 1Superose-12 Moxo-S at et25 18 injection S2ml/min ~cnoxs al sequenced, cm Analysis 19amino Q (C) 20 min from OSolvent BPFwx 21of Yand then acid oHr~txcn 22 of sample, fractions the El the B, Y3sequences Aathereby kept 24 (me~cnoxs temperature was B anrwlata the 2Slinear W ~eoM 18-19 christyi (C) 028solvent 27 correspond DBtocompleting Superose-12 28 Va45% 28-29) was TFA was toxin, annulata 30°C Bheld and atwhich to the

1322

S. .

.

Peak 1 ME(C)H Peak 1 2 4 I R I T

S

D Q H V PEAK n

PEAK

Fta. . .~-ptL+se Moxo-S The solvent .06% The at . 30 peaks fractions contained contained structures .

. .

;

.

. . .46

.

.

.0

. .

.

T K T

.06%

.1 .

.

. .

. .

Lethal Toxins and Neutralization of

Boutertgerina

Venoms

1323

The primary structures of these toxins are shown in Fig. 3. The toxins are composed of toxin) or 62 (B . christyi toxin) residues and both have four half~ystines . The structural data suggest that these toxins can be classified into the short-chain neurotoxin group and will therefore be referred to hereafter as B. a. annulata and B. christyi short-chain neurotoxins 1 . Based on the sequence data, the mol. wt of B. a. annulata short-chain neurotoxin 1 was 6819, and that of B. christyi short-chain neurotoxin 1 was 6876. The sequences indicated that 1 mole of each toxin contained a single tryptophan residue and (by homology with other short-chain neurotoxins) four disulfide bonds. The B. a. annulata toxin had two tyrosines per mole compared with a single tyrosine residue present in the B. christyi toxin . Thus, the molar extinction coefficients calculated from these data were 8730 (B . a. annulata short-chain neurotoxin 1) and 7450 (B. christyi short-chain neurotoxin 1). It is noteworthy that there was close agreement between protein concentrations calculated from molar extinction coefficients and the BCA assay. Reverse-phase analysis resolved two peaks from each Superose-121ethal pool from Mono-S fractions 28-29 and 27-28 from venoms of B. christyi and B. a. annulata, respectively . The profile of B. a. annulata lethal pool indicated that Peak A (HPLC fraction 13) represented substantially less of the pool than Peak B (HPLC fractions 27-30) (Fig. 4A). The profile obtained from B. christyi lethal pool showed that Peak A (HPLC fractions 9-10) was similar in height to Peak B (HPLC fractions 21-25) (Fig. 4B). N-Terminal amino acid sequences indicated that Peak A from both B. a. annulata and B. christyi corresponded to a short-chain neurotoxin sequence homologous to short-chain neurotoxin 1 from B. christyi venom (Figs 3 and 4). This suggests that Peak A represents an isotoxin species shared by these venoms . Peak B of both samples contained sequences that had identical N-termini. Although these peaks were broad and appeared to have shoulders (Fig. 4), the partial sequences of fractions taken from the positive or negative slope (e.g. fractions 21 and 27, or 25 and 30) indicated no difference in the first 28 residues . These toxins were not derivatized and thus half-cysteines could not be identified . However, when cysteines were assigned according to sequence homology with other toxins, it became apparent that these toxins belonged to the long-chain neurotoxin group (Fig. 4). Preparative analysis of 7201eg of B. a. annulata Superose-12 lethal pool protein yielded 371eg of HPLC-purified Peak B, and 9001eg of B. christyi Superose-12 lethal pool protein yielded 117 leg of purified material . Boulengerina christyi and B. a. annulata Peak B represented 13 and 5% of Superose-121ethal pool protein recovered, respectively . Because of the marked homology of the partial sequences of these toxins with long-chain neurotoxins, these toxins were termed B. a. annulata and B. christyi long-chain neurotoxins 1 . The l,n~ of B. a. annulata long-chain neurotoxin 1 was 861eg/kg (95% confidence limits = 75.5-98.5), and that of B. christyi was 90 ug/kg (95% confidence limits = 78.5-104 .5). Lethal doses produced in mice marked tachypnea, rapid collapse and prostration, which led rapidly to death. Necropsy was unremarkable. 61 (B. a. annulata

Non-radioactive acetylcholine receptor assay A non-radioactive ACh receptor assay (STII.ES, 1991) indicated that short- and longchain neurotoxins 1 purified from both Boulengerina venoms, bound to ACh receptor derived from the electric organ of Torpedo californica. Competition studies performed

with the cholinergic antagonist, carbamylcholine, showed that long-chain neurotoxin 1 from either venom produced an inhibition (64%) very similar to that observed with cobrotoxin (59%)from venom of N. n. atra . Absorbance (Aaoso,) values from the assays

1324

S . A. WEINSTEIN et al.

using long-chain Boulengerina neurotoxin 1 with and without carbamylcholine chloride present were 0.247 (S .D . f 0.010) and 0.695 (S .D.±0.019), respectively, whereas the values for cobrotoxin were 0.939 (S.D .±0.049) and 2.270 (S.D .t0.147), respectively . In the presence of carbamylcholine chloride, the binding of short~hain neurotoxin 1 from B. a. annulata to AChR was inhibited by 39% and that of B. christyi was inhibited 58%. Experiments investigating receptor-toxin binding with and without antagonist present gave A~5 values for the B, a. annulata toxin of 1.209 (S.D . ± 0.025) and 1 .978 (S.D. f0.116), respectively, whereas the toxin from B. christyi produced values of 0.657 (S.D. f0.033) and 1 .579 (S.D .±0.054), respectively . These data support the sequence data and lethal potency values, which indicate that these toxins are members of the short- and long-chain classes of postsynaptic neurotoxins.  ,fl

Immunoreactivity of venoms from Boulengerina and other African elapds with SAIMR polyvalent antivenom Double immunodiffusion with SAIMR polyvalent antivenom demonstrated extensive cross-reactivity among venoms from a number of representative African elapid taxa. Of the eight venoms tested, five (B. a. annulata, B. christyi, D. polylepis, H. haemachatus, and N. melanoleuca) produced at least three precipitin arcs, and the remaining venom samples (N. nigricollis, N. haje and N. nivea) produced one or two arcs. Both Boulengerina venoms produced one arc each, which fused with an arc shared by venoms of H. haemachatus, N. melanoleuca, N. nivea and D. polylepis . This arc also fused between both Boulengerina venoms . Reaction of either Boulengerina short-chain neurotoxins 1 with SAIMR polyvalent resulted in a single, faint precipitin arc . The long-chain neurotoxins 1 from either venom did not produce a visible arc. Structural comparisons A search of the SWISS-Prot (University of Geneva, Switzerland) and protein synthesis databases, with computer programs from Intelligenetics Corp. (Mountain View, CA, U.S.A .), found that Naja nigricollis a-toxin was over 80% homologous with short-chain neurotoxins 1 from both Boulengerina venoms . The partial sequences (28 residues) determined for long-chain neurotoxins 1 from both Boulengerina venoms, were over 70% homologous with N. nivea a-toxin. DISCUSSION

The lethal potencies of African elapid venoms range generally between 0.2 and 1 .5 mg/kg. Ct-nus~NSEN (1971) reported a marine i .v. LD P Of O.2 mg/kg for B. annulata (probably B. a. stormst) venom. To our knowledge, this is the only r n~ value published previously for any Boulengerina venom. Our lethal toxicity studies indicate that the marine i.p. LDP of B. a. annulata and B. christyi venoms (0.143 and 0.120 mg/kg, respectively) are the lowest for venoms from African elapid taxa studied to date . Both venoms elicited debilitating tachypnea which led rapidly to prostration and death. Injecting B. christyi venom into mice caused protracted tachypnea with an onset more rapid than that observed with B. a. annulata venom. Our venom neutralization studies showed that polyvalent commercial antivenom (SAIMR) efficiently neutralized B. a. annulata venom, but had significantly less neutra-

Lethal Toxins and Neutralization of Boulengerina Venons

1325

lining capacity for B. christyi venom. The degree of protection against B. a. annulata was about 50% less than that observed with D. polylepis venom, one of the homologous venoms of the antivenom studied. Neutralization of N. nivea venom was about 69% less than that of D. polylepis. Vrss~t and CxnPrtnN (1978) studied 35 cases of African elapid envenomations. Of this number, 25 were considered life threatening and eight were fatal. Fourteen of the patients reportedly showed a rapid, positive clinical response to SAIMR polyvalent. Interestingly, MoxnMm et al. (1977) described poor neutralization of D. polylepis and N. nigricollis venoms (3 and 34 marine i .p. l,n~ neutralized/ml serum, respectively) by a polyvalent antivenom prepared against venoms of seven elapids (including N. nivea and N. nigricollis), six viperine viperids and one crotaline viperid. These workers also reported considerable antigenic distances between African elapid venoms, as evidenced by a small number of fused precipitin lines in immunodiffusion assay, an observation noted in our study as well. To our knowledge, there are no data available describing the use of any antivenom in human envenomations inflicted by Boulengerina, even though such accidents have probably occurred. Envenomations inflicted by these snakes are infrequent because of the restricted habitat and mild, retreating nature of these animals. However, considering the potential large adult size (2.5-2.7 m) of Boulengerina and average venom yield range (approximately 85-250 mg) of adult bungarine elapids reaching comparable size, these snakes could clearly produce a serious snake bite accident . Boulengerina christyi venom contains larger amounts than B. a. annulata venom of a toxin that is homologous with N. nivea alpha toxin. Naja nivea alpha toxin is a 71 amino acid polypeptide (mol. wt 7897) with five disulfides and a marine i.v. Ln~ of 0.076 mg/kg (BOTES, 1971 ; BOTES et al., 1971). BOTFS et al. (1971) obtained a 5% yield of this toxin from crude venom and found that it was immunologically distinct from N. nivea beta and delta toxins . Considering that i .p. Ln~ of long-chain neurotoxins 1 from B. a. annulata (0.086 mg/kg) and B. christyi (0.090 mg/kg) are very similar to that reported for N. nivea alpha toxin, it is possible that these toxins contribute to the relative difficulty in neutralizing some of these venoms . Injecting lethal doses of long-chain neurotoxins 1 isolated from either Boulengerina venom resulted in marked tachypnea that led rapidly to prostration and death. These toxins are clearly responsible for the aforementioned symptoms observed ft`om crude venom injections . The higher proportion of the toxin B. christyi venom probably accounts for the more rapid development of tachypnea elicited by this venom. The major lethal toxins of both Boulengerina venoms are 61 or 62 residue polypeptides, which clearly belong to the short-chain neurotoxin class. These toxins comprise about 12% of Boulengerina venom protein, and they are over 80% homologous with N. nigricollis short-chain toxin 1 . KOPEYAN et al. (1973) described N. nigricollis shortchain toxin 1 as a 61 amino acid (mol. wt 6796) polypeptide, with four disulfides and an s.c. l,n~ of 0.036 mg/kg. These authors reported a low yield (1 .6%) of this toxin from the crude venom. In comparison, short-chain neurotoxins 1 from both Boulengerina venoms are present in significantly greater amounts. This probably contributes to the marked differences between Boulengerina and N. nigricollis venom lethality (0.12-0.14 and 1 .0-1 .1 mg/kg, respectively). It is also noteworthy that B. christyi venom contains a larger amount of the characterized short-chain neurotoxin than B. a. annulata venom and it is also apparent that other short-chain isotoxins are shared by both venoms . Interestingly, the shared isotoxin partially characterized in this study appeared to represent a sequence from B. christyi venom. In comparison with B. a. annulata venom, B. christyi venom

132 6

S . A . WEINSTEIN et al.

contained a larger proportion of this isotoxin species. Protein concentrations calculated from molar extinction coefficients of all the toxins above agreed closely with those determined by the BCA assay. We have found previously (~EINSTEIN et al., 1991) that toxins of lower mol. wt (e.g. Trimeresurus wagleri, peptides 1 and 2, mol. wts 2504 and 2530, respectively) were detected by the BCA assay in erroneously low amounts, when compared with concentrations calculated from molar extinction coefficients . Further study of these venoms and snakes is warranted. Venom yield data for each Boulengerina species are desirable . Continued investigation of Boulengerina venom toxins could provide information important for improving the neutralizing efficiency of antivenoms prepared against African elapid venoms that contain large proportions of potent postsynaptic neurotoxins. Study of the natural histories, morphology, serum chemistries and behavior of these ophidians could provide information useful in determining the taxonomy of Boulengerina in relation to the tribe Najini, the status of which is unclear. BOGERT (1943) found that Boulengerina possessed dentitional characteristics that clearly distanced this genus from Naja. Further study of these snakes and their venoms could provide data that might clarify the evolutionary radiation and specialization of some African elapids. Acknowledgements-The acetylcholine (ACh) receptor assay was kindly performed by Dr Bx~w SCI FC, USAMRIID, Pathophysiology Division, Fort Detrick, Frederick, MD . The technical assistance of Ct.eAt DewiTT and Fw~rtc~s SsxroN is gratefully acknowledged.

REFERENCES BOGERT, C. M . (1943) Dentitional phenomena in cobras and other elapids with nota on adaptive modifications of fangs . Bull. Am . Mus. Nat . Hilt. 81, 285-360. Bons, D . P. (1971) The amino acid sequences of toxins alpha and beta from Naja nivea venom and the disulfide bonds of toxin alpha . J. biol . Chem . 246, 7383-7394. Bons, D . P., $TRYDOM, D . l., AxnExsotv, C. G . and Cxx~stExseiv, P. A. (1971) Purification and properties of three toxins from N. nivea (Linnaeus) (Cape cobra) venom and the amino acid sequence of toxin delta . J. biol. Chem . 246, 3132-3139. Bout.exaeR, G. A . (1904) Descriptions of two new elapine snake from the Congo . Ann . Mag. Nat . Hist . 14, 14-15 . Buct~ot.z, W . and Peretes, W . (1876) Eine zweite Mittheilung liber die von Hrn Prof. Buchholz in Westafrika gesammelten Amphibien . Mber. dt . Akad. Wins . Berl . 1876, 117-123 . Cevnvs, J . F. and Fxn=.n~nr, M . (1970) An internal standard for amino acid analysis : S-beta-(4-pyridylethyl)L-cysteine . Analyt . Biochem . 35, 489-493. Ctmtsrnvsex, P. A. (1971) The venoms of Central and South Africa. In : Venomous Animals and Their Venoms, Vol . 1, pp. 43762 (BUCHERL, W ., Deuwreu, V . and Bucxt.EV, E . E. Eds) . New York : Academic Press . DoLw, L. (1886) Notice sur les Reptiles et Batraciens recueillis par M . le Capitaine Em. Storms dans la region du Tanganyika . Bull . Mus . R. Hist . Nat. Belg. 4, 151-160. KOPEYAN, C ., Vert Rn:rscxoTex, J ., Max~nxez, G ., Roc,-EUr, H . and MIRANDA, F . (1973) Characterization of five neurotoxins isolated from the venoms of two Elapidae snaky Naja haje and Naja nigricollis . Eur. !. Biochem. 35, 244-250 . Mottw~n, A . H ., Ktw.a., F . K . and Baser, A. A . (1977) Immunological studies on polyvalent and monovalent snake antivenins . Toxicon 15, 271-275 . Sci-n~tror, K. P . (1923) Contributions to the herpetology of the Belgian Congo based on the collation of the American Museum Congo expedition ; 1909-1915 . Part 2 . Snake . Bull. Am. Mus. Nat . Hilt. 49, 1-146. Sct~nor, J. J . and Mtuut.eexoox, J. L . (1989) Purification, sequencing and characterization of pseudexin phospholipasa A2 from Pseudechis porphyriacus (Australian red-bellied black snake). Toxicon 27, 805-818 . Swat, P. K., Kxotnr, R. L, HExst~xSON, G. T ., MALLIA, A. K., GnRrrme, F. H ., Pxovrxzexo, M . D ., Funeso~ro, E. K ., Goet~, N. M ., Otsox, B. J . and Kt.t:xtc, D . C . (1985) Measurement of protein using bicinchoninic acid . Analyt . Biochem . 150, 76-85 .

Lethal Toxins and Neutralization of Boulengerina Venoms

132 7

B. G . (1991) A non-radioactive receptor assay for snake venom post-synaptic neurotoxins . Toxicon 29, 503-510. Vim, J . and Ct~uPr~twv, D . S . (1978) Snakes and Snakebite. Venomous Snakes and Management ojSnakebite in Southern Africa, pp. 80-86. Cape Town : Struik . WmvsrnrN, S . A., MnsroN, S . A . and Wtr.ne, C. E. (1985) The distribution among ophidian venoms of a toxin isolated from venom of the mojave rattlesnake (Crotahcr sctttulatus scutulatus) . Toxicon 23, 825-844. Wsnvsreix, S. A ., Scr~nor, J . J., BERNHE~II~.R, A . W. and Scent, L . A . (1991) Characterization and amino acid sequences of two lethal peptides isolated from venom of Wagler's pit viper, Trimeresurus wagleri . Toxicon 29, 227-236. Wotet,o H~c.rtt Oao~xtznnoN (1981) Progress in the characterization of venoms and standardization of antivenoms. WHO Offset 13eb1. 58, 23-24 . $77LE4,

rox t9~rr-c