The venom gland transcriptome of Latrodectus tredecimguttatus revealed by deep sequencing and cDNA library analysis

Abstract

Latrodectus tredecimguttatus, commonly known as black widow spider, is well known for its dangerous bite. Although its venom has been characterized extensively, some fundamental questions about its molecular composition remain unanswered. The limited transcriptome and genome data available prevent further understanding of spider venom at the molecular level. In the present study, we combined next-generation sequencing and conventional DNA sequencing to construct a venom gland transcriptome of the spider L. tredecimguttatus, which resulted in the identification of 9,666 and 480 high-confidence proteins among 34,334 de novo sequences and 1,024 cDNA sequences, respectively, by assembly, translation, filtering, quantification and annotation. Extensive functional analyses of these proteins indicated that mRNAs involved in RNA transport and spliceosome, protein translation, processing and transport were highly enriched in the venom gland, which is consistent with the specific function of venom glands, namely the production of toxins. Furthermore, we identified 146 toxin-like proteins forming 12 families, including 6 new families in this spider in which α-LTX-Lt1a family2 is firstly identified as a subfamily of α-LTX-Lt1a family. The toxins were classified according to their bioactivities into five categories that functioned in a coordinate way. Few ion channels were expressed in venom gland cells, suggesting a possible mechanism of protection from the attack of their own toxins. The present study provides a gland transcriptome profile and extends our understanding of the toxinome of spiders and coordination mechanism for toxin production in protein expression quantity.

Figure 1. The pipeline of data process and analysis.

Figure 2. Examples of new toxins.

A) Sequence alignment of seven members of the α-LTX-Ltla-1/2 families. P23631 is the Uniprot ID for α-LTX-Ltla-2, a well-known toxin of Latrodectus tredecimguttatus. The other six proteins are new potential toxins found in our core dataset. Amino acid residue point mutations are marked in green; Residues conserved across two families are marked in red. The two families are indicated by pink and cyan backgrounds. B) The secondary structure of three new toxins (P_206187, EST_P_221, P_141871, EST_P_151 and P_208737, members of the ctenitoxin family) is shown. The amino acids forming an alpha helix are colored in blue; red rectangles indicate predicted signal peptides; purple lines represent disulfide bridges.

Figure 3. Statistical analysis of BLAST searches.

A) Dot plot of ML/BL vs. identity of the BLAST queries, which searched against known spider EST datasets with assembled cDNA sequences (see methods). B) Length distribution of transcripts shared by the de novo assembly and EST sequencing datasets. C) Statistical analysis of protein homologues identified by a BLAST search against the Uniprot database performed with all translated protein sequences. Figures C1, C5 and C9 are bar charts for ML/BL, ML/PL and identity distribution. Other dot plots represent the pairwise correlations among them.

Figure 4. The RPKM distribution of transcripts in different categories and three GO namespaces.

For each GO category, the sum of RPKMs, protein number and RPKM average were calculated and shown as blue, red and green bars, respectively.

Figure 5. Cluster of the toxinome of Latrodectus tredecimguttatus.

Figure 6. Domain architecture of toxins.

Domain architectures of toxins were predicted by the SMART and Pfam servers (http://smart.embl.de and http://pfam.janelia.org) [59,60]. All toxins were grouped based on the respective family classification. The character “-F” appended to protein ID numbers indicates that these sequences are fragments and not a full-length protein. The character “EST_” appended to protein ID numbers indicates that these sequences were extracted from the EST dataset. Proteins sharing the same domain architecture were combined. The members of the Ank superfamily and related legends are listed in the middle figure. For the orphan family, multiple sequence alignments of potential toxins and known toxins (green labels) are shown and the matched cysteine domains are indicated. The abbreviations of domain names are as follows: ankyrin repeats (ANK, SMART ID: SM00248); TY (SMART ID: SM00211); trypsin (Pfam ID: PF00089); SCP (SMART ID: SM00198); EGF (SMART ID: SM00181); RHO (SMART ID: SM00174); toxin 35 (Pfam ID: PF10530).

Figure 7. Evolutional relationship of species and toxins.

A. Phylogenetic tree of the 26 species. It was constructed by MAGE software using sequences of single-copy nuclear protein-coding genes from these species. Latrodectus tredecimguttatus was grouped with other arthropods (insects) and highlighted in red. B. Phylogenies of Arthropod. Latrodectus tredecimguttatus was clustered with other three spider species (Theraphosidae, Amblypygi and Schizomidae) in Arachnida, which are marked with different colors: purple, Hexapoda; brown, Crustacea; silver gray, Oligostraca; orange, Myriapoda; blue, Pycnogonida; The full names of the species can be found in Tables S7 and S8 File S1. C. Phylogenies of the ANK superfamily, in which pink indicates α-LTX-Lt1a family1; blue indicates α-LTX-Lt1a family2; green indicates δ-LIT-Lt1a family; red indicates α-LIT-Lt1a family; brown indicates ANK family.