Description of Bacterial RNA Transcripts Detected in Mycobacterium tuberculosis Infected Cells from Peripheral Human Granulomas using Single Cell RNA Sequencing

Philip J Moos,Allison F Carey,Jacklyn Joseph,Stephanie Kialo,Joe Norrie,Julie M Moyarelce,Anthony Amof,Hans Nogua,Albebson L Lim,Louis R. Barrows

DOI: https://doi.org/10.1101/2024.08.20.608852

2024-08-20

Abstract:Mycobacterium tuberculosis (Mtb) remains a global human health threat and a significant cause of human morbidity and mortality. We document here the capture of Mtb transcripts in libraries designed to amplify eukaryotic mRNA. These reads are often considered spurious or nuisance and are rarely investigated. Because of early literature suggesting the possible presence of polyadenylated transcripts in Mtb RNA, we included the H37Rv Mtb reference genome when assembling scRNA seq libraries from fine needle aspirate samples from patients presenting at the TB clinic, Port Moresby General Hospital, Papua New Guinea. We used 10X Genomics single-cell RNA sequencing transcriptomics pipeline, which initiates mRNA amplification with poly-T primers on ~30-micron beads designed to capture, in this case, human mRNA associated with individual cells in the clinical samples. Utilizing the 10X Genomics Cell Ranger tool to align sequencing reads, we consistently detected bacterial small and large ribosomal subunit RNA sequences (rrs and rrl, respectively) and other bacterial gene transcripts in the cell culture and patient samples. We interpret Mtb reads associated with the host cell's unique molecular identifier (UMI) and transcriptome to indicate infection of that individual host cell. The Mtb transcripts detected showed frequent sequence variation from the reference genome, with greater than 90% of the rrs or rrl reads from many clinical samples having at least 1 sequence difference compared to the H37Rv reference genome. The data presented includes only bacterial sequences from patients with TB infections that were confirmed by the hospital pathology lab using acid-fast microscopy and/or GeneXpert analysis. The repeated, non-random nature of the sequence variations detected in Mtb rrs and rrl transcripts from multiple patients, suggests that, even though this appears to be a stochastic process, there is possibly some selective pressure that limits the types and locations of sequence variation allowed. The variation does not appear to be entirely artefactual, and it is hypothesized that it could represent an additional mechanism of adaptation to enhance bacterial fitness against host defenses or chemotherapy.

Microbiology