If you make a discovery and at first people tell you that it can’t be right, and then they eventually switch to telling you ‘we knew that all along,’ then you are probably on to something.” It’s a quip that has stuck in Clifford Brangwynne’s mind. For the biophysicist at Princeton University, that is “exactly what happened with our findings on intracellular liquid phases.”
Think of liquids with different properties that don’t really mix but, under specific circumstances, cluster and separate like the shifting blobs in a lava lamp. That phenomenon, also known as liquid-liquid phase separation, was once considered to be an exclusively chemical process. But less than a decade ago, Brangwynne became one of the first to observe it happening inside cells as well, and ever since then, biologists have been trying to learn its significance.
Now scientists are beginning to understand that evolution has tuned certain proteins to act in aggregate like liquids. Through phase separation, they spontaneously self-assemble into dynamic, membrane-free, dropletlike structures that can perform needed tasks in cells.
“Somehow, no one thought that this kind of ability of molecules could be harnessed by evolution to achieve functionality or regulate functions,” said Simon Alberti, a biologist at the Max Planck Institute of Molecular Cell Biology and Genetics (MPICBG) in Dresden, Germany. “The focus was on the individual molecule and not on the collective.”
The breakthrough has big implications for our understanding of cellular organization and function, said Vasily Zaburdaev, a biophysicist at the Max Planck Institute for the Physics of Complex Systems, also in Dresden. One of the latest findings is that phase separation allows certain types of cells to cheat death when they are deprived of nutrients or otherwise put under stress. Phase separation enables the cells to turn a large part of their cytoplasm from a liquid to a solid—essentially putting themselves into a hardy condition of stasis until the nutrients return.
Nineteenth-century cell biologists coined the term organelle (“little organ” in Latin) to describe the tiny components they saw inside cells. Even then, pioneers in the field such as the American cell biologist Edmund Beecher Wilson suspected that the jellylike cytoplasm filling cells might hold various liquids “like suspended drops … of different chemical nature.” That early insight found little purchase in biology for almost a century, however: Researchers simply assumed that any droplet-shaped cellular organelles must have an encapsulating lipid membrane to prevent their contents from remixing with the cytoplasm.
Still, electron microscopy by researchers such as L. Dennis Smith of the University of California, Irvine, and Edward Mitchell Eddy of the National Institute of Environmental Health Sciences in the 1960s and early 1970s showed that some organelles simply didn’t seem to have any membrane at all. More membraneless structures continued to be found, such as the nucleolus, a dense structure in the cell nucleus. Yet until 2009, how and why they were forming wasn’t clear.
That year, when Brangwynne was a young postdoc at MPICBG, he, his colleague Christian Eckmann and his supervisor Tony Hyman saw something unexpected. They were looking at the uneven and inconsistent distribution of organelles called P granules inside cells of the roundworm Caenorhabditis elegans. P granules were widely assumed to be dense pellets of RNA and protein. But Brangwynne, Eckmann and Hyman saw that the granules were not solid at all. Instead, they appeared to be droplets of liquid that were coalescingat times to form bigger drops, like oil in a well-shaken vinaigrette.
“It was a serendipitous discovery,” Brangwynne said. “When we discovered that they were liquids, a number of quantitative measurements that we had been taking suddenly made perfect sense.” It also changed biologists’ understanding of how cells work.
That initial work by Brangwynne, Eckmann and Hyman triggered an avalanche of papers investigating the assembly and dispersal of various cytoplasmic proteins under various conditions. The evidence was getting stronger that cells had evolved a fine-tuned mechanism for organizing some of their internal structure and processes through phase separation—that is, letting proteins self-assemble into structures that could perform distinct functions.
In some instances, however, researchers are learning that the process is not reversible—and that this failure represents a malfunction of proteins associated with a broad range of diseases, including neurodegenerative disorders and cancer. For example, Zaburdaev observed that several mutant forms of a protein linked to certain diseases showed abnormal phase-separation behavior. “Instead of forming nice drops, they form very strange hedgehog structures,” he said.
SOLIDIFYING FOR SURVIVAL
Intrigued, Zaburdaev and several of his colleagues, including Alberti, decided to check what happens to proteins when cells are subjected to stresses such as falling temperatures and the sudden disappearance of nutrients. The surprising result they uncovered was that phase separation can be part of a cell’s survival mechanism.
The cells’ behavior could be likened to hibernation for bears. The animal lays still in a dormant state for weeks, minimizing its expenditure of energy. At a cellular level, phase separation helps the gelatinous cytoplasm make a protective transition into something more solid. “In this ‘solidified’ state, a cell can survive starvation,” Zaburdaev said.
The researchers studied this phenomenon by depriving yeast and amoebas of nutrients. No nutrients means no energy, and yeast cells need energy to pump protons out of their cytoplasm to maintain the neutral pH essential for their biochemistry. “By starving, cells acidified,” Zaburdaev said. Under the more acidic conditions, proteins readily went from a dissolved state to a more condensed and solid one, and the well-mixed cytoplasm separated into clusters of gelatinous blobs.
Simply by varying the acidity of the cells’ environment, the scientists could induce them to switch into this survival state, even without taking away the cells’ nutrients. The cells could rest this way for hours or even days. “We found that the cells are so rigid that they keep their shape” instead of being deformable, Alberti said. They “transition into a completely different material state.”
When their normal pH was later restored, the cells returned to normal, “dividing and living happily,” Zaburdaev said.
The scientists found that they could also trigger phase separation and solidification by completely dehydrating the yeast through osmosis. Different types of stresses seem to induce slightly different solid states, however. Exactly how that works is “something we don’t yet understand,” Alberti said.
Nevertheless, the survival mechanism that the experiments revealed was very simple, Alberti said: When there is stress, extensive phase separation leads to the rigidification of the entire cytoplasm, and the cell turns off its metabolism, like a hibernating bear settling down for the winter.
The comparison to hibernation may be more than figurative. “The cells of hibernating mammals may also solidify inside,” Alberti said. “It’s a perfect way of dealing with these kinds of environmental changes because solidification comes for free. The energy comes out of the temperature change or the drop in pH.” However, the hypothesis that phase separation is involved still needs to be tested, he said.
IMMOBILIZED FOR METABOLIC CONTROL
Most recently, Alberti’s team has been probing the phase-separation response of cytoplasmic proteins to stress at the molecular level. Their particular interest is in how it relates to control over cellular metabolism.
The perfect way to turn something off, Alberti said, is to put it into a solid material that can reversibly immobilize it until it’s needed again. “It’s a way of protecting molecules from damage, but also turning them off, storing them for later use.”
The team found that when a protein has a certain identifiable domain or region, the protein will form easily reversible gels. In the absence of this domain, the protein forms an irreversible type of assembly—permanently removing it from further use.
In effect, this domain modifies the protein’s phase behavior and keeps it reusable. “The domain provides a new possibility, for that protein to assemble into a benign kind of gel and not something from which you cannot come back,” Alberti said.
In one test-tube experiment, the researchers took a solution containing a single type of protein and lowered its pH. They saw molecules of the protein phase-separate from the solution and form gel-like blobs. Then they brought the pH back to neutral, getting the gels to dissolve, “showing exactly what we saw in cells,” Alberti said.
Such results imply that nature has designed the domain sequences to tune the proteins’ material properties. That’s very beneficial, said Dustin Updike, a biologist at the MDI Biological Laboratory in Bar Harbor, Maine, because it gives cells “a mechanism to respond to abrupt stress, such as heat shock, pH or osmotic stress.” Regulatory mechanisms in cells often work at the genetic level, he explained, meaning that they depend on signals reaching the nucleus, initiating gene transcription and the manufacture of an appropriate enzyme. But those events take time. In contrast, phase separations are very rapid—and can provide an almost immediate response to stress.
DOES IT REALLY MATTER?
Understanding the precise mechanism and effects of phase separation in cells could be highly relevant for a whole range of big biological challenges — from organ preservation and aging research all the way to space travel, according to Zaburdaev.
Recently, for example, the neuroscientist Pietro De Camilli at Yale University and his colleagues found evidence that phase separation might be involved in the controlled release of neurotransmitters at synapses. It had been observed that vesicles containing neurotransmitters routinely hover in clusters near the presynaptic membrane until they are needed. De Camilli’s team showed that a scaffolding protein called synapsin 1 condenses into a liquid phase, along with other proteins, to bind the vesicles into these clusters. When the synapsin is phosphorylated, the droplet rapidly dissipates and the vesicles are freed to spill the neurotransmitters into the synapse.
It’s still early days, though. When Brangwynne and his colleagues published their paper a decade ago, biologists reacted with either total incredulity or hope for a brand-new direction of research. As Updike noted, it can be hard for cell biologists to go from thinking about a phenomenon in terms of protein aggregation to the more complex problem of liquid phase separation, which requires fluid dynamics to describe.
“To me, Cliff’s work was a huge advance that better described the nature of P granules and what we were seeing,” Updike said, in part because it also explained why P granules had evaded biochemical purification for over two decades. “You can purify a granule, but purifying something more similar to an oil droplet is much more of a challenge.”
As ever more scientific papers back up the concept of phase separation as a cellular mechanism, the number of skeptics keeps on dropping, according to Updike and Brangwynne. Questions still remain, though.
“One of the criticisms is that some people say every protein can do this,” Alberti said. It’s common knowledge in science that concentrating proteins under various conditions can sometimes make them solidify or liquefy. “But there was never this idea that this could actually be used by cells, that evolution would actually act and use this ability of biomolecules to achieve a functional change such as down-regulating metabolism.”
Susan Wegmann, a biologist at the German Center for Neurodegenerative Diseases (DZNE), said, “So far it has not been shown that phase separation of proteins actually occurs in a living multicellular organism.” The relevance of phase separation in cells to complex problems in neuroscience and other areas is therefore uncertain. “We and others are trying to make that link, but it is of course very difficult and technically challenging. And if it turns out that protein condensation is linked to human diseases such as neurodegeneration, then we have to find smart ways to interfere in a specific manner with it.”
Tim Mitchison, a professor of systems biology at Harvard Medical School, is skeptical about whether phase separation is a generally important concept in biology. “I haven’t seen much evidence for phase separation in the cytoplasm of cells except for a few specific examples, like stress granules,” he said. The concept has seemingly not found much of an audience outside of cell biology: Many researchers either have not yet heard of phase separation or are ignoring the research.
“Maybe [they’re] waiting until there is more functional evidence,” Mitchison said. He noted that with enough of the right solvent, almost any protein or RNA can be made to phase separate. “But it’s not clear how much of this is physiologically relevant. I’m totally convinced phase separation is a thing, perhaps especially in RNA-protein biology,” he said. “I’m less clear how general it is.”
Brangwynne seems unperturbed by that reservation. He thinks that some skeptics “are asking very valid questions about what this all means for cell function and dysfunction, which is still not well understood.” Others might still be warming up to the idea of predictive quantitative models, he said, “but that is the future of biology.”