MBE Advance Access originally published online on August 13, 2008
Molecular Biology and Evolution 2008 25(11):2391-2407; doi:10.1093/molbev/msn179
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Research Articles |
Evolution of Snake Venom Disintegrins by Positive Darwinian Selection
,1,
* Instituto de Biomedicina de Valencia, Consejo Superior de Investigaciones Biológicas, Valencia, Spain
Instituto Cavanilles de Biodiversidad y Biología Evolutiva, Universidad de Valencia, Polígono La Coma s/n, Valencia, Spain
en Epidemiología y Salud Pública, Valencia, Spain
E-mail: jcalvete{at}ibv.csic.es.
 |
Abstract |
|---|
 TOP Abstract IntroductionMaterials and MethodsResults and DiscussionConcluding Remarks and...Supplementary MaterialAcknowledgementsReferences |
|---|
PII-disintegrins, cysteine-rich polypeptides broadly distributed in the venoms of geographically diverse species of vipers and rattlesnakes, antagonize the adhesive functions of β1 and β3 integrin receptors. PII-disintegrins evolved in Viperidae by neofunctionalization of disintegrin-like domains of duplicated PIII-snake venom hemorrhagic metalloproteinase (SVMP) genes recruited into the venom proteome before the radiation of the advanced snakes. Minimization of the gene (loss of introns and coding regions) and the protein structures (successive loss of disulfide bonds) underpins the postduplication divergence of disintegrins. However, little is known about the underlying genetic mechanisms that have generated the structural and functional diversity among disintegrins. Phylogenetic inference and maximum likelihood–based codon substitution approaches were used to analyze the evolution of the disintegrin family. The topology of the phylogenetic tree does not parallel that of the species tree. This incongruence is consistent with that expected for a multigene family undergoing a birth-and-death process in which the appearance and disappearance of loci are being driven by selection. Cysteine and buried residues appear to be under strong purifying selection due to their role in maintaining the active conformation of disintegrins. Divergence of disintegrins is strongly influenced by positive Darwinian selection causing accelerated rate of substitution in a substantial proportion of surface-exposed disintegrin residues. Global and lineage-specific sites evolving under diversifying selection were identified. Several sites are located within the integrin-binding loop and the C-terminal tail, two regions that form a conformational functional epitope. Arginine-glycine-aspartic acid (RGD) was inferred to represent the ancestral integrin-recognition motif, which emerged from the subgroup of PIII-SVMPs bearing the RDECD sequence. The most parsimonious nucleotide substitution model required for the emergence of all known disintegrin's integrin inhibitory motifs from an ancestral RGD sequence involves a minimum of three mutations. The adaptive advantage of the emergence of motifs targeting β1 integrins and the role of positively selected sites located within nonfunctional disintegrin regions appear to be difficult to rationalize in the context of a predator–prey arms race. Perhaps, this represents a consequence of the neofunctionalization potential of the disintegrin domain, a feature that may underlie its recruitment into the venom proteome followed by its successful transformation into a toxin.
Key Words: molecular evolution • snake venom disintegrins • adaptive evolution • positive Darwinian selection • phylogeny
|
Introduction |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResults and DiscussionConcluding Remarks and...Supplementary MaterialAcknowledgementsReferences |
|---|
Venom, produced by a pair of specialized glands in the upper jaw (Kochva 1987
; Jackson 2007), represented a key innovation in ophidian evolution that allowed advanced snakes to transition from a mechanical (constriction) to a chemical (venom) means of subduing and digesting prey larger than themselves. Venom toxins likely evolved from endogenous proteins with normal physiological functions that were recruited into the venom proteome before the radiation of the advanced snakes (Fry and Wüster 2004; Fry 2005; Fry et al. 2006). Toxic venom proteins play a number of roles, such as immobilizing, paralyzing, killing, liquefying prey, and deterring competitors. Venoms from Viperinae (vipers) and Crotalinae (pitvipers) subfamilies of Viperidae snakes contain proteins that interfere with the coagulation cascade, the normal hemostatic system and tissue repair, and human envenomations are often characterized by clotting disorders, hypofibrinogenemia, and local tissue necrosis (Markland 1998; Fox and Serrano 2005a). Despite being complex mixtures, viperid venom proteins belong to only a few major protein families, including enzymes (serine proteinases, Zn2+-metalloproteases, L-amino acid oxidase, group II PLA2) and proteins without enzymatic activity (disintegrins, C-type lectins, natriuretic peptides, ohanin, myotoxins, CRISP toxins, nerve and vascular endothelium growth factors, cystatin and Kunitz-type protease inhibitors) (reviewed by Calvete, Juárez, and Sanz 2007). Notably, most venom toxins are extensively cross-linked by disulfide bonds and have flourished into functionally diverse, toxin multigene families that exhibit interfamily, intergenus, interspecies, and intraspecific variability. The occurrence in the same venom of a diversity of isoforms of proteins belonging to the same family but differing from each other in their pharmacological effects likely results from gene duplication followed by accelerated evolution by positive selection and neofunctionalization of duplicated gene copies and suggests an important role for balancing selection in maintaining high levels of functional variation in venom proteins within populations (reviewed by Richman 2000).
In line with studies on a number of animal toxins (Kordis et al. 2002), snake venom toxin gene families, such as PLA2s (Lynch 2007 and references cited; Gibbs and Rossiter 2008), C-type lectin-like proteins (Ogawa et al. 2005), and serine proteinase inhibitors (Zupinski et al. 2003), are reported to have evolved new functions by the process of neofunctionalization following gene duplication and evolution under strong positive adaptive selection (Ohno 1970). The evolutionary pressure acting to promote high levels of variation in venom proteins may be part of a predator–prey arms race that allows a sit-and-wait predator, such as a snake, to adapt to a variety of different prey, each most efficiently subdued with a different venom formulation (Greene 1983; Daltry et al. 1996; Ménez 2002).
Snake venom hemorrhagic metalloproteinases (SVMPs) are evolutionarily close to mammalian matrix-degrading metalloproteinases and proteins of the ADAM (a disintegrin and metalloproteinase) reprolysin subfamily of Zn2+-metalloproteinases (PFAM family PF01421; http://pfam.sanger.ac.uk/family) (Moura-Da-Silva et al. 1996). The phylogenetic distribution of the mammalian and snake venom proteins indicates that ADAMs and SVMPs have evolved relatively late from a common ancestor gene both through speciation (after mammals and reptiles diverged) and gene duplication, followed by divergence of the copies mediated by positive Darwinian selection (Glassey and Civetta 2004). Snake venom hemorrhagins have been classified according to their domain structure (Fox and Serrano 2005b). The PIII class comprises the closest homologues of cellular ADAMs and are large multidomain toxins (60–100 kDa) built up by an N-terminal metalloproteinase domain and C-terminal disintegrin-like and cysteine-rich domains. The class PII metalloproteinases (30–60 kDa) contain a disintegrin domain at the carboxyl terminus of the metalloproteinase domain. PI metalloproteinases (20–30 kDa) are single-domain proteins.
Disintegrins, a family of small (40–100 amino acids) cysteine-rich polypeptides, are broadly distributed in the venoms of geographically diverse species of vipers and rattlesnakes. Disintegrins are released in viper venoms by proteolytic processing of Serpentes-specific PII-SVMP precursors (Kini and Evans 1992) or synthesized from messenger RNAs lacking the metalloprotease-coding region (Okuda et al. 2002) and selectively block the function of cell surface adhesive receptors of the integrin family (Calvete et al. 2005; Sanz et al. 2006). Disintegrins can be classified according to their length and number (4–7) of disulfide bonds into long, medium-sized, dimeric, and short subfamilies (Calvete et al. 2003), and their integrin inhibitory activity critically depends on the appropriate pairing of cysteines that determines the conformation of the inhibitory loop that harbors an active tripeptide located at the apex of a mobile loop protruding 14–17 Å from the protein core (Monleón et al. 2003, 2005; Calvete et al. 2005). Hence, disintegrins have found numerous applications in studies on a variety of biological processes in which integrins play pivotal roles (Niewiarowski et al. 2002; Marcinkiewicz 2007). Selective blockage of integrins is a desirable goal for the therapy of a number of pathological conditions, including acute coronary ischemia and thrombosis (
IIbβ3), tumor metastasis, osteoporosis, restenosis, rheumatoid arthritis (vβ3), bacterial infection, vascular disease (Vβ3), inflammation, autoimmune diseases (4β1, 4β7 and 9β1), and tumor angiogenesis (1β1 y vβ3). The relevant integrin receptors involved in the above listed pathologies are among the targets of many disintegrins (fig. 1D).
|
The current view, schematically depicted in figure 1B and C, is that amino acid substitutions along with minimization of both the gene organization (i.e., loss of introns and coding regions) and the protein structure (including the selective loss of pairs of cysteine residues engaged in the formation of disulfide bonds) underpin the evolution of the snake venom disintegrin family (Calvete et al. 2003
, 2005; Bazaa et al. 2007). However, little is known about the evolutionary history and the molecular basis underlying the adaptive evolution of disintegrins, and the sites evolving under either stochastic mutational process or accelerated amino acid substitutions have not been elucidated. Therefore, the aim of this study was to investigate the evolutionary process that resulted in the structural and functional diversification within the disintegrin family. Here, we report the identification of a subset of amino acid sites that are targets of positive Darwinian selection causing accelerated structural and functional diversification among, and within, the different disintegrin subfamilies. Phylogenetic clustering within the snake species tree and topology mapping of positively selected sites to structural and functional regions of disintegrin molecules allows us to define molecular and temporal features along the diversification pathway of disintegrin lineages. |
Materials and Methods |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResults and DiscussionConcluding Remarks and...Supplementary MaterialAcknowledgementsReferences |
|---|
Sequence Alignment and Phylogenetic Reconstruction
Protein and DNA sequences of short, medium, long, and dimeric disintegrins and PIII-SVMPs were retrieved from the National Center for Biotechnology Information database (http://www.ncbi.nlm.nih.gov). Partial sequences, sequences with indels that caused frameshifts, and sequences with premature stop codons were excluded from the analysis. Depending on the analysis, two sequence data sets were used (table 1): one including 119 distinct amino acid sequences (98 derived by protein analysis and 21 deduced by DNA cloning) from 19 genera and another comprising 70-nt sequences from 13 genera. In order to obtain an accurate codon-based alignment from the nucleotide data set, needed for the positive selection analyses, the corresponding translated amino acid sequences were aligned using ClustalW (Thompson et al. 1994
). However, due to the structural differences among the different families, a manual refinement was necessary to correctly identify homologous positions. The amino acid data set was also aligned using ClustalW and manually refined using the previously described amino acid alignment (Calvete et al. 2003) as a guide. Phylogenetic inference analyses were carried out based on the amino acid sequence alignment, which combined protein-derived sequences and DNA-translated amino acid sequences. The maximum likelihood (ML) phylogenetic tree was obtained using the PHYML program (Guindon and Gascuel 2003). The Jones, Taylor, and Thornton model (JTT) for amino acid substitutions incorporating a gamma distribution with eight rate categories and a fraction of invariant sites to account for substitution rate heterogeneity among sites was used. The robustness of the topology was evaluated through 500 bootstrap pseudorandom replicates. This topology was used for the ensuing positive selection analyses but converting branch lengths to substitutions per codon site as implemented in baseml of the PAML package (Yang 1997). Additional support values were obtained through Bayesian inference of phylogeny using MrBayes 3.1.2 (Ronquist and Huelsenbeck 2003) using the same evolutionary model (JTT + I + G) and two runs of four chains during 1.5 million generations. Convergence of the parameters was evaluated using Tracer 1.4 (Rambaut and Drummond 2007) in order to ensure that the effective sample size of all parameters was larger than 100. The first 10% of trees were discarded as burn-in, and a majority rule consensus tree was obtained from the remaining samples after the two runs were combined.
|
Comparing Substitution Rates across Lineages
For computing substitution rate differences between the different disintegrin lineages, the nucleotide alignment was analyzed using program RRTree 1.0 (Robinson-Rechavi and Huchon 2000). The assumed null hypothesis was that the two lineages are not accumulating substitutions at a significantly different rate. Sequences were divided into four groups according to their structure assignment (long, medium, dimeric, and short), and all possible pairwise comparisons of substitution rates between the four groups were performed. The PIII sequences were chosen as the common outgroup. The significance of the difference in number of substitutions for each comparison was assessed at the 0.05 level.
Testing Functional Divergence among PIII and PII Sequences
To study the relationship between functional divergence and structural differences after the emergence of PII sequences from PIII sequences, the method of Gu (1999) implemented in version 1.04 of program DIVERGE (detecting variability in evolutionary rates among genes) (Gu and VanderVelden 2002) was applied. It was used to detect evolutionary rate pattern shifts coupled with the emergence of a new phylogenetic cluster and to identify which amino acid sites contributed most to this functional divergence. This method allows looking for significant changes in the rates of evolution after a split event (i.e., duplication, speciation) which results in the differentiation of two clearly defined monophyletic clusters in a phylogeny. A functional divergence coefficient (
) is calculated. Briefly, in a phylogeny with two monophyletic clusters, this coefficient summarizes the probability that a fraction of sites in the alignment belong to different evolutionary rate categories due to a different rate in the evolution of the two clusters. When all sites belong to the same rate category, there is no effect of the differential phylogenetic clustering and therefore equals 0. When the coefficient is significantly different from 0, as revealed by a likelihood ratio test (LRT), then there is evidence for functional divergence between the two clusters. In these cases, it is possible to identify, through a Bayesian approach, which sites are most likely to be evolving at different rates in the two clusters, establishing empirically a cutoff value of significance by progressively removing the most likely sites until reaching a equal to 0.
Positive Selection Tests
To examine the pattern of selection acting in the different disintegrin gene families at global and lineage-specific levels, several selection tests implemented in the PAML package (Yang 1997) were applied to the codon-based alignment of PIII and PII sequences. To detect sites under positive selection in all lineages, we applied two codon-based ML substitution models that are site-specific but assume the same selection pattern for a site in all lineages. At the genomic level, whether a particular amino acid within a gene is under accelerated evolution or remains functionally constrained can be detected by comparing the ratio of nonsynonymous nucleotide substitutions per nonsynonymous site (dN) with that of synonymous substitutions per synonymous sites (dS) (
= dN/dS) for each codon in the alignment. Codons evolving with > 1 are presumed to evolve under positive selection (functional diversification), whereas < 1 indicates that the codon evolves under the influence of purifying selection. Sites exhibiting values that do not differ significantly from 1 are thought to evolve neutrally. We performed a series of tests combining information from site-based and lineage-specific analyses in order to determine the most likely groups on which positive selection has been operating.
Estimates of dN/dS are sensitive to the assumed underlying mutation model (Yang et al. 2000). First, a pair of nested, site-based models (M7 and M8) were used as recommended (Yang et al. 2005). These models make no consideration about lineages and thus are suitable to detect sites under positive selection in all lineages. M7 and M8 are based on a continuous beta distribution whose shape can change depending on the estimated proportions of sites under neutral and purifying selection. Model M8 adds an extra category of sites to the M7 model allowing for sites with greater than one and hence evolving under positive selection. A significantly better fit to model M8 over M7 is indicative of positive selection acting on the corresponding gene. An LRT was used to compare the pair of nested models with the corresponding statistics being approximated by a chi-squared distribution. Posterior probabilities of the different site classes were calculated using a Bayes empirical Bayes (BEB) procedure (Yang et al. 2005) and those sites with an a posteriori probability higher than 0.95 of having an > 1 were considered to have evolved under positive selection. These tests were applied to the whole PIII- and PII-SVMP sequence data sets. The length of the amino acid sequences used varies from 100 (disintegrin-like domain of PIII-SVMP) to 50 residues (short disintegrins). Although longer sequences exhibit an increased probability of detecting adaptive evolution, the LRT power increases with sequence divergence, the use of a large number of sequences, and a strong (presumably) signal of positive selection (Anisimova et al. 2001).
To investigate the existence of sites evolving under positive selection in only a specific lineage, we applied the branch-site model described by Zhang et al. (2005) to the same alignment used for the site-based models. The model allowing for positive selection is denoted model A and the lineage to be tested is the foreground lineage, whereas the remaining ones are the background lineages. The model considers four site categories with 
1 for all lineages and two lineage-specific categories corresponding to those positions evolving under purifying or neutral selection in the background lineages but which are allowed to evolve under positive selection in the foreground lineage. This model A was compared with model A1, in which the same proportions of sites evolving under different conditions are maintained but the previous class of sites evolving with > 1 in the foreground lineage is now restricted to evolve with 0 < < 1 and = 1. An LRT was used to compare branch-site models A and A1 as described above. In those cases where the tests were significant, sites under positive selection were identified using a BEB procedure. We have carried out four different lineage-specific analyses specifying each time as foreground lineage one of the disintegrin gene families with a relevant number of sequences (PIII-disintegrin–like domains of SVMPs, PII-disintegrins, PII medium-sized disintegrins, and PII-dimeric disintegrins).
Two major factors can introduce confounding effects on the different positive selection tests. Occasionally, the same site was recovered as having evolved under positive selection both in site-based and in lineage-specific analyses. In those cases, the same site-based test was carried out after removing the lineage detected under positive selection thereby assuring that the original analysis was not influenced by lineage-specific selection. Also some detected sites presented gaps due to the shorter sequence size in some of the lineages. In these cases, the corresponding lineages were removed, and the same positive selection tests were carried out in order to corroborate original analyses. This approach allowed us to unequivocally determine groups bearing positive selection sites, as explained in the Results.
Ancestral Reconstruction of the Integrin-Binding Site
In order to analyze the evolution of the integrin-binding loop from PII-disintegrins, we used the phylogenetic topology to infer ancestral states by maximum parsimony and ML analyses of the positions involved using MESQUITE 2.0 (Maddison WP and Maddison DR 2007).
Mapping Residues Evolving under Positive Selection in the Three-Dimensional Structures of Disintegrins
To infer structure–function correlations, amino acid residues on which positive Darwinian selection has been operating were mapped onto the three-dimensional (3D) structures of the following proteins: disintegrin–like/cysteine-rich domains of the PIII-SVMP vascular apoptosis protein-2 (VAP2) from Crotalus atrox (2DW0) and PII-disintegrins bitistatin (long disintegrin from Bitis arietans) (modeled as in Calvete et al. [1997]), 1J2L (trimestatin, a medium-sized disintegrin from Trimeresurus flavoviridis), 1TEJ (heterodimeric disintegrin from Echis carinatus), and 1RO3 (short disintegrin from E. carinatus). Coordinates were retrieved from the Protein Data Bank addressing http://www.rcsb.org/pdb, and figures were generated using Phymol (http://phymol.sourceforge.net). For computing the proportion of conservative (Nc) and nonconservative (Nnc) amino acid replacements, amino acids were grouped according to their side chain physicochemical features in aromatic (Trp, Tyr, Phe), hydrophobic (Leu, Ile, Val, Met, Ala), turn forming or polar (Gly, Pro, Asn, Gln, Ser, Thr), positively charged (His, Arg, Lys), and negatively charged (Asp, Glu).
|
Results and Discussion |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResults and DiscussionConcluding Remarks and...Supplementary MaterialAcknowledgementsReferences |
|---|
Phylogenetic Analysis of Snake Venom Disintegrins
To study the molecular evolution of snake venom disintegrin genes, we carried out phylogenetic inference analyses based on amino acid sequence alignments and inferred their evolutionary history using ML and Bayesian methods. The ML topology was rooted with the PIII sequences, the ancestral family from which PII sequences are thought to have derived (Moura-Da-Silva et al. 1996
). The four resulting phylogenetic clades (fig. 2A) were congruent with the disintegrin classification based on their structural properties (fig. 1A). Only the PII medium disintegrin clade was paraphyletic though this might be due to the low reliability of the internal nodes. Two observations are worth noting: on the one hand, the phylogenetic history is compatible with previous evolutionary and structural studies indicating that the PII-short and PII-dimeric represent the latest diverging lineages of disintegrins (Calvete et al. 2003, 2005). On the other hand, the PII-dimeric disintegrin cluster shows long branches, an indication of fast-evolving lineage with a large number of structural changes accumulating on them.
|
Rapid Evolution of Dimeric and Short Disintegrins
Comparison of the rate of synonymous (KS) and nonsynonymous substitutions (KA) in duplicated genes provides information on the relative role of natural selection and genetic drift in the functional diversification of the disintegrin family. An excess of nonsynonymous over synonymous substitutions is a molecular indicator for adaptive evolution operating as the major driving force creating diversity at the amino acid level (Hughes 1994
). In a previous work (Calvete et al. 2008), we determined the strength and direction of selection from genomic DNA sequences of short and dimeric disintegrins from Echis ocellatus and Macrovipera lebetina transmediterranea. In almost, all pairwise nucleotide comparisons KA/KS values were >1, suggesting a rapid evolution between the genes coding for these disintegrin subfamilies. To provide further evidence for the hypothesis that the disintegrin family, like toxins from other venoms (Duda and Palumbi 1999; Kordis et al. 2002; Ohno et al. 2003), has evolved rapidly by adaptive evolution, pairwise nucleotide comparisons of available cDNA sequences coding for disintegrins and disintegrin-like domains (table 1) were performed. We found that for all the comparisons, the number of transitions were saturated and thus KS values could not be computed (data not shown). The analysis (table 2) revealed that the dimeric disintegrin lineage exhibits accelerated nonsynonymous substitutions with respect to medium-sized (P value = 0.026) and long disintegrins (P value = 0.029) and that short disintegrins are not accelerated with respect to the other disintegrin lineages.
|
Dimeric disintegrins exhibit the largest sequence diversity in their integrin-binding motifs (fig. 2B). Among the integrin inhibitory motifs, RGD blocks the β1 and β3 integrins
8β1, 5β1, vβ1, vβ3, and IIbβ3; MLD targets the 4β1, 4β7, 3β1, 6β1, 7β1, and 9β1 integrins; VGD and MGD impair the function of the 5β1 integrin; KGD inhibits the IIbβ3 integrin with a high degree of selectivity; WGD has been reported to be a potent inhibitor of the RGD-dependent integrins 5β1, vβ3, and IIbβ3; and KTS and RTS represent selective 1β1 inhibitors (Calvete 2005; Calvete et al. 2005; Sanz et al. 2006) (fig. 1D). Dimeric disintegrins are widely distributed in Echis and Vipera venoms (table 1), which are, in addition, rich sources of monomeric disintegrins. It is worth mentioning that non-RGD dimeric disintegrins are present in venoms that also contain RGD disintegrins (Calvete et al. 2003), usually other dimeric or short disintegrins. The coexistence in the same snake species of disintegrins with a conserved RGD motif and disintegrins with variable non-RGD sequences (orthologs) support the hypothesis that, following gene duplication, one copy of the gene (i.e., that coding for an RGD disintegrin) evolved divergently under pressure dictated by an ancestral function. The duplicated gene, freed from functional constraints, would accumulate selectively neutral mutations until by chance a series of mutations confer a new function, such as inhibition of non-RGD–dependent integrin receptors, and become eventually fixed by random genetic drift (Ohno 1973). Venoms from genus Echis often possess the combination of short RGD-bearing and dimeric RGD and/or non-RGD disintegrins, most probably reflecting the evolutionary emergence of short disintegrins from a short-coding RGD dimeric disintegrin precursor by two nucleotide mutations (Juárez, Wagstaff, Sanz, et al. 2006). Moreover, although a detailed analysis of more venoms from snakes representing different geographically and ecologically taxonomic groups is required, the fact that dimeric disintegrins are widely distributed in Viperinae and Crotalinae, whereas short disintegrins appear to be restricted to African and Asian Echis and Eristicophis species indicates that emergence of dimeric disintegrins represents an early evolutionary event predating the Viperinae–Crotalinae split, whereas short disintegrins have evolved much more recently, that is, after radiation of Viperinae. According to Lenk et al. (2001), viperines originated in the Oligocene (35–22 MYA) in Africa and successively underwent a first radiation leading to the five basal groups, Bitis, Cerastes, Echis, the Atherini, and the Eurasian viperines.
The topology of the phylogenetic tree does not parallel that of the species tree (fig. 2C). Clades consist of sequences from different taxa. This evolutionary pattern is consistent with that expected for a multigene family whose members are undergoing a birth-and-death process in which the appearance and disappearance of particular loci are being driven by selection (Nei et al. 1997). Indeed, the number of cases to which this model applies is rapidly increasing (Fry et al. 2003; Nei and Rooney 2005).
Functional Divergence between PIII and PII Sequences
Functional divergence may occur through a variety of evolutionary processes, including relaxed selective constraints, neutral evolution, and positive selection (reviewed by Fay and Wu 2003; Nei 2005). We have explored the possibility and sign of the functional divergence within the disintegrin gene family. Unlike PII-disintegrins, whose biological activity as integrin receptor antagonists is firmly established, the functionality of PIII-disintegrin–like domains remains controversial (reviewed in Calvete et al. 2005). From the crystal structure of PIII-SVMP VAP1 from C. atrox (Takeda et al. 2006), it was clearly observed that the 68ECD70 sequence motif (numbering as in fig. 3), which had been suggested to be involved in integrin 2β1 binding by this domain (reviewed by Lu et al. 2007), is sterically unavailable for interacting with the receptor. Our working hypothesis was that the emergence of the integrin inhibitory activity of PII-disintegrins could have been coupled to a shift in their evolutionary rates with respect to their PIII homologs, which might be characterized by a stronger pressure for sequence conservation, whereas the structure of the new PII family would change more freely. The coefficient of functional divergence estimated for the PIII/PII comparison was found to be statistically significant ( = 0.3088, LRT = 19.5097, P value < 0.01). This result supported our initial hypothesis that functional changes had occurred in PII sequences, perhaps coupled with their progressive reduction in length (fig. 1C).
|
To identify which sites were critical for this functional divergence, a posteriori probabilities of
for each site were computed. To obtain a cutoff value, helping us to decide which probabilities were significant, sites with the highest probability value were progressively removed from the alignment until the removal of one further site yielded a nonsignificant value. In this case, the cutoff value was 0.91. The analysis revealed that amino acids at only two positions (numbered 58 and 70 in the alignment displayed in fig. 3) accounted for the emergence of functional divergence between the PIII-disintegrin–like domains and proper PII-disintegrins. These amino acids are characterized by being extraordinarily conserved in the PIII family, whereas they are highly variable in PII-disintegrins. Position 58 is invariably occupied by Ala or Ser residue in all known PIII-disintegrin–like domains (fig. 4). Alanine and serine are also found in the corresponding position in some PII-disintegrins, whereas in other PII-disintegrins, this position is occupied by different amino acids such as Lys, Glu, Pro, and Val, and no subfamily pattern is evident, indicating that after divergence this position has evolved under type II functional divergence (Gu 2001), that is, accumulating amino acid replacements at random. Indeed, position 58 is located far away from the active tripeptide motif of PII-disintegrins (fig. 4) to be of functional relevance. Moreover, Ala/Ser-58 is surface exposed and thus its structural role remains elusive. On the other hand, conservation of Asp70 in the disintegrin-like domains of PIII-SVMPs indicates an important action of purifying selection at this site. Asp70 is adjacent to Cys69, which forms the PIII-SVMP–specific disulfide bond with Cys94 (Calvete et al. 2000). It has been proposed that deletions of both gene regions coding for the cysteine-rich domain and the Cys69–Cys94 bond represented key events along the structural diversification pathway of disintegrins (Calvete et al. 2003; Juárez, Wagstaff, Oliver, et al. 2006) (see also fig. 1B). Hence, removal of the structural constraint imposed by the Cys69–Cys94 linkage may have paved the way for the emergence and subsequent evolution of the integrin-binding loop of PII-disintegrins (fig. 4). Disintegrins have evolved a restricted panel of integrin blocking sequences that adapted to the ligand-binding architecture of their target integrin receptors (Sanz et al. 2006). The crystal structure of the extracellular segment of integrin vβ3 in complex with an RGD ligand (Xiong et al. 2002) showed that the peptide fits into a crevice between the V propeller and the β3 A-domain. The conserved aspartate residue (position 68 in fig. 2) is responsible for the binding of disintegrins to the integrin receptor β subunit, whereas the two other residues of the integrin-binding motif at positions 66 and 67 (RG, KG, MG, WG, ML, VG) dictate the integrin specificity through interaction with the subunit.
|
Position 70 is highly polymorphic and displays branch specificity among PII-disintegrins (fig. 3). This finding is in line with previous reports that the inhibitory effect of disintegrins is finely tuned by residues flanking the active tripeptide. A number of studies support a functional role for the amino acids flanking the integrin-recognition tripeptide in determining the receptor-binding characteristics (Niewiarowski et al. 1994
; Lu et al. 1996; McLane et al. 1996). Disintegrins with RGD motifs C-terminally flanked by an aromatic residue (tryptophan [W] or phenylalanine [F], i.e., RGD[W/F]) and a polar residue (asparagine [N] or aspartic acid [D], RGD[N/D]) are quite selective in inhibiting native ligand binding to integrins IIbβ3 and vβ3, respectively. Molecular modeling and nuclear magnetic resonance structure determination of cyclic RGD peptides (Pfaff et al. 1994) showed the importance of the amino acid immediately C-terminal to the RGD sequence in determining the conformation of the RGD loop. The width and shape of the integrin-binding loop represents an important structural feature that modulates integrin-binding selectivity, and the distance between the Cβ atoms of Arg66 and Asp68 (numbering as in fig. 3A) distinctly affects the fitting of disintegrins in the binding pocket of integrins IIbβ3 and vβ3, which share the β3 subunit: the optimum distance is in the range of 7.5–8.5 Å for IIbβ3 and at or below 6.7 Å for vβ3 (and 5β1) (Pfaff et al. 1994). Furthermore, the occurrence of convergent amino acid Asp70/Asn changes in different clades of the PII topology (fig. 2) highlights the narrow link between changes at the structural level and adaptive evolution at the nucleotide level.
Positive Selection in the Evolution of Disintegrin Gene Families
To investigate the impact of positive selection during the evolutionary history of disintegrin lineages, we examined the nonsynonymous-to-synonymous substitution rates ( = dN/dS). This ratio is an indication of the presence and type of selection. Site-based models for the distribution of values were used in order to detect which fraction, if any, of the codons in the gene have been evolving under positive selection pressure. A pair of nested models (M7–M8) were applied to the complete PIII and PII codon-based alignments. The test was highly significant revealing the existence of a relevant fraction of sites evolving under > 1. A Bayesian approach allowed us to identify which sites are most likely evolving under positive selection (fig. 3B; for a full alignment and site-by-site distribution of values, see supplementary figure [Supplementary Material online]. This information is also available from the authors upon request). Globally, 13 sites were detected to have evolved under positive selection by this approach: 14, 24, 39, 45, 48, 56, 57, 64, 66, 71, 74, 86, and 98. These residues are shown in figure 3A encased by hollow boxes, and figure 4 displays their location in the 3D structures of PIII- and PII-disintegrins. Except for residues at positions 45, 48, and 74, which are buried to a high degree in the disintegrin domain structure, the positively selected sites exhibit surface-exposed side chains (fig. 4F), indicating that hot spots of amino acid substitutions occur mostly on the surface of the disintegrin domain. This is likely due to their role as "pharmacological sites," which play a key role in protein–protein interactions between venom proteins and prey tissues.
A substantial number of disintegrin residues (14–18%) are subjected to diversifying selection. Four of these sites (64, 66, 71, and 74) are located within the integrin-binging loop, and residue at position 86 is located in the C-terminal tail (fig. 4). These two regions display concerted motions and form a conformational functional epitope that is engaged in extensive interactions with target integrin receptors (Monleón et al. 2005). The functional role of the other sites remains elusive, and hence the real reason for their accelerated evolution is unclear. However, it should be mentioned that whereas sites 14, 24, 39, 45, 48, 56, and 57 accumulate overwhelmingly conservative amino acid replacements (type I functional divergence, Gu 2001), in both medium-sized and dimeric disintegrins, the type of divergence at the functional epitope sites 64, 66, 71, 74, and 86 depart between these two disintegrin groups. Hence, in medium-sized disintegrins sites 64 and 66 exhibit similar number of conservative (Nc) and nonconservative (Nnc) amino acid replacements, whereas Nnc > Nc at sites 71, 74, and 86. On the other hand, among dimeric disintegrins, except for site 66 where Nnc 
Nc, all these positions accumulate conservative substitutions. This implies a stronger pressure to preserve the qualitative property of the amino acid residue side chains within the disintegrin-binding loop of dimeric than of medium-sized disintegrins.
The integrin inhibitory activity of disintegrins depends on the appropriate pairing of cysteines, which determine the conformation of the inhibitory loop (Niewiarowski et al. 1994). As expected, invariant cysteine residues are negatively selected (0.06 < < 0.4). Non-cysteine residues subjected to strong negative selection ( < 0.135) include the absolutely conserved residues Glu13, Asp19, Gly47, Gly77, and Pro83. Other highly conserved positions are Asp/Asn32, Ala/Pro33, Ser/Thr35 (0.25 < < 0.35), Ile/Phe55 ( = 0.47), and Asp/Glu73 ( = 0.41). Except for cysteines 18 and 36 and residues at positions 19, 32, 33, and 35, negatively selected sites comprise protein surface inaccessible residues (fig. 4F) that may be under strong purifying selection due to their structural contribution in maintaining the biologically active conformation of the disintegrin domain.
Because different combinations of disintegrins coexist in different snake venoms, it seems reasonable to postulate that adaptive evolution processes involving different sites and lineages took place along the structural diversification pathways of disintegrins. To assess this point, we carried out branch-site tests of positive selection for the following four groups of sequences: 1) PIII-SVMPs, 2) PII-disintegrins, 3) PII-dimeric, and 4) PII medium disintegrins. These tests allowed us to investigate the possible occurrence of a dN/dS () > 1 in a specific lineage with respect to the remaining lineages and, if this were the case, which sites were involved in this adaptive process. All four tests were highly significant at the 0.05 level after Bonferroni correction (P value < 0.0125), indicating that each group underwent a process of independent adaptive evolution after divergence. Furthermore, in each group, different sites appear to be involved in the process of adaptive evolution driven by positive selection, including the surface-exposed residues at positions 16, 21, 70, 85, and 87 in medium-sized disintegrins and sites 63 and 65 in dimeric disintegrins (figs. 3 and 4). Except for the residue at position 70, which is highly variable (discussed above), all these sites exhibit Nc > > > Nnc. No lineage-specific site was shared with the other lineages, thus indicating that positive selection in each lineage has followed different evolutionary routes in response to different selective conditions, although preserving in more than 95% of the cases the physicochemical character of the amino acid side chains.
Amino acids at positions 39, 57, 74, and 98 of PIII-SVMPs, 64 of PII medium disintegrin sequences, and 66 of PII-dimeric disintegrin sequences were identified to be under positive selection by both the sites-only and the branch-site approaches. To test whether detection of these residues in the whole data set was an artifact due to the influence of lineage-specific selection on the final result, the positive selection test was carried out after removing the corresponding lineages as explained in Materials and Methods, and thus the confounding factor of lineage-specific selection, from the analysis. Applying this restriction, all these sites were still detected as evolving under positive selection (data not shown). This test was useful not only to discard possible artifacts but also for establishing a temporal frame for different positive selection events. Those amino acid positions detected by the sites-only models (hollow columns in fig. 3B) are under positive selection in all lineages, thus indicating that their neofunctionalization potential has remained intact during the whole evolutionary history of disintegrins. Lineage-specific adaptations that appeared after structural diversification splits represent more recent events of positive selection that allowed the different members of the disintegrin subfamilies to refine their interactions with the corresponding integrin target. At this respect, it is worth noting that positions 63 and 66, located within the integrin-binding loop of dimeric disintegrins (fig. 4D), have evolved under positive selection. Dimeric disintegrins exhibit the largest sequence diversity in their integrin-binding motifs (fig. 2B). The analysis of nonsynonymous nucleotide substitution rates clearly indicated an accelerated evolution of this lineage with respect to medium and long disintegrins, and this conclusion is further supported by the KA/Ks values higher than 1 (Calvete et al. 2008). As a whole, these data support the birth-and-death model of protein evolution (Hughes 1994, 2000; Fry et al. 2003).
The results and conclusions outlined above are in line with the evolutionary scheme displayed in figure 1 and supported by phylogenetic analysis in conjunction with biochemical and genetic data. According to this evolutionary model, structural and functional diversification within the disintegrin family occurred through the successive losses of the cysteine-rich domain and lineage-specific disulfide bonds, followed by the emergence of the integrin inhibitory motifs under natural selection favoring specific changes at the amino acid level (Hughes 2000; Fry et al. 2003).
Reconstructing the Evolution of the Integrin-Binding Site
The evolution of the integrin-binding motif of PII-disintegrins was inferred by parsimony and likelihood analyses of positions 66–68 (fig. 3A) using the ML phylogenetic tree displayed in figure 2. Both approaches yielded the same results indicating that RGD represents the ancestral integrin-recognition domain, which emerged from the subgroup of PIII-SVMPs bearing the 66RDECD70 sequence after deletion of the PIII-lineage–specific Cys69. Conversion of 66RDE68 into 66RGD68 can be accomplished with a minimum of two mutations: GAT or GAC for GGT or GGC and GAA or GAG for GAT or GAC. Except for the KTS/RTS short disintegrins (Calvete, Marcinkiewicz, and Sanz 2007), whose evolution departs from the canonical pathway shown in figure 1B, the most parsimonious nucleotide substitution events required for the emergence of all known disintegrin's integrin-recognition motifs (fig. 1D) from an ancestral RGD sequence involves a minimum of three mutations (fig. 5). The PII-dimeric disintegrin branch shown in figure 2B is paraphyletic and does not parallel the phylogenetic tree (fig. 2C), thus precluding a reliable evolutionary pathway to be resolved with the available sequence data. Complete snake genome sequences are required to analyze the phylogenetic relationships among the multigene disintegrin family members in single species and between species within monophyletic taxonomic clades.
|
|
Concluding Remarks and Hypotheses |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResults and DiscussionConcluding Remarks and...Supplementary MaterialAcknowledgementsReferences |
|---|
Snake venom toxins likely evolved from endogenous proteins with normal physiological functions that were recruited into the venom proteome before the radiation of the advanced snakes (Fry and Wüster 2004
; Fry 2005; Fry et al. 2006). PII-disintegrins represent integrin inhibitors, which evolved by neofunctionalization of disintegrin-like domains of duplicated PIII-SVMP genes. The fact that PIII-SVMPs are widely distributed in the five families of Colubroidea, Viperidae, Elapidae, Atractaspididae, and Colubridae, whereas PII-disintegrins have been only found in venoms of vipers and rattlesnakes (Viperidae), strongly suggests that disintegrins emerged after the split of Viperidae and Elapidae but before the separation of the Viperidae subfamilies, Viperinae and Crotalinae, about 12–20 MYA (Campbell and Lamar 2004). Structural and functional diversification of disintegrins may have occurred early in the evolution of Viperidae because basal genera like Echis and Bitis contain molecules from different disintegrin classes, which exhibit a broad spectrum of integrin-binding motifs (fig. 2C). On the other hand, disintegrin subfamilies segregate distinctly in African and Eurasian Viperinae, expressing mainly dimeric and short disintegrins, and in Asian and New World Crotalinae, in whose venoms large and medium-sized disintegrins are predominantly found (fig. 2C). These distinct geographic distributions of lineages points to biogeographical factors (population expansion, population bottlenecks, vicariance, and migration) as historical forces shaping disintegrin evolution. Though the inventory of disintegrins among Viperidae genera is still fragmentary, the pattern of disintegrin distribution strongly suggests that the emergence of dimeric (and subsequently of short) disintegrins seems to have occurred twice independently, as a major evolutionary route in Viperinae, after its split from Crotalinae, and occasionally in genus Agkistrodon (Crotalinae). The structural and functional complexity of disintegrins contrasts with their small molecular size. The large proportion of residues under positive selection in all lineages, and the finding that in addition each subfamily exhibits lineage-specific sites (figs. 3 and 4), highlights that the rate of amino acid substitutions, and hence the neofunctionalization potential, of disintegrins has increased during evolution. Currently, 18% of all amino acid residues of long disintegrins are under diversifying selection, and this figure increases to 23%, 21%, and 26% in medium-sized, dimeric, and short disintegrins, respectively.
The accelerated evolution of disintegrins might be linked to adaptation to the environment, including feeding habits. The currently recognized biological activity of PII-disintegrins is steric inhibition of the adhesive interactions of most β1 and β3 integrin receptors (fig. 1D). Blockage of the platelet's fibrinogen receptor, integrin IIbβ3, resulting in persistent bleeding clearly represents an evolutionary advantage for prey capture and thus may increase the predator population's fitness. It is therefore not surprising that RGD represents the ancestral integrin inhibitory motif and that some disintegrins have evolved KGD and WGD motifs, which confer enhanced specificity and increased potency toward integrin IIbβ3 (Scarborough et al. 1993; Calvete et al. 2002). However, the advantage of the emergence of other tripeptide motifs targeting the β1 integrin receptors seems to be difficult to rationalize in the context of a predator–prey arms race. β1 integrins bind to extracellular matrix molecules (Brakebusch and Fässler 2005). Per instance, integrins 1β1, 2β1, 10β1, and 11β1 are collagen receptors. Integrins 1β1, 2β1, 3β1, 6β1, and 7β1 bind to laminin. Integrins 4β1, 5β1, 8β1, and Vβ1 are fibronectin receptors and 9β1 and Vβ1 bind, respectively, to tenascin C and vitronectin. Some β1 integrins also interact with cellular receptors: 4β1 and 9β1 bind to VCAM-1 and 4β1 to MadCAM-1. Adhesion of cells to the surrounding extracellular matrix and to other cells is essential for tissue integrity. In addition to this structural function, cell adhesion induces intracellular signaling mechanisms that regulate proliferation, apoptosis, cell polarity, and differentiation. None of these processes appears to be associated with defense mechanisms against an external aggression, like a snakebite, that requires an immediate response for animal survival. It is tempting to hypothesize that inhibition of 1 integrins could be a consequence of the great neofunctionalization potential of a protein scaffold requiring weak functional constraints and evolving under accelerated evolution. Perhaps, this feature of the disintegrin domain underlies its recruitment into the venom proteome followed by its successful transformation into a toxin. Supporting this view, disintegrins, like many venom toxins from other vertebrates and invertebrates, are small protein domains bearing a compact and highly disulphide-cross-linked core from which functional loops protrude, and increasing evidence indicates that these venom toxins belong to rapidly evolving gene families under strong positive selection (Vita et al. 1995; Ménez 2002; Fry et al. 2003; Zupinski et al. 2003; Ferrat and Darbon 2005; Ogawa et al. 2005; Olivera 2006; Lynch 2007; Fry et al. 2008; Gibbs and Rossiter 2008).
The rate of molecular evolution is a function of the rate of neutral, deleterious, and advantageous mutations, their selection coefficients, and the effective population size. A major tenet of the neutral mutation–random drift theory (proposed independently by Kimura [1968] and King and Jukes [1969]) is that functionally important sites will remain constrained over time, whereas neutral sites will evolve at a much faster rate. Conserved cysteines and buried residues appear to be under strong purifying selection due to their role in maintaining the active conformation of disintegrins, whereas residues under diversifying selection are scattered throughout the whole disintegrin domain surface. Though mutations provide the ground on which natural selection operates to create functional innovations, accumulation of amino acid substitutions per se does not necessarily lead to functional divergence of paralogous genes. Positively selected residues located within currently nonfunctional regions may represent neutral polymorphisms with respect to fitness and function and may therefore be regarded as "orphan sites searching for a function." Alternatively, disintegrins may possess unrecognized biological activities. Research on disintegrins is relevant for understanding the biology of viper venom toxins, but it also provides information about new structural determinants involved in integrin recognition that may be useful in both, basic and clinical research. Clearly, determination of the mechanism and significance of the changes driven by positive selection, which is required for unraveling structure–function correlations of disintegrins, remains a challenge for future investigations.
|
Supplementary Material |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResults and DiscussionConcluding Remarks and...Supplementary MaterialAcknowledgementsReferences |
|---|
Supplementary figure is available at Molecular Biology and Evolution online (http://www.mbe.oxfordjournals.org/).
|
Acknowledgements |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResults and DiscussionConcluding Remarks and...Supplementary MaterialAcknowledgementsReferences |
|---|
This study has been financed by grants from the Ministerio de Educación y Ciencia (BFU2004-01432/BMC) and Ministerio de Ciencia e Innovación (BFU2007-61563), Madrid, Spain. Mariano Polo (Instituto de Biomedicina de Valencia, Valencia) is gratefully acknowledged for helping with the 3D models.
|
Footnotes |
|---|
1 These authors contributed equally to this work and may both be considered first authors.
Marcy Uyenoyama, Associate Editor
|
References |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResults and DiscussionConcluding Remarks and...Supplementary MaterialAcknowledgementsReferences |
|---|
Angulo Y, Escolano J, Lomonte B, Gutiérrez JM, Sanz L, Calvete JJ. Snake venomics of Central American pitvipers: clues for rationalizing the distinct envenomation profiles of Atropoides nummifer and Atropoides picadoi. J Proteome Res (2008) 7:708–719.[CrossRef][Web of Science][Medline]
Anisimova M, Bielawski JP, Yang Z. Accuracy and power of the likelihood ratio test in detecting adaptive molecular evolution. J Mol Evol (2001) 18:1585–1592.
Bazaa A, Juárez P, Marrakchi N, Lasfer ZB, El Ayeb M, Harrison RA, Calvete JJ, Sanz L. Loss of introns along the evolutionary diversification pathway of snake venom disintegrins evidenced by sequence analysis of genomic DNA from Macrovipera lebetina transmediterranea and Echis ocellatus. J Mol Evol (2007) 64:261–271.[CrossRef][Web of Science][Medline]
Brakebusch C, Fässler R. β1 integrin function in vivo: adhesion, migration and more. Cancer Metastasis Rev (2005) 24:403–411.[CrossRef][Web of Science][Medline]
Calvete JJ. Structure-function correlations of snake venom disintegrins. Curr. Pharm. Des (2005) 11:829–835.[CrossRef][Web of Science][Medline]
Calvete JJ, Fox JW, Agelan A, Niewiarowski S, Marcinkiewicz C. The presence of the WGD motif in CC8 heterodimeric disintegrin increases its inhibitory effect on IIbβ3, Vβ3, and 5β1 integrins. Biochemistry (2002) 41:2014–2021.[CrossRef][Web of Science][Medline]
Calvete JJ, Juárez P, Sanz L. Snake venomics. Strategy and applications. J Mass Spectrom (2007) 42:1405–1414.[CrossRef][Web of Science][Medline]
Calvete JJ, Juárez P, Sanz L, Forthcoming. Snake venomics and disintegrins. Portrait and evolution of a family of snake venom integrin antagonists. In: Reptile venom and toxins—Mackessy SP, ed. (2008) Francis & Taylor.
Calvete JJ, Marcinkiewicz C, Monleón D, Esteve V, Celda B, Juárez P, Sanz L. Snake venom disintegrins: evolution of structure and function. Toxicon (2005) 45:1063–1074.[Medline]
Calvete JJ, Marcinkiewicz C, Sanz L. KTS and RTS-disintegrins: anti-angiogenic viper venom peptides specifically targeting the 1β1 integrin. Curr Pharm Des (2007) 13:2853–2859.[CrossRef][Web of Science][Medline]
Calvete JJ, Moreno-Murciano MP, Sanz L, Jürgens M, Schrader M, Raida M, Benjamin DC, Fox JW. The disulfide bond pattern of catrocollastatin C, a disintegrin-like/cysteine-rich protein isolated from Crotalus atrox venom. Protein Sci (2000) 9:1365–1367.[Web of Science][Medline]
Calvete JJ, Moreno-Murciano MP, Theakston RDG, Kisiel DG, Marcinkiewicz C. Snake venom disintegrins: novel dimeric disintegrins and structural diversification by disulphide bond engineering. Biochem J (2003) 372:725–734.[CrossRef][Web of Science][Medline]
Calvete JJ, Schrader M, Raida M, McLane MA, Romero A, Niewiarowski S. The disulphide bond pattern of bitistatin, a disintegrin isolated from the venom of the viper Bitis arietans. FEBS Lett (1997) 416:197–202.[CrossRef][Web of Science][Medline]
Campbell JA, Lamar WW. The venomous reptiles of the western hemisphere (2004) Ithaca (NY): Cornell University Press.
Castoe TA, Parkinson CL. Bayesian mixed models and the phylogeny of pitvipers (Viperidae: serpentes). Mol Phylogenet Evol (2006) 39:91–110.[CrossRef][Web of Science][Medline]
Daltry JC, Wüster W, Thorpe RS. Diet and snake venom evolution. Nature (1996) 379:537–540.[CrossRef][Web of Science][Medline]
Duda TF Jr, Palumbi SR. Developmental shifts and species selection in gastropods. Proc. Natl. Acad. Sci. USA (1999) 96:6820–6823.
Fay JC, Wu C-I. Sequence divergence, functional constraint, and selection in protein evolution. Annu Rev Genomics Hum Genet (2003) 4:213–235.[CrossRef][Web of Science][Medline]
Ferrat G, Darbon H. An overview of the three dimensional structure of short spider toxins. Toxin Rev (2005) 24:361–383.[CrossRef]
Fox JW, Serrano SMT. Special issue: snake toxins and hemostasis. Toxicon (2005a) 45:951–1181.[Medline]
Fox JW, Serrano SMT. Structural considerations of the snake venom metalloproteinases, key members of the M12 reprolysin family of metalloproteinases. Toxicon (2005b) 45:969–985.[Medline]
Fry BG. From genome to "venome": molecular origin and evolution of the snake venom proteome inferred from phylogenetic analysis of toxin sequences and related body proteins. Genome Res (2005) 15:403–420.
Fry BG, Scheib H, van der Weerd L, Young B, McNaughtan J, Ramjan SF, Vidal N, Poelmann RE, Normam JA. Evolution of an arsenal: structural and functional diversification of the venom system in the advanced snakes (Caenophidia). Mol Cell Proteomics (2008) 7:215–246.
Fry BG, Vidal N, Norman JA, et al, (14 co-authors). Early evolution of the venom system in lizards and snakes. Nature (2006) 439:584–588.[CrossRef][Web of Science][Medline]
Fry BG, Wüster W. Assembling an arsenal: origin and evolution of the snake venom proteome inferred from phylogenetic analysis of toxin sequences. Mol Biol Evol (2004) 21:870–883.
Fry BG, Wüster W, Kini RM, Brusic V, Khan A, Venkataraman D, Rooney AP. Molecular evolution and phylogeny of elapid snake venom three-finger toxins. J Mol Evol (2003) 57:110–129.[CrossRef][Web of Science][Medline]
Gibbs HL, Rossiter W. Rapid evolution by positive selection and gene gain and loss: PLA2 venom genes in closely related Sistrurus rattlesnakes with divergent diets. J Mol Evol (2008) 66:151–166.[CrossRef][Web of Science][Medline]
Glassey B, Civetta A. Positive selection at reproductive ADAM genes with potential intercellular binding activity. Mol Biol Evol (2004) 21:851–859.
Greene HW. Dietary correlates of the origin and radiation of snakes. Am Zool (1983) 23:431–441.[Web of Science]
Gu X. Statistical methods for testing functional divergence after gene duplication. Mol Biol Evol (1999) 16:1664–1674.[Abstract]
Gu X. Maximum-likelihood approach for gene family evolution under functional divergence. Mol Biol Evol (2001) 18:453–464.
Gu X, VanderVelden K. DIVERGE: phylogeny-based analysis for functional-structural divergence of a protein family. Bioinformatics (2002) 18:500–501.
Guindon S, Gascuel O. A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Syst Biol (2003) 52:696–704.
Hughes AL. The evolution of functionally novel proteins after gene duplication. Proc Biol Sci (1994) B256:119–124.
Hughes AL. Adaptive evolution of genes and genomes (2000) New York: Oxford University Press.
Jackson K. The evolution of venom-conducting fangs: insights from developmental biology. Toxicon (2007) 49:975–981.[Medline]
Juárez P, Wagstaff SC, Oliver J, Sanz L, Harrison RA, Calvete JJ. Molecular cloning of disintegrin-like transcript BA-5A from a Bitis arietans venom gland cDNA library: a putative intermediate in the evolution of the long chain disintegrin bitistatin. J Mol Evol (2006) 63:142–152.[CrossRef][Web of Science][Medline]
Juárez P, Wagstaff SC, Sanz L, Harrison RA, Calvete JJ. Molecular cloning of Echis ocellatus disintegrins reveals non-venom-secreted proteins and a pathway for the evolution of ocellatusin. J Mol Evol (2006) 63:183–193.[CrossRef][Web of Science][Medline]
Kimura M. Evolutionary rate at the molecular level. Nature (1968) 217:624–626.[CrossRef][Web of Science][Medline]
King J, Jukes T. Non-Darwinian evolution. Science (1969) 164:788–798.
Kini R, Evans HJ. Structural domains in venom proteins: evidence that metalloproteinases and nonenzymatic platelet aggregation inhibitors (disintegrins) from snake venoms are derived by proteolysis from a common precursor. Toxicon (1992) 30:265–293.[Medline]
Kochva E. The origin of snakes and evolution of the venom apparatus. Toxicon (1987) 25:65–106.[Medline]
Kordis D, Krizaj I, Gubensek F. Functional diversification of animal toxins by adaptative evolution. In: Perspectives in molecular toxinology—Ménez A, ed. (2002) Chichester (UK): John Wiley & Sons Ltd. 401–419.
Lenk P, Kalyabina S, Wink M, Joger U. Evolutionary relationships among the true vipers (Reptilia: Viperidae) inferred from mitochondrial DNA sequences. Mol Phylogenet Evol (2001) 19:94–104.[CrossRef][Web of Science][Medline]
Lu X, Lu D, Scully MF, Kakkar VV. Structure-activity relationship studies on ADAM protein-integrin interactions. Cardiovasc Hematol Agents Med Chem (2007) 5:29–42.[Medline]
Lu X, Rahman S, Kakkar VV, Authi KS. Substitutions of proline 42 to alanine and methionine 46 to asparagine around the RGD domain of the neurotoxin dendroaspin alter its preferential antagonism to that resembling the disintegrin elegantin. J Biol Chem (1996) 271:289–294.
Lynch WJ. Inventing an arsenal: adaptative evolution and neofunctionalization of snake venom phospholipase A2 genes. BMC Evol Biol (2007) 7:1–14.[CrossRef][Medline]
Maddison WP, Maddison DR. Mesquite: a modular system for evolutionary analysis. Version 2 0 1 [Internet] (2007) Available from: http://mesquiteproject.org.
Marcinkiewicz C. Pharmacological impact of snake venom compounds. Curr Pharm Des (2007) 13:2851–2922.[CrossRef][Web of Science][Medline]
Markland FS. Snake venoms and the hemostatic system. Toxicon (1998) 36:1749–1800.[Medline]
McLane MA, Vijay-Kumar S, Marcinkiewicz C, Calvete JJ, Niewiarowski S. Importance of the structure of the RGD-containing loop in the disintegrins echistatin and eristostatin for recognition of IIbβ3 and Vβ3 integrins. FEBS Lett (1996) 291:139–143.[CrossRef]
Ménez A. Perspectives in molecular toxinology (2002) Chichester (UK): John Wiley & Sons.
Monleón D, Esteve V, Kovacs H, Calvete JJ, Celda B. Conformation and concerted dynamics of the integrin-binding site and the C-terminal region of echistatin revealed by homonuclear NMR. Biochem J (2005) 387:57–66.[CrossRef][Web of Science][Medline]
Monleón D, Moreno-Murciano MP, Kovacs H, Marcinkiewicz C, Calvete JJ, Celda B. Concerted motions of the integrin-binding loop and the C-terminal tail of the non-RGD disintegrin obtustatin. J Biol Chem (2003) 278:45570–45576.
Moura-Da-Silva AM, Theakston RDG, Crampton JM. Evolution of disintegrin cysteine-rich and mammalian matrix-degrading metalloproteinases: gene duplication and divergence of a common ancestor rather than convergent evolution. J Mol Evol (1996) 43:263–269.[Web of Science][Medline]
Nei M. Selectionism and neutralism in molecular evolution. Mol Biol Evol (2005) 22:2318–2342.
Nei M, Gu X, Sitnikova T. Evolution by the birth-and-death process in multigene families of the vertebrate immune system. Proc Natl Acad Sci USA (1997) 94:7799–7806.
Nei M, Rooney AP. Concerted and birth-and-death evolution of multigene families. Annu Rev Genet (2005) 39:121–152.[CrossRef][Web of Science][Medline]
Niewiarowski S, Marcinkiewicz C, Wierzbicka-Patynowski I, McLane MA, Calvete JJ. Structure and function of disintegrins and C-lectins: viper venom proteins modulating cell adhesion. In: Perspectives in Molecular Toxinology—Ménez A, ed. (2002) New York: John Wiley & Sons. 327–340.
Niewiarowski S, McLane MA, Kloczewiak M, Stewart GJ. Disintegrins and other naturally occurring antagonists of platelet fibrinogen receptors. Sem. Hematol (1994) 31:289–300.[Web of Science][Medline]
Ogawa T, Chijiwa T, Oda-Ueda N, Ohno M. Molecular diversity and accelerated evolution of C-type lectin-like proteins from snake venoms. Toxicon (2005) 45:1–14.[Medline]
Ohno S. Evolution by gene duplication (1970) Berlin (Germany): Springer Verlag.
Ohno S. Ancient linkage groups from frozen accidents. Nature (1973) 244:259–262.[CrossRef][Web of Science][Medline]
Okuda D, Koike H, Morita T. A new gene structure of the disintegrin family: a subunit of dimeric disintegrin has a short coding region. Biochemistry (2002) 41:14248–14254.[CrossRef][Web of Science][Medline]
Olivera BM. Conus peptides: biodiversity-based discovery and exogenomics. J Biol Chem (2006) 281:31173–31177.
Pfaff M, Tangemann K, Müller B, Gurrath M, Müller G, Kessler H, Timpl R, Engel R. Selective recognition of cyclic RGD peptides of NMR defined conformation by IIbβ3, Vβ3, and 5β1 integrins. J Biol Chem (1994) 269:20233–20238.
Rambaut A, Drummond AJ. Tracer v1.4 [Internet] (2007) Available from: http://beast.bio.ed.ac.uk/Tracer.
Richman A. Evolution of balanced genetic polymorphism. Mol Ecol (2000) 9:1953–1963.[CrossRef][Medline]
Robinson-Rechavi M, Huchon D. RRTree: relative-rate tests between groups of sequences on a phylogenetic tree. Bioinformatics (2000) 16:296–297.
Ronquist F, Huelsenbeck JP. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics (2003) 19:1572–1574.
Sanz L, Bazaa A, Marrakchi N, Pérez A, Chenik M, Bel Lasfer Z, El Ayeb M, Calvete JJ. Molecular cloning of disintegrins from Cerastes vipera and Macrovipera lebetina transmediterranea venom gland cDNA libraries: insight into the evolution of the snake venom integrin-inhibition system. Biochem J (2006) 395:385–392.[CrossRef][Web of Science][Medline]
Scarborough RM, Naughton MA, Teng W, Rose RW, Phillips DR, Nannizzi L, Arfsten A, Campbell AM, Charo IF. Design of potent and specific integrin antagonists. Peptide antagonists with high specificity for glycoprotein IIb-IIIa. J Biol Chem (1993) 268:1066–1073.
Takeda S, Igarashi T, Mori H, Araki S. Crystal structures of VAP1 reveal ADAMs’ MDC domain architecture and its unique C-shaped scaffold. EMBO J (2006) 25:2388–2396.[CrossRef][Web of Science][Medline]
Thompson JD, Higgins DG, Gibson TJ. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res (1994) 22:4673–4680.
Vita C, Roumestand C, Toma F, Ménez A. Scorpion toxins as natural scaffolds for protein engineering. Proc Natl Acad Sci USA (1995) 92:6404–6408.
Xiong J-P, Stehle T, Zhang R, Joachimiak A, Frech M, Goodman SL, Arnaout MA. The crystal structure of the extracellular segment of integrin vβ3 in complex with an Arg-Gly-Asp ligand. Science (2002) 296:151–155.
Yang Z. PAML: a program package for phylogenetic analysis by maximum likelihood. Comput Appl Biosci (1997) 13:555–556.
Yang Z, Nielsen R, Goldman N, Pedersen AMK. Codon-substitution models for heterogeneous selection pressure at amino acid sites. Genetics (2000) 155:431–449.
Yang Z, Wong WSW, Nielsen R. Bayes empirical Bayes inference of amino acid sites under positive selection. Mol Biol Evol (2005) 22:1107–1118.
Zhang J, Nielsen R, Yang Z. Evaluation of an improved branch-site likelihood method for detecting positive selection at the molecular level. Mol Biol Evol (2005) 22:2472–2479.
Zupinski V, Kordis D, Gubensek F. Adaptative evolution in the snake venom Kunitz/BPTI protein family. FEBS Lett (2003) 547:131–136.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||