The deep-rooted origin of disulfide-rich spider venom toxins

  1. Naeem Yusuf Shaikh
  2. Kartik Sunagar  Is a corresponding author
  1. Evolutionary Venomics Lab, Centre for Ecological Sciences, Indian Institute of Science Bangalore, India

Abstract

Spider venoms are a complex concoction of enzymes, polyamines, inorganic salts, and disulfide-rich peptides (DRPs). Although DRPs are widely distributed and abundant, their bevolutionary origin has remained elusive. This knowledge gap stems from the extensive molecular divergence of DRPs and a lack of sequence and structural data from diverse lineages. By evaluating DRPs under a comprehensive phylogenetic, structural and evolutionary framework, we have not only identified 78 novel spider toxin superfamilies but also provided the first evidence for their common origin. We trace the origin of these toxin superfamilies to a primordial knot – which we name ‘Adi Shakti’, after the creator of the Universe according to Hindu mythology – 375 MYA in the common ancestor of Araneomorphae and Mygalomorphae. As the lineages under evaluation constitute nearly 60% of extant spiders, our findings provide fascinating insights into the early evolution and diversification of the spider venom arsenal. Reliance on a single molecular toxin scaffold by nearly all spiders is in complete contrast to most other venomous animals that have recruited into their venoms diverse toxins with independent origins. By comparatively evaluating the molecular evolutionary histories of araneomorph and mygalomorph spider venom toxins, we highlight their contrasting evolutionary diversification rates. Our results also suggest that venom deployment (e.g. prey capture or self-defense) influences evolutionary diversification of DRP toxin superfamilies.

Editor's evaluation

This is an important survey of disulfide-rich peptides (DRPs), which comprise a large fraction of the most functionally important components of spider venom. While spider DRPs were thought to have evolved independently numerous times throughout the spider tree of life, the authors make a solid case for the idea that they all stem from a single common ancestral protein. The study makes a significant advance towards formalizing the diversity of spider venoms, which will be of interest both to scientists working on protein evolution and to those working on functional venomics.

https://doi.org/10.7554/eLife.83761.sa0

eLife digest

The majority of spiders rely on their venom to defend themselves, to hunt, or both. Armed with this formidable weapon, they have managed to conquer every continent besides Antarctica since they first emerged about 495 million years ago.

A closer look at spider venoms hints at an intriguing evolutionary history which has been rarely examined so far. The venom of other animals, such as snakes or scorpions, is usually formed of a wide range of unrelated toxins; in contrast, spiders rely on a single class of proteins, known as disulfide-rich peptides, to create their deadly venom cocktail. This family of molecules is impressively diverse, with each peptide having a distinct structure and mode of action. Its origins, however, have remained elusive.

To fill this knowledge gap, Shaikh and Sunagar scanned the sequences of all disulfide-rich peptides generated to date, bringing together a dataset that includes 60% of all modern-day spiders. The analyses allowed the identification of 78 new superfamilies of spider toxins. They also revealed that all existing peptides originate from a single molecule, which Shaikh and Sunagar named after the powerful Hindu goddess Adi Shakti. This ancestral toxin was present 375 million years ago in the last common ancestor of modern-day spiders.

The work also highlighted that disulfide-rich peptides evolved under different pressures in various groups of spiders; this may be because some species primarily use their venom for hunting, and others for defence. While the ‘hunters’ may need to constantly acquire toxins with new roles and structures to keep their edge over their prey, those that rely on venom to protect themselves may instead benefit from relying on tried-and-tested toxins useful against a range of infrequent predators. Finally, the analyses revealed that the disulphide-rich peptides of Mygalomorphae tarantulas, which form one of the three major groups of spiders, are much more diverse than the related toxins in other spiders. The underlying reason for this difference is still unclear.

Several life-saving drugs currently on the market are based on toxins first identified in the venoms of snakes, cone sails or lizards. Similar discoveries could be unlocked by better understanding the range of deadly molecules used by spiders, and how these came to be.

Introduction

With their killer instinct and deadly toxins, spiders have been at the centre of many myths and folktales from times immemorial. They are an archetypal arthropod group with mid-Cambrian or early Ordovician origin, nearly 495 million years ago (MYA; Lozano-Fernandez et al., 2016). Because of their unique ability to secrete silk and venom, spiders have successfully colonised diverse ecological niches. They are amongst the most successful predators on the planet, with over 50,000 species and 129 families described to date (King, 2004; WSC, 2022). The majority of spiders are equipped with chelicerae harbouring venom glands, with Symphytognathidae, Uloboridae, and certain Mesothelae species being the only exceptions (King, 2004; Mullen and Vetter, 2019).

Spider venoms are a concoction of enzymes, polyamines, nucleic acids, inorganic salts and disulfide-rich peptides (DRPs) (Senji Laxme et al., 2019b; King and Hardy, 2013). They are predominantly rich in DRPs that are characterised by a diversity of structural motifs, including Kunitz (Yuan et al., 2008), disulfide-directed β-hairpin (Wang et al., 2000), disulfide-stabilised antiparallel β-hairpin stack (DABS; Pineda et al., 2020) and inhibitor cystine knot (ICK) – also known as knottins (Pallaghy et al., 1994; Undheim et al., 2016). Despite the fact that DRPs constitute three-quarters of spider venom, our evolutionary understanding of their origin and diversification has remained elusive. This knowledge gap stems from a lack of sequence and structural data for DRPs from diverse spider lineages and the prevalence of significant sequence divergence in these toxins.

Here, we examined DRP sequences from the spiders of the Mygalomorphae infraorder (includes funnel-web spiders and tarantulas), family Theridiidae (includes the red-back spiders), and the Retrolateral Tibial Apophysis (RTA) clade from Araneomorphae, which constitutes over 58% of spider genera (2527 genera) described to date (Figure 1A). A molecular phylogenetic framework implemented in this study resulted in the identification of 78 novel toxin superfamilies and suggests a deep-rooted origin of venom DRPs in spiders. Our findings also highlight the role of distinct prey capture strategies of Araneomorphae and Mygalomorphae in shaping the recruitment and diversification of venom DRPs. Furthermore, by comparatively evaluating spider venom toxins employed for anti-predatory defense and prey capture, we also unravel the impact of the purpose of venom deployment on the evolution of spider venoms. Thus, sequence, phylogenetic, structural and evolutionary assessments in this study have provided insights into the fascinating origin and early diversification of this predominant spider venom component.

Schematic representation of Araneae phylogeny and their venom superfamilies.

Panel A here shows a cladogram of Araneae with lineages under investigation indicated in red. In panel B, representative signal peptide alignments of toxin superfamilies are shown with sequence conservation of >90% highlighted in blue.

Results

Novel spider toxin superfamilies

Superfamilies (SF) of venom toxins in spiders have been classified based on their signal peptide and propeptide sequences (Pineda et al., 2014). This premise was first used to describe the Shiva superfamily of toxins from Atracidae spiders (Pineda et al., 2014). Recently, using a similar approach, 33 novel spider toxin superfamilies have been identified from the venom of the Australian funnel-web spider, Hadronyche infensa (Pineda et al., 2020). Since gene phylogenies have not been extensively utilised while classifying spider venom toxins, our understanding of their origin and diversification has been severely limited.

In this study, we relied on the strong conservation of signal peptide and propeptide regions in identifying several novel spider venom toxin superfamilies, following the same strategy as [Figure 1B; Pineda et al., 2020; Pineda et al., 2014]. Blast searches were used to identify the homology between largely divergent toxin superfamilies. Toxin sequences were found to share strong sequence conservation within a superfamily. Cysteine residues, which are involved in the formation of disulphide bonds and, thereby, are extremely vital in determining protein structure and function, were used as guides to manually refine sequence alignments. This approach enabled the identification of 33 novel toxin superfamilies along the breadth of Mygalomorphae (Figure 2—figure supplements 2 and 3). Among these, 31 superfamilies belonged to the DRP class, whereas the other two were enzymatic non-DRP toxins, including the first report of Neprilysin (SF103) and CAP (CRiSP/Allergen/PR-1; SF104) from Atracidae spiders (Figure 4—source data 1).

Moreover, analyses of Araneomorphae toxin sequences using the strategy above resulted in the identification of 45 novel toxin superfamilies from Araneomorphae, all of which but one (SF109) belonged to the DRP class of toxins (Figure 3—figure supplements 2 and 3). Overall, among all novel spider toxin superfamilies identified in this study, the majority (n=75) were DRPs, reinstating the dominance of this toxin type in spider venoms. Based on the arrangement of cysteine residues involved in the formation of disulphide bonds, these DRPs could be further segregated into ICK-like (n=28), DABS (n=13) and novel disulphide patterned non-ICK (n=34) superfamilies (Pineda et al., 2020). We named these novel spider toxin superfamilies after deities of death, destruction, and the underworld (Supplementary file 1), following a nomenclature system introduced by Pineda et al., 2014.

The identification of novel toxin superfamilies was further supported by phylogenetic and principal component analyses. Reconstruction of evolutionary histories using Bayesian inference (BI) and maximum-likelihood (ML) approaches retrieved monophyletic groups of toxin superfamilies (Figures 2 and 3; node support: ML:>90/100; BI:>0.95; refer to figure supplement provided for complete phylogeny with branch lengths). Interestingly, the plesiotypic DRP scaffold seems to have undergone lineage-specific diversification in Mygalomorphae, where the selective diversification of the scaffold has led to the origination of novel toxin superfamilies corresponding to each Linnaean genus (Figure 2). In our Bayesian and maximum-likelihood phylogenetic tree reconstructions, these toxin scaffolds were found to form distinct clades, further supporting this claim (Figure 2—figure supplement 1; node support: ML: ML:>90/100; BI:>0.95).

Figure 2 with 4 supplements see all
The Bayesian phylogeny of mygalomorph spider venom toxin superfamilies.

This figure represents the Bayesian phylogeny of Mygalomorphae spider toxin superfamilies, where branches with high (BPP >0.95) and low (BPP <0.95) node supports are shown in thick black and thin grey lines, respectively. Coloured spheres alongside tree tips represent the spider genera, while the coloured outer circle indicates the spider family in which the respective toxin superfamily has been identified (Atracidae [red], Barychelidae [orange], and Theraphosidae [blue]).

Figure 2—source data 1

This zip archive contains sequence alignment used to perform phylogenetic analyses and the Bayesian phylogeny of Mygalomorphae DRP toxin superfamilies.

https://cdn.elifesciences.org/articles/83761/elife-83761-fig2-data1-v2.zip
Figure 3 with 4 supplements see all
The Bayesian phylogeny of araneomorph spider venom toxin superfamilies.

This figure represents the Bayesian phylogeny of Araneomorphae spider toxin superfamilies, where branches with high (BPP >0.95) and low (BPP <0.95) node supports are shown in thick black and thin grey lines, respectively. Coloured spheres, alongside tree tips, represent the spider genera, while the coloured outer circle indicates the spider family (Agelenidae [red], Ctenidae [green], multiple araneomorph families [purple]: Ctenidae, Gnaphosidae, Homalonychidae, Lycosidae, Oxyopidae, Pisauridae, Psechridae, Salticidae, Thomisidae, Xenoctenidae) in which the respective toxin superfamily has been identified.

Figure 3—source data 1

This zip archive contains sequence alignment used to perform phylogenetic analyses and the Bayesian phylogeny of Araneomorphae DRP toxin superfamilies.

https://cdn.elifesciences.org/articles/83761/elife-83761-fig3-data1-v2.zip

A similar pattern was also observed in the case of Araneomorphae, where certain toxin SFs (n=12) were found to have diversified within individual genera, corresponding to the Linnaean taxonomy (Figure 3). However, we also documented a large number of DRP toxins (n=32) that were found to have diversified in a family-specific manner, wherein, a toxin scaffold seems to have a more ancient recruitment, corresponding to the Linnean family, rather than the Linnaean genus. As a result, and in contrast to mygalomorph DRPs, araneomorph toxin superfamilies were found to be scattered across spider lineages (Figure 3; Figure 3—figure supplement 1; node support: ML:>90/100; BI:>0.95). Moreover, Principal component analysis (PCA) of toxin sequences further provided evidence for the monophyly of mygalomorph and araneomorph SFs, where each toxin superfamily formed a distinct group in PCA plots (Figure 2—figure supplement 4; Figure 3—figure supplement 4).

Furthermore, sequence alignments of DRPs clearly highlighted the homology among DRP toxin superfamilies (Figure 4; Figure 4—figure supplement 1; node support: ML:>90/100; BI:>0.95). Six cysteine residues were found to be nearly universally conserved across 101 DRP toxin SFs (Figure 4B; Figure 4—figure supplement 2). Our findings enabled us to trace the origin of spider venom DRPs in Opisthothelae, the clade that encompasses Araneomorphae and Mygalomorphae (Magalhaes et al., 2020). Thus, we highlight for the first time that all DRP toxins in spiders may have had a common molecular origin, nearly 375 MYA. It should be noted, however, that functional analyses have been performed only on a handful of mygalomorph toxins, with even fewer studies focusing on araneomorph toxin superfamilies, and that it would be inaccurate to speculate on the functions of these toxins based on homology.

Figure 4 with 2 supplements see all
The Bayesian phylogeny and cysteine framework representation of spider venom DRPs.

This figure depicts the Bayesian phylogeny and alignment of representative sequences of Araneae DRP toxin superfamilies, where branches with high (BPP >0.95) and low (BPP <0.95) node supports are shown in thick black and thin grey lines, respectively. The coloured outer circle in panel A indicates the infraorder of spiders (Mygalomorphae and Araneomorphae shown in dark and light brown, respectively) in which the respective DRP superfamily was identified. In panel B, cysteine framework conserved across toxin SFs is highlighted in blue.

Figure 4—source data 1

This zip archive contains sequence alignment used to perform phylogenetic analyses and the Bayesian phylogeny of Araneae DRP toxin superfamilies.

A list of accession IDs for sequence data utilised in this study has also been provided.

https://cdn.elifesciences.org/articles/83761/elife-83761-fig4-data1-v2.zip

Molecular evolution of spider venom DRP toxins

To evaluate the nature and strength of the selection that has shaped spider venom DRPs, we employed site-specific models that detect selection across sites. Our findings suggest that the majority of Mygalomorphae toxin superfamilies (12/19 SFs) have evolved under the influence of positive selection (ω ranging between 1.1 and 2.9; positively selected sites [PS]: 0–26), while the remaining few have experienced negative or purifying selection (ω ranging between 0.7 and 0.8; PS: 0–13; Figure 5, Figure 5—source data 1). In stark contrast, nearly all of the Araneomorph toxin superfamilies that we investigated here were found to have evolved under a strong influence of negative selection (ω ranging between 0.2 and 1.0; PS: 0–10; Figure 5, Figure 5—source data 1). We further assessed whether these changes documented across sites have a significant effect on the biochemical and structural properties of amino acids using TreeSAAP (Figure 5—source data 1). Outcomes of these analyses revealed the accumulation of replacement changes in Mygalomorphae toxin superfamilies that result in radical shifts in amino acid properties, potentially influencing their structure and function (Figure 5—source data 1).

Figure 5 with 1 supplement see all
Molecular evolution of spider toxin superfamilies.

This figure shows the distribution of ω values (Y-axis) for araneomorph and mygalomorph spider venom toxin superfamilies (X-axis). The horizontal dotted black line represents neutral evolution (ω=1), with ω values above and below it indicating positive (ω>1) and negative (ω<1) selection, respectively.

Figure 5—source data 1

Molecular evolution of toxin superfamilies.

(a): Positively selected sites detected by the Bayes Empirical Bayes approach implemented in M8; (b): Fast Unconstrained Bayesian AppRoximation (c): Sites detected as experiencing episodic diversifying selection (0.05 significance) by the Mixed Effects Model Evolution (MEME). Sites detected at 0.99 and 0.95 significance are indicated in the parenthesis; (d): number of sites under pervasive diversifying selection at the posterior probability ≥0.9 (FUBAR); (e): Number of sites under pervasive purifying selection at the posterior probability ≥0.9 (FUBAR); ω: mean dN/dS. Biochemical properties evaluated: Equilibrium Const. – ionisation, COOH (pK); Hydropathy (h); Long-range n.b. energy (El); Polarity (p); Total n.b. energy (Et). Structural properties: α−helical tendencies (Pα); β-structure tendencies (Pβ); Average # surrounding residues (Ns); Bulkiness (Bl); Chromatographic index (RF); Coil tendencies (Pc); Compressibility (K°); Helical contact energy (Ca); Mean r.m.s. fluctuation displacement. (F); Molecular volume (Mv); Molecular weight (Mw); Partial specific volume (V°); Polar requirement (Pr); Power to be – C-term. α-helix (αc); Power to be – middle, α-helix (αm); Power to be – N-term., α-helix (αn); Refractive index (µ); Solvent accessible reduct. ratio (Ra); Thermodynamics transfer hydrophobicity (Ht).

https://cdn.elifesciences.org/articles/83761/elife-83761-fig5-data1-v2.zip

To comparatively evaluate the nature of selection that shapes venom components deployed either for prey capture or antipredator defence, we employed maximum-likelihood and Bayesian approaches. In these analyses, we identified toxin superfamilies SF74, SF77, SF79, SF89, SF90, SF92, and SF99 as predatory toxins (i.e. toxins deployed for prey capture – refer to the discussion section for the principle considered for this classification), whereas SF13 (i.e. Ares SF) was classified as a defensive spider venom toxin superfamily (i.e. toxins deployed for antipredator defence) as described previously (Herzig et al., 2020). Assessment of molecular evolutionary regimes identified a significant influence of positive selection on venom toxins that are employed for prey capture (ω ranging between 1.2 and 2.9; PS: 0–11, Figure 5—source data 1, Figure 5—figure supplement 1), relative to those that are chiefly or exclusively used for antipredatory defence (ω=0.8; PS: 3; Figure 5—source data 1, Figure 5—figure supplement 1).

Discussion

The deep evolutionary origin and diversification of the primordial knot

Prior attempts to explore the phylogenetic and evolutionary histories of spider venom DRPs have hypothesised independent origin and lineage-specific diversification of DRP venom toxins (Rodríguez de la Vega, 2005). In contrast, recent literature, primarily focusing on Hadronyche infensa, suggests that the diverse disulfide-rich venom arsenal of this Australian funnel-web spider is a derivative of an ancestral ICK motif that underwent several rounds of duplication and diversification (Pineda et al., 2020). Often restricted to a specific spider lineage, or given the inconsistent ways of classifying spider venom toxins, previous attempts have failed to provide a broader perspective on the evolution of these peptides (Ferrat and Darbon, 2005; Chen et al., 2008). Given their very long evolutionary histories, genes encoding DRP toxins have undergone significant diversification, making it difficult to precisely trace their phylogenies. Together with the lack of structural and functional data for these toxins, all of the aforementioned factors have impeded our understanding of the origin and evolution of this predominant spider venom component.

To address this knowledge gap, we employed sequence comparisons, phylogenetic inferences and evolutionary analyses, which enabled the identification of 76 novel spider venom toxin superfamilies (45 from Araneomorphae [44 DRP and 1 non-DRP] and 33 from Mygalomorphae spiders [31 DRP and 2 non-DRP]). Our findings strongly suggest a deep-rooted origin of DRP spider venom toxin superfamilies (Figure 4A), possibly from a single ancestral DRP or knottin scaffold, which we name ‘Adi Shakti’, after the original creator of the universe according to Hindu mythology. We propose that all of the extant spider disulfide-rich peptide toxin superfamilies (n=101) in Mygalomorphae and Araneomorphae, which include those that were previously reported (n=26), as well as the ones identified in the present study (n=75), have originated from this ‘primordial knot’, further undergoing lineage-specific gene duplication and diversification (Figures 24). The origin and diversification of these superfamilies can be explained by a mechanism that is similar to the combinatorial peptide strategy, wherein certain venomous animals, such as cone snails, generate a remarkable diversity in their mature toxin peptides while preserving the signal and propeptide regions (Escoubas and Rash, 2004; Zhu et al., 2000; Olivera et al., 1995). Rapid events of diversification, preceded by repeated rounds of gene duplication, form the basis of the combinatorial peptide library strategy (Sollod et al., 2005). These hyper-mutational events have been previously shown to be restricted to the mature peptide regions of toxins (Conticello et al., 2001). In contrast, the signal and propeptide regions, which are vital for the precise secretion and folding of proteins, respectively, evolve under the strong influence of negative selection pressures (Duda and Palumbi, 1999) - a molecular evolutionary trend also reported in venom coding genes of snakes (Brust et al., 2013). Spider venom coding genes appear to have followed a similar strategy. However, unlike the cone snail venom coding genes that have a recent evolutionary origin (<35–50 MYA; Olivera, 1997; Duda and Kohn, 2005), spider venom toxins have likely originated from an ancestral scaffold in Opisthothelae, the clade that encompasses Mygalomorphae and Araneomorphae spiders (over 99% of all spider genera described to date), nearly 375 MYA (Magalhaes et al., 2020). Given their significant sequence divergence since their deep-rooted evolutionary origin, the entire protein-coding gene, including the signal and propeptide regions, has accumulated significant differences. Consistent with this hypothesis, the majority of positively selected sites (~96%) identified in spider venom DRP toxins (all sites in Araneomorphae, and all but two sites in Mygalomorphae) were restricted to the mature peptide region, whereas the signal and propeptide regions harboured a minor proportion of these sites (1% and 3%, respectively; Figure 5—source data 1).

It has been theorised that the plesiotypic (or ancestral) DRP scaffold comprised of eight cysteines that formed four disulfide bonds (Pineda et al., 2020; Cole and Brewer, 2021). However, the evolutionary history of DRP scaffold in spider venom has been riddled with events of duplication and diversification, which may have resulted in multiple gain and loss of structural and functional residues, including cysteines. As a result, we find that DRP toxins in extant spiders are comprised of distinct scaffolds with a range of cysteine pairs (4–12 cysteines forming 2–6 disulfide bonds). For example, an individual spider from the Haplopelma genus may contain SF90, SF91, SF99 and SF100 toxin superfamilies in its venom with 6, 8, 12, and 10 cysteines, respectively. This makes it very difficult to trace the nature and cysteine skeletal structure of the plesiotypic scaffold – something that could be answered in future with the help of comparative genomics and synteny analyses. Our extensive phylogenetic and evolutionary analyses provide insight into the common origin of DRP toxins in spiders, dating back to the common ancestor of Mygalomorphae and Araneomorphae. However, we refrain from speculating on the exact nature of this plesiotypic scaffold.

Contrasting weaponisation strategies: recruitment versus innovation

Venom is an intrinsically ecological trait that has underpinned the evolutionary success of many animals (Suranse et al., 2022). The ability of venomous organisms to incapacitate prey and predators emanates from toxins that exhibit an array of biochemical activities and target divergent pathways. Many venomous lineages deploy a wide range of toxins from phylogenetically unrelated superfamilies. Venomous snakes, for example, have ‘recruited’ a myriad of toxins, including snake venom metalloproteinases, snake venom serine proteases, three-finger toxins, phospholipase A2s, L-amino acid oxidases, Kunitz-type serine protease inhibitors, kallikreins, lectins, DNases, and hyaluronidases (Casewell et al., 2013; Faisal et al., 2021; Dutta et al., 2017; Senji Laxme et al., 2019a; Casewell et al., 2020; Figure 6). Similarly, spider venoms typically possess many forms of enzymes (e.g. phospholipases, proteases and chitinases), polyamines, salts, and disulphide-rich toxins (King and Hardy, 2013; Figure 6). However, spider venom DRPs with diverse ion channel targeting activities, such as sodium, potassium, calcium, and chloride ion channels, predominate in the venoms of nearly all spiders, constituting three-quarters of the venom (Figure 6). Phylogenetic and evolutionary assessments in this study trace the evolutionary origin of DRPs in Opisthothelae, the suborder that includes the majority of spiders described to date. This strategy, wherein, a molecular scaffold with a single deep-rooted evolutionary origin constitutes the major content of the venom, is unique to spiders. Venoms of most other animals are, instead, composed of unrelated toxin types, derived from distinct scaffolds/gene superfamilies in varying proportions. Thus, instead of recruiting distinct toxins with diverse functions into their venoms like the majority of venomous animals, spiders seem to have diversified a single molecular template to generate a commensurate functional diversity in their venoms. These findings not only shed light on the fascinating evolutionary history of spider venoms but also highlight an unrealized potential of molecular scaffolds in underpinning the dramatic structural and functional diversification of the venom arsenal.

Distinct toxin scaffold recruitment strategies in spiders and snakes.

This figure depicts distinct toxin scaffold recruitment strategies in (A) spiders and (B) advanced snakes. The Araneae phylogeny highlights the domination of disulfide-rich peptide toxins in spiders [Atracidae: Atrax sp.; Theraphosidae: Poecilotheria formosa; Theridiidae: Latrodectus mactans; Ctenidae: Phoneutria nigriventer: e.g., Palagi et al., 2013; Oldrati et al., 2017; Diniz et al., 2018], whereas venoms of advanced snakes are constituted by diverse phylogenetically unrelated toxin superfamilies (Viperidae: Daboia russelii, Elapidae: Naja naja, Colubridae: Spilotes sulphureus: e.g., Senji Laxme et al., 2021a; Senji Laxme et al., 2021b; Modahl et al., 2018). Doughnut charts, portraying the major molecular scaffolds in venom are also shown disulfide-rich peptides (yellow), snake venom metalloproteinases (SVMP, red), phospholipase A2 (PLA2, green), three-finger toxins (3 FTx, blue) and other minor components (black). Structures of the major scaffolds are also shown, with helices coloured in green, β-strands in blue and disulfide bonds in orange.

Distinct recruitment and diversification of spider venom superfamilies in Mygalomorphae and Araneomorphae

In addition to suggesting the common evolutionary origin of DRP toxins, Bayesian and maximum-likelihood phylogenies provided intriguing insights into the early diversification of DRPs in spiders. Mygalomorph DRP toxin superfamilies formed lineage-specific toxin clades (63/66) that suggested the recruitment of unique DRP scaffolds at the level of Linnaean genera (Figure 2), while the majority of unique DRP scaffolds seemed to be recruited ancestrally at the level of Linnaean families in Araneomorphae (Figure 3). Only a minor fraction (6/38) of araneomorph toxin superfamilies were recruited at the level of Linnaean genera.

When the nature and strength of selection on venom DRPs were assessed, a strong influence of positive selection was identified on the evolution of these toxin superfamilies in mygalomorph spiders. Only a minority of these toxin superfamilies were found to be evolving under negative selection (6/19), or under near neutral evolution (1/19), while the majority (12/19) experienced diversifying selection (ω between 1.19 and 2.95; PS: 0–26, Figure 5). In complete contrast, the evolution of venom superfamilies (21/22) in Araneomorphae was constrained by purifying selection (ω between 0.03 and 0.97; PS: 1–3, Figure 5), and a single superfamily was found to be evolving nearly neutrally (ω of 1.0; PS: 10). We further investigated the impact of these amino acid replacements on the structure and function of spider venom toxins. Outcomes of these evaluations suggest that the majority of replacements in mygalomorph spiders (between 0 and 29 properties) had a radical effect on the structure and/or biochemical property of the encoded toxin, while none were identified in most toxin superfamilies of Araneomorphae. Only a minor proportion of non-synonymous substitutions in two toxin superfamilies (SF40 and SF68) of araneomorph lineage were reported to be radically different (Figure 5—source data 1). Differences in the evolutionary histories of mygalomorph and the araneomorph DRP toxin superfamilies became apparent as we further evaluated them for the signatures of episodic diversification. We detected a greater prevalence of episodic diversifying selection on mygalomorph DRP toxin superfamilies than their araneomorph counterparts (0–34 versus 0–6 events, respectively).

Such starkly contrasting phylogenetic and evolutionary patterns are indicative of differential recruitment and diversification of DRPs in spiders. We postulate the following hypotheses that could possibly explain this unique pattern of spider venom evolution.

Scenario 1. Distinct recruitment hypothesis

Mygalomorphae spiders may have recruited individual toxin superfamilies or unique toxin scaffolds post the emergence of spider family members corresponding to the Linnaean taxonomy (Figure 7A). This could explain why toxin superfamilies in Mygalomorphae form lineage-specific clades that correspond to individual spider genera (Linnaean taxonomy) in our phylogenetic analyses (Figure 2). In contrast, Araneomorphae may have recruited unique toxin scaffolds prior to the divergence of family members, insinuating an ancestral recruitment event (Figure 7B). This, perhaps, explains why araneomorph toxin SFs are scattered across spider families and form Linnaean family-specific groups in phylogenetic trees (Figure 3).

Hypotheses explaining the stark differences in recruitment and diversification of toxin SFs in Araneomorphae and Mygalomorphae.

This figure depicts various hypotheses that explain distinct toxin SF recruitment and diversification in spiders. Scenario 1 depicts genus- or family-specific recruitment of spider toxin SFs in Mygalomorphae and Araneomorphae, respectively, while scenario 2 highlights the implications of differential rates of diversification.

Scenario 2. Differential molecular diversification rate hypothesis

The apparent Linnaean taxa-specific (genus- and family-) diversification of spider venom toxin superfamilies can also be explained by the differential rate of molecular evolution in the spider infraorders. In this schema, the recruitment of toxin SFs could have happened in the common ancestor of Mygalomorphae and Araneomorphae. However, the contrasting rates of diversification, wherein, mygalomorph toxin SFs underwent extensive diversification under positive selection, while araneomorph toxin SFs were very well conserved under negative selection, resulted in the contrasting patterns of DRP diversification that we see today (Figures 5, 7C and D).

Scenario 3. Prey-capture strategies and toxin recruitment hypothesis

The use of webs for prey capture in araneomorphs versus the sit-and-wait predation strategy of mygalomorphs may have resulted in the selective diversification of toxin superfamilies. Since most araneomorph spiders heavily rely on their foraging web for prey capture, and because these spiders mostly prey on insects (Pérez-Miles and Perafán, 2017), we speculate that their venom DRPs exhibit relatively lower sequence diversity (Figure 5, Figure 5—source data 1). In complete contrast, venom DRPs in mygalomorph spiders that mostly rely on venom and not silk, being either ambush or sit-and-wait predators to capture a much diverse prey base, appear to have experienced a significantly greater influence of the diversifying selection (Beydizada et al., 2022; Figure 5, Figure 5—source data 1). However, the current literature and our investigation are limited to the most diverse lineage in Araneomorphae - the RTA clade. Since this lineage does not employ silk webs for predation, Scenario 3 is unlikely to explain the current observations. Surprisingly, however, despite being the most speciose spider lineage, and having a significantly higher genomic diversification rate in comparison to other araneomorphs (Fernández et al., 2018), the lack of toxin sequence diversity in the RTA clade is intriguing (Figure 5, Figure 5—source data 1). It should also be noted that venom toxins from the foraging web-building araneomorphs outside the RTA clade are very poorly studied (e.g. only a handful of species are investigated from a biodiscovery perspective, and not a single toxin has been sequenced at the nucleotide level to date).

Deployment dictates spider venom evolution

The current literature is replete with findings that support the strong influence of positive selection on genes encoding venom toxins in diverse animal lineages (Juárez et al., 2008; Sunagar et al., 2012; Sunagar et al., 2013; Župunski and Kordiš, 2016). Venom proteins are theorised to follow a ‘two-speed’ mode of evolution, wherein they readily diversify in animals that experience drastic shifts in ecology and/or environment - a prominent feature of evolutionarily younger lineages [e.g. cone snails and advanced snakes with evolutionary origins dating back to <35–50 MYA (Sunagar and Moran, 2015)]. This rapid expansion, or the ‘expansion phase’, is shaped by a strong influence of positive selection that underpins the transition of organisms into novel ecological niches. Post these adaptive changes, the influence of diversifying selection is replaced by the effects of purifying selection (the ‘purification phase’) that preserve potent toxins generated during the expansion phase. This, perhaps, explains the contrasting evolutionary regimes documented in evolutionarily younger and ancient lineages (Sunagar and Moran, 2015). Venom coding genes in evolutionarily ancient lineages are said to re-enter the expansion phase if they re-encounter dramatic shifts in ecology and environment. The only exceptions to this hypothesis are toxins that non-specifically interact with their molecular targets or those that are deployed for antipredatory defence (Sunagar and Moran, 2015). The latter hypothesis, however, mostly stems from the analyses of venom proteins that are deployed for predation. A dearth of sequence information for venom components majorly employed for antipredator defence has impeded our understanding of their evolutionary diversification.

Spiders of the genera, Hadronyche and Atrax (family Atracidae), are known to deploy their DRP toxin superfamily (SF13: Ares) predominantly for antipredatory defence (Herzig et al., 2020). In contrast, tarantulas of the family Theraphosidae are known to mostly employ their venom to capture prey animals. This provided us with a unique opportunity to comparatively investigate the molecular evolution of spider venom proteins chiefly deployed for predation (SF74, SF77, SF79, SF89, SF90, SF92, and SF99 from Theraphosidae) and self-defence (SF13 from Atracidae). Our analyses of the molecular evolutionary histories of theraphosid spider venom DRPs deployed for prey capture reveal a strong influence of diversifying selection (ω: 1.2–2.9; PS: 0–11; Figure 5—source data 1, Figure 5—figure supplement 1), whereas those employed for self-defence in Atracidae spiders were constrained by negative selection (ω: 0.8; PS: 3; Figure 5—source data 1, Figure 5—figure supplement 1). Outcomes of FUBAR and MEME analyses further corroborated these findings. FUBAR identified numerous sites (~10%) in defensive toxins as evolving under the pervasive influence of negative selection, while MEME detected several episodically diversifying sites (~22%) in theraphosid toxins deployed for prey capture (Figure 5—source data 1).

Such contrasting modes of diversification could be attributed to the ‘two-speed’ mode of venom evolution, where the offensive toxins gain an evolutionary advantage over prey by amplifying their sequence and functional diversity (Sunagar and Moran, 2015). In contrast, as defensive venoms are infrequently deployed, or have evolutionarily conserved molecular targets across predatory lineages, they experience relatively reduced effects of diversifying selection. In the absence of a need for sequence variation, purifying selection pressures instead ensure the preservation of broadly effective toxins (Sunagar and Moran, 2015).

Methods

Sequence data curation and assembly

Nucleotide datasets consisting of Mygalomorphae DRP sequences were assembled from the National Center for Biotechnology Information’s Non-redundant and Transcriptome Shotgun Assembly databases using manual search and exhaustive BLAST iterations (Altschul et al., 1990). Sequences for Araneomorphae toxins were retrieved using a similar strategy, while additional sequences from the RTA clade were derived from Cole and Brewer, 2021. A list of sequence data analysed in this study has been provided as Figure 4—source data 1. Translated sequences were aligned in MEGA X (v 10.2.6) using MUSCLE (Edgar and Batzoglou, 2006; Kumar et al., 2018) before back-translation to nucleotides. Alignment was further refined by using structurally conserved cysteines as guides.

Toxin superfamily identification and nomenclature

Spider toxin superfamilies were identified based on the strong conservation observed in signal peptide and propeptide sequences as illustrated by Pineda et al., 2020; Pineda et al., 2014. Additional support for uniqueness of each superfamily was obtained by our extensive phylogenetic (both Bayesian and Maximum likelihood) and Principal Component Analyses of toxin sequences. All the novel toxin superfamilies were labelled after gods/deities of death, destruction and underworld based on a nomenclature system as described before (Pineda et al., 2014). A list of novel superfamilies identified has been included as a part of supplementary file (Supplementary file 1).

Phylogenetic analyses

Phylogenetic histories of toxin families were reconstructed for whole length toxin nucleotide sequence using Bayesian and maximum-likelihood inferences implemented in MrBayes 3.2.7 a (Altekar et al., 2004; Ronquist et al., 2012) and IQ-TREE v1.6.12 (Nguyen et al., 2015; Chernomor et al., 2016 f), respectively. All alignments utilised have been made available in Figure 2—source data 1, Figure 3—source data 1 and Figure 4—source data 1, respectively. Bayesian analyses were run for a minimum of ten million generations using twelve Markov chains across four runs, sampling every 100th tree. Twenty-five percent of the total trees sampled were discarded as burn-in. The log-likelihood score for each tree was plotted against the number of generations to assess whether the analysis has reached an asymptote. A stop value of 0.01 was used for the average standard deviation of split frequencies. Bayesian Posterior Probability (BPP) was used to evaluate node support for the branches of Bayesian trees. ML analyses were performed using IQ-TREE with an edge-proportional partition model and 100 Bootstrap replicates. The best partition scheme for each partition was determined by utilising the inbuilt ModelFinder plugin in IQ-TREE (Kalyaanamoorthy et al., 2017). Phylogenetic trees were rooted with non-venom nucleolar cysteine-rich protein sequences from Mastigoproctus giganteus, Stenochrus portoricensis, Prokoenenia wheeleri, Phrynus marginemaculatus and Cryptocellus centralis from the class Arachnida that fall outside of the suborder Opisthothelae.

Principal component analysis

PCA of signal peptide sequences from spider toxin superfamilies was performed in R (v 4.1.2; R Development Core Team, 2021) using a previously published script (Konishi et al., 2019; https://github.com/TomokazuKonishi/direct-PCA-for-sequences; (McClellan and McCracken, 2001). Sequences were aligned using MUSCLE in MEGA X (v 10.2.6) (Edgar and Batzoglou, 2006; Kumar et al., 2018) and further digitising in R utilising boolean vectors. The scaled principal component values (sPC) were calculated using conventional PCA prior to plotting.

Assessment of molecular evolution

The nature of selection shaping the evolution of DRP toxins was determined using a maximum-likelihood inference implemented in CodeML of the PAML package (v 4.9 j) (Yang, 2007). Superfamilies with a minimum of 15 sequence representatives were further down-selected for this analysis to avoid inaccurate estimation of omega values when analysing smaller datasets. The ratio of non-synonymous substitutions (nucleotide changes that alter the coded amino acid) to synonymous substitutions (nucleotide changes that do not alter the coded amino acid), also known as omega (ω), was estimated. A likelihood ratio test (LRT) for the nested models - M7 (null model) and M8 (alternate model) - was performed to assess the statistical significance of the findings. The Bayes Empirical Bayes (BEB) approach implemented in M8 was used to calculate the posterior probabilities for site classes (Yang et al., 2005). Amino acid sites with a posterior probability of over 95% (PP ≥95%) were inferred as positively selected. The episodic and pervasive nature of selection was determined using the Mixed Effect Model of Evolution (MEME; Murrell et al., 2012) and the Fast Unconstrained Bayesian AppRoximation (FUBAR; Murrell et al., 2013), respectively.

Evaluation of selection on amino acid properties

The influence of positive selection on the biochemical and structural properties of amino acids was evaluated using TreeSAAP (v 3.2; Woolley et al., 2003). TreeSAAP estimates the rate of selection using a modified MM01 model (McClellan and McCracken, 2001). Statistical probabilities corresponding to a range of properties were further calculated for each amino acid. BASEML was set to run with the REV model and eight evolutionary pathway categories were defined for evolutionary pathway analyses with a sliding window size set to one. Data acquired from TreeSAAP was further visualised and processed with IMPACT_S (Maldonado et al., 2014).

Structural analyses

Structural homologues of spider toxin superfamilies were identified via blast searches against the RCSB Protein Data Bank (https://www.rcsb.org/) and subsequently modelled using the SWISS-MODEL web server via user template mode (Waterhouse et al., 2018). The resultant models were validated using MolProbity (v 4.4; https://github.com/rlabduke/MolProbity; Williams et al., 2018) and general Ramachandran plot. Regimes of evolutionary selection pressures were evaluated and mapped onto homology models using the Consurf webserver (Ashkenazy et al., 2016, http://consurf.tau.ac.il/). PyMOL v2.5.2 (Schrödinger, LLC, USA) was used to visualise and generate the images of homology models.

Data availability

All data generated or analysed during this study are included in the manuscript and supporting files. Source data files have been provided for Figures 2, 3, 4 and 5.

References

  1. Book
    1. Mullen GR
    2. Vetter RS
    (2019) Spiders (araneae)
    In: Mullen GR, editors. Medical and Veterinary Entomology. Elsevier. pp. 507–531.
    https://doi.org/10.1016/B978-0-12-814043-7.00025-X
  2. Book
    1. Pérez-Miles F
    2. Perafán C
    (2017) Behavior and biology of mygalomorphae
    In: Viera C, Gonzaga M, editors. Behaviour and Ecology of Spiders. Springer. pp. 29–54.
    https://doi.org/10.1007/978-3-319-65717-2_2
  3. Software
    1. R Development Core Team
    (2021) R: A language and environment for statistical computing
    R Foundation for Statistical Computing, Vienna, Austria.
    1. Suranse V
    2. Iyer A
    3. Jackson TN
    4. Sunagar K
    (2022)
    The Origin and Early Evolutionary History of Snakes
    248–268, Origin and early diversification of the enigmatic squamate venom cocktail, The Origin and Early Evolutionary History of Snakes, 10.1017/9781108938891.016.
  4. Software
    1. WSC
    (2022) World spider catalog, version Version 23.0
    Natural History Museum, Bern.

Decision letter

  1. Ariel Chipman
    Reviewing Editor; The Hebrew University of Jerusalem, Israel
  2. Christian R Landry
    Senior Editor; Université Laval, Canada
  3. Ariel Chipman
    Reviewer; The Hebrew University of Jerusalem, Israel
  4. Michael Brewer
    Reviewer; East Carolina University, United States

Our editorial process produces two outputs: (i) public reviews designed to be posted alongside the preprint for the benefit of readers; (ii) feedback on the manuscript for the authors, including requests for revisions, shown below. We also include an acceptance summary that explains what the editors found interesting or important about the work.

Decision letter after peer review:

Thank you for submitting your article "The primordial knot: the deep-rooted origin of the disulfide-rich spider venom toxins" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, including Ariel Chipman as Reviewing Editor and Reviewer #1, and the evaluation has been overseen by Christian Landry as the Senior Editor. The following individual involved in the review of your submission has agreed to reveal their identity: Michael Brewer (Reviewer #3).

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.

Essential revisions:

All three reviewers judge that this is an important body of work. However, in the discussion there was also agreement that some claims need to be better substantiated and that some of the analyses will need to be re-done unless you can make a convincing case for why they were done as they were. Please note the following key points:

1) Explain why not all the data from the cited references were included (most notably ref. 38). If you cannot justify not including all the data, please repeat the analysis with all available data.

2) Modify your conclusions to be specific about the representation of spider diversity included in the study.

3) Give more details in the Methods section to make your conclusions stronger and better supported.

4) Correct errors in phylogenetic nomenclature and usage of terminology (e.g. confusion between clade names and ancestors).

5) Explain the names and references from mythology and popular culture to make them clear to a culturally diverse readership.

Reviewer #1 (Recommendations for the authors):

The manuscript is rich in cultural references and metaphors. This is not a problem on its own, but in many cases, the reason for the metaphor or cultural reference is unclear. I am fully in favor of authors bringing in their own cultural background, but it should be put in context for the reader. I assume the "ancestral knot" in the manuscript's title refers to knottin proteins, but this is not made clear, and I may be missing a deeper reference. The decision to name the ancestral protein "Adi Shakti" is confusing until one realizes that all the protein super-families were given names from Indian mythology (and also a few from other mythologies?) Are these names that were given by the authors as part of the preparation of this paper, or did some of them already exist? Is there a convention to name venom proteins after figures from Indian mythology? I don't know, and I assume the reader won't either. Please give this background explicitly somewhere. Similarly, is there a deeper context to "many ways to skin a cat"? There might be something I am missing, but this seems to me to be a metaphor out of place. This is a short paper, you can afford to expand a bit on some of the metaphors and references.

On a completely different note – while the results are interesting and important, there is a bit too much use of hyperbolic language. I suggest cutting a few "fascinating" and "astonishing".

Reviewer #2 (Recommendations for the authors):

This interesting study looks into the evolution of putative spider venom toxins, specifically disulfide-rich peptides. The results and interpretations presented are a promising start, but I don't consider them to be sufficiently convincing yet to warrant publication in eLife in their current state. Below I make suggestions for some further analyses that can make the conclusions more robust.

The study builds upon previous work, especially references 9, 12, and 38. However, I think that the authors need to discuss more clearly how their results and conclusions relate to those reached in these papers. They should do this on several levels to enable readers, who are not spider venom experts, to better understand this contribution. For instance, on the most superficial level, it would be helpful to better explain that the strategy of naming toxin superfamilies for deities, which struck this reviewer as peculiar, was actually established in previous works. More importantly, clarification is needed about the sources of the sequences. The paper discusses the results from the perspective of spider venom. Yet, all the non-ctenid araneomorph data seem to be derived from whole-body transcriptomes generated by reference 38. Therefore, the results have to be presented with caution because the tissue source of the putative toxin sequences in these taxa is uncertain.

Moreover, it is difficult for me to understand how the authors identified new DRP superfamilies when the phylogenetic analyses that they provide do not seem to incorporate all previously published relevant data. For instance, of the 26 DRP superfamilies identified in reference 9, only 14 are included in the analyses presented here. This makes it impossible to assess how these relate to the new superfamilies reported here, or if some of the new superfamilies may be a subset of previously established superfamilies or vice versa. Why were these data excluded from the phylogenetic analyses? This is puzzling because they are included in the supplementary alignment S10. It makes sense to me to also include these excluded superfamilies in the phylogenetic analyses.

– I suggest removing the reference to Adi Shakti from the abstract or including an explanation because few non-Hindus will understand what this refers to.

– This study attempts to trace the evolutionary origin of DRPs in a phylogenetically diverse clade with many spider genera and families. It would be a very helpful context for readers if the authors clarify on a schematic phylogeny of spiders exactly from which genera and families the data in their paper comes and which higher-level clades these represent. For instance, figure 2 includes genera from several araneomorph families. Which families? This gives readers a better sense of how comprehensively the analyses data represent spider phylogeny as a whole.

– Figures 1 and 2; please insert clade support values. Also please clarify if whole toxin sequences or just signal and propeptide sequences were used to construct the trees. Also, which of the included superfamilies are new? Indicate that in the tree or legend and amend the title. Also, where did the data for the taxon Pisaurina come from? It is not present in reference 38, which is mentioned as the source of the araneomorph sequences.

– Figure 3B: this alignment suggests that there are no gaps. Is that correct? If not, please indicate alignment gaps.

– Lines 90-93: it is more accurate to also mention that the cysteine framework was used to identify superfamilies as well.

– Lines 140-141: Opisthothelae denotes a clade comprising araneomorphs and mygalomorphs, not their common ancestor, so the analysis traces DRPs to the origin of this clade.

– Line 145: 'speculate on'.

– Lines 188-189: you refer to figure 3a when mentioning Adi Shakti. I don't understand how this relates to the figure. Do you want to call the ancestral cysteine scaffold Adi Shakti? If so, say this specifically. And also, what does this ancestral scaffold look like? References 9 and 38 reconstruct the ancestral DRP as having 8 cysteines. Do your analyses agree with this? Was this tested?

– Lines 189-192: this should be rephrased because there certainly are other spider venom toxins that do not belong to the DRP group.

– Lines 204-205: rephrase. You have reconstructed the origin of DRPs in the last common ancestor of Opisthothelae, which is a clade, not an ancestor.

– Line 208: positively selected sites or amino acids.

– Line 224: predominate in the venoms.

– Lines 238-240 (also lines 128-130, 262-264): I am unconvinced by the arguments here and by the data claimed to support this. Genera and families are Linnean ranks of arbitrary phylogenetic depth. They do not meaningfully coincide with biologically relevant lineages. Moreover, the taxa sampled only cover a small fraction of the phylogenetic breadth of spider lineages. As a result, interpretational difficulties emerge. For example, in Figure 1 some of the genera represent families (Trittame is the only taxon representing Barychelidae), and some genera are not split (Atrax and Hadronyche are colour coded the same, and together represent the family Atracidae), while in Figure 2 several of the supposed family-level taxa are only represented by a single genus, including Salticidae, Thomisidae, Lycosidae, and Homalonychidae. It, therefore, becomes impossible in many instances to distinguish whether toxin recruitments happened on the level of genera, families, or even higher-level taxa. To use Figure 1 to argue that mygalomorph toxin recruitments happened "post divergence of family members" (lines 263) is unconvincing since the included genera represent just three families that diverged between 150-100 million years ago (see Opatova et al., 2019, Syst. Biol. 69: 671). To get a more accurate estimate of when toxin families were recruited the authors would have to do ancestral state reconstruction on an established phylogeny.

– Lines 265-266: I don't understand this. Do the araneomorphs have fewer DRP superfamilies than the mygalomorphs?

– Line 329: were the phylogenetic analyses done on the basis of nucleotide or amino acid data? Also, what substitution models were used to generate the trees, and how were these chosen?

– Line 393: correct spelling of Mygalomorphae.

– Line 415: the dotted line is black.

– Figure 5: you need to reverse the colour labeling for 3FTXs and SVMPs, which should be blue and green, respectively, and you need to pair the 3FTXs with the elapids and the SVMPs with the vipers. Also, do the donut charts depict proteomic or transcriptomic data, and are they for specific species (if so, which?) or some sort of average (if so, based on what species and how were the proportions calculated?)? Moreover, in contrast to what you write in the legend and main text (e.g. lines 227-228), it seems that for each of the major groups of venomous snakes one or two types of toxins dominate venom composition as well, just like in the spiders. Please discuss this.

– lines 228-229: this is also true for spiders. Their venoms do not exclusively contain DRPs.

– line 243 and further (and related text in the Results section): the contrast between mygalomorphs and araneomorphs in terms of the evolutionary selection pressures governing toxin evolution is indeed striking at first sight. However, before trying to explain these in general terms (lines 261 and further), the authors need to ask if these results can actually be generalized. 11 of the 12 identified superfamilies of mygalomorph toxins that evolve under the influence of positive selection are limited to just two genera: Hadronyche and Haplopelma. How representative are just these two genera for conclusions about mygalomorph evolution generally? Are the results not equally well explained (or better) in terms of the specific biology of these two taxa? Similarly, 16 of the 20 DRP superfamilies in the selection analyses for araneomorphs are restricted to just 1 araneomorph family (Ctenidae), which detracts from how general the results can be claimed to be. Moreover, it is not clear to me why the selection analyses were performed on only a subset of the spider DRPs included in this study and previously published work. How were these superfamilies selected, and why were others excluded? Could the picture change if more superfamilies were analysed? This makes it impossible to judge how general the results may be as indicators for evolutionary mechanisms in the two clades.

– Figure S2: I don't understand the significance of the 'novel' cysteines in green. Are these depicted separately because they are part of the scaffold? Please clarify. Also, are the names given to these new superfamilies, like Seker, the official new superfamily names you propose or are they there to enliven the trees? Please clarify. Also, make clear that these exclude the novel non-DRP toxins.

– Supplementary figures 5, 6, and 9: please make sure that the published trees have legible clade support measures.

Reviewer #3 (Recommendations for the authors):

I greatly enjoyed reading your work and am grateful to be asked to review it. I hope my comments improve your already well-done study.

Please see the comments in my public review regarding method documentation. Perhaps a GitHub repository documenting all of the analyses and code along with a Figshare repository containing all of the data, results, and supplements would be enough.

I hope I did not misunderstand your use of only our ctenid data to represent araneomorphs. I hope the example datasets I provided are useful, and I apologize if my interpretation is incorrect.

Targeted comments:

The abstract states that "these lineages constitute over 50% of the extant spiders" following a discussion of the Mygalomorphae and Araneomorphae. These two lineages comprise far more than 50% of spider species. Are you referring to the Mygalomorphae + RTA-clade only?

Line 63 – "primitive Mesothelae species" Species are not primitive, but individual traits can be.

Lines 118 and 124: "monophyletic clade" is redundant.

Lines 236 – 241: Is the comparison of taxonomic ranks employed here fair? Are spider families and genera separated by 350 million years assumed to be comparable? Perhaps evolutionary ages or relative species diversity within the ranks could standardize the comparisons. I am not sure how best to approach this, but it stands out to me as potentially problematic, but I am not sure it greatly impacts your overall conclusions.

Lines 264 – 266: While it is true that many araneomorphs rely on prey capture webs, none of the species included in this work do. Making these claims without including species that do utilize prey-capture webs is tenuous.

I greatly enjoyed this work and think it will advance the field forward. I am honored that you utilized the data presented in my former PhD student's and my work so prominently and appreciate the novel directions you took with them. I hope my comments are useful and look forward to seeing a future version of the manuscript.

https://doi.org/10.7554/eLife.83761.sa1

Author response

Essential revisions:

All three reviewers judge that this is an important body of work. However, in the discussion there was also agreement that some claims need to be better substantiated and that some of the analyses will need to be re-done unless you can make a convincing case for why they were done as they were. Please note the following key points:

1) Explain why not all the data from the cited references were included (most notably ref. 38). If you cannot justify not including all the data, please repeat the analysis with all available data.

We have, indeed, analysed all spider toxin sequences available to date. We have relied on the signal and propeptide regions for identifying novel superfamilies, which is an accepted convention: Pineda et al. (2014 BMC Genomics); Pineda et al. (2020 PNAS). Although many additional superfamilies can be identified, we have only retained those sequences for which there were at least 5 representatives for the identification of toxin superfamilies, and 15 representatives for selection analyses to ensure robustness. This filtering step ensured that the generated alignments, phylogenetic trees and evolutionary assessments were robust and devoid of noise that stems from single-representative groups. Adding in those sequences would have enabled us to identify many more superfamilies, solely based on the signal and propeptide examination, but it wouldn’t have been possible to support them with other lines of evidence that were provided for all other superfamilies in this study, jeopardising the overall quality of the manuscript. Nonetheless, there is strong evidence that the left-out sequences are also related to the ones analysed in this study (Figure 4 —figure supplement 2). In future, when more transcriptomes are sequenced, it would be possible to designate these newer toxin superfamilies with much stronger support.

Please note that, in response to some of the reviewer’s comments, we have now included additional sequences and reanalysed data, which has now led to the identification of 5 additional toxin superfamilies.

2) Modify your conclusions to be specific about the representation of spider diversity included in the study.

We were attempting to explain the contrasting rates of diversification of toxin superfamilies in mygalomorphs and araneomorphs analysed in this study, as providing no explanation for this very interesting trend would have left the readers wondering. We had listed all possible scenarios that explain the observed trend, including the ability to spin webs by the araneomorphs. However, we did clarify that this explanation lacks support at this stage since we have only analysed sequences from non-web-building RTA clade.

Line 371: “However, the current literature and our investigation are limited to the most diverse lineage in Araneomorphae – the RTA clade.”

We understand now that this can be confusing to the reader. Hence, we have completely modified the title of this section to “Distinct recruitment and diversification of spider venom superfamilies in Mygalomorphae and Araneomorphae”. Moreover, we have also analysed an additional toxin superfamily from web-building spiders (the only one in the literature for which there are sequences available), and the outcome of this analysis supports our hypothesis. However, since toxin sequences belonging to diverse DRP superfamilies from web-building araneomorphs (amino acids or nucleotides) have not been sequenced or deposited to date, this hypothesis has limited support at this stage. We have made this very clear in the revised version of the manuscript. We have also provided alternative explanations for observing this contrasting trend, and avoided highlighting the web-building vs non-web-building spider comparisons throughout the paper (now only discussed as one of the three hypotheses in the discussion).

3) Give more details in the Methods section to make your conclusions stronger and better supported.

We have now further expanded our methods section for better reproducibility and have included all the possible details.

4) Correct errors in phylogenetic nomenclature and usage of terminology (e.g. confusion between clade names and ancestors).

We thank the reviewers for their valuable suggestions regarding nomenclature. We apologise for the inconsistency in the usage of some of the terminologies. We have now made appropriate changes.

5) Explain the names and references from mythology and popular culture to make them clear to a culturally diverse readership.

We have now included a supplementary table describing the names and their cultural references (Supplementary File 1).

Reviewer #1 (Recommendations for the authors):

The manuscript is rich in cultural references and metaphors. This is not a problem on its own, but in many cases, the reason for the metaphor or cultural reference is unclear. I am fully in favor of authors bringing in their own cultural background, but it should be put in context for the reader. I assume the "ancestral knot" in the manuscript's title refers to knottin proteins, but this is not made clear, and I may be missing a deeper reference. The decision to name the ancestral protein "Adi Shakti" is confusing until one realizes that all the protein super-families were given names from Indian mythology (and also a few from other mythologies?) Are these names that were given by the authors as part of the preparation of this paper, or did some of them already exist? Is there a convention to name venom proteins after figures from Indian mythology? I don't know, and I assume the reader won't either. Please give this background explicitly somewhere. Similarly, is there a deeper context to "many ways to skin a cat"? There might be something I am missing, but this seems to me to be a metaphor out of place. This is a short paper, you can afford to expand a bit on some of the metaphors and references.

On a completely different note – while the results are interesting and important, there is a bit too much use of hyperbolic language. I suggest cutting a few "fascinating" and "astonishing".

Spider toxin superfamilies have been named after gods/deities of death, destruction and the underworld based on nomenclature introduced by Pineda et al. (2014 BMC genomics). This convention was also followed in subsequent papers introducing novel toxin superfamilies Pineda et al. (2020 PNAS). We have attempted to include names from diverse cultures, and not only named our new superfamilies after Hindu gods and goddesses, but also from Egyptian, Greek, and other mythologies. We have now included this explanation in the manuscript under the methods and Results sections. We have also provided additional details pertaining to this nomenclature in Supplementary File 1. We have also left out unclear references (e.g., many ways to skin a cat) and reinforced efforts to expand key references (such as: Adi Shakti, knottin, etc.) to make this manuscript more reader friendly.

Reviewer #2 (Recommendations for the authors):

This interesting study looks into the evolution of putative spider venom toxins, specifically disulfide-rich peptides. The results and interpretations presented are a promising start, but I don't consider them to be sufficiently convincing yet to warrant publication in eLife in their current state. Below I make suggestions for some further analyses that can make the conclusions more robust.

We thank the reviewer for their constructive inputs and suggestions. We hope that our responses have met the reviewers’ expectations.

The study builds upon previous work, especially references 9, 12, and 38. However, I think that the authors need to discuss more clearly how their results and conclusions relate to those reached in these papers. They should do this on several levels to enable readers, who are not spider venom experts, to better understand this contribution. For instance, on the most superficial level, it would be helpful to better explain that the strategy of naming toxin superfamilies for deities, which struck this reviewer as peculiar, was actually established in previous works.

We thank the reviewer for their suggestion. Novel Spider toxin superfamilies identified in this study were named after gods/deities of death, destruction and the underworld based on nomenclature as introduced by Pineda S, et al. (2014, BMC genomics; 2020, PNAS). We have now included a list of names with appropriate cultural references in the supplementary text (Supplementary File 1) and also clarified this nomenclature in the methods.

We have not tested the four-disulfide bond stabilised ancestral DRP hypothesis as theorised by Pineda et al. (2020 PNAS) and Cole and Brewer (2020 bioRxiv). It would be difficult to trace the exact nature and type of this ancestral scaffold given the long evolutionary history and their molecular divergence. Multiple events of duplication and diversification throughout their evolutionary history have resulted in events of gain and loss of cysteines and functional residues. As a result, we find diverse DRP scaffolds exhibiting a range of cysteine pairs – 2 to 6 disulfide bonds (i.e 4 to 12 cysteines) within the same spiders at times (e.g., an individual spider from the Haplopelma genus may contain SF90, SF91, SF99 and SF100 toxin superfamilies in its venom with 6, 8, 12 and 10 cysteines, respectively). In our opinion, this could be only identified by the examination of the genomic architectures (synteny) which is beyond the scope of this study. Our extensive phylogenetic and evolutionary analyses provide a deeper insight into the common origin of these DRP toxins in spiders dating as old as the split of Mygalomorphae and Araneomorphae, but we cannot comment on the nature and type of those scaffolds.

More importantly, clarification is needed about the sources of the sequences. The paper discusses the results from the perspective of spider venom. Yet, all the non-ctenid araneomorph data seem to be derived from whole-body transcriptomes generated by reference 38. Therefore, the results have to be presented with caution because the tissue source of the putative toxin sequences in these taxa is uncertain.

We would like to clarify that the non-ctenid araneomorph sequences utilised in this study were identified as putative toxins based on homology and DRP motif conservation (Cole and Brewer 2021, bioRxiv). This is further supported by our sequence alignments and phylogenetic analyses, wherein these toxin sequences are nested with annotated and experimentally characterised toxin sequences (Figures 3, 4 and respective figure supplements). We have also modified our methods section with regards to clarification of sequence data used. We previously included a supplementary dataset (now Figure 4 – Source data 2) with all sequences analysed in this study.

Moreover, it is difficult for me to understand how the authors identified new DRP superfamilies when the phylogenetic analyses that they provide do not seem to incorporate all previously published relevant data. For instance, of the 26 DRP superfamilies identified in reference 9, only 14 are included in the analyses presented here. This makes it impossible to assess how these relate to the new superfamilies reported here, or if some of the new superfamilies may be a subset of previously established superfamilies or vice versa. Why were these data excluded from the phylogenetic analyses? This is puzzling because they are included in the supplementary alignment S10. It makes sense to me to also include these excluded superfamilies in the phylogenetic analyses.

We had previously described the criteria that was used to define SFs in the Results section (line 85). We have relied on the signal and propeptide regions for identifying novel superfamilies, which is an accepted convention: Pineda et al. (2014 BMC Genomics); Pineda et al. (2020 PNAS). Evidence was further generated to support the monophyly of these toxin superfamilies using various means, including phylogenetics, sequence clustering and PCA. The excluded superfamilies do not have the same signal and propeptide regions as any other superfamily described here. Hence, it is just not possible that they are a subset of the superfamilies described here. In support of this, we previously included a supplemental alignment Figure S10 (now Figure 4 —figure supplement 2) of all toxin superfamily sequences analysed in this study.

Line 85 “Superfamilies (SF) of venom toxins in spiders have been classified based on their signal peptide and propeptide sequences (12)”

The reason for excluding certain previously described superfamilies from our phylogenetic analysis was the lack of nucleotide or amino acid sequence data. Adding in such highly divergent sequences, where there aren’t enough representatives, only adds phylogenetic noise and prevents the generation of robust phylogenies. As explained above, because of the differences in signal and propeptide regions, it is just not possible that the excluded superfamilies are a part of any of the superfamilies being described here. Therefore, adding them to our phylogenies would have only presented their relative position in the tree. We have, indeed, attempted to build such large phylogenies where every sequence was included. However, as explained above, because of the extreme sequence divergence in these toxins and the lack of enough representative sequences from some of the superfamilies, the topology of the tree was completely unresolved.

– I suggest removing the reference to Adi Shakti from the abstract or including an explanation because few non-Hindus will understand what this refers to.

While we completely agree that this may not be very clear to many, including many Hindus as it is a Sanskrit world, we feel that it is an important reference to the primordial origin of spider venoms. Following the reviewer’s suggestion, we have provided clarity on this in the abstract.

– This study attempts to trace the evolutionary origin of DRPs in a phylogenetically diverse clade with many spider genera and families. It would be a very helpful context for readers if the authors clarify on a schematic phylogeny of spiders exactly from which genera and families the data in their paper comes and which higher-level clades these represent. For instance, figure 2 includes genera from several araneomorph families. Which families? This gives readers a better sense of how comprehensively the analyses data represent spider phylogeny as a whole.

We thank the reviewer for this valuable suggestion. We have now introduced Figure 1 which depicts the schematic phylogeny of spiders and the sampled genera.

– Figures 1 and 2; please insert clade support values. Also please clarify if whole toxin sequences or just signal and propeptide sequences were used to construct the trees. Also, which of the included superfamilies are new? Indicate that in the tree or legend and amend the title. Also, where did the data for the taxon Pisaurina come from? It is not present in reference 38, which is mentioned as the source of the araneomorph sequences.

We have now made appropriate changes to the figure making it more informative. We have used the whole toxin sequence to construct phylogenies and this has been clearly described in the methods now. We did not add node support values to the trees as they are unreadable and uninformative. Instead, we have used thick dark (BPP>0.95; ML>90) and light grey (BPP<0.95; ML<90) branches to indicate node support. This is now clearly indicated in the legend of all phylogenetic figures.

We have retrieved Pisaurina DRP toxin sequences from the dataset by Cole and Brewer (2021, bioRxiv, Ref. 38). We request the reviewer to refer to Table 4 of that paper.

– Figure 3B: this alignment suggests that there are no gaps. Is that correct? If not, please indicate alignment gaps.

In this schematic representation (not an alignment), we intend to highlight the strongly conserved cysteine framework. The actual alignment can be viewed in Figure 4 —figure supplement 2. We have now further elaborated on this in the figure legend.

– Lines 90-93: it is more accurate to also mention that the cysteine framework was used to identify superfamilies as well.

We did not employ the cysteine framework to identify superfamilies. Superfamilies were identified using signal and propeptide as mentioned above. The criteria used to define SFs was previously described under the Results section (line 85). We have now further expanded on this to provide more clarification. The cysteine framework was rather used to highlight the homology across spider venom toxin superfamilies.

Line 85 “Superfamilies (SF) of venom toxins in spiders have been classified based on their signal peptide and propeptide sequences (12)”

– Lines 140-141: Opisthothelae denotes a clade comprising araneomorphs and mygalomorphs, not their common ancestor, so the analysis traces DRPs to the origin of this clade.

We thank the reviewer for pointing out this oversight. We have now corrected it to suborder Opisthothelae as the clade encompassing Mygalomorphae and Araneomorphae spiders.

– Line 145: 'speculate on'.

We have incorporated the suggested change.

– Lines 188-189: you refer to figure 3a when mentioning Adi Shakti. I don't understand how this relates to the figure. Do you want to call the ancestral cysteine scaffold Adi Shakti? If so, say this specifically. And also, what does this ancestral scaffold look like? References 9 and 38 reconstruct the ancestral DRP as having 8 cysteines. Do your analyses agree with this? Was this tested?

We apologise for the confusion caused. We confer the name ‘Adi Shakti’ to the ancestral DRP scaffold suggesting a deep-rooted origin. We have made appropriate changes to the text under the Discussion section to reflect this with more clarity:

Line 227 “Our findings strongly suggest a deep-rooted origin of DRP spider venom toxin superfamilies (Figure 4A), possibly from a single ancestral DRP or knottin scaffold, which we name ‘Adi Shakti’, after the original creator of the universe according to Hindu mythology”.

We have not tested the four disulfide bond stabilised ancestral DRP hypothesis as theorised by Pineda et al. (2020 PNAS). It would be difficult to trace the exact nature and type of this ancestral scaffold given the long evolutionary history and their molecular divergence. Multiple events of duplication and diversification throughout their evolutionary history have resulted in events of gain and loss of cysteines and functional residues. As a result, we find diverse DRP scaffolds exhibiting a range of cysteine pairs – 2 to 6 disulfide bonds (i.e., 4 to 12 cysteines) within the same spiders at times (e.g., an individual spider from the Haplopelma genus may contain SF90, SF91, SF99 and SF100 toxin superfamilies in its venom with 6, 8, 12 and 10 cysteines, respectively). In our opinion, this could be only identified by the examination of the genomic architectures (synteny) which is beyond the scope of this study. Our extensive phylogenetic and evolutionary analyses provide a deeper insight into the common origin of these DRP toxins in spiders dating as old as the split between Mygalomorphae and Araneomorphae, but we cannot comment on the nature and type of those scaffolds.

– Lines 189-192: this should be rephrased because there certainly are other spider venom toxins that do not belong to the DRP group.

We have now modified the statement appropriately to distinctly focus on DRP toxin superfamilies.

– Lines 204-205: rephrase. You have reconstructed the origin of DRPs in the last common ancestor of Opisthothelae, which is a clade, not an ancestor.

We thank the reviewer for pointing out this oversight. We have now corrected it to suborder Opisthothelae as the clade encompassing Mygalomorphae and Araneomorphae spiders.

Line 244: “… originated from an ancestral scaffold in Opisthothelae, the clade that encompasses Mygalomorphae and Araneomorphae spiders”.

– Line 208: positively selected sites or amino acids.

We thank the reviewer for their suggestion and have incorporated the suggested change.

Line 249: “Consistent with this hypothesis, the majority of positively selected sites (~96%) identified in spider venom DRP toxins …”

– Line 224: predominate in the venoms.

We have incorporated the suggested change and modified the statement accordingly.

– Lines 238-240 (also lines 128-130, 262-264): I am unconvinced by the arguments here and by the data claimed to support this. Genera and families are Linnean ranks of arbitrary phylogenetic depth. They do not meaningfully coincide with biologically relevant lineages.

We respectfully disagree with this argument. Certain adaptive traits (e.g., the origin of potent toxins or resistance to venom toxins) can drive speciation and diversification rates. Therefore, when adaptive traits are mapped onto species phylogenies, one can trace the origin of such traits and determine if the trait originated convergently or in the common ancestor. Mapping of a similar adaptive trait (resistance to cardiac glycosides) in various animal lineages demonstrated that, while some lineages have evolved and inherited this trait at the family level (e.g., Elapidae, Viperidae and Colubridae family of snakes), others (e.g., bufonid toads or hedgehogs) have independently acquired the resistant mutations at a genera-level, depending on their feeding ecology (Ujavari et al. 2015, PNAS; Mohammadi et al. 2016, Royal Society B).

Moreover, the taxa sampled only cover a small fraction of the phylogenetic breadth of spider lineages. As a result, interpretational difficulties emerge. For example, in Figure 1 some of the genera represent families (Trittame is the only taxon representing Barychelidae), and some genera are not split (Atrax and Hadronyche are colour coded the same, and together represent the family Atracidae), while in Figure 2 several of the supposed family-level taxa are only represented by a single genus, including Salticidae, Thomisidae, Lycosidae, and Homalonychidae. It, therefore, becomes impossible in many instances to distinguish whether toxin recruitments happened on the level of genera, families, or even higher-level taxa. To use Figure 1 to argue that mygalomorph toxin recruitments happened "post divergence of family members" (lines 263) is unconvincing since the included genera represent just three families that diverged between 150-100 million years ago (see Opatova et al., 2019, Syst. Biol. 69: 671). To get a more accurate estimate of when toxin families were recruited the authors would have to do ancestral state reconstruction on an established phylogeny.

We would like to point out that the phylogenies in Figures 2 and 3 highlight the differential recruitment events. Lines 118 and 143 state that this may not only be a result of recruitment and could arise from differential rates of diversification (also evident in other analyses presented in Figures 5 and Figure 5 – Source data 1).

Line 118 “Interestingly, the plesiotypic DRP scaffold seems to have undergone lineage-specific diversification in Mygalomorphae, where the selective diversification of the scaffold has led to the origination of novel toxin superfamilies corresponding to each genus (Figure 2).”

Line 135 “However, we also documented a large number of DRP toxins (n=32) that were found to have diversified in a family-specific manner, wherein, a toxin scaffold seems to be recruited at the level of the spider family, rather than the genus. As a result, and in contrast to mygalomorph DRPs, araneomorph toxin superfamilies were found to be scattered across spider lineages (Figure 3; Figure 3 —figure supplement 1; node support: ML: >90/100; BI: >0.95).”

Adding any number of missing lineages will neither change the fact that araneomorphs ‘appear’ to have recruited these superfamilies at the genera level, nor the family-level recruitment of toxin superfamilies in a large number of examined mygalomorphs. This outcome arises from the evident differences in rate evolution. We have now introduced a new figure (Figure 7) that highlights the different scenarios that explain the observed differences in the evolution of mygalomorph and araneomorph spider toxins. We have also expanded this section in the revised version of the manuscript.

To address the comment regarding the limited taxa investigated, we have now incorporated a new figure (Figure 1) depicting the spider phylogeny and the lineages analysed in this study. This figure demonstrates that we have representation from groups across the spider phylogeny.

We believe that it would be difficult to trace the exact nature and type of ancestral DRP scaffold given their long evolutionary history, riddled with events of duplication and diversification. Further difficulties identifying the exact ancestral knottin in spiders arise from the large diversity of DRP scaffolds (containing 4 to 12 cysteines) recovered from a single spider lineage, or even individuals (e.g., an individual spider from the Haplopelma genus may contain SF90, SF91, SF99 and SF100 toxin superfamilies in its venom with 6, 8, 12 and 10 cysteines, respectively). Synteny analyses, such as the ones we have performed in the past (Casewell et al. 2019, PNAS), can shed further light on this but are beyond the scope of this study. Our extensive phylogenetic and evolutionary analyses provide a deeper insight into the common origin of these DRP toxins in spiders, dating as old as the split of Mygalomorphae and Araneomorphae, but we cannot comment on the nature and type of those scaffolds We have now provided more clarification regarding conclusions from literature and their limitations in the revised version of our manuscript.

We have also modified the colour scheme in mygalomorph toxin SF phylogeny (Figure 2), depicting Atrax and Hardronyche with distinct colours.

– Lines 265-266: I don't understand this. Do the araneomorphs have fewer DRP superfamilies than the mygalomorphs?

In our statement “… their venom DRPs may have become relatively less diverse.”, we are referring to the sequence diversity, and not the number of toxin superfamilies in araneomorphs. Our findings suggest that all araneomorph DRP toxin superfamilies remain extremely well-conserved as opposed to the observed sequence diversity in mygalomorph DRP toxin superfamilies. As a result of such extreme sequence conservation, DRP motifs remain conserved within each araneomorph spider family. In contrast, because mygalomorph DRP toxins diversify under positive selection, a conserved motif can only be identified within a genus. We have now modified the statement to “… venom DRPs exhibit relatively lower sequence diversity” for clarity.

– Line 329: were the phylogenetic analyses done on the basis of nucleotide or amino acid data? Also, what substitution models were used to generate the trees, and how were these chosen?

We have utilised nucleotide data for both phylogenetic and evolutionary analyses. We have employed ModelFinder built within the IQ-TREE (v1.6.12) package to determine the best substitution model. All of this is now clearly explained in the manuscript.

– Line 393: correct spelling of Mygalomorphae.

We have incorporated the change suggested.

– Line 415: the dotted line is black.

We have now corrected the mistake.

– Figure 5: you need to reverse the colour labeling for 3FTXs and SVMPs, which should be blue and green, respectively, and you need to pair the 3FTXs with the elapids and the SVMPs with the vipers. Also, do the donut charts depict proteomic or transcriptomic data, and are they for specific species (if so, which?) or some sort of average (if so, based on what species and how were the proportions calculated?)?

We have now modified the colour scheme used to avoid confusion. These representative doughnut charts are based on our assembled proteomic data. It is well-known that Elapidae snake venoms are rich in 3FTx, while viper venoms are usually rich in SVMPs (Casewell 2020, Trends in pharmacological sciences; Sunagar 2013, Toxins; Laxme 2019, PNTD; Laxme 2021a, PNTD; Laxme 2021b, PNTD; Rashmi 2021, Journal of Proteomics). References and the source organisms are now cited in the figure legend to provide more clarity.

Moreover, in contrast to what you write in the legend and main text (e.g. lines 227-228), it seems that for each of the major groups of venomous snakes one or two types of toxins dominate venom composition as well, just like in the spiders. Please discuss this.

– lines 228-229: this is also true for spiders. Their venoms do not exclusively contain DRPs.

A single DRP scaffold (or its derivatives) dominates the venoms of most spiders around the world (King and Hardy 2013, Annual review; Lamxe et al. 2019, Toxicon). This is in contrast to snake venoms that contain multiple unrelated gene superfamilies (e.g., PLA2, 3FTX, SVMP as indicated in the figure; some snakes also have LAAO, Kunitz, Serine Protease, etc.). While spider venoms contain non-DRP components, such as amines, glutamate, protease, etc., they constitute only a very small proportion of their venom (<10%) (King and Hardy 2013, Annual review; Lamxe et al. 2019, Toxicon). We have modified the relevant text under the Discussion section to make this distinction clear.

Line 310: "Venoms of most other animals are, instead, composed of unrelated toxin types, derived from distinct scaffolds/gene superfamilies in varying proportions”.

– line 243 and further (and related text in the Results section): the contrast between mygalomorphs and araneomorphs in terms of the evolutionary selection pressures governing toxin evolution is indeed striking at first sight. However, before trying to explain these in general terms (lines 261 and further), the authors need to ask if these results can actually be generalized. 11 of the 12 identified superfamilies of mygalomorph toxins that evolve under the influence of positive selection are limited to just two genera: Hadronyche and Haplopelma. How representative are just these two genera for conclusions about mygalomorph evolution generally? Are the results not equally well explained (or better) in terms of the specific biology of these two taxa? Similarly, 16 of the 20 DRP superfamilies in the selection analyses for araneomorphs are restricted to just 1 araneomorph family (Ctenidae), which detracts from how general the results can be claimed to be.

We would like to point out that in our molecular evolutionary analyses, we find strong support for the influence of diversifying selection shaping the evolution of Mygalomorphae toxin superfamilies, despite them being separated by over 107 MYA (Barychelidae – Theraphosidae) to >150 MYA (Atracidae – Theraphosidae; Opatova et al. 2020, Systematic Biology 69.4). Similarly, we find evidence for signatures of purifying selection driving the evolution of Araneomorphae venom toxin superfamilies. We have now analysed additional datasets from Araneomorphae, including representatives from families Agelenidae, Pisauridae and Theridiidae, having diverged <70, <50 and <200 MYA respectively (Magalhaes 2020, Biological Reviews 95.1). We firmly believe that further addition of toxin sequence data from other groups will not deviate from the general trend of molecular evolution observed in both these lineages across such large period of time; barring certain certain exceptions (such as SF13 a defensive toxin identified from Hadronyche experiencing purifying selection; Volker, et al. 2020 PNAS).

Moreover, it is not clear to me why the selection analyses were performed on only a subset of the spider DRPs included in this study and previously published work. How were these superfamilies selected, and why were others excluded? Could the picture change if more superfamilies were analysed? This makes it impossible to judge how general the results may be as indicators for evolutionary mechanisms in the two clades.

We have, indeed, analysed all spider toxin sequences available to date. We have relied on the signal and propeptide regions for identifying novel superfamilies, which is an accepted convention: Pineda et al. (2014 BMC Genomics); Pineda et al. (2020 PNAS).

Although many additional superfamilies can be identified, we have only retained those sequences for which there were at least 5 representatives for the identification of toxin superfamilies, and 15 representatives for selection analyses to ensure robustness. This filtering step ensured that the generated alignments, phylogenetic trees and evolutionary assessments were robust and devoid of noise that stems from single-representative groups. Adding in those sequences would have enabled us to identify many more superfamilies, solely based on the signal and propeptide examination, but it wouldn’t have been possible to support them with other lines of evidence that were provided for all other superfamilies in this study, jeopardising the overall quality of the manuscript. Nonetheless, there is strong evidence that the left-out sequences are also related to the ones analysed in this study (Figure 4 —figure supplement 2). In future, when more transcriptomes are sequenced, it would be possible to designate these newer toxin superfamilies with much stronger support.

– Figure S2: I don't understand the significance of the 'novel' cysteines in green. Are these depicted separately because they are part of the scaffold? Please clarify. Also, are the names given to these new superfamilies, like Seker, the official new superfamily names you propose or are they there to enliven the trees? Please clarify. Also, make clear that these exclude the novel non-DRP toxins.

We thank the reviewer for their suggestion. We have now removed the reference to novel cysteines upon reviewers’ suggestion to avoid confusion. We have also modified the figure legend to clarify the exclusion of non-DRP toxins.

Novel Spider toxin superfamilies identified in this study were named after gods/deities of death, destruction and the underworld based on nomenclature as introduced by Pineda S, et al. (2014, BMC genomics; 2020, PNAS). We have now included a list of names with appropriate cultural references in the supplementary text (Supplementary File 1) and clarified this nomenclature in the methods.

– Supplementary figures 5, 6, and 9: please make sure that the published trees have legible clade support measures.

We thank the reviewer for their suggestion towards making figures SF 5, 6 and 9 legible and informative (now included as Figure supplement for Figure 2, 3 and 4 respectively). These phylogenies are large and dense. Hence it is difficult to incorporate more information into figures. Instead, we have highlighted the branches with Bayesian Posterior Probability lower than 0.95 in the Bayesian Inference tree and Bootstrap lower than 90 in the Maximum Likelihood tree in thin grey lines while those with higher support in thick black lines, respectively. This has also been made clear in the figure legend.

Reviewer #3 (Recommendations for the authors):

I greatly enjoyed reading your work and am grateful to be asked to review it. I hope my comments improve your already well-done study.

Please see the comments in my public review regarding method documentation. Perhaps a GitHub repository documenting all of the analyses and code along with a Figshare repository containing all of the data, results, and supplements would be enough.

We have made efforts to elaborate our methods section and additional details associated with each step are now provided.

I hope I did not misunderstand your use of only our ctenid data to represent araneomorphs. I hope the example datasets I provided are useful, and I apologize if my interpretation is incorrect.

We had initially excluded non-ctenid datasets from our analyses on account of poor sequence annotation and lack of enough representative sequences. However, we have now incorporated Dolomedes mizhoanus (DRP) (Jiang et al. 2013 Toxins) and Latrodectus tredecimguttatus (non-DRP) (He et al. 2013 PLoS ONE) toxin datasets into our analyses, following your suggestion, which has led to identification of 5 novel superfamilies, providing additional support to our spider venom evolution hypothesis.

Targeted comments:

The abstract states that "these lineages constitute over 50% of the extant spiders" following a discussion of the Mygalomorphae and Araneomorphae. These two lineages comprise far more than 50% of spider species. Are you referring to the Mygalomorphae + RTA-clade only?

Yes, we agree with the reviewer that Mygalomorphae and Araneomorphae together constitute about 99% of spider diversity described to date. In our statement “these lineages constitute over 50% of the extant spiders”, we restrict ourselves to the diversity in Mygalomorphae and RTA-clade alone that were the subject of our investigation – which wasn’t clear before. We have modified this statement to clarify the same as below.

As the lineages under evaluation constitute about 59% of the extant spiders, our findings provide fascinating insights into the early evolution and diversification of the spider venom arsenal.

Line 63 – "primitive Mesothelae species" Species are not primitive, but individual traits can be.

Apologies for the oversight. We have removed the word “primitive”.

Lines 118 and 124: "monophyletic clade" is redundant.

Apologies for the oversight. We have modified this as ‘monophyletic group’

Lines 236 – 241: Is the comparison of taxonomic ranks employed here fair? Are spider families and genera separated by 350 million years assumed to be comparable? Perhaps evolutionary ages or relative species diversity within the ranks could standardize the comparisons. I am not sure how best to approach this, but it stands out to me as potentially problematic, but I am not sure it greatly impacts your overall conclusions.

We would like to clarify that certain adaptive traits (e.g., the origin of potent toxins or resistance to venom toxins) can drive speciation and diversification rates. Therefore, when adaptive traits are mapped onto species phylogenies, one can trace the origin of adaptive traits and determine if the adaptive trait is common or independent in origin. Mapping of a similar adaptive trait (resistance to cardiac glycosides) in various animal lineages demonstrated that, while some lineages have evolved and inherited this trait at the family level (e.g., Elapidae, Viperidae and Colubridae family of snakes), others (e.g., bufonid toads or hedgehogs) have independently acquired the resistant mutations at a genera-level, depending on their feeding ecology (Ujavari et al. 2015, PNAS; Mohammadi et al. 2016, Royal Society B). Based on the similar premise we can confidently trace the recruitment of toxin SFs at taxon level. We have now also provided more explanation on all possible scenarios leading to differential recruitment.

Regarding the reviewer’s suggestion pertaining to species diversity influencing recruitment of toxin SFs, we believe that this may or may not correlate with the rate of venom diversification. For instance, we find lower levels of toxin sequence diversity in the RTA clade, despite them being speciose and possessing higher genetic diversification rates (Fernández 2018, Current Biology).

Lines 264 – 266: While it is true that many araneomorphs rely on prey capture webs, none of the species included in this work do. Making these claims without including species that do utilize prey-capture webs is tenuous.

We were attempting to explain the contrasting rates of diversification of toxin superfamilies in mygalomorphs and araneomorphs analysed in this study, as providing no explanation for this very interesting and contrasting trend would have left the readers wondering. We had listed all possible scenarios that explain the observed trend, including the ability to spin webs by the araneomorphs. However, we did clarify that this explanation lacks support at this stage since we have only analysed sequences from non-web-building RTA clade.

Line 371: “However, the current literature and our investigation are limited to the most diverse lineage in Araneomorphae – the RTA clade.

We understand now that this can be confusing to the reader. Hence, we have completely modified the title of this section to “Distinct recruitment and diversification of spider venom superfamilies in Mygalomorphae and Araneomorphae”. Moreover, we have also analysed an additional toxin superfamily from web-building spiders (the only one in the literature for which there are sequences available), and the outcome of this analysis supports our hypothesis. However, since toxin sequences belonging to diverse DRP superfamilies from web-building araneomorphs (amino acids or nucleotides) have not been sequenced or deposited to date, this hypothesis has limited support at this stage. We have made this very clear in the revised version of the manuscript.

We have now also included a new figure (Figure 7) and have provided explanations for possible scenarios leading to differential recruitment of toxins in spiders under the Discussion section.

I greatly enjoyed this work and think it will advance the field forward. I am honored that you utilized the data presented in my former PhD student's and my work so prominently and appreciate the novel directions you took with them. I hope my comments are useful and look forward to seeing a future version of the manuscript.

We thank the reviewer for their very kind and encouraging comments. Their extremely constructive feedback has definitely added value to our publication. Overall, this has been a wonderful experience in revising our manuscript.

References

1. Pineda, S. S., Sollod, B. L., Wilson, D., Darling, A., Sunagar, K., Undheim, E. A., … and King, G. F. (2014). Diversification of a single ancestral gene into a successful toxin superfamily in highly venomous Australian funnel-web spiders. BMC genomics, 15(1), 1-16.

2. Pineda, S. S., Chin, Y. K. Y., Undheim, E. A., Senff, S., Mobli, M., Dauly, C., … and King, G. F. (2020). Structural venomics reveals evolution of a complex venom by duplication and diversification of an ancient peptide-encoding gene. Proceedings of the National Academy of Sciences, 117(21), 11399-11408.

3. Sunagar, K., Jackson, T. N., Undheim, E. A., Ali, S. A., Antunes, A., and Fry, B. G. (2013). Three-fingered RAVERs: Rapid Accumulation of Variations in Exposed Residues of snake venom toxins. Toxins, 5(11), 2172-2208.

4. Cole, T. J., and Brewer, M. S. (2021). Killer Knots: Molecular evolution of Inhibitor Cystine Knot toxins in wandering spiders (Araneae: Ctenidae). bioRxiv.

5. Herzig, V., Sunagar, K., Wilson, D. T., Pineda, S. S., Israel, M. R., Dutertre, S., … and Fry, B. G. (2020). Australian funnel-web spiders evolved human-lethal δ-hexatoxins for defense against vertebrate predators. Proceedings of the National Academy of Sciences, 117(40), 24920-24928.

6. Sunagar, K., and Moran, Y. (2015). The rise and fall of an evolutionary innovation: contrasting strategies of venom evolution in ancient and young animals. PLoS genetics, 11(10), e1005596.

7. Sunagar, K., Johnson, W. E., O’Brien, S. J., Vasconcelos, V., and Antunes, A. (2012). Evolution of CRISPs associated with toxicoferan-reptilian venom and mammalian reproduction. Molecular biology and evolution, 29(7), 1807-1822.

8. Yang, Z. (2007). PAML 4: phylogenetic analysis by maximum likelihood. Molecular biology and evolution, 24(8), 1586-1591.

9. Ujvari, B., Casewell, N. R., Sunagar, K., Arbuckle, K., Wüster, W., Lo, N., … and Madsen, T. (2015). Widespread convergence in toxin resistance by predictable molecular evolution. Proceedings of the National Academy of Sciences, 112(38), 11911-11916.

10. Mohammadi, S., Gompert, Z., Gonzalez, J., Takeuchi, H., Mori, A., and Savitzky, A. H. (2016). Toxin-resistant isoforms of Na+/K+-ATPase in snakes do not closely track dietary specialization on toads. Proceedings of the Royal Society B: Biological Sciences, 283(1842), 20162111.

11. Casewell, N. R., Petras, D., Card, D. C., Suranse, V., Mychajliw, A. M., Richards, D., … and Turvey, S. T. (2019). Solenodon genome reveals convergent evolution of venom in eulipotyphlan mammals. Proceedings of the National Academy of Sciences, 116(51), 25745-25755.

12. Casewell, N. R., Jackson, T. N., Laustsen, A. H., and Sunagar, K. (2020). Causes and consequences of snake venom variation. Trends in pharmacological sciences, 41(8), 570-581.

13. Laxme, R. S., Khochare, S., de Souza, H. F., Ahuja, B., Suranse, V., Martin, G., … and Sunagar, K. (2019). Beyond the ‘big four’: Venom profiling of the medically important yet neglected Indian snakes reveals disturbing antivenom deficiencies. PLoS neglected tropical diseases, 13(12), e0007899.

14. Senji Laxme, R. R., Attarde, S., Khochare, S., Suranse, V., Martin, G., Casewell, N. R., … and Sunagar, K. (2021). Biogeographical venom variation in the Indian spectacled cobra (Naja naja) underscores the pressing need for pan-India efficacious snakebite therapy. PLoS neglected tropical diseases, 15(2), e0009150.

15. Senji Laxme, R. R., Khochare, S., Attarde, S., Suranse, V., Iyer, A., Casewell, N. R., … and Sunagar, K. (2021). Biogeographic venom variation in Russell’s viper (Daboia russelii) and the preclinical inefficacy of antivenom therapy in snakebite hotspots. PLoS neglected tropical diseases, 15(3), e0009247.

16. Rashmi, U., Khochare, S., Attarde, S., Laxme, R. S., Suranse, V., Martin, G., and Sunagar, K. (2021). Remarkable intrapopulation venom variability in the monocellate cobra (Naja kaouthia) unveils neglected aspects of India's snakebite problem. Journal of Proteomics, 242, 104256.

17. King, G. F., and Hardy, M. C. (2013). Spider-venom peptides: structure, pharmacology, and potential for control of insect pests. Annual review of entomology, 58, 475-496.

18. Laxme, R. S., Suranse, V., and Sunagar, K. (2019). Arthropod venoms: Biochemistry, ecology and evolution. Toxicon, 158, 84-103.

19. Opatova, V., Hamilton, C. A., Hedin, M., De Oca, L. M., Král, J., and Bond, J. E. (2020). Phylogenetic systematics and evolution of the spider infraorder Mygalomorphae using genomic scale data. Systematic Biology, 69(4), 671-707.

20. Magalhaes, I. L., Azevedo, G. H., Michalik, P., and Ramírez, M. J. (2020). The fossil record of spiders revisited: implications for calibrating trees and evidence for a major faunal turnover since the Mesozoic. Biological Reviews, 95(1), 184-217.

21. Jiang, L., Liu, C., Duan, Z., Deng, M., Tang, X., and Liang, S. (2013). Transcriptome analysis of venom glands from a single fishing spider Dolomedes mizhoanus. Toxicon, 73, 23-32.

22. He, Q., Duan, Z., Yu, Y., Liu, Z., Liu, Z., and Liang, S. (2013). The venom gland transcriptome of Latrodectus tredecimguttatus revealed by deep sequencing and cDNA library analysis. PLoS One, 8(11), e81357.

23. Fernández, R., Kallal, R. J., Dimitrov, D., Ballesteros, J. A., Arnedo, M. A., Giribet, G., and Hormiga, G. (2018). Phylogenomics, diversification dynamics, and comparative transcriptomics across the spider tree of life. Current Biology, 28(9), 1489-1497.

https://doi.org/10.7554/eLife.83761.sa2

Article and author information

Author details

  1. Naeem Yusuf Shaikh

    Evolutionary Venomics Lab, Centre for Ecological Sciences, Indian Institute of Science Bangalore, Bengaluru, India
    Contribution
    Conceptualization, Data curation, Formal analysis, Investigation, Visualization, Methodology, Writing - original draft
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-9903-8484
  2. Kartik Sunagar

    Evolutionary Venomics Lab, Centre for Ecological Sciences, Indian Institute of Science Bangalore, Bengaluru, India
    Contribution
    Conceptualization, Resources, Formal analysis, Supervision, Funding acquisition, Validation, Investigation, Visualization, Methodology, Writing - original draft, Project administration, Writing - review and editing
    For correspondence
    ksunagar@iisc.ac.in
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-0998-1581

Funding

The Wellcome Trust DBT India Alliance (IA/I/19/2/504647)

  • Kartik Sunagar

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication. For the purpose of Open Access, the authors have applied a CC BY public copyright license to any Author Accepted Manuscript version arising from this submission.

Acknowledgements

This work was supported by the DBT/Wellcome Trust India Alliance Fellowship [grant number IA/I/19/2/504647] awarded to KS. NYS is thankful to Vivek Suranse and Senji Laxme R R (Indian Institute of Science) for insightful discussions.

Senior Editor

  1. Christian R Landry, Université Laval, Canada

Reviewing Editor

  1. Ariel Chipman, The Hebrew University of Jerusalem, Israel

Reviewers

  1. Ariel Chipman, The Hebrew University of Jerusalem, Israel
  2. Michael Brewer, East Carolina University, United States

Version history

  1. Received: September 28, 2022
  2. Preprint posted: October 7, 2022 (view preprint)
  3. Accepted: February 8, 2023
  4. Accepted Manuscript published: February 9, 2023 (version 1)
  5. Version of Record published: March 15, 2023 (version 2)

Copyright

© 2023, Shaikh and Sunagar

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 1,877
    Page views
  • 331
    Downloads
  • 0
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Naeem Yusuf Shaikh
  2. Kartik Sunagar
(2023)
The deep-rooted origin of disulfide-rich spider venom toxins
eLife 12:e83761.
https://doi.org/10.7554/eLife.83761

Further reading

    1. Ecology
    2. Genetics and Genomics
    Franziska Grathwol, Christian Roos ... Gisela H Kopp
    Research Advance

    Adulis, located on the Red Sea coast in present-day Eritrea, was a bustling trading centre between the first and seventh centuries CE. Several classical geographers--Agatharchides of Cnidus, Pliny the Elder, Strabo-noted the value of Adulis to Greco--Roman Egypt, particularly as an emporium for living animals, including baboons (Papio spp.). Though fragmentary, these accounts predict the Adulite origins of mummified baboons in Ptolemaic catacombs, while inviting questions on the geoprovenance of older (Late Period) baboons recovered from Gabbanat el-Qurud ('Valley of the Monkeys'), Egypt. Dated to ca. 800-540 BCE, these animals could extend the antiquity of Egyptian-Adulite trade by as much as five centuries. Previously, Dominy et al. (2020) used stable istope analysis to show that two New Kingdom specimens of P. hamadryas originate from the Horn of Africa. Here, we report the complete mitochondrial genomes from a mummified baboon from Gabbanat el-Qurud and 14 museum specimens with known provenance together with published georeferenced mitochondrial sequence data. Phylogenetic assignment connects the mummified baboon to modern populations of Papio hamadryas in Eritrea, Ethiopia, and eastern Sudan. This result, assuming geographical stability of phylogenetic clades, corroborates Greco-Roman historiographies by pointing toward present-day Eritrea, and by extension Adulis, as a source of baboons for Late Period Egyptians. It also establishes geographic continuity with baboons from the fabled Land of Punt (Dominy et al., 2020), giving weight to speculation that Punt and Adulis were essentially the same trading centres separated by a thousand years of history.

    1. Ecology
    Jingxuan Li, Chunlan Yang ... Zhong Wei
    Research Article Updated

    While bacterial diversity is beneficial for the functioning of rhizosphere microbiomes, multi-species bioinoculants often fail to promote plant growth. One potential reason for this is that competition between different species of inoculated consortia members creates conflicts for their survival and functioning. To circumvent this, we used transposon insertion mutagenesis to increase the functional diversity within Bacillus amyloliquefaciens bacterial species and tested if we could improve plant growth promotion by assembling consortia of highly clonal but phenotypically dissimilar mutants. While most insertion mutations were harmful, some significantly improved B. amyloliquefaciens plant growth promotion traits relative to the wild-type strain. Eight phenotypically distinct mutants were selected to test if their functioning could be improved by applying them as multifunctional consortia. We found that B. amyloliquefaciens consortium richness correlated positively with plant root colonization and protection from Ralstonia solanacearum phytopathogenic bacterium. Crucially, 8-mutant consortium consisting of phenotypically dissimilar mutants performed better than randomly assembled 8-mutant consortia, suggesting that improvements were likely driven by consortia multifunctionality instead of consortia richness. Together, our results suggest that increasing intra-species phenotypic diversity could be an effective way to improve probiotic consortium functioning and plant growth promotion in agricultural systems.