sign in sign up
<<
Home , Soft matter

technology feature befriends soft matter separation creates complex condensates in eukaryotic cells. To study these mysterious droplets, many disciplines come together. Vivien Marx

n many cartoons, eukaryotic cells have When the right contacts are made, phase a tidy look. A taut membrane surrounds separation occurs. Ia placid, often pastel-hued lake with well-circumscribed organelles such as Methods multitude ribosomes and a few bits and bobs. It’s a In the young condensate field, Amy helpful simplification of the jam-packed Gladfelter of the University of North cytoplasm. In some areas, there’s even more Carolina at Chapel Hill says her lab intense milling about1–4. To get a sense of intertwines molecular-biology-based these milieus, head to the kitchen with approaches, modeling, and in vitro Tony Hyman of the Max Planck Institute reconstitution experiments using simplified of Molecular Cell Biology and Genetics systems. “You can’t build models that have and Christoph Weber and Frank Jülicher of all the parameters that exist in a cell,” she the Max Planck Institute for the Physics of says. But researchers need to consider that Complex Systems, who note5: “For instance, in vitro work or overexpressing proteins when you make vinaigrette and leave it, you will not do complete justice to condensates come back annoyed to find that the oil and in situ. “Anything will phase separate vinegar have demixed into two different in a test tube,” she says, highlighting phases: an oil phase and a vinegar phase.” Just Eukaryotic cells contain mysterious, malleable, experiments in which crowding agents as oil and vinegar can separate into two stable -like condensates in which there is intense such as polyethylene glycol are used to phases, such liquid–liquid phase separation activity. Credit: Y. Drozdova/Alamy Stock Vector. induce phase separation. Another widely can take place in the cytoplasm. This used method involves measurement of creates condensates, which are malleable, fluorescence recovery after photobleaching. liquid-like areas where proteins, peptides giving way to rigorous approaches and tools Such FRAP-ing, “that’s just not telling you and RNAs congregate that might otherwise to study “native biological condensates” and that much.” Better measurements are needed be on opposite sides of Lake Cytoplasm. manipulate them, he says. “Methods are still to, for example, characterize viscosity P-bodies and stress granules are some of pretty underdeveloped in this field,” he says. in a condensate, she says. New types of the condensates in the cytoplasm. Nucleoli Given that molecular forces and biophysical density probes will be useful, as will dyes and Cajal bodies are some of those in the factors drive condensate formation and that fluoresce to reflect how crowded a nucleus. Unlike organelles, condensates have maintenance, advances need to draw on condensate is. Her lab applies rotor dyes to no membrane, but there is a “non-membrane several disciplines. Cell and molecular assess how much ‘tumbling’ is taking place boundary,” says Hyman, with, for example, biologists, proteomics labs and those in in a condensate. One organism Gladfelter density and concentration unlike those of physics are intrigued by condensates. studies is the fungus Ashbya, which has the surrounding cytoplasm or nucleoplasm. Brangwynne, then a postdoctoral fellow multinucleated cells, as do other organisms Condensates are a little like flash mobs, says in the Hyman lab, discovered that P granules and human muscle cells, too. In her Woods Hyman. When the music is on, people come phase-separate in the posterior of the Hole summer lab, she began collaborating together, and when the music stops, they embryo as part of germ cell development in on condensates with Brangwynne, Hyman, disperse. Condensates are dense locales. A Caenorhabditis elegans1. During a research Michael Rosen of the University of Texas eukaryotic cell typically contains billions stint at the Marine Biological Laboratory Southwestern Medical Center and others. of proteins. A condensate might be one in Woods Hole, Massachusetts, and using In Ashbya’s non-membrane-bound areas, to two micrometers in diameter and hold quantitative particle tracking, he and one she came across Whi3, an RNA-binding around one million proteins. It’s akin, he of the students in the physiology course says, to situating the world’s population in tagged P granules and showed how they greater Los Angeles. And everyone in this move, condense and coalesce to bring RNAs megalopolis stays on the move. and proteins together. Hyman, Jülicher Components in a condensate tumble and colleagues were excited about the around one another, touch briefly and move finding, says Brangwynne. “Such phase on, which leads to a dynamic of interactions, transitions may represent a fundamental says Princeton University researcher Clifford physicochemical mechanism for structuring Brangwynne. A wide range of biochemical the cytoplasm,” the authors noted. P activities take place in condensates, such as granules may look granular, but they are chromatin organization or stress responses. actually phase-separated liquid droplets. Early papers in the condensate field were What is known, says Hyman, is that the cell’s more qualitative and descriptive, but that is proteins interact and ‘sample’ one another.

Nature Methods | VOL 17 | June 2020 | 567–570 | www.nature.com/naturemethods 567 technology feature protein. Her team in collaboration with Sua versus intrinsic disorder have given way to the Myong, now at Johns Hopkins University, concept of multivalency, says Pappu, which used techniques such as single- the model represents. Adapting concepts fluorescence resonance transfer from , the model considers (FRET) and found that Whi3 binding segments along a protein’s chain of amino changes the conformational dynamics of the acids to be more or less prone to interactions: RNAs it targets. RNAs and the RNA–protein some behave like ‘stickers’, others as ‘spacers’. interactions shape condensate formation “The balancing of sticker-to-spacer ratio and maintenance and give condensates becomes quite important,” he says. With distinct identities. Condensates will differ only stickers, a protein can end up a goopy in their Whi3 levels and location. Some mess, which can be a hallmark of disease. condensates form at the cell tips where Phase separation, says Pappu, may play a growth is taking place. Gladfelter explores role in Huntington’s disease. A mutation in molecular mechanisms of condensate the huntingtin gene leads to a polyglutamine dynamics to understand how sequestering repeat that contributes to fibrillary tangles a biochemical reaction in a condensate has in cells. Working on huntingtin led him advantages for the cell. Since condensates to condensates, says Pappu. Using online are so dynamic, classic structural biology Optogenetics methods can be used to study the resources, scientists can find out how sticky approaches to studying proteins and protein of phase separation. Shown here, the individual residues are, but these resources complexes don’t address all of a condensate’s nucleus of a living human cell at different times lack context for the particular sequence one complexity, says Hyman. Mass spectrometry after optogenetic activation with blue light. is looking at, says Pappu. This motivates him can help labs establish the metabolites in a Credit: D. Bracha, C. Brangwynne, Princeton U. to contribute to the community’s efforts to condensate. And when condensates can be build “a rigorous grammar” of condensates purified and they stay intact, says Gladfelter, to help predict, from sequence, how proteins mass spec helps to identify proteins, It’s a way to get a quantitative readout of will interact. which has worked with stress granules, the intrinsically disordered proteins that This prediction is informed by for example. For the RNAs, labs can use might be driving phase separation in cells. experiments such as single-molecule FRET RNA sequencing. For the condensates she Brangwynne is particularly fond of Corelet measurements that capture resonance studies, however, “we have not been able to because it helps query how oligomerized an transfer efficiencies in disordered proteins. purify the condensates and do that.” Nuclear environment has to be for a condensate to Pappu’s then postdoctoral fellow Kiersten magnetic resonance (NMR) form. “It lets us map these phase diagrams,” Ruff worked out how to give FRET has been applied to some proteins relevant he says. Corelet involves a self-assembled measurements a molecular context with to phase separation. She points to the wealth GFP-tagged particle, with a 24-mer ferritin all- simulations and machine-learning of tools from the Brangwynne lab and says, core in which each core is fused to an approaches to characterize protein “I personally don’t think any one thing optogenetic domain. As the particles ‘touch’ ensembles. Calculated FRET efficiencies is sufficient.” The method optoDroplet6, proteins with internally disordered regions, are calibrated with experimental ones. developed in the Brangwynne lab, uses light more or fewer proteins attach, depending on What results is calculated conformational to trigger protein association and selectively conditions. Corelet is like having an “ heterogeneity of proteins in condensates. control a condensate. Researchers can knob” on cells, says Brangwynne. Just as They have a general “tadpole architecture,” look at intrinsically disordered proteins in temperature changes a system’s entropy, so he says, with a more globular head and a condensates, initiate intracellular protein condensates can have altered entropy levels. tail where the spacers are. In the protein clustering and study phase separation. Probing the interactions of proteins with huntingtin, the glutamines are stickers, corelets, researchers find out how much but in a different protein, fused in sarcoma these components “want to stay together,” he (FUS), glutamines and asparagines are says. “That’s a technique I really like for that not as potent stickers. “Who is a spacer reason.” Recently, his lab developed a way and who is be sticker will become context to extract thermodynamic parameters from dependent,” says Pappu. Using this model, microscopy images of condensates9. To do the team has built a computational pipeline so, the team applied Corelet to manipulate with which people can delineate their proteins in several condensates. stickers and spacers. It takes into account post-translational modifications, which Stickers and spacers modify the strengths of the stickers and By iterating between experiment and spacers. Experimental measurements feed computational modeling, Rohit Pappu and into this computational pipeline, as do his team at Washington University developed scattering, single-molecule FRET, NMR and the ‘stickers and spacers’ model, a framework hydrogen deuterium exchange data. “The that enables one to describe and predict prospecting happens with computation,” he CasDrop7 from the Brangwynne lab is condensate formation on the basis of protein says. When researchers apply this framework an optogenetic approach to address genes attributes4. Intrinsically disordered regions of to computationally identify stickers and of choice in a condensate. One might, says proteins have long been considered drivers spacers, it helps them decide which protein Gladfelter, use it to place a condensate of condensate formation and maintenance. region to query experimentally. They in the nucleus. To study multivalency of The Pappu lab, the Rosen lab and others can map out condensate changes using proteins in a condensate, one might choose have shown that folded proteins play a role, calculated phase diagrams. Altering stickers Corelet8, also from the Brangwynne lab. too, says Pappu. Discussions over foldedness and spacers changes valency and thus

568 Nature Methods | VOL 17 | June 2020 | 567–570 | www.nature.com/naturemethods technology feature

shapes a dynamic in a condensate. “This is a biologists focus on amino acids as the level Nuclear Signalling speckle field where you cannot become a one-trick that evolutionary forces act upon. There may granule Cajal pony,” says Pappu. Given the diversity of be a ‘state of the polymer’ as a function of body DNA challenges, one needs a “kitchen sink” individual amino acids. “We need to try to damage of methods. Yet an individual lab might understand the range of protein dynamics Receptor Stress focus cluster granule P-body Balbiani not have the time, resources and skills to in a particular context,” says Hyman. New body readily use, for example, small-angle X-ray types of FRET measurement may help, as Nucleolus Germ scattering (SAXS), NMR and hydrogen– might label-free imaging techniques such granule deuterium exchange. “You don’t want to as the light-scattering process that say to every cell biologist: you have to do underlies Brillouin microscopy and other The cytoplasm and nucleoplasm of eukaryotic NMR and SAXS and single-molecule FRET,” approaches used to study the elasticity of cells reveal many types of condensates that he says. Perhaps, he says, this predictive condensed matter. sequester a variety of biochemical reactions. computational engine can lead to ‘codifiable What is emerging is that in the Credit: Pappu Lab, Washington U. Reproduced rules’ to target experiments and reveal the interactions between proteins and RNAs, says with permission from Annu. Rev. Biophys. 49 molecular grammar of condensates. Brangwynne, sometimes disordered regions © 2020 Annual Reviews. Stickers and spacers is a useful in proteins play more modulatory roles. representation and a way to reduce In his work, he draws on network-based the complexity of factors that shape a methods and concepts from soft matter Lemke, is fueled by an interest in biophysics condensate and prepare how to test them physics. “It’s about connectivity,” he says. and the need for useful in-cell fluorescent experimentally, says Hyman. A sticker, says “The field has been pretty protein-centric for labels for studying intrinsically disordered Brangwynne, has affinity for one type of a while,” he says. “We’re to blame for that, and proteins. The contrast in imaging readouts molecule and a different affinity for another. others as well,” he says. RNA can modulate a can be faint in experiments that involve “You put them all together and ask: what condensate’s protein–protein network, “but genetic code expansion, in which synthetic does the network of interactions look like?” RNA can play an important scaffolding role.” biologists expand beyond the standard This is one way condensate research draws Ultimately one needs a sufficient number protein repertoire of 20 amino acids. Cells do on approaches in . of interactions, he says, which is where the not readily use non-natural amino acids, and concept of valency comes in. off-target effects result. “It’s just the tool that RNAs, too works really well for chemists, but it really As condensate research emerged, RNAs were Applied condensates doesn't work that well for biologists yet,” rather neglected, says Gladfelter, but that is Condensates might propel research on says Lemke. It’s challenging to encapsulate changing10,11. “It’s pretty fair game to say that neurodegeneration and cancer. Hyman this different and complex translation. almost any RNA molecule is in a condensate and Whitehead Institute researcher Rick That’s when it occurred to him to use phase at some moment in time in its life.” Proteins Young cofounded Dewpoint Therapeutics, separation to do so. He and his team believe interact with one another and with RNAs, a company based in Boston and Dresden. the system is scalable. They used FUS, a which renders condensate dynamics even In November 2019, Bayer invested $100 well-studied protein that always phase more complex, says Hyman. Getting at this million in the company. In a statement, separates, and then they engineered a binding richness takes input from different fields: Joerg Moeller, who directs research and site for their intended RNA. The system is molecular and cell biology, chemistry, development in Bayer’s Pharmaceuticals readily reproducible, he says, and it recruits physical chemistry and polymer physics, he Division, said new analytic tools and a needed components into the system. A lack says. Whereas labs working in physics and growing understanding of biomolecular of membrane means ribosomes can “scoot polymer physics are happy with the concept condensates could shed light on cellular over.” Although a ribosome or a tRNA might of ‘states’ that a material can have, molecular functions that drug development scientists leave the designated area, the concentration have not previously considered and help gradient prevents that for the most part. to identify pharmacological targets. “Basically our organelle is a heat-sink for our Hyman believes condensates can be tRNA,” he says. It’s where the tRNA’s codon used to explore a drug candidate’s cell is translated differently than outside this biological mechanisms, to characterize area. The scientists made many binding sites Folded Intrinsically proteins disordered the and kinetics of the within a droplet for the tRNA by actively proteins condensate and the actions of a drug whose targeting its synthetase, the enzyme that target might be in a condensate. catalyzes the reaction to charge a tRNA with For Edward Lemke, condensates were a specific amino acid. The approach shows part of the solution to a problem he and his that translation likely involves around 100 team at the European Molecular Biology factors, which travel into and then out of Laboratory, Johannes Gutenberg University the condensate. It shows “you can basically and Institute of Molecular Biology Mainz switch how translation works, simply by were trying to solve: how to precisely position phase separating,” he says. “Chemistry inside Stickers Spacers non-canonical amino acids in engineered the phase-separated organelle can be very, proteins12. They used condensates to bring very, very different from the cytoplasm.” The Pappu lab developed the ‘stickers and together, in a eukaryotic cell, the components The approach to expand the genetic spacers’ model for computationally predicting of an ‘orthogonally translating synthetic lexicon beyond 20 amino acids was condensates on the basis of protein attributes. designer organelle’. The team targets a developed by Peter Schultz at Scripps Credit: Pappu Lab, Washington U. Reproduced particular tRNA and mRNA to an assembly Research and University of California, with permission from Annu. Rev. Biophys. that also contains proteins, such as FUS and Berkeley and his team, and it’s a powerful 49 © 2020 Annual Reviews. kinesin motor proteins. His motivation, says synthetic biology tool, says Lemke. In

Nature Methods | VOL 17 | June 2020 | 567–570 | www.nature.com/naturemethods 569 technology feature

raises some valid points. There have been some “handwavy” papers that only purport but do not test the validity of a quantitative biophysical framework. “If we’re trying to test quantitative biophysical models, let’s use quantitative biophysical approaches to do it.” It’s insufficient, for example, to state that one has detected a “diffraction-limited spot, so it must be phase separated.” The condensate field needs collaboration across disciplinary divides, which is never easy, says Hyman. He has built a group with team leaders from different disciplines. He notices that researchers might use the same words that have different meanings and that different words can mean the same thing in different fields. “Right now it’s a little bit of a Tower of Babel,” he says. Over time and with successful collaborations, this will change. Lemke’s group includes physicists, chemists, biologists and bioengineers. Half of his group does biophysics and the other half chemical biology and synthetic biology. Some members of the Gladfelter lab joined the lab as mathematical modelers, others focus on in vitro work for the biophysical analysis of phase separation The cellular factory makes proteins with canonical amino acids. Within a condensate, the Lemke and others do in vivo experiments, such lab built an artificial organelle to also make proteins with noncanonical amino acids. Credit: E. Lemke, as confocal image analysis in live cells. She G. Estrada Girona, IMB/U Mainz/EMBL trains them each to be involved in at least two of those areas. One of her students, an applied mathematician who built a this experiment, Lemke’s lab tweaked the membrane. “That’s probably, I would guess, condensate model, has been trying his hand approach to work well in eukaryotes, where how the nucleus must have evolved,” he says. at experiments and doing well, she says. codon specificity has been a challenge. The field needs better ways to explore the “It’s always easier to engineer an organelle Beyond a Tower of Babel compositional complexity of condensates than a eukaryotic host,” he says. “We In a recent paper13, Robert Tjian at the in situ. “There’s a lot of challenges ahead decoupled host engineering from organelle University of California, Berkeley and that necessitate technical development,” she engineering.” The approach opens up many colleagues point out that various mechanisms says. Concerning the fact that labs are using possibilities for work in eukaryotic cells, can lead to condensates and that the tests many different methods to characterize says Lemke. More label colors are needed commonly used to probe liquid–liquid phase condensates, “I think that’s actually one of to track proteins, and off-target issues separation (LLPS) “are insufficient to rule out the cool things about the field.” ❐ need to be avoided. The artificial organelle other mechanistic interpretations.” Among should be smaller and less toxic to the cell, other aspects, the team notes, “we urge the Vivien Marx ✉ and he has ideas how to make that happen. application of more stringent criteria and Technology editor for Nature Methods. “Then I think the actual cell biologist will more appropriate experimental approaches ✉e-mail: [email protected] appreciate the technology that we have,” to understand the functional role of LLPS he says. Lemke launched a startup, Araxa condensates in cellular organization.” Published online: 20 May 2020 Biosciences, focused on protein engineering Although some in the condensate https://doi.org/10.1038/s41592-020-0855-3 tools for antibody therapeutics, where he community have reacted with irritation, is chief scientific officer. The company has Hyman calls the paper “perfectly legitimate” References an investor and it licensed the organelle in its advice for caution. “He asks some very 1. Brangwynne, C. P. et al. Science 324, 1729–1732 (2009). 2. Boeynaems, S. et al. Trends Cell Biol. 28, 420–435 (2018). technology from his lab. Perhaps there is good questions,” says Hyman, such as what 3. Alberti, S., Gladfelter, A. & Mittag, T. Cell 176, 419–434 (2019). no limit to the complexity of the tasks an confers size stability on condensates in the 4. Choi, J.-M., Holehouse, A. S. & Pappu, R. V. Annu. Rev. Biophys. organelle can pull off. Translation is not just a nucleus, which is not trivial from a physics 49, 107–133 (2020). 5. Hyman, A. A., Weber, C. A. & Jülicher, F. Annu. Rev. Cell Dev. simple enzymatic reaction, he says. Perhaps, point of view. Hyman says he reminds his lab Biol. 30, 39–58 (2014). he speculates, the method of sequestering that “the last thing you need is cheerleaders.” 6. Shin, Y. et al. Cell 168 159–171.e14 (2017). reactions in condensates has been a Rather, it’s naysayers who push researchers 7. Shin, Y. et al. Cell 175 1481–1491.e13 (2018). 8. Bracha, D. et al. Cell 175 1467–1480.e13 (2018). driving force in evolution in eukaryotic to keep devising experiments to drive insight 9. Riback, J. A. et al. Nature https://doi.org/10.1038/s41586-020- cells. Membrane-bound organelles likely to answer the question. He and Brangwynne 2256-2 (2020). emerged when cytoplasmic housekeeping saw condensates for the first time in 2009. 10. Langdon, E. M. et al. Science 360, 922–927 (2018). needed more selectivity. For example, the Over time, “you begin to revise things you 11. Deniz, A. A. Cell 181, 228–230 (2020). 12. Reinkemeier, C. D. et al. Science 363, 1415 (2019). lysosome, the cell’s waste disposal system, didn’t quite understand in the beginning.” 13. McSwiggen, D. T., Mir, M., Darzacq, X. & Tjian, R. Genes Dev. 33, has advantages when it is encapsulated by a Brangwynne agrees that the Tjian team 1619–1634 (2019).

570 Nature Methods | VOL 17 | June 2020 | 567–570 | www.nature.com/naturemethods