Targeted RNA-Seq Reveals the M. tuberculosis Transcriptome from an In Vivo Infection Model

Abstract

The study of host-pathogen interactions using in vivo models with intracellular pathogens like Mycobacterium tuberculosis (Mtb) entails technical limitations, such as: (i) Selecting an efficient differential lysis system to enrich the pathogen cells; (ii) obtaining sufficient high-quality RNA; and (iii) achieving an efficient rRNA depletion. Thus, some authors had used flow cytometers to separate infected cells or significantly increase the sequencing depth of host-pathogen RNA libraries to observe the pathogens' gene expression. However, these options carry additional expenses in specialized equipment typically not available for all laboratories. Here, we propose an experimental protocol involving differential cell lysis and a probe-based ribosomal depletion to determine the gene expression of Mtb and its host during in vivo infection. This method increased the number of observed pathogen-expressed genes from 13 using the traditional RNA-seq approach to 702. After eliminating rRNA reads, we observed that 61.59% of Mtb sequences represented 702 genes, while 38.41% represented intergenic regions. Some of the most expressed genes codified for IS1081 (Rv2512c) transposase and eight PE-PGRS members, such as PGRS49 and PGRS50. As expected, a critical percent of the expressed genes codified for secreted proteins essential for infection, such as PE68, lppN, and LpqH. Moreover, three Mtb ncRNAs were highly expressed (small RNA MTS2823, transfer-messenger RNA RF00023, and ribozyme RF00010). Many of the host-expressed genes were related to the inflammation process and the expression of surfactant proteins such as the Sftpa and Sftpc, known to bind Mtb to alveolar macrophages and mi638, a microRNA with no previous associations with pulmonary diseases. The main objective of this study is to present the method, and a general catalog of the Mtb expressed genes at one point of the in vivo infection. We believe our method represents a different approach to the existing ones to study host-pathogen interactions in tuberculosis and other similar intracellular infections, without the necessity of specialized equipment.

Keywords: RNA-seq; host-pathogen; in vivo infection; transcriptome; tuberculosis.

Figure 1

(a) Experimental strategies to obtain the Mtb (S1, S2, and S3) and mouse (M) transcriptome profiles. Pearson correlations between three biological replicates (A, B, and C) and strategies (S) are shown in yellow heat maps. (b) Assembled genome of the M. tuberculosis infectious strain. Regions with depth coverage lower and above average (788×) are shown as purple and yellow inner circles, respectively. In addition, the statistic parameters of the assembly were reported in the table.

Figure 2

Zihel-Neelsen staining of the infected lung homogenate before and after differential cellular lysis. The red arrow shows the bacilli per field.

Figure 3

Final strategy to obtain Mtb gene expression profile and genome location of the twenty most expressed genes. (a) Detailed experimental procedure used to obtain the transcriptome profile in Strategy 3. (b) Genomic location of the Top 20 most expressed Mtb genes (green vertical lines) and the three most expressed intergenic regions (red vertical lines) across the Mtb-sequenced genome. The red horizontal line indicates the 1138 kb region of the Mtb genome (from 3,005,716 bp to 4,144,624 bp) that concentrated twelve of the top 20 most expressed genes. The tables indicate each expressed gene ranked from the most to the least expressed and the percentage of reads mapped to each intergenic region.

Figure 4

Venn diagram comparing the expressed genes between Talaat et al. [6], Pisu et al. [5], and Cornejo-Granados in vivo studies.