In this issue we feature several articles that explore advances in the study of phase separation. They highlight some recently reported mechanistic features and progress in the methodology used to study it within cells, and they delve into the implications that phase separation has for select cellular functions.

The spontaneous separation of biomolecules, such as proteins and nucleic acids, from a homogeneous mixture into two distinct liquid phases in which they have different concentrations is known as biomolecular phase separation. This compartmentalization facilitates select biochemical processes and is important for various cellular functions. Efforts to understand the underlying molecular mechanisms of phase separation and how it influences biological processes have been underway since it was first observed in biology over 15 years ago1. Numerous reports highlight that various factors promote this process, such as the regulation of bulk water2, the presence of intrinsically disordered regions and a selection of key non-covalent interactions from contributing amino acids3,4,5,6. Yet, progress in methodology and additional molecular-level insight continues to emerge, both of which we feature in this issue.

For some perspective on the field, we quizzed a leading researcher in a Q&A. From it we learn how Tanja Mittag’s training on protein NMR — alongside other biophysical and biochemical techniques — to study the dynamics underlying protein–ligand binding primed her to study phase separation in her independent career. Her group studies how the properties of select proteins influence their phase separation and ultimately how this impacts their biological function. She notes that we have a general understanding of how disorder, sequence identity, valency and phase interfaces influence this physical transition. However, despite knowing that phase separation facilitates many fundamental biological processes, Mittag highlights that the underlying mechanisms are not entirely clear. She notes that the theoretical frameworks derived from polymer physics have been useful for describing biomolecular phase separation, but methods that can facilitate the quantitative determination of condensate properties — both in vitro and in cells — will be important to decode the richer phase behaviours of more complex systems and potentially derive understanding of new physics.

Studying biomolecular, proteome-wide condensate formation in cells is challenging because isolating such complex and dynamic components at the endogenous level is non-trivial. To circumvent these issues, Bi-Feng Liu, Yiwei Li and colleagues develop in their Article a platform for identifying endogenously expressed biomolecular condensates by modulating phase separation using volumetric compression, which tunes protein concentrations and molecular crowdedness. They tracked changes in the partition of proteins into condensates using gradient ultracentrifugation coupled to quantitative mass spectrometry. They found that more than 16% of the proteome with involvement across a broad range of biological functions is responsive to phase separation and identified more than 500 endogenous-expression condensate proteins that were previously unreported. In the accompanying Research Briefing, expert Sindhuja Sridharan notes that this platform complements existing system-wide approaches and presents an orthogonal strategy to microscopy-based readouts.

Scaling down to the amino acid level, an Article by Benjamin Schuster, Kristi Kiick, Jeetain Mittal and colleagues explores whether previously overlooked interactions in an intrinsically disordered protein’s (IDP) sequence are important for the formation and material properties of condensates. Through molecular dynamics simulations and in vitro phase separation studies, the authors identified that previously overlooked polar and non-polar residues can stabilize the condensed phase and that multiple interactions from diverse residue pairs drive phase separation; that is, it is the sum of interactions that tip the balance towards separation. They found that all condensates behave similarly to viscous fluids but with different viscosities, which demonstrates that interactions that drive phase separation may not necessarily contribute to rheology.

Meanwhile, in related Articles from Mittal, Matthew Good and colleagues and José Avalos, Clifford Brangwynne and colleagues — also covered in an associated News & Views by Michael Phillips and Kingshuk Ghosh — the key principles that regulate whether different macromolecules undergo co-partitioning or exclusion during phase separation are characterized both in solution and in cells. In their Article, Mittal, Good and colleagues studied two IDPs that selectively partition both in cells and in solution, suggesting that this process is sequence-encoded and does not result from cellular factors. Through a combination of modelling and mutagenesis, they identified how amino acids that are key to selectivity can be mutated to promote co-partitioning between two IDPs and showed that charge partitioning is a major factor in intrinsically disordered region mixing. In a separate but complementary Article, Avalos, Brangwynne and colleagues investigated how the oligomerization of two different IDPs during condensation influences exclusion versus association. They found that an IDP with a higher state of oligomerization affords enhanced immiscibility and multiphase formation; that is, there is an asymmetry when two IDPs have different homotypic interaction strengths.

In their News and Views article, Phillips and Ghosh highlight that homotypic and heterotypic interactions in condensates can be designed by modulating two reported parameters: increasing oligomerization and sequence alteration. They also note that the two studies from Mittal, Good and colleagues and Avalos, Brangwynne and colleagues highlight that understanding how equilibrium physics principles applied to living systems that operate out of equilibrium can help narrow the knowledge gap. This agrees with Mittag’s insight that such efforts could expand our understanding of new physics.

Looking towards an interesting heterotypic example in biology, an Article from Paolo Arosio and colleagues characterizes how protein–RNA interactions tune both condensate formation and the liquid-to-amyloid transition of a protein involved in amyothropic lateral sclerosis (hnRNPA1A). The authors find that in the absence of RNA, hnRNPA1A condensate formation potentiates its aggregation, leading to fibril formation. Meanwhile, they also show that the presence of RNA tunes the transition of hnRNPA1A amyloids from their soluble form through different pathways, which depend on the RNA:protein stoichiometry. They find that when RNA concentrations are low, condensation and amyloid formation are both promoted such that RNA plays a sort of catalytic role at the condensate interface. By contrast, high RNA concentrations suppress condensate formation — through the re-entrant phase behaviour — but hnRNPA1A fibrils eventually form after longer incubation times. Interestingly, they found that RNAs of different lengths and sequence identity had similar effects on these phase transitions. Despite evidence suggesting a generic RNA effect here, Arosio and colleagues highlight that additional sequences and structures of RNA would need to be tested to better characterize this theory.

The past 15 years have afforded substantial advances in our understanding of biomolecular phase separation that highlight that many factors are at play for this physical process to take place, such as biomolecule sequence, structure and dynamics, and the homotypic versus heterotypic interactions that are involved. Although tying these molecular insights back to specific cellular functions remains challenging, advances in experimental methodologies and theoretical frameworks will continue to help fill these knowledge gaps and tap into our curiosity of this intriguing interdisciplinary research area.