Comparative studies on three rattlesnake toxins

Toxicon, Vol. 23, No. 3, pp. 361-374, 1985 . Printed in Grat Britain .

0041-0101/BS 53 .00+ .00 ® 1985 Perpmon Preaf Ltd .

COMPARATIVE STUDIES ON THREE RATTLESNAKE TOXINS D.

STEVEN AIRD and IVAN I. KAISER* Department of Biochemistry, University of Wyoming, University Station, Boz 3944, Laramie, WY 82071, U .S .A . (Accepted for publication 1 August 1984) S . D . Auto and I . I . ICAISF-R. Comparative studies on three rattlesnake toxins . Toxicon 23, 361- 374, 1985 . - Toxins from the venoms of Crotalus durissus te~cus, Crotalus s. srutuiatus and Crotalus viridis cortcolor were compared using gel filtration, ion~xchange chromatography on DEAF-Sephacel and denaturing and non~enaturing polyacrylamide gel electrophoresis . The three heterodimeric native toxins behaved similarly on each of the separation media, except that the C. d. terr(fuvs toxin displayed a pronounced tendency to dissociate on DEAF-Sephacel, even in the absence of urea. In the presence of 6M urea, subunit dissociation was quantitative for all three tOldns . Recombination of purified subunits resulted in toxins which eluted from the gel filtration cohunn in identical fashion to native toxins . Non-denaturing polyacrylamide gel electrophoresic patterns of recombined toxins actually showed greater band resolution than did the native toxins . Six hybrid toxins were generated on polyatuylamide gels from cross-combinations of purified subunits, each with different mobilities than the parental toms . Mobilities of the hybrid toxins depended principally upon the mobilities of the basic subunits. All three purified native toxins showed comparable Ln,a's in female mice (0 .039-0 .061 Ng/g) . The C. d. terr(}uvs acidic x C. s. m4tulatus basic hybrid toxin showed toxicity identical to that of the C. s. srutulatus rxombined toxin . Phospholipase activity is associated with the basic subunit in all three toxins . Intact toxins show a distinctive lag in phospholipase activity which is not seen with purified basic subunits alone. These results indicate that the principal toxins in these three venoms are homologous . INTRODUCTION T`HE DISCOVERY Of `crotoxin', a toxin from the venom of the South American rattlesnake (Crotalus durissus tutus), by SLOTTA and FRAENKEL-CONRAT (1938) was an important milestone in snake venom research . This heterodimeric protein proved to be a potent enzymatic neurotoain . For nearly the next four decades, the false dogma persisted that neurotoxins were the exclusive province of elapid and hydrophüd venoms . Crotalid venoms were thought to possess only necrotizing components and C. d. terrificus was regarded as something of an enigma . Nearly thirty years after the isolation of crotoxin, RUSSELL (1967) reported exceptionally low LDP values for the venom of the mojave rattlesnake (Crotalus s. scutulatus). Seven years later BIEBER and Tu (1974) isolated an acidic toxin from this venom which was later designated a cardiotoxin (BIEBER et al., 1975). In the same year, HENDON (1975) identified its principal mode of action as neurotoxity . Over the years, evidence accumulated to suggest great similarity between the C. d. te~cus and C. s. scutulatus toxins in both structure and function . Three recent reviews provide a detailed " To whom reprint requests should be addressed. 361

362

STEVEN D. AIRD and IVAN I. KAISER

history of these events (HABERMANN and BREITHAUPT, 1982 ; FYtAENKEL-CONRAT, 1982-83) . POOL slid BIEBER (1981) reported the

1978 ; HENDON

and BEIBER,

presence of a component with properties similar to those of crotoxin in the venom of the midget faded rattlesnake (Crotales viridis concolor). The present study concerns itself with the homology of the three toxins and investigates this matter in terms of chromatographic and electrophoretic behavior, and with respect to biological activity . MATERIALS AND METHODS Lyophilized crude venom of Crotalus durissus tutus (lots CJSBJ and CJ13VZ) and Crotalus scutelatus scetulatus (lot CS 11B-1Z) were purchased from the Miami Serpentarium. Venom of Crotales virldis concolor was obtained from a colony of midget faded rattlesnakes maintained in our own venom production facility . Spectrapor and viscose process dialysis tubing were purchased from VAty WnTExs and RoaEtts. They were boiled in 0.1 M NaHCO,-0 .01 M EDTA, washed in deionized H,O and stored in 309k ethanol at 4°C until used . Sephacryl 5-200 and DEAF-Sephacel were obtained from Pha**narin , r.-a-Phosphatidylcholine, Type V-E from frozen egg yolk, was purchased from Sigma Chemical Co . All other chemicals were either analytical or reagent grade. Deionized water was used to prepare all solutions. Reagent grade urea (Baker Chemical) was dissolved to make a slightly greater than 8 M solution and stirred overnight at room temperature with Bio-ltad AO11A8 resin (1 g/100 ml). Following filtration through a 1.2 Ean Millipore filter, the volume was adjusted to give a final concentration of 8 M urea from which buffered chromatographic solutions were prepared on the day of their use. Fractionation of crude venoms Lyophilized crude venom was suspended in 4-8 ml of 0.1 M sodium acetate (pH 4.0) buffer and stirred at room temperature . Occasionally, if necessary, minimal amounts of 1 M HCl were added to enhance wlubilization of venom components . The sample was centrifuged (20°C) at 10,000 x ~ for 10-13 min to remove small amounts of precipitate. Chromatography was carried out at room temperature on either 1 .5 or 2.3 x 110 cm columns of Sephacryl 5-200 equilibrated and run with 0.1 M sodium acetate (pH 4.0). Flow rates were about 20 and 36 ml/hr on the 1.5 and 2.5 cm columns, respectively. Hybrid toxins to be used in t,n  studies were also purified by chromatography on 5-200 at room temperature. DEAF-Sephacel chromatography was carried out at room temperature on 1.5 x 34 cm columns at a pumped flow rate of 15 -20 ml/hr. A linear gradient of NaCI from 0 to 0.3 M, buffered with SO mM Trice- HCl (pH 8.3) was used in all elutiona with a total gradient volume of 400 ml . Recoveries of A,-absorbing material generally ran greater than 90% on both gel filtration and ion~acchange columns. Abaorbance measurements were made in a Gilford mode1240 spectrophotometer using silica cells with 1 cm path length . Pooled fractions were either wncentrated first in an Amicon unit using UM 10 membranes and/or dialyzed directly against deionized H,O at 4°C. Exhaustively dialyzed samples were recovered by lyophilization . Some pooled samples from 5-200 were dialyzed directly against theinitial equilibrating DEAF-Sephatel buffer (SO mM Trig-HCI, pH 8.3). Dialysis against deionized H,O invariably resulted in a pronounced white precipitate which redisaolved readily with an increase in solution ionic strength . No chromatographic differences on DEAF Sephacel were apparent between samples dialyzed against the ionexchange column buffer and those that were dialyzed against H,O, tyophilized and resuspended in the equilibrating buffer. Electrophorrsis of taut jractions Purified, native toxins, their subunits and hybrid toxins were analyzed with polyacrylamide gel electrophoreaia (PAGE) using 13~ resolving and 4.4Sto stacking gels . Bis-acrylamide constituted 2.7sJo of total acrylamide . Four discontinuous buffer systems were utilized. Sodium dodecyl sulfate (SDS) gels employed buffer strengths of 0.3 M Tris -HCl (pH 8.8) in the resolving gel and 0.1 M Tris -HCl (pH 6.8) in the stacking gel, both with 0.14ti SDS. Reservoir buffers were 0.025 M Tria-0.2 M glycine (pH 9.4) with 0.1% SDS. Non-denaturing buffer systems included ~-alanine-acetate, a reversed-polarity system, Tris-glycine and a modification of the Tris-alanine system of PAxtQt~tsoty d al. (1981) . In the ß-alanine-acetate system, the resolving and stacking gel buffers were 0.088 M acetic acid-NaOH (pH 3.80) and 0.038 M acetic acid-NaOH (pH 6.63), respectively . The upper, or anodal, reservoir buffer was 0.04 M ß-alanine, adjusted to pH 4.13 with glacial acetic acid . The lower, or cathodal, buffer was 0.032 M acetic acid-NaOH (pH 4.15). Resolving and stacking gel buffers for both the Tria-glycine and Tris-alanine systems were 0.373 M Tris -HCl (pH 8 .8) and 0.1 M Tris -HCl (pH 6.8), respectively . Reservoir buffers for the Tris-glycine system

Rattlesnake Toxins

363

contained 0.023 M Tris and 0.2 M glycine, (pH 9.4). Those for the Tris-alanine system contained 0.08 M Tris and 0.16 M alanine (pH 8.22). Gel buffer ionic strengths employed in this study were reduced from those employed in most comparable buffer systems reported in the literature . This reduction enhanced protein resolution . Gels were stained with Coomaasie brilliant blue G-250 using a modification of the formaldehyde fixation technique of S~recx et ol. (1980). The stain was addified to pH 2.3 with concentrated HCI. Destaining reagent, which also doubled as the gel storage preservative, contained 10 ml of 37ß'e formaldehyde, 100 ml of glycerol, 200ml of glacial antic acid, S00 ml of 9Sß4 ethanol and 1190 ml of deionized H,O. Lethality assays

A colony of min was established with stock obtained from the Center for Disease Control in Fort Collies, Colorado. The CDC colony originally came from NIH and was identified as "the NIH all-purpose strain". The lyophilized toxin was dissolved in phosphate-buffered saline pH 7.1 (8 .0 g NaCI, 0.2 g KCI, 0.2 g KH,1?O 0.15 g Na,HPO, per liter, pH adjusted to 7.1 with NaHPO,) for injection. Stock solutions for injection were diluted to 0.1 times the intended dose. That is, 0.005 Ng/pl stock solution was used for injections at the dosage of 0.03 pg/g mouse. The volume igjetted per mouse (in Fl) was then 10 x the weight of themouse in grams (e .g. a 20 g mouse received 200 pl of a 0.005 pg/Nl solution resulting in a dose of O.OS Ng/g, based on dry weight of toxin) . Injections were made i.v . in either of the dorsolateral caudal veins, using a Sh" x 27 guage needle and a tuberculin syringe, and were considered accurate to the nearest 10 pl. A total of 300 mice, ranging from 20-40 g, were employed in the present study. Ten mice of the appropriate sex were injected at the indicated toxin concentration (Fig. 6) . Mice were placed in a 37°C incubator for 3-S min prior to injection in order to induce slight vasodilation. This procedure greatly enhanced the injection success. Animals were immobilized for injection by placing them in a SO ml polyethylene centrifuge tube with the tip rtmoved for ventilation. The mouth of the tube was then blocked with a notched cork so that only the mouse's tail protruded. After injection, each test animal was scored as to whether all, part or none of the material was injected into the vein . The time of injection was recorded and the outcome (survival or death) was recorded at 24 hr postinjection. Partially successful injections were countedif the mouse died, but not if it survived . The reasoning for this was that if a mouse died after receiving a partial done, it certainly would not have survived a full dose . Partials that survived, however, could not be assigned reliably to either outcome class and they were excluded from the date analysis . The excluded animals resulted in sex-dose cell sizes ranging from 7-10 mice . The data were analyzed using BMDP programs 1D and LR (Department of Biomathematics, University of California, Los Angeles, California) on a Control Data Corporation Cyber 760 computer . The stepwise logistic regression analysis was weighted per thenumber of storable injections persex-dose cell in order to ascent for variability due to smell all sizes. Stepping computations were based on the asymptotic covariance estimation procedure. Phospholipase assays

Phospholipase activity against egg yolk t,-a-phosphatidylcholine wes determined using a Radiometer PHM 82 pH meter equipped with a TTT 80 titrator, an ABU 80 auto burette and a TTA 80 titration asembly. Fatty acids released by the reaction were titrated to pH 8.0 with dilute NaOH (c. 0.04 M) under nitrogen at 37°C. The reaction mixture was unbuffered . The substrate was prepared using a modification of the procedures outlined by Cormaae.+ et al. (1980) . Ninety milligrams of ~-o-phosphatidylcholine, dried under N,, was suspended in 9.0 ml of 12 .5 mM TYiton X-100 (estimated mol. wt of 646) using a Virsonic Cell Disrupter, Model 16-830 . (Vutis Company, Gardener, Nln: When the suspension was complete, 1.0 ml of 0.1 M CaCI, was added and the substrate -CaCI, mixture wes then briefly resonicated. This yielded a roughly 2:1 molar ratio of Tritron X-100:phospholipid. Phospholipase activity was assayed for 20 min after the addition of 2.5-40 hg of enzyme. RESULTS Chromatographic Pu~cations Crudè venoms from Crotales d. tutus, Crotales s. scutulatus (the most toxic form only, GLENN et al., 1983) and Crotales v. concolor all have reported i.v . LD values in the range 0.20-0.35 Etg/g of body weight in mice (GLENN and STRAIGHT, 1982). Initial fractionation on Sephacryl S-200 revealed a toxic component (the only fraction with an LDP less than 3 x the LDP of the crude venom), which eluted similarly in all three venoms (Fig. 1) . In C. d. ter~cus venom (Fig . lA) this fraction represented 65% of the recovered A,~-absorbing material, which is consistent with the high crotoxin content of this venom

STEVEN D. AIRD and IVAN I. KAISER

364

~10 C. d. terrificus

t

A

N S W U 6 Z

a

O

ma

4 2 °0

10

20

~o ao so so

70

~s N4

ui U3 z a ~2 a 0

0

10 20 30 40 50 80 70 FRACTION No. (9ml/tube)

FIO . 1 . GEL FILTRATION OF CRUDE VENOMS ON SEPHACRYL S-i0o . (A) C. d. te~cus venom (400 mg); (B) C. s. srutulatus venom (399 mg); (C) C. v. rnncolor venom (213mg). The venoms were passed over a column of Sephacryl S-200 as described in the experimental section. Protein concentrations were determined by absorption at 2g0 nm . Fractions pooled for subsequent chromatography on DEAE~ephacel 'are indicated by the bars .

reported by FRAENKEL-CoNRAT (1982 - 83). In C. s. scutulatus and C. v. concolor venoms the corresponding peak represented 37% and 19%, respectively, of the recovered A,so-absorbing material (Figs . 1 B, 1C) . The indicated tubes (Fig . 1) were pooled and fractionated further on DEAF-Sephacel . Ion-exchange elution profiles for all three species are illustrated in Fig. 2. All three

3(,3

Rattlesnate Toxins

ôo.a 0 z t 0 .4

a 0 0

so

4o so so FRACTION No. (4 .amVtub~)

~oo

0

tso

FIG. 2. DEAE-SEPHACEL COLUMN CHROMATOGRAPHY OF POOLED FRACI70NSFROM SEPHACAYL S-200 RUNS (FIG . 1) . (A) C. d. t~cus (SO mg ; 73 A,-units); (B) C. s. srutulatus (150 A,-units); (C) C. v. cnncolor (47 .9 Ae-units) . Fractions were dissolved in SO mM Tris-HCI (pH 8 .3) buffer, pumped on to a DEAE-Sephacel column, and doted as described in the experimrntal section. Absorbance of each fraction at 280 run was determined. The indicated fractions were pooled, dialyzed against ddonized H,O and lyophilized.

absorbance profiles were basically similar, with an early-eluting fraction that exhibited little binding to the cohunn (tubes 10 - 30), a second peak of variable size that eluted shortly after initiation of the NaCI gradient (tubes 40 - 60), and a third major peak or pair of peaks that constituted the major component (tubes 62 -100). A fourth peak was recovered when the column was flushed with 0.6 M NaCI. The major component from each species exhibited an i.v. LDso of less than 0.1 Etg/g in mice . PAGE experiments with the DEAE-column fractions from C. s. scutulatus and C. v. concolor suggested that the first, second and fourth peaks constituted protein contaminants unrelated to the

366

STEVEN D. AIRD and IVAN I. KAISER

E e 0 ao a ui

v z m a~ 0 m m

S

10 -

16

Z

FRACTION No. (S .4m1/tubs) FIG. 4. DEAF-SEPHACELCHROMATOGRAPHY OF PURIFIED C. Y. COnCOiOr TOXIN IN THE PRESENCE OF 6 M UREA. C. v. contolor toxin (15 mg ; 22 .7 A,-units) from a DEAE-Sephacel run in the absence of urea (Fig . 2C) was dissolved in 2.2m1 starting buffer (SO mM Tris-HCI, pH 7.2, 0.1 M NaCI, 6 M urea) and applied to a column (1 .5 x 7.5 cm) of DEAE~ephacel at room temperature. Following a brief wash with the starting buffer to elute the basic subunit, a steep linear gradient waa started to elute the acidic subunit. The gradient reservoir contained SO mM Tris-HCl (pH 7.2), 2 M NaCI, 6 M urea and a total gradient volume of 1SO ml . Flow rate was about 20 ml/hr. Recovery of A,-material was 974'0 . Absorbance of each fraction was determined at 280 nm and the indicated fractions were pooled, dialyzed against deionized water using Spectrapor 3 tubing and lyophilized .

FIG . 6. INTRAVENOUS LD,o ASSAYS OF NATIVE TOXINS, RECOMBINED AND HYBRID TOXINS IN MICE.

Open circles denote assays in male mice . Closed circles denote female mice .

FIa . 3. A COMPOSTIB OF SAMPLES FROM C. d.

tEiIjfICüS ON 1S % POLYACRYLAMIDE SLAB GELS STAINED wITfI COOMASSIE BLUE .

Toxins in Tris-shrine (pH 8.8) gel. Well 1: early~luting half of doublet from DEAESephaoel ; 18 .5 Ng (tubes 65 - 70). Well 2: later~luting half of doublet ; 19 .8 Pg (tubes 71- 77). Well 3: blank. Well 4: combination of suto~lissociated basic and acidic suborns; 14.8 Ng (peaks 1 and 4, respectively). Note that only later~luting-like material results . (B) Tris-glycine (pH 8 .8) gel. Fifty micrograms of protein per well . Well 1 : crude venom. Well 2: basic suborn (does not enter gel) . Well 3: polymorphic acidic auburn . Small band at top of lane is a contar+irnr t. Well 4: mixed acidic and basic suborns. Note the production of intact toxin and the equal depletion of both acidic auburn forms . Suborns used were autodissociated. (C) ~-Alanine-acetate gel (pH 3 .8). Fifty micrograms of protein per well . Well 1 : crude venom. Well 2: basic auburn. Note multiple forms. Well 3: acidic auburn (does not enter gel). Well 4: mixed acidic and basic suborns. Note the formation of dimeric toxin and the proportional disappearance of all forms of the basic suborn. Suborns used were auto-dissociated . (A)

368 1

2

1

3

2

4 5 8

3

4

5

7 8

8 7

5 1011 121314 15 181718

8 5101112131418161718

FIG. S. NON-DENATURING PAGE OF CROTALID TOXINS AND THEIR SUBUNITS ON 1S% OEIS ; TRIS-ALANINE (PH 8 .8) (A) OR ß-ALANINE-ACETATE (PH 3 .H) (B~ .

Twenty-two to 32 Ng of protein was added to each well . Contents of wells : 1 . native C. d. terr~cus; 2 . C. d. ter~cus recombined ; 3 . C. d, terr;/`tt~us acidic subunit; 4 C. d. fucus basic subunit ; 5 C. d. fucus acidic x C. s. scutulatus basic hybrid ; 6 . C. d. terr~cus acidic x C. v . concolor basic hybrid ; 7 . native C. s . scutulatus; 8 . C. s. scutulatus recombined ; 9 . C. s. scutulatus acidic subunit ; 10. C. s, scutulatus basic subunit; 11 . C. s. scutulatus acidic x C. d. ter~cus basic hybrid ; 12 . C. s. scutulatus acidic x C. v. concolor basic hybrid; 13 . native C. v. concolor, 14 . C. v. concolor recombined ; 1 S . C. v. concolor acidic subunit ; 16 . C. v . conrnlor basic subunit ; 17 . C. v. concolor acidic x C. d. terr~cus basic hybrid ; 18 . C. v. concolor acidic x C. s . scutulatus basic hybrid .

Rattleanal~e To~na

369

major toxin. Mixing of the above components failed to produce a component with the electrophoretic characteristics of the intact toxin (no interaction between the apparent cOIItAminan tg was observed), suggesting that these lesser peaks did not contain dissociated subunits of the toxin. Recombination experiments with the C. d. tutus DEAF fractions were, however, quite different. In this case, peaks one and four (Fig . 2A) consisted of the dissociated basic and acidic subunits, respectively. A component having the mobility of the intact toxin was consistently observed on polyacrylamide gels following mixing of these two peaks. DEAF peak two from C. d. te~cus was a minor contaminant . All DEAF separations of C. s. scutulatus and C. v. concolor venoms yielded consistent A, profiles. However, because of the tendency of the C. d. tutus toxin to dissociate spontaneously, the DEAF fractionations of the latter venom were difficult to reproduce consistently. Results ranged from virtually no dissociation of the toxin to nearly total dissociation . Other workers have indicated similar difficulties (HENDON and F~tAENKHL-CONRAT, 1971). Different lots of C. d. te~cus venom also behaved differently. For instance, lot no . CJ13VZ (Miami Serpentarium) yielded a clear doublet (Fig . 2A) instead of a single peak . Recombination experiments on gels with peaks one and four yielded two bands, with mobilities corresponding to bands present in the later-eluting

Fto . 7.

PHOBPHOLiPA38 A9 .4AY3 ON DIFPEBHPTC Quexrrr~s oF

ivw~ ~roxnv (A) ~xn sueurnr (B) trnoM Crotalus vtrtdit croneolor.

a+s~c

370

STEVEN D. AIRD and IVAN I. KAISER

half of the doublet recovered from DEAF-Sephacel (Fig . 2A). Essentially, none of the earlier-eluting peak of the doublet was generated (Fig . 3A). The implications of this experiment will be discussed later. Separation of the subunits was achieved by passing the toxic fraction from DEAESephacel over a second DEAE-Sephacel column in the presence of 6 M urea (HENDON and FRAENKEL-CONRAT, 1971 ; LATE and aIEBER, 1978). All three toxins exhibited elution profiles similar to that for the toxin from C. v. concolor venom (Fig . 4) . No differences between the C. d. terr~cus subunits isolated in this manner and those which were isolated by virtue of their spontaneous dissociation could be detected with PAGE . Electrophoretic analyses of subunits, native toxins, recombined toxins and hybrid toxins Electrophoretic analyses on 15% gels indicated the presence of four to five structural variants in each of the three toxins (Fig . S) . Based on the mobilities of the toxins from the three taxa, the C. s. scutulatus toxins are, on the average, the most acidic, those of C. v. concolor tend to be the most basic, while those of terr~cus exhibit intermediate charge density, assuming similar molecular shapes and volumes. Isoelectric focusing gels run on these intact toxins confirm that the relative isoelectric points are the same as suggested by their mobilities (A . L. BEIBER, personal communications). Because of the apparent range of pI's of the isomers in any one species, there is some overlap between the isomers from the three taxa (Fig . S) . Recombined toxins (reassociated subunits originating from the same native toxin) displayed mobilities indentical to those of the native toxins . Reassociations employed a slight excess of the acidic subunit. As a result, free acidic subunit can be clearly seen in these wells (Fig . SA, wells 2, 8 and 14). Purified subunits behaved as anticipated on non-denaturing PAGE . The acidic subunts migrated rapidly in Tris - alanine (pH 8.8) gels (Fig . SA), but did not enter ßalanine - acetate (pH 3.8) gels (Fig . SB). The acidic subunit, by virtue of its low mol. wt (9500), was readily leached from the gels by prolonged destaining. The basic subunits, which were highly polymorphic, like the intact toxins (Fig . 3C, well no . 2), exhibited greater mobilities than the intact toxins in ß-alanine - acetate gels (Fig . SB), but they did not enter alkaline gels (Fig . SA). All isomers of the basic subunit recombined readily with acidic subunit to form the dimeric toxin (Fig . 3C, well no . 4) . These results are consistent with those reported by LATE and BIEBER (1978) for the C. s. scutulatus toxin. The acidic subunts appear to be entirely monomorphic when examined on Tris - alanine gels (Fig . SA), but alkaline gels run in a Tris - glycine (pH 9.4) buffer system resolved two components in all samples (Fig . 3B, well nos. 3 and 4) . The two components ran identically in all three species. Both components appear to have similar affinity for the basic subunit, since they both disappear at equal rates during titration with basic subunit to form intact toxin (Fig . 3B, well nos . 3 and 4) . All possible combinations of subunits from the three species were also examined, and all combinations yielded heterodimeric proteins that behaved electrophoretically like the native toxins . Under alkaline conditions, for any given subunit, mobility of the hybrid toxin reflected that of the basic subunit employed (Fig . SA). The relative mobilities of the hybrid toxins thus formed parallel those of the native toxins, C. s. scutulatus ~ C. d. terr~cus ~ C. v. concolor. For any given basic subunit, varying the acidic subunit did not affect the mobility of the hybrid toxin, except in the case of the C. v. concolor acidic subunit. The latter appeared to cause a slight reduction in mobility . Under acidic gel conditions, free basic subunts exhibited the same relative mobilities as the native toxins (C. v. concolor ~ C. d. terrificus ~ C. s. scutulatus; Fig. SB). When hybrid toxins were formed with the acidic subunit from one species and basic subunts

Rattlesnake Toxins

371

from the other two, hybrids composed of C. v. concolor basic subunit migrated faster than those containing C. d. ter~cus basic subunits, which in turn migrated faster than hybrids containing C. s. scutulatus basic subunits . Varying the acidic subunit for any given basic subunit again had a negligible effect on the mobility of the hybrid toxins, except where the C. v. concolor acidic subunit was concerned. It appeared to increase the mobility of the hybrid toxins slightly . These electrophoresic comparisons of hybrid and reconstituted toxins indicate that the basic subunit has a greater influence on mobility than the acidic subunit. SDS-gel electrophoresis in the absence of sulfhydryl reagents resulted in dissociation of all three toxins into separate basic and acidic subunits . There was no evidence that the subunits of the three toxins differed in mol. wt . Disintegration of acidic subunits into component polypeptide chains did not occur in the absence of sulfhydryl reagents . Lethality assays The i.v . r.n,~ for the C. s. scutulatus toxin (0.030 pg/g in male mice and 0.039 in female mice) has been previously reported (AIRD and KAISER, 1985). The i.v . Ln,~ for C. v. concolor toxin in female mice was determined to be 0.045 pg/g, while that for C. d. ter~cus in male mice was 0.047 pg/g (Fig . 6) . The latter corresponds to an i.v . r n,~ in female mice of 0.061 Ng/g, or about 1 .6 x the C. s. scutulatus r.D,~ . These values for the C. s. scutulatus and C. d. terr~cus toxins correspond well with those previously reported in the literature (LATE and BIEBER, 1978 ; RussA~x et al., 1971), however, the value for the C. v. concolor toxin is significantly lower than the 0.12 ~g/g reported by Poor. and BIEBER (1981) for partially purifiod material . The recombined C. s. scutulatus toxin, formed by dissolving the two subunits in phosphate-buffered saline and then combining the two volumes of saline, had an i.v . Ln,~ of 0.066 pg/g, nearly twice that of the native toxin. A hybrid toxin, produced in the same fashion as the C. s. scutulatus recombined toxin but utilizing the C. d. te~ce~s acidic subunit rather than that from C. s. scutulatus, had an Ln~, identical to that of the C. s. scutulatus recombined toxin (0 .066 kg/g) (Fig . 6) . A second hybrid toxin, utilizing C. v. concolor acidic subunit and C. s. scutulatus basic subunit, was also tested . Unlike the former hybrid, this one exhibited no lethality at doses as high as 0.27 F~g/g. Phospholipase assays Phospholipase assays demonstrated that the C. v. concolor toxin, like the other two, possesses phospholipase activity which is associated with the basic subunit, but not the acidic subunit (HENDON and FRaE11xEL-CoxRnT, 1971 ; CnTE and BiESER, 1978 ; Pool and BIEBER, 1981). When the intact toxin was assayed against L-a-phosphatidylcholine from egg yolk, a sigmoidal relationship between paroles of fatty acid released and time was evident (Fig . 7A). The basic subunit displayed a hyperbolic relationship (Fig . 7B), with no lag period at the outset . These findings are consistent with those reported by LATE and BIEBER (1978) for the C. s. scutulatus toxin and by RuBSAMEN et al. (1971) for the C. d. te~cus toxin. initial rates of hydrolysis were calculated for the basic subunits of all three toxins during the first minute of catalysis only, in order to avoid the inhibitory effect of free fattyacids liberated by the reaction . Quantities of basic subunit larger than 10 Ng (per ml of reaction mixture) resulted in significant nonlinearity during the first minute . Under our experimental conditions with 10 pg of each basic subunit, C. v. concolor hydrolyzed the substrate most rapidly (179 pcnoles/min/mg of enzyme), followed by C. s. scutulatus (154 moles/min/mg) and C. d. tet~cus (85 pmoles/min/mg) .

372

STEVEN D. AIRD and IVAN I . KAISER DISCCISSION

The two-step purification procedure employed in this study proved efficacious for the C. v. concolor and C. s. scutulatus toxins, but the fractionations were not consistently reproducible with the toxin from C. d. tutus venom. Electrophoretic analyses of crude venoms and DEAE-Sephacel fractionation of the third 5-200 fraction reveal that the venoms of C. s. scutulatus and C. v. concolor are similar in terms of complexity. Both of them appear to contain many more components than does the venom of C. d. tutus. This relationship is not overly surprising in the light of currently accepted rattlesnake phylogeny (KLAUBER, 1972), although the relative simplicity of C. d. te~cus venom is not easily explained . The simplicity of the latter venom is fortuitous considering the tendency of the neurotoxin to dissociate on DEAE-Sephacel at slightly alkaline pH values, a130 noted by HENDON and FRAENKEL-CONRAT (1971) on carboxymethyl cellulose at pH 4.0 . Different lots of C. d. terr~cus venom behaved differently on the ion exchange column . In most cases, the intact toxin eluted as a single peak, which sometimes displayed a slight leading shoulder . One batch, however, eluted as a clear doublet (Fig . 2A). When the autodissociated subunits from that material were recombined on gels, only the component present in the trailing half of the doublet was generated (Fig . 3A). A plausible explanation for this phenomenon and for the apparently greater stability of the C. s. scutulatus and C. v. concolor toxins may be primary sequence differences . Certainly, a few charge differences in the subunit binding regions could alter the affinity of the acidic subunit for the basic. Sequence work, currently in progress, may enhance our understanding of subunit interactions . JENG and FRAENKEL-CONRAT (1976) reported that the toxicity of the C. d. terr~cus basic subunit could be enhanced Crfold by complexing it with volvatoxin A=, the smaller subunit of a heterodimeric mushroom cardiotoxin (mol . wt X25,000; pI~S). Complex formation between these two different toxin components was considered specific since the C. d. te~cus basic subunit was not complexed by any other proteins with isoelectric points similar to that of volvatoxin A, . Still, the maximum toxicity was only one-fourth that obtained by recombining the acidic and basic subunits of crotoxin . Electrophoretic gels showed considerable polymorphism in each of the pooled samples, which may be attributed to both the acidic and basic subunits (Fig . 3C, well no . 2) . The polymorphism in this and other venom proteins (LEVY and BDOLAH, 197 appears to be genetic in origin (S . D. Aird, Ph .D . dissertation in preparation) . Earlier work employing PAGE has often failed to detect multiple alleles of a given toxin because of poor resolution stemming from excessive gel ionic strengths and resolving gel concentrations which were too low to provide adequate sewing . Separation on DEAF-urea columns and non-denaturing and SDS electrophoresis gels provide the first evidence that the toxin from C. v. concolor venom has a subunit structure identical to that of the other two toxins . Recombination experiments indicate that any of the three acidic subunits is able to complex any of the basic subunits, suggesting great similarity (Fig . 5) . Homology of these three toxins is further suggested by the demonstration of full biological activity of one of the hybrid toxins (Fig . ~. Immunological evidence (A . L. Hieber, personal communication) and the hybrid compatibility lend credence to the suggestion of BIEBER et al. (1975) that this toxin may be widely distributed among rattlesnake venoms . The reason for the reduced lethality of the C. s. scutulatus recombined toxin and of the C. d. te~cus acidic - C. s. scutulatus basic hybrid toxin is not entirely clear. It seems

Rattlesnake To~dna

373

likely that simple mixing of the subunits in buffers of low ionic strength results in rapid complex formation that does not permit minor structural adjustments essential for full activity. It is conceivable that if the hybrid toxins were formed in the presence of urea and then dialyzed, the resultant protein would show normal toxicity . LATE and BIEBER (1978) found that subunits of the C. s. scutulatus toxin recombined under such conditions resulted in a dimer with full toxicity . The absence of any observable deleterious effects of urea on biological activity of the toxins is probably attributable to the high degree of disulfide cross linking in each subunit. We are unable to explain the apparent lack of toxicity for the C. v. concolor acidic x C. s. scutulatus basic hybrid toxin. The levels of phospholipase activity, which are comparable for the three toxins and which agree well with those reported by both RUBSAMEN et al. (1971) and CATS and BIEBER (1978), lend additional support to the hypothesis that they are homologous . Results of physical and chemical studies currently in progress also suggest great similarity . Considering these findings, it would be surprising if the toxin from C. v. concolor did not turn out to be a neurotoxin. Acknowledgements-We thank Dr A. L. BmHeR of Arizona State University for ruining isoelcctric focusing gels on several of our samples and Mrs JwcQue Ww~rws for typing the manuscript and for drawing the figures. This work was supported in-part by a Faculty Research grant-in-sid from the University of Wyoming. REFERENCES Arno, S. D. and ICwrsert, I. I. (1985) Toxicity assays . Toxicon 23, 11 . Be~se, A. L. and Tu, A. T. (1974) Purification from an acidic toxin protein from the venom of the mojave rattlesnake. Fedn Proc. Fedn Am. Socs exp. Biol. 33, 1564 . Bmaett, A. L., Tu, T. and Tu, A. T. (1975) Studies of acidic cardiotoxin isolated from the venom of mojave rattlesnake (Crotahrs srutulatrrs) . Blochim. blophys. Acts 400, 178. Cwre, R. L. and Bn?aeA, A. L. (1978) Purification and characterization of mojave (Crotales scutelatus) toxin and its subunits . Archs Biochim. Biophys. 1g9, 397. Coxoaew, E., YANC3, C . C ., and RosarreEAa, P. (1980) Comparison of a relatively toxic phospholipaae A, from Ngja nigrlc»llis snake venom with that of a relatively non-toxic phospholipase A, from Hemaehotes haemachatus snake venom-1 . Enzymatic activity on free and membrano-bound substrates . Biochem. Pharmac. 29, 1555 . Fhr~etv~c~±.-Cor~nwT, H. (1982-83) Snake venom toxins related to phoapholipese A, . J. Toxic.-Toxin Rev. 1, 205. F L-Cor~xwr, H. (1983) Crotoxin from 1938 to 1983 . Fedre Proc. Fedn Am. Socs exp. BioL 42, 2186 . GLExx, J. L. and STawtaNr, R. C. (1982) The rattlesnakes and their venom yield and lethal toxicity . In : Rattlawrake Yenoms. Their Actions and Tirotment, p. 3 fht, A. T., Ed .) New York : Marod Dekker . GLSrav, J. L., STAwtattz, R. C., Wor.Fe, M. C. and HwxnY, D. L. (1983) Geographical variation in Crotahrs scutulatus scretulatus (mojave rattlesnake) venom properties . Toxirnn 21, 119. Ilwsantwwtv, E. and Ba~tTttwurT, H. (1978) The crotoxin complex-an example of biochemical and pharmacological protein complementation. Toxicon 16, 19. HBrwox, R. A. (1975) Preliminary studies on the toxin in the venom of Crotahrs acvtulatus (mojave rattlesnake) . Toucan 13, 477. Hsxnox, R. A. and Bmeatt, A. L. (1982) Preaynapdc toxins from rattlesnake venoms . In: Rattlesnake Yenoms: TheirActions and 7Yeatment, p. 211 fIt~, A. T., Ed .) New York : Marcel Dekker . Haxnox, R. A. and Fnw~r.~oxtuT, H. (1971) Biological roles of the two components of crotoxin. Proc. natn. Ac~l. Sci. U.S.A . 6i, 1560 . Java, T .-W. and Fnwwt~t.-Co~tnwT, H. (1976) Activation of crotoxin B by volvatoxin A, . Biochem. blophys. Res. Commun. 70, 1324 . Kr.wueea, L. M. (1972) Rattlesnakes. 77rdr habits, Ljje Histories, acrd Ighuenae on Mankinal, p. 168, Berkeley: University of California Press. Levy, Z. and BnoLwx, A. (1976) Multiple molavlar forma of snake venom phosphodiesterase from T?pea palesttnae. Toxicon 14, 389. Pwaxnvsox, A. M., Donx, A. R., MAPLHS, P. H. and Bnovi.~, R. H. (1981) Improved polyacrylamide gel eledrophoreaia with different amino acids as the trar7ing constituent. Analyt. Blochem. 117, 6. Pool ., W. R. and Bmaea, A. L. (1981) Fractionation of midget faded rattlesnake (Crotahrs viridis rnncolor) venom: lethal fractions and en~yn±~+~~ activities. Toxirnn lf, 517.

374

STEVEN D . AIRD and IVAN I . KAISER

BaerrEtnupr, H . J . and HnaERAUtarr, E. (1971) Biochemistry and pharmacology of the crotoxin complex - I . Subfractionation and recombination of the crotoxin complex . Naueyn-Schmiedebags Archs Pharmac. 270, 274 . Russat.t., F. E . (1%7) Pharmacology of animal venons . Clin . Pharnrar. Then. a, 849 . St.o1-re, K . H . and FAwartr~L~orrx~r, H . (1938) Schlangengifte III . Mitt . Reiningung and kirstallisation des Iclappaschlangen-giften . Ba. dt. them. Ges . 71, 1076 . Sz~c, ß ., L$[rrttNtn, P . and BOttx, R . R . (1980) Detection of basic proteins and low molecular weight peptides in polyacrylamide gels by formaldehyde fixation. Arralyt. Biochem. 107, 21 . RUHSAMCN, K .,