Snake Venom Toxins

The venom of the black mamba (Dendroaspis poZyZepis) was fractionated by chromatography on Amberlite CG-50. Two of the toxins were purified, and their amino acid sequences were determined. Toxin QI is a 60-residue protein cross-linked by four disulfide bridges; it is novel among snake venom toxins in containing 4 tyrosyl residues as opposed to the usual single tyrosyl residue. Toxin y is a ‘72residue protein with five disulfide bridges and is the first snake venom toxin isolated to contain 2 tryptophanyl residues. Toxins LY and y have subcutaneous LD50 values of 0.09 and 0.12 pg per g of mouse, respectively.


SUMMARY
The venom of the black mamba (Dendroaspis poZyZepis) was fractionated by chromatography on Amberlite CG-50. Two of the toxins were purified, and their amino acid sequences were determined.
Toxin QI is a 60-residue protein cross-linked by four disulfide bridges; it is novel among snake venom toxins in containing 4 tyrosyl residues as opposed to the usual single tyrosyl residue.
Toxin y is a '72residue protein with five disulfide bridges and is the first snake venom toxin isolated to contain 2 tryptophanyl residues. Toxins LY and y have subcutaneous LD50 values of 0.09 and 0.12 pg per g of mouse, respectively.
The black I~~al~~biL (Dev~dronspis polylepis) is popularly the most dreaded snake in .Ifrica, l)resumably because of the reportedly short death time of \-ictims, the size of the snake, and the fierceness of its attacks.
Very little ia knows ahout its venom and even less about the toSill ill the venom.
The sample \olrimr clicl not exceed 1 ml, and the columns were eluted at a rate of 15 ml per hour; 3-ml fractions were collected after mollitorina at 240, 260, and 280 nm with a Beckman Spectrochrom model 130.
Reduction and X-Carboxymethylation of Toxins-Tosins o( and y were reduced with P-mercaptoethanol and S-carboxymethylated with iodoacetate as described (4). When dithiothreitol (14) was used as reducing agent, the same degree of reduction was obtained. Tozicity Test-Column eluates were checked for toxicity, and toxicity of the pure fractions was determined, by subcutaneous injections into white mice weighing 20 to 25 g. Death within 24 hours was taken as indicating toxicity.
.lmino Acid Analyses--Proteins and peptides were hydrolyzed with constant boiling HCl at 110" for 18 hours as described (4). Values for the serine and threonine contents of the peptides were corrected by 8% and 4%, respectively, to compensate for destruction during hydrolysis.
Half-cystine was determined as S-car-bosymethylcyst~eine after reduction and S-carboxymethylation.
E&nation of Trypfophan Content-The tryptophan content of t.osins o( and y were determined by using ultraviolet difference spectra between these toxins and toxins 01 of Nuja huje and Q( of ll--crja nivea, respect'ively.
The latter two toxins have been characterized completely with respect to amino acid composition and sequence (4,5). Their similarity in size to toxins rr and y of D. polylepis venom enabled the contributions to the ultraviolet spectra by the invariant tyrosyl and tryptophanyl residues to be subtracted from the ultraviolet spectra of toxins LY and y. With the tyrosine content of both toxins (Y and y known from amino acid analyses, the t.ryptophan contents were easily obtained. In some cases the tryptophan content of peptides was determined by a combination of pronase and aminopeptidase M digestion, followed by analysis on the amino acid analyzer. Digestion with Proteolytic Enzymes-Digestion with trypsiu and chymotrypsin was carried out at 37" in 2y0 ammonium bicarbonate solution at pH 8. Tryptic digests were carried out with 1 y0 (w/w) trypsin for 2 hours. Toxin 01 was digested with 2% (w/w) chymotrypsin for 3 hours, and toxin y digested exhaustively with 3% (w/w) chymotrypsin for a total of 7 hours. Thermolysin (lo/ w/v) digestion of peptides was carried out at 50" in 2% ammonium bicarbonate solution at pH 8. Papain was used for digestion of peptides at pH 6.3 as described (4).
Asparagine and Glutamine-Asparagine and glutamine contents of peptides were determined by examining the electrophoretie mobilities of appropriate peptides at pH 6.5. Alternatively, it was determined by digestion with aminopeptidase M or a combination of pronase and aminopeptidase M, and analysis of the enzymatic hydrolysates on the automatic amino acid analyzer, Blank runs were included as a check on autodigestion.
Puriification of Pepticks-Peptides were purifred and their homogeneity tested by descending paper chromatography and i~omenclature-l'el~tides derived from tryl)t,ic :IIK~ ohymotr~~~tic digests of t.osins ar and y are prefised T :wd (y, resljectiwly, and are numbered consequently accordiug t.o their posit ion:: in the polypept.ide chain. Pept.ides derived from the original lqtidc:: by further degradatiou are similarly distinguislletl by appendiug TL (thermolysin) or P (papaiu) to the symbol for the parent prptide. In all tables the numbers given in pnrent,heses after the analytical values signify the assumed int.egral \-alues for thr residues per molecule of pure peptide.

Fraction&ion
of TI-hole Benom-Cllt,omatograi~~t~-of D. polylepis venom on Atnberlite CG-50 gave the chrotnatogram of Fig. 2. Toxicity was widely spread among the fractions as is indicated on t.he clworttu.togram. The same effecta noticed with fractions of D. crngusfieeps venom (a), were found for Fractions B, C, I, and Ii, namely Ion: toxicity, long death times, a "fluffed" appearance, and festering of the eyes. This places these fractions in a different rategory from the other toxic. fractions.
Fraction E caused dent,h within 4 t,o 4) min, by subcutaneous injection of a large overdose. Such short death times have not been found for any of the otlter toxins from snake \-enoms so far examined, the shortest death times being in the range of 7 to 8 min. Fractions E attd F (containing toxins (Y and y, respectively) were lgophilized.
Putification of Z'ozin y-The lyophilized Fraction F was eluted through Sephadex G-50 in a similar manner as Fraction E above. Both of the fractions obtained from G-50 (Fig. lb) were toxic. The least retarded peak contained toxin y. After rechromatography of this fraction through carboxymethylcellulose, pure toxin y was obtained (Fig. 3)  was found for toxin (Y by the rapid sedimentation equilibrium method of Yphantis (12). The amino acid analyses of toxins cy and y are presented in Table I. The difference spectrum of toxin a and toxin o( of Naja haje venom showed that mamba toxin a! contained 3 more tyrosyl residues than Seja haje toxin (I! and also only 1 tryptophanyl residue. The spertrum of 1 tryptophanyl residue remained when the spectrmll of t)orin (Y of Naja ni vea venom was subtracted from the spectrum of toxin y. Toxin y therefore contains 2 tryptophanyl residues. No S-carboxymethylcysteine could be demonstrated in hydrolysates of unreduced S-carboxymethylated toxins (Y and y. The Tryptic and Chymotryptic Peptides of Reduced and S-Carboxymethylated Toxin a-Dansyl-Edman degradat'ion of toxin o( yielded Arg-Ile-Cys(Cm)-Tyr-Asx as NHz-terminal sequence. Tritium labeling of the reduced and S-carboxymethylated toxin gave tyrosine as COOH-terminal amino acid. The tryptic digest of toxin OL was fractionated by paper chroma,tography with Solvent II. Peptides T-3, T-4, and T-6 required no further purification.
Peptides T-l and T-2 mere separated from each other by electrophoresis at pH 1.9, while the fraction containing Peptides T-3a, T-5, T-7, and T-S was fractionated by electrophoresis at the same pH, to yield pure Peptides T-3a and T-5. Peptides T-7 and T-8 had the same mobility at pH 1.   compositions and some properties of t,hese peptides are summarized in Table IT. The chymotryptic digest of toxin o( was fractionated by gel filtration through Sephades G-25. The elution diagram is presented in Fig. 4. Fractions S2, 53, S5, and S8 were further fract'ionnted by paper chromatography in Solvent II, Fraction S4 by electrophoresis at pH 6.5, and Fraction S6 by paper chromatography in Solvent I. This resulted in the purif% cation of all of the peptides except Pept,ide C-3a of Fraction S2 \Thich had to be further purified by electrophoresis at pH 1.9. The amino acid composition and some properties of the chymotryptic peptides are summarized in Table III. The complete amino acid composition of toxin cy could be accounted for with both the tryptic or the chymotryptic peptides.
Sequence studies on the tryptic and chymotryptic peptides of toxin (Y are summarized in Table IV. The sequence of Peptide C-l is obvious from the NHz-terminal sequence of tosin (Y. The tryptophan residue of Peptide C-5a was destroyed upon acid hydrolysis of t'he tritium-labeled peptide so that the radioact.ivity of t,his residue could not be determined. Neither lysine nor tyrosine was labeled and, together with the unique composition of Peptide C-5, when compared to Peptide C-5, the sequence must be as stated.
The sequence of Peptide C-6 follows from Pept)ides T-4 and T-5. The amides of the 2 aspartyl residues in Peptide C-8 were determined on peptides isolated from a papain digest of Peptide C-8. The sequence studies on this peptide are summarized in Table V. Complete Sequence of Toxin a-The unique alanyl residue present in peptide C-3a, together with its high glutamic acid content, showed this peptide to be an overlap for Peptides T-2 and T-3. Peptide C-5a with it's unique lysyl-lysyl sequence presents an overlap for peptides T-3 and T-4. The occurrence of the unique tryptophanyl residue in Pept,ide C-6 furnishes the overlap between Peptides T-4 and T-5. If Peptides T-6, T-7, and the sequenced 7 residues of Peptide T-8 are coupled, we find that the resultant peptide has the amino acid composition of Peptide C-7 minus arginine and the COOH-terminal histidine of Peptide C-7. When Peptide C-7a is esamined in conjunction with these, we find that only one unique sequence is possible for Peptides 'M, T-7, and T-8. The COOH-terminal histidine of l'eptides C-7 and C-7a must follow the last isoleucyl residue of the sequenced half of Peptide T-8 and must be followed by the known sequence of Peptide C-8 which accounts for the rest of the amino acid von1position of Peptide T-8. The arginyl, glycyl, and threon$ residues by which Peptide C-7a is shorter than Peptide (r-7 provide the evidence for placing Peptide T-6 to the NI-IQ-ternlinnl end of C-7. Peptide T-8 must provide the COO%ternlinal part of Peptides C-7 or C-7+ leaving Peptide T-i to be between l'eljtides T-6 and T-8. The result is that we have three regions of toxin LY that overlap to the extent of only 1 residue (arginine) with each other. Pep tide T-l must be NHz-terminal from its resemblance t,o the S& terminal of the tosin.
The region T-6-T-i-T-X ends in tyrosine and therefore has to be COOH-terminal.
Peptides T-2 and T-6, and T-lb and T-S were sepnlatrtl, and Peptide T-4 was purified by paper chromatography in Sol\-ent I. Paper chromatography in Solvent II was used to separate l'rl~tides T-3 and T-9. The amino acid compo&ion and Some prop erties of these peptides are summarized in Table VI. The thymotryptic digest of toxin y was fractionat.ed by gel filtrat.ion through Sephadex G-25 as is shown in Fig. 6. Two of the sis fractions were pure peptides, namely Fraction Sl (Peptide CTa) and Fraction S5 (Peptide C-4). Peptides C-3 and C-8 in Fraction S2 were separated by chromatography in Solvent I. Fraction S3 yielded four subfractions by electrophoresis at pH 1.9. The fastest migrating fraction was pure Peptide C-10. The other three fractions were fractionated by chronlnt~ogral~ll~ ill Solvent 11 to yield pure peptides C-6, C-6b, C-tic, :md C-7. Fraction S4 was fractionated into five major fractions upon chromatography in Solvent 1. Three peptides were pure, i.c~. Peptides C-l, C-2, and C-(id. Peptides C-61) :~nd ('-10, \vhich were also found in Fraction S3, were separated, and I'el)t,ide C-6a was purified by chromatography in Solvent, II. Fractioll S6 contained peptides C-4, C-5, and C-9, which were aepar:ltetl 1)~ electrophoresis at $1 I .9. The amino acid composition and some properties of the chymotryptic peptides are presented in Tal:le VII.
The full amino acid composition of toxin y could 1~ :ICcounted for by both the tryptic and chymotryptic peptides.
Sequence studies for most of the tryptic peptides are ~11marized in Table VIII. Peptides T-6a and T-6 (which is Peptide T-6a + Arg) were pooled and digested with papain.
One of the pure peptides obtained from this digest had the composition Thr 0.87 (l), Cys(Cm), 0.91 (l), Pro 1.11 (l), and Lys 1.10 (1). Two steps of dansy-Edman degradation ga.ve Thr-Cys(C'm). Lysine must be the COOH-terminal of Peptide T-63 (tryptic specificity) and therefore also of this papain peptide.
Together I\-it11 the Edman degradation studies on Peptide T-Fa (Table VII), the complete sequence of Pept,ide T-6:1 is as above. The amidr intent of Peptide T-9 was ascertained by nluillol'rl'ticl:lse 11 digt+   Table IX. The sequences of Peptides C-l and C-2 by Peptide C-6. Together with Peptide C-6d, Peptide C-Ta were obvious from the sequence studies on reduced and X-car-equals Peptides T-6 + T-7 + T-8 + T-9 + Phe. This phenylboxymethylated toxin y. Peptide C-3 was digested with papain, alanine can only come from Peptide T-10. Peptides C-7a TL1 and from studies on the papain peptides, which are summarized and -TLz provide the overlap for Peptides T-7 and T-8. Peptide in Table X, the complete sequence followed.
Two peptides were T-6 has to provide the NHz-terminal part of Peptide C-7a. The isolated from a thermolysin digest of Peptide C-7a. Peptide amino acid composition of Peptide C-8 shows that T-8 and T-9 C-7a TL1 (Val 1.02 (l), Lys 1.06 (l), Ala 0.92 (1)) had COOH-are joined, so that the sequence of Peptide C-ia must be conterminal alanine, and Peptide C-7a TL2 (Val 0.99 (l), Lys 0.98 tained in the sequence T-6-T-7-T-8-T-g-T-10. Thus a unique (l), Ala 1.10 (l), Gly 0.93 (1)) had COOH-terminal glycine. amino acid sequence can be written for toxin y as is presented in These two peptides gave an overlap for Peptide T-7 and T-8. Considering the data presented here, it is obvious that toxin a! so far examined in this laboratory, but toxin cy of black marnba of D. polylepis venom is a 60-amino acid residue polypeptide chain venom killed mice in 4 to 43 min. This had tended t,o confirm with a molecular weight of 6907. This differs from an estimate the Sephadex estimates for the molecular weight of this toxin. of 3500 to 4000 based on the elution from Sephadex G-50 (2).
We cannot account for the physiological action of this tosin, but The death time of mice after subcutaneous injections of toxins is the retardation on Sephadex is explained by adsorption of the probably dependent on the diffusion rate of the tosins.
A minitoxin because of the exceptionally high content of aromatic amino acid residues (as compared \vith other Elupidae toxins). That even :I medium of high ionic strength (0.3 RI potassium phosphate, I = 0.6) camlot prel-ent this adsorbance is surprising, although the adsorbance of aromatic compounds to destran gels at low ionic strength is known (21,22). A similar effect has recently been noticed with a protease from rat mast cells (23).
TIP use of Sephatles G-25 columns for t,he initial separat'ion of hnme of the enzyme digests proved to be more advantageous than l~:~~~er electropliore&-:md paper chromatography. Ammonium avetatr IV:W used :ls ~1 relatively volatile buffer rather than the illore volatile acetic avid. Neutral so1Utions were preferred because ol the smaller risk of deanlidation of peptides during chro-mntogra1)hy.
The Ion-ionic strength of the 0.05 ti1 ammonium acetate eluent facilitated the separation of tryptophanyl peptides from the reht, by adhorption to the destran gel. Since the spectrol'llotolnetl,ic system used at, 210, 260, and 280 nm for monitoring the eluent flow \vas not compatible with pyridine solutions, this solvent wns not, considered.
On both of the reduced and S-carboq-methylated toxins, tryl'sin exhibited norn~al specificity.
Only lysyl and arginyl bouds were hydrolyzed, whereas the COOH-terminal bonds of Lyn-45, Lys-57, and Arg-1 of toxin 01 and Lys-70 of toxin y were not hydrolyzed.
*Irg-1 of toxin a is KHz-terminal and should therefore not be suscel)t,ible to attack by the endo-I)eptidasr, trypsin, alt~hough Arg-1 of tosin y was partially hydrolyzed.
The hydrolysis of the Y&(30)-Arg(31) bond of toxin OL by chymotrypsin is quite suq)rising since thi:: bond, which is present in all of the Elapidar short tosins, has not been found to be susceptible in any of the earlier cases.
The COOH-terminal trit'ium labeling method of RIatsuo et al. which could prevent the oxazolone formation during the labeling procedure.
The sequences of these two toxins have a few interesting features.
The occurrence of homodipeptides (i.e. places in the sequence where a residue is immediateIy repeated) is quite high in other toxins; toxins a! of N. huje and N. nigricollis have nine such pairs, and N. melanoleuca toxin dl has 10 homodipeptide sequences. These have been suggested to have some structural significance (4). Although toxin c11 of D. polylepis has sis pairs of homodipeptides, this is appreciably lower than other toxins and it does seem as if the relatively high level of occurrence of this type of oddity in snake venom toxins is of little critical significance in structure-function relationships since these homodipeptides are found at different positions in the chains of toxins from different species. The part of the chain between the second and third half-cystinyl residues needs an extra deletion when compared with the known short toxins of 61-and 62amino acid residues. Together with the sequence of the erabutoxins in this region, this part of the sequence of toxin cr differs so much from the corresponding region of the other Elapidae toxins (which are mutually remarkably similar in this respect) that we can conclude that its structure has little or no influence on the toxicity of the protein. This does not necessarily mean that this region is without influence on the structure of the active site of the toxins, since it is conceivable that these different toxins can have different physiological activities at the molecular level and still retain comparable lethalities. The "extra" 3 tyrosyl residues (as compared with the average known toxins containing one tyrosine), are accommodated in three novel positions in the chain, replacing His-4 or Phe-4, appearing next to the invariant trypt,ophanyl residue in a position which seems to be extremely variable, and appearing as the COOH-terminal residue, replacing a basic amino acid or asparagine.
Tyrosyl residues have therefore now been found in six different positions in the short tosin chains.
Tosin y differs considerably from the other two long toxins of known amino acid sequence; the main differing portions are the 20 NHp-terminal residues and the 7 residues at the COOHterminus.
The most prominent features of this toxin are the occurrence of a second tryptophanyl residue and an extra amino a.cid between half-cystines two and three as compared with e.g. Naja nivea toxin LY.
The occurrence of a second tryptophanyl residue in toxin y yields the first positive evidence that the rrlong'r line of toxins was genetically derived from a "short" line as was postulated by Strydom (24). When the alignment chart of the proterogl-yphae toxins (Table XI) is inspected, a gap of four positions (alignment positions 33 to 36) is seen in the short toxins.
Corresponding t,o this gap is the sequence Trp-Cys-Ser-Gln in toxin y. The sequence of the preceding 4 residues of toxin y is Trp-Q-s-Asp-Ala.
Such a situation could easily have come about as follows. Bearing in mind the existence of seryl residues in position 30 (alignment chart) of the erabutoxins (9), H.
haemachatus toxin -I (8) and N. wzelanoleuca toxin d,i this position could conceivably contain a cysteinyl residue (single-base change) in a short toxin. If the gene coding for this tosin underwent unequal crossing over to the extent of 12 bases, we 29 33 would find the sequence TrpCys-Asp-His-TrpCys-Asp-His 38 Arg-Gly in the resultant protein.
The 2 cysteinyl residues could then form a disulfide bridge.
In actual fact, in toxin a of Naja r;ivea venom, they do form a disulfide bridge (5), the rest of the disulfide bridges being identical with those of the erabutoxins (25) and cobrotoxin (26). Normal mutations would then lead to the sequences we find today. In the long toxins of N. nivea and N. melanoleuca the second tryptophanyl residue has been replaced by a phenylalanyl residue.