Signaling dynamics control cell fate in the early Drosophila embryo

Heath E Johnson, Stanislav Y Shvartsman, Jared E Toettcher

Preprint posted on June 08, 2018

Article now published in Developmental Cell at

How can the same molecule specify different dev outcomes? By systematically inducing Erk signalling using optogenetics, a preprint by @toettch lab suggests that differences in the dynamics of signalling underlie different cell fates in fly embryos

Selected by Yara E. Sánchez Corrales


The Erk mitogen-activated protein kinase is required for cells to adopt different fates during early embryogenesis in the fruit fly at least in three different contexts:

  • Anterior pole, where Erk is implied in the formation of head structures
  • Ventral midline, where Erk is involved in the patterning underlying neural precursors
  • Posterior pole, where Erk influences posterior midgut invagination

 How could the same signalling molecule specify different cell fates? An attractive hypothesis is called dynamic control which posits that a single signal can be interpreted differently depending on its dynamic features such as amplitude, duration or frequency of activation.

In order to explore whether Erk signalling in the developing embryo is decoded by dynamic control, the authors use a recent tool, the OptoSOS system, that permits the precise manipulation of Erk signalling using light. Upon blue light stimulation, a Ras activator goes to the membrane and this activates a kinase cascade that culminates in the activation of Erk signalling (see Johnson et al., 2017 for more details). Using this tool in combination with genetic manipulations, the authors are able to test different hypotheses related to the dynamic decoding of Erk signalling.

Key findings

The authors start by inducing Erk signalling in the whole embryo using homogeneous illumination. This perturbation induces tissue contractions that are explained by the ectopic expression of myosin II. Importantly, upon light stimulation the receptor-ligand pair mist and folded gastrulation (fog), involved in tissue contractility in the embryo, are also upregulated. In order to address which network is involved in ectopic contraction, the authors assess the expression pattern of some transcription factors known to promote contractility during development. They find that the light-induced Erk is able to expand the expression domain of tailess (tll) and huckebein (hkb), two transcription factors associated with posterior midgut (PMG) invagination. Thus, ectopic Erk signalling is sufficient to drive cells to adopt a contractile PMG-like fate.

Interestingly, the 85% posterior part of the embryo shows higher level of myosin activation compared to the anterior, suggesting that contractility mediated by Erk is suppressed in the anterior part of the embryo. The authors show that high levels of Bicoid (Bcd) block Erk-dependent contraction behaviour because when Bcd is removed (ie. bcd mutant), Erk induces uniform contractility. This suggests that the combination of signals, Bcd and Erk, underlies head structures in the anterior part of the embryo. Thus, it is an example of combinatorial control.

The authors next tested the effects of the time and duration of Erk signaling that is known to be different in normal development. For instance, in the PMG, Erk expresses early during the first nuclear cycles while in the ventral midline it expresses just before gastrulation during the 14th nuclear cycle. Moreover, in the PMG, Erk is sustained for over 1h while in the ventral midline is a transient 20 min pulse. Using the OptoSOS tool, the authors systematically vary the time and the duration of Erk signalling (i.e light induction for 30, 45, 60 or 120 min through different developmental times). They find that the duration of stimuli, regardless of the specific developmental window, predicts the response to Erk stimulation very well. Interestingly, short time stimulation (30 min) leads to increased levels of ind (associated with neuron specification) and longer stimulation (120 min) caused ectopic expression of mist, associated with contractility. Thus, the duration of Erk signalling is able to direct cell fate towards neuron precursors or contractile behaviour.

How is Erk signalling decoded within the cell? Two different models have been proposed: 1) persistence detector and 2) cumulative load sensor (Figure 1). In the former, persistent pulses of Erk stimuli are needed for downstream responses. In the second, distinct cell fates are triggered when the total signal reaches a threshold (Figure 1). These mechanisms can be distinguished by the phenotypic response to a pulsed versus a continuos stimulus. Erk induction divided into three equal pulses (3×5 or 3×15 min) leads to similar phenotypes to those observed in the case of single doses (15 or 45 min), suggesting that it is the total or cumulative Erk activity that is decoded.

Figure 1. The persistence detector and the cumulative load sensor are two mechanisms that could decode Erk signalling (a). In order to distinguish between them, Erk can be induced in pulses or as continuos stimulus. The cumulative sensor mechanism predicts the same response regardless of the shape of the stimuli (b). Induction of Erk signalling leads to similar phenotypic responses regardless of the shape of stimuli ( b), supporting the the cumulative load mechanism. (Figure 4 of the original preprint).

The authors also explore the type of genes that could be involved in decoding Erk signalling under the cumulative detector mechanism. They propose two different decoding types of genes: the accumulators and the thresholders. The accumulators will integrate the signal over time, thus their expression level will be proportional to the intensity of the stimuli. The thresholders will act as a switch and would increase abruptly above a stimulus threshold. The authors find that the terminal gap genes tll and hkb respond as predicted for accumulator genes, while hkb and mist act as a thresholders genes turning on abruptly after 30-45 min of stimulation.

Figure 2. Summary of Erk signalling dynamics implied in different cell fates during Drosophila embryo development (Figure 5 of the original preprint).

What I like about this preprint and open questions

 In this study, the use of an optogetic tool makes it possible to systematically control the spatial and temporal dynamics of Erk signalling. In my opinion, the combination of these kinds of quantitative tools and genetic perturbations make the Drosophila embryo a remarkable system for unravelling the dynamic features of both signalling and genetic networks.

I am convinced that quantitation and precise manipulation of dynamic signalling features (such as duration, intensity and time of stimuli) adds another level of understanding to signalling pathways by incorporating the temporal decoding of signalling.

I would like to know if the authors also tried to modulate the amounts of signalling levels by controlling the amplitude of the signal. If the cumulative load mechanism is driving the process, would a single high intensity stimulus produce the same effects as a prolonged low intensity stimuli?

I also wonder about the role of pulse frequency and periodicity. Could the frequency and period of activation be important? Presumably, other response genes could be influenced by a pulsatile behaviour.

I find it very interesting that induction of Erk increased contractility and induced myosin II ectopic expression. It is know that myosin II can be junctional or show pulses in the apico-medial domain. The former is associated with junction shrinking and the latter with decrease in cell area. I wonder if Erk overexpression will also change these distinct contractile behaviours.

I also really like the idea of unravelling the network behaviour that could decode signalling. I wonder whether “thresholders” and “accumulators” could also be part of a persistence decoding mechanism.

Finally, I find this preprint very inspiring and hope that approaches like this will help unravel more examples of dynamic control in development.


Tags: cell fate, contractility, myosin ii, optogenetics

Posted on: 18th October 2018 , updated on: 22nd October 2018

Read preprint (No Ratings Yet)

  • Author's response

    Jared Toettcher shared

    We are honored to be featured in preLights! We are also quite impressed
    with the depth and accuracy with which our paper was presented; for us,
    this discussion really underscores the value of publishing preprints.

    The open questions raised in the preLight are all excellent ones. First:
    your are correct that the “cumulative load” model predicts an
    equivalence between amplitude and duration. Put simply: if the total
    dose is all that the cell cares about, then a brief, bright pulse of
    light should give the same phenotype as a long, dim one. Indeed, this is
    one experimental direction we have pursued since publishing the preprint
    – and so far, the results appear to confirm our model. We hope to
    publish a revised manuscript soon!

    A second question revolves around the frequency of stimulation. This is
    something we are *very* interested in, especially as Erk activity has
    been shown to pulse (or oscillate) in many mammalian model systems. In
    mammalian cells, we have shown that repeated pulses of Erk can turn on
    some genes more efficiently than a single, constant stimulus (Wilson et
    al, Mol Cell 2017). Intriguingly, Erk pulses have not yet been observed
    in the fly – but I strongly suspect that they soon will…

    Have your say

    Your email address will not be published. Required fields are marked *

    This site uses Akismet to reduce spam. Learn how your comment data is processed.

    Sign up to customise the site to your preferences and to receive alerts

    Register here