Thread by @connor_rosen

Here’s a bit of a dive into some interesting recent work on #PhaseSeparation in cells - focusing on the role of ATP as a protein stabilizer in cells, independent of its energy-providing function. It all starts with a paper in @ScienceMagazine from 2017, science.sciencemag.org/content/356/63….

This Science paper showed that in vitro, ATP had hydrotope activity - the ability to solubilize proteins, and dissolve protein aggregates or liquid phase droplets. This function occurred at millimolar concentrations of ATP, near normal cellular levels.

There was some pretty quick push-back against the interpretation about any cellular role - pubpeer.com/publications/8…. One major point of criticism is that even at 5 mM ATP, there’s only about 1 ATP molecule per protein in a cell. So it’s hard to imagine proteome-scale activity.

Still, the paper has picked up over a hundred citations - scholar.google.com/scholar?cites=…. Most of them, however, are “drive-by” citations that don’t actually confirm the findings but simply cite hydrotope activity as a role of ATP.

Those that are directly assaying ATP activity are somewhat split - plenty of individual examples in vitro or in cells of no role for ATP in stabilization, but also some suggesting, but not fully proving, a hydrotope role in cells.

So far, I think it’s pretty tough to say this is a really well-established fact that can be applied to cellular homeostasis. One big hurdle is disassociating ATP activity as a hydrotope vs as an energy source for chaperones.

And so stories like cdn.elifesciences.org/articles/35224… and sciencedirect.com/science/articl… end up arguing for multiple roles of ATP, including chaperone-dependent roles in protein stabilization in cells. And it's not really clear how to reconcile them with a hydrotope model.

That’s why I really like this new paper in @NatureComms from @savitski_lab (nature.com/articles/s4146…). They use some really good proteome-wide mass spec techniques to profile both thermal stability and solubility in response to ATP.

The basic principle is to measure soluble protein abundance in cell lysates by mass-spec at a range of temperatures (or detergent solubilizing agents) and ATP concentrations, allowing building of melting-temp and solubility curves in response to ATP.

The experiments seem very well set up and controlled, and are pretty strongly suggestive of an energy-independent role for ATP stabilization in cells. Importantly, it’s not for all proteins, which gets around the concentration issues discussed on pubpeer.

Many proteins that are stabilized are ATP-binding proteins, and ligand binding is known to (in general) stabilize proteins, so that’s not exactly a pure “hydrotope” role. It is certainly an energy-independent role, with large effects on the proteome, but not a hydrotope.

But I think there’s sufficient evidence that a substantial number of non-ATP-binding proteins (especially IDPs and highly charged proteins, known to be prone to phase separation) are stabilized and solubilized in a manner that is more consistent with a hydrotope activity.

There’s a lot more data in the paper, certainly worth careful reading. It seems like a really solid paper and I think it really argues for a broad, but not universal, role of ATP as a biological hydrotope that previous work hadn’t fully established.

As a quick note, it was also posted early as a #preprint (biorxiv.org/content/10.110…). @NatureComms also posts peer review comments, and the feedback was very positive. This is a good example of two trends in publishing I like, preprints and open peer review!