Comparison of Sea Snake (Hydrophiidae) Neurotoxin to Cobra (Naja) Neurotoxin

Both sea snakes and cobras have venoms containing postsynaptic neurotoxins. Comparison of the primary structures indicates many similarities, especially the positions of the four disulfide bonds. However, detailed examination reveals differences in several amino acid residues. Amino acid sequences of sea snake neurotoxins were determined, and then compared to cobra neurotoxins by computer modeling. This allowed for easy comparison of the similarities and differences between the two types of postsynaptic neurotoxins. Comparison of computer models for the toxins of sea snakes and cobra will reveal the three dimensional difference of the toxins much clearer than the amino acid sequence alone.


Introduction
In tropical and subtropical regions, snakebites are a serious public health problem. The symptoms that follow snake envenomation are different depending on the species of snake. This is because the venoms are not entirely identical [1]. It has been known that venoms of Elapidae (cobras, kraits) land snakes are similar to that of Hydrophiidae (sea snakes) [2]. Venoms of both families contain potent postsynaptic neurotoxins. Close examination of the primary structures of neurotoxins obtained from Elapidae and Hyrophiidae showed considerable structural difference, although the backbone of the disulfide bonds is similar. In this investigation, we compared the neurotoxins of the two families by molecular model methods including the toxin obtained from Prasecutata viperina ( Figure 1) that was newly isolated.

Materials and Methods
Specimens of the sea snake, Praescutata viperina, were captured in the Gulf of Thailand. The venom glands were removed and dried at room temperature. Crude venom was extracted from venom glands using distilled water and the insoluble tissue debris was removed by centrifugation (3,860 × g, 30 min, 4 °C), then the supernatant liquid was lyophilized for storage.

Isolation procedure
All of the purification procedures were performed at 4 °C. Crude venom (20 mg) was dissolved in 5 mL of 10 mM potassium phosphate buffer (pH 7.8) and applied to a CM52-cellulose column. The toxin was eluted with the same buffer at a flow rate of 13.5 mL per hour and it was monitored at 280 nm. Fractions of 3 mL were collected, and pooled fractions were tested for toxicity. The homogeneity of preparation was checked using SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and reversed-phase HPLC.

Toxicity test
The toxicity tests were done by injecting by 0.1 mL of toxin at various concentrations intravenously into ddY mice weighing 18 g each. At each of the five dosage levels, five mice were used. After 24 h, the number of mice that had died was observed. The toxicity was determined statistically using the method of Litchfield and Wilcoxon [3] and expressed as the lethal dosage 50%, the LD 50 value (micrograms of toxin per gram of body weight of mouse). Experimental protocols concerning the use of laboratory animals were approved by the committee of Meijo University.

Biochemical characterization
The molecular mass of purified toxin was determined by MALDI/TOF-MS (matrix-assisted laser desorption/ionization time-of-flight mass spectrometry) with a Nihon Perseptive Biosystems/Voyager R.P., and by SDS-PAGE using the method of Weber and Osborn [4]. Isoelectric point was determined by the Pharmalyte concentration 3% (w/v) with a pH range of 3-10. Isoelectric foucusing was carried out using a constant potential of 200 V at 5 °C for 4 h.

Sequence analysis
The amino-terminal sequence of native toxin and the enzymatically cleaved fragments of toxin were analyzed by an Applied Biosystems 491 protein sequencer. The phenylthiohydantoin (PTH) derivatives of amino acids were identified with an Applied Biosystems Model 610A PTH analyzer in accordance with the manufacturer's instructions.

Homology analysis by computer modeling
MOE TM (Molecular Operating Environment), a molecular simulation and modeling software purchased from Ryoka Systems Inc. was used for construction of protein models.

Isolation and purification
Praescuta toxin was isolated by a CM52-cellulose column chromatography ( Figure 2) and lethal activity was found in fraction 5. Homogeneity of the toxin was established by two independent methods: SDS-PAGE ( Figure 2A) and reversed-phase HPLC ( Figure 2B). This purified toxin was named praescuta toxin and the yield of the praescuta toxin from 20 mg Praescutata viperina venom was found to be approximately 2.0 mg.

Toxicity
The LD 50 of the praescuta toxin in mice was found to be 0.41 (0.31~0.55) μg/g by intravenous injection. Since the LD 50 values of sea snake neurotoxins from Pelamis platurus and Hydrophis ornatus were reported to be 0.13 μg/g (intravenous injection) and 0.09 μg/g (intramuscular injection) respectively [12,13], the toxicity of Praescutata viperina toxin is relatively weak compared to sea snake neurotoxins.
The effect of temperature on toxicity was also examined. A 50 μg of toxin preparation was heated for ten minutes at 70, 85 and 100 °C individually, and injected into mice (n = 3) intravenously. The results of injection are shown here: It is well known that sea snake toxins are relatively stable during heat treatment because of their compact molecular structure with a relatively small size and a peptide backbone held together by four disulfide bonds [14]. The result shows the high thermal stability of the toxin. At 70 °C, all mice in the group died, which meant that the toxin retained full potency. Even at 85 °C, one out of three mice in the group died, indicating the toxin is still active even at high temperatures. The toxin was denatured in the test at 100 °C, as indicated by the survival of all three test animals.

Biochemical properties
The molecular mass of praescuta toxin determined by SDS-PAGE was found to be 7,400 Da, while the MALDI/TOF/MS method gave a molecular weight of 6,674.9 Da. The isoelectric point was determined by electrophoresis and was found to be higher than 10.0.
In order to ascertain that praescuta toxin is not an enzyme and does not show hemorrhagic and clotting activity, several assays were performed. The final preparation did not show phospholipase A 2 activity, elastase activity using STANA (Suc-Ala-Ala-Ala-pNA) as the substrate, arginine ester hydrolase activity using TAME (tosyl-L-arginine methyl ester), arylamidase activity using L-leucine-β-naphtylamide as the substrate, proteinase activity using dimethylcaesin, fibrinogen, casein, and insulin B chains as substrates, collagenase activity using collagen Type IV as the substrate, and hemorrhagic activity. These results provide considerable evidence that this sea snake neurotoxin is a non-enzymatic type toxin. This conclusion is consistent with earlier findings that sea snake toxins bind to the acetylcholine receptor and are competitive inhibitors of acetylcholine [12,15].

Reduction of disulfide bonds
It is known that all sea snake toxins contain four disulfide bonds. Disulfide bonds are especially important for small proteins in order to hold the peptide backbone together. The reagent dithiothreitol (DTT) is known to cleave disulfides by reducing the bond. When 170 μg of praescuta toxin at a concentration of 17 μg /mL was incubated with 1% DTT at 37 °C for 30 min and then injected into three mice intraperitoneally, all of them survived. This indicates that the four disulfide bonds are essential for toxicity because the cleavage of these bonds with the reducing agent resulted in a loss of toxicity.

Direct amino acid sequence analysis and sequence analysis of digested fragments
Praescuta toxin was digested with endoproteinase Arg-C or endoproteinase Lys-C and different fragments were isolated by reversed-phase HPLC as shown in Figure 3A and 3B. By combining all the data obtained from the intact toxin and various fragments, the complete amino acid sequence of praescuta toxin was determined (Figure 4). From these results, praescuta toxin is composed of 60 amino acids and the molecular mass of the protein portion of praescuta toxin was calculated to be 6675.29 Da. The molecular mass of praescuta toxin based on the amino acid sequence was identical to the molecular mass (6674.9 Da) obtained by MALDI/TOF/MS.

Comparison of primary structure within Hydrophiinae toxins
The amino acid sequence of toxins isolated from Lapemis, Pelamis and Acalptophis venoms are extremely homologous and the positions of 9 cysteine residues are conserved ( Figure 5), indicating that the three-dimensional structure of these toxins are similar. Praescuta toxin also possesses a similar primary structure with these toxins except for the position of Cys(27) and Trp(39). These substitutions of reactive amino acids might be the reason for the relatively weak toxicity of praescuta toxin.
The differences are few and are summarized here.

Comparison of Hydrophiinae and Laticaudinae toxins
Only two toxins from the subfamily Laticaudinae are quoted for comparison ( Figure 5). There are still many similarities, especially the disulfide bond backbone, but there are clearly more differences between the two subfamilies when analyzing the amino acid sequence of the neurotoxins than through an analysis of the sequences of toxins within the Hydrophiinae.

Comparison with land snake (Elapidae) toxins
Two cobra toxins are presented in Figure 5 for comparison. Elapidae toxins are quite different from all Hydrophiinae toxins. Comparison of the toxin sequences shown here reveals that their primary structure is related to the phylogenicity of the snakes: the more closely related, the more similar the amino acid sequence.
Recently Fry et al. [16] published a paper discussing the evolution of sea snakes extensively. In this paper, they mentioned that genus Laticauda is quite similar to many Elapidae land snakes. In our paper, our objective is more focused on chemical structure rather than evolution. We followed the classification of Smith [17], who mentioned that Hydrophiidae (sea snakes) contain two subfamilies of Hydrophiinae and Laticaudinae.    Figure 5, the disulfide bonds of praescuta toxin are predicted to be formed between Cys (3) and (22)