Gene expression is a fundamental process in all organisms, and we have a great deal of information about this process at the biochemical and structural level. However, we still have a poor understanding of transcription at the cellular level. A recent and fascinating discovery about gene expression relates to stochastic noise : transcription is not constant but fluctuates randomly over time, even for genes at steady-state. Different cells of a clonal population will therefore express different amount of any given mRNA, and this is very important as it creates phenotypic variability. However, because of the lack of appropriate technologies, these phenomena are not well understood.
Measuring transcription at the level of single cells requires tools able to detect and quantify single molecules of mRNAs, both in live and fixed cells. Single molecule FISH and MS2-tagging of RNAs provide such powerful tools. Building on this, we have developped microscopy approaches meeting these single molecule requirements. Thus, it is now possible to label specific mRNAs in live or fixed cells and to detect every molecule with a high spatial and temporal resolution. By measuring in real-time the amount of nascent RNAs at the transcription site of the gene of interest, it is possible to follow the stochastic variations of the corresponding promoter and to measure activity of RNA polymerase II in real-time. This data provide important insights into the mechanisms of transcription and gene expression.