Ion Torrent-based transcriptional assessment of a Corynebacterium pseudotuberculosis equi strain reveals denaturing high-performance liquid chromatography a promising rRNA depletion method

Abstract

Corynebacterium pseudotuberculosis equi is a Gram-positive pathogenic bacterium which affects a variety of hosts. Besides the great economic losses it causes to horse-breeders, this organism is also known to be an important infectious agent to cattle and buffaloes. As an outcome of the efforts in characterizing the molecular basis of its virulence, several complete genome sequences were made available in recent years, enabling the large-scale assessment of genes throughout distinct isolates. Meanwhile, the RNA-seq stood out as the technology of choice for comprehensive transcriptome studies, which may bring valuable information regarding active genomic regions, despite of the still impeditive associated costs. In an attempt to increase the use of generated reads per instrument run, by effectively eliminating unwanted rRNAs from total RNA samples without relying on any commercially available kits, we applied denaturing high-performance liquid chromatography (DHPLC) as an alternative method to assess the transcriptional profile of C. pseudotuberculosis. We have found that the DHPLC depletion method, allied to Ion Torrent sequencing, allows mapping of transcripts in a comprehensive way and identifying novel transcripts when a de novo approach is used. These data encourage us to use DHPLC in future transcriptional evaluations in C. pseudotuberculosis.

Figure 1

Chromatograms of a RNA sample before and after DHPLC fractionation for rRNA removal.A. Total RNA sample chromatographic profile. The two highest spikes correspond to ribosomal RNAs, and the red dashed lines delimitate the interval of run corresponding to the collected sample (mRNA) for downstream applications.B. Chromatographic profile of sample submitted to depletion with DHPLC. The green dashed lines virtually represent the previously selected area for sample collection. The proportion between the selected peak and the rRNA-related spikes appears to be more balanced in this case.

Figure 2

Percentages of 16S rRNA depletion for each of four independently prepared biological replicates, in relation to the sigA gene mRNA levels. DHPLC and RiboMinus treatments were performed in a parallel way for each replicate. The asterisks indicate the cases where the levels of depletion observed for DHPLC were significantly different from those seen for RiboMinus, according to the software REST 2009.

Figure 3

Classification of reads mapped against the CDS of C. pseudotuberculosis 31, according to terms of Gene Ontology.A. The identified transcripts were grouped by the biological processes to which their products are associated with. The horizontal bars indicate the amounts of grouped transcripts (genes) for each category, considering transcriptome results for both RiboMinus- and DHPLC-treated samples.B. In a similar way as before, the identified transcripts were sorted by their association with different classes of molecular functions.