2. Condensates in the transcriptional and post-transcriptional responses of plants to water deficit

A few landmark studies showing the implication of condensates in regulating mRNA abundance under WD were published over the recent years (Figure 1).

Figure 1. Key examples of biomolecular condensates regulating mRNA fate in plants under osmotic stress conditions. Note: In addition to their role in mRNA sequestration, biomolecular condensates are involved at various steps of the mRNA life cycle in plants confronted with osmotic stresses. For example, SEU, a transcription regulator, and STM form nuclear condensates under osmotic stress to facilitate transcriptional responses (Cao et al., Reference Cao, Du, Guo, Wang and Jiao2023; Wang et al., Reference Wang, Zhang, Huai, Peng, Wu, Lin and Fang2022). OsFKBP20-1b colocalizes with UBP1b in stress granules and with SR45, a protein involved in splicing (Park et al., Reference Park, You, Lee, Jung, Jo, Oh, Kim, Lee, Kim and Kim2020). DCP5, meanwhile, exhibits rapid assembly dynamics that are dependent on osmolarity to regulate the transcriptome (Wang et al., Reference Wang, Yang, Zhang, Lu, Wang, Pan, Qiu, Men, Yan and Xiao2024b). CBC, cap binding complex; Pol II, RNA polymerase II; PABPN, nuclear poly(A) binding protein; PABPC, cytoplasmic poly(A) binding protein; eIF4E, eukaryotic translation initiation factor 4E.

SEUSS (SEU) is a transcriptional adaptor involved in regulating transcription and may play a role in plant growth and development (Bao et al., Reference Bao, Azhakanandam and Franks2010), as well as in the response to hyperosmotic and salt stress (Wang et al., Reference Wang, Zhang, Huai, Peng, Wu, Lin and Fang2022). Upon osmotic treatments such as sorbitol, KCl or NaCl, SEU rapidly forms nuclear condensates through LLPS (Wang et al., Reference Wang, Zhang, Huai, Peng, Wu, Lin and Fang2022). This condensation is reversible, as SEU-GFP condensates quickly dissolve after stress removal. This mechanism allows SEU to directly sense osmotic changes through molecular crowding and induce its own condensation, which appears to be essential for cellular adaptability to stress. The formation of SEU condensates depends on the presence of intrinsically disordered regions (IDRs), particularly the N-terminal IDR1 domain. Removal of this domain prevents SEU condensation, even though the protein remains functional under normal growth conditions. Indeed, the expression of a truncated SEU variant (SEUΔIDR1) rescues the developmental defects of seu-6 mutant, suggesting that SEU’s transcriptional function is not impaired in the absence of stress. However, this mutated variant fails to restore stress tolerance, highlighting the critical role of SEU condensation in stress adaptation. SEU mutant lines exhibit marked hypersensitivity to NaCl and mannitol, as well as an exacerbated response to abscisic acid (ABA), suggesting that SEU functions either upstream or independently of the ABA signaling pathway during osmotic stress. SEU homologs, present in all land plants and characterized by high conservation of an LDB domain and α-helices within the N-terminal IDR, form condensates in response to hyperosmotic stress. Analysis of differentially expressed genes (DEGs) in seu-6 mutants reveals that SEU regulates a distinct set of genes under normal and stress conditions; more than half of the DEGs identified under osmotic treatment are stress-specific, indicating that SEU plays a different role depending on environmental conditions. Among these genes, several are directly involved in the osmotic response, and their expression is significantly reduced in seu-6 under sorbitol or NaCl treatment, at both early and late stress stages. Thus, SEU acts as both a sensor and a transcriptional regulator of osmotic stress by forming dynamic condensates in response to environmental changes. This condensation ability, which is evolutionarily conserved, is essential for activating transcriptional responses that enable plants to adapt to hyperosmotic stress conditions.

STM (SHOOTMERISTEMLESS) plays a key role in the initiation and maintenance of the shoot apical meristem (SAM) in Arabidopsis (Cao et al., Reference Cao, Du, Guo, Wang and Jiao2023). While stresses such as hyperosmotic stress, hydrogen peroxide (H₂O₂), temperature variations and abscisic acid have minimal effects, high concentrations of NaCl (≥ 0.2 M) significantly stimulate the formation of cytoplasmic STM-Venus condensates in vivo. The disordered central PrD domain of STM, located in the N-terminal region, promotes this condensation through electrostatic and hydrophobic interactions, which are essential for the organization and regulation of components necessary for meristem activity. In the absence of this domain, as in a truncated STM variant, the ability to form condensates is lost, resulting in a uniform distribution of the protein. Plants with truncated STM exhibit increased sensitivity to salt stress and reduced SAM size. Furthermore, STM interacts with BELL-class TALE homeodomain transcription factors (BEL1-like), a family of plant-specific regulators including ARABIDOPSIS THALIANA HOMEOBOX 1 (ATH1), REPLUMLESS (RPL) and POUNDFOOLISH (PNF) which, through their prion-like (PrD) domains, contribute to the stabilization and regulation of STM-formed condensates. They thus facilitate meristem regulation, particularly of inflorescence meristems, with RPL playing a central role in this process. STM condensates also colocalize with MEDIATOR SUBUNIT 8 (MED8), thereby enhancing its transcriptional activity. In fact, NaCl treatment, which promotes condensation, increases the expression of STM-regulated genes, while treatment with 1,6-hexanediol, which dissolves the condensates, reduces this activation. This suggests that condensate formation enhances its transcriptional regulatory activity and thus promotes meristem maintenance. Thus, the formation of STM biomolecular condensates plays a crucial role in transcriptional regulation and environmental acclimation, acting as an adjustment mechanism to determine cell fate and response to stress conditions.

OsFKBP20-1b (Park et al., Reference Park, You, Lee, Jung, Jo, Oh, Kim, Lee, Kim and Kim2020), a protein belonging to the immunophilin family, is expressed at low levels throughout all developmental stages and in all tissues, with some variations during the development of rice plants. In leaves of 4-week-old plants, OsFKBP20-1b localizes to nuclear and cytoplasmic foci, whereas in leaves of 5-week-old plants it redistributes predominantly to the nucleoplasm and cytoplasm, indicating dynamic compartmental mobility during development. OsFKBP20-1b colocalizes with the SG marker OLIGOURIDYLATE BINDING PROTEIN 1b (UBP1b) in the cytoplasm and, additionally, with SR45 (a serine/arginine-rich nuclear ribonucleoprotein and splicing factor) in both nuclear and cytoplasmic speckles. Under salt stress (200 mM NaCl), heat stress (42°C) and other abiotic stresses, the expression of OsFKBP20-1b increases and its subcellular localization becomes more dynamic, with an increase in the number of foci. Consistent with its high reactivity to stress conditions, OsFKBP20-1b plays a key role in regulating stress-responsive genes. Indeed, the loss-of-function mutant shows a significant reduction in the expression of stress-responsive genes such as LATE EMBRYOGENESIS ABUNDANT 3 (OsLEA3) and RESPONSIVE TO ABA 16 (OsRAB16), whereas their transcription levels are tripled in OsFKBP20-1b overexpression lines compared to wild-type plants in response to ABA. Through its interaction with SR45, OsFKBP20-1b also modulates the alternative splicing of several pre-mRNAs, including those of OsbZIP60, OsNAC5, OsDREB2B, OsTFIIIA and OsRS31, in response to ABA treatment as well as heat and cold stress, as marked differences in their alternative splicing patterns are observed between the osfkbp20-1b mutant and wild-type under these conditions. Moreover, OsFKBP20-1b functions as a molecular chaperone by preventing the degradation of the OsSR45 protein, which is dependent on the 26S proteasome, thereby contributing to the stability and regulation of stress-related proteins. Transgenic seedlings overexpressing OsFKBP20-1b display enhanced tolerance to high salinity and drought compared to wild-type plants, while loss-of-function mutants exhibit growth retardation and increased sensitivity to stress. In summary, RNA processing mediated by OsFKBP20-1b is essential for rice adaptation to stressful environmental conditions, but the role of condensation for its function remains to be determined.

Also in rice, among genes located in a genomic region associated with drought tolerance, DROUGHT RESISTANCE GENE 9 (DRG9) stood out as a key candidate due to its strong induction under water stress and drought sensitivity of drg9 mutants and drought resistance of DRG9 over-expressors (Wang et al., Reference Wang, Ye, Guo, Yao, Tu, Wang, Zhang, Wang, Li and Li2024a). The protein is able to bind dsRNA and the 3’-UTR of OsNCED4 mRNA in particular, which is involved in ABA synthesis. DRG9-YFP forms condensates under mannitol treatment in the root tip through an IDR located in its N-terminus. Mutation within the IDR that disrupts phase separation also abolishes DRG9’s ability to stabilize its RNA targets, indicating that LLPS is essential for its biological function. Of note, natural variation of DRG9 in the rice population studied influences mRNA binding capacity rather than condensation properties. Altogether, results in this study position DRG9 as both an mRNA-binding protein and a phase-separation mediator with the capacity to recruit stress-related mRNAs into SG, thereby preserving them from degradation. This mechanism highlights a novel layer of post-transcriptional regulation crucial for plant survival under water-deficit conditions.

Along the same lines, the role of EVOLUTIONARILY CONSERVED C-TERMINAL REGION 8 (ECT8), a protein recognizing the N6-methyladenosine modification (m⁶A) in mRNAs, was identified for its role in WD signaling in Arabidopsis in 2024 (Wu et al., Reference Wu, Su, Zhang, Zhang, Wong, Ma, Shao, Hua, Shen and Yu2024). In this study, cytoplasmic condensation of ECT8-GFP in SG was observed in root tip cells under various WD-related treatments, including osmotic stresses. This protein contains 3 IDRs, of which IDR #2 is responsible for the condensation properties of ECT8 after polyethylene glycol (PEG) treatment in vitro and mannitol in vivo. ect8 mutants show hypersensitivity to ABA and improved fitness upon rewatering after drought. Those phenotypes are not complemented in an ect8 line expressing the IDR2-deleted ECT8 protein, showing that condensation of ECT8 under stresses is required for its function. Mutations in the m⁶A recognition domain also impact the capacity of ECT8 to form condensates. ECT8 appears to recruit m⁶A mRNAs into SG, and the mRNA of the ABA receptor PYL7 in particular, reducing its protein level under ABA treatment and providing a negative feedback loop on ABA signaling.

Also in Arabidopsis, the decapping activator/translation repressor DECAPPING 5 (DCP5) was reported to promote condensate formation in response to osmotic stress (Wang, Yang, et al., Reference Wang, Yang, Zhang, Lu, Wang, Pan, Qiu, Men, Yan and Xiao2024b). DCP5 was previously characterized as a component of P-bodies, which can associate with SGs in response to heat stress or even promote nuclear condensates controlling flowering in Arabidopsis (Blagojevic et al., Reference Blagojevic, Baldrich, Schiaffini, Lechner, Baumberger, Hammann, Elmayan, Garcia, Vaucheret and Meyers2024; Wang et al., Reference Wang, Wang, Wang, Ma, Wang, Tao, Liu, Li, Hu and Gu2023; Xu & Chua, Reference Xu and Chua2009). In 2024, Wang et al. characterized DCP5 as a molecular crowding sensor that directly responds to water deficiency (Wang, Yang, et al., Reference Wang, Yang, Zhang, Lu, Wang, Pan, Qiu, Men, Yan and Xiao2024b). Within the IDR of DCP5, a region enriched in highly hydrophobic residues primarily contributes to its condensation ability. While DCP5 homologs in yeast (Scd6) and animals (LSM14A) lack most of this region, it is well conserved in land plants. In Arabidopsis, the dcp5-1 mutant exhibits increased susceptibility to hyperosmolarity and desiccation, and resistance can only be partially restored by DCP5 lacking this region. In response to molecular crowding induced by cell volume changes under hyperosmotic stresses, DCP5 undergoes LLPS and forms a distinct type of SG, so-called DCP5-enriched osmotic stress granules (DOSGs). DOSGs recruit typical SG components such as mRNAs, translation initiation factors, nucleocytoplasmic transporters and proteins related to the proteasome. Hyperosmotic stress triggers rapid DCP5 condensation within 15 minutes, resulting in altered translation efficiency of thousands of genes. Gene Ontology enrichment analysis revealed a downregulation of growth-promoting genes, along with an upregulation of genes involved in ion transport and stress response. At the transcriptional level, a DCP5-dependent reshaping of the transcriptome can be observed after one hour of hyperosmotic stress, with several hundred genes down-or upregulated. Consistent with the translational regulation, those genes are mainly involved in stress response and growth or developmental processes, respectively. This reprogramming of the transcriptome may be explained by the sequestration of nucleocytoplasmic transporters and transcription factors such as BASIC TRANSCRIPTION FACTOR 3 (BTF3), which were found in the proteome of DOSGs. Additionally, nuclear-localized DCP5 has been reported to directly participate in transcriptional regulation: DCP5 co-condensates with SISTER OF FCA (SFF), inhibiting its ability to promote the transcription of target genes. Both DCP5 and SFF contain a prion-like domain (PrLD), which is essential for condensation. Whether DCP5 interacts with other transcriptional regulators in the nucleus and whether this occurs under osmotic stress has not yet been specifically described, which remains a potential direction for further investigation to better understand DCP5’s role in stress adaptations.

Taken together, those studies illustrate the potential of biomolecular condensates in sensing WD and in regulating various transcriptional and post-transcriptional steps of gene expression in response to WD. Detailed molecular insights are still required, however. In particular, an added or altered functionality of the condensates compared to the solute phase needs to be assessed to support a model in which condensation can transduce a WD signal (Glauninger et al., Reference Glauninger, Wong Hickernell, Bard and Drummond2022; Putnam et al., Reference Putnam, Thomas and Seydoux2023) -which is already addressed in most of the recent work we reported above. Moreover, a plethora of distinct molecular mechanisms may have been selected to mediate those changes in gene expression and have to be taken into account. These can range from transcriptional regulation to the modulation of mRNA stability or the sequestration of factors or mRNAs away from the translation machinery. In the next section, we propose to add another layer of complexity for future studies, as we discuss why it also seems important to consider other, more biophysical, aspects of water deficit.

3. Can physicochemical cues of plant cells under water deficit trigger the formation of condensates?

Biomolecular condensate formation has almost always been associated with an increase in molecular crowding, Π, or volume change. However, proteins can also condensate in response to P (Cinar et al., Reference Cinar, Fetahaj, Cinar, Vernon, Chan and Winter2019), and a drop in P is characteristic of the initial phase of WD in plant cells, during which we unraveled quantitative transcriptomic regulations (Crabos et al., Reference Crabos, Huang, Boursat, Maurel, Ruffel, Krouk and Boursiac2023). In this section, we therefore explore the ranges of inner P or Π at which condensate dynamics have been observed in plant cells, and we evaluate whether a condensate-based sensing could occur for P or Π variations upon WD.

As a technical note, we should stress that both parameters (as well as Ψ) are challenging to measure in vivo at the cellular level. Direct or indirect measurements of P with a cell pressure probe, pico gauges (Hüsken et al., Reference Hüsken, Steudle and Zimmermann1978; Knoblauch et al., Reference Knoblauch, Mullendore, Jensen and Knoblauch2014) or indenters and related techniques (Beauzamy et al., Reference Beauzamy, Fourquin, Dubrulle, Boursiac, Boudaoud and Ingram2016; Zimmermann et al., Reference Zimmermann, Reuss, Westhoff, Gessner, Bauer, Bamberg, Bentrup and Zimmermann2008) are not amenable to all cell types and are of low throughput. Techniques using exogenously applied chemical probes are promising but still require solid foundations for the interpretation of their signals when applied to plant cells (Michels et al., Reference Michels, Gorelova, Harnvanichvech, Borst, Albada, Weijers and Sprakel2020; Ryder et al., Reference Ryder, Lopez, Michels, Eseola, Sprakel, Ma and Talbot2022) or remain to be effective inside cells (Jain et al., Reference Jain, Liu, Zhu, Chang, Melkonian, Rockwell, Pauli, Sun, Zipfel and Holbrook2021). FRET-based genetically encoded sensors are also very promising, given the increasing knowledge of intra or inter-protein interactions in relation to physicochemical cues and one, Sensor Expressing Disordered Protein 1 (SED1), has been reported for Π but has yet to be used in planta (Cuevas-Velazquez et al., Reference Cuevas-Velazquez, Vellosillo, Guadalupe, Schmidt, Yu, Moses, Brophy, Cosio-Acosta, Das and Wang2021).

A bibliographic survey for “condensates” and “hydrostatic pressure” retrieved no relevant results in plants, which suggests that the sensitivity of condensates to variations in turgor pressure in plant cells has not been reported yet. Condensates have been shown to be sensitive to hydrostatic pressures in vitro, but for variations in the range of a few to hundreds of MPa (Cinar et al., Reference Cinar, Fetahaj, Cinar, Vernon, Chan and Winter2019). This exceeds by far what could be physiologically relevant for land plants. Indeed, the root cortical cells of Arabidopsis, barley or maize typically exhibit P in the range of 0.18 to 0.70 MPa (Crabos et al., Reference Crabos, Huang, Boursat, Maurel, Ruffel, Krouk and Boursiac2023; Ding et al., Reference Ding, Milhiet, Couvreur, Nelissen, Meziane, Parent, Aesaert, Van Lijsebettens, Inzé, Tardieu, Draye and Chaumont2020; Javot et al., Reference Javot, Lauvergeat, Santoni, Martin-Laurent, Guclu, Vinh, Heyes, Franck, Schaffner and Bouchez2003; Rygol et al., Reference Rygol, Pritchard, Zhu, Tomos and Zimmermann1993). The drop in P due to an osmotic treatment depends on the nature of the solute and the selectivity of the membrane, but is roughly proportional to the concentration of the solute. For example, we have shown that there is a quasi-linear relationship of the drop in P of cortical cells in the young part of the root, in response to 25–150 mM NaCl or sorbitol, or 75–150 g/l PEG8000 (Crabos et al., Reference Crabos, Huang, Boursat, Maurel, Ruffel, Krouk and Boursiac2023). Figure 2 (black curve) shows a characteristic turgor pressure change in response to an outer Ψ drop. From the literature cited above and other works related to condensates in response to WD (Cao et al., Reference Cao, Du, Guo, Wang and Jiao2023; Chong et al., Reference Chong, Foo, Lin, Wong and Verslues2019; Khan et al., Reference Khan, Garbelli, Grossi, Florentin, Batelli, Acuna, Zolla, Kaye, Paul and Zhu2014; Soma et al., Reference Soma, Takahashi, Suzuki, Shinozaki and Yamaguchi-Shinozaki2020; Wang et al., Reference Wang, Zhang, Huai, Peng, Wu, Lin and Fang2022; Wang et al., Reference Wang, Ye, Guo, Yao, Tu, Wang, Zhang, Wang, Li and Li2024a, Reference Wang, Yang, Zhang, Lu, Wang, Pan, Qiu, Men, Yan and Xiao2024b), we extracted quantitative indications on the dynamics of molecular condensates and plotted them as a function of Ψ in the same figure. It turns out that almost all the data we retrieved fit into a stress intensity zone where root cortical cells would be under plasmolysis. One of the only conditions compatible with a positive P was 150mM mannitol treatment on DCP5-GFP expressing plants, which showed a transient formation of condensates (Wang et al., Reference Wang, Yang, Zhang, Lu, Wang, Pan, Qiu, Men, Yan and Xiao2024b). Therefore, a variation in P in published experiments is unlikely to have driven LLPS and formation of microscopically visible condensates. The absence of a report on condensate formation during milder treatments does not rule out the possibility that it may occur, but this should be tested through dedicated experiments.

Figure 2. Schematic turgor pressure (black line, left axis) and osmotic pressure (green line, right axis) versus water potential of cells during a hyper-osmotic shock. Greyed zone represents the turgid state of the cell, beyond this zone, the cell is plasmolyzed, the turgor pressure is zero and the water potential is equal to –Π. Horizontal segments and points represent osmotic shock ranges or values found for plant cells in the literature (cf main text). P and Π were calculated using ϵ = 2.23MPa and an initial turgor pressure of 0.41MPa. The initial osmotic pressure was set equal to the initial P. Π was calculated using the mass balance into the cell, where the relative variation of solute concentration, and so the relative osmotic pressure variation, is the inverse of the relative volume variation. P is calculated from the elastic modulus definition, and $\varPsi =P\hbox{--} \varPi$ . Note that some observations were made in the root apical meristem, which could lose turgor at higher osmotic treatments. For more information, see Supplementary Material S1.

The other physicochemical parameter related to water in plant cells is Π (or its opposite, the osmotic potential Ψ _Π). It is roughly related to the number of solutes within the given volume of a cell. Upon WD, Π will be affected in a biphasic way. Because the cell wall is elastic, a change in turgor pressure also impacts cell volume and therefore will impact Π. Considering the definition of the elastic modulus ϵ, $dP=\epsilon \frac{dV}{V}$ , V being the cell volume and the data obtained with the cell pressure probe on Arabidopsis root cortical cells, ϵ = 2.23MPa (Javot et al., Reference Javot, Lauvergeat, Santoni, Martin-Laurent, Guclu, Vinh, Heyes, Franck, Schaffner and Bouchez2003) and P=0.41 MPa (Crabos et al., Reference Crabos, Huang, Boursat, Maurel, Ruffel, Krouk and Boursiac2023), we were able to calculate the variation in Π together with the variation in P while P remains positive, i.e. under a mild water deficit treatment (Figure 2, green curve in the gray zone). On average, the maximal volume variation in cortical cells due to turgor loss, thus before plasmolysis, is around 17%, which translates into an increase in Π of about 20%. As mentioned earlier for P, only one study reported the possible transient formation of DCP5-GFP condensates upon “mild water deficit” (Wang et al., Reference Wang, Yang, Zhang, Lu, Wang, Pan, Qiu, Men, Yan and Xiao2024b). With more severe water deficit treatments, P is lost and water efflux from the cell will lead to plasmolysis (if solutes can cross the cell wall, such as NaCl or sorbitol) or cytorrhesis (in the case of solutes blocked by the cell wall, such as PEG8000) and, in such case, the increase in Π will be directly proportional to the outer osmotic potential (Figure 2, water potentials lower than -0.5MPa). Almost all published works on molecular condensates under WD treatments fall within that range (Figure 2, bottom). Papers cited herein typically have reported increases in Π of more than 75 % (see Chong et al., Reference Chong, Foo, Lin, Wong and Verslues2019) Figure 2). Hence, the formation of condensates reported thus far is clearly more correlated to Π and related parameters than to P.

The question of the exact physico-chemical cue underlying our use of Π in this review is vast and will not be addressed here. However, we want to point out that several works cited herein have linked molecular crowding sensing to condensation or phase separation (Boyd-Shiwarski et al., Reference Boyd-Shiwarski, Shiwarski, Griffiths, Beacham, Norrell, Morrison, Wang, Mann, Tennant and Anderson2022; Meneses-Reyes et al., Reference Meneses-Reyes, Rodriguez-Bustos and Cuevas-Velazquez2024; Wang et al., Reference Wang, Zhang, Huai, Peng, Wu, Lin and Fang2022, Reference Wang, Yang, Zhang, Lu, Wang, Pan, Qiu, Men, Yan and Xiao2024b). The relationship between WD and molecular crowding or volume change could be straightforward since an efflux of water from a cell with a selective membrane directly results in a volume reduction and an increase in solute and protein concentrations. But besides the fact that a biological system is far more complex than an in vitro system, such as a dialysis setup, the interpretation of condensation might have to go beyond. Indeed, two driving force categories control LLPS or condensation: entropic forces linked to solute concentrations, and enthalpic forces related to noncovalent interactions, and making the distinction between both is difficult in cells (Romero-Perez et al., Reference Romero-Perez, Dorone, Flores, Sukenik and Boeynaems2023). Moreover, Watson and co-authors (Watson et al., Reference Watson, Seinkmane, Styles, Mihut, Krüger, McNally, Planelles-Herrero, Dudek, McCall and Barbiero2023) addressed the formation of condensates from the perspective of Ψ_Π . These authors showed, using two different macromolecules of similar sizes, PEG and Dextran, that PEG (which decreases Ψ_Π ) led to bovine serum albumin (BSA) condensation while dextran (which has a low impact on Ψ_Π ) did not. This result suggests that molecular crowding alone did not provoke LLPS. Furthermore, the condensate state does not always follow the “obvious” molecular crowding way. For instance, at natural expression levels in plant seeds, FLOE1 droplets dissolve during the desiccation process, even though the loss of water increases its intracellular concentration (Dorone et al., Reference Dorone, Boeynaems, Flores, Jin, Hateley, Bossi, Lazarus, Pennington, Michiels and De Decker2021; Romero-Perez et al., Reference Romero-Perez, Dorone, Flores, Sukenik and Boeynaems2023). Beyond concentration and Ψ_Π , Watson and colleagues also highlighted the complexity of physical cues acting on LLPS. In particular, using BSA and PEG solutions, they have shown that an increase in temperature could revert PEG-induced BSA condensation. The authors then concluded that molecular crowding could not be the primary condensation factor, but rather the solvent thermodynamics (for instance, temperature, see Figure 4g and h of Watson et al., Reference Watson, Seinkmane, Styles, Mihut, Krüger, McNally, Planelles-Herrero, Dudek, McCall and Barbiero2023). Therefore, there is no unique way to understand and anticipate condensate formation in plant cells.

From the works cited above, Π rather than P seems thus far to be the physicochemical parameter to work with to better understand and anticipate biomolecular condensate formation. However, solvent thermodynamics influence protein hydration which involves the organization of a few layers of water molecules around macromolecules. This led to the idea of two “kinds” of water: structured water molecules that are influenced by their (in-)direct interactions with macromolecules, and free/bulk water molecules that are “available” for biochemical functions (Romero-Perez et al., Reference Romero-Perez, Dorone, Flores, Sukenik and Boeynaems2023; Watson et al., Reference Watson, Seinkmane, Styles, Mihut, Krüger, McNally, Planelles-Herrero, Dudek, McCall and Barbiero2023). Hydration of a protein is modified, among others, by its phosphorylation status, which is influenced by osmolarity or temperature changes and implicates condensate formation (Watson et al., Reference Watson, Seinkmane, Styles, Mihut, Krüger, McNally, Planelles-Herrero, Dudek, McCall and Barbiero2023). Moreover, it was shown in the animal field that hydrostatic pressure at physiological ranges (80–190kPa) activates WITH NO LYSINE KINASES (WNK, Humphreys et al., Reference Humphreys, Teixeira, Akella, He, Kannangara, Sekulski, Pleinis, Liwocha, Jiou and Servage2023). Whether post-translational modifications, mediated through P-dependent proteins, can shift the condensation parameters of other proteins to ranges of P observed in plant cells should be investigated. And then, this raises again the interesting possibility that P can be a physical factor acting on the formation of functionally relevant biomolecular condensates in vivo.