View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector Fooyin J Health Sci 2009;1(2):57−64

REVIEW ARTICLE

Origin of Functional Diversities in Taiwan Banded Krait (Bungarus multicinctus) Three-finger Proteins

Long-Sen Chang*, Pei-Hsiu Kao

Institute of Biomedical Sciences, National Sun Yat-Sen University, Kaohsiung, Taiwan

Received: September 5, 2009 Revised: October 3, 2009 Accepted: October 7, 2009

Taiwan banded krait (Bungarus multicinctus) neurotoxins and neurotoxin homologues, including α-bungarotoxin (Bgt), κ-Bgt, γ-Bgt, BM8, BM10-1, BM10-2 and BM14, have been reported. These proteins have a common three-finger scaffold and conserved cysteine residues at homologous positions. Nevertheless, these proteins show functional diversity and sequence variations in loop regions. The genomic DNAs encoding the precursors of α-Bgt, κ-Bgt, γ-Bgt, BM10-1 and BM14 are organized with three exons and two introns. The intron regions of these genes have a high degree of sequence identity, but the protein-coding regions are highly variable with the exception of the signal peptide region. These findings suggest that B. multicinctus three-finger proteins share a common evolutionary origin, and the evolution of snake venom proteins shows a tendency to diversify their functions, which may be beneficial for catching prey. Given that a multitude of functional diversities is noted with three-finger toxins, protein engineering in highly variable regions without distorting the three-finger scaffold may result in the development biopharmaceutical agents with novel functions of scientific and therapeutic interest.

Key Words: Bungarus multicinctus; functional diversities; genetic organization; snake venom; three-finger protein

Introduction neuromuscular transmission.2,3 To date, more than 120 postsynaptic α-neurotoxins have been isolated in Snake venom contains a number of proteins that the pure state from elapid and hydrophid venoms. have a variety of biochemical and pharmacologic Short α-neurotoxins contain 60−62 amino acid resi- functions. Among them, the neurotoxins are the dues with four disulfide bonds, and long α-neurotoxins primary toxic proteins in cobra, krait, tiger snake comprise 71−74 residues with five disulfide bonds. and sea snake venoms, which block neuromuscular The pairings of four disulfide bonds are similar in transmission and cause death by respiratory paral- both short and long neurotoxins. The extra disulfide ysis. These snake neurotoxins are classified into two bond in the long neurotoxins pinches off a short distinct types, postsynaptic and presynaptic neuro- pentapeptide section in the second disulfide loop, toxins, in relation to the neuromuscular junction.1 thereby shortening the loop to approximately the Postsynaptic α-neurotoxins bind specifically to the same length as the short neurotoxins. Moreover, long nicotinic acetylcholine receptor (nAChR) at the motor neurotoxins extend several amino acid residues endplate and produce a nondepolarizing block of beyond the carboxyl terminus of the short toxins.

*Corresponding author. Institute of Biomedical Sciences, National Sun Yat-Sen University, Kaohsiung 804, Taiwan. E-mail: [email protected]

©2009 Fooyin University 58 L.S. Chang and P.H. Kao

X-ray and nuclear magnetic resonance analyses pharmacologic activities and adopt a three-finger show that all postsynaptic α-neurotoxins adopt a structure.7,8 However, the molecular mechanism similar three-finger structure, but differ in details for exerting their biologic activities still remains such as the extent of the secondary structure and to be resolved. the positions of invariant side-chains.2−4 Searches on sequence similarity of neurotoxins Snake venom κ-neurotoxins exhibit a potent against the GenBank and SWISS-PROT databases effect on the neuronal nicotinic receptor and block have shown the existence of snake venom neuro- transmission in several neuronal systems in which toxin homologues with 10 cysteine residues. The neu- α-neurotoxins have no effect.5 Although diversities rotoxin homologues are usually composed of 65−72 in amino acid sequences and biologic activities are amino acid residues. Unlike the extra cysteine res- noted for α-neurotoxins and κ-neurotoxins, they idues noted in the long α-neurotoxins or κ-neuro- possess a similar three-finger structure (Figure 1).6 toxins, the two additional cysteine residues in the Similar to long α-neurotoxins, κ-neurotoxins con- neurotoxin homologues appear in the N-terminal tain an extra disulfide linkage at the tip of the region and form a disulfide linkage.9 Circular di- second loop (Figure 2). Other snake proteins that chroism measurement shows that their secondary are similar to α-neurotoxins and κ-neurotoxins structures are dominated by a β-sheet similar to that in chain length and disposition of cysteine residues noted with snake venom three-finger proteins.10 are cardiotoxins and cardiotoxin-like basic proteins Moreover, the family of snake venom three-finger (CLBPs) from elapid snake venoms. Cardiotoxins proteins also include muscarinic toxin, anticholineste- and CLBPs have been found to exhibit multiple rase toxin, disintegrin, blocker of L-type calcium channel, blocker of adrenergic receptors, and anti- coagulation toxin.3,11 Obviously, three-finger proteins affect a broad range of molecular targets. Taiwan banded krait (Bungarus multicinctus) three-finger proteins, i.e. α-bungarotoxin (Bgt) and κ-Bgt, have been well studied.2,5 However, ex- tensive analyses on the components of B. multi- cinctus venom suggest that the venom contains neurotoxin homologues with a three-finger struc- ture. Given that protein sequences and biologic α-Bgt κ1-Bgt γ-Bgt activities of these toxic proteins differ, this review Figure 1 Three-dimensional structure of α-bungarotoxin aims to clarify the origin of their function diversi- (Bgt) (PDB code 1ikc), κ1-Bgt (PDB code 1kba), and γ-Bgt ties and the biologic implication with multiple (PDB code 1mr6). toxins in B. multicinctus venom.

α-Bgt IVCHTTA--TSPISA-VTCPPGENLCYRKMWCDAFCSSRG------KVVELGCAATCPS-KKPYEEVT--CCSTDKCNPHPKQRPG κ1-Bgt RTCLISP--SS--TP-QTCPNGQDICFLKAQCDKFCSIRG------PVIEQGCVATCPQFRSNYRSLL--CCTTDNCNH------

BM10-2 KTCFNDDLTNPKTTEL--CRHSMYFCFKNSWIAG-GV------ERIERGCSLTCPDIKYNGKYIY--CCTRDNCNA------Erbc RICFNHQSSQPQTT--KTCSPGESSCYHKQWSDFRG------TIIERGCG--CPTVK-PGINLS--CCESEVCNN------

γ-Bgt MQCKTCSFYTCP-NS-ETCPDGKNICVKRSWTAVRGDGPK------REIRRECAATCPPSKLGL-TVF--CCTTDNCNH------BM10-1 MKCKICHFDTCRAGELKVCASGEKYCFKESWREARG------TRIERGCAATCPKGSVYGLYVL--CCTTDDCN------

BM14 EMCNMCVRPYPFMSS--CCPEGQDRCYKSYWVNENGKQKKYHGKYPVILERGCVTACTGPGSGSIYNLYTCCPTNRCGSSST--SG BM8 EMCNMCVRPYPFMSS--CCPEGQDRCYKSYWVNENGKQEAYHGKYPVILERGCVTACTGPGSGSIYNLYTCCPTNRCGSSST--SG Loop I Loop II Loop III

Figure 2 Sequence alignments of α-bungarotoxin (Bgt), κ1-Bgt, γ-Bgt, BM10-1, BM10-2, BM8, BM14, and Laticauda semifasciata erabutoxin c (Erbc). Laticauda semifasciata erabutoxin c is a short-chain α-neurotoxin. The gray boxes indicate cysteine residues. The disulfide linkages and the segments contributing to the three loops are outlines. Taiwan banded krait three-finger proteins 59 a-Bgt from N. naja atra, N. naja sputatrix and Laticauda semifasciata also comprised three exons and two α-Bgt is a long neurotoxin from Taiwan banded introns.16,17 Their nucleotide sequences share a krait (B. multicinctus) and has been widely used to high degree of sequence homology. Noticeably, the study the structure and function of nAChR because size of intron 2 is more conserved than that of in- of its specific and irreversible binding. Extensive tron 1 during the evolution of snake neurotoxin chemical modifications have been carried out to genes (Figure 4). It is evident that the evolution of study the relationships between structure and α-neurotoxin genes is accompanied by nucleotide function of α-Bgt.1 Lenz et al12 found that purified segment insertion/deletion in intron 1. Nevertheless, rabies virus glycoprotein competes with α-Bgt for the intronic sequences neighboring the splicing binding with AChR. Moreover, human immunodefi- sites are highly conserved. Given that genomic DNAs ciency virus (HIV)-1 glycoprotein 120 inhibits the encoding α-Bgt (Ala31) and α-Bgt (Val31) were iden- binding of α-Bgt with muscle-like nicotinic receptor tified from a single snake, this suggests that the in the human rhabdomyosarcoma cell line TE671, sug- two genes arise from a common ancestor by gene gesting that HIV infection of muscular and neuronal duplication and coexist in the snake genome. cells is mediated through AChR.13 These observa- tions imply that α-Bgt potentially exhibits activity in blocking AChR-mediated viral infection. Moreover, k-Bgt nAChR is expressed in normal bronchial epithelial and non-small cell lung cancer cells (NSCLCs) and In contrast to the finding that α-Bgt specifically is involved in cell growth regulation. Grozio et al14 binds with α7-nAChR, κ-Bgt recognizes α2β4 and found that Naja kaouthia α-cobratoxin, a potent α3β4-nAChR.18 κ-Bgt exhibits a potent effect on α7-nAChR antagonist, decreases NSCLC A549 cell the neuronal nicotinic receptor and blocks trans- growth both in vitro and in vivo. The finding that mission in several neuronal systems in which α-Bgt α-Bgt preferentially binds with α7-nAChR reflects has no effect. Structure−function studies have that α-Bgt, as well as α-cobratoxin, may be useful shown that the residues Arg34 and Pro36 of κ-Bgt adjuvants for treatment of NSCLC. These observa- are involved in the binding of toxin molecules to the tions suggest the potential biomedical application neuronal receptor.18 These findings provide a basis of α-Bgt, in addition to its research interest. for elucidating the contribution of structural vari- Two α-Bgt isotoxins, α-Bgt (Val31) and α-Bgt ations between α-Bgt and κ-Bgt to their effect and (Ala31), have been isolated from the venom of receptor specificities. Almost all known three-finger B. multicinctus. Alignment of the deduced comple- proteins exist as a monomer, but κ-Bgt is a noncova- mentary DNA sequences of α-Bgt (Ala31) and α-Bgt lently linked homodimer. Hemachatus haemacha- (Val31) have demonstrated that one amino acid sub- tus hemextin AB complex is a noncovalently linked stitution results from one base substitution.15 The heterotetramer of two different three-finger sub- genomic DNAs encoding α-Bgt (Ala31) and α-Bgt units,19 while Boiga irregularis irditoxin is a disulfide- (Val31) have three exons, and the three exons are linked heterodimer. 11 The target of hemextin AB is interrupted by two introns (Figures 3 and 4). Struc- coagulation factor VIIa, while irditoxin shows a high turally, the geno mic DNAs of the short neurotoxins affinity toward muscle but not neuronal nAChR.

Exon 1 Exon 2 Exon 3 α-Bgt MKTLLLTLVVVTIVCLDLGYT IVCHTTA--TSPISA-VTCPPGENLCYRKMWCDAFCSSRG------KVVELGCAATCPS-KKPYEEVT--CCSTDKCNPHPKQRPG κ1-Bgt MKTLLLTLVVVTIVCLDLGYT RTCLISP--SS--TP-QTCPNGQDICFLKAQCDKFCSIRG------PVIEQGCVATCPQFRSNYRSLL--CCTTDNCNH------γ-Bgt MKTLLLTLVVVTIVCLDLGYT MQCKTCSFYTCP-NS-ETCPDGKNICVKRSWTAVRGDGPK------REIRRECAATCPPSKLGL-TVF--CCTTDNCNH------BM10-1 MKTLLLTLVVVTIVCLDLGYT MKCKICHFDTCRAGELKVCASGEKYCFKESWREARG------TRIERGCAATCPKGSVYGLYVL--CCTTDDCN------BM10-2 MKTLLLSLVVLTIACLDLGYT KTCFNDDLTNPKTTEL--CRHSMYFCFKNSWIAG-GY------ERIERGCSLTCPDIKYNGKYIY--CCTRDNCNA------BM14 MKTLLLTLVVVTIICLDLGYT EMCNMCVRPYPFMSS--CCPEGQDRCYKSYWVNENGKQKKYHGKYPVILERGCVTACTGPGSGSIYNLYTCCPTNRCGSSST--SG BM8 MKTLLLTLVVVTIICLDLGYT EMCNMCVRPYPFMSS--CCPEGQDRCYKSYWVNENGKQEAYHGKYPVILERGCVTACTGPGSGSIYNLYTCCPTNRCGSSST--SG Loop I Loop II Loop III

Figure 3 Alignment of the sequences of protein precursors encoded by α-bungarotoxin (Bgt), κ1-Bgt, γ-Bgt, BM10-1, and BM14 genes. α-Bgt, κ1-Bgt, γ-Bgt, BM10-1, and BM14 genes share the same genetic organization. The signal peptide region is underlined and amino acid residues encoded by exons 1−3 are outlined. The genes encoding BM10-2 and BM8 have not been determined yet. 60 L.S. Chang and P.H. Kao

α-Bgt γ-Bgt Intron 1 Intron 2 Intron 1 Intron 2 (1793 bp) (537 bp) (1282 bp) (517 bp)

(1) (33) (33) (74) (1) (34) (34) (68)

κ1-Bgt BM14 Intron 1 Intron 2 Intron 1 Intron 2 (1065 bp) (537 bp) (1188 bp) (544 bp)

(1) (31) (31) (63) (1) (34) (34) (82)

BM10-1 Erabutoxin c Intron 1 Intron 2 Intron 1 (197 bp) Intron 2 (1203 bp) (547 bp) (538 bp)

(1) (36) (36) (66) (1) (34) (34) (62)

Figure 4 Schematic representation of the structures of α-bungarotoxin (Bgt), κ1-Bgt, γ-Bgt, BM10-1, BM14 and Laticauda semifasciata erabutoxin c genes. The white boxes and black boxes represent the signal peptide region and mature protein-coding regions, respectively. The gray boxes represent the 5Ј- and 3Ј-noncoding regions of the genes. The size of introns is indicated in italics within parentheses. The non-italicized numbers in parentheses indicate the number of amino acid residues encoded by mature protein-coding regions.

The oligomeric structure is believed to be essential that the size of intron 2 is similar in these toxin for the activities of these toxic proteins. Formation genes, there is a notable variation with the size of of the dimeric structure leads to an α-cobratoxin intron 1. dimer that not only interacts with α7-nAChR but, in contrast to the α-cobratoxin monomer, also blocks α3β2-nAChR.20 In addition to amino acid substitu- g-Bgt tions, oligomerization further increases the func- tional diversities of three-finger toxins. In 1999, Aird et al22 isolated γ-Bgt from B. multicinc- To date, six κ-Bgts have been identified from tus venom. It has been shown that γ-Bgt contains an polymerase chain reaction-amplified B. multicinc- extra disulfide linkage within its N-terminal region, tus venom gland complementary DNAs.21 However, which is in contrast to an extra disulfide linkage in only κ1-, κ2- and κ3-Bgt have been isolated from α-Bgt and κ-Bgt that is located at the tip of the sec- B. multicinctus venom. Statistical analyses on the ond loop (Figure 2). Moreover, γ-Bgt has an Arg-Gly- numbers of nucleotide substitutions per synony- Asp (RGD) motif, which suggests its interaction with mous site and nonsynonymous site in the protein- integrin. Similar to α-Bgt and κ-Bgt, γ-Bgt also adopts coding region suggest that an adaptive selection was a three-finger structure (Figure 1). Mutagenesis stud- involved in the evolution of κ-Bgt. Two genomic ies have suggested that the amino acid residues DNAs encoding κ1-Bgt and κ3-Bgt have been cloned flanking the RGD motif might control the width of from the B. multicinctus genome. κ-Bgt genes and the RGD loop and affect the interaction specificity α-Bgt genes share the same exon-intron organiza- of γ-Bgt with integrin.23 Given that integrin−ligand tion and a high degree of nucleotide sequence iden- interactions play an important role in regulating tity (Figures 3 and 4), supporting the notion that biologic processes such as angiogenesis and metas- these genes might have originated from the same tasis of cancer, snake venom RGD-containing pro- ancestor.21 Comparative analyses on the sequence teins may be employed as a structural scaffold in identity for the common regions, of κ-Bgt and α-Bgt developing agents with anticancer activity. Alterna- genes show that the protein-coding regions are tively, γ-Bgt was found to interact with Torpedo more diversified than the intron regions, except for nAChR and muscarinic AChR (mAChR).24 Likewise, the signal peptide region. In contrast to the findings N. kaouthia weak toxin interacts with all mAChR Taiwan banded krait three-finger proteins 61 subtypes and α7-nAChR.25 Thus, γ-Bgt and N. kaouthia weak toxin could be regarded as multifunctional three-finger toxins. Since targets are available in several organs of prey or in different prey species, the significance of multitargeted toxins for prey capture may be greater than that of highly specific toxins. The gene encoding the γ-Bgt precursor is organ- ized with three exons and two introns, which is virtu- ally identical to that of α-Bgt and κ-Bgt (Figure 4).24 BM10-1 BM10-2 Moreover, protein-coding regions of γ-Bgt, κ-Bgt and α-Bgt are interrupted by introns in similar posi- tions (Figure 3). These findings support the idea that B. multicinctus three-finger proteins share a common origin.

Taiwan Banded Krait Neurotoxin Homologues, BM10-1, BM10-2, BM8 BM8 BM14 and BM14 Figure 5 Models of BM10-1, BM10-2, BM8 and BM14. Homology models of BM10-1 and BM10-2 were built using Bungarus multicinctus venom can be separated κ 24 the structure of 1-bungarotoxin (PDB code 1kba) as a into 23 fractions on an SP-Sephadex C-25 column. template under the homology program. Homology mod- The results of sequence determination and mass els of BM8 and BM14 were built using the structure of analyses have shown that fractions 8 and 14 con- α-bungarotoxin (PDB code 1ikc) as a template. tain novel proteins. The proteins purified from fractions 8 and 14 are designated as BM8 and BM14, respectively.26 After separation on a reverse phase α-Bgt. The response to carbachol did not signifi- high-performance liquid chromatography column, cantly revert after being extensively washed with fraction 10 was verified to contain two toxic pro- Krebs solution to remove BM10-2 and α-Bgt. In teins, BM10-1 and BM10-2.27 BM8 and BM14 are iso- contrast, the inhibition of carbachol-induced mus- meric proteins and are nearly identical except for cle contraction caused by BM10-1 could be recov- the substitution of Glu37-Ala38 in BM8 for Lys37- ered by thoroughly washing out the toxin. Thus, Lys38 in BM14 (Figure 2). BM10-1, BM8 and BM14 con- BM10-1 reversibly binds with postsynaptic acetyl- tain 10 cysteine residues, while BM10-2 has eight choline receptors, while BM10-2 binds with postsy- cysteine residues. Alignment of the amino acid se- naptic acetylcholine receptors in an irreversible quence shows that eight out of the 10 cysteine resi- manner. Nirthanan and Gwee9 suggested that the dues in BM10-1, BM8 and BM14 are located at the reversibility of α-neurotoxin action at the neuromus- same conserved positions as those in BM10-2, α-Bgt, cular junction is not always a reflection of its bind- κ-Bgt and γ-Bgt (Figure 2). The two additional cys- ing affinity to the receptor and may perhaps be teine residues of BM8 and BM14 are at positions associated with a specific area of interaction on 6 and 16, while the two extra cysteine residues of the toxin molecule, distinct from the receptor rec- BM10-1 are at positions 6 and 11, which is similar ognition site. It is evident that amino acid substitu- to γ-Bgt. Although three-dimensional structures of tions along with the evolution of three-finger toxins BM10-1, BM10-2, BM8 and BM14 have not been de- alter the binding affinity as well as the interaction termined yet, circular dichroism measurement has mode with their targets. shown that the secondary structure of BM10-1, It has been shown that injection of BM8 and BM10-2, BM8 and BM14 are dominated by a β-sheet BM14 is not toxic to mice at doses up to 3 mg/kg.26 structure.26,27 Computer modeling shows that BM10- The two proteins are unable to compete with radio- 1, BM10-2, BM8 and BM14 adopt a “three-loop” iodinated α-Bgt for binding with Torpedo nAChR.26 folded structure (Figure 5). The inability of BM8 and BM14 to bind with rat BM10-1, BM10-2 and α-Bgt exhibit activity for in- cortex M1 mAChR subtype has been observed, hibiting carbachol-induced contraction of chicken but BM14 exhibits an affinity for binding with the biventer cervicis muscle.27 The dose for achieving rat ventricle M2 mAChR subtype.26 Taken together, 50% inhibition was 89.3nM and 23.7nM for BM10-1 these results indicate that B. multicinctus neuro- and α-Bgt, respectively. The inhibition of carbachol- toxin homologues exhibit species-specific neurotox- induced muscle contraction by BM10-2 is approxi- icity. Previous studies have shown that three-finger mately a two-order magnitude higher than that of toxins Boiga dendrophila denmotoxin and Boiga 62 L.S. Chang and P.H. Kao irregularis irditoxin exhibit a taxon-specific le- has not been fairly delineated, the diverse functions thality toward birds and are nontoxic toward of three-finger proteins may have a selective ad- mice.11,28 vantage for snakes in catching prey and for defense Although the genetic organization of BM10-2 against predators. A recent study supports the idea and BM8 genes remains elusive, the BM10-1 and that predator−prey interactions contribute to the BM14 genes are organized with three exons and two diversity and evolution of snake venom components.38 introns (Figures 3 and 4). The percentage similarity Moreover, dietary habits have been found to alter between the common regions of BM10-1, BM14, venom composition of Aipysurus eydouxii.39 Thus, α-Bgt, κ-Bgt and γ-Bgt genes indicates that the it is conceivable that snake venom proteins have protein-coding regions are more diversified than evolved to functionally alter the physiologic activi- the intron regions, except for the signal peptide ties along with predator−prey interaction, and conse- region.25−27 Alignment of the promoter sequences of quently snake venom has become a complex mixture BM10-1, BM14, α-Bgt, κ-Bgt and γ-Bgt genes suggests of pharmacologically active polypeptide toxins. that their promoter regions are highly conserved Several multigene families including phospholi- and contain some consensus transcriptional factor pase A2 (PLA2) enzymes, Kunitz-type protease inhibi- binding sites.17 The putative transcriptional factor tors, three-finger toxins, snake venom hemorrhagic binding sites include NF-1, TATA box, CACCC-binding metalloproteinases and disintegrins have been factor, and EFII. The functional involvement of CACCC- determined.38,40−46 Analyses on the genetic struc- binding factor in the regulation of N. naja sputatrix ture of these genes suggest that these families have cardiotoxin gene has been reported previously.29 evolved by gene duplication, followed by functional This finding indicates that the expression of these diversification by positive selection. Previous stud- genes, at least in part, is regulated under the same ies on the genes encoding snake PLA2 from differ- transcriptional mechanism. Comparative analyses on ent species have revealed that the protein-coding 40,41 three-finger toxin genes suggest an evolution mode regions of PLA2 genes are unusually variable. with sequential steps including gene duplication, In contrast, the introns of venom gland PLA2 genes divergence of the snake subfamily, divergence of evolved at a similar rate to neutral evolution and snake species, and functional divergence of snake are highly conserved. Thus, accelerated evolution venom proteins.30 With respect to the genetic di- of exon regions have been proposed to be involved vergence of snake toxins, the variations in gene in the evolution of PLA2 genes to acquire diverse transcription may represent another event in evo- functions. Recently, Doley et al47 proposed that snake lutionary diversification of toxin genes. venom toxin superfamilies evolve through an accel- erated segment switch in exons to alter targeting, resulting in drastic changes in functionally impor- The Evolution of B. multicinctus tant surface regions, followed by accelerated point Three-finger Proteins mutations in those regions that fine-tune the target specificity.47 The presence of fused protein in the To date, three-finger proteins, including BM10-1, Sistrurus catenatus edwardsii venom gland transcrip- BM10-2, BM8, BM14, α-Bgt, κ-Bgt and γ-Bgt, have tome indicates that exon shuffling and transcrip- been purified from B. multicinctus venom. The genes tional splicing contribute to generate the diversity encoding these proteins share a common genetic of toxins and toxin isoforms observed among snake structure and high degree of sequence identity, venoms.48 The highly conserved introns in snake suggesting that these genes might originate from a venom gene families probably facilitate segment common evolutionary origin followed by sequen- switch in exons and/or exon shuffling via homolo- tial gene duplication and divergence. In addition gous recombination. Taken together, the evolution of to genomic DNAs encoding B. multicinctus three- the B. multicinctus three-finger protein gene family finger proteins, N. naja atra, N. naja sputatrix and might have resulted from genetic divergence and Pseudonaja textilis α-neurotoxin, cardiotoxin, CLBP in response to predator-prey interactions. and neurotoxin homologue genes have been deter- mined.31−37 The genomic DNAs of these three-finger proteins have the same genetic organization and Conclusion high degree of sequence identity. Accordingly, the other three-finger proteins, including muscarinic The interaction of three-finger proteins with their toxin, anticholinesterase toxin, disintegrin, L-type molecular targets has been well-described and found calcium channel blocker, anticoagulation toxin and to involve one or more different regions. Structural adrenergic receptor blocker, are believed to have the variability in loop regions contributes to the activ- same genomic organization. Although the biologic ity of three-finger proteins and is the basis of the significance with multiple toxins in snake venom functional diversity of the three-finger fold.9 Thus, Taiwan banded krait three-finger proteins 63 it has been proposed that the three-finger scaffold 11. Pawlak J, Mackessy SP, Sixberry NM, et al. Irditoxin, a novel is used by snakes to “create” different combinations covalently linked heterodimeric three-finger toxin with high taxon-specific neurotoxicity. FASEB J 2009;23:534−45. of functional groups, generating different target 12. Lentz TL, Burrage TG, Smith AL, et al. Is the acetylcholine 9 specificities. The evolution of three-finger pro- receptor a rabies virus receptor? Science 1982;215:182−4. teins along with functional diversity is in line with 13. Neri P, Bracci L, Rustici M, et al. Sequence homology between the goal of protein engineering, leading to the pro- HIV gp120, rabies virus glycoprotein, and snake venom neu- duction of mutated toxins that have a conserved rotoxins. Is the nicotinic acetylcholine receptor an HIV receptor? Arch Virol 1990;114:265−9. structural scaffold but different structural flexibil- 14. Grozio A, Paleari L, Catassi A, et al. Natural agents targeting ity and biologic activities. The grafting of loop I the α7-nicotinic-receptor in NSCLC: a promising prospective and part of loop II of fasciculin (anticholinesterase in anti-cancer drug development. Int J Cancer 2008;122: toxin) into the homologous location in toxin α 1911−5. α 15. Chang LS, Lin SK, Huang HB, et al. Genetic organization ( -neurotoxin) contributes to the appearance of an- α 49 of -bungarotoxins from Bungarus multicinctus (Taiwan ticholinesterase activity with the chimeric protein. banded krait): evidence showing that the production of This suggests the possibility that generation of ra- α-bungarotoxin isotoxins is not derived from edited mRNAs. tionally designed novel activities use three-finger Nucleic Acids Res 1999;27:3970−5. scaffolds to develop proteins with novel functions of 16. Yee JSP, Nanling G, Afifiyan F, et al. Snake postsynaptic scientific and therapeutic interest. Thus, understand- neurotoxins: gene structure, phylogeny and applications in research and therapy. Biochimie 2004;86:137−49. ing the origin of the functional diversity of B. multi- 17. Chang LS. Genetic diversity in snake three-finger proteins cinctus three-finger proteins may contribute to the and phospholipase A2 enzymes. Toxin Reviews 2007;26: development of biomedical applications. 143−67. 18. Chiappinelli VA, Weaver WR, Mclame KE, et al. Binding of native κ-neurotoxins and site-directed mutants to nicotinic acetylcholine receptors. Toxicon 1996;34:1243−56. Acknowledgments 19. Banerjee Y, Mizuguchi J, Iwanaga S, et al. Hemextin AB complex: a unique anticogulant protein complex from The work was supported by the National Science Hemachatus haemachatus (African Ringhals cobra) venom Council, Taiwan. The authors express sincere ap- that inhibits clot initiation and factor VIIa activity. J Biol Chem 2005;280:42601−11. preciation to their colleagues for their contribu- 20. Osipov AV, Kasheverov IE, Makarova YV, et al. Naturally tion to this study. occurring disulfide-bound dimers of three-fingered toxins: a paradigm for biological activity diversification. J Biol Chem 2008;283:14571−80. References 21. Chang LS, Chung C, Lin J, et al. Organization and phylo- genetic analysis of κ-bungarotoxin genes from Bungarus multicinctus (Taiwan banded krait). Genetica 2002;115: 1. Yang CC, Chang LS. Biochemistry and molecular biology of 213−21. − snake neurotoxin. J Chinese Chem Soc 1999;46:319 32. 22. Aird SD, Womble GC, Yates III JR, et al. Primary structure 2. Endo T, Tamiya N. Structure-function relationship of post- of γ-bungarotoxin, a new postsynaptic neurotoxin from the synaptic neurotoxins from snake venoms. In: Harvey AL, ed. venom of Bungarus multicinctus. Toxicon 1999;37:609−25. − Snake Toxins. New York: Pergamon Press, 1991:165 222. 23. Shiu JH, Chen CY, Chang LS, et al. Solution structure of 3. Tsetlin VI. Snake venom α-neurotoxins and other “three- γ-bungarotoxin: the functional significance of amino acid finger” proteins. Eur J Biochem 1999;264:281−6. residues flanking the RGD motif in integrin binding. Proteins 4. Kini RM. Molecular moulds with multiple missions: functional 2004;57:839−49. sites in three-finger toxins. Clin Exp Pharmacol Physiol 2002; 24. Chang LS, Chung C, Wu BN, et al. Characterization and gene 29:815−22. organization of Taiwan banded krait (Bungarus multicinctus) 5. Chiappinelli VA. κ-Neurotoxins and α-neurotoxins: effects on γ-bungarotoxin. J Protein Chem 2002;21:223−9. neuronal nicotinic acetylcholine receptors. In: Harvey AL, 25. Mordvintsev DY, Polyak YL, Rodionov DI, et al. Weak toxin ed. Snake Toxins. New York: Pergamon Press, 1991:223−58. WTX from Naja kaouthia cobra venom interacts with both 6. Dewan JC, Grant GA, Sacchettini JC. Crystal structure of nicotinic and muscarinic acetylcholine receptors. FEBS J kappa-bungarotoxin at 2.3 Å resolution. Biochemistry 1994; 2009;276:5065−75. 33:13147−54. 26. Chung C, Wu BN, Yang CC, et al. Muscarinic toxin-like pro- 7. Sivaraman T, Kumar TKS, Jayaraman G, et al. Characterization teins from Taiwan banded krait (Bungarus multicinctus) of a partially structured state in an all-beta-sheet protein. venom: purification, characterization and gene organization. Biochem J 1997;321:457−64. Biol Chem 2002;383:1397−406. 8. Sun YJ, Wu WG, Chiang CM, et al. Crystal structure of car- 27. Chang LS, Chung C, Liou JC, et al. Novel neurotoxins from diotoxin V from Taiwan cobra venom: pH-dependent con- Taiwan banded krait (Bungarus multicinctus) venom: puri- formational change and a novel membrane-binding motif fication, characterization and gene organization. Toxicon identified in the three-finger loops of P-type cardiotoxin. 2003;42:323−30. Biochemistry 1997;36:2403−13. 28. Pawlak J, Mackessy SP, Fry BG, et al. Denmotoxin, a three- 9. Nirthanan S, Gwee MCE. Three-finger α-neurotoxins and the finger toxin from the Colubrid snake Boiga dendrophila nicotinic acetylcholine receptor, forty years on. J Pharmacol (Mangrove Carsnake) with bird-specific activity. J Biol 2004;94:1−17. Chem 2006;281:29030−41. 10. Chang LS, Lin SR, Wang JJ, et al. Structure-function studies 29. Ma D, Armugam A, Jeyaseelan K. Expression of cardiotoxin-2 on Taiwan cobra long neurotoxin homolog. Biochim Biophys gene: cloning, characterization and deletion of the promoter. Acta 2000;1480:293−301. Eur J Biochem 2001;268:1844−50. 64 L.S. Chang and P.H. Kao

30. Tamiya T, Fujimi TJ. Molecular evolution of toxin genes in 40. Ohno M, Menez R, Ogawa T, et al. Molecular evolution of Elapidae snakes. Mol Divers 2006;10:529−43. snake toxins: is the functional diversity of snake toxins asso- 31. Fuse N, Tsuchiya T, Nonomura Y, et al. Structure of the ciated with a mechanism of accelerated evolution? Prog short-chain neurotoxin, erabutoxin c, precursor gene. Eur J Nucleic Acid Res Mol Biol 1998;59:307−64. Biochem 1990;193:629−33. 41. Kordis D, Gubensek F. Adaptive evolution of animal toxin 32. Chang LS, Chou YC, Lin SR, et al. A novel neurotoxin, cobro- multigene families. Gene 2000;261:43−52. toxin b, from Naja naja atra (Taiwan cobra) venom: purifi- 42. Zupunski V, Kordis D, Gubensek F. Adaptive evolution in the cation, characterization and gene organization. J Biochem snake venom Kunitz/BPTI protein family. FEBS Lett 2003; 1997;122:1252−9. 547:131−6. 33. Chang LS, Lin J, Chou YC, et al. Genomic structures of car- 43. Calvete JJ, Marcinkiewicz C, Monleon D, et al. Snake venom diotoxin 4 and cobrotoxin from Naja naja atra (Taiwan disintegrins: evolution of structure and function. Toxicon cobra). Biochem Biophys Res Commun 1997;239:756−62. 2005;45:1063−74. 34. Lachumanan R, Armugam A, Tan CH, et al. Structure and 44. Soto JG, White SA, Reyes SR, et al. Molecular evolution of organization of the cardiotoxin genes in Naja naja sputatrix. PIII-SVMP and RGD disintegrin genes from the genus Crotalus. FEBS Lett 1998;433:119−24. Gene 2007;389:66−72. 35. Afifiyan F, Armugam A, Tan CH, et al. Postsynaptic α- 45. Lynch VJ. Inventing an arsenal: adaptive evolution and neurotoxin gene of the spitting cobra, Naja naja sputatrix: neofunctionalization of snake venom phospholipase A2 structure, organization and phylogenetic analysis. Genome genes. BMC Evol Biol 2007;7:2. Res 1999;9:259−66. 46. Gibbs HL, Rossiter W. Rapid evolution by positive selection 36. Gong N, Armugam A, Jeyaseelan K. Molecular cloning, char- and gene gain and loss: PLA2 venom genes in closely related acterization and evolution of the gene encoding a new Sistrurus Rattlesnakes with divergent diets. J Mol Biol 2008; group of short-chain α-neurotoxins in an Australian elapid, 66:151−66. Pseudonaja textilis. FEBS Lett 2000;473:303−10. 47. Doley R, Mackessy SP, Kini RM. Role of accelerated segment 37. Gong N, Armugam A, Mirtschin P, et al. Cloning and charac- switch in exons to alter targeting (ASSET) in the molecular terization of the pseudonajatoxin b precursor. Biochem J evolution of snake venom proteins. BMC Evol Biol 2009; 2001;358:647−56. 9:146. 38. Pahari S, Bickford D, Fry BG, et al. Expression pattern of 48. Pahari S, Mackessy SP, Kini RM, The venom gland trans- three-finger toxin and phospholipase A2 genes in the venom criptome of the desert massasauga rattlesnake (Sistrurus glands of two sea snakes, Lapemis curtus and Acalyptophis catenatus edwardsii): towards an understanding of venom peronii: comparison of evolution of theses toxins in land composition among advanced snakes (Superfamily snakes, sea kraits and sea snakes. BMC Evol Biol 2007;7:175. Colubroidea). BMC Mol Biol 2007;8:115. 39. Li M, Fry BG, Kini RM. Egg-only diet: its implications for the 49. Le-Du MH, Ricciardi A, Khayati M, et al. Stability of a struc- toxin profile changes and ecology of the marbled sea snake tural scaffold upon activity transfer: X-ray structure of a (Aipysurus eydouxii). J Mol Evol 2005;60:81−9. three-finger chimeric protein. J Mol Biol 2000;296:1017−26.