Evolution stings: the origin and diversification of scorpion toxin peptide scaffolds

Abstract

The episodic nature of natural selection and the accumulation of extreme sequence divergence in venom-encoding genes over long periods of evolutionary time can obscure the signature of positive Darwinian selection. Recognition of the true biocomplexity is further hampered by the limited taxon selection, with easy to obtain or medically important species typically being the subject of intense venom research, relative to the actual taxonomical diversity in nature. This holds true for scorpions, which are one of the most ancient terrestrial venomous animal lineages. The family Buthidae that includes all the medically significant species has been intensely investigated around the globe, while almost completely ignoring the remaining non-buthid families. Australian scorpion lineages, for instance, have been completely neglected, with only a single scorpion species (Urodacus yaschenkoi) having its venom transcriptome sequenced. Hence, the lack of venom composition and toxin sequence information from an entire continent's worth of scorpions has impeded our understanding of the molecular evolution of scorpion venom. The molecular origin, phylogenetic relationships and evolutionary histories of most scorpion toxin scaffolds remain enigmatic. In this study, we have sequenced venom gland transcriptomes of a wide taxonomical diversity of scorpions from Australia, including buthid and non-buthid representatives. Using state-of-art molecular evolutionary analyses, we show that a majority of CSα/β toxin scaffolds have experienced episodic influence of positive selection, while most non-CSα/β linear toxins evolve under the extreme influence of negative selection. For the first time, we have unraveled the molecular origin of the major scorpion toxin scaffolds, such as scorpion venom single von Willebrand factor C-domain peptides (SV-SVC), inhibitor cystine knot (ICK), disulphide-directed beta-hairpin (DDH), bradykinin potentiating peptides (BPP), linear non-disulphide bridged peptides and antimicrobial peptides (AMP). We have thus demonstrated that even neglected lineages of scorpions are a rich pool of novel biochemical components, which have evolved over millions of years to target specific ion channels in prey animals, and as a result, possess tremendous implications in therapeutics.

Figure 1

Bayesian phylogenetic reconstruction of the Na_V-CSα/β clade. Outgroups were the K_V-CSα/β Q0GY40 Hadrurus gertschi and Q95NK7 Mesobuthus martensi. *Chaerilus tricostatus contig sequence is from [71].

Figure 2

The two alternate scenarios of the cysteine relationships between DDH and ICK peptides. Sequences presented: 1. B8QG00 Hadrurus gertschi; 2. P59868 Pandinus imperator; 3. B8XH22 Buthus occitanus israel; 4. P0DJL0 Isometrus maculatus; 5. P0C5F2 Liocheles australasiae; 6. F8W670 Liocheles australasiae; 7. GALI01000016 Urodacus manicatus; 8. C5J894 Opisthacanthus cayaporum; 9. GALI01000015 Urodacus manicatus; 10. P0DJ08 Liocheles waigiensis; 11. SmpIT2 Scorpio maurus palmatus [66] and 12. GALI01000017 Urodacus manicatus. ICK connectivity schematic image adopted from [63]. Alignment scenario 1 is that proposed previously [63,64,65] while alignment scenario 2 is the alternative proposed in this study to better reflect charge molecule distribution.

Figure 3

Bayesian phylogenetic reconstruction of the SV-SVC, ICK and DDH clade. Outgroups were the non-toxin SVC peptides B4M772 Drosophila virilis and B4NQ53 Drosophila willistoni. * SmpIT2 Scorpio maurus palmatus is from [66]. Alignment scenario 1 is that proposed previously [63,64,65] while alignment scenario 2 is the alternative proposed in this study to better reflect charge molecule distribution.

Figure 4

Sequence alignment of cytotoxic linear peptides: (1). GALK01000016 Isometroides vescus; (2). GALK01000016 Isometroides vescus; (3). GALL01000023 Lychas buchari; (4). D9U2B7 Lychas mucronatus; (5). Q9Y0X4 Mesobuthus martensii; (6). C9X4J0 Tityus discrepans; (7). P0CF38 Isometrus maculatus; (8). P83312 Parabuthus schlechteri; (9). Q9GQW4 Mesobuthus martensii; (10). B8XH50 Buthus occitanus israelii; (11). I0DEB4 Vaejovis mexicanus smithii; (12). GALH01000010 Cercophonius squama; (13). P0C8W1 Hadrurus gertschi; (14). C5J886 Opisthacanthus cayaporum; (15). P0DJ03 Heterometrus petersii; (16). L0GCV8 Urodacus yaschenkoi; (17). P0DJO3 Scorpiops tibetanus; (18). GALH01000009 Cercophonius squama; (19). GALI01000003 Urodacus manicatus; (20). GALI01000004 Urodacus manicatus; (21). GALI01000007 Urodacus manicatus; (22). GALI01000005 Urodacus manicatus; (23). GALH01000008 Cercophonius squama; (24). GALI01000006 Urodacus manicatus; (25). L0GCI6 Urodacus yaschenkoi; (26). H2CYR5 Pandinus cavimanus; (27). G8YYA6 Androctonus amoreuxi; (28). B9UIY3 Lychas mucronatus; (29). GALL01000021 Lychas buchari; (30). GALK01000015 Isometroides vescus; (31). Q5G8B3 Tityus costatus; (32). E4VP60 Mesobuthus eupeus; (33). Q5G8B5 Tityus costatus; (34). D9U2B8 Lychas mucronatus; (35). C7B247 Isometrus maculatus; (36). G1FE62 Chaerilus tricostatus; (37). GALL01000022 Lychas buchari. Signal peptide and C-terminal cleaved propeptides are shown in lowercase. BPP domain shown in black and the cytotoxic posttranslationally processed peptide is highlighted in gray. ‘>’ indicates incomplete sequence.

Figure 5

Mid-point rooted Bayesian phylogenetic reconstruction of the cytotoxic linear peptides. * Chaerilus tricostatus and C. tryznai contig sequences are from [71].

Figure 6

Molecular evolution of scorpion toxins. Three dimensional homology models of various scorpion CSα/β and non-CSα/β toxins, depicting the locations of positively selected sites are presented. Site-model 8 computed omega and the total number of positively selected sites (PS) detected by its Bayes Empirical Bayes (BEB) approach (PP ≥ 0.95) are indicated, along with the number of episodically diversifying sites (Epi) detected by MEME (at 0.05 significance). PDB codes used for modelling are: α-Na_V-CSα/β: 1DJT; β-Na_V-CSα/β: 2I61; Cl_V-CSα/β: 1SIS; DDH: 2KYJ; ICK: 1IE6; short-K_V-CSα/β: 1PVZ and SVC: 1U5M).

Figure 7

Surface accessibility of hypermutable sites. A plot of amino acid positions (x-axis) against accessible surface area (ASA) ratio (y-axis) indicating the locations of amino acids (exposed or buried) in the crystal structure of various scorpion toxins is presented. Positively selected residues are presented as large dots, while the remaining sites are presented as small dots in the plot. Residues with an ASA ratio greater than 50% are considered to be exposed, while those with an ASA ratio less than 20% are considered to be buried to the surrounding medium (ASA of 21%–39%: cannot be assigned to buried/exposed class; ASA of 40%-50% are likely to have exposed side chains). Three dimensional homology models of various scorpion toxin types, depicting the locations of positively selected (PS) sites along with model 8 omega values and the number of exposed and buried positively selected sites are also presented. PDB codes used for modelling are: α-Na_V-CSα/β: 1DJT; β-Na_V-CSα/β: 2I61; Cl_V-and CSα/β: 1SIS.