Our project aims to uncover mechanisms by which noise is controlled and even exploited in gene regulatory networks at different organizational scales, from individual promoters to entire cells, from bacteria to yeast and mammalian cells. The different approache we use are described here in more detail.

A - Transcriptional kinetics and stochasticity of mammalian circadian regulatory elements - PI: Felix Naef

Transcription of individual genes does not occur with a uniform rate. It is an intrinsically stochastic process in which bursts of transcription are interspersed with periods of “quietness”. This stochastic dynamics can be superimposed on processes with a remarkable regularity such as the circadian rhythm. In this project we want to characterize how the bursting characteristics are encoded in the transcription regulatory sequences. For this, we established a short-lived luciferase reporter system that we use to measure the transcriptional output driven by a large number of circadian regulatory elements in single cells for (Fig. 1).


Figure 1. Reporter constructs. A promoter inserted into the Gateway (GW) cassette drives the expression of the nuclear luciferase fused with sequences that reduce the protein and mRNA half-lives. The FRT cassette allows stable integration of the construct into genomic DNA in FlpIn recombination compatible cells.

B - Temperature-sensing by RNA-mediated mechanisms (PI: David Gatfield)

Organisms and cells carry out their functions in fluctuating environmental conditions. Temperature variations can be assumed to represent a ubiquitous challenge to the intracellular network of biochemical reactions, whose kinetics and thermodynamics will generally display strong temperature dependence. Interestingly, also in warm-blooded organism temperature is not constant, if one considers for example naturally occurring body temperature rhythms, the temperature changes between the body’s core and periphery, or fever and hypothermia. Nucleic acids – in particular single-stranded RNA – has ideal properties to sense and react to cellular temperature changes through the formation of intra- and intermolecular base pairing that could directly control the accessibility to regulators (e.g. RNA binding proteins, miRNAs) or the ability of the ribosome to initiate translation (similar to the well-established RNA thermometers in bacteria). In this project we aim to uncover and characterize RNA elements that confer temperature-dependent regulation of protein expression in mammalian cells.

C - Combination of non-linearities and stochastic effects in multi-loop networks (PI: Attila Becskei)

The structure of the transcriptional regulatory networks that control the expression of individual genes is known to be important for the dynamics of gene regulation. Interestingly, although highly conserved among yeasts, some genes such as the galactokinase (GAL1), are controlled by regulatory circuits of different structures. In this project we aim to characterize the dynamical properties that are conferred by circuits of different structures. This work relies on the development of very accurate methods to quantify expression of individual transcripts in single cells (Fig. 2), a field to which we also contribute.

Figure 2. Single molecule FISH images of the GAL1 RNA. DNA is stained by DAPI. Arrows mark the active sites of transcription (nascent RNA). The scale bar is equal to 2 μm. The nascent mRNA spot intensity is larger than that of the mature mRNAs and is thought to reflect the burst size in stochastic gene expression. It is also affected by the rates of transcriptional elongation and termination.

D - Stochastic dynamics of E. coli’s gene regulatory network (PI: Erik van Nimwegen)

The main goal of this project is to investigate how and to what extent the activities of different TFs in single E. coli cells are coordinated in response to changing conditions. That is, on one end of the spectrum of possibilities one could imagine that each E. coli TF modulates its activity in response to a separate signal, without much regard to the activities of other TFs, whereas on the other end of the spectrum one could imagine that TF activities are tightly integrated such that single cells can only take on a very low-dimensional subspace of possible TF activity patterns. To investigate this we are developing a microfluidic setup, based on a published design called “Mother Machine” (Wang et al. 2010) that, in combination with time lapse microscopy and automated image analysis, allows us accurately track division times, growth rates and gene expression (using fluorescent reporters) of lineages of single E. coli cells. This allows studying how fluctuations in the activities of TFs couple to phenotypic variable such as growth and replication rates. Second, using multi-color fluorescent reporters in combination with flow-cytometry we will quantitatively characterize the joint distributions of the activities of pairs and triplets of TFs across single cells.  Such comparisons should provide insight into whether the cell's regulatory circuits impose control or exploitation of noise in the gene expression of individual regulators.

E – The impact of miRNAs on the cell-to-cell variability in target expression (PI: Mihaela Zavolan)

MiRNAs are small regulatory RNAs that have been shown to increase the decay rates of their targets and to reduce their rate of translation. It has been hypothesized that one of their functions is to reduce the cell-to-cell variability in the expression of their targets, which is also viewed as a “fine-tuning” function. In this project we want to take advantage of a system that we have previously established in the lab - human embryonic kidney (HEK) 293 cells in which the expression of a miRNA with doxycycline (Fig. 3) – to measure the impact of miRNA expression in single cells.

Figure 3. Design of the experiment for measuring the effect of miRNA expression in single cells. A. MiRNA expression can be induced by doxycyclin together with that of the green fluorescence protein from a bi-directional promoter. B. Snapshots of cells that were induced for 8 days with different concentrations of doxycycline. C. Relationship between the GFP mRNA expression and miRNA expression in cell populations. The data indicates that GFP mRNA level could be a good proxy for the miRNA expression.

F - Transcriptional and post-transcriptional regulation of cell state transitions during pluripotency induction in somatic cells (PI: Matthias Lutolf)

The deterministic and highly efficient reprogramming of somatic cells into induced pluripotent stem cells (iPSCs) is one of the important goals of modern cellular biology.  However, it has been reported already that the entrance into a “reprogrammable” state is stochastic. In this project we would like to study how gene expression patterns change along cell genealogies. For this purpose we develop a microfluidic technology to trap and analyze cell progeny of single somatic mother cells that are generated over multiple rounds of cell divisions during the earliest steps in reprogramming. In order to facilitate transcriptomics analyses of captured daughter cells, the reliable extraction of mRNA from single cells is key. To this end, we have been constructing a microfluidic chip module for isolation of mRNA and as well as synthesis of cDNA directly on chip (Fig. 4).

Figure 4: Schematics for the on-chip single cell mRNA-capture and RT module.