Phase behaviour of disordered proteins underlying low density and high permeability of liquid organelles
Nat. Chem., (Online) (2017)
Wei, M.-T*., Elbaum-Garfinkle, S.*, Holehouse, A.S.*, Chen, C.C.-H., Feric, M., Arnold, C.B., Priestley, R.D., Pappu, R.V., and Brangwynne, C.P.
* - Co-first authors
This paper represents a massive set of work from myself, Ming-Tzo (Steven) Wei, and Shani Elbaum-Garfinkle. To cut a long and fairly technical story short, Steven developed and built a new type of Fluorescence Correlation Spectroscopy (FCS) microscope that provides a way to obtain calibration-free measurements of protein concentration within a confocal volume. Using this ultrafast scanning FCS (usFCS) Steven measure the concentration of protein inside phase-separated droplets of LAF-1, a protein that is involved in the formation of P granules in C. elegans. To everyone’s surprise, the concentration of protein inside these droplets was on the order of 4-8 mg/ml – i.e. around 100 fold more dilute than had been expected. We found that these dilute droplets could not be explained using the theoretical frameworks typically used to describe simple homopolymers that phase separate, but that by using an advanced theory developed by Murugappan Muthukumar in 1986 we were able to quantitatively describe this behaviour. This theoretical formalism explicitly captures the fact that the underlying polymers undergo large-scale conformational fluctuations, providing a direct theoretical explanation for why disorder might be important in driving phase separation. Moreover, this theory makes quantitative predictions regarding the mesh-size (an emergent property of the droplet) and the extent of the fluctuations experienced by an individual disordered protein.
These predictions set up a testable framework; using all atom simulations and diffusion measurements we found that the extent of fluctuations are entirely consistent with the predicted fluctuations. Moreover, the mesh size was tested using both rheological approaches and probe accessibility experiments, again showing quantitative agreement with the predictions from theory. This mesh size was tested in vitro, but was also examined for P-granules in vivo, where it was found to match. Taken together, these results paint a surprising picture in which disordered proteins can drive the formation of hyper-dilute liquid droplets.
It is important to emphasize that we are not suggesting that all liquid-like droplets should show the same kind of dilute behaviour. In fact, we have multiple examples of proteins we believe behave in fundamentally different ways, all of which can be explained within a framework that explicitly considers the extent of conformational fluctuations. However, an important result from this work is that for heteropolymers, a complete decoupling between the critical point, the low concentration arm, and the high concentration arm of the co-existence curve (binodal) appears possible. This means that many of the mean-field homopolymeric theories developed to describe phase behaviour and aggregation in disordered proteins may not necessarily be applicable. Many questions remain to be answered with this system – of upmost importance is a quantitative description of how (and why) RNA modulates these droplets’ phase behaviour.
This project represents one of several collaborations between myself and the Brangwynne lab, and it has been an absolute privilege to work with such an exceptional group of individuals.