An interview with Mingjie Zhang: phase separation in biological systems

Abstract Mingjie Zhang is a distinguished structural biologist whose research interest is the protein assemblies in cells, particularly those at the synapse, where two neurons meet and transmit nervous impulses from one to the other. In recent years, his group discovered the phenomenon of phase separation in both the pre- and post-synaptic sites of neurons and elucidated its regulation and physiological function. In this interview, NSR Executive Editor-in-Chief Mu-ming Poo talks with Prof. Zhang on the fast-developing research area of biological phase separation, as well as Zhang's new role as the Founding Dean of the School of Life Sciences at Southern University of Science and Technology (SUSTech) in Shenzhen.

Zhang: Phase separation is a well-established phenomenon in physical sciences, and has been observed for a long time in biology. More than 100 years ago, Ramón Cajal discovered condensed bodies in the nucleus, which are now known as Cajal bodies and their formation is a phase separation process.
The 2009 paper published by Anthony Hyman and Clifford Brangwynne [Brangwynne C et al. Science 2009;324: 1729-32] is probably the landmark paper that made biological phase separation a hot topic. They showed that in the developing C. elegans embryo, P granules form in a certain region via phase separation and this process is important for embryo development, making a connection between phase separation and physiological functions. Then suddenly, many people began to think that more biological phenomena that have been observed may be linked to phase separation.

Poo: What exactly is biological phase separation?
Zhang: In cells or other biological systems, biological molecules, such as proteins and nucleic acids, are able to spontaneously assemble into a state with lower entropy or higher order, forming a thermodynamic equilibrium with a condensed phase and a dilute phase co-existing in a system. That is the phenomenon of biological phase separation.
The phenomenon is exactly the same as when you pour black vinegar into olive oil. You will see beautifully separated two liquid phases. But in biological systems, the beauty is that the process happens in aqueous solution under biological conditionsneutral pH, as well as physiological salt and biomolecule concentrations.
Importantly, some of the condensed phases have biological functions and are considered to be non-membranous organelles. And these organelles can interact or communicate with the membranous organelles, thus greatly expanding the number of organelles that a cell can use.

Poo:
In biology, we have learnt about lipid phase separation within the cell membrane for a long time. We knew that cholesterol can be enriched in one domain to form a 'lipid raft'. But phase separation occurring in aqueous cytoplasm is a relatively new idea.
Zhang: That's right. In the past, it has been a little bit difficult to imagine that soluble cytoplasmic molecules can condense into droplet-like phases. separation-physiological phase separation that is functional and highly reversible, and the non-reversible phase separation that will very likely lead to functional defects. Disease-related aggregates of beta amyloids or short peptide repeats-containing proteins (e.g. polyQ, polyGR, polyPR, etc.) belong to the category of non-reversible phase separation, sometimes referred to as phase transition.
Poo: What is the major driving force of biological phase separation, specific inter-molecular interaction or non-specific interaction?
Zhang: There are actually two views on this issue. Some people have been saying that phase separation is driven by intrinsically disordered proteins. Another camp, which is relatively a minority now but is also my view, is that biological phase separation is driven by the combination of specific molecular interactions as well as non-specific interactions among intrinsically-disorderedsequence-containing proteins. From a thermodynamic point of view, intrinsically disordered proteins provide the general interaction driven by free energy gain, while specific molecular interactions provide specificity of condensates formation.
According to polymer chemistry theory, any polymer, including proteins, nucleic acids and their mixtures, is able to undergo phase separation under certain conditions. But in the biological systems, specific interactions are needed to drive the formation of condensed phases under physiological conditions. The specific interactions often involve multiple interaction sites and can form a network of molecules. When the network reaches a certain complexity, phase separation will happen spontaneously. Also, specific interactions can provide more space for the regulation of the condensed phases.
Poo: Biological systems need energy input to maintain their structural order and functions. Do the formation and regulation of phase separation need energy input?
Zhang: Phase separation per se does not need energy input. It happens spontaneously once the biomolecule concentration reaches a certain threshold. But the synthesis and turnover of the biomolecules, which build up the concentration, need energy. It's the same for regulation. The modification or concentration change of certain biomolecules needs energy input, and after that, the modulation of the condensed phase formation or dispersion would happen spontaneously.

Poo: Phase separation is useful for cells because it can enrich certain molecules to a specific location. But if this localization
Phase separation is a very good way for cells to increase the efficiency and specificity of signal transduction. It will be surprising if cells do not use this kind of strategy more often than we have realized.
-Mingjie Zhang Phase separation in cell, showing the interactions and communications among membranous organelles and non-membranous organelles. Depicted by Xiandeng Wu.
process can be accomplished merely by specific molecular interactions, then why do cells need phase separation?
Zhang: That's a really great question. The point is that, if the molecules are purely localized and enriched by specific interactions, their distribution would be governed by the diffusion law, meaning that there will be a concentration gradient throughout the cell. That means the cell has to synthesize a large amount of molecules in order to fulfill the requirement in some specific locations. But with phase separation, the concentration difference between the two phases can be as large as several hundred-fold, or even larger, so we do not need that many molecules to be synthesized. This is just one example. There are many features of biological condensates formed via phase separation and these features cannot be explained by the canonical stoichiometric and specific molecular interactions occurring in dilute solution.

Examples
Poo: Would you give some examples of biological phase separation?

Zhang:
One famous example is that in a fertilized egg, maternal and paternal materials are deposited in very condensed phases and are located in specific areas within the egg. That is important for embryo development. Another example is, in a 2002 paper [Galkin O et al. PNAS 2002;99: 8479-83], the authors talked about hemoglobin phase separation in red cells. They found that with gene mutation in sickle cell anemia patients, hemoglobin molecules undergo liquid-liquid phase separation much more easily and are more likely to form polymers that are found in sickle cells. That was a beautiful paper that linked phase separation with human diseases.
In the nucleus, one theory is that the partitioning of chromatin, or the formation of euchromatin and heterochromatin, is a process of phase separation. But there is strong resistance to this theory and we are still unsure if it's a real phase separation process or not. I think if it's proven to be true, it will provide many new angles in understanding chromosome organization and regulation.

Poo: You had been a structural biologist and bumped into phase separation at the synapse in recent years. How did this happen?
Zhang: In the last 25 years I have been trying to understand how synapses form and work. When looking at the electron microscope pictures of the synapse, you can see a high density of protein assembly right beneath the postsynaptic membrane. This protein assembly, which is called postsynaptic density (PSD), is so stable that it can be chemically purified, and so dynamic that it can be modulated by changes in synaptic activity. I really wanted to understand the chemical, structural and physical basis of the formation and regulation of PSD, and that has been the central theme of research in my lab.
We meticulously studied the molecular interactions of all major PSD proteins. When we were studying the interaction between PSD-95 and SynGAP, two major proteins in PSD, one of my students told me that when mixing these two proteins, some droplet-like phenomenon was observed. That was really a surprise and we suddenly realized that it may be a phase separation process that drives the formation of PSD. That was our initial discovery [Zeng M et al. Cell 2016;166: 1163-75].

Poo: Would you explain more about PSD?
Zhang: PSDs were observed as electron-dense thickenings beneath postsynaptic plasma membranes and thus are open to the cytoplasm of dendritic spines. PSD contains many different proteins, and these proteins form interconnected disc-shaped molecular assemblies. Based on our studies, such disc-shaped molecular assemblies of PSD are likely formed via phase separation. Importantly, the PSD assemblies are highly dynamic in response to synaptic activity changes. Mechanistically, the dynamic PSD assemblies are regulated by various enzymes such as Ca 2+ -calmodulin protein kinase II and protein phosphatases.

Poo:
Are there phase separation events at the presynaptic terminal?
Zhang: Yes. In mature synapses, the presynaptic structures are fairly stable and contain a huge number of synaptic vesicles (SVs). These SVs mostly exist in two pools-a reserve pool in the presynaptic bouton, and a readily releasable pool docked on the active zone right above the presynaptic membrane. Both pools are condensed phase organized by phase separation, and SVs can exchange between these two pools.
The reserve pool is mainly organized by a protein called synapsin, which can undergo phase separation to form a condensed phase. Moreover, synapsin directly interacts with the SV For the field of biological phase separation, we urgently need proper theory to guide future studies.
-Mingjie Zhang surface, thus clustering the SVs into the condensed phase, forming coacervated condensates. On the other hand, the active zone is a condensate formed by a different set of proteins, including RIM and RIM-BP. SVs can physically coat the surface of RIM-RIM-BP, thus docking on the active zone. Actually, this phenomenon perfectly explains the observation that the number of docked SVs in each bouton is proportional to the area of active zone.
I think the presynaptic terminal is a wonderful example showing different modes of interactions and communications between membranous organelles and non-membranous organelles. In this case, they can also be regulated by physiological activity, namely the action potentials.

Directions
Poo: Phase separation can localize a large number of biomolecules in a very specific region of the cell. I guess many signal transduction pathways could be carried out in the separated phases. Is there any evidence for this?
Zhang: There are actually a number of studies being done in this direction. For example, in receptor tyrosine kinase signaling, people have produced evidence that phase separation can help with pathway efficiency and specificity.
In our study, we found that two enzymes, GIT and β-PIX, interact with each other very strongly and specifically. In living cells, they are able to form condensates, via phase separation, that act as modular functional entities. In different cell types, the condensates respond to different upstream signals and perform different activities. In synapses, they bind to Shank; in cell-cell junctions, they bind to Scribble; and during cell migration, they respond to the integrin pathway by binding to paxillin.
I think phase separation is a very good way for cells to increase the efficiency and specificity of signal transduction. It will be surprising if cells do not use this kind of strategy more often than we have realized.
Poo: What do you think will be the future direction of biological phase separation research? Zhang: I have always believed that any biological process has to obey physics laws. All the theories governing molecular interaction that we use in biology are based on dilute solution systems, but now, phase separation brings biological systems into the realm of soft matter physics. However, currently available soft matter physics theories break down in the living cells due to the complexity.
My own training in physics is limited, so I think moving forward, one of the very important directions for myself is that I need to collaborate with people who understand soft matter physics. For the field of biological phase separation, we urgently need proper theory to guide future studies.