Extrinsic vs. Intrinsic noise-induced bimodality in post transcriptional regulation experiments

Extrinsic noise vs. intrinsic noise-induced bimodality in post-transcriptional regulation experimentsMicroRNAs (miRNAs) are a class of nearly 22 nucleotide-long, non-coding RNA molecules involved in the post-transcriptional regulation of their target RNAs. MiRNAs and their targets interact via a titration-like mechanism characterized by threshold effects and cross-talk among targets. Threshold has to be understood as a value of mRNA transcription rate such that above this value many target molecules are available for translation while below it mRNA are bound to miRNA, rapidly degraded and cannot be translated. It has been discovered that different RNA species may compete for miRNAs binding with this inducing indirect interactions amongst the miRNA targets which reciprocally influence their expression levels. Moreover, also the fluctuations on their levels of expression are coupled through miRNAs. It is of particular interest that an increase in noise leads to an increase in cell variability and this may lead to bimodal cell population distributions with high and low expression states of specific miRNA targets. This mechanism is particularly relevant near the threshold, where stochastic fluctuations play an important role for the cell fate. It is worth noting that bimodal expression of genes may lead to very different phenotypes, where the modes of the distribution underlies healthy or sick cells or two different stages of differentiation. An open question is whether these bimodal distributions at the target level are due to extrinsic fluctuations on miRNA pools or to intrinsic fluctuations on miRNA-target interaction strength. My thesis aims to answer this question combining both experiments and theoretical modelling. The experimental part consists of transfection experiments of bidirectional plasmids in epithelial human cells (HEK293). The plasmids, circular DNA molecules, code for two fluorescent proteins, a yellow fluorophore named eYFP and a red one named mCherry. The sequence coding for mCherry contains a varying number of miRNA binding sites, while the eYFP sequence is left unchanged. mCherry and eYFP are thus respectively proxies for target expression and its constitutive activity. To understand if bimodal distributions are due to extrinsic noise in the miRNA pool or to intrinsic noise in the miRNA-target interaction strength we need to distinguish between two scenarios: (i) a single cell scenario, in which bimodality is given by single cells near the threshold that jump from one state to the other over time (this scenario is due to intrinsic noise on the miRNA-target interaction strength); (ii) a population scenario, for which bimodality is given by cells whose target may be in one state or the other because of different miRNAs basal values (this is a situation of extrinsic noise in the miRNA pool). These two situations can be distinguished by measuring transition times between the two states when observing time trajectories of the target. Cells are therefore observed over time with time-lapse microscopy experiments that allow to record their fluorescence. Acquired images should be analysed to detect every single cell, track it over time and obtain temporal trajectories of its fluorescence. It can be expected that single cell transition times are distributed according to exp(-kt): therefore we can estimate k for every time trajectory and, depending on the distribution of k found, be able to understand in which scenario we are.