Unusual accelerated rate of deletions and insertions in toxin genes in the venom glands of the pygmy copperhead (Austrelaps labialis) from Kangaroo island

Abstract

Background: Toxin profiling helps in cataloguing the toxin present in the venom as well as in searching for novel toxins. The former helps in understanding potential pharmacological profile of the venom and evolution of toxins, while the latter contributes to understanding of novel mechanisms of toxicity and provide new research tools or prototypes of therapeutic agents.

Results: The pygmy copperhead (Austrelaps labialis) is one of the less studied species. In this present study, an attempt has been made to describe the toxin profile of A. labialis from Kangaroo Island using the cDNA library of its venom glands. We sequenced 658 clones which represent the common families of toxin genes present in snake venom. They include (a) putative long-chain and short-chain neurotoxins, (b) phospholipase A2, (c) Kunitz-type protease inhibitor, (d) CRISPs, (e) C-type lectins and (f) Metalloproteases. In addition, we have also identified a novel protein with two Kunitz-type domains in tandem similar to bikunin.

Conclusion: Interestingly, the cDNA library reveals that most of the toxin families (17 out of 43 toxin genes; approximately 40%) have truncated transcripts due to insertion or deletion of nucleotides. These truncated products might not be functionally active proteins. However, cellular transcripts from the same venom glands are not affected. This unusual higher rate of deletion and insertion of nucleotide in toxin genes may be responsible for the lower toxicity of A. labialis venom of Kangroo Island and have significant effect on evolution of toxin genes.

Figure 1

Composition of a cDNA library from A. labialis venom gland. A) Relative abundance of genes sequenced from the cDNA library; B) Relative abundance of the toxin genes in the cDNA library.

Figure 2

Three-finger toxins in A. labialis venom gland. A) Alignment of long-chain neurotoxins with α-cobratoxin (α-cobra) [64]. The changed in amino acids are highlighted in grey which is either due to addition or deletion of nucleotide. ◆, conserved Cysteine residues; *, residues involved in binding to nicotinic acetylcholine receptor (nAChR); and □, residues Ala28 and Lys49 involved in binding to α7 receptor and nAChR receptor, respectively. Residues that are different from consensus sequence of A. lablialis toxins are highlighted. The truncated transcripts and the elongated protein product (clone 112; for brevity, functionally unimportant parts of the sequence are not shown and the dots indicate the missing segments). The number of clones is shown in italics and the predicted signal peptide using SignalP 3.0 is underlined (The signal peptide for α-cobrotoxin is not available). B) Alignment of two cDNA-deduced peptide representatives of one cluster and one singleton of putative short-chain neurotoxin with Erabutoxin a (Erabtx_a) [65]. The gaps in clone 6A and 12B are represented by blank spaces. *, residues involved in binding to nicotinic acetylcholine receptor (nAChR). The substitution of Thr to Ala in clone 12B is highlighted. Phe32 (□) of erabutoxin a involved in binding to nAChR receptor which is substituted with His32 in both the clones of A. labialis is highlighted. The number of clones is shown in italics and the predicted signal peptide using SignalP 3.0 is underlined.

Figure 3

Phylogenetic tree of elapid three-finger toxins. Full length A. labialis long-chain and short-chain neurotoxins are encircled. Phylogram was generated by PHYML using maximum likelihood method.

Figure 4

Phospholipase A₂ enzymes in A. labialis venom gland. Alignment of PLA₂ enzymes from A. labialis with A. superbus (Aussup) [5] is shown. Residues that are different from A. superbus PLA₂ are highlighted. The three Trp (W) residues present in the β-wing of the A. labialis PLA₂ enzyme are underlined. Clone 518 which encodes a truncated protein is also shown. The number of clones is shown in italics and the predicted signal peptide using SignalP 3.0 is underlined.

Figure 5

Serine protease inhibitors in A. labialis venom gland. A) Alignment of Kunitz-type serine protease inhibitors. ■, conserved six cysteine residues in kunitz domain. Residues that are different from the consensus sequence of A. labialis proteins are highlighted. The number of clones is shown in italics and the predicted signal peptide using SignalP 3.0 is underlined. B) Alignment of clone 655 (serine protease inhibitor containing two Kunitz domains) with bikunin [29]. Signal peptide of clone 655 is indicated by a vertical bar ( | ), ■, conserved six cysteine residues; , N-glycosylation site. Conserved residues are highlighted. C) Alignment of isoforms of two Kunitz-type domains. The presence of two Kunitz-type domains was further confirmed by PCR using gene specific primer and reverse primer from cDNA construction kit (for details, see Materials and methods). *, consensus residues and residues that are different from the consensus sequence are highlighted. The number of clones is shown in italics and the predicted signal peptide using SignalP 3.0 is underlined.

Figure 6

Alignment of CRISPs from A. labialis Notechis scutatus, Hoplocephalus stephensii, Oxyuranus microlepidotus [66]. ■, conserved cysteine. Residues that are different from the majority of sequences are highlighted. The number of clones is shown in italics and the predicted signal peptide using SignalP 3.0 is underlined.

Figure 7

Alignment of A. labialis C-type lectin (clone 132) with L1Bf (Bungarus fasciatus), LBm (B. multicinctus), BFL2 (Bungarus fasciatus) and TsL (Trimeressurus stejneger) [47,53] Gln-Pro-Asp (QPD) sequence essential for galactose interaction is indicated by box and region flanked by (▶) and (◀) are residues essential for Ca²⁺ interaction [67]. The number of clones is shown in italics and the predicted signal peptide using SignalP 3.0 is underlined.