RNA sequencing and de novo assembly of the digestive gland transcriptome in Mytilus galloprovincialis fed with toxinogenic and non-toxic strains of Alexandrium minutum

Abstract

Background: The Mediterranean mussel Mytilus galloprovincialis is marine bivalve with a relevant commercial importance as well as a key sentinel organism for the biomonitoring of environmental pollution. Here we report the RNA sequencing of the mussel digestive gland, performed with the aim: a) to produce a high quality de novo transcriptome assembly, thus improving the genetic and molecular knowledge of this organism b) to provide an initial assessment of the response to paralytic shellfish poisoning (PSP) on a molecular level, in order to identify possible molecular markers of toxin accumulation.

Results: The comprehensive de novo assembly and annotation of the transcriptome yielded a collection of 12,079 non-redundant consensus sequences with an average length of 958 bp, with a high percentage of full-length transcripts. The whole-transcriptome gene expression study indicated that the accumulation of paralytic toxins produced by the dinoflagellate Alexandrium minutum over a time span of 5 days scarcely affected gene expression, but the results need further validation with a greater number of biological samples and naturally contaminated specimens.

Conclusion: The digestive gland reference transcriptome we produced significantly improves the data collected from previous sequencing efforts and provides a basic resource for expanding functional genomics investigations in M. galloprovincialis. Although not conclusive, the results of the RNA-seq gene expression analysis support the classification of mussels as bivalves refractory to paralytic shellfish poisoning and point out that the identification molecular biomarkers of PSP in the digestive gland of this organism is problematic.

Figure 1

Transcripts integrity plot. The transcripts integrity analysis is based on BLASTx similarities of the M. galloprovincialis digestive gland contigs with the ortholog proteins predicted from the fully sequenced genome of Crassostrea gigas (with a cut-off p-value of 1x10⁻⁵). Ortholog coverage is displayed on the X axis, the percentage of contigs falling into each integrity category are shown on the Y axis.

Figure 2

Gene expression profiles at the early (T1) and late (T2) phase of PSP contamination. Scatter plots displaying gene expression profiles at T1 and T2; gene expression values (displayed as log₂ normalized read counts) are plotted for the AL1T-fed (X-axis) and AL9T-fed (Y-axis) samples. Genes identified as putative PSP biomarkers by the Kal’s Z-test on proportions at both time points are highlighted.

Figure 3

Comparison between the recent sequencing approaches in M. galloprovincialis. Venn diagram summarizing the overlap between the M. galloprovincialis reference transcriptome generated in the present work, Mytibase (Venier et al., 2009) and the pyrosequencing datasets produced by Craft et al. (2010), Philipp et al. (2012) and Suarez-Ulloa et al. (2013). Numbers shown in the graph correspond to the number of contigs showing a positive mapping hit in each sequencing set.