Counting single molecules to study the control of gene expression in development

It is an open question how, during embryonic development, spatial patterns of gene expression are induced with high precision. By counting single mRNA molecules during vulva induction in the nematode worm Caenorhabditis elegans, we established that the induced cells dynamically control the level of gene expression by increasing their sensitivity to the external inductive signal.

HFSP Career Development Award holder Jeroen Van Zon and colleagues
authored on Tue, 21 July 2015

During development, spatial patterns of cell fate are often induced by gradients of signaling proteins called morphogens. It is increasingly recognized that morphogen gradients can induce complex dynamic gene expression programs in the cells responding to the morphogen. Understanding how the quantitative information contained in morphogen gradients, coupled with its readout by the downstream gene regulatory network, generates such dynamics is challenging.  Quantitative analysis of the induced gene expression dynamics can provide novel insights into these questions. In particular, quantitative measurements can be used to test and constrain mathematical models of the underlying gene regulatory network. We used such an approach to study vulva induction in the nematode C. elegans, a classical model of spatial pattern formation by a morphogen gradient.

Click on image to enlarge

Figure: (A) Visualization of single-mRNA molecules of the Notch ligands lag-2 (red) and apx-1 (green) by single-molecule FISH in fixed animals at different stages of vulva induction. Each diffraction-limited spot corresponds to a single mRNA molecule. Nuclei are stained by DAPI (blue). The anchor cell (AC) as well as the vulva precursor cells P5.p, P6.p and P7.p are labeled. Scale bar, 10 μm. (B) Detail of lag-2 and apx-1 mRNA expression in P6.p, corresponding to the dashed boxes in panel (A). (C) lag-2 mRNA levels in P5.p (red), P6.p (green) and P7.p (blue) as a function of time The magenta, cyan and yellow markers indicate levels in P6.p of the animals in panel (A). (D) Final model of the signaling pathway controlling the dynamics in lag-2 expression, taking into account the external inductive ligand LIN-3, the receptor LET-23, the transcription factor LIN-1 and a (currently unknown) activator A. The rise in lag-2 expression is due to an increase of activity of the activator A from early (blue) to late (magenta) induction, independent of the external LIN-3 signal.

During vulva induction, it is thought that an LIN-3 gradient emanating from the anchor cell induces expression of the Notch ligands lag-2 and apx-1 in the nearby vulva precursor cells P5.p, P6.p and P7.p (Fig. 1A). Subsequently, Notch signaling between the vulva precursor cells restricts the expression of Notch ligands to P6.p exclusively. However, the expression dynamics of these Notch ligands had not been studied quantitatively. We used single-molecule fluorescence in situ hybridization (smFISH) to visualize individual lag-2 and apx-1 mRNA molecules at different stages of induction (Fig. A and B). Surprisingly, we found that lag-2 and apx-1 expression was highly dynamic.

By counting the number of mRNAs in the P5.p, P6.p and P7.p at different stages of vulva induction (Fig. C), we found that the expression dynamics occurred in three stages: an early stage at which lag-2 and apx-1 were present in P5.p, P6.p and P7.p, with the highest levels in P6.p, the cell closest to the anchor cell; a middle stage in which expression of lag-2 and apx-1 was restricted exclusively to P6.p; and a late stage at which the expression level of apx-1 and especially lag-2 rose strongly. We found a similar strong rise in lag-2 and apx-1 expression in different wild C. elegans strains, indicating that this feature of the expression dynamics is likely to be important for vulva induction.

We examined the mechanism behind the rise in lag-2 and apx-1 expression by systematically comparing quantitative expression data in wild-type and mutant animals with different mathematical models of the underlying signaling network (Fig. D). This analysis showed that the rise in lag-2 and apx-1 expression is due to an increase in time of the sensitivity of the P5.p, P6.p and P7.p cell to the external LIN-3 signal. Surprisingly, this change in sensitivity was independent of the presence of external LIN-3 and likely reflects an intrinsic temporal program executed by the vulva precursor cells. In general, this combination of quantitative smFISH data and mathematical modeling can be a powerful tool to dissect the dynamics of signaling pathways in development.       


Cells change their sensitivity to an EGF morphogen gradient to control EGF-induced gene expression. J.S. van Zon, S. Kienle, G. Huelsz-Prince, M. Barkoulas & A. van Oudenaarden. Nature Communications 6, 7053, doi:10.1038/ncomms8053 (2015)

Link to Nature Communications article

Pubmed link