Summary of discussion on 16/7/2009

Circadian clock and cell cycle

Experimental facts

We went through the main results of Matsuo et al., and Nagoshi et al. Main findings are that in liver the circadian clock gates M phase but not S-phase (at least not in the first wave of cell division, actually they never show a second wave, or do they?)

In the fibroblast cells, no molecular markers were used for cell division, instead time of nuclear envelope breakdown (late prophase) was found to be gated by the clock (in three windows, data is still a bit weak, We should still check if this makes sense with respect to maximal Wee1 activity (mRNA is ZT14 in Matsuo, fig. 1D).

Overall there is still not much data on the question. We also looked at the DNA damage phase shifting paper.

There also clear evidence that cell-cycle desynchronizes cell populations. (Cf. Hastings in Journal of biological rhythms JBR, old cells have tighter rhythms) 

Existing Models

Zamborszky et al. model effects of circadian wee1 on cell cycle. Use Tyson Novak model, cell cell length is given by minima of wee1 expression. Model is stochastic (chemical langevin). Claim to reproduce the tri-modal distribution when mass doubling time (DMT) is lower than 24 hours.

Kang et al. model the other direction, i.e. the effect of stalled transcription on the circadian clock. Analysis is so so...

BTW: it looks like they can entrain the circadian clock only for periods <24, which does not really make sense, right? Our data suggests that it goes both ways?

Ideas for experiments

Let’s start discussing this next time.

Ideas for models

1.     Establish phase diagram for bi-directional coupling, but otherwise free-running oscillators (mimicks the fibroblasts). Open question: do we treat the cell-cycle like a true oscillator. To start i would say yes.

2. Introduce entrainement of the circadian clock (this is more like the liver)

3. Generalize to noisy dynamics, two natural models would be to:

a.     Model cell cycle as relaxation oscillator

b.     We could use our previous frequency model for the circadian noise (drifting frequencies)

c.      Notice that it may not be trivial to generalize the notion of synchro in the noisy case, cf. Pikovsk

4. Discuss experimental data in relationship with phase diagrams. Naively liver looks synchro (expression data), while fibroblasts don’t.

5. Illustrate regime using explicit models (is  this necessary?)

Implementation

Nacho: make sure you fix the relative phases right, e.g. wee1 is expressed at ZT13, etc.