Transcriptome analysis of 454-seq data and assemblage of transcript A, B, C1, C2 and D families

All so far on proteomic site identified cupiennins and CsTx-16¹⁵ were used to search with BlastP (E-threshold, 0.0001) against the above mentioned 454-transcriptome (454-seq). Firstly, the obtained contigs were analyzed and sorted in terms of identical peptides, belonging to different cupiennin families. Secondly, a further sorting was done by taking specific linkers, which are in N- or C-terminal position to the peptides, into account. A linker sequence is characterized N-terminally by a PQM and C-terminally by an iPQM motif. With the obtained specific N-terminal- or C-terminal linker information of the peptides, peptide chains were assembled and elongated in both directions until N-terminally the signal peptide or C-terminally a stop codon was identified.

For the assemblage of the transcript families A, B, C1, C2, and D (see ^S1A-E Fig), only transcriptomic data from 454-seq with the following conditions were used: every peptide/linker unit had to be identified at least in two contigs, composed at least of five reads. Mainly completely identified peptide sequences, built at least of a linker/peptide/linker or a peptide/linker/peptide structure, were used for overlapping elongation of the protein sequence. Counting of the identified peptides at a certain position in the transcript families was done by read counts of the corresponding contigs. For the determination of the content of individual peptides, reads of only fully premature sequences including a C-terminal glycine for amidation, which is characteristic for several cupiennin families (^S1 Table), were counted (Fig. 1, ^S1A-E Fig). The transcriptomic data analysis workflow, and especially the manual construction of the transcript families, is given in detail in (^S5 Fig).

All known LP sequences obtained earlier by Edman-sequencing of purified peptides from C. salei venom and of 454-seq identified contigs, encoding signal peptides, propeptides and LPs including the linker regions, were used as query and blasted against the Illumina 3000-seq venom gland transcriptomes of C. salei and C. getazi. Analysis of the obtained contigs were focused on the identification of identical, related and new LPs.

LPs, which were accepted as identical or related peptides had to be identified in one contig exhibiting full N-terminal and C-terminal peptides linkers, or at least in two contigs, exhibiting at least 12 bps of both linkers (iPQM and PQM motif). Complete new peptides were characterized by full N-terminal and C-terminal linkers and at least of two contigs. Additionally, all transcriptomes were blasted against a set of so far published small LPs identified in zodariids, oxyopids and lycosids (S5 Fig).

All identified LPs from the Illumina3000 transcriptome of C. getazi were sorted to the corresponding cupiennin families created for C. salei. Alignments of peptide, signal and propeptide sequences belonging to different cupiennin families of C. salei and C. getazi were done by Clustal omega (www.ebi.ac.uk)⁴². New peptide families were created for peptides which differ in their biochemical characteristics (^S1 Table). In a last step, all obtained peptide sequences from C. salei and C. getazi were used as query and blasted against the UniProtKB database.

The data for this study have been deposited in the European Nucleotide Archive (ENA) at EMBL-EBI under accession number PRJEB42022 (https://www.ebi.ac.uk/ena/browser/view/PRJEB42022). Accession numbers as well as the characteristics of the deposited cDNA sequences (N = 283) are summarized in ^S2Table.

Signal peptides were identified using SignalP v. 5.0⁴³, biochemical characterization of the peptides was done with Expasy/ProtParam⁴⁴, peptide secondary structure prediction with the GOR method⁴⁵, peptide logos were generated with WebLogo (Vers. 2.8.2)²⁶, and the visualization of the α-helical peptides as presented in²⁷.

Cupiennins 3a, 3b, 4a, 6a, 6b, 6d, 7b, 8, and the two-chain CsTx-16a1b1, and CsTx-16c1d1 were synthesized (all C-terminally amidated) by GeneCust, (Dulange, Luxembourg). Bioassays using Drosophila flies were performed as described elsewhere²⁵. Briefly, all peptides were dissolved in 0.1 M ammonium acetate, pH 6.1 and 0.05 µl injected intrathoracically into the flight muscles of female flies (1- to 7-days old). For each peptide, in minimum five different peptide doses were injection into the flies (each n = 20) and also a negative control with only buffer was injected (n = 10). Calculations of the lethal dose (LD₅₀) were done with GraphPad PRISM Vers. 6.07 (GraphPad Software, San Diego, CA, USA).