*Editorial* **Biological Liquid–Liquid Phase Separation, Biomolecular Condensates, and Membraneless Organelles: Now You See Me, Now You Don't**

**Vladimir N. Uversky**

Department of Molecular Medicine and Byrd Alzheimer's Research Institute, Morsani College of Medicine, University of South Florida, Tampa, FL 33612, USA; vuversky@usf.edu

Liquid–liquid phase separation (LLPS, also known as biomolecular condensation) and the related biogenesis of various membraneless organelles (MLOs) and biomolecular condensates (BMCs) are now considered fundamental molecular mechanisms governing the spatiotemporal organization of the intracellular space [1–5]. Physically, LLPS represents a special case of phase transition taking place in a homogeneous solution, i.e., a system including a solvent (in this case, water) and homogenously distributed solutes (e.g., proteins and nucleic acids). Under specific conditions, this homogeneous solution undergoes a process of spontaneous separation into two (or more) distinct immiscible liquids reflected in the emergence of dense and dilute phases that, respectively, contain more and less specific solutes. Although as a result of this phase transition, the chemical nature of solvent is not changed, and dense and dilute phases are still filled with water, the resulting separated phases are characterized by noticeable changes in the solvent properties of water [6]. Therefore, in addition to the consideration of the peculiarities of the interaction between different solutes undergoing LLPS, one should keep in mind that water plays a crucial role in biological LLPS and in the biogenesis of various MLOs [6].

When LLPS of biopolymers takes place in a cell, it leads to the emergence of MLOs or BMCs, which, being liquid droplets, represent specific compartments within a cell that are not enclosed by a lipid membrane [7–10]. Therefore, the biogenesis of MLOs is linked to the ability of biopolymers (e.g., intrinsically disordered proteins (IDPs) [11–13] and nucleic acids [14,15]) to separate into phases under specific conditions. LLPS is known to be controlled by various external factors and environmental cues, such as changes in the temperature, pH, ionic strength of the solution, posttranslational modifications [16], small molecules [17], and a number of other causes [18]. Furthermore, most LLPS processes are reversible, and many MLOs exist transiently and show "now you see me, now you don't" behavior, rapidly emerging when conditions are changed and rapidly disintegrating as soon as the original conditions are restored.

MLOs/BMCs are many (to date, about a hundred different MLOs are known [19]) and can be found in eukaryotes, archaea, bacteria, and, likely, viruses. They exist as liquid droplets (or cellular bodies, puncta, etc.) in the cytoplasm, nucleoplasm, mitochondrial matrix, and stroma of chloroplasts. Figure 1 displays the dramatic increase in the appreciation of the importance of MLOs in recent literature. Despite being almost unknown to researchers until quite recently, MLOs are rapidly becoming mainstream in modern cellular research. In line with this ever-increasing interest in MLOs, this Special Issue includes seven research articles and three reviews considering different aspects related to the LLPS, MLOs, and BMCs. These articles are briefly outlined below.

**Citation:** Uversky, V.N. Biological Liquid–Liquid Phase Separation, Biomolecular Condensates, and Membraneless Organelles: Now You See Me, Now You Don't. *Int. J. Mol. Sci.* **2023**, *24*, 13150. https://doi.org/ 10.3390/ijms241713150

Received: 10 August 2023 Accepted: 18 August 2023 Published: 24 August 2023

**Copyright:** © 2023 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

analysis.

**Figure 1.** Increase in the number of publications dealing with the different aspects of membraneless organelles. Plot represents the output of a PubMed search for "membraneless organelles" conducted on 9 August 2023. The actual annual publication rate is shown by gray bars, whereas the hatched gray bar shows extrapolated data based on the number of papers published in 2023 by the date of **Figure 1.** Increase in the number of publications dealing with the different aspects of membraneless organelles. Plot represents the output of a PubMed search for "membraneless organelles" conducted on 9 August 2023. The actual annual publication rate is shown by gray bars, whereas the hatched gray bar shows extrapolated data based on the number of papers published in 2023 by the date of analysis.

In their research article, Mohtadin Hashemi, Siddhartha Banerjee, and Yuri L. Lyubchenko investigated the effects of membrane and free cholesterol on the early stages of the aggregation process of amyloid β (Aβ) [20]. The authors showed that the aggregation of this protein, which is related to the formation of neurotoxic species in Alzheimer's disease, is dramatically enhanced via direct Aβ-membrane interactions. Importantly, this interaction promotes oligomer assembly on the lipid bilayer at physiologically low concentrations of the Aβ monomer. The process is strongly dependent on the membrane composition, and the presence of cholesterol in the membranes significantly enhances the aggregation kinetics. Furthermore, in agreement with previous studies where the presence of free cholesterol in amyloid plaques was reported, the authors found that free cholesterol can further accelerate the aggregation process and promote fast formation of aggregates of significantly larger sizes (as evidenced by an in-solution time-lapse Atomic Force Microscopy (AFM) analysis). Furthermore, free cholesterol accelerated the dissociation of Aβ In their research article, Mohtadin Hashemi, Siddhartha Banerjee, and Yuri L. Lyubchenko investigated the effects of membrane and free cholesterol on the early stages of the aggregation process of amyloid β (Aβ) [20]. The authors showed that the aggregation of this protein, which is related to the formation of neurotoxic species in Alzheimer's disease, is dramatically enhanced via direct Aβ-membrane interactions. Importantly, this interaction promotes oligomer assembly on the lipid bilayer at physiologically low concentrations of the Aβ monomer. The process is strongly dependent on the membrane composition, and the presence of cholesterol in the membranes significantly enhances the aggregation kinetics. Furthermore, in agreement with previous studies where the presence of free cholesterol in amyloid plaques was reported, the authors found that free cholesterol can further accelerate the aggregation process and promote fast formation of aggregates of significantly larger sizes (as evidenced by an in-solution time-lapse Atomic Force Microscopy (AFM) analysis). Furthermore, free cholesterol accelerated the dissociation of Aβ oligomers from the surface and their accumulation in bulk solution [20].

oligomers from the surface and their accumulation in bulk solution [20]. Juliet F. Nilsson, Hakima Baroudi, Frank Gondelaud, Giulia Pesce, Christophe Bignon, Denis Ptchelkine, Joseph Chamieh, Hervé Cottet, Andrey V. Kajava, and Sonia Longhi also studied the formation of amyloid fibrils, but analyzed the aggregation behavior of an interesting viral protein, the phosphoprotein (P protein) from the Nipah and Hendra viruses (NiV and HeV), which is an essential polymerase cofactor [21]. A peculiar feature of this system is that the gene encoding the P protein in both NiV and HeV also encodes the V and W proteins. These P, V, and W proteins have unique C-terminal domains (CTD) but share an intrinsically disordered N-terminal domain (NTD). A short PTN3 region found within the shared NTD is capable of amyloid-like structure formation. The HeV PNT3 containing the amyloidogenic motif (EYYY) was used to analyze the relevance of each of three contiguous tyrosine residues to the fibrillation process. This analysis revealed that the ability to form fibrils is dramatically reduced by the removal of a single tyrosine independently of its position. Furthermore, the authors showed that PNT3 fibrillation can be regulated by the C-terminal half of this protein, showing an inhibitory effect Juliet F. Nilsson, Hakima Baroudi, Frank Gondelaud, Giulia Pesce, Christophe Bignon, Denis Ptchelkine, Joseph Chamieh, Hervé Cottet, Andrey V. Kajava, and Sonia Longhi also studied the formation of amyloid fibrils, but analyzed the aggregation behavior of an interesting viral protein, the phosphoprotein (P protein) from the Nipah and Hendra viruses (NiV and HeV), which is an essential polymerase cofactor [21]. A peculiar feature of this system is that the gene encoding the P protein in both NiV and HeV also encodes the V and W proteins. These P, V, and W proteins have unique C-terminal domains (CTD) but share an intrinsically disordered N-terminal domain (NTD). A short PTN3 region found within the shared NTD is capable of amyloid-like structure formation. The HeV PNT3 containing the amyloidogenic motif (EYYY) was used to analyze the relevance of each of three contiguous tyrosine residues to the fibrillation process. This analysis revealed that the ability to form fibrils is dramatically reduced by the removal of a single tyrosine independently of its position. Furthermore, the authors showed that PNT3 fibrillation can be regulated by the C-terminal half of this protein, showing an inhibitory effect on fibril formation [21].

on fibril formation [21].

Aleksandra E. Badaczewska-Dawid, Vladimir N. Uversky, and Davit A. Potoyan reported the development of a convenient web platform, Bioinformatics Analysis of LLPS Sequences (BIAPSS) [22]. The need for such a tool is based on the premises that despite a broad acceptance of the importance of biological LLPS, MLOs, and BMCs, there is a remarkable gap in the current knowledge which prevents a complete understanding of the sequence "codes" of phase separation required for the design of new phase-separating sequences of fundamental, medical, and technological importance. Therefore, the goals of this tool were to enhance apprehension of the interplay between the primary sequences of proteins and their capability to undergo spontaneous LLPS and thereby to uncover the sequence-encoded signals of the LLPS potential and related biogenesis of numerous functional MLOs and cellular bodies [22]. Researchers can use this web server in on-the-fly analysis as BIAPSS provides a useful tool for the visualization and interpretation of the physicochemical and structural features for the superset of curated LLPS proteins [22].

Yoon-Jeong Choi, Yujin Lee, Yuxi Lin, Yunseok Heo, Young-Ho Lee, and Kiwon Song dedicated their research article to the investigation of a P-body-associated protein from *Saccharomyces cerevisiae*, Nst1 [23]. The authors emphasized that not all highly promiscuous proteins found within MLOs are made equal, with only a few of them acting as key scaffolds. Being capable of inducing condensation of the components of P-bodies (PBs), the multidomain scaffolding protein Nst1 serves as an important subject for targeted analysis of the roles of different domains in biomolecular condensation processes. To this end, a series of Nst1 domain deletion mutations were prepared and investigated [23]. This analysis revealed several important features. For example, deletion of the aggregation-prone domain (APD) significantly inhibited self-condensation, whereas the Nst1 mutant with the deleted multivalent polyampholyte domain (PD) within the intrinsically disordered region (Nst1∆PD) was able to form self-condensates but failed to interact and condensate Dcp2, a decapping protein and PB component. Furthermore, the Nst1∆PD deletion mutant was also unable to condensate other PB components, such as Xrn1, Dhh1, and Edc3, indicating that the PD of the IDR in Nst1 functions as a hub domain interacting with other PB components [23].

The research team of Tamami Miyagi, Rio Yamazaki, Koji Ueda, Satoshi Narumi, Yuhei Hayamizu, Hiroshi Uji-i, Masahiko Kuroda, and Kohsuke Kanekura investigated how the patterns of charged amino acids and the net charge within proteins undergoing LLPS determine their targeting of the specific MLOs, such as nuclear speckles and the nucleolus [24]. This was an important exercise aimed at understanding the basic mechanisms underlying the distributions of the LLPS-prone proteins with charged low-complexity domains (LCDs) to specific MLOs. The authors used proteins with Arg-enriched mixedcharge domains (R-MCDs) primarily composed of R and Asp (D) and known to accumulate in nuclear speckles via LLPS. They demonstrated that the distribution of R-MCD can be shifted from nuclear speckles to the nucleolus by redistributing their R and D residues from an alternately sequenced pattern to uneven blocky sequences [24]. Furthermore, it was established that the blocky R-MCD peptide showed affinity to RNA, acidic poly-Glu, and the acidic nucleolar protein nucleophosmin and was capable of efficient phase separation. On the other hand, the R-MCD peptide with alternating amino acids did not undergo LLPS. Furthermore, localization of the R-MCDs to the MLOs and their accumulation in the nucleolus were promoted by the incorporation of the basic residues into the R-MCDs. Based on these observations, it was concluded that the proximal positioning of D and R linked to the mutual neutralization of their charges is required for the distribution of proteins to nuclear speckles [24].

A series of two papers from the group headed by Ji-Long Liu is dedicated to the analysis of cytoophidia, filamentous structures formed by the CTP synthase (CTPS) [25,26]. Cytoophidia, or "cellular snakes", are evolutionary conserved MLOs found in the cells of many species in all three domains of life and represent an interesting case of metabolic regulation via enzyme filamentation and resulting compartmentalization [27,28]. In the first paper of this miniseries, Yi-Fan Fang, Yi-Lan Li, Xiao-Ming Li, and Ji-Long Liu used

fluorescence recovery after photobleaching (FRAP) to study the dynamic characteristics of cytoophidium in human cell lines and also utilized stimulated emission depletion (STED) microscopy to analyze the super-resolution structure of these cellular snakes [25]. This analysis revealed that cytoophidia are dynamic and reticular, with the reticular structure of CTPS cytoophidia potentially providing space for other components [25], such as inosine monophosphate dehydrogenase (IMPDH), another enzyme capable of cytoophidia formation [29]. Curiously, in their structural analysis, the authors also observed CTPS granules with tentacles [25].

In the second paper of the miniseries, Qiao-Qi Wang, Dong-Dong You, and Ji-Long Liu used the female reproductive system of Drosophila as a model for studying the physiological function of cytoophidia [26]. Utilization of a CTPSH355A mutant with a diminished cytoophidium-forming ability revealed the ingression and increased heterogeneity of follicle cells in the CTPSH355A egg chambers, indicating that cytoophidia may play a role in upholding the integrity of the follicle epithelium [26].

This Special Issue concludes with three interesting reviews regarding important biological implementations of LLPS. Lin Zhang, Shubo Wang, Wenmeng Wang, Jinming Shi, Daniel B. Stovall, Dangdang Li, and Guangchao Sui considered the associations between aberrant LLPS, misbehaving MLOs, and the development of various pathological conditions [30]. The authors systemized the properties of different MLOs and BMCs and summarized the multiple LLPS-regulated biological processes. They also emphasized that although normally functioning LLPS controls the biogenesis and composition of dozens of MLOs and BMCs in the cell, the onset and progression of various diseases, including neurodegenerative disorders and cancers, may be associated with (or even driven by) altered physiological conditions or genetic mutations, because of which phase-separated condensates may undergo aberrant formation, maturation, or gelation [30].

Woei Shyuan Ng, Hendrik Sielaff, and Ziqing Winston Zhao focused their review on phase-separation-mediated chromatin organization and dynamics [31]. The authors show how cells can utilize specific physico-chemical properties of chromatin-based phase condensates with various regulatory functions in a spatially and temporally controlled manner. In addition to presenting some key recent findings on the mechanistic roles of phase separation in regulating the organization and dynamics of chromatin-based molecular processes and illuminating the complex phase-separation-mediated interplay between chromatin and diverse chromatin-interacting molecular species, the authors also focused on the quantitative characterizations of these condensates using advanced imagingbased approaches. They emphasized that such phase-separation-mediated chromatin organization defines an emerging multifaceted, multimodal, and multiscale landscape responsible for the hierarchical regulation of the genome. The authors also discussed some deficiencies in existing studies and emphasized the need for multiparametric approaches for the in-depth characterization of chromatin-based phase separation in close-to-native cellular contexts [31].

Finally, Olga I. Povarova, Iuliia A. Antifeeva, Alexander V. Fonin, Konstantin K. Turoverov, and Irina M. Kuznetsova discussed the important roles of LLPS in the spatiotemporal regulation of the cytoskeleton assembly/disassembly, including the formation of actin filaments [32]. The authors argued that the LLPS leading to the formation of coacervates of actin-binding proteins can increase local concentration of G-actin and thereby initiate its polymerization. Furthermore, coacervates can acts as biological reactors, wherein, in addition to this local increase in the G-actin concentration, integration of proteins controlling actin polymerization, such as N-WASP, Arp2/3, and Cortactin, into such MLOs enhances the activity of these actin-binding proteins, thereby providing additional means for the efficient formation of actin filaments [32].

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** No new data were created or analyzed in this study.

**Conflicts of Interest:** The author declares no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.

#### **References**


**Disclaimer/Publisher's Note:** The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

## *Review* **The Role of Liquid–Liquid Phase Separation in Actin Polymerization**

**Olga I. Povarova † , Iuliia A. Antifeeva † , Alexander V. Fonin , Konstantin K. Turoverov and Irina M. Kuznetsova \***

> Laboratory of Structural Dynamics, Stability and Folding of Proteins, Institute of Cytology, Russian Academy of Sciences, 4 Tikhoretsky Ave., 194064 St. Petersburg, Russia

**\*** Correspondence: imk@incras.ru

† These authors contributed equally to this work.

**Abstract:** To date, it has been shown that the phenomenon of liquid–liquid phase separation (LLPS) underlies many seemingly completely different cellular processes. This provided a new idea of the spatiotemporal organization of the cell. The new paradigm makes it possible to provide answers to many long-standing, but still unresolved questions facing the researcher. In particular, spatiotemporal regulation of the assembly/disassembly of the cytoskeleton, including the formation of actin filaments, becomes clearer. To date, it has been shown that coacervates of actin-binding proteins that arise during the phase separation of the liquid–liquid type can integrate G-actin and thereby increase its concentration to initiate polymerization. It has also been shown that the activity intensification of actin-binding proteins that control actin polymerization, such as N-WASP and Arp2/3, can be caused by their integration into liquid droplet coacervates formed by signaling proteins on the inner side of the cell membrane.

**Keywords:** actin; actin polymerization; actin-binding proteins; liquid–liquid phase separation (LLPS); coacervate; membrane; signaling proteins

#### **1. Introduction**

Actin exists in nature in a monomeric globular form (G-actin) and a polymeric fibrillar form (F-actin). In the cytoplasm of cells, F-actin, which is the main part of the cytoskeleton, has a functional role. The actin cytoskeleton is necessary for a variety of processes in cells, including the establishment of cell polarity, the activation of cell migration, the launch of cytokinesis, and the positioning of intracellular organelles. Despite extensive research during the last 100 years, there is no complete understanding of the dynamic nature of the cytoskeleton and its regulation in cells and tissues. Recent work has shown a significant role of biomolecular condensates in the regulation of the formation, functioning, and assembly/disassembly of the cytoskeleton [1–5].

The presence of coacervates in protoplasm was also observed a long time ago [6]; however, the significance and universality of this phenomenon of the most diverse intracellular processes became clear only after the work of Brangwynne et al. [7]. The study of LLPS, and its role in cell life, is one of the most rapidly developing areas of molecular and cellular biology. It becomes obvious that the phase separation of biomacromolecules plays a significant role in almost all intracellular processes.

It is believed that intracellular liquid–liquid phase separation leads to the formation of coacervates (also called droplets, membraneless organelles, bodies, and granules), which do not have a membrane separating them from the rest of the cellular space, and which include proteins and nucleic acids. A key role in the formation of membraneless organelles is played by proteins containing intrinsically disordered regions (IDRs), which are mainly scaffold proteins of these structures [8,9].

**Citation:** Povarova, O.I.; Antifeeva, I.A.; Fonin, A.V.; Turoverov, K.K.; Kuznetsova, I.M. The Role of Liquid–Liquid Phase Separation in Actin Polymerization. *Int. J. Mol. Sci.* **2023**, *24*, 3281. https://doi.org/ 10.3390/ijms24043281

Academic Editor: Pavel Draber

Received: 5 January 2023 Revised: 30 January 2023 Accepted: 2 February 2023 Published: 7 February 2023

**Copyright:** © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

The IDRs of proteins are important for the fluid-like behavior of coacervates. Studies have shown that proteins rich in IDRs are more likely to separate into phases that retain the dynamic properties of the IDPs forming them [10]. Post-translational modifications (PTMs), such as phosphorylation and SUMOylation, also have a significant impact on LLPS, allowing cells to dynamically regulate LLPS in response to various cellular signals [11–13].

A high degree of multivalence, also inherent in IDPs, is necessary for recruiting other components into membraneless organelles and thus increasing their local concentration [14]. Liquid–liquid phase separation (LLPS) has been recognized as a general cellular mechanism of control for the spatiotemporal dynamics of many important signaling pathways [14,15].

The most important, functionally significant features of membraneless organelles are that only weak interactions are necessary for their formation, they are easily assembled and disassembled after performing their function, the absence of a rigid membrane allows them to freely exchange molecules with the environment, and weak impacts can significantly change their properties. At the same time, mature, membraneless organelles can be complex structures that include hundreds of protein molecules, including globular proteins and RNA [16]. The components of condensates are subdivided into scaffold proteins responsible for phase separation and client proteins, which are included in membraneless organelles to fulfill their function. In coacervates, the concentration of client proteins, including actin, can significantly exceed the concentration in the environment surrounding the organelle.

It is well known that a high concentration of actin is required in a solution to initiate the polymerization of actin monomers. This is primarily due to the fact that the formation of a polymerization nucleus requires the interaction of at least three actin monomers [17]. Coacervates working as biological reactors can contribute to an increase in the likelihood of this event.

#### **2. Model Coacervates as Reactors of Actin Polymerization Initiation**

The possibility of actin polymerization through increasing its concentration in model coacervates was shown for the first time in prior work [18], which aimed to demonstrate the possibility of creating a bioreactor that provides a significant increase in the reaction rate. It is known that charged homopolymers (polyelectrolytes) form liquid phases through complex coacervation [19] and can include charged proteins [20–22] and small molecules. To demonstrate the operation of a bioreactor based on poly-L-lysine (pLK) polycation and poly-(L, D)-glutamic acid (pRE) polyanion, two proteins were used: monomeric G-actin and bovine serum albumin (BSA) [18]. G-actin and BSA monomers are globular proteins; they are similar in size (42 and 66 kDa, respectively) and carry a comparable negative charge (isoelectric points of 5.23 and 5.60, respectively).

The proteins were visualized using fluorescently labeled protein-TMR-actin (green) and Alexa-647-BSA (orange). It turned out that both proteins accumulated in the created coacervates model; moreover, actin formed fibrils that were localized at the phase boundary, and BSA filled the coacervates evenly (Figure 1). Using fluorescently labeled phalloidin (Alexa Fluor 647-phalloidin)—a small, uncharged toxin known to bind specifically to F-actin—fibrils, formed by actin, near the coacervate surface were shown to be F-actin (Figure 1D). The extent to which the coacervate microenvironment influenced the rate of actin polymerization was investigated by recording pyrene fluorophore fluorescence.

A study of the polymerization of 1.5 mM of actin in a solution showed the presence of a characteristic lag phase, indicating a slow stage of filament nucleation [23] followed by a phase of rapid growth and then saturation after reaching a steady state [24]. At an actin concentration of 1.5 mM, the initial lag phase is usually 10 min, and the steady state is reached after 120 min. The presence of pLK/pRE coacervates eliminates the lag phase, and a steady state is established within 10 min. Thus, actin incorporation in pLK/pRE coacervates significantly stimulated its polymerization.

**Figure 1.** Actin polymerization in the coacervates formed by poly-L-lysine (pLK) and the polyanion poly-(L,D)-glutamic acid (pRE) at the phase boundary. (**A**,**B**) Microphotographs, obtained using confocal fluorescence microscopy (left and middle) and differential interference contrast (DIC) microscopy (right), show polypeptide coacervates containing TMR-actin (green) and Alexa 647-BSA (orange). The focal plane is at the interface of the coacervates and the substrate (surface (**A**)) or near the midplane of the droplet (**B**), indicated by the dashed yellow line. Scale bar, 5 mm. (**C**) Normalized intensity line scans are provided along the dashed yellow lines, as indicated in (**B**). (**D**) Falsecolored fluorescence images are provided in (**A**) and (**B**) from the surface (upper row) and midplane (bottom row). The figure represents panels (A), (B), and (D) of Figure 1 and panels (C) and (D) of Figure 2 from the article McCall et al., Biophysical Journal 114, 1636–1645, © 2018 [18]. **Figure 1.** Actin polymerization in the coacervates formed by poly-L-lysine (pLK) and the polyanion poly-(L,D)-glutamic acid (pRE) at the phase boundary. (**A**,**B**) Microphotographs, obtained using confocal fluorescence microscopy (left and middle) and differential interference contrast (DIC) microscopy (right), show polypeptide coacervates containing TMR-actin (green) and Alexa 647-BSA (orange). The focal plane is at the interface of the coacervates and the substrate (surface (**A**)) or near the midplane of the droplet (**B**), indicated by the dashed yellow line. Scale bar, 5 mm. (**C**) Normalized intensity line scans are provided along the dashed yellow lines, as indicated in (**B**). (**D**) False-colored fluorescence images are provided in (**A**,**B**) from the surface (upper row) and midplane (bottom row). The figure represents panels (A), (B), and (D) of Figure 1 and panels (C) and (D) of Figure 2 from the article McCall et al., Biophysical Journal 114, 1636–1645, © 2018 [18].

A study of the polymerization of 1.5 mM of actin in a solution showed the presence of a characteristic lag phase, indicating a slow stage of filament nucleation [23] followed by a phase of rapid growth and then saturation after reaching a steady state [24]. At an actin concentration of 1.5 mM, the initial lag phase is usually 10 min, and the steady state is reached after 120 min. The presence of pLK/pRE coacervates eliminates the lag phase, and a steady state is established within 10 min. Thus, actin incorporation in pLK/pRE coacervates significantly stimulated its polymerization. The authors showed that the mechanism underlying the increased assembly rate is an increase in the local concentration of actin in coacervates. The threshold monomer con-The authors showed that the mechanism underlying the increased assembly rate is an increase in the local concentration of actin in coacervates. The threshold monomer concentration/critical concentration required for the polymerization of Mg-ATP-actin is 0.1 mM [25]. If actin is concentrated 30-fold in coacervate droplets, then the expected assembly of actin within coacervates should be observed at total actin concentrations of 0.003 mM. However, F-actin formation in coacervates was experimentally observed at total actin concentrations of 0.05 mM. Interestingly, a peripheral shift of F-actin in pLK/pRE coacervates was observed at all actin concentrations studied—even at the lowest concentration (0.01 mM), at which the filaments are almost indistinguishable.

centration/critical concentration required for the polymerization of Mg-ATP-actin is 0.1 mM [25]. If actin is concentrated 30-fold in coacervate droplets, then the expected assembly of actin within coacervates should be observed at total actin concentrations of 0.003

A little later, actin polymerization was examined in a model of the two-phase poly(ethylene glycol) (PEG)/dextran system [26]. It was shown that globular actin was distributed evenly between the phases, while F-actin was concentrated inside the droplets and F-actin bundles were distributed along the periphery of the droplets, deforming them. In addition, it was shown in this work that, in the two-phase PEG/dextran system, the threshold concentration of KCl (which causes actin polymerization) decreased by about one order of magnitude, and the threshold concentration of MgCl<sup>2</sup> (which causes the formation of F-actin bundles) decreased by a third [26].

The authors of [18] consider three physical mechanisms of F-actin localization on the periphery of coacervates that do not exclude each other: the bulging of filaments, the depletion of macromolecules, and interfacial adsorption. However, it is noted that the results available to date do not provide a definitive answer for what determines the peripheral localization of F-actin; this could be an interesting problem for future research.

#### **3. Initiation of Actin Polymerization from abLIM1 Coacervates**

For the first time, the phase separation of the actin-binding protein, abLIM1, which leads to the formation of coacervates (i.e., reactors for the polymerization of cortical actin) was shown in work by Yang et al. [27]. The nonerythroid actin-binding protein, abLIM1, is known to be critical for stable interactions between the plasma membrane and the actin cortex under mechanical stress. This actin-binding protein has a C-terminal region that is 40% identical to dematin, an erythroid protein [28,29], which consists of a large, N-terminal, intrinsically disordered region (IDR) and a short actin-binding region. This region is identical to the villin headpiece (VHP) domain [30–33]. It is known that abLIM1 is expressed as three isoforms; namely, the long (abLIM-L, which is very rare), medium (abLIM-M, the widely expressed main isoform), and short (abLIM-S, a dematin-like isoform) isoforms [29]. These isoforms differ in a number of N-terminal LIM domains that mainly serve in protein–protein interactions [34]. The proteins abLIM1 and dematin are important for the stable attachment of the cortex to the plasma membrane in non-erythroid cells [28] and erythrocytes [31,35,36], respectively (Figure 2A).

Previously, it was shown that bacterially expressed His-tagged GFP-∆LIM (3 µM) induces the formation of dense networks of bound F-actin in vitro [28], such as dematin [30], and asters ~10 µm in diameter. Usually, asters contain an amorphous core involving His-GFP-∆LIM, from which numerous actin filaments radiate, decorated with GFP-∆LIM. It was shown that the asters' sizes depended upon ∆LIM concentration, and aster fibers contained densely packed actin filaments (Figure 2C). This allowed the authors of the work [27] to suggest that ∆LIM condensates could serve as the centers of the actin filament asters' organization in vitro. To prove this idea, it was necessary to rule out the presence of F-actin prior to aster formation (Figure 2C).

To this end, G-actin in a storage buffer (containing 0.2 mM of CaCl2) was mixed with GFP-∆LIM (6 µM) followed by an actin polymerization buffer. Then, TRITC phalloidin (tetramethylrhodamine B isothiocyanate) was used to visualize F-actin. It appeared that GFP-∆LIM-decorated actin bundles, especially GFP-∆LIM punctate condensates, were shown to gradually emerge after about 15 min and expand rapidly with time (Figure 2C).

The network (web) of actin fibrils turned out to be quite stable and remained preserved for hours without signs of disassembly. The web had a thickness of 6 to 7 µm and was located above an array of His-GFP-∆LIM condensates at the bottom of the substrate, while numerous F-actin asters of various sizes floated in the solution along with sporadic bundles. Thus, it was shown that ∆LIM was capable of the self-organization of asters and webs of actin bundles in vitro [27] (Figure 2F).

**Figure 2.** ΔLIM induces asters and webs of F-actin bundles in vitro. (**A**) Isoforms of human abLIM1. (**B**) Effects of GFP-ΔLIM on F-actin organizations in vitro. Arrows indicate representative astral structures. This presents the results of experiments in which the GFP-ΔLIM concentration was 0.1 μM. (**C**) The scheme of the experiment and the results of live imaging. The final concentration of Gactin was 6 μM. Phalloidin-TRITC (final concentration: 4 μM) was used to label F-actin. Representative time-lapse images showing the formation of His-GFP-ΔLIM-induced actin webs. Z-stack images, at 1 μm intervals, were captured for the GFP autofluorescence to cover a depth of 6 μm close to the bottom of the substratum via spinning disk microscopy at ~3.6 s intervals. The time started immediately after the addition of G-actin. Arrows denote two asters integrated into the web. (**D**) Massive actin polymerization from liquid droplets of GFP-ΔLIM. The experiments were performed without phalloidin-TRITC to show that the aster formation was not due to the presence of phalloidin. (**E**) Positive size–length correlations between the GFP-ΔLIM droplets and astral F-actin bundles. The size (volumes) of each droplet and the mean length of its astral bundles were measured from time-lapse images. (**F**) Illustrations showing ΔLIM (green)-induced F-actin-based asters, bundles, webs, and their relationships. The figure represents panels A of Figure 1; line 5 of panel B; panels E, F, and H of Figure 2; and panels A and C of Figure 4 from Yang et al., PNAS 119 (29), e2122420119 [27]. Copyright © 2022 the Yang et al., published by PNAS. This article is distributed under Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND). **Figure 2.** ∆LIM induces asters and webs of F-actin bundles in vitro. (**A**) Isoforms of human abLIM1. (**B**) Effects of GFP-∆LIM on F-actin organizations in vitro. Arrows indicate representative astral structures. This presents the results of experiments in which the GFP-∆LIM concentration was 0.1 µM. (**C**) The scheme of the experiment and the results of live imaging. The final concentration of G-actin was 6 µM. Phalloidin-TRITC (final concentration: 4 µM) was used to label F-actin. Representative time-lapse images showing the formation of His-GFP-∆LIM-induced actin webs. Z-stack images, at 1 µm intervals, were captured for the GFP autofluorescence to cover a depth of 6 µm close to the bottom of the substratum via spinning disk microscopy at ~3.6 s intervals. The time started immediately after the addition of G-actin. Arrows denote two asters integrated into the web. (**D**) Massive actin polymerization from liquid droplets of GFP-∆LIM. The experiments were performed without phalloidin-TRITC to show that the aster formation was not due to the presence of phalloidin. (**E**) Positive size–length correlations between the GFP-∆LIM droplets and astral F-actin bundles. The size (volumes) of each droplet and the mean length of its astral bundles were measured from time-lapse images. (**F**) Illustrations showing ∆LIM (green)-induced F-actin-based asters, bundles, webs, and their relationships. The figure represents panels A of Figure 1; line 5 of panel B; panels E, F, and H of Figure 2; and panels A and C of Figure 4 from Yang et al., PNAS 119 (29), e2122420119 [27]. Copyright © 2022 the Yang et al., published by PNAS. This article is distributed under Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND).

Previously, it was shown that bacterially expressed His-tagged GFP-ΔLIM (3 μM) induces the formation of dense networks of bound F-actin in vitro [28], such as dematin [30], and asters ~10 μm in diameter. Usually, asters contain an amorphous core involving His-GFP-ΔLIM, from which numerous actin filaments radiate, decorated with GFP-ΔLIM. A comprehensive study of the properties of ∆LIM condensates in astral centers has been made (Figure 2B), and it has been proven that they have the properties of coacervates. To this end, in particular, the change in the turbidity of the solution with a change in temperature was monitored [27]. Increasing the temperature of the GFP-∆LIM samples

It was shown that the asters' sizes depended upon ΔLIM concentration, and aster fibers contained densely packed actin filaments (Figure 2C). This allowed the authors of the

from 0 ◦C to 25 ◦C over 5 min resulted in an increase in turbidity. Moreover, the turbidity was reversible [27]. In addition, droplets were shown to increase in both size and density depending on the concentration of GFP-∆LIM and could be clearly observed at a concentration of 1 µM. They could easily merge into larger ones, which also confirms their liquid properties. Fluorescence recovery after photobleaching (FRAP) assays revealed a rapid recovery of droplet fluorescence, indicating a dynamic exchange of ∆LIM molecules with the environment. Finally, in the presence of polyethylene glycol (PEG), a crowding agent known to promote the separation of protein phases [37,38], droplet formation was observed even at 8 nM [27]. The ability to achieve the phase separation of various modifications of His-GFP-DHU-VHP, including His-GFP-DHU and His-GFPHis-GFP-VHP, was considered. The results of these studies concluded that the IDR region homologous to dematin mediated LLPS.

Finally, since aromatic amino acids (Fs and Ys) are known to be critical for the LLPS of some IDRs [14,27,39], DHU mutants that had a serine (S) residue instead of aromatic residues were investigated, and it was shown that that the resulting mutants, DHU25S and ∆LIM25S, did not form coacervates. Thus, it was proven that ∆LIM can, indeed, be separated into liquid droplets due to its IDR. In order to prove that ∆LIM liquid droplets can function as actin polymerization sites for the direct production of de novo radial actin filaments, GFP-∆LIM liquid droplets were examined through live imaging immediately after the addition of G-actin. Numerous hair-like protrusions were shown to continuously grow from droplets (Figure 2D,E), indicating the massive nucleation and elongation of F-actin. To prove that the rays were formed by F-actin, the slides were stained with 4 µM of phalloidin-TRITC. Although the presence of phalloidin can potentially stabilize actin filaments, it does not affect the rate of actin polymerization [40] and, thus, does not interfere with the process of aster formation. Indeed, phalloidin-TRITC rapidly congressed to the outer surface of the droplets within 30 s after the G-actin addition, after which, GFPpositive F-actin arrays gradually appeared around all of the droplets (Figure 2A,D). It is noteworthy that some of the drops gradually became hollow during the growth of the aster, which suggests that the elongation of the astral bundle absorbs the contents of the drops. Accordingly, larger droplets produced longer bundles of astral actin (Figures 2D and 3A).

It was also shown that aster formation was not observed in the presence of Cytochalasin D (CytoD) or LatrunculinA (LatA), drugs that inhibit actin polymerization and prevent actin assembly in the plus-end [24] or sequestering G-actin [41].

To find out whether the formation of aster was simply the result of a directed arrangement of abLIM1 molecules on a spherical surface and not their phase separation, a control experiment was carried out with magnetic beads. To perform this experiment, HEK293T cells were overexpressed with GFP-∆LIM and concentrated with anti-GFP, antibody-conjugated magnetic beads ~1.5 µm in diameter. Thus, the ∆LIM molecules on the beads were oriented so that their actin-binding VHP domain was directed radially from the center of the bead. Direct imaging showed that the fluorescence intensity of the GFP beads was comparable to that of similarly sized liquid GFP-∆LIM droplets (Figure 3B), suggesting a similar density of GFP-∆LIM molecules on the surface of the beads and the droplets. However, in the presence of G-actin, the beads could not produce visible actin filaments, while liquid droplets induced aster formation. Therefore, the formation of ∆LIM coacervates is necessary for aster formation [27].

Thus, the authors of the work [27] proved that ∆LIM, consisting mainly of a disordered fragment of DHU, can form coacervates, which are actin polymerization reactors, due to LLPS. At the same time, it was shown that both DHU and VHP are necessary for actin aster formation since no aster formation was observed in experiments with GFP-DHU liquid droplets.

**Figure 3.** Nucleation, elongation of polymerization, formation of asters, and F-actin filaments in the presence of ΔLIM or abLIM-M liquid droplets. (**A**) Intense actin polymerization from GFP-ΔLIM liquid droplets in real time. The initial moment of time was taken as the time when G-actin was introduced into the actin polymerization buffer containing preliminarily formed GFP-ΔLIM liquid drops. (**B**) F-actin does not polymerize from GFP-ΔLIM immobilized with an anti-GFP antibody on magnetic beads (upper panel); however, in the presence of preformed His-GFP-ΔLIM droplets, actin filaments grow rapidly (lower panel). (**C**) abLIM-M liquid droplets can also generate asters. Although the LIM domain inhibits droplet formation, GFP-abLIM-M forms droplets in the presence of 1% PEG. Drops of GFP-abLIM-M, as well as GFP-ΔLIM, induce aster formation after the addition of G-actin. (**D**) G-actin polymerization requires the presence of GFP-ΔLIM or GFP-abLIM-M droplets. Individually, neither GFP-DHU (IDR domain) nor GFP-VHP (actin-binding domain) affects the rate of actin polymerization. (**E**) A model illustrating the role of abLIM1 in actin polymerization. At least, its abLIM-S or abLIM-M isoform undergoes LLPS to form condensates that promote actin concentration, initiate polymerization, and cross-link actin filaments to form asters and sporadic bundles that further develop into the F-actin web. The figure represents panels B, E, and G of Figure 4 and panels C and F of Figure 5 of Yang et al., PNAS 119 (29), e2122420119, 2022, [27]. Copyright © 2022 Yang et al., published by PNAS. This article is distributed under Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND). **Figure 3.** Nucleation, elongation of polymerization, formation of asters, and F-actin filaments in the presence of ∆LIM or abLIM-M liquid droplets. (**A**) Intense actin polymerization from GFP-∆LIM liquid droplets in real time. The initial moment of time was taken as the time when G-actin was introduced into the actin polymerization buffer containing preliminarily formed GFP-∆LIM liquid drops. (**B**) F-actin does not polymerize from GFP-∆LIM immobilized with an anti-GFP antibody on magnetic beads (upper panel); however, in the presence of preformed His-GFP-∆LIM droplets, actin filaments grow rapidly (lower panel). (**C**) abLIM-M liquid droplets can also generate asters. Although the LIM domain inhibits droplet formation, GFP-abLIM-M forms droplets in the presence of 1% PEG. Drops of GFP-abLIM-M, as well as GFP-∆LIM, induce aster formation after the addition of G-actin. (**D**) G-actin polymerization requires the presence of GFP-∆LIM or GFP-abLIM-M droplets. Individually, neither GFP-DHU (IDR domain) nor GFP-VHP (actin-binding domain) affects the rate of actin polymerization. (**E**) A model illustrating the role of abLIM1 in actin polymerization. At least, its abLIM-S or abLIM-M isoform undergoes LLPS to form condensates that promote actin concentration, initiate polymerization, and cross-link actin filaments to form asters and sporadic bundles that further develop into the F-actin web. The figure represents panels B, E, and G of Figure 4 and panels C and F of Figure 5 of Yang et al., PNAS 119 (29), e2122420119, 2022, [27]. Copyright © 2022 Yang et al., published by PNAS. This article is distributed under Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND).

rangement of abLIM1 molecules on a spherical surface and not their phase separation, a control experiment was carried out with magnetic beads. To perform this experiment, Since actin-binding proteins are the entire abLIM family, it was necessary to find out how the ordered LIM fragment affects LLPS and whether these proteins could form

HEK293T cells were overexpressed with GFP-ΔLIM and concentrated with anti-GFP,

To find out whether the formation of aster was simply the result of a directed ar-

coacervates. The tendency to precipitate His-GFP-abLIM-L [28] made it difficult to perform experiments with this form. However, His-GFP-abLIM-M, a widely expressed major isoform (Figure 1A) [29], was soluble, so the study of whether its liquid droplets could also generate aster was carried out with this isoform. It was found that GFP-abLIM-M did not phase separate, even at 105 µM, consistent with the inhibitory role of the LIM region. However, it was subjected to LLPS at 6 µM in the presence of 1% PEG (Figure 3C).

Live imaging showed that, after the addition of G-actin, liquid droplets mediated aster formation in a similar manner (Figure 3C). Therefore, the separated abLIM-M and abLIM-S phases were able to facilitate local actin polymerization and binding to form complex aster.

In [27], the rate of actin polymerization in the presence of abLIM1 was also estimated. For this, the authors used the fluorescence of pyrene. An essential increase in the actin polymerization rate was shown in the presence of abLIM1. It is known that the stage limiting rate of actin polymerization is the formation of nuclei, i.e., the formation of initial actin trimers [42,43]. To quantify how abLIM1 promotes actin nucleation, the G-actin concentration was chosen to be 2 µM, which excludes spontaneous actin nucleation [44,45]. This made it possible to evaluate the activity of nucleation mediated by nucleators [46,47]. It turned out that while pyrene fluorescence in GFP samples only slightly increased over time; it steadily increased in 3 µM GFP-∆LIM samples and stabilized after about 25 min (Figure 3D). The steady-state pyrene intensity was more than five times higher than that of the GFP samples (Figure 3D). However, separately, neither GFP-DHU nor GFP-VHP had a clear effect on actin polymerization kinetics compared to GFP (Figure 3D), confirming the need for both regions. As expected, the GFP∆LIM25S that did not undergo LLPS was unable to stimulate actin polymerization (Figure 3D).

Thus, based on all in vitro studies, it can be concluded that LLPS-induced ∆LIM condensates are capable of strongly stimulating actin nucleation. The resulting actin filaments are then stitched together to form stable asters and sporadic beams that further develop into stable networks (Figure 3E).

Condensates of abLIM1 likely move along actin filaments to regionally nucleate and cross-link actin filaments together to self-organize elastic, interconnected networks of F-actin cortical bundles so that cells properly resist mechanical stress. The tendency of Factin polymerizing arrays to consume liquid droplets and the positive correlation between the length of actin astral bundles and the size of liquid droplets suggest fluid fluidity of ∆LIM and abLIM-M along F-actin in vitro. In [27], it was concluded that the absence of the F-actin association of the phase-defective ∆LIM25S and abLIM25S-M mutants, the appearance of liquid droplets of abLIM-M and ∆LIM during actin depolymerization in cells, their resorption into repolymerizing actin cytoskeletons, and their constant propagation along intracellular F-actin bundles confirmed that the intracellular molecules, abLIM-M and abLIM-S, associated with F-actin were also form coacervates. Moreover, previous studies demonstrate that dense, interconnected cortical actin networks in RPE1 and U2OS cells become sparse and rich in thick linear filaments when abLIM1 is depleted, leading to membrane swelling during cell proliferation or migration [28]. In addition, because abLIM1 is highly enriched in the Z-disc of sarcomeres [29,48], it may also assist in the construction of actin networks in striated muscle cells through the nucleation and crosslinking of actin filaments in the Z-disc. This is also consistent with the ∆LIM-induced efficient polymerization of Ca2+-G-actin in the absence of Mg2+ .

#### **4. LLPS of Signaling Proteins Leads to the Formation of Coacervates Concentrating Actin-Binding Proteins, N-WASP, Arp2/3, and Cortactin**

A recent review on the role of phase transitions in the formation of the cytoskeleton [3] analyzed the role of phase transitions in actin polymerization and briefly mentioned several studies showing that the phase separation of signaling proteins modulates the functioning of proteins associated with actin polymerization.

It was shown in [1] that the phase transition of the neuronal Wiskott–Aldrich syndrome protein (N-WASP)—which interacts with its biological partners, Nck and phosphorylated nephrin—leads to the formation of coacervates, which include complexes (actin-related proteins), Arp2/3, which are actin nucleation factors. It has been shown that actin polymerization increases dramatically when coacervates of signaling proteins are formed and decreases when the LLPS process or the interaction of Arp2/3 with N-WASP is disrupted. In [49], a mechanism was proposed for the intensification of the Arp2/3 complex in coacervates. It has been shown that the phase separation of Nephrin/Nck/N-WASP signaling proteins on lipid bilayers leads to an increase in the residence time of N-WASP and the Arp2/3 complex on the membrane and, therefore, intensifies actin assembly.

Another example of condensates that enhance actin polymerization is activated T-cell receptors [2]. The authors biochemically recreated a 12-component signaling pathway on model membranes, starting with T-cell receptor (TCR) activation and ending with actin assembly. When TCR phosphorylation was triggered, signaling proteins spontaneously separated into fluid-like clusters that facilitated signaling both in vitro and in human Jurkat T cells. The reconstituted clusters were enriched in kinases, but excluded phosphatases and enhanced the assembly of actin filaments by recruiting and organizing actin polymerization regulators, N-WASP-Arp2/3. Thus, it was demonstrated that the LLPS of proteins can create separate physical and biochemical compartments that facilitate signaling.

Another example of this phenomenon was considered in [50], where signaling in synapses was considered. Obviously, the correct formation and rapidity of responses to synaptic stimulation are fundamental for the functions of the mammalian brain; however, the molecular basis that governs the formation and modulation of separated synaptic ensembles has, so far, remained unclear. Using a biochemical reconstruction approach, the authors showed that post-synaptic scaffold proteins at physiological concentrations can form highly condensed, self-assembling postsynaptic density protein (PSD)-like assemblies via liquid–liquid phase separation (LLPS). Such PSD scaffold condensates can cluster glutamate receptors, incorporate synaptic enzymes, and promote the formation of actin bundles; however, they do not allow gephyrin, which is the main scaffold protein for almost all inhibitory synapses, to pass through.

When creating model coacervates, the authors used the main PSD scaffold proteins, including PSD-95, GKAP, Shank, and Homer, which serve to connect ion channels/receptors on the postsynaptic plasma membrane to the actin cytoskeleton in the PSD cytoplasm. It is known that Shank directly interacts with regulatory proteins of the actin cytoskeleton, such as cortactin [51] and subunits of the Arp2/3 complex [52].

Cortactin has an N-terminal acidic domain that binds to and stimulates the Arp2/3 complex, followed by tandem cortactin repeats (CR) that bind to F-actin, and a C-terminal SH3 domain that binds to Shank3. It was shown that when cortactin, the Arp2/3 complex, and G-actin were introduced into the 63 PSD system assembled on lipid bilayers, polymerized structures of F-actin began to appear in the condensed PSD assemblies 15 min after the onset of the phase transition. Actin bundles colocalized with PSD condensates gradually became thicker and longer. Phalloidin staining confirmed that the bundles were actin filaments. Without the addition of the five PSD components but with the presence of cortactin, the Arp2/3 complex, and G-actin, actin bundles were not observed. Finally, it has been shown that PSD can also promote the formation of actin bundles on lipid bilayers in the absence of Arp2/3. Thus, it has been shown that PSD condensates can promote F-actin assembly by increasing the concentration of G-actin and cortactin.

#### **5. Conclusions**

To date, it is becoming more and more obvious that the LLPS of proteins is of a universal nature and, apparently, mediates all known cellular processes [9]. The process of the formation of F-actin, which is one of the main components of the cytoskeleton, is no exception.

Figure 4 shows the coacervates that are known, to date, to arise from LLPS, which plays a crucial role in actin polymerization. This model shows:

**5. Conclusions** 

exception.

3);

− coacervates of LAT signaling proteins capable of integrating actin-binding proteins, which, as noted in the literature, leads to the intensification of their work (see Section 3); monomers and thereby increases their local concentration, thus leading to the initiation of polymerization and the growth of F-actin fibrils in the form of asters (see

polymerized structures of F-actin began to appear in the condensed PSD assemblies 15 min after the onset of the phase transition. Actin bundles colocalized with PSD condensates gradually became thicker and longer. Phalloidin staining confirmed that the bundles were actin filaments. Without the addition of the five PSD components but with the presence of cortactin, the Arp2/3 complex, and G-actin, actin bundles were not observed. Finally, it has been shown that PSD can also promote the formation of actin bundles on lipid bilayers in the absence of Arp2/3. Thus, it has been shown that PSD condensates can pro-

To date, it is becoming more and more obvious that the LLPS of proteins is of a universal nature and, apparently, mediates all known cellular processes [9]. The process of the formation of F-actin, which is one of the main components of the cytoskeleton, is no

Figure 4 shows the coacervates that are known, to date, to arise from LLPS, which

− coacervates of LAT signaling proteins capable of integrating actin-binding proteins, which, as noted in the literature, leads to the intensification of their work (see Section

− coacervates of the actin-binding protein, abLIM1, which integrates G-actin

mote F-actin assembly by increasing the concentration of G-actin and cortactin.

	- − VASP coacervates, which trigger the spontaneous, self-sustaining growth of actin bundles and are analyzed in detail in the yet-unpublished work of Graham et al. [53]. Actin polymerizes inside the VASP droplets, and elongating filaments are distributed along the periphery of the droplet, forming an actin-rich ring inside the droplet. As actin polymerizes and the ring thickens, its rigidity increases and eventually overcomes the surface tension of the drop, deforming the drop and transforming it into a linear bundle. The resulting bundles contain long, parallel actin filaments growing from their tips (Figure 4). The length of the final F-actin bundle can be many times greater than the size of the initial droplet. Actin polymerizes inside the VASP droplets, and elongating filaments are distributed along the periphery of the droplet, forming an actin-rich ring inside the droplet. As actin polymerizes and the ring thickens, its rigidity increases and eventually overcomes the surface tension of the drop, deforming the drop and transforming it into a linear bundle. The resulting bundles contain long, parallel actin filaments growing from their tips (Figure 4). The length of the final F-actin bundle can be many times greater than the size of the initial droplet.

*Int. J. Mol. Sci.* **2023**, *24*, x FOR PEER REVIEW 10 of 13

plays a crucial role in actin polymerization. This model shows:

**Figure 4.** The currently discovered, coacervate-dependent models of actin polymerization and bundling. The figure was created based on the results of the works [2,27,53]. **Figure 4.** The currently discovered, coacervate-dependent models of actin polymerization and bundling. The figure was created based on the results of the works [2,27,53].

Actin filaments are known to be associated with the membrane. Within the framework of new ideas of actin polymerization, it becomes clear that this is energetically favorable since the threshold concentration of coacervate formation on the membrane is an order of magnitude lower than in free space [54]. New ideas about actin polymerization Actin filaments are known to be associated with the membrane. Within the framework of new ideas of actin polymerization, it becomes clear that this is energetically favorable since the threshold concentration of coacervate formation on the membrane is an order of magnitude lower than in free space [54]. New ideas about actin polymerization reveal new directions in the study of the cytoskeleton and formulate new challenges for researchers. It remains to be seen whether some other actin-binding proteins can form coacervates during phase separation and under which scenario they will work, similar to how the abLIM1 or VASP coacervates interact with formed filaments [3]. It is interesting to evaluate the existence of proteins that can form coacervates similar to signaling proteins in response to internal cell stimuli.

Together with earlier work on the mechanism of microtubule formation through the concentration of tubulin in centrosomes [4], recent works on actin polymerization highlight the universal role of LLPS in the formation of the cytoskeleton. Researchers are now challenged to explore the dynamic cytoskeletal organization within the frame of this new paradigm.

**Author Contributions:** Writing—original draft preparation O.I.P., I.A.A., A.V.F., K.K.T. and I.M.K.; writing—review and editing, O.I.P., I.A.A., A.V.F., K.K.T. and I.M.K. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not application.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


**Disclaimer/Publisher's Note:** The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
