Role of accelerated segment switch in exons to alter targeting (ASSET) in the molecular evolution of snake venom proteins
© Doley et al. 2009
Received: 08 January 2009
Accepted: 30 June 2009
Published: 30 June 2009
Skip to main content
© Doley et al. 2009
Received: 08 January 2009
Accepted: 30 June 2009
Published: 30 June 2009
Snake venom toxins evolve more rapidly than other proteins through accelerated changes in the protein coding regions. Previously we have shown that accelerated segment switch in exons to alter targeting (ASSET) might play an important role in its functional evolution of viperid three-finger toxins. In this phenomenon, short sequences in exons are radically changed to unrelated sequences and hence affect the folding and functional properties of the toxins.
Here we analyzed other snake venom protein families to elucidate the role of ASSET in their functional evolution. ASSET appears to be involved in the functional evolution of three-finger toxins to a greater extent than in several other venom protein families. ASSET leads to replacement of some of the critical amino acid residues that affect the biological function in three-finger toxins as well as change the conformation of the loop that is involved in binding to specific target sites.
ASSET could lead to novel functions in snake venom proteins. Among snake venom serine proteases, ASSET contributes to changes in three surface segments. One of these segments near the substrate binding region is known to affect substrate specificity, and its exchange may have significant implications for differences in isoform catalytic activity on specific target protein substrates. ASSET therefore plays an important role in functional diversification of snake venom proteins, in addition to accelerated point mutations in the protein coding regions. Accelerated point mutations lead to fine-tuning of target specificity, whereas ASSET leads to large-scale replacement of multiple functionally important residues, resulting in change or gain of functions.
Snake venoms contain a mixture of proteins and polypeptides which exhibit various biochemical and pharmacological functions. These proteins and polypeptides are classified into non-enzymatic and enzymatic proteins which belong to a small number of superfamilies, such as three-finger toxins (3FTx), Kunitz-type serine protease inhibitors, phospholipase A2(PLA2) enzymes, serine proteases and metalloproteases [1–12]. Members of these superfamilies have similar protein scaffolds but, at times, differ markedly in their biological effects. For example, members of 3FTx family exhibit a wide variety of specific pharmacologic effects by targeting various receptors and ion channels with high affinity and specificity. Short chain and long chain α-neurotoxins antagonize muscle nicotinic acetylcholine receptors [5, 12], κ-bungarotoxins recognize neuronal nicotinic receptors , muscarinic toxins are selective agonists/antagonists of distinct sub-types of muscarinic acetylcholine receptors , fasciculins inhibit acetylcholinesterase , calciseptine and related toxins block the L-type Ca2+ channels [16, 16, 17], cardiotoxins/cytotoxins exert their toxicity by forming pores in cell membranes , and dendroaspins are antagonists of various cell-adhesion processes .
Similarly, other venom proteins, such as the Kunitz-type serine protease inhibitors, have a conserved fold and are structurally similar to bovine pancreatic trypsin inhibitor (BPTI) . They have been reported to inhibit proteolytic activity of trypsin or chymotrypsin specifically [3, 21, 21–23]. In addition, some inhibitor-like proteins specifically block potassium and calcium channels [24–27]. Snake venom PLA2 isoenzymes, also characterized by a highly conserved fold, are known to induce various pharmacological activities such as neurotoxic, myotoxic, cardiotoxic, anticoagulant, and antiplatelet effects through specific interaction with their target proteins (for a review see ). Thus many subfamilies and isoforms of snake venom serine proteases and metalloproteases act on various components of the coagulation cascade and induce procoagulant or anticoagulant effects, as well as affect platelet aggregation, fibrinolytic and kallikrein-kinin systems [29–35]. The isoforms of the different superfamilies are known to evolve through a process of gene duplication followed by accelerated point mutations in the protein coding regions.
In venom proteins, comparisons of cDNA and gene sequences have shown that nonsynonymous nucleotide substitutions (leading to change in amino acid residues) are commonly greater than synonymous nucleotide substitutions (not producing change in amino acid residues) in the protein coding region compared to the non-coding (UTRs) and intron regions [36–38]. Thus, protein coding regions of genes encoding 3FTxs [39–41], Kunitz-type serine protease inhibitors , PLA2 enzymes [7, 43, 44] and serine proteases  appear to be undergoing accelerated point mutations, resulting in numerous isoforms. Individual point mutations affect one residue at a time, leading to small change in the surface characteristics of a protein. Therefore, point mutations may contribute to fine tuning of toxin specificities by (a) improving the specificity towards a particular receptor or ion channel; (b) altering the specificity towards a closely related receptor or ion channel; and (c) modifying the species specificity. However, accelerated point mutations may not be sufficient to explain drastic changes in the molecular surface needed for the observed targeting of toxins with conserved scaffolds to diverse receptors or ion channels.
In a recent study, we identified five transcripts encoding 3FTxs from the cDNA library of venom gland tissues of Sistrurus catenatus edwardsii . These transcripts showed very low sequence similarity with elapid 3FTxs except for the conserved signal peptide and the number and position of cysteines. A systematic comparison of their sequences revealed that some of the segments in the mature proteins were 80–100% identical, whereas other segments were only 12.5–50% similar . Some segments in the protein coding region seem to be exchanged with distinctly different segments, keeping the structural fold intact during their evolution. Interestingly, the segments in the introns of genes encoding these same proteins show high similarity (>85%) when present; and profound differences in segments appear to be restricted to exons only. Such switching of segments in the exon alters the surface topology and charge of the mature protein, which might alter the molecular targets of 3FTxs and contribute to the evolution of novel function. Therefore, we proposed that the phenomenon of a ccelerated s egment s witch in e xons to alter t argeting (ASSET) might play an important role in the evolution of 3FTxs in viperid snake venoms .
Here we have analyzed isoforms of 3FTxs from elapid snake venoms, as well as toxins from other protein superfamilies, to evaluate whether ASSET plays a role in their evolution. Elapid 3FTxs have been found to undergo changes due to ASSET as observed for viperid 3FTxs. Due to such exchange of segments, functionally important residues are changed, which might significantly affect their function. In some of these toxins, such change has lead to the evolution of demonstrated novel functions [19, 47], and thus the 3FTx toxin family seem to be functionally evolving through ASSET. In the Kunitz-type serine protease inhibitor family, ASSET does not seem to play an important role in evolution, even though there are multiple isoforms. The enzymatic families, such as PLA2 and metalloproteases, appear to be evolving more through accelerated point mutations rather than ASSET. In these families, some of the segments seem to be exchanged during their evolution, but functional implication of such changes is not clearly understood. However, in the serine protease family, three segments near the substrate binding region have been found to be undergoing accelerated exchange of segments, and at least one of them may have significant implications for their substrate specificity.
3FTxs form a well-characterized superfamily of non-enzymatic proteins. They have a canonical three-finger fold of extending β- sheeted loops that is stabilized by four conserved disulphide bridges in the core region. Until recently, this family of proteins was thought to be present only in elapid venom . However, 3FTxs have now been reported in colubrid venoms and in viperid venom gland transcriptomes as well [45, 49–52]. Different isoforms of 3FTxs bind to various receptors/acceptors and exhibit diverse pharmacological functions despite their similar folding (for a review see ). Functionally important residues that are involved in interacting with the target receptors/ion channels generally reside in the tip of the loops . Therefore the amino acid sequences, length and conformation of the loops play important role in their functional specificity (for reviews see [53, 55].
The three-finger folds of 3FTxs are held together by four conserved disulfide bridges. However, some 3FTxs have a fifth disulfide bridge in either the second loop (long chain neurotoxins and κ-neurotoxins ) or the first loop (non-conventional toxins ). The insertion of the fifth disulfide bridge in the second loop of long chain 3FTx is due to a change in the intron-exon boundary. This alteration in the intron-exon boundary is due to an insertion of a single nucleotide "A" in intron 2 which causes a shift in the splicing site , leading to the insertion of a short segment (S5) containing a cysteine residue. In the S4 segment, there is also a frame shift due to the deletion of a nucleotide, leading to a completely different sequence which also contains a new cysteine residue. Both cysteine residues form the fifth disulphide bridge and a cyclic structure in the second loop that is important to their binding to α7 receptors with high affinity . The insertion of this short segment in long chain 3FTxs leads to a new function – binding to α7 receptors. In contrast, the fifth disulphide bridge in the first loop of short chain 3FTxs is due to exchange of segments (ASSET). This additional fifth disulphide bridge does not change the overall fold but it causes subtle changes in the first loop which are known to have functional implications . Further, the number of amino acid residues in this segment differs among the toxins and hence would lead to change in the length of the loop.
Short chain and long chain α-neurotoxins are known to antagonize muscle nicotinic acetylcholine receptors, resulting in flaccid paralysis [5, 12]. The structure-function relationships of α-neurotoxins have been thoroughly studied using both chemical modification and genetic engineering approaches [54–61]. Unlike dendroaspin and calciseptine (Figure 1), the functional site in the neurotoxins is discontinuous and is distributed on all three loops . In erabutoxin a (BAC78199), the important functional residues involved in binding to Torpedo electroplax or to muscle nAChR (α2βγδ) are Lys27, Trp29 (S3), Asp31, Phe32, Arg33 (S5) and Lys47 (S7) (underlined in Figure 1) . Although Lys27, Trp29 and Arg33 are conserved in all Laticauda toxins, the other critical residues (Asp31, Phe32 and Lys47) are replaced via exchange of segments. We hypothesize that these segment exchanges may have a direct impact on their ability to bind to Torpedo or muscle (α2βγδ) receptors.
It is also important to note that there are minor changes in amino acid residues within the identical segments (highlighted in white in Additional file 1) and these changes are due to an accelerated rate of point mutations. Both ASSET and accelerated point mutations have contributed to the functional diversity of elapid 3FTXs; ASSET leads to major changes in the surface properties, resulting in targeting of new receptors, while accelerated point mutations lead to fine-tuning of binding to the same receptors through minor alterations of the surface charge and topology.
Snake venom Kunitz-type serine protease inhibitors have been reported from both elapid and viperid venoms. Structurally, they are similar to Kunitz/BPTI inhibitors with a conserved fold stabilized by three disulphide bridges . As with other toxin families, the isoforms are encoded by a multigene family and have evolved through gene duplication and positive selection . The isoforms from the same genus are grouped together as above and analyzed for ASSET and accelerated point mutations (Additional file 2).
Though the snake venom Kunitz-type serine protease inhibitor family contains multiple isoforms, functionally they are not as diverse as other venom protein superfamilies, and they can be divided into either non-neurotoxic or neurotoxic homologs. Non-neurotoxic homologs inhibit either trypsin or chymotrypsin, while neurotoxic homologs act as calcium and potassium channel blockers which do not have protease inhibitory activity [26, 27, 42]. Structurally, both groups have a conserved fold similar to BPTI, but the inhibitor binding loops and the β turn regions have undergone adaptive evolution, resulting in new biological activities . Analysis of the amino acid sequences of the isoforms shows that there is no radical change in the amino acid residues in the mature proteins, as observed in 3FTxs. However, they have undergone adaptive evolution through accelerated point mutation (Additional file 2). In calcicludine and dendrotoxin-I, the N-terminal part and overall conformation play a significant role in calcium and potassium channel-blocking activity (Additional file 2). This has been demonstrated by synthesizing chimeras containing the N-terminal (1–30) of calcicludine and C-terminal (31–60) of dendrotoxin-I, and vice versa . However, there are not enough Kunitz-type serine protease inhibitors and dendrotoxin sequences from Dendroaspis species in the database in order to determine if they have evolved through ASSET. Similarly, the B chain of β-bungarotoxin (from Bungarus) is also a Kunitz-type serine protease inhibitor but does not have protease inhibitory activity; however, it contributes to neurotoxicity . The interaction of the B chain with the potassium channel was predicted to be localized opposite of the anti-protease loop, between residues 27–30 . In addition to this, the mature protein shows accelerated point mutations which resulted in the introduction of a cysteine residue at the C-terminal end (underlined in Additional file 2). This extra cysteine residue forms the disulphide bridge with chain A . Further, the C-terminal region of chain B shows a conformational change due to its interaction with the chain A and accounts for the lack of protease inhibitor activity . Unlike 3FTxs, where ASSET has played an important role in the evolution of new functions, deviation of some members of Kunitz-type serine protease inhibitors from protease inhibitory activity is mainly due to accelerated point mutations. This might explain the low functional diversity in this group of toxins, even though they have multiple isoforms.
PLA2 enzymes are one of the best-studied hydrolytic enzymes and are found abundantly in nature. Snake venoms are a good source of these enzymes and often contain multiple isoenzymes. In addition to a role in the digestion of prey, they induce a wide variety of pharmacological effects in prey/victims (for a review see ). It has been well documented that accelerated point mutations have occurred in the protein coding regions, and this adaptive mode of evolution might also be responsible for acquisition of new functions . We analyzed the elapid and viperids PLA2 isoenzymes to determine if ASSET has played any role in the functional evolution of these toxins.
In addition to this exchange of a segment near the calcium binding region, the presence or absence of another segment has been observed in exon III of Austrelaps superbus PLA2 enzymes (Figure 3, shown in green color). This segment represents the pancreatic loop, an ancestral feature found in the pancreatic PLA2 enzymes. Pancreatic PLA2 enzymes show low hydrolytic activity due to the presence of this loop, and the deletion of this loop in porcine PLA2 results in 16 times higher catalytic activity . Thus, the deletion of pancreatic loop in venom PLA2 enzymes plays an important role in the evolution of catalytically more active enzymes.
Accelerated point mutations in the mature protein of PLA2 enzymes are known to play important roles in functional evolution [37, 43, 69, 70]. These substitutions appear to occur mostly in the surface residues and thus alter the specificity of targeting to various tissues or cells, resulting in distinct pharmacological effects . Though we observed ASSET near the calcium-binding region, its role in functional evolution of PLA2s is not yet clear.
Snake venom serine proteases (SVSPs) are one of the well characterized families of enzymes that affect the hemostatic system. They act on various components of the coagulation cascade, fibrinolytic and kallikrein-kinin systems as well as on platelets to cause significant perturbance of the haemostatic system [31, 72–75]. This family of enzymes are believed to have evolved from glandular kallikrein and trypsin-like enzymes, as they have similar gene structure and share common three-dimensional structure . Similar to other multigene families, they have evolved through accelerated evolution in the protein coding region . In the present study we aligned the SVSPs from Trimeresurus species, Crotalus species, Sistrurus catenatus edwardsii and Bothrops jararaca obtained from the database to analyze for ASSET.
Previously, we discussed the possible molecular mechanisms of ASSET, including splicing variation, recombination, accumulation of point mutations and independent recruitment events . Nevertheless, we believe that none of these explanations are satisfactory. Splicing variations, such as alternative splicing and changes in the splicing site, can lead to insertion/deletion of alternate segments in the mature protein. However, all but one segment change occur within the exons and not at the intron-exon boundaries. In the long chain 3FTxs only, the insertion of a segment occurs at the intron-exon boundary due to a shift in the splicing site (discussed above). Genetic recombination might also give rise to replacement of segments in the mature protein. However, the segment exchanges observed in the venom proteins are too small, and canonical recombination processes cannot explain exchange of short segments. The possibility of accumulation of point mutations producing the observed change in segments cannot be ruled out unequivocally, as venom proteins have been well-documented to evolve through accelerated point mutations [7, 8, 37, 39–43, 84]. In such circumstances, this would have to occur over many generations to attain the observed change in segments, and intermediates would have to be selected via positive selection. Further, the same point mutations would have to occur independently in several unrelated lineages to produce the same segment composition. Thus, odds are against the accumulation of point mutations as an explanation. The 3FTxs showing these exchanges could have occurred through independent recruitment events, but significant similarities in protein sequence and gene structure (particularly high similarity among intron sequences) show that they have evolved from a common ancestor. The observed changes in protein segments could also be due to insertion/deletion of 1–2 nucleotides, resulting in a frame shift and hence altered protein sequence. In such a case, a similar number of nucleotide(s) must be removed/added in a downstream region, respectively, to get back into the original open reading frame. Therefore, we carefully analyzed each of these segment exchanges at the nucleotide level. Only one segment, RKCHNSPLSLVYQ (S2 in Q8UUK0), is changed to RKCNKLVPL-FYK in P01443 (Figure 1A) due to deletion and insertion of nucleotides. In this case, there is an insertion of A at the 78th position (from the start codon) and a deletion of four nucleotides downstream (Additional file 3). Therefore, none of the above possibilities explain the observed exchange of segments in the exons. It is important to note that in spite of the segment exchanges, the cysteine residues, which maintain the three dimensional fold, are conserved in 3FTxs. Although the molecular mechanism of the exchange of segments is not yet understood, these events clearly play a significant role in the functional evolution of some snake venom proteins.
Surface residues of a protein molecule are important for their physicochemical properties as well as for their interactions with biomolecules, including other proteins. Accordingly, the alteration of the conformation and surface properties indeed affects the pharmacological properties of protein toxins. In an earlier paper, we showed that in PLA2 enzymes the surface residues have undergone natural substitution 2.6–3.5 times faster than the buried residues and proposed that accelerated point mutations preferentially target the surface residues in PLA2 enzymes, leading to the evolution of new isoforms with distinct functions . As shown here, ASSET also targets surface residues in 3FTXs, PLA2 enzymes, serine proteases and metalloproteases (Figure 2, 4, 7 and 9). Accelerated point mutations result in finer modifications to the surface topology and/or electrostatic potential, whereas ASSET drastically alters the surface, essentially instantaneously producing large-scale changes in the ligand interaction site(s). The molecular mechanisms of both accelerated point mutation and ASSET are not clearly understood, but both phenomena play a crucial role in the evolution of snake venom proteins.
Elapid 3FTxs, similar to viperid 3FTxs (Doley et al., 2008), evolve by both ASSET and accelerated point mutations. ASSET affects the entire mature protein of 3FTxs except for segment S8, which is highly conserved. In serine proteases, three of the surface segments are changed rapidly by ASSET, but the rest of the mature protein evolves only by accelerated point mutations. In PLA2 enzymes and metalloproteases, only one and three surface segments (respectively) are changed via ASSET. In all these superfamilies of toxins, ASSET most likely affects their functional properties. However, serine protease inhibitors have evolved by only accelerated point mutations. We propose that ASSET occurs first, resulting in drastic changes in functionally important surface regions, followed by accelerated point mutations in those regions which fine-tune the target specificity. Although the molecular mechanisms of ASSET and accelerated point mutations are unknown, both contribute to the evolution of snake venom toxins and both help to explain the observed functional diversity of toxins and the evolution of new functions in snake venom protein superfamilies.
The protein and cDNA sequences were obtained from the NCBI database. Sequence alignments were done using the DNAMAN program and by manual examination. The intron-exon boundary (marked by a red dashed line in the figures) was identified by comparing the gene and their respective cDNA sequences. In those toxins whose gene structure is not available, the intron-exon boundary was identified by comparing with other toxins whose boundary is known. These segments are identified by comparing with the corresponding sequences, and the point of deletion in amino acid sequence was identified as the boundary of most of the segments. We have analyzed the sequences in different species, but there is no absolute trend in defining the segment, other than high sequence identity. Color coding was used to distinguish segments with distinct % identity; segments with >50% identity are shown in the same color, whereas segments with <50% identity are shown in different colors. Ribbon and surface models were generated from PDB files using DS ViewerPro software.
ASSET: Accelerated Segment Switch in Exon to alter Targeting.
This work was supported by a grant to R.M. Kini from the Biomedical Research Council, Agency for Science, Technology and Research, Singapore.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.