Microbial metatranscriptomics : towards understanding microbial gene expression and regulation in natural habitats

Abstract

Metagenomic research has paved the way for a comprehensive understanding of the microbial gene parts list in nature, but a full understanding of microbial gene expression, regulation, and ecology remains a challenge. In this thesis, I present the methodological foundations and applications of deep sequencing-based metatranscriptomics, for profiling community transcriptomes on spatial and temporal scales. Several findings and relevant hypotheses have emerged from this work. I show that transcripts of house-keeping genes necessary for the maintenance of basic cellular machinery are abundant and readily detectable. Habitat-specific transcripts are also discernible when comparing community transcriptomes along distinct geochemical conditions. Normalization of detected transcripts to their corresponding gene abundance suggests that numerically less abundant microorganisms may nevertheless contribute actively to ecologically relevant processes. Along the same lines, it is a recurrent observation that many transcripts are of unknown function or phylogenetic origin, and have not been detected in genomic/metagenomic data sets. These novel sequences may be derived from less abundant species or variable genomic regions that are not represented in sequenced genomes. Furthermore, I applied metatranscriptomics in a microcosm experiment, where a deep water mixing event was simulated and community transcriptomes were monitored over the course of 27 hours. Relative to the control, the treatment sample showed signals of stimulated photosynthesis and carbon fixation by phytoplankton cells, enhanced chemotactic, motility, and growth responses of heterotrophic bacteria, as well as possibly altered phage-host interactions. Such experimental metatranscriptomic studies are well suited to reveal how microorganisms respond during the early stages of environmental perturbations. Finally, I show that metatranscriptomic data sets contain a wealth of highly expressed small RNAs (sRNAs), transcripts that are not translated to proteins but instead function as regulators. I propose a bioinformatics pipeline for identifying these sRNA elements, characterizing their structures and genomic contexts, and predicting possible regulatory targets. The extraordinary abundance of some of the identified sRNAs raises questions about their ecological function, which warrants further biochemical and genetic studies. Overall, this work has extended our knowledge of functional potentials and in situ gene expression of natural microbial communities.