Spider genomes provide insight into composition and evolution of venom and silk

Abstract

Spiders are ecologically important predators with complex venom and extraordinarily tough silk that enables capture of large prey. Here we present the assembled genome of the social velvet spider and a draft assembly of the tarantula genome that represent two major taxonomic groups of spiders. The spider genomes are large with short exons and long introns, reminiscent of mammalian genomes. Phylogenetic analyses place spiders and ticks as sister groups supporting polyphyly of the Acari. Complex sets of venom and silk genes/proteins are identified. We find that venom genes evolved by sequential duplication, and that the toxic effect of venom is most likely activated by proteases present in the venom. The set of silk genes reveals a highly dynamic gene evolution, new types of silk genes and proteins, and a novel use of aciniform silk. These insights create new opportunities for pharmacological applications of venom and biomaterial applications of silk.

Figure 1. Proteomics.

Venn diagrams with number of identified proteins based on 194 LC-MS/MS analyses (Supplementary Table 3). The silk analyses are based on in-solution trypsin digestion, while the body analyses are based on in-gel trypsin digestion. The venom analyses are based on a combination of the two methods. (a) Velvet spider. (b) Tarantula.

Figure 2. Comparative genomics.

(a) Phylogenetic tree based on the amino-acid sequence of 452 one-to-one orthologous genes in eight species. (b) Estimated divergence times using a relaxed molecular clock and fossil calibration time ranges. Red bars are 95% credibility intervals. (c) The 1-1 orthologous genes between the two spider species. (d) The number of gene families shared among the species using the TreeFam classification scheme among spiders, mites, ticks and insects.

Figure 3. Venomics.

(a) Coomassie-blue-stained SDS-gel of venom from tarantula and velvet spider, respectively, (Supplementary Fig. 16). The major protein(s) in the bands are: T1, membrane venom metalloendopeptidase-a; T2, putative cysteine-rich venom protease; T3, genicutoxin A1; V1, membrane venom metalloendopeptidase and venom aminopeptidase-a; V2, venom pancreatic-like triacylglycerol lipase-a and c; V3, cysteine-rich venom protease-a; V4, venom phospholipase A2-a; and V5, S.m. Sp2b. In addition to S.m. Sp2b, the V5 band also contains several protoxins. The composition of the visual blue bands was specifically interrogated using the relevant LC-MS/MS analyses of bands from the gels shown in Supplementary Figs 4 and 5. The data are a sub-fraction of the data shown in Supplementary Data 1 and 2, where the merged result of LC-MS/MS analyses of gel bands from a complete gel lane is presented. (b) The table summarizes quantitative analyses of venom proteins, excluding protoxins (mainly present in the lower bands on the gel in Fig. 3a). The reason to exclude the protoxins from this analysis is described in the Supplementary Note 2. The proteases are shown in red, the lipases in grey and the other proteins in shades of blue. Numbers in parentheses refer to the number of variants of the particular protein that were quantified. Individual proteins constituting <1% of the venom protein content are not included. These quantitative LC-MS/MS analyses are based on a gel-free approach and extracted ion chromatography (XIC). The table is an extract of the full quantitative analyses (Supplementary Data 7 and 8). (c) The genomic localization of the protoxin families in the velvet spider, the stegotoxins, is depicted. The letters (A–F) indicate the family, based on sequence similarities, and the numbers distinguish between the different toxins in the same family (Supplementary Note 2). The ‘A-homologue’ refers to a sequence homologous with the A-family of toxins, but without proteomic support. The arrows indicate the direction of transcription. Introns are shown in red and coding sequences in black. In the A-family cluster, two non-related predicted protein-coding genes are present. These are shown as blue rectangles. The figure shows that toxins with sequence similarities cluster on the genome.

Figure 4. Silk genes in the velvet spider.

(a) Schematic overview of exon-intron and repeat structure of the identified silk genes. The repeats from a single locus share evolutionary history and are easy to align, except for S.m. Sp2c that consists of two sets of repeats (grey and blue) with different evolutionary histories. S.m. PiSp consists of sequences obtained from two scaffolds. PCR verified that they come from the same locus. (b) The evolutionary history of the major ampullate C-terminal domain, including the recent events of gene conversion, whole-gene duplication (~10 MYA) and pseudogenization* (Supplementary Note 3). Scale bar represents number of nucleotide differences per site. (c) Major ampullate repeat evolution over ~10 MY since a whole-gene duplication. The evolutionary relationship of repeats from two recently duplicated loci is depicted by a phylogenetic tree and repeat structures. Similar coloured repeats are evolutionarily most related. Scale bar represents the number of nucleotide differences per site. (d) Use of aciniform silk in the velvet spider compared with previously studied species that use aciniform silk to wrap their prey. Velvet spiders do not wrap their prey, but use aciniform silk for dragline silk in combination with major ampullate silk. Velvet spiders also use aciniform silk for egg case construction as do previously studied species.