Unlike animals, plants cannot move to escape harsh conditions. Consequently, they must continuously monitor their environment and, when exposed to high temperatures, quickly adjust their expression of developmental and growth-related genes. Writing in Nature, Jung et al.1 describe a molecular process that might underlie this temperature responsiveness.
The expression of developmental and growth-related genes in animals and plants typically occurs in a rhythmic fashion over a 24-hour cycle. Such daily oscillations are controlled by a molecular loop of protein activity that provides what is termed the circadian clock. Clock-induced transcriptional changes enable plants to anticipate daily environmental changes.
In the model plant species Arabidopsis thaliana, one component of the circadian clock is a protein assembly called the evening complex. It is maximally active at dusk and represses the expression of many genes important for plant development. The evening complex comprises the transcription-factor protein ELF3 (Fig. 1), a small peptide known as ELF4 and a protein called LUX. Plants with mutations that disable the gene encoding ELF3 flower earlier than normal during development and grow long embryonic stems termed hypocotyls, suggesting that ELF3 has a key developmental role.
Temperature fluctuations are known to affect the circadian rhythm of plants. The growth of A. thaliana at 22 °C is normally restricted to the period around dawn, because of the repressive action of the evening complex at other times of day2. However, at 27 °C, this growth repression is relieved2, and plants show accelerated flowering and rapid hypocotyl elongation compared with growth at 22 °C. Yet the mechanism underlying such temperature-regulated growth has remained a mystery. Jung and colleagues propose that a physical process called phase separation is at the heart of plant responsiveness to heat.
To investigate, the authors focused on ELF3. They engineered A. thaliana so that ELF3 was replaced with a related version from two plants that do not show temperature-accelerated flowering: Solanum tuberosum (potato) and Brachypodium distachyon (a grass). The resulting A. thaliana plants were indistinguishable from wild-type A. thaliana at moderate temperatures, but were unable to accelerate flowering at a higher temperature, suggesting that ELF3 has a key role in temperature responsiveness.
To investigate further, Jung and colleagues focused on a region of ELF3 that is enriched in polar (hydrophilic) amino-acid residues, depleted of charged residues and predicted to be intrinsically disordered. Such protein regions are known as prion-like domains (PrDs), and have been proposed to mediate environmental responses in budding yeast (Saccharomyces cerevisiae)3,4. Jung et al. engineered A. thaliana to express a chimaeric protein, in which its normal PrD was replaced with the corresponding region of ELF3 from B. distachyon. The authors report that the engineered plant did not display temperature-accelerated flowering, indicating that this ELF3 domain might have a key role in establishing temperature responsiveness.
The PrD of A. thaliana contains continuous stretches of the amino acid glutamine that are called polyglutamine (polyQ) repeats. The authors note a correlation between plant species that have long polyQ repeats in this domain and accelerated growth at warm temperatures, suggesting that the polyQ repeats modify ELF3 temperature responsiveness. Because repeat expansions generally evolve rapidly compared with non-repetitive sequences, this correlation suggests a potential way for plants to adapt to the predicted higher temperatures arising from global warming.
To investigate the molecular changes underlying ELF3 temperature responsiveness, the authors used a range of biochemical, biophysical and cell-biological tests. They observed that, at low temperatures, ELF3 was diffusely distributed inside the cell, but when the temperature rose it assembled into microscopically visible ‘dots’ called puncta. This outcome depended on the presence of the PrD, and the number of observed puncta increased with the length of the polyQ repeats. Crucially, the formation of these puncta was reversed if the temperature fell, suggesting that this represents a normal assembly mechanism in response to heat, rather than an irreversible protein-aggregation event.
Previous work3–5 led to the proposal that PrD-containing proteins in budding yeast undergo a stimulus-dependent phenomenon called phase separation. This is a process by which a well-mixed protein solution ‘demixes’ into a dense phase (or condensate) and a dilute phase, comparable to the way that oil and water are partitioned into different phases6,7. To test whether the ELF3 PrD forms condensates, the authors performed in vitro experiments using a fragment of A. thaliana ELF3 containing the PrD. Indeed, this fragment showed temperature-dependent phase separation with a threshold for condensate formation at approximately 28 °C. By contrast, the corresponding fragment of ELF3 from B. distachyon did not form condensates under the same conditions. This indicates that the A. thaliana ELF3 PrD forms condensates in vitro in a temperature-dependent manner. However, whether the heat-induced assemblies observed in cells are condensates of inactive ELF3 remains to be established.
Next, the authors focused on ELF4, which binds to ELF3 in the vicinity of its PrD. Jung et al. found that ELF4 inhibits the temperature responsiveness of ELF3: plants that were engineered to express higher-than-normal levels of ELF4 were unable to respond to warm temperatures with accelerated flowering. This suggests that the binding of ELF4 to ELF3 modulates condensate assembly by ELF3. The regulation of phase separation by the action of a binding ligand is a widespread phenomenon termed polyphasic linkage8. However, more in vitro and in vivo experiments will be needed to determine whether polyphasic linkage of ELF3 and ELF4 underlies the inhibition of temperature-accelerated flowering.
The polyQ repeats modify the temperature responsiveness of ELF3, but alone they are probably insufficient to drive this responsiveness, and the identity of the amino-acid residues responsible for driving this property of ELF3 is unknown. Work so far in other systems9 to understand phase separation of PrD-containing proteins has focused mainly on those that undergo phase separation on cooling, such as a human protein called FUS. Amino-acid residues that are aromatic (those that contain a benzene ring, or an analogue thereof), polar or basic provide cohesive forces for intramolecular interactions in FUS that enable phase separation9. ELF3 undergoes phase separation when the temperature rises, rather than falls, and previous studies5,10 suggest that heat-induced phase separation of elastomeric proteins (flexible proteins with biomechanical functions) often depends on hydrophobic amino-acid residues. Indeed, the ELF3 PrD contains several hydrophobic amino-acid residues, such as methionine, but their role in condensate assembly is unknown.
Another crucial point to establish is how polyQ repeats modify condensate assembly. One possibility is that these repeats alter ELF3 solubility, thus shifting the temperature at which phase separation occurs.
This study raises some exciting questions for the future. For example, what are the properties and composition of these ELF3 condensates in plant cells? Can the ELF3 PrD respond to signals besides temperature, such as other physico-chemical cues? How widespread is this mechanism in plants, and do organisms other than plants regulate components of their circadian clocks through stimulus-dependent phase separation? Repeats of the amino acids threonine and glycine in a transcription-factor protein modulate the temperature responsiveness of the circadian clock in the fruit fly Drosophila melanogaster11. This suggests that phase separation could have a much broader role in coupling environmental inputs to biological rhythms than had been thought.
Nature 585, 191-192 (2020)