Temperature regulation in plants: From molecular mechanisms to climate‐resilient crop improvement

ABSTRACT

Temperature is a fundamental environmental determinant of plant growth, development, reproduction, and yield, and increasing thermal variability poses a major threat to global food security. Plants have evolved multilayered thermosensory systems that perceive cold and heat, and convert these cues into coordinated physiological, molecular, and developmental responses through interconnected regulatory networks operating across cellular and chromatin levels. Beyond stress adaptation, temperature also controls key developmental programs. Thermomorphogenesis confers architectural plasticity under moderately elevated temperatures through the integrated actions of hormones, light signaling, the circadian clock, and chromatin remodeling. Temperature‐sensitive genic male sterility links RNA metabolism, translational fidelity, and protein quality control to reproductive thermosensitivity, providing the genetic basis of two‐line hybrid breeding systems. Vernalization represents a temperature‐encoded epigenetic memory, in which prolonged cold establishes stable chromatin states that repress FLC in Arabidopsis and activate VRN1 in cereals, ensuring seasonal flowering competence while requiring resetting in the next generation. This review summarizes recent advances in temperature perception, signaling, regulatory networks, and epigenetic memory, and discusses how natural variation, genome editing, and AI‐assisted prediction can accelerate molecular design breeding for climate‐resilient crops.

This review summarizes advances in thermosensory and signaling networks, thermomorphogenesis, male sterility, and vernalization, and discusses how natural variation, genome editing, and AI‐assisted approaches enable molecular design breeding of climate‐resilient crops.

INTRODUCTION

Global climate change, characterized by rising average temperatures and an increasing frequency of extreme weather events, poses a profound threat to ecosystem stability and agricultural productivity worldwide (Challinor et al., 2014). As sessile organisms, plants cannot escape adverse environments and have therefore evolved intricate mechanisms to perceive and respond to temperature fluctuations, which critically shape their growth, development, and geographic distribution (Kan et al., 2023).

Temperature influences nearly every aspect of plant biology, from cellular homeostasis to developmental programming (Figure 1). At the extremes of the thermal spectrum, both cold and heat stress disrupt physiological balance, impair photosynthetic efficiency, and ultimately lead to substantial reductions in crop yield and quality. It is estimated that each 1°C rise in the global mean temperature reduces the productivity of major staple crops by 3.1%–7.2% (Zhao et al., 2017a). Moreover, extreme temperature events have caused cumulative economic losses exceeding USD 69 billion globally between 1975 and 2023 (Kang et al., 2025). These impacts underscore the urgency of elucidating how plants perceive, transduce, and adapt to temperature stress signals in the context of climate‐resilient agriculture.

Figure 1.

Temperature orchestrates diverse regulatory networks governing plant development and stress adaptation across the thermal spectrum

Temperature shapes virtually all aspects of plant growth and development through multiple, interconnected regulatory pathways spanning cold stress, vernalization, thermomorphogenesis, thermosensitive genic male sterility (TGMS), and heat stress. Cold stress: Upon perception by cold sensors (e.g., COLD1), Ca²⁺‐dependent signaling cascades activate upstream transcription factors that induce DREB1/CBFs, thereby promoting the expression of cold‐responsive (COR) genes and enhancing freezing and chilling tolerance. Vernalization: Prolonged exposure to low temperature promotes flowering via distinct species‐specific modules, including VIN3‐FLC‐FT in Arabidopsis and the VRN1–VRN2–VRN3 pathway in cereals, both converging on floral activators. Thermomorphogenesis: Elevated ambient temperatures are integrated by the transcription factor PIF4, which coordinates thermal, hormonal, and circadian signals to stimulate auxin biosynthesis and signaling, driving adaptive changes in plant architecture. TGMS: High temperatures compromise the function of the RNase Z homolog TMS5, leading to defects in RNA metabolism and triggering a temperature‐dependent transition from pollen fertility to male sterility. Heat stress: Plants deploy multilayered defense mechanisms to maintain cellular homeostasis under acute or prolonged heat stress. These include membrane‐ and lipid‐associated signaling modules such as TT3.1 and the DGK‐TT2 pathway, HSFA1b‐mediated stomatal closure, and endoplasmic reticulum stress responses mediated by NTL3, bZIP74, and bZIP60 that activate the unfolded protein response (UPR). In parallel, the canonical heat shock response (HSR) is initiated through heat shock factor (HSF)‐dependent induction of heat shock proteins (HSPs), collectively safeguarding protein stability and cell viability.

Beyond stress responses, moderately elevated temperatures induce a suite of adaptive morphological changes collectively referred to as thermomorphogenesis. These include hypocotyl and petiole elongation, leaf hyponasty, and accelerated flowering (Quint et al., 2016; Park et al., 2021; Zhang et al., 2021b). Such architectural adjustments optimize plant cooling capacity and reproductive success under warm conditions, reflecting a finely tuned developmental plasticity in response to ambient temperature. Temperature also plays a decisive role in reproductive development. This property has been ingeniously exploited in modern agriculture through the generation of thermosensitive genic male‐sterile (TGMS) lines in rice, which show male sterility at high temperatures but regain fertility under cooler conditions. The molecular mechanisms underlying this reversible fertility transition not only provide a compelling model for studying plant thermosensitivity but also the foundation of the two‐line hybrid breeding system, a major innovation that has greatly enhanced hybrid seed production efficiency (Figure 1).

At the opposite end of the thermal continuum lies vernalization, a process by which prolonged exposure to low, non‐freezing temperatures confers flowering competence after winter (Zhang et al., 2019b). Vernalization exemplifies the remarkable ability of plants to sense, memorize, and integrate thermal information to synchronize reproductive development with favorable seasonal cues (Amasino, 2010; Figure 1).

Importantly, temperature signaling does not operate in isolation. Instead, it is tightly interconnected with other endogenous regulatory networks, including phytohormones and the circadian clock, as well as with external environmental cues such as light and photoperiod. The molecular basis and physiological consequences of these multilayered interactions have been comprehensively summarized in several recent reviews (Ding et al., 2020; Qi et al., 2022; Li et al., 2023a). Therefore, these aspects will not be discussed in detail here. In this review, we focus on recent advances in understanding plant responses to temperature across the full thermal continuum. We first describe the molecular and cellular mechanisms underlying cold and heat perception, signaling, and adaptation. We then discuss temperature‐regulated developmental programs, including thermomorphogenesis and TGMS, which bridge environmental response and developmental regulation. Finally, we summarize current knowledge on vernalization, emphasizing its perception and epigenetic regulation. By linking these fundamental mechanisms with molecular design breeding, natural variation, and AI‐assisted prediction, we highlight emerging strategies to translate temperature biology into agricultural innovation and the development of climate‐resilient crops.

PHYSIOLOGICAL IMPACTS AND PLANT RESPONSES TO TEMPERATURE EXTREMES

Temperature extremes can be broadly categorized into cold stress, encompassing chilling (0°C–15°C) and freezing (< 0°C), and heat stress, which occurs when temperatures exceed species‐specific thresholds, typically 10°C–15°C above the optimal growth range. Both cold and heat stress impose severe and often irreversible constraints to plant growth, development, and overall fitness across multiple developmental stages.

During germination and early seedling establishment, temperature stress markedly suppresses seed germination and seedling growth. Cold stress frequently induces wilting of leaves and hypocotyls that can progress to tissue necrosis, whereas heat stress often results in albino cotyledons and shortened radicles. Both stresses impair root development, promote leaf wilting, and may ultimately cause seedling mortality (Zhou et al., 2022a; Li et al., 2023a). The reproductive stage represents one of the most temperature‐sensitive phases of the plant life cycle. Both cold and heat stress compromise anther development and dehiscence, reduce stigma receptivity, and inhibit pollen germination and pollen tube elongation (Tazib et al., 2015; Shi et al., 2018a), collectively leading to reduced fertilization efficiency and seed set.

Cellular targets and damage mechanisms under temperature stress

At the cellular level, membrane integrity constitutes a primary target of temperature stress. Chilling stress decreases membrane fluidity (Kang et al., 2025), whereas heat stress increases it (Cano‐Ramirez et al., 2021). Despite these opposing biophysical effects, both conditions elevate membrane permeability, promote electrolyte leakage, disrupt transmembrane potential, and compromise cellular compartmentalization (Zhou et al., 2025b).

Photosynthesis is another highly temperature‐sensitive process. Under cold stress, photosynthetic efficiency declines due to inhibition of the triose phosphate translocator cycle and reduced ATP synthesis, accompanied by disruptions in thylakoid electron transport and carbon fixation (Wang et al., 2024c). In contrast, heat stress damages chloroplast thylakoid membranes, leading to grana disintegration and structural destabilization (Wang et al., 2017). In both cases, photosystem II (PSII) activity is markedly decreased, pigment contents decline, and the overall photosynthetic capacity is severely impaired (Wahid et al., 2007).

Temperature extremes also induce oxidative stress through excessive accumulation of reactive oxygen species (ROS). Elevated ROS production arises from disturbances in photosynthetic and mitochondrial electron transport chains, as well as from enhanced activity of NADPH oxidases. Excess ROS damage lipids, proteins, and nucleic acids, thereby impairing organelle function and cellular viability (Niu and Xiang, 2018; Liu et al., 2025c).

Physiological mechanisms underlying plant responses to temperature stress

In response to temperature extremes, plants activate a complex and highly coordinated network of physiological and biochemical responses to maintain cellular homeostasis. This adaptive capacity, referred to as stress tolerance, is not instantaneous but develops progressively through a process known as acclimation. Acclimation describes the ability of plants to enhance their tolerance to cold or heat stress via gradual physiological, biochemical, and molecular adjustments without changes to genomic sequences (Thomashow, 1998; Charng et al., 2023). Importantly, acclimation confers acquired tolerance, enabling plants to withstand subsequent exposure to temperatures beyond their basal tolerance range. Together, basal and acquired tolerance constitute an integrated thermal defense strategy.

During acclimation, plants undergo extensive physiological reprogramming, including cell wall remodeling, modulation of membrane composition and fluidity, accumulation of protective proteins, and activation of osmoprotectant and antioxidant systems. The following sections summarize these key processes and their roles in establishing temperature stress tolerance.

Cell wall remodeling

The plant cell wall serves as the first line of defense against temperature stress, as it is directly exposed to environmental fluctuations prior to the plasma membrane. Under cold conditions, the cell wall represents the initial site of ice crystal formation, and its structural properties strongly influence freezing tolerance. For example, increased pectin galactan content enhances freezing tolerance by reducing wall extensibility and increasing rigidity (Kutsuno et al., 2023; Takahashi et al., 2024). In rice, the cell wall extension protein COG2 (COLD‐TOLERANCE IN GENG RICE 2) negatively regulates cold tolerance by altering cell wall composition, including pectin and cellulose content (Feng et al., 2023).

In contrast, under heat stress, activation of pectin methylesterase (PME) and endo‐polygalacturonase promotes cell wall acidification and loosening, facilitating the release of calcium ions into the cytoplasm and triggering downstream heat stress signaling (Wu and Jinn, 2010; Wu et al., 2018). Thus, cold stress and heat stress elicit distinct, yet adaptive cell wall remodeling strategies that contribute to cellular protection.

Modulation of membrane composition and fluidity

The plasma membrane is another critical target of temperature stress, with its fluidity largely determined by the ratio of unsaturated to saturated fatty acids. Under cold stress, fatty acid desaturases (FADs) convert saturated fatty acids into unsaturated forms, thereby maintaining membrane fluidity and enhancing cold tolerance (Soria‐García et al., 2019). In rice, loss of COLD8 function increases the accumulation of unsaturated fatty acids, such as 16:1 and 18:3, resulting in improved chilling tolerance (Zheng et al., 2025). Consistently, mutations in membrane remodeling genes, including acyl‐CoA binding protein 1 (ACBP1), freezing‐sensitive protein 2 (SFR2), and ACYL‑LIPID DESATURASE 2 (ADS2), confer hypersensitivity to cold stress (Du et al., 2010; Chen et al., 2012; Chen and Thelen, 2013; Wang et al., 2016b).

Conversely, heat stress is associated with increased lipid saturation. Levels of monogalactosyldiacylglycerol (MGDG) and phosphatidylglycerol (PG) decline, whereas triacylglycerol (TAG) and diacylglycerol (DAG) accumulate to stabilize membrane structure (Munnik, 2014; Hou et al., 2016). In addition to their structural roles, lipid‐derived molecules such as phosphatidic acid (PA) act as second messengers in temperature stress signaling pathways (Munnik, 2014). Together, lipid composition and dynamic remodeling are crucial determinants of plant responses to temperature extremes.

Accumulation of protective proteins

Plants synthesize a diverse array of protective proteins in response to temperature fluctuations. Under cold stress, antifreeze proteins (AFPs) inhibit ice crystal formation, while dehydrins such as COR15A (COLD‐RESPONSIVE 15A) enhance freezing tolerance by stabilizing chloroplast membranes (Artus et al., 1996). Other dehydrins, including COR47, LTI29, and LTI30, preserve thylakoid membrane integrity and chloroplast function during low‐temperature conditions (Bozovic et al., 2013).

In contrast, heat stress induces the accumulation of heat shock proteins (HSPs), such as HSP70, HSP90, and HSP101, which function as molecular chaperones to prevent protein misfolding and aggregation, thereby substantially improving plant survival under high temperatures (Hahn et al., 2011; Kim and An, 2013; Lin et al., 2014).

Activation of osmoprotectant and antioxidant systems

To mitigate osmotic imbalance and oxidative damage, plants accumulate osmoprotectants and activate antioxidant defense systems. Soluble sugars stabilize proteins, regulate osmotic potential, and scavenge ROS (Nishizawa et al., 2008). Proline acts as an osmolyte, antioxidant, and signaling molecule that activates stress‐responsive gene expression (Zhou et al., 2025b). Similarly, glycine betaine contributes to enhanced tolerance to temperature stress (Park et al., 2004; Quan et al., 2022).

Excess ROS generated under temperature stress are detoxified by antioxidant enzymes such as superoxide dismutase (SOD), catalase (CAT), and ascorbate peroxidase (APX), which maintain cellular redox homeostasis and protect cellular structures from oxidative injury (Kang et al., 2025).

In summary, plants use an integrated common suite of physiological strategies, including cell wall reinforcement, membrane remodeling, biosynthesis of protective proteins, osmoprotectant accumulation, and activation of antioxidant defenses, to preserve cellular homeostasis and enhance survival under either cold or heat stress conditions.

THE PERCEPTION AND EARLY SIGNAL TRANSDUCTION OF TEMPERATURE STRESS IN PLANTS

Plants have evolved sophisticated mechanisms to perceive temperature fluctuations and initiate appropriate signaling responses. This section summarizes current advances in understanding how plants sense cold and heat stress and activate early signal transduction cascades (Figure 2).

Figure 2.

Putative temperature‐sensing and signal transduction pathways in plants

Plants perceive temperature fluctuations through multiple sensing modules at the plasma membrane and within intracellular compartments. Membrane‐localized COLD1 acts as the primary thermosensor that initiates downstream Ca²⁺ signaling. Cold and heat stresses evoke distinct cytosolic Ca²⁺ signatures via Ca²⁺ channels (CNGCs) and transports (ANNs), which are decoded by Ca²⁺ sensors and associated proteins to regulate COR and HSR gene expression. Under cold stress, the plasma membrane receptor complex comprising COLD6 and the cold‐induced protein OSM1 generates the second messenger 2′,3′‐cAMP to activate downstream defense pathways. Cold exposure also promotes the nuclear translocation of NLP7, CPK17, and 14–3–3 proteins to fine‐tune cold responses. In contrast, heat stress induces relocalization of the thermosensor TT3.1 from the plasma membrane to endosomes to mediate TT3.2 degradation for protecting chloroplasts. Heat stress is also sensed by diacylglycerol kinase 7 (DGK7) at the plasma membrane, which activates a phosphatidic acid‐dependent pathway that modulates nuclear transcription via cAMP degradation, a process negatively regulated by the G protein subunit OsTT2. HSFA1b acts as a heat sensor, translocating from the cytosol to the nucleus, relieving its inhibition of OST1 kinase activity, thereby promoting stomatal closure under high temperature while activating heat signaling. Additionally, heat‐dependent nuclear translocation of GAPC, NTL3, and NLP3 further activates HSR genes. The BAM1–PBS1 receptor complex facilitates rapid heat signaling by promoting H₂O₂ production and HSFA1b activation, a process amplified by the CLE40 peptide. Chloroplasts act as a central signaling and metabolic hub that integrates multiple molecular pathways to fine‐tune rice adaptation under temperature stress. Arrows denote positive regulation and T‐bars denote negative regulation.

Membrane‐associated temperature perception and early second messenger‐mediated early signaling

Alteration in plasma membrane fluidity represents one of the earliest and most conserved events in plant temperature perception. Temperature‐induced changes in lipid composition and membrane organization affect the conformation and activity of membrane‐associated proteins, thereby initiating downstream signaling. One immediate consequence of these biophysical changes is the rapid influx of Ca²⁺ across the plasma membrane (Knight et al., 1996), establishing Ca²⁺ as a central second messenger that links temperature perception to intracellular responses.

In rice, the G‐protein regulatory factor COLD1 (chilling tolerance divergence 1) exemplifies this coupling between membrane sensing and Ca²⁺ signaling. COLD1 localizes to the plasma membrane, where it interacts with the Gα subunit RGA1 to mediate cold perception and induce Ca²⁺ influx, thereby enhancing chilling tolerance (Ma et al., 2015). In contrast, under heat stress, the Gγ subunit OsTT2 (Thermo‐tolerance 2) negatively regulates thermotolerance by modulating the Ca²⁺‐dependent signaling pathway and altering wax biosynthesis (Kan et al., 2022). Extending this paradigm, recent studies demonstrate that heat‐induced plasma membrane lipid remodeling itself functions as a primary thermal signal. Diacylglycerol kinase 7 (DGK7) senses elevated temperatures at the plasma membrane and activates a PA‐dependent pathway that modulates nuclear transcription (Kan et al., 2025; Figure 2). Together, these findings highlight the plasma membrane as both a thermosensory interface and a regulatory hub for second messengers, including Ca²⁺ and cAMP.

Multiple Ca²⁺‐permeable channels and transporters contribute to temperature stress‐induced cytosolic Ca²⁺ elevation. These include mechanosensitive channels (MCA1 and MCA2), cyclic nucleotide‐gated channels (CNGCs), and annexins (Qiao et al., 2015; Mori et al., 2018; Liu et al., 2021). In Arabidopsis, CNGC5, CNGC6, and CNGC20, and in rice, OsCNGC9, enhance freezing and chilling tolerance by facilitating Ca²⁺ influx during cold exposure (Wang et al., 2021; Peng et al., 2024b; Ming et al., 2025). Notably, OsCNGC14 and OsCNGC16 confer tolerance to both cold and heat stress (Cui et al., 2020), suggesting that CNGCs may serve as shared nodes within temperature‐responsive signaling networks (Figure 2). Whether these channels act as direct thermosensors or as downstream signal amplifiers remains an open question.

Decoding Ca²⁺ signals and organelle‐based early responses

Elevated cytosolic Ca²⁺ levels are decoded by Ca²⁺‐binding sensor proteins, including calcium‐dependent protein kinases (CPKs) and calcineurin B‐like protein (CBL)‐CBL interacting protein kinase (CIPK) modules. These sensors convert Ca²⁺ signatures into phosphorylation cascades that regulate gene expression and stress responses. In Arabidopsis, CPK28 transduces cold‐induced Ca²⁺ signals by phosphorylating the transcription factor NLP7 (NIN‐LIKE PROTEIN 7), thereby enhancing freezing tolerance (Ding et al., 2022). Similarly, in rice, OsCPK24 acts as a positive regulator of cold tolerance (Liu et al., 2018b).

Under cold stress, endoplasmic reticulum (ER)‐localized OsCRT3 (Calreticulin 3) undergoes a conformational change that strengthens its interaction with OsCIPK7, promoting kinase activation and cold signal transduction (Zhang et al., 2019a; Guo et al., 2023). In parallel, chloroplast‐associated signaling also plays a critical role. The putative calmodulin‐binding protein, COG3 (CHILLING‐TOLERANCE IN GENG/JAPONICA RICE 3), interacts with chloroplast protease OsFtsH2 (FILAMENTATION TEMPERATURE‐SENSITIVE H 2) to facilitate the removal of cold‐damaged D1 protein, thereby maintaining photosynthetic efficiency under chilling stress (Liu et al., 2024a). Moreover, a chloroplast‐localized vitamin E–vitamin K1 subnetwork downstream of COLD1 modulates natural variation in chilling tolerance among rice subspecies (Luo et al., 2021; Figure 2). These findings position the chloroplast as a central signaling and metabolic hub in early cold stress responses.

ROS signaling and its integration with Ca²⁺ dynamics

ROS are continuously produced in plant cells but are normally maintained at low levels by antioxidant systems (Mittler et al., 2011; Pucciariello et al., 2012). Temperature stress disrupts this balance, leading to ROS accumulation. Importantly, controlled and transient ROS bursts act as early signaling cues that regulate gene expression and promote stress adaptation (Pucciariello et al., 2012; Zandalinas et al., 2020).

Under heat stress, membrane‐localized respiratory burst oxidase homologs (RBOHs) constitute the primary sources of ROS (Evans et al., 2016). In Arabidopsis, rboh mutants accumulate less H₂O₂ under heat stress but display markedly reduced survival (Miller et al., 2009; Liu et al., 2025c). Specifically, mutations in AtRBOHB and AtRBOHD impair the binding of the master heat shock factor HSFA1a to heat shock elements (HSEs), while exogenous H₂O₂ directly activates HSFA1a (Volkov et al., 2006). In addition, heat‐induced H₂O₂ activates MAPK kinase kinase ANP1, which phosphorylates MPK3 and MPK6; the activated MPK6 subsequently enhances HSFA2 activation and nuclear accumulation, reinforcing HSP transcription (Kovtun et al., 2000; Mittler et al., 2011).

ROS signaling is strongly interconnected with Ca²⁺ dynamics. Elevated H₂O₂ levels induce the expression of the annexin gene OsANN1 in rice (Qiao et al., 2015), while CPKs phosphorylate and activate RBOHD, thereby promoting ROS production (Ogasawara et al., 2008). Conversely, exogenous H₂O₂ rapidly induces cytosolic Ca²⁺ influx, further amplifying Ca²⁺ signaling pathways (Ogasawara et al., 2008). Notably, all nine rice RBOH proteins harbor EF‐hand Ca²⁺‐binding domains, and most are heat‐inducible (Wong et al., 2007), supporting the existence of a self‐reinforcing Ca²⁺–ROS feedback loop during temperature stress (Figure 2).

Emerging roles of cAMP in temperature stress signaling

Beyond Ca²⁺ and ROS, accumulating evidence suggests that cyclic adenosine monophosphate (cAMP) acts as an additional second messenger in plant temperature responses. Under cold stress, rice COLD6 forms a plasma membrane complex with the cold‐induced protein osmotin‐like 1 (OSM1), promoting the production of 2′,3′‐cAMP to activate downstream cold defense pathways (Luo et al., 2024). Under heat stress, 3′,5′‐cAMP signaling homeostasis is dynamically regulated, and perturbation of this signaling pathway reshapes nuclear transcriptional programs associated with thermotolerance (Kan et al., 2025). Interestingly, reports have shown that HSFA1b, beyond its role as a transcription factor, can serve as an adenylate cyclase to sense heat signals and modulate stomatal closure (Zhang et al., 2025f; Figure 2).

Collectively, early temperature perception in plants is initiated at the plasma membrane and transduced through an interconnected network of second messengers, including Ca²⁺, ROS, and cAMP. These signaling modules converge on transcriptional and physiological reprogramming, enabling plants to rapidly sense, integrate, and adapt to thermal fluctuations.

Receptor‐like kinases in temperature stress signaling

Plasma membrane‐localized receptor‐like kinases (RLKs) function as key molecular sentinels that perceive extracellular cues and relay them into intracellular signaling cascades, enabling plants to dynamically respond to environmental fluctuations (Liang and Zhou, 2018). Accumulating evidence indicates that both RLKs and receptor‐like cytoplasmic kinases (RLCKs) play pivotal roles in modulating plant responses to temperature extremes by integrating extracellular perception with cytoplasmic and nuclear signaling networks.

RLKs in cold stress signaling

Multiple RLKs have been identified as important regulators of cold tolerance. In Arabidopsis, PSY1R (PLANT PEPTIDE CONTAINING SULFATED TYROSINE 1 RECEPTOR) is activated by low temperature and phosphorylates CNGC20, thereby enhancing Ca²⁺ channel activity and promoting cold‐induced Ca²⁺ influx (Peng et al., 2024b). Another well‐characterized regulator, CRLK1 (Ca²⁺/CaM‐regulated receptor‐like kinase 1), enhances freezing tolerance through dual modulation of MAPK signaling. CRLK1 suppresses the activation of MPK3/6 while stimulating the MEKK1–MKK2–MPK4 cascade, which in turn inhibits MPK3/6 activity (Yang et al., 2010; Zhao et al., 2017b). Because MPK3/6 act as a negative regulator of freezing stress responses, CRLK1 effectively fine‐tunes MAPK signaling to promote cold tolerance (Li et al., 2017a; Zhao et al., 2017b). By contrast, CRPK1 (cold‐responsive protein kinase 1) acts as a negative regulator of freezing tolerance. CRPK1 phosphorylates 14–3–3 proteins, facilitating their nuclear translocation and promoting degradation of CBF transcription factors (Liu et al., 2017). CRPK1 also interacts with the RLK KOIN (KINASE ON THE INSIDE) to regulate root cortex cell proliferation and primary root growth under chilling stress (Zhang et al., 2025d; Figure 2).

In rice, OsSERL2 (SOMATIC EMBRYOGENESIS RECEPTOR KINASE‐LIKE 2) forms a complex with the leucine‐rich repeat receptor‐like protein COG1 (CHILLING TOLERANCE IN GENGDAO/JAPONICA RICE 1), activating OsMPK3 and promoting chilling tolerance (Zhang et al., 2017a; Xia et al., 2023; Figure 2). Collectively, these studies illustrate the diverse strategies by which RLKs modulate cold‐responsive signaling pathways.

RLKs in heat stress signaling

RLKs also act as important regulatory hubs in thermotolerance pathways. In Arabidopsis, the ER‐localized SDF2 (stromal cell‐derived factor 2)–ERdj3B (endoplasmic reticulum‐localized DnaJ family 3B)–BiP complex ensures proper plasma membrane localization of ERECTA‐family RLKs, thereby supporting ovule development and heat tolerance (Leng et al., 2022). Recent work has uncovered a direct RLK–RLCK signaling axis in early heat perception. Heat stress rapidly induces the formation and activation of the BAM1 (BARELY ANY MERISTEM 1)–PBS1 (AVRPPHB SUSCEPTIBLE 1) complex, triggering RBOHD‐dependent H₂O₂ production, an early hallmark of heat stress signaling (Li et al., 2025a). This process is further amplified by CLE40 peptide binding to BAM1. The resulting ROS signal oxidatively activates HSFA1b, promoting its nuclear accumulation and transcriptional activation of heat‐responsive genes (Li et al., 2025a; Figure 2). This pathway establishes a direct mechanistic link between plasma membrane‐associated RLKs and nuclear transcriptional control during heat stress.

In crop species, RLKs also play critical roles in reproductive thermotolerance. In rice, TMS10 (Thermo‐Sensitive Genic Male Sterile 10) and its homolog TMS10L are indispensable for maintaining male fertility at elevated temperatures (Yu et al., 2017). In wheat, TaSERL2 phosphorylates the transcription factor TaBZR2, which subsequently regulates the expression of HSFs and HSPs to enhance thermotolerance (Hao et al., 2025).

Collectively, RLK‐mediated signaling modules function as pivotal nodes linking extracellular cues to intracellular transcriptional and physiological responses. Future studies should clarify the ligand–receptor specificity, downstream phosphorylation networks, and whether certain RLKs function as bona fide thermosensors. These insights will be important for exploiting RLK pathways in crop improvement.

Phase separation as a mechanism for temperature sensing and response

Temperature also modulates protein behavior through liquid–liquid phase separation (LLPS), an entropy‐driven process that enables rapid and reversible reorganization of biomolecules. In Arabidopsis, phytochrome B (phyB) acts not only as a red‐light receptor but also as a thermosensor across ambient temperatures (12°C–27°C). Elevated temperatures accelerate conversion of phyB from the active Pfr into the inactive Pr state (Jung et al., 2016; Legris et al., 2016). Notably, LLPS mediated by the intrinsically disordered N‐terminal extension (NTE) directly senses temperature, distinct from light‐induced photobody formation driven by the C‐terminal self‐association (Chen et al., 2022). Phytochromes also contribute to freezing tolerance (Franklin and Whitelam, 2007; Jiang et al., 2020), underscoring their broad thermoregulatory roles.

Additional thermoresponsive proteins undergo LLPS in response to elevated temperatures. ELF3 (EARLY FLOWERING 3) forms nuclear condensates at warm temperatures, relieving repression of flowering genes such as FLOWERING LOCUS T (FT) (Jung et al., 2020). At 37°C, TWA1 (THERMO‐WITH ABA‐RESPONSE 1) relocalizes into nuclear condensates, where it interacts with jasmonate‐associated transcription factors and co‐repressors to fine‐tune HSFA2 and HSP expression (Bohn et al., 2024). Similarly, FUST1 promotes heat stress granule formation via condensation (Geng et al., 2025).

Temperature‐dependent LLPS also influences plant immunity. GUANYLATE BINDING PROTEIN‐LIKE 3 (GBPL3) forms condensates at 23°C to activate CBP60g expression and salicylic acid biosynthesis, whereas elevated temperature (28°C) disrupts this process, leading to compromised immune responses (Huang et al., 2021; Kim et al., 2022).

RNA‐binding proteins, including RBGD2, RBGD4, and ALAB (acetylation lowers binding affinity), and CP29A, undergo LLPS during heat stress to regulate mRNA stability (Tong et al., 2022; Zhu et al., 2022). Conversely, cold stress induces LLPS of chloroplast‐localized CP29A, facilitating mRNA splicing and translation of specific COR genes (Legen et al., 2024). Together, these studies establish LLPS as a versatile and dynamic mechanism linking temperature perception to transcriptional, developmental, and stress responses.

Subcellular relocalization of proteins in response to temperature stress

Temperature‐induced protein relocalization provides another rapid and reversible mechanism for stress response and adaptation (Zhu et al., 2023). Chloroplasts are among the most heat‐sensitive organelles, and their protection is essential for plant survival. In rice, the TT3 locus encodes TT3.1, a transmembrane E3 ligase, and TT3.2, a chloroplast precursor protein. Upon heat stress, TT3.1 relocalizes from the plasma membrane to endosomes, promoting TT3.2 degradation and thereby safeguarding chloroplast integrity (Zhang et al., 2022a; Figure 2).

Several transcription factors also undergo temperature‐dependent nuclear translocation. In rice, membrane‐anchored OsNTL3 (NAC WITH TRANSMEMBRANE MOTIF 1‐LIKE 3) and cytoplasmic ONAC023 translocate to the nucleus during heat stress to regulate genes involved in ER protein folding, ROS homeostasis, water transport, and alternative splicing (Liu et al., 2020; Chang et al., 2024). In Arabidopsis, NTL6 relocates to the nucleus under cold stress to activate pathogen‐related genes (Seo et al., 2010). Similarly, AtNLP7 and OsNLP3 translocate to the nucleus upon cold and heat exposure, respectively, to induce stress‐responsive gene expression (Ding et al., 2022; Zhu et al., 2025a; Figure 2).

Temperature‐dependent relocalization also affects metabolic enzymes and kinases. Under heat stress, the glycolytic enzyme GAPC translocates to the nucleus, where it activates the transcription factor NF‐YC10 and facilitates heat signal transduction (Kim et al., 2020a). In maize, heat‐activated ZmCPK7 relocates from the plasma membrane to the cytoplasm to phosphorylate sHSP17.4 and RBOHB (Zhao et al., 2021b), whereas ZmCPK17 translocates from the cytoplasm to the nucleus under cold stress to phosphorylate the transcription factor COOL1 (COLD‐RESPONSIVE OPERATION LOCUS 1), enhancing its protein stability and attenuating cold tolerance (Zeng et al., 2025a; Figure 2). [Correction added on 16 June 2026, after first online publication: In the preceding sentence, “reducing” has been replaced with “enhancing”.] These dynamic relocalizations enable rapid translation of thermal cues into precise transcriptional and metabolic outputs.

Other potential thermosensing mechanisms

Temperature perception also occurs at the chromatin and RNA levels. In Arabidopsis, the histone variant H2A.Z is proposed to participate in thermal sensing within the 12°C–27°C range (Kumar and Wigge, 2010). Elevated temperatures promote HSFA1‐mediated eviction of H2A.Z from nucleosomes, leading to rapid gene activation, while cooling restores H2A.Z occupancy and transcriptional repression (Cortijo et al., 2017).

Temperature additionally influences RNA structure and translation. High temperatures induced hairpin formation in the 5′‐UTR of PIF7 mRNA, enhancing translation of growth‐promoting proteins (Chung et al., 2020). Conversely, RNA G‐quadruplexes formed in the 5′‐UTRs of specific transcripts stabilize mRNAs under cold stress, contributing to enhanced cold tolerance (Yang et al., 2022).

Together, these studies reveal that temperature perception and early signaling constitute a multilayered sensing architecture, spanning membrane dynamics, protein phase behavior, subcellular relocalization, chromatin remodeling, and RNA thermo‐responsiveness. This integrated system enables plants to perceive, transduce, and integrate thermal cues with remarkable precision and plasticity. Advancing mechanistic insights, particularly through structural, biophysical, and single‐molecule approaches, will be crucial for identifying bona fide thermosensors. Such knowledge holds considerable promise for the rational engineering of temperature‐responsive signaling networks to enhance crop resilience under global climate change.

MOLECULAR REGULATORY NETWORKS OF PLANT RESPONSES TO TEMPERATURE STRESS

Transcriptional and post‐transcriptional regulatory mechanisms

Plants respond to temperature stress through multilayered transcriptional and post‐transcriptional regulatory networks that confer both robustness and flexibility to gene expression programs (Figure 3). These networks integrate conserved core modules with species‐specific regulators to fine‐tune stress adaptation across developmental stages and environmental contexts.

Figure 3.

Multi‐layered molecular regulatory networks of plant responses to temperature stress

Plants deploy complex regulatory networks to achieve thermal tolerance, integrating transcriptional, translational, and post‐translational mechanisms to achieve thermal adaptation. The CBF‐dependent pathway functions as a central regulatory hub in the cold stress response network. Upon cold stress, CBFs/DREB1s are rapidly induced and activate downstream COR genes to enhance freezing tolerance. In heat stress networks, HsfA1s act as master regulators that orchestrate HSR gene expression, ensuring thermotolerance. At the post‐transcriptional level, alternative splicing governed by specific splicing factors diversifies the transcriptome and proteome to promote plant thermotolerance. Temperature stress also impairs mRNA translation and proteostasis, making proper ribosome biogenesis and rRNA processing critical for cellular function. Post‐translational modifications, including phosphorylation and ubiquitination, dynamically modulate protein activity and stability, while the ubiquitin–proteasome and SUMOylation pathways remove misfolded or damaged proteins. Epigenetic mechanisms, such as DNA methylation, histone modifications, and non‐coding RNA‐mediated regulation, further fine‐tune temperature responses. Additionally, heat stress specifically activates the UPR, which restores endoplasmic reticulum proteostasis through two major signaling branches, thereby mitigating ER stress under high temperature. Arrows denote positive regulation and T‐bars denote negative regulation.

Transcriptional regulation under cold stress

During cold acclimation, the CBF‐dependent pathway represents the most extensively characterized regulatory module. In Arabidopsis, cold stress rapidly induces C‐REPEAT/DEHYDRATION RESPONSIVE ELEMENT (DRE)‐BINDING FACTORs (CBFs/DREB1s), which activate a large set of COR genes to enhance freezing tolerance (Jaglo‐Ottosen et al., 1998). This CBF regulon is highly conserved across angiosperms, constituting a core cold‐response module (Shi et al., 2018b; Zhang et al., 2019b; Nie et al., 2022; Yang et al., 2023c).

CBF expression is controlled by a complex and dynamic upstream network. Transcriptional activators include ICE1/2 (INDUCER OF CBF EXPRESSION 1/2), BZR1 (BRASSINAZOLE‐RESISTANT 1), CESTA, and CAMTAs (CALMODULIN‐BINDING TRANSCRIPTION ACTIVATORs), whereas EIN3, MYB15, WRKY41, and PIF4/7 act as repressors, enabling precise temporal and amplitude control of CBF induction (Chinnusamy et al., 2003; Agarwal et al., 2006; Doherty et al., 2009; Fursova et al., 2009; Lee and Thomashow, 2012; Shi et al., 2012; Kim and Park, Gilmour, et al., 2013; Eremina et al., 2016; Li et al., 2017b; Wang et al., 2023d). In crops, this regulatory architecture is further diversified. In major cereals such as rice and maize, lineage‐specific regulators, including ZmbZIP68, COOL1, OsMYBS3, and OsERF52, fine‐tune DREB1/CBF expression, thereby modulating cold responsiveness in a species‐ and genotype‐dependent manner (Su et al., 2010; Jiang et al., 2022; Kan et al., 2023; Xu et al., 2024; Zeng et al., 2025a). In tomato, SlMPK1/2 phosphorylate SlBBX17 to enhance its interaction with HY5 and promote CBF expression (Song et al., 2023) Together, these studies reveal a conserved, yet highly adaptable CBF‐centered transcriptional network that has been extensively rewired during crop diversification (Figure 3).

Transcriptional regulation under heat stress

In contrast, heat stress responses are predominantly governed by HSFs, with HsfA1s acting as master regulators (Yoshida et al., 2011; Ohama et al., 2017; Andrasi et al., 2021). Under non‐stress conditions, HsfA1s are sequestered in an inactive complex with HSP70/90 chaperones. Heat stress induces chaperone dissociation, allowing HsfA1s' nuclear translocation and activation of downstream heat‐responsive genes (Ohama et al., 2015, 2017; Figure 3).

Among HsfA1 targets, HsfA2 is essential for acquired thermotolerance, whereas HsfB1 and HsfB2b act as negative feedback regulators to prevent overactivation (Ikeda et al., 2011; Ohama et al., 2017). Another key HsfA1 target, DREB2A, functions as a central integrator of heat and drought signaling. Its expression and activity are controlled by multiple regulators, including JUB1, MBF1c, NFXL1, and cofactors such as NF‐YC10 and DPB3‐1, forming a multilayered regulatory hierarchy in Arabidopsis (Liu et al., 2011; Yoshida et al., 2011; Sato et al., 2014; Ding et al., 2020; Zhu et al., 2024; Figure 3).

In crops, heat transcriptional networks show additional layers of integration and diversification. In rice, four AP2/ERF transcription factors (ERF74, 77, 108, and 125) synergistically activate HsfA2c, thereby amplifying thermotolerance (Luo et al., 2025). Moreover, the nitrogen sensor NLP3 links nutrient status with heat responses by modulating HsfA3 and HsfA7 expression (Zhu et al., 2025a), highlighting crosstalk between metabolic and thermal signaling pathways.

Beyond HSFs, diverse transcription factor families contribute to thermotolerance. For instance, in rice, NAT1 negatively regulates bHLH110, which activates wax biosynthesis genes CER1 (ECERIFERUM 1) and CER1L (ECERIFERUM 1‐Like), thereby modulating cuticular responses to heat stress (Lu et al., 2025). In maize, ZmMYB104 enhances heat tolerance via ZmCAT2 regulation, while rice ONAC023 coordinates growth and thermotolerance. In tomato, SlJA2L integrates thermotolerance with fruit ripening via ethylene signaling (Chang et al., 2024; Liang et al., 2025; Zhang et al., 2025a). These findings underscore extensive integration of heat signaling with developmental and metabolic programs across plant species.

Post‐transcriptional regulation under temperature stress

Post‐transcriptional regulation adds a critical layer of plasticity to temperature stress responses. Alternative splicing (AS) is rapidly and extensively reprogrammed under both cold and heat stress, reshaping the functional transcriptome (Lee and Rio, 2015). Cold‐induced AS in Arabidopsis and rice is mediated by splicing factors, including OsRS33, OsRS2Z38, LSM2‐8 (sm‐like 8), STA1 (stabilized 1), and RCF1 (regulator of CBF gene expression 1), which collectively ensure proper expression of COR genes (Lee et al., 2006; Guan et al., 2013a; Carrasco‐López et al., 2017; Calixto et al., 2018; Zhong et al., 2024). Notably, CBF proteins interact with the spliceosome component SKIP (SKI‐INTERACTING PROTEIN) to form nuclear condensates, enhancing the splicing efficiency of COR transcripts and reinforcing acquired freezing tolerance (Fu et al., 2025; Figure 3).

AS is equally important during heat stress (Ling et al., 2021). Temperature‐dependent splicing generates functionally distinct isoforms of key regulators, such as HSFA2d in Arabidopsis and HTG3 (HSFA2d) in rice, which are required for thermotolerance at multiple developmental stages (Cheng et al., 2015; Wu et al., 2022b). In maize, ZmHSFA2B produces full‐length and truncated isoforms that fine‐tune heat responses (Song et al., 2025a). RNA‐binding proteins such as OsGRP3 and OsGRP162 prevent heat‐induced exon skipping, whereas AS of OsbZIP58 under high temperature disrupts starch and protein metabolism, leading to chalkiness (Xu et al., 2020a; Yang et al., 2023a; Figure 3).

Overall, temperature responses are orchestrated through tightly integrated transcriptional and post‐transcriptional regulation. Conserved core modules ensure robustness, while species‐specific regulators and alternative splicing confer plasticity, enabling plants and crops to fine‐tune gene expression across diverse thermal environments.

Translation and post‐translational regulation

Temperature stress profoundly perturbs protein synthesis and homeostasis, necessitating multilayered regulatory mechanisms to safeguard proteome integrity and cellular function.

Translational regulation under cold stress

During cold acclimation, translational reprogramming is essential for sustaining protein synthesis. Ribosome biogenesis factors of the REIL (REI1‐LIKE) family promote ribosome remodeling and the accumulation of functional cytoplasmic ribosomal subunits under chilling conditions (Beine‐Golovchuk et al., 2018). REIL2/STCH4 (sensitive to chilling 4) enhances translation of CBF transcripts, thereby strengthening freezing tolerance (Yu et al., 2020). In chloroplasts, RNA‐binding proteins such as RBD1 (RNA‐binding domain 1), CP29A, and CP31A, along with the rRNA maturation factor NUS1 (N utilization substance 1), coordinate rRNA processing and translation under cold stress (Kusumi et al., 2011; Kupsch et al., 2012; Wang et al., 2016c; Figure 3).

Epitranscriptomic regulation further fine‐tunes translational efficiency. For instance, MTA1 (mRNA adenosine methylase A)‐mediated m⁶A modification enhances translation of DGAT1 (diacylglycerol acyltransferase 1) without altering transcript abundance, improving cold tolerance (Govindan et al., 2022; Wang et al., 2023c; Figure 3).

Translational control under heat stress

By contrast, heat stress disrupts ribosome biogenesis and rRNA processing, frequently leading to global translational repression (Shanmugam et al., 2021). The DEAD‐box RNA helicase TOGR1 (Thermotolerant Growth Required 1) safeguards ribosome integrity by stabilizing precursor rRNA under heat stress (Wang et al., 2016a).

tRNA modifications represent another critical determinant of translational thermotolerance. In rice, AET1 (Adaptation to Environmental Temperature 1), a tRNA^His guanylyltransferase, modifies tRNA^His and interacts with ribosome‐associated proteins RACK1A (Activated C‐Kinase 1 A) and eIF3h (Eukaryote Initiation Factor 3 h) to maintain decoding fidelity under heat stress (Chen et al., 2019). Similarly, SLG1 (a tRNA 2‐thiolation protein)‐mediated tRNA thiolation contributes to natural variation in thermotolerance between indica and japonica rice (Xu et al., 2020b), while the chloroplast‐localized serine–tRNA ligase TSCD11 (Temperature‐Sensitive Chlorophyll‐Deficient 11) preserves chloroplast function under heat stress (Fang et al., 2020). These findings highlight tRNA integrity as a key component of translational adaptation to heat.

Post‐translational regulation of temperature responses

Post‐translational modifications (PTMs) provide a rapid and highly dynamic layer of temperature regulation. The ICE1–CBF module exemplifies PTM‐mediated control of cold responses. In Arabidopsis, ICE1 is regulated by phosphorylation (via OST1, MPK3/6, and BIN2), ubiquitination (by HOS1 and PUB25/26), and SUMOylation (by SIZ1; for Sap and MIZ1), collectively fine‐tuning CBF expression (Dong et al., 2006; Miura et al., 2007; Ding et al., 2015; Li et al., 2017a; Ye et al., 2019; Mei et al., 2023). CBF protein stability is further modulated by OST1‐dependent phosphorylation of nascent polypeptide‐associated complex subunits BTF3/BTF3L, whereas CRPK1 promotes CBF destabilization via phosphorylation of 14–3–3 proteins (Liu et al., 2017; Ding et al., 2018b). Additionally, the redox protein Trx‐h2 converts inactive CBF oligomers into active monomers, enhancing COR gene expression (Lee et al., 2021a; Figure 3).

In maize, PTMs exert both positive and negative effects on cold tolerance. ZmMPK8 phosphorylates and stabilizes ZmbZIP68, a repressor of ZmDREB1 genes (Li et al., 2022), while simultaneously promoting degradation of ZmRR1, a positive regulator of cold tolerance (Zeng et al., 2021). Similarly, ZmCPK17 stabilizes the transcription factor COOL1, thereby attenuating cold tolerance in maize (Zeng et al., 2025a). These examples illustrate how PTMs dynamically reconfigure stress‐responsive transcriptional networks (Figure 3).

Protein quality control under heat stress

Protein homeostasis under heat stress is primarily maintained by the ubiquitin–proteasome system (UPS) and SUMOylation pathways. In rice, the thermotolerance QTL (Quantitative trait locus) TT1 (Thermo‐tolerance 1) encodes the α2 subunit of the 26S proteasome, facilitating clearance of heat‐denatured proteins (Yu et al., 2024). In Arabidopsis, the E3 ligase XBAT31 targets the transcriptional repressors HsfB2a and HsfB2b for degradation, thereby enhancing reproductive thermotolerance (Zhang et al., 2021c). Additional E3 ligases, including HIRP1 and HTAS, also contribute positively to heat tolerance in rice (Liu et al., 2016; Kim et al., 2019; Figure 3).

Substrate specificity is further refined through N‐terminal acetylation by the N‐terminal acetyltransferase A (NatA) complex. The NatA‐interacting protein SUF1 (STRESS‐RELATED UBIQUITIN‐ASSOCIATED DOMAIN PROTEIN FACTOR 1) promotes thermotolerance by associating with the NatA subunit NAA15 (Aksnes et al., 2016; Song et al., 2022). Moreover, heat‐induced accumulation of SUMO conjugates and overexpression of OsSIZ1 both enhance thermotolerance in rice (Li et al., 2013).

Collectively, translational control, tRNA modification, post‐translational regulation, and protein quality control form an integrated regulatory network that safeguards proteome stability under temperature extremes. By dynamically coordinating these molecular layers, plants achieve a finely balanced response that reconciles growth, development, and stress resilience.

Unfolded protein response (UPR)

Heat stress disrupts protein folding within ER, leading to the accumulation of misfolded proteins and activation of the UPR (Liu and Howell, 2016). The UPR serves as a central ER quality‐control system that restores proteostasis by coupling transcriptional reprogramming with enhanced protein folding and degradation capacity. In plants, this response is primarily governed by ER‐localized bZIP and NAC transcription factors (Sun et al., 2021).

In rice, the UPR signaling is organized into two major branches. The first is mediated by inositol‐requiring enzyme 1 (IRE1), which catalyzes the unconventional splicing of OsbZIP74 (OsbZIP50) mRNA (Lu et al., 2012). This splicing event removes the transmembrane domain, producing an active bZIP74 isoform that translocates into the nucleus (Hayashi et al., 2012; Lu et al., 2012). Under heat stress, nuclear bZIP74 directly induces OsNTL3 expression, which in turn upregulates bZIP74 expression and activates genes involved in ER protein folding and ROS detoxification. This positive feedback loop amplifies ER protection and enhances thermotolerance (Liu et al., 2020; Figure 3).

The second branch operates through proteolytic activation of bZIP60. Under non‐stressed conditions, bZIP60 is retained at the ER membrane via interactions with the chaperone BiP (luminal binding protein). ER stress triggers its release and translocation to the Golgi apparatus, where site‐2 protease (S2P) cleavage generates a nuclear‐localized active form, analogous to bZIP28 activation in Arabidopsis (Liu et al., 2007; Hayashi et al., 2013). Nuclear bZIP60 then activates a suite of target genes that restore ER proteostasis. Together, these two branches function synergistically to maintain ER function and confer thermotolerance in rice (Zhou et al., 2022b; Figure 3).

Recent studies highlight the agricultural significance of UPR regulation. The rice locus QT12 negatively regulates grain quality thermotolerance by overactivating the UPR, thereby disrupting endosperm storage metabolism and exacerbating heat‐induced quality defects (Li et al., 2025b). This finding highlights that optimal rather than maximal UPR activation is required to balance stress protection with yield and quality.

In summary, the UPR constitutes a key adaptive module that enables plants to sense and mitigate ER stress under heat stress. Its dual signaling arms and interconnected feedback regulation safeguard proteome stability and stress resilience, and yet, excessive or prolonged activation can compromise productivity. Therefore, precise modulation of UPR intensity emerges as a critical consideration for molecular design breeding aimed at achieving both heat resilience and high crop performance.

Epigenetic regulation

Epigenetic mechanisms, including DNA methylation, histone modification, non‐coding RNAs, and RNA modifications, play important roles in plant responses to temperature stress by dynamically linking environmental cues with transcriptional regulation and stress memory.

During cold acclimation, coordinated changes in DNA methylation and histone acetylation facilitate activation of cold‐responsive gene networks. In Arabidopsis, the RNA‐directed DNA methylation component RDM4 associates with RNA polymerase II at CBF2 and CBF3 promoters to enhance their transcription (Chan et al., 2016). The CUL4–HOS15 complex targets the histone deacetylase HD2C for degradation, leading to increased H3 hyperacetylation at COR loci and facilitating CBF access, thereby enhancing freezing tolerance (Park et al., 2018).

Chromatin remodeling also contributes to thermotolerance and post‐stress recovery. In rice, the Polycomb Repressive Complex 2 (PRC2) represses MADS82 and MADS87 through H3K27me3 deposition, contributing to temperature‐dependent developmental regulation (Dhatt et al., 2021). The PRC2 subunit FIE1 (Fertilization Independent Endosperm 1) further regulates grain filling via temperature‐dependent promoter demethylation (Cheng et al., 2020; Dhatt et al., 2021). During recovery from heat stress, the H3K4 methyltransferases SDG25 (Set Domain protein 25) and ATX1 (Arabidopsis homolog of Trithorax 1) dynamically adjust H3K4me3 levels and DNA methylation to re‐establish appropriate expression of heat‐responsive genes (Song et al., 2021). Moreover, HsfA2 integrates transcriptional and epigenetic memory by recruiting histone methyltransferases to deposit H3K4me3 at target loci, enabling sustained thermomemory (Lämke et al., 2016; Figure 3).

Non‐coding RNAs provide an additional epigenetic regulation. Under cold stress, long non‐coding RNAs (lncRNAs) modulate chromatin structure and transcription. The Arabidopsis locus, initially termed SVALKA recently redefined as CAS (CBF ANTISENSE TRANSCRIPT), promotes cold acclimation by forming R‐loops that reduce nucleosome occupancy at CBF1 and CBF3 loci, thereby facilitating transcriptional activation (Kindgren et al., 2018; Sun et al., 2025). Small RNAs are also involved in heat responses: miR156 suppresses SPL (SQUAMOSA PROMOTER BINDING PROTEIN‐LIKE) transcripts to activate heat shock genes and establish thermomemory (Stief et al., 2014), while miR398 enhances ROS signaling by inhibiting SOD translation (Guan et al., 2013b). In rice, miR444b.2 directly targets HsfA1 transcripts, forming a heat‐responsive regulatory module (Qiu et al., 2025; Figure 3).

Epitranscriptomic modifications further fine‐tune RNA fate under temperature stress. Cold stress reduces acetylation of the N⁶‐methyladenosine (m⁶A) reader OsECT3 (EVOLUTIONARILY CONSERVED C‐TERMINAL REGION 3), enhancing RNA‐binding activity and cold tolerance in rice (Ma et al., 2025). Conversely, the mRNA 5‐methylcytosine (m⁵C) methyltransferase NSUN2 (NOP2/SUN Domain Family Member 2) stabilizes transcripts associated with photosynthesis and detoxification under heat stress (Tang et al., 2020; Figure 3). These reversible RNA modifications provide rapid, flexible control that complements chromatin‐ and transcription‐based regulation.

In summary, epigenetic regulation provides a multilayered framework through which plants integrate temperature signals into gene expression programs. By coordinating chromatin remodeling, non‐coding RNA function, and RNA modification, epigenetic mechanisms enable rapid transcriptional reconfiguration and stress memory, thereby enhancing plant adaptation to fluctuating thermal environments.

Shared regulatory mechanisms and network interactions between low‐ and high‐temperature stress responses

Although cold and heat stress impose contrasting physiological challenges, increasing evidence indicates that their response pathways are interconnected through partially shared regulatory nodes and network architectures. Rather than operating as discrete modules, low‐ and high‐temperature signaling systems form a continuum that enables plants to dynamically cope with fluctuating thermal environments.

At the systems level, convergence is largely mediated by common signaling hubs and multifunctional regulators. Core components such as Ca²⁺ signaling modules, ROS‐producing and scavenging systems, and conserved transcription factor families (e.g., AP2/ERF, NAC, and NLP) participate in both cold‐ and heat‐responsive networks, but elicit distinct outputs depending on the cellular context, developmental stage, and stress intensity (Ding et al., 2020; Ding and Yang, 2022; Kan et al., 2023). This shared architecture allows temperature cues of opposite polarity to be processed through common molecular frameworks while maintaining response specificity.

Integration is further reinforced at higher‐order regulatory layers. Chromatin remodeling, histone modifications, and RNA‐based regulation provide temperature‐responsive plasticity that is not strictly stress‐specific, but instead modulates transcriptional competence, RNA processing, and translational efficiency across a broad thermal range. Such mechanisms are well suited for coordinating long‐term adaptation and stress memory under recurring or compound temperature stresses.

From an applied perspective, the existence of shared regulatory nodes provides opportunities to improve dual tolerance to both low and high temperatures. Network intersection points, rather than terminal, stress‐specific effectors, represent promising targets for molecular design breeding and genome editing. Precise tuning of these integrative regulators may enable broad‐spectrum thermal resilience while minimizing trade‐offs between cold and heat tolerance.

GENETIC AND ENVIRONMENTAL REGULATION OF THERMOMORPHOGENESIS

Beyond episodic temperature extremes, gradually increasing ambient temperatures pose persistent challenges to plant growth and ecosystem stability. To optimize fitness and reproductive success, plants continuously sense subtle thermal fluctuations and adjust their developmental programs accordingly. Under moderately elevated, non‐stressful temperatures, plants show coordinated morphological and physiological responses, including hypocotyl and petiole elongation, leaf hyponasty, and accelerated flowering, collectively termed thermomorphogenesis. These responses enhance evaporative cooling, optimize canopy architecture, and promote reproductive efficiency in warming environments.

At the core of thermomorphogenesis lie the transcription factors PHYTOCHROME INTERACTING FACTOR 4 (PIF4) and ELONGATED HYPOCOTYL 5 (HY5), which integrate thermal cues with light and circadian signals (Figure 4). PIF4, together with other PIF family members, predominantly drives shoot elongation, whereas HY5 plays a central role in root thermoresponses and whole‐plant coordination (Koini et al., 2009; Kumar et al., 2012; Lau et al., 2018; Fiorucci et al., 2020; Gaillochet et al., 2020; Lee et al., 2021b, 2021c; Burko et al., 2022). Regulation of PIF4 occurs at transcriptional, post‐transcriptional, and post‐translational levels, enabling precise integration of temperature, light quality, and circadian timing (Zhang et al., 2021b; Figure 4).

Figure 4.

Molecular regulation of thermomorphogenesis in Arabidopsis

Plants use coordinated regulatory networks, spanning environmental sensing, hormonal signaling, and cellular responses, to reshape their architecture under warm ambient temperatures. In shoots, thermosensors, including PHYB, ELF3, and PIF7, perceive elevated temperature and modulate the stability and activity of PIF4. Concurrently, light conditions, circadian rhythms, and endogenous factors fine‐tune PIF4 expression and function at both transcriptional and post‐transcriptional levels. As a central growth coordinator, PIF4 activates thermomorphogenic responses primarily through auxin and gibberellin (GA) pathways. Other hormonal signals, such as salicylic acid (SA), brassinosteroids (BR), and ethylene (ETH), also contribute to thermal adaptation. Systemically, the transcription factor HY5 translocates from shoots to roots, facilitating interorgan communication and enhancing root elongation. At the cellular level, membrane‐localized proteins such as PIN2, AHA1/2, and plasma membrane H⁺‐ATPases induce apoplastic acidification, promoting cell wall loosening and cell elongation required for thermomorphogenic growth. Arrows denote positive regulation and T‐bars denote negative regulation.

Plant hormones as central integrators of thermomorphogenesis

Phytohormones constitute the principal regulatory axis that integrates thermal cues into growth outputs. Auxin, brassinosteroids (BRs), gibberellins (GAs), ethylene (ET), jasmonates (JAs), and salicylic acid (SA) form an interconnected network that modulates thermoresponsive development (Lu et al., 2021; Figure 4).

Auxin acts as a primary driver of heat‐induced elongation. Elevated temperatures enhance auxin biosynthesis, transport, and signaling, largely through direct activation of auxin biosynthetic genes by PIF4 (Gray et al., 1998; Sang et al., 2023). Fine‐tuning of auxin sensitivity is mediated by regulatory modules such as miR156‐SPL9, ensuring appropriate growth amplitude (Sang et al., 2023). In roots, warm temperatures promote PIF4‐dependent PIN2 trafficking to the plasma membrane via SNX1 (SORTING NEXIN 1), increasing local auxin accumulation (Hanzawa et al., 2013; Kim et al., 2020b). Concurrently, CCaP proteins modulate H⁺‐ATPase activity to enhance cell wall extensibility during elongation (Wang et al., 2024a).

BR signaling synergizes with auxin through BZR1, which directly activates PIF4 expression. Together with ARF6 (AUXIN RESPONSE FACTOR 6), BZR1 and PIF4 form a cooperative transcriptional module that amplifies elongation growth (Bouré et al., 2019). This pathway is restrained by E3 ligases such as UPL3, which promotes BZR1 degradation under warm conditions (Zhu et al., 2025b). In parallel, the membrane‐localized kinase MAP4K4/TARGET OF TEMPERATURE 3 (TOT3) mediates BR‐dependent thermomorphogenesis independently of PIF4 and light, underscoring pathway diversification (Vu et al., 2021).

GAs promote thermoresponsive growth by relieving DELLA‐mediated repression (Stavang et al., 2009). TEOSINTE BRANCHED 1/CYCLOIDEA/PCF (TCP) transcription factors interact with PIF4 to activate GA biosynthetic genes, whereas DELLAs antagonize PIF4 activity via direct protein–protein interaction (de Lucas et al., 2008; Feng et al., 2008; Li et al., 2016; Ferrero et al., 2019; Zhou et al., 2019). Circadian gating of this interaction is achieved through GIGANTEA, which stabilizes DELLAs under long‐day conditions to prevent excessive elongation (Park et al., 2020).

In contrast, ethylene generally suppresses thermomorphogenesis. Stabilization of EIN3 through the degradation of EBFs (ETHYLENE‐INSENSITIVE 3‐BINDING F‐BOX PROTEINs) induces downstream repressors, reducing H⁺‐ATPase activity and cell elongation (Hao et al., 2021; Kim et al., 2021). Ethylene signaling also interfaces with chromatin remodeling, reinforcing transcriptional repression at thermoresponsive loci (Shao et al., 2024).

JA and SA provide additional modulatory layers. Elevated temperatures accelerate JA turnover, releasing growth repression by reducing JAZ protein abundance (Zhu et al., 2021b). Conversely, suppression of SA biosynthesis under warm conditions alleviates its inhibitory effect on thermomorphogenesis, in part by preventing ELF3 phase separation (Chen et al., 2025b).

Collectively, thermomorphogenesis emerges from a finely balanced hormonal network that converts modest temperature elevations into coordinated developmental plasticity.

Environmental regulation of thermomorphogenesis

Light acts both as an energy source and as a dominant environmental modulator of PIF4 activity. PIF3 directly activates PIF4 transcription under warm conditions, and the two factors cooperatively induce elongation‐related genes (Das et al., 2025). The de‐etiolated 1 (DET1)‐constitutive photomorphogenic 1 (COP1)‐HY5 module further shapes thermoresponses by promoting HY5 degradation and enhancing PIF4 expression and PIF4 stability, both depending on and independent of HY5 (Delker et al., 2014; Gangappa and Kumar, 2017). Blue‐light signaling antagonizes thermomorphogenesis through CRY1‐mediated inhibition of PIF4 activity, whereas TCP transcription factors (TCP5/13/17) promote PIF4 expression, with TCP17 being released from CRY1 repression under warm temperatures (Ma et al., 2016; Zhou et al., 2019). Belowground, HY5 predominates by directly regulating auxin‐ and BR‐related genes to drive root thermomorphogenesis (Lee et al., 2021b; Figure 4).

Thermomorphogenesis is tightly gated by the circadian clock. Morning‐phased regulators, including CIRCADIAN CLOCK ASSOCIATED 1 (CCA1) and SHORT HYPOCOTYL‐UNDER BLUE 1 (SHB1), sustain PIF4 expression and promote the initiation of thermoresponsive growth, whereas the Evening Complex, composed of EARLY FLOWERING 3 (ELF3), ELF4, and LUX ARRHYTHMO (LUX), represses PIF4 and PIF5 during early evening to prevent excessive elongation (Nusinow et al., 2011; Sun et al., 2019; Figure 4).

Warm temperatures attenuate EC‐mediated repression by weakening its DNA‐binding capacity and promoting ELF3 degradation through E3 ligases XBAT31/35 and COP1, and additional factors such as HSP90, thereby releasing PIF4 activity and enabling heat‐induced elongation (Yu et al., 2008; Zhang et al., 2021a, 2021c; Zeng et al., 2023). This process is further reinforced by multiple BBX proteins, which directly interact with ELF3 to dampen its repressive function under warm conditions (Ding et al., 2018a; Malakar et al., 2025). Consistent with its central regulatory role, natural variation in ELF3 alleles contributes to adaptive divergence in thermosensitivity and photoperiodic growth responses across environments (Raschke et al., 2015).

The circadian–temperature interface is further refined at the transcriptional level. CCA1 and REVEILLE (RVE) family members dynamically modulate EC activity by competitively binding to the evening element (EE) motif in the ELF4 promoter. RVE5 and CCA1 fine‐tune ELF4 expression under warm conditions, while RVE7 reinforces repression, collectively adjusting the amplitude and timing of thermomorphogenic growth (Tian et al., 2022; Li et al., 2023b). In parallel, the evening‐phased clock component TIMING OF CAB EXPRESSION 1 (TOC1) directly interacts with PIF4 to constrain its activity at night, ensuring that elongation responses remain temporally restricted (Zhu et al., 2016; Figure 4).

Thermomorphogenesis is also modulated by environmental context. Nutrient deficiency in nitrogen or phosphorus suppresses root thermomorphogenesis via HY5 accumulation, which inhibits the expression of the nitrate transporter NRT1.1 (Lee et al., 2024). Shade and warm synergistically promote elongation through PIF7‐dependent auxin signaling (Burko et al., 2022). Drought and soil moisture signals intersect with temperature pathways through SA accumulation and SnRK2–COP1–HY5 modules, respectively, ensuring that growth decisions reflect integrated environmental inputs (Hayes et al., 2025).

Collectively, environmental factors such as light, temperature, nutrients, and water availability form an integrated regulatory network that dynamically modulates thermomorphogenesis, ensuring coordinated and adaptive growth in fluctuating environments.

Epigenetic regulation of thermomorphogenesis

Epigenetic mechanisms add a crucial regulatory layer of spatial and temporal control over thermomorphogenesis. Histone methylation, acetylation, chromatin remodeling, and histone variant dynamics collectively fine‐tune transcriptional outputs in response to ambient temperature.

Histone methylation plays a pivotal role in modulating gene activity during thermal responses. Demethylases such as JMJ14/15/17/18 restrict thermoresponsive gene activation, with JMJ17/18 recruited by ELF3 to repress PIF4 and PIF5 (Tian et al., 2025). In contrast, the transcriptional co‐regulator SEUSS (SEU) promotes H3K4me3 deposition at auxin biosynthetic genes, facilitating elongation (Huai et al., 2018). This activation is reinforced by the INO80 complex, which cooperates with PIF4 to evict the repressive histone variant H2A.Z and enhance transcriptional competence (Xue et al., 2021).

Additional histone marks also contribute to temperature‐dependent regulation. H3K36me3 influences temperature‐sensitive AS and flowering, while chromatin reader MRG2 links this mark to FT activation (Xu et al., 2014; Pajoro et al., 2017). Removal of repressive H3K27me3 by JMJ30 and REF6/JMJ12 activates thermoresponsive and flowering‐related genes, whereas the chromatin remodeler PICKLE (PKL) connects circadian regulation to chromatin remodeling at auxin‐responsive loci (Zha et al., 2017). Histone acetylation provides an additional regulatory dimension. MRG1 and MRG2 promote activating H4K5ac at PIF4 targets (Zhou et al., 2024b), whereas histone deacetylases (HDACs) such as HDA9 and HDA15 restrain elongation by counteracting acetylation and interacting with negative regulators like HFR1 (van der Woude et al., 2019; Xiang et al., 2019).

In summary, these epigenetic mechanisms coordinate with transcriptional and hormonal networks to translate modest temperature increases into adaptive developmental programs.

GENETIC REGULATION OF TGMS

The application of male sterility is a cornerstone of hybrid crop breeding. Male sterility systems are generally classified into cytoplasmic male sterility (CMS) and genic male sterility (GMS). While the classical three‐line hybrid breeding system relies on CMS, maintainer, and restorer lines, the two‐line system relies on photoperiod‐ or temperature‐sensitive genic male sterility (P/TGMS). In these systems, plants are male‐sterile under restrictive environmental conditions but regain fertility under permissive environments, greatly simplifying hybrid seed production.

Thermosensitive genic male sterility (TGMS) represents a prominent two‐line strategy, in which fertility is reversibly switched by ambient temperature. Each TGMS line is defined by a critical sterility‐inducing temperature (CSIT), which determines the threshold for the fertility–sterility transition (Peng et al., 2023).

Molecular basis and the TMS5‐centered pathway

At the molecular level, TGMS arises from temperature‐dependent perturbation of essential cellular processes during pollen development, thereby unmasking latent defects caused by TGMS‐associated mutations (Wang et al., 2025). To date, nine TGMS genes have been characterized in rice, including UGP1, TMS5, OsMS1/TMS9‐1, TMS10, TMS12‐1, OsTMS15, OsTMS16, OsTMS18, and OsTMS19, which collectively define multiple, partially interconnected pathways governing reproductive thermosensitivity (Wang et al., 2025).

Among these, TMS5 represents the best‐characterized and central regulatory node. TMS5 encodes an RNase Z enzyme essential for RNA metabolism during pollen development. Loss of TMS5 function leads to defective processing of UbL40 mRNAs, resulting in pollen abortion and temperature‐dependent male sterility (Zhou et al., 2014). More recent work revealed an additional, mechanistically unifying role of TMS5 as a tRNA 2′,3′‐cyclic phosphatase. Mutations in TMS5 cause the accumulation of aberrant tRNAs lacking 3′CCA tails and cyclic phosphate ends, particularly tRNA‐Ala, with defects that are markedly exacerbated at high temperatures (Yan et al., 2024). TMS5 expression is directly activated by the transcription factor OsbHLH138, ensuring proper UBL40 mRNA processing and fertility maintenance under permissive temperatures (Wen et al., 2019).

Protein quality control has emerged as a critical modifier of the CSIT in tms5‐based TGMS lines. Ribosome‐associated quality control (RQC) components, including the E3 ubiquitin ligases, CSIT1 and CSIT2, facilitate degradation of misfolded nascent peptides. Mutations in these factors elevate the CSIT, rendering plants sterile only at higher temperatures (Peng et al., 2023, 2024a). Additional RQC regulators, such as OsHel2, OsRqc2, and the OsRqc1–OsVms1 module, modulate ribosome stalling, tRNA homeostasis, and translational fidelity, thereby fine‐tuning the temperature sensitivity of male fertility (Zhou et al., 2025a; Liu et al., 2025b). Together, these findings define a TMS5–RQC regulatory axis that links RNA metabolism, tRNA integrity, and translational surveillance to temperature‐dependent fertility control.

TGMS pathways beyond TMS5

TGMS can also arise from disruption of additional cellular processes essential for pollen development. Ugp1, encoding UDP‐glucose pyrophosphorylase, is required for callose deposition during meiosis of pollen mother cells; its downregulation leads to abnormal transcript accumulation and TGMS (Chen et al., 2007). OsMS1, a histone‐binding protein crucial for tapetal development, harbors a natural thermosensitive allele (OsMS1 ^wenmin1 ) in which elevated temperature compromises nuclear localization and interaction with the tapetal regulator TDR, leading to male sterility (Wu et al., 2022a). TMS10, encoding an RLK, ensures proper tapetal degeneration and pollen viability under high temperatures (Yu et al., 2017). TMS12‐1 encodes a noncoding RNA that produces a 21‐nucleotide small RNA; a single‐nucleotide mutation disrupting its processing leads to male sterility (Zhou et al., 2012). OsTMS18, a GMC oxidoreductase required for pollen wall formation, harbors a heat‐sensitive amino acid substitution and is transcriptionally regulated by the tapetal transcription factor OsMS188 (Zhang et al., 2022b). OsTMS19, encoding a pentatricopeptide repeat protein, affects mitochondrial function and ROS homeostasis, with excessive ROS accumulation under high temperature leading to pollen abortion (Zhou et al., 2024).

Collectively, TGMS reflects a temperature‐dependent breakdown of multiple, pollen‐essential cellular processes, including RNA processing, protein quality control, carbohydrate metabolism, and redox balance. The integration and buffering capacity of these pathways determine the CSIT and thus the fertility–sterility switch. Understanding these molecular frameworks not only deepens insight into the genetic–environmental interplay governing reproductive thermosensitivity but also provides strategic targets for engineering environmentally stable, high‐performance two‐line hybrid systems in crops.

VERNALIZATION

The concept of vernalization originated from early agricultural observations in the 19th century and was formalized in the early 20th century through studies showing that prolonged cold exposure accelerates flowering in winter cereals such as wheat and rye (Chouard, 1960). Subsequent work across diverse plant species established vernalization as a conserved developmental strategy that synchronizes flowering with seasonal temperature cycles. Vernalization refers to the requirement for prolonged low‐temperature exposure to confer flowering competence, allowing plants to overwinter in a vegetative state and flower under favorable spring conditions (Amasino, 2010; Zhang et al., 2019b).

Vernalization is widespread among both monocots and dicots, and proceeds through three coordinated phases: initiation, memory, and resetting, in which plants perceive and record the duration of cold exposure, maintain this “vernalized” state through cell divisions, and then reset it in the next generation (Tao et al., 2017; Figure 5).

Figure 5.

Vernalization‐induced flowering in wheat and Arabidopsis

Prolonged cold exposure promotes flowering by activating floral inducers and repressing floral inhibitors, leading to stable induction of the florigen FT. In wheat, vernalization regulates VRN1, VRN2, and VRN3 (FT1) in a coordinated manner. Cold triggers strong VRN1 expression via epigenetic regulation, which suppresses the floral repressor VRN2 and permits VRN3 activation in leaves and meristems to promote flowering. Yin–Yang modifications, phosphorylation (P) and O‐GlcNAcylation (G), fine‐tune the activity of vernalization regulators to accelerate floral transition. Dynamic histone marks (H3K4me3 and H3K27me3) and the lncRNA VAS promote chromatin opening at VRN1, facilitating its transcriptional activation. In Arabidopsis, vernalization gradually induces VIN3 expression, which recruits the FRI complex, PRC2 components, and cofactors to stably repress the floral repressor FLC. The long noncoding RNAs (COLDWRAP, COLDAIR, and COOLAIR) guide chromatin remodeling, mediating a switch from active (H3K4me3, H3/H4 acetylation) to repressive (H3K27me3) chromatin states at FLC. Upon return to warm conditions, stable silencing allows FT and SOC1 activation, promoting flowering. Arrows denote positive regulation and T‐bars denote negative regulation.

Genetic regulation of vernalization in cereals

In temperate cereals, vernalization requirement is a key determinant of flowering time and geographic adaptation. In wheat (Triticum aestivum), natural variation in this trait is largely governed by allelic diversity at vernalization (VRN) loci, VRN1, VRN2, and VRN3, which together form an interdependent regulatory circuit modulating flowering competence (Sharma et al., 2020). Despite sharing nomenclature with Arabidopsis vernalization genes, the cereal VRN system comprises distinct molecular components, reflecting independent evolutionary solutions to cold‐induced flowering in monocots and dicots (Figure 5).

VRN1 encodes a MADS‐box transcription factor that functions as the primary floral promoter in cereals (Yan et al., 2003). Prolonged cold exposure robustly induces VRN1 expression in leaves and the shoot apical meristem, and this elevated expression is stably maintained after the return to warm conditions (Fu et al., 2005; Trevaskis, 2010). In hexaploid wheat, dominant alleles at the three homoeologous VRN1 loci (VRN‐A1, VRN‐B1, and VRN‐D1) confer a spring growth habit through constitutive or enhanced expression, generating extensive phenotypic diversity that has been widely exploited in breeding (Mizuno et al., 2022).

By contrast, VRN2 encodes a zinc finger‐CCT domain protein that acts as a floral repressor by inhibiting VRN3, the cereal ortholog of FLOWERING LOCUS T (FT) (Distelfeld et al., 2009). Vernalization progressively suppresses VRN2 expression, releasing repression of VRN3 and enabling flowering. Accordingly, loss‐of‐function mutations in VRN2 or constitutive expression of VRN1 bypass the vernalization requirement and convert winter genotypes into spring types (Yan et al., 2004; Chen and Dubcovsky, 2012).

VRN3 encodes the mobile florigen produced in leaves and transported to the shoot apex to initiate reproductive development (Yan et al., 2006). Its expression is repressed by VRN2 and promoted by VRN1, establishing a self‐reinforcing VRN1–VRN3 positive feedback loop following vernalization (Li and Dubcovsky, 2008; Chen and Dubcovsky, 2012; Oliver et al., 2013; Deng et al., 2015). At the meristem, VRN3 associates with FLOWERING LOCUS D‐like 2 (FDL2) and 14–3–3C to activate floral identity genes and irreversibly commit the apex to flowering (Li and Dubcovsky, 2008; Li et al., 2015a). A broadly similar regulatory framework operates in barley, underscoring the conservation of vernalization mechanisms across cereals (Greenup et al., 2010; Deng et al., 2015).

In summary, cereal vernalization relies on a genetically compact but highly effective VRN1–VRN2–VRN3 regulatory module that converts prolonged winter cold into stable flowering competence. Natural variation within this circuit provides the molecular basis for adaptation across diverse agroecological zones and represents a central target for crop improvement.

Genetic regulation of vernalization in Arabidopsis

In dicots, Arabidopsis thaliana has served as the principal model for elucidating the molecular basis of vernalization. At the core of this pathway lies FLOWERING LOCUS C (FLC), a MADS‐box transcription factor whose expression is promoted by FRIGIDA (FRI) and whose repression by prolonged cold enables flowering (Ream et al., 2012; Figure 5).

FLC acts as a potent floral repressor by inhibiting the floral integrators FT and SOC1, either directly or via heterodimerization with SHORT VEGETATIVE PHASE (SVP) (Sheldon et al., 2000; Richter et al., 2019). Importantly, FLC expression is quantitatively preset during early embryogenesis through antagonistic co‐transcriptional regulation by FRI and the RNA‐processing factor FCA. This presetting establishes transcriptional thresholds that are maintained in seedlings through antisense transcription and a CPF‐like‐Polycomb module, linking RNA processing to the establishment of a permissive chromatin state prior to vernalization (Schon et al., 2021; Mateo‐Bonmatí et al., 2024; Menon et al., 2024).

Prolonged cold exposure initiates a stepwise repression program at the FLC locus. During the early phase of vernalization, transcription initiation is attenuated in a manner largely decoupled from RNA polymerase II elongation and splicing (Maple et al., 2026). This phase is mediated by cold‐induced lncRNAs, COLD ASSISTED INTRONIC NONCODING RNA (COLDAIR), COLD OF WINDTER‐INDUCED NONCODING RNA FROM THE PROMOTER (COLDWRAP), and COOLAIR, which act at distinct regions of the FLC locus to dampen transcription (Csorba et al., 2014; Kim and Sung, 2017; Kim et al., 2017). Several transcription factors, including NTM1‐LIKE (NTL8), CBFs, and WRKY63, promote COOLAIR expression, reinforcing early cold‐dependent repression (Zhao et al., 2020; Hung et al., 2022; Jeon et al., 2023).

As cold exposure persists, FLC repression transitions from a transcriptional to a chromatin‐based mode. VRN‐class regulators, including VRN1 and VERNALIZATION INSENSITIVE3 (VIN3)–VRN5 complexes, mediate the nucleation of PRC2 at FLC, establishing H3K27me3 marks that underpin stable epigenetic silencing (Gendall et al., 2001; Levy et al., 2002; Jean Finnegan et al., 2011; Antoniou‐Kourounioti et al., 2018; Zhao et al., 2020; Fiedler et al., 2022; Franco‐Echevarria et al., 2023; Schulten et al., 2025). This silenced state is subsequently reinforced and maintained by chromatin regulators, such as LHP1/TERMINAL FLOWER 2 (TFL2) and the methyltransferase CURLY LEAF (CLF), ensuring maintenance of flowering competence throughout vegetative growth (Sung et al., 2006; Turck et al., 2007; Doyle and Amasino, 2009).

Upstream of FLC, FRI activates transcription through a dedicated multiprotein complex comprising four FLC‐specific regulators: FRI‐LIKE1 (FRL1), FLC EXPRESSOR (FLX), SUPPRESSOR OF FRI4 (SUF4), and FRE ESSENTIAL 1 (FES1). Loss of individual components reduces FLC expression and accelerates flowering (Choi et al., 2011; Ream et al., 2012; Zhang and Jimenez‐Gomez, 2020). Notably, cold exposure promotes FRI nuclear condensate formation, which correlates with reduced FRI occupancy at the FLC promoter and contributes to temperature‐dependent modulation of FLC repression (Zhu et al., 2021a).

Together, vernalization in Arabidopsis is orchestrated by a multilayered regulatory network centered on FLC, integrating transcriptional presetting, non‐coding RNA‐mediated repression, Polycomb‐dependent chromatin silencing, and temperature‐sensitive nuclear organization. This system converts prolonged winter cold into stable epigenetic memory, ensuring precise seasonal control of flowering.

The perception of vernalization

Vernalization is initiated by the perception of prolonged cold, a process fundamentally distinct from the transient responses to short‐term temperature fluctuations. Earlier physiological analyses and in situ hybridization studies demonstrated that extended cold exposure selectively induces vernalization‐induced expression of genes such as VER2 in immature leaves, identifying young leaf tissue as a primary site of cold perception and signal initiation (Yong et al., 2003; Xiao et al., 2022). These findings support a model that vernalization begins locally in developing leaves and subsequently engages systemic signaling to enable the floral transition.

A central question in vernalization biology is how plants quantitatively measure cold duration and convert it into a stable molecular memory. Accumulating evidence, particularly from cereals, indicates that PTM‐based signaling provides a key mechanism for temporal integration of cold exposure. In wheat, vernalization induces a dynamic interplay between phosphorylation and O‐GlcNAcylation of the glycine‐rich RNA‐binding protein GRP2. These modifications alter GRP2 interaction with the RIP‐3 region of VRN‐A1, thereby modulating VRN1 activation and flowering initiation (Kippes et al., 2018; Xu et al., 2019). Consistently, the O‐GlcNAc transferase TaOGT1 accelerates flowering by promoting GRP2 O‐GlcNAcylation, reinforcing the functional relevance of this modification (Fan et al., 2021).

Additional studies support a broader role for O‐GlcNAc‐based signaling in vernalization sensing. The lectin‐like protein VER2 undergoes cold‐induced phosphorylation and is associated with a global increase in protein O‐GlcNAcylation during vernalization, and its overexpression accelerates flowering (Yong et al., 2003; Xing et al., 2009). More recently, vernalization was shown to reduce the phosphorylation while increasing O‐GlcNAcylation of fructose‐1,6‐bisphosphate aldolase HtL1/FBA10, enhancing its enzymatic activity and promoting VRN1 transcription through increased histone acetylation at the VRN1 locus (Zhang et al., 2025a). These studies suggest that PTMs act as molecular integrators of cold duration, enabling plants to translate prolonged cold exposure into graded transcriptional activation of flowering genes (Figure 5).

In Arabidopsis, vernalization perception relies on a distinct but conceptually analogous thermosensory module centered on VIN3 (Zhao et al., 2020). VIN3 expression increases gradually during extended cold exposure and rapidly declines upon rewarming, closely mirroring the duration of winter cold (Jean Finnegan et al., 2011). This progressive induction is governed by the cold‐dependent accumulation of the transcription factor NTL8, whose levels increase under low temperatures due to reduced growth‐mediated protein dilution. NTL8 directly activates VIN3, which in turn promotes Polycomb‐mediated silencing of FLC, thereby coupling cold duration sensing to epigenetic memory formation (Sung and Amasino, 2004; Zhao et al., 2020).

Collectively, vernalization sensing is a multi‐layered process involving tissue‐level cold perception, dynamic post‐translational signaling, and growth‐coupled transcriptional control to ensure accurate timing of flowering after winter.

Epigenetic mechanism underlying vernalization memory

Epigenetic regulation during vernalization is fundamentally distinct from classical stress responses, as it establishes a chromatin‐based memory that allows plants to record prolonged winter cold and activate flowering only after favorable conditions return. This memory is encoded through stable, heritable changes in chromatin state at key floral regulators, thereby converting environmental history into long‐term transcriptional competence.

Vernalization memory in cereals

In temperate cereals such as wheat and barley, vernalization memory is primarily established through cold‐induced epigenetic activation of VRN1. Prolonged cold leads to the progressive removal of the repressive H3K27me3 and enrichment of activating modifications, including H3K4me3 and H3K36me3, resulting in sustained VRN1 expression after plants return to warm conditions (Oliver et al., 2009; Diallo et al., 2012; Kosová et al., 2025). Both the promoter and the first intron of VRN1 serve as key cis‐regulatory regions for this epigenetic switch, integrating cold duration into stable transcriptional output (Yan et al., 2004; Muterko et al., 2016). In addition, the cold‐induced lncRNA VAS enhances VRN1 transcription, illustrating how RNA‐mediated regulation cooperates with chromatin remodeling to reinforce vernalization memory in cereals (Xu et al., 2021; Figure 5).

In Brachypodium distachyon, the overall framework of vernalization memory is broadly conserved, with VRN1 and FT acting as floral activators and VRN3 serving as an additional epigenetic node that integrates dynamic changes in H3K4me3 and H3K27me3. However, the absence of VRN1‐mediated repression of VRN2 highlights lineage‐specific rewiring of the monocot vernalization circuitry (Ream et al., 2012; Huan et al., 2013, 2018; Bettgenhaeuser et al., 2017).

Vernalization memory in Arabidopsis

In Arabidopsis, FLC silencing constitutes the canonical model of vernalization‐induced epigenetic memory. Prior to cold exposure, the FRI complex maintains an active chromatin state at FLC, characterized by H3K4me3, H3K36me3, and histone acetylation (Jiang et al., 2009, 2011). Prolonged winter cold recruits PRC2 to deposit H3K27me3, establishing stable repression that persists after rewarming, with CK2‐dependent PRC2 phosphorylation enhancing silencing (De Lucia et al., 2008; Yang et al., 2014; Zeng et al., 2025b; Figure 5).

Importantly, FLC repression proceeds through a multistep silencing trajectory that integrates transcriptional attenuation and chromatin‐based memory. Even under warm conditions, co‐transcriptional RNA processing establishes a low‐FLC pre‐state that primes PRC2 recruitment and Polycomb memory (Mateo‐Bonmatí et al., 2024; Maple et al., 2026). Prolonged cold further reinforces repression through the action of multiple lncRNAs, including COOLAIR and COLDWRAP, which modulate transcriptional dynamics and chromatin organization at the FLC locus (Zhao et al., 2021a). Cold also alters co‐transcriptional splicing and RNA Polymerase II activity via an SMU1 (SUPPRESSORS OF MEC‐8 AND UNC‐52 1)‐dependent mechanism, promoting the production of two distinct COOLAIR isoforms that together repress FLC transcription (Long et al., 2024). These lncRNA dynamics occur alongside cold‐induced phase separation of the FRI complex and reduced transcriptional activity, buffering transient temperature fluctuations (Kim and Sung, 2017; Nielsen et al., 2024). Concurrently, extended cold exposure induces biophysical and structural chromatin changes, including reduced nucleosome dynamics, repositioning of the +1 nucleosome, and deposition of H2Aub and H3K27me3, thereby establishing a Polycomb‐competent chromatin environment (Montez et al., 2025).

In summary, vernalization establishes a stable epigenetic memory centered on VRN1 in cereals and FLC repression in Arabidopsis. Despite clear lineage‐specific implementations, conserved principles emerge, including the use of histone modifications, Polycomb‐mediated chromatin regulation, and chromatin‐associated non‐coding RNAs, to translate prolonged cold exposure into durable transcriptional states.

Resetting of the vernalization requirement

The epigenetic memory established during vernalization must be erased or reset in the offspring to prevent transgenerational inheritance of the vernalized state. This resetting process is essential to re‐establish the vernalization requirement in the next generation, ensuring that plants remain responsive to seasonal cues.

In cereals, resetting the vernalization memory is closely associated with embryogenesis and early seed development. After fertilization, VRN1 is re‐silenced through re‐establishment of the repressive histone mark H3K27me3, whereas VRN2 remains transcriptionally inactive until light‐dependent germination (Xiang et al., 2019; Xiao et al., 2022; Niu et al., 2024). This coordinated reprogramming restores the pre‐vernalized chromatin state and reinstates the requirement for prolonged cold in the next generation. A key component of this process is the H3K27me3 reader RVR, which recognizes repressive chromatin at the VRN1 locus and facilitates stable transcriptional silencing during embryogenesis, thereby resetting flowering competence (Niu et al., 2024). Despite these advances, the upstream signals and chromatin‐remodeling machineries that initiate VRN1 re‐silencing in cereal embryos remain poorly defined, representing an important gap in our understanding of seasonal epigenetic resetting.

In Arabidopsis, resetting of the vernalized state occurs through active chromatin reprogramming during early embryogenesis. The seed‐specific transcription factor LEAFY COTYLEDON 1 (LEC1) plays a central role by promoting reactivation of FLC and establishing an active chromatin configuration. LEC1 acts in concert with B3‐domain transcription factors LEC2, FUSCA3 (FUS3), and ABSCISIC ACID‐INSENSITIVE 3 (ABI3), which collectively displace transcriptional repressors, including VAL1, VAL2, and Polycomb group proteins (Tao et al., 2017, 2019; Xu et al., 2022). This process involves the removal of repressive histone marks and restoration of FLC transcription inherited from the gametes. Reactivated FLC persists throughout embryogenesis and early seed development, reinstating the vernalization requirement for the next generation.

Together, these findings illustrate how embryonic chromatin reprogramming acts as a developmental reset system that erases vernalization memory, ensuring that flowering is repressed until the plant experiences a new cycle of winter cold.

GENETIC AND MOLECULAR BASIS OF TEMPERATURE ADAPTATION

Temperature adaptability fundamentally shapes plant geographic distribution, cultivation range, and yield stability. During evolutionary and domestication, crops have developed diverse molecular strategies to perceive, tolerate, and adapt to temperature fluctuations. Recent advances in rice, maize, wheat, and model species reveal that temperature adaptation is governed by the interplay of natural allelic variation, epigenetic regulation, and introgression from wild relatives, providing a mechanistic foundation for molecular design breeding aimed at climate resilience.

Genetic and molecular basis of cold adaptation

In rice, extensive progress has been made in elucidating the genetic architecture of cold adaptation across developmental stages. At the seedling stage, COLD1 confers cold tolerance, with the COLD1 ^SNP2 allele forming the genetic basis of japonica rice adaptation to high‐latitude regions (Ma et al., 2015). Similarly, the plasma membrane‐localized COLD6–OSM1 complex promotes cold tolerance by activating 2′,3′‐cAMP signaling; the COLD6 ^jap variant, derived from rare alleles in South Asia, was selectively enriched in temperate japonica varieties (Luo et al., 2024). Another domestication‐associated polymorphism, bZIP73^Jap, interacts with bZIP71 to modulate ABA levels and ROS homeostasis, facilitating japonica establishment in cool climates (Liu et al., 2018a).

Epigenetic mechanisms also contribute to cold adaptation. Natural DNA hypomethylation at the Acquired Cold Tolerance 1 (ACT1) locus promotes both the acquisition and stable inheritance of cold tolerance, with methylation states correlating with geographic distribution and climatic origin (Song et al., 2025b). These findings highlight the combined roles of genetic and epigenetic variation in the latitudinal expansion of japonica rice.

Cold tolerance at the reproductive stage is controlled by a distinct set of loci critical for yield stability. CTB4a, CTB2 ^SNP(A), and CTB5 encode key regulators of booting‐stage cold tolerance, with adaptive alleles arising from domestication or wild introgression and becoming enriched in high‐altitude and high‐latitude regions (Zhang et al., 2017b; Li et al., 2021; Guo et al., 2025a). These examples illustrate the stage‐specific genetic basis of reproductive stability under low temperatures.

In contrast, maize, originating from tropical regions, has acquired cold adaptability primarily through selection on regulatory variation. The transcription factor COOL1 acts as a negative regulator of maize cold tolerance, and promoter haplotypes affecting HY5 binding generate expression differences. The cold‐tolerant allele COOL1 ^HapA predominates in northern landraces, illustrating how cis‐regulatory evolution facilitated maize expansion into temperate zones (Zeng et al., 2025a).

Genetic and molecular basis of heat adaptation

Thermal variables also exert strong selective pressure on crop genomes. In soybean, environmental temperature explains a major proportion of genomic diversity among landraces (Bandillo et al., 2017). In grapevine, TTC4 (thermotolerance on chromosome 4), encoding a WRKY transcription factor, acts as a key determinant of thermotolerance. Natural intronic variant modulates SPL13 binding and TTC4 expression, leading to genotype‐dependent differences in heat tolerance (Chen et al., 2025a).

In cereals, naturally occurring allelic variants similarly contribute to heat adaptation. In Indian dwarf wheat, the TaSG‐D1(E286K) allele improves thermotolerance by stabilizing the heat‐responsive regulator TaPIF4 through phosphorylation (Cao et al., 2024). In rice, allelic variation in SRL10 (SEMI‐ROLLED LEAF 10), a double‐stranded RNA‐binding protein involved in miRNA biogenesis, modulates transcript abundance and protein stability, conferring differential heat tolerance among cultivars (Wang et al., 2023a). Likewise, SLG1 (Slender Guy 1), a tRNA 2‐thiolation protein, shows promoter and coding sequences' divergence between indica and japonica, with higher expression and thiolation activity conferring enhanced thermotolerance in indica rice (Xu et al., 2020b). These studies reveal how standing natural variations form a rich genetic reservoir for breeding heat‐resilient cultivars.

Thermal regulation of flowering and vernalization adaptation

Temperature adaptation also critically shapes reproductive timing. In temperate cereals, vernalization ensures flowering synchrony with seasonal cycles. VRN1 serves as the central determinant of vernalization requirement, with allelic combinations of VRN1 and VRN3 distinguishing winter and spring growth habits (Fu et al., 2005; Kippes et al., 2018; Xu and Chong, 2018; Afshari‐Behbahanizadeh et al., 2024). Copy number variation at VRN‐A1 further fine‐tunes vernalization duration, enabling wheat adaptation across diverse agroclimatic zones (Díaz et al., 2012; Shcherban and Salina, 2017; Dixon et al., 2019). In Arabidopsis, natural variation at FRI and FLC similarly underlies flowering‐time diversity across global populations (Li et al., 2014; Zhang and Jimenez‐Gomez, 2020). Together, these examples demonstrate that thermal regulation of flowering integrates genetic variation, epigenetic memory, and environmental sensing to optimize life‐history strategies under diverse climates.

MECHANISMS OF PLANT RESPONSES TO COMBINED STRESS

The inhibitory effects of combined stress on plant growth and development are not just a simple additive result of individual stresses. Instead, they arise from the integration of multiple stress signals into interconnected regulatory networks that reprogram physiological, metabolic, and molecular responses. Plants achieve tolerance to combined stress through coordinated modulation of a limited set of core processes, among which ABA signaling, ROS dynamics, and osmotic adjustment represent central hubs for signal convergence and crosstalk.

ABA functions as a master regulator of plant responses to diverse abiotic stresses and plays a pivotal role in coordinating combined‐stress responses. Stress‐induced ABA accumulation activates extensive transcriptional programs involving LEA (late embryogenesis abundant) proteins, DREB/CBF factors, ABFs (ABA‐responsive element binding factors), MYB proteins, and associated regulators (Shinozaki and Yamaguchi‐Shinozaki, 2007). Beyond transcriptional control, ABA regulates stomatal behavior through guard cell ion fluxes and mitochondrial H₂O₂ production, thereby linking hormonal signaling with redox regulation (Chen et al., 2020; Postiglione and Muday, 2023).

ROS play a dual role in combined stress responses, acting both as damaging agents and as essential signaling molecules. Excessive ROS accumulation under stress can impair proteins, lipids, and nucleic acids, necessitating tight control of ROS production and scavenging (Wang et al., 2024b). At the same time, stress‐induced ROS bursts—often mediated by NADPH oxidases such as RBOHs—function as signals that activate downstream defense pathways. For example, RBOHD‐dependent ROS signaling enhances heat tolerance by inducing heat‐responsive genes in Arabidopsis (Mittler et al., 2022; Li et al., 2025a). ABA‐mediated regulation of ROS homeostasis therefore serves as a key integrative mechanism that enables plants to balance oxidative signaling and cellular protection under combined stress conditions.

At the transcriptional level, plants use multifunctional transcription factor families that integrate signals from distinct stress pathways. NAC transcription factors represent prominent examples, with many members participating in responses to drought, salinity, temperature extremes, oxidative stress, and nutrient limitation (Xiong et al., 2025). In rice, OsNAC023 acts as a mobile stress integrator that translocates from the cytoplasm to the nucleus under drought or heat stress, activating genes involved in water transport, ROS homeostasis, and alternative splicing, thereby enhancing tolerance to combined stress conditions (Chang et al., 2024).

Heat shock transcription factors (HSFs) also contribute broadly to combined‐stress adaptation beyond their canonical role in heat responses (Feng and Xia, 2025). HSFA1b functions as a thermosensory regulator that modulates ABA signaling by suppressing OST1‐mediated stomatal closure through its adenylate cyclase activity (Zhang et al., 2025f). Similarly, HvHSFA2e enhances drought and heat tolerance by coordinating heat‐responsive gene expression with ABA and flavonoid biosynthesis pathways (Mishra et al., 2024). These findings underscore the role of HSFs as nodes of convergence between temperature, hormonal, and metabolic signaling.

Chromatin remodeling further refines transcriptional integration under combined stress. During early salt or heat stress, the light signaling regulator HY5 dissociates from the histone deacetylase HDA9, alleviating repression at the HsfA2 promoter and enhancing tolerance to both salt stress and heat shock (Yang et al., 2023b). Such dynamic chromatin–transcription factor interactions provide an additional regulatory layer that enables flexible and rapid transcriptional reprogramming in complex stress environments.

Beyond transcription factors, several protein kinases and signaling components function as hubs conferring tolerance to multiple stresses. In rice, OsMAPK3 enhances tolerance to cold, heat, and salt stresses, highlighting its role as a central signaling node (Zhang et al., 2017a, 2018; Deng et al., 2025). Similarly, VuDREB2A improves cowpea productivity under combined drought and high‐temperature conditions (Kumar et al., 2022), while ZmSnRK2.10 positively regulates both salt and drought tolerance in maize (Jiang et al., 2025; Zhang et al., 2025c). These examples illustrate that targeting shared regulatory hubs can yield broad‐spectrum stress resilience.

Future research should focus on identifying core nodes of signal integration under combined stress, elucidating regulatory hierarchies among hormonal, redox, transcriptional, and chromatin‐based pathways, and defining how these layers interact dynamically over time. Particular attention should be paid to the synergistic actions between transcription factors, protein kinases, and chromatin remodelers. Leveraging gene‐editing and molecular design breeding approaches to fine‐tune such key nodes, rather than single stress‐specific genes, offers a promising strategy for developing crop varieties with durable tolerance to increasingly complex environmental stresses under climate change.

MOLECULAR DESIGN BREEDING STRATEGIES FOR CROP TEMPERATURE RESILIENCE

Advances in molecular breeding have enabled the systematic identification and deployment of temperature‐resilience genes with high breeding value. Marker‐assisted backcrossing and allele pyramiding have successfully generated novel germplasms with enhanced thermotolerance and yield stability. In rice, pyramiding favorable alleles such as DTH2, ghd7, ltt1, or ipa1‐2D synergistically enhances cold tolerance while maintaining yield potential (Guo et al., 2025b). Likewise, superior haplotypes of cold tolerance‑related genes (CTB4a, Ctb1, bZIP73, and OsAPX1) show additive effects in multi‑gene pyramiding backgrounds (Guo et al., 2020). Notably, these elite alleles are enriched in high‑latitude or high‑altitude germplasm, providing a clear roadmap for cold‑tolerance improvement.

Wild relatives represent an invaluable, yet underexploited reservoir of thermal adaptability. Introgression of TT1 alleles from African rice (Oryza glaberrima) markedly improves thermotolerance in Asian cultivars (Li et al., 2015b), while bZIP68 alleles from teosinte enhance cold tolerance in maize. These examples underscore the value of expanding breeding pools beyond domesticated germplasm to capture adaptive variation shaped by natural selection.

Genome‐editing technologies further enable precise manipulation of temperature‐resilience traits. In rice, targeted editing of NAT1, which controls leaf surface wax deposition, improves thermotolerance at both the seedling and reproductive stages, conferring yield advantages under field conditions (Lu et al., 2025). Similarly, precise insertion of a 10‐bp heat shock element into promoters of cell‐wall‐invertase genes enhances carbon allocation to grain and fruits in rice and tomato, markedly increasing yield under heat stress without compromising quality (Lou et al., 2025). In maize, editing of COOL1, CPK17, and TIP4;3 improves cold tolerance (Zeng et al., 2024, 2025a), demonstrating the feasibility of genome‐guided thermal adaptation in warm‐origin crops.

A major challenge in temperature‐resilience breeding is the frequent trade‐off between stress tolerance and growth. Classical cold regulators such as CBFs impose growth penalties when constitutively activated through repression of gibberellin (GA) biosynthesis and enhanced DELLA accumulation (Jaglo‐Ottosen et al., 1998; Achard et al., 2008). However, recent studies demonstrate that precise tuning of transcription factors and hormonal homeostasis can decouple stress resistance from growth inhibition. For example, DREB1C overexpression in rice enhances nitrogen‐use efficiency and photosynthesis, leading to higher yield and a shortened growth cycle (Wei et al., 2022). Likewise, chilling‐induced phosphorylation stabilizes IPA1 (IDEAL PLANT ARCHITECTURE 1), which directly activates OsDREB1C transcription to confer chilling tolerance while maintaining optimal plant architecture (Jia et al., 2022). In maize, co‐overexpression of both ZmHSF12‐1 and ZmHSF12‐2 enhances thermotolerance while mitigating growth penalty, highlighting the importance of cooperative regulatory balance (Qi et al., 2025). Fine‐tuning GA biosynthesis genes such as ATT1/2 similarly enables simultaneous enhancement of yield and tolerance to combined alkali and heat stress (Guo et al., 2025c).

Collectively, these findings indicate that precision regulation rather than maximal activation of key regulators is central to designing crops that are both stress‐resilient and high‐yielding. Future strategies should prioritize synergistic gene modules, smart promoter design, post‐translational modification tuning, and multi‐gene stacking to achieve yield stability, or even yield enhancement, under thermal stress.

APPLICATIONS OF ARTIFICIAL INTELLIGENCE IN CROP SCIENCE

Artificial intelligence (AI) is rapidly transforming crop science by enabling data‐driven discovery, prediction, and decision‐making across breeding, phenotyping, and stress adaptation. By integrating heterogeneous datasets with advanced learning algorithms, AI provides powerful tools to address the complexity of climate resilience.

Smart breeding and genomic design

AI is reshaping crop breeding by enabling high‐resolution decoding of complex genotype–phenotype relationships. AI‐driven genomic selection models, including random forests and gradient boosting, leverage genome‐wide marker data to predict breeding values with high accuracy, substantially shortening breeding cycles and accelerating selection gain (Crossa et al., 2017). More recently, deep learning frameworks, such as graph neural networks, have further improved prediction of polygenic traits by capturing nonlinear effects and gene–gene interactions that are often missed by conventional statistical models (Wang et al., 2023b).

Beyond predictive breeding, AI is increasingly integrated into gene discovery and molecular design. Natural language processing tools extract gene–trait associations from large‐scale literature and databases to construct knowledge graphs, facilitating target prioritization. Meanwhile, deep learning‐based sequence and structure models enhance protein structure prediction and genome‐editing target design (Jumper et al., 2021), streamlining functional validation and precision engineering.

AI is also driving automation in hybrid breeding. The development of genome‐edited “robot‐compatible” structural male‐sterile lines, together with the autonomous pollination robot GEAIR powered by deep learning algorithms, represents a milestone toward fully automated hybrid seed production (Xie et al., 2025).

A recent study provides a compelling example of AI‐assisted molecular design. The maize E3 ligase NLA (Nitrogen Limitation Adaptation) was identified as a key regulatory hub linking cold signaling and phosphate homeostasis. Under cold stress, NLA enhances JA‐mediated cold tolerance by degrading JAZ11, while repressing phosphate uptake via ubiquitination of the phosphate transporter PT4, creating a trade‐off between cold resilience and phosphorus‐use efficiency. Using AI‐guided structural modeling and genome editing, the NLA^Δ12 variant was engineered to uncouple PT4 regulation from JA‐mediated cold signaling. This rational redesign simultaneously improves cold tolerance, phosphorus‐use efficiency, and yield, illustrating how AI‐enabled precision design can overcome physiological trade‐offs (Liao et al., 2026).

Collectively, these advances signal a transition toward closed‐loop, intelligent breeding systems that integrate prediction, rational design, genome editing, and automated implementation, paving the way for climate‐resilient and resource‐efficient crop improvement.

High‐throughput phenotyping and growth monitoring

AI‐integrated sensing technologies are overcoming long‐standing bottlenecks in field phenotyping by enabling automated, high‐throughput, and quantitative trait acquisition. Unmanned aerial vehicles (UAVs) and ground‐based platforms equipped with multispectral, hyperspectral, and RGB sensors facilitate real‐time monitoring of crop canopies, while convolutional neural network‐based image analysis allows precise extraction of morphological and physiological traits, including plant architecture, biomass, and stress‐related indices (Jiang et al., 2021). Beyond growth traits, AI models have been successfully applied to early stress diagnostics, particularly for the rapid detection of diseases and pest infestations under field conditions (Liu et al., 2024b). Below‐ground phenotyping, traditionally constrained by technical limitations, has also advanced markedly through the integration of X‐ray computed tomography and deep learning tools such as RootPainter, enabling automated reconstruction and quantitative analysis of three‐dimensional root system architecture (Lucas et al., 2023).

Stress response and climate adaptation

AI provides powerful new avenues for dissecting stress‐response networks and predicting climate adaptability. Machine learning has identified key regulators such as OsRAV11/12 that coordinate drought tolerance and flowering time in rice (Zhang et al., 2025e). By integrating multi‐omics datasets, deep learning models further enable the systemic identification of gene modules and metabolic pathways underlying drought and temperature stress responses (Liu et al., 2025a; Zhang et al., 2025b). Importantly, AI‐based climate adaptability prediction enables matching varieties to future climate scenarios, reducing production risk under environmental uncertainty (Fradgley et al., 2023). Together, these approaches support a transition from empirical breeding toward predictive, demand‐driven climate‐resilient crop design.

FUTURE PERSPECTIVES

Global climate change, characterized by rising average temperatures and frequent extremes, poses an escalating threat to global food security. Development of temperature‐resilient cultivars of major crops has therefore become an urgent priority. While major advances have been made in elucidating the genetic, molecular, and epigenetic mechanisms of temperature responses, most insights have been derived from model species and controlled laboratory conditions. Translation of these discoveries into stable yield gains under fluctuating field environments remains a central challenge.

Future research must integrate temperature perception, signaling, and memory into unified frameworks. Epigenetic mechanisms, exemplified by vernalization, offer paradigms for how plants store and recall thermal information; yet, how these processes operate across tissues, developmental stages, and environmental contexts remains incompletely understood. Emerging technologies such as single‐cell and spatial multi‐omics, live‐cell imaging, and long‐term field phenomics will be critical for bridging this gap.

In nature, temperature stress rarely occurs in isolation but interacts with drought, salinity, nutrient limitation, and biotic challenges. Deciphering the molecular logic of combined stress responses and validating key loci under realistic field conditions will be essential for robust breeding outcomes. Large‐scale multi‐environment trials, coupled with genomic and phenomic analyses, should become standard practice.

Finally, the convergence of AI, molecular breeding, and environmental modeling heralds a new era of climate‐smart agriculture. By integrating mechanistic understanding with predictive analytics and precision breeding, it will be possible to design crop varieties that combine resilience, productivity, and stability. Bridging laboratory insights with field complexity through AI‐assisted design represents one of the most promising pathways for sustaining global agriculture under climate change.

CONFLICTS OF INTEREST

The authors declare no conflicts of interest.

AUTHOR CONTRIBUTIONS

K.C., J.L., and S.Y. conceived the study. R.Z., C.Y., W.L., and L.L.Z. drafted the manuscript and prepared the figures. K.C., J.L., and S.Y. critically revised and edited the manuscript. All authors have read and approved the contents of this paper.

Biographies

Zeng, R. , Yang, C. , Luo, W. , Zhang, L.‐L. , Chong, K. , Liu, J.‐X. , and Yang, S. (2026). Temperature regulation in plants: From molecular mechanisms to climate‐resilient crop improvement. J. Integr. Plant Biol. 68: 2454–2488.