Sequence Analysis

A schematic of the bioinformatics workflow is presented in Figure 2. A Tachyglossus aculeatus venom gland transcriptome was sequenced on an Illumina GAIIx instrument using previously describe methods for sample preparation and sequencing [7]. The venom gland was kindly provided by Frank Grutzner under University of Adelaide Animal Ethics Committee project number S-032-2008. The sequence reads have been deposited under accession number SRP027593 in the SRA database at NCBI. Quality filtered reads were assembled with the Velvet-Oases pipeline (kmer length = 31bp and –ins_length = 260) [12], [13]. To improve the transcriptome assembly, the Scaffolding using Translation Mapping (STM) strategy [14] was used to scaffold Oases contigs using Ensembl predicted proteins, an experimentally derived dataset of platypus venom proteins [5], [7], [15], and all human Refseq proteins as references. Data available from the Dryad Digital Repository: http://doi.org/10.5061/dryad.4qq0v. BLAST [16] searches were performed against a database containing all Tox-Prot proteins [17], additional platypus venom toxins that have not been yet included in the Tox-Prot database [7] and human Refseq sequences using scaffolds derived from STM to identify possible homologies between echidna and other venom toxin proteins.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. Schematic of bioinformatics workflow for the assembly, annotation and quantitation of echidna transcripts.

https://doi.org/10.1371/journal.pone.0079092.g002

Expression Quantification and Comparison

Expression levels for each scaffold were derived by first mapping the raw sequencing reads to transcript scaffolds using Bowtie [18]. We then generated digital counts for each scaffold by counting reads that map back to transcripts. Multi-mapping reads were discarded. To compare the most highly expressed proteins between echidna and platypus venom, we ranked the expression values (normalized using Reads Per Kilobase per Million (RPKM) [19]) of the echidna transcripts, from highest to lowest, and compared it with similarly ranked list from a platypus transcriptomic dataset described in [20]. Echidna transcripts were annotated with respect to both human and platypus genes. The top 100 genes from both lists were compared and functional enrichment analysis was performed using Ontologizer [21] with orthologous human gene names as input. GoStat2 [22] was used to compare overrepresented ontology terms from highly expressed echidna genes versus the same in platypus. False discovery rate (FDR) [23] corrections were used for all functional term enrichment analyses. Ranked gene lists for comparisons only included genes that map to a homologous protein and did not include unnamed genes. If multiple scaffolds map to the same gene name, the name is only included once in the list.

Tox-Prot Comparison

We compared our scaffolds to sequences in a toxin database, Tox-Prot, to identify putative venom proteins based on homology. As venom proteins have typically evolved from body proteins, to distinguish between venom proteins and related body proteins we also searched against the human Refseq sequences. Sequences were considered as possible toxins if they were more similar to known toxins than a homologous protein from human. 30 transcripts from 15 genetic regions matched 12 unique toxin sequences with a lower BLAST e-value than to a human protein (Table 1). Five transcripts of proteins found in platypus venom were identified: CD55, amide oxidase, corticotrophin-releasing factor binding protein, a Kunitz domain-containing protease inhibitor and a cysteine-rich secretory protein (Table 2). Echidna sequences also shared similarity with a subunit stonefish toxin with haemolytic activities [24], snake nerve growth factors, protease inhibitors from snake and sea anemone venom with ion channel inhibitory activity (sea anemone) [25], and myotoxic and neurotoxic snake phospholipase A2s [26], [27].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Echidna transcripts that display higher similarity to known toxins than human proteins.

https://doi.org/10.1371/journal.pone.0079092.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Known platypus venom proteins and their homologs in the echidna transcriptome.

https://doi.org/10.1371/journal.pone.0079092.t002

We further examined the expression levels of these echidna transcripts to gain insights into their putative function. We found that all identified echidna protein-coding transcripts with homology to known toxins were lowly expressed in the transcriptome (Table 2). Pharmacologically characterized venom proteins have been found to be some of the most highly expressed genes in the platypus venom gland transcriptome [7], [20]. Using this logic the low expression of echidna toxin transcripts suggests that venom production is not a key role of this gland.

Analysis of most Highly Expressed Genes

To provide insight into the function of echidna venom, we examine the most highly expressed genes in the echidna venom gland. Quantification of expression levels involved the counting of raw reads that map to assembled scaffolds. The numbers were normalized with respect to the length of the scaffolds. This method allowed us to examine the most highly expressed genes and allowed comparison with similar experiments in platypus. None of the top 50 most highly expressed echidna genes were found in the top 50 of platypus. Similar results were obtained when human protein-coding transcripts instead of platypus ones were used to assemble the contigs. Comparison of the top 200 most abundant transcripts from the two datasets identified three proteins common to both echidna and platypus glands: a mitochondrial phosphate carrier protein (SLC25A3), mesencephalic astrocyte-derived neurotrophic factor (MANF) and a ribosomal protein (RPL9). The most highly expressed genes in the echidna venom gland are shown in Table 2. The most abundant venom components in platypus venom, venom defensin-like peptides and C-type natriuretic peptides [20], were not identified in echidna (Table 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. The most highly expressed proteins in the echidna transcriptome as annotated by platypus Ensembl proteins.

https://doi.org/10.1371/journal.pone.0079092.t003

Gene Ontology Analysis

We compared gene ontology (GO) enrichment statistics between the expressed genes (RPKM>10) in the echidna and platypus datasets to examine and infer biological differences between the two species by restricting the analysis to the top 200 most highly expressed genes in both datasets. 182 out of the top 200 echidna genes were associated with 1,537 GO annotations that represented 580 unique terms. Compared to the platypus dataset, we detected five significantly enriched GO terms that included adenyl ribonucleotide binding, adenyl nucleotide binding, ATP binding, post-translational modification and plasma membrane (Table 4).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Gene ontology (GO) terms that were enriched in the top 200 most highly expressed echidna genes compared to the platypus using GoStat2.

https://doi.org/10.1371/journal.pone.0079092.t004

We also performed functional term analyses on the top 200 most highly expressed genes in both echidna and platypus datasets. No gene ontology terms were found to be significant at p<0.05 after multiple testing correction (FDR) in independent tests for each species.

These analyses confirm that the key role of the echidna crural gland is not venom production. The life history of echidnas may point to the likely functional role of this gland. Echidnas are usually solitary, but during the breeding season (June to September) they are believed to use chemical cues to attract mates [1], [28]. Both sexes exude a pungent, musky odour possibly from the cloaca to advertise their reproductive status [29]. Male echidnas also exude a white milky substance from the base of their non-erectable spurs during the breeding season, while females, who lose their spurs before maturity, produce a small amount of solid secretion from the pits previously occupied by the spurs [28]. Several authors have speculated that the primary function of this white exudates acts as a chemical attractant during mating [2], [28].

The recent study by Harris et al. [28] investigated the composition of waxy secretions from the base of the spur using a mass spectrometry strategy to identify non-peptide compounds. The study revealed a variety of large molecules in produced by the tissue around the spur collar including sterols, fatty acids and methyl esters supporting their function in some form of scent communication. Consistent with this finding, stearoyl-CoA desaturase (SCD) – a protein involved in fatty acid biosynthesis was highly expressed in the echidna transcriptome. Other proteins involved in fatty acid metabolism including acyl-CoA dehydrogenase (ACADVL), acetyl-CoA carboxylase beta (ACACB), acyl-CoA synthetase long-chain family member 1 (ACSL1) as well as lecithin-cholesterol acetyltransferase (LCAT), sortilin-related receptor, L(DLR class) Repeats-containing protein (SORL1) and the Niemann-Pick disease protein (NPC1) are also all involved in sterol metabolic processes and expressed in the echidna crural gland transcriptome. Similar sterols and fatty acids are commonly found in scent gland secretions from other animals and can function to prolong the life of a scent mark (sterol esters) and encode complex informational content (long-chain fatty acids) (reviewed in [28]).

The major components of secretory granules in echidna crural glands are proteins [2] and it may be that the spur secretions possess additional, as yet, unknown roles. The most highly expressed gene in the echidna crural gland is most similar to butyrylcholinesterase, a serine hydrolase, whose physiological function in humans is still unclear, although it has been found to be an effective detoxification enzyme [30], [31]. Further research involving functional assays is required to determine the true biological roles of the most highly expressed proteins in echidna spur secretions.

Our results suggest that the primary role of the echidna crural gland is not to produce venom for offensive or defensive purposes. The majority of known platypus toxins were not detected in the transcriptome of the echidna crural gland. 30 echidna scaffolds matched 12 known toxins from a number of venomous species, but these were only lowly expressed in the gland and perhaps provide a remnant of the glands life history. Today the transcriptomes of the echidna and platypus crural glands have distinct expression profiles sharing few similarities in their most highly expressed genes. Should the echidna genome become available a review of the evolutionary remnants of venom genes can be conducted. The echidna gland appears to have evolved to play a role in chemical communication during the breeding season. Although the function of the crural gland has diverged significantly since the last common monotreme ancestor, each lineage continues to use the gland to aid reproduction.