I first got interested in the cell biology of Myxococcus xanthus in 2013 after meeting Tâm Mignot at a bacteriology conference. After seeing the work I was doing in Martin Howard's group on simplifying the cell cycle of Caulobacter crescentus, he wondered whether a similar approach would be useful for figuring out how polarity reversals are controlled in M. xanthus. Indeed, this turned out to be precisely the kind of problem for which mathematical modelling can be most useful.
Myxococcus xanthus is a swarming rod-shaped bacterium that has a highly complex life cycle involving swarming, predation and fruiting body formation and is a model organism for the study of collective behaviours and multicellular development. It possesses two types of motility machineries: one based on extension and retraction of a pilus and another involving gliding machinery and focal adhesion complexes. Both mechanisms are unidirectional but cells can change their direction of movement by inverting their polarity and this ability is essential for proper multi-cellular behaviours.
The polarity of the cell, and hence the location of the motility machineries, is controlled by a module containing MglA, a Ras-like GTPase that binds to the leading cell pole and MglB, a GTPase activating protein (GAP) that binds to the lagging pole, where it excludes MglA by stimulating GTP hydrolysis. During reversals, these two polar complexes rapidly switch poles in a mysterious process controlled by the Frz chemosensory-like system. This was the process we wanted to understand.
After working on the problem for a while, Martin and I came to the conclusion that the interactions amongst MglA, MglB and the third polarity protein RomR are such that they behave as a relaxation oscillator. However, for this to work, it required that the dynamics of RomR be much slower than that of MglA and MglB. Mathilde Guzzo, a PhD student in Tâm’s lab, performed FRAP experiments to check this and indeed found a substantial difference in timescale.
At that point, we were very happy and thought we would just need to a few more confirmatory experiments and we could start writing a paper. But as often happens, the system turned out to be much more complicated than expected. The model predicted that a reversal occurs when RomR accumulates sufficiently at the lagging pole. However, when we looked at more images, we found that RomR levels are often stable at the lagging pole. It seemed like its accumulation was not the only requirement for reversals. Additionally, RomR had been previously suggested to be a target of the Frz chemosensory system but Mathilde could not find any phenotype when mutating the critical phosphorylatable residue in RomR. The lab of Virginie Molle in the University of Montpellier also tested in vitro the phosphorylation of RomR by Frz, and even though this test was very successful for other Frz targets, she couldn’t detect any phosphorylation of RomR. Of course, negative results are always difficult to interpret but at this point we had accumulated a lot of evidence that RomR was not the central target of Frz and we started to search for an additional regulator.
Mathilde and Tâm began a genomic search for possible missing factors and discovered the missing response regulator FrzX. After a lot of genetics, microscopy, image analysis and modelling, and with the help of Eugénie Martineau and Sébastien Lhospice (and other people), we established that FrzX is absolutely critical for reversals and acts as a gate in controlling the MglA-MglB-RomR oscillator. This was the critical discovery, both on the genetic and modelling sides, that unlocked this project!
Signalling from the Frz chemosensory system is read by FrzX which switches the oscillator on (opens the gate) and MglA and MglB rapidly switch poles. However, the accumulation of RomR at the (new) lagging pole is still required so that subsequent reversals can only occur after a time-delay given by the time it takes for RomR to switch poles. This delay gives the system time to switch off the oscillator (close the gate). In this way, cells are able to flip their polarity in response to stimulus (or indeed stochastically), while maintaining it stably otherwise. Interestingly, this type of regulatory topology has, to our knowledge, not previously been found in biology. However, as this system involves a Ras-like GTPase, a family highly conserved throughout all Eukaryotes, we would not be surprised if similar mechanisms are found elsewhere.
The article 'A gated relaxation oscillator mediated by FrzX controls morphogenetic movements in Myxococcus xanthus' was published this week in Nature Microbiology (free view-only link: https://rdcu.be/232Y).
- This post was written with the help of Mathilde Guzzo.