ABF1 Positively Regulates Rice Chilling Tolerance via Inducing Trehalose Biosynthesis

Chilling stress seriously limits grain yield and quality worldwide. However, the genes and the underlying mechanisms that respond to chilling stress remain elusive. This study identified ABF1, a cold-induced transcription factor of the bZIP family. Disruption of ABF1 impaired chilling tolerance with increased ion leakage and reduced proline contents, while ABF1 over-expression lines exhibited the opposite tendency, suggesting that ABF1 positively regulated chilling tolerance in rice. Moreover, SnRK2 protein kinase SAPK10 could phosphorylate ABF1, and strengthen the DNA-binding ability of ABF1 to the G-box cis-element of the promoter of TPS2, a positive regulator of trehalose biosynthesis, consequently elevating the TPS2 transcription and the endogenous trehalose contents. Meanwhile, applying exogenous trehalose enhanced the chilling tolerance of abf1 mutant lines. In summary, this study provides a novel pathway ‘SAPK10-ABF1-TPS2’ involved in rice chilling tolerance through regulating trehalose homeostasis.


Introduction
As staple food for more than half of the world's population, rice (Oryza sativa L.) is highly sensitive to chilling stress [1]. With an expanding rice cultivation area from the original tropical and sub-tropical regions to high-latitude areas, chilling stress becomes a major factor that damages crop growth and development, restricts geographical distribution, and eventually constrains yield and quality [2]. Therefore, it is of great value to uncover the molecular mechanism in response to chilling stress for breeding chilling-tolerant varieties.
In the past decades, evidence has shown that trehalose enhances plants' growth and development in response to cold stress [3][4][5]. The synthesis of trehalose mainly includes two steps, the first is the conversion of UDP (uridine diphosphate)-glucose and glucose-6-phosphate to T6P (Trehalose-6-phosphate) and UDP by TPS (Trehalose-6-phosphate synthase), then TPP (Trehalose-6-phosphate phosphatase) dephosphorylates T6P to trehalose, eventually trehalase (TRE) degraded trehalose into two glucose [6,7]. In Arabidopsis thaliana, over-expression of TPPI, a member of the TPP family, improves seedling chilling tolerance through accumulating Jasmonic acid and soluble sugar, such as trehalose, while tppI mutant lines are sensitive to low temperatures [8]. Overexpression of TPS1 or TPP1 increases trehalose content and strengthens seedlings chilling adaptation in rice [3,9]. Similarly, the over-expression of wheat TPS11 in Arabidopsis shows a higher survival rate in response to chilling stress compared with the wild type [10]. On the other hand, applying exogenous trehalose could enhance chilling tolerance by increasing endogenous trehalose levels, root growth, and water uptake [3,5,11]. The application of trehalose could also improve the efficiency and quantum yield of PS-II of the photosynthetic system, increase NO and H 2 O 2 levels to enhance antioxidant activity and suppress membrane lipid peroxidation, whereas reducing the electrolyte leakage to maintain plant growth and harvest in response to cold stress [4].
The basic leucine zipper (bZIP) transcription factors have a bZIP domain composed of two structural features: a DNA-binding base region and a leucine zipper dimerization region [12]. Numerous bZIP transcription factors have been reported to play essential roles in plants' survival under cold environments [13][14][15]. In rice, the expression of bZIP52 is strongly induced by low temperatures. Several abiotic adversity-related genes such as LEA3, TPP1, and Rab25 are down-regulated in bZIP52 over-expression (OxbZIP52) lines, and OxbZIP52 shows increased cold sensitivity, suggesting that bZIP52 acts as a negative regulator of cold stress response [16]. bZIP73 regulates rice cold tolerance not only in the seedling period, but also during the heading period. Japonica type bZIP73 Jap could interact with bZIP71 to control the balance of abscisic acid (ABA) and reactive oxygen species, thereby improving rice chilling tolerance [13,17]. In wheat, bZIP60 is induced by ethylene glycol, salt, cold, and ABA, and over-expression of bZIP60 in Arabidopsis significantly improves the seedling chilling tolerance [18]. bZIP96 is reported to interact with ICE1 (inducer of CBF expression) to positively regulate wheat cold adaptation [19]. In maize, bZIP68, a negative regulator inhibiting seedling chilling tolerance, is phosphorylated by MPK8 to promote protein stability and the DNA binding affinity to the promoter of DREB 1.7 under chilling stress [14]. In chrysanthemums, Bai et al. (2022) found that bZIP3 could interact with bZIP2 to activate the expression of POD, and then promote the peroxidase activity to modulate the balance of reactive oxygen species to improve the cold stress tolerance [15].
The above research shows that both trehalose and bZIP transcription factors are vital for plant chilling tolerance. However, whether a relationship exists between them in response to cold stress is still uncovered. This study identified that ABF1 (bZIP12) acts as a positive regulator in rice chilling tolerance by inducing trehalose biosynthesis through the novel pathway 'SAPK10-ABF1-TPS2'. Briefly, abf1 mutant lines exhibited increased chilling sensitivity by interfering with trehalose biosynthesis. SnRK2 protein kinase SAPK10 could phosphorylate ABF1 and promote the latter's binding ability to TPS2, a member of the trehalose-6-phosphate synthase family, ultimately inducing trehalose biosynthesis under chilling stress.

ABF1 Acts as a Positive Regulator of Rice Chilling Tolerance
In this study, we first analyzed the expression level of ABF1 in response to cold treatment at 0, 1, 3, 6 and 12 h. Upon cold treatment, the transcription level of ABF1 sharply rose to about 3 fold at 1 h, and kept steady until 6 h, then dropped slowly to about 2.5 fold at 12 h, which indicated that ABF1 might participate in rice chilling adaption ( Figure S1). To explore the biological role of ABF1, two independent homozygous mutant lines (abf1-1 and abf1-2) and two ABF1 over-expression lines (OxABF1-1 and OxABF1-9) were generated ( Figure S2). The abf1-1 and abf1-2 mutant lines harbored a TC deletion and a TCGC deletion in the first exon and shifted the open reading frame of ABF1, respectively ( Figure S2). As shown in Figure 1, under normal conditions, no significant difference in seedling growth was distinguished between the wild-type and mutant lines ( Figure 1). However, the leaves of the mutant lines showed more severe wilting and crimping than the wild-type under cold treatment for 5 days after 7-days recovery, with the survival rates (less than 23%) notably lower than that of the wild type (average 32%), while the survival rates of ABF1 over-expression lines (average 49%) were significantly higher than the wild type (Figures 1 and 2A,B). Previous reports have shown that chilling stress affects numerous physiological and cellular processes in plants, including inhibiting photosynthesis, regulating calcium signals, and altering the organelle's ultrastructure, especially the chloroplast lipid membrane [20]. Accordingly, plants employ several strategies when recovering from chilling to normal temperatures, such as inducing the expression of photosynthetic enzymes, enhancing the linoleic acid metabolism and inhibiting phenylpropanoid biosynthesis and ribosome synthesis [21]. Among these processes, maintaining the cell membrane integrity and stability is vital for chilling tolerance, and ion leakage reflects the cell membrane's damage degree, while proline declines the electrolyte leakage level to protect membrane stability [22,23]. In accordance with the increased cold sensitivity of abf1, the mutant lines' electrolyte leakage was higher than the wild type, accompanied by reduced proline contents ( Figure 2B-D). In contrast, ABF1 over-expression lines exhibited the opposite tendency under cold stress ( Figure 2B-D). These results demonstrated that ABF1 positively participated in rice chilling tolerance. than 23%) notably lower than that of the wild type (average 32%), while the survival rates of ABF1 over-expression lines (average 49%) were significantly higher than the wild type ( Figures 1 and 2A,B). Previous reports have shown that chilling stress affects numerous physiological and cellular processes in plants, including inhibiting photosynthesis, regulating calcium signals, and altering the organelle's ultrastructure, especially the chloroplast lipid membrane [20]. Accordingly, plants employ several strategies when recovering from chilling to normal temperatures, such as inducing the expression of photosynthetic enzymes, enhancing the linoleic acid metabolism and inhibiting phenylpropanoid biosynthesis and ribosome synthesis [21]. Among these processes, maintaining the cell membrane integrity and stability is vital for chilling tolerance, and ion leakage reflects the cell membrane's damage degree, while proline declines the electrolyte leakage level to protect membrane stability [22,23]. In accordance with the increased cold sensitivity of abf1, the mutant lines' electrolyte leakage was higher than the wild type, accompanied by reduced proline contents ( Figure 2B-D). In contrast, ABF1 over-expression lines exhibited the opposite tendency under cold stress ( Figure 2B-D). These results demonstrated that ABF1 positively participated in rice chilling tolerance. Figure 1. ABF1 positively regulates rice seedling chilling tolerance. abf1 mutant lines are less tolerant to chilling stress with more severe wilting and crimping leaves than the wild type. ABF1 overexpression lines showed the opposite tendency. Seven-day-old seedlings were hydroponically cultured under chilling stress (4 °C) for 5 days and photographed. Red arrow indicated the wilting and crimping leaves. Bar = 3 cm. Figure 1. ABF1 positively regulates rice seedling chilling tolerance. abf1 mutant lines are less tolerant to chilling stress with more severe wilting and crimping leaves than the wild type. ABF1 overexpression lines showed the opposite tendency. Seven-day-old seedlings were hydroponically cultured under chilling stress (4 • C) for 5 days and photographed. Red arrow indicated the wilting and crimping leaves. Bar = 3 cm.
Previous reports have shown that ABF1 is involved in various abiotic stress responses, such as hypoxia, salt, drought, and cold adversity stress [24]. abf1 mutant lines are hypersensitive to drought and salt stresses, and ABF1 over-expression lines significantly improve drought tolerance [25,26]. Moreover, ABF1 acts as a negative regulator for flowering transition and is functionally redundant with bZIP40, which may delay heading by directly activating WRKY104 and indirectly inhibiting Ehd1 under drought conditions [26]. Our lab recently reported that ABF1 regulates plant height and seed germination by inhibiting gibberellin synthesis in rice [27]. Overexpression of ABF1 results in a typical gibberellin deficient phenotype, exhibiting semi-dwarfism and retarded seed germination, which could be restored by applying exogenous GA3 [27]. In this study, we have detected that ABF1 was induced by chilling stress, which was consistent with the described previously, and further found a novel role of ABF1 involved in elevating rice chilling adaption, indicating ABF1 could play pleiotropic roles in regulating various abiotic stresses tolerance. Survival rates (%) = the number of seedlings that survived after chilling stress treatment/total number of the seedlings (%). (B) Survival rates of wild type, ABF1 overexpression lines under chilling stress (4 • C) for 5 days. (C) Ion leakage of seedling leaves from wild type, abf1 mutant lines, and ABF1 overexpression lines under chilling stress (4 • C) for 5 days. (D) Proline contents of seedling leaves from wild type, abf1 mutant lines and ABF1 overexpression lines under chilling stress (4 • C) for 5 days. Error bars indicate SD with biological triplicates (n = 3). Tukey's test with one-way analysis of variance (ANOVA). Different letters indicate statistical differences at p < 0.05.
Previous reports have shown that ABF1 is involved in various abiotic stress responses, such as hypoxia, salt, drought, and cold adversity stress [24]. abf1 mutant lines are hypersensitive to drought and salt stresses, and ABF1 over-expression lines significantly improve drought tolerance [25,26]. Moreover, ABF1 acts as a negative regulator for flowering transition and is functionally redundant with bZIP40, which may delay heading by directly activating WRKY104 and indirectly inhibiting Ehd1 under drought conditions [26]. Our lab recently reported that ABF1 regulates plant height and seed germination by inhibiting gibberellin synthesis in rice [27]. Overexpression of ABF1 results in a typical gibberellin deficient phenotype, exhibiting semi-dwarfism and retarded seed germination, which could be restored by applying exogenous GA 3 [27]. In this study, we have detected that ABF1 was induced by chilling stress, which was consistent with the described previously, and further found a novel role of ABF1 involved in elevating rice chilling adaption, indicating ABF1 could play pleiotropic roles in regulating various abiotic stresses tolerance.

SAPK10 Directly Interacts with and Phosphorylates ABF1
After screening a rice seedling-derived cDNA library using a yeast two-hybrid system, we found that SAPK10, a member of the specific Ser/Thr SnRK2 (Sucrose non-fermenting-1-related protein kinase 2) type kinase family, could interact with ABF1 ( Figure 3A). Intriguingly, the transcription level of SAPK10 was also significantly elevated under cold treatment ( Figure S1). A bimolecular fluorescence complementation (BiFC) assay detected no YFP fluorescence signals when YN-ABF1 or YC-SAPK10 was expressed alone, and ABF1 and SAPK10 could interact to generate YFP fluorescence in the nuclei of tobacco leaf cells ( Figure 4). In the GST pull-down assay, the purified His-SAPK10 was pulled down with GST-ABF1, but not with GST, and Coimmunoprecipitation (Co-IP) assay exhibited a similar result, ABF1-GFP was immunoprecipitated by SAPK10-FLAG, while GFP was not ( Figure 3B,C). Moreover, in vitro kinase assays detected a phosphorylation band of purified GST-ABF1 (p-GST-ABF1) when GST-ABF1 co-expressed with His-SAPK10 in E. coli, and the relative strength of p-GST-ABF1 band was reduced about 23% of that of GST-ABF1 band after CIAP treatment ( Figure 3D). Noteworthy, the band shift mobility of p-GST-ABF1 is significantly slower than that of GST-ABF1, further supporting the highly specific of the SAPK10-mediated ABF1 phosphorylation ( Figure 3D). Additionally, we also detected the auto-phosphorylation band of SAPK10-FLAG, which was consistent with previous work [28].

SAPK10 Directly Interacts with and Phosphorylates ABF1
After screening a rice seedling-derived cDNA library using a yeast two-hybrid system, we found that SAPK10, a member of the specific Ser/Thr SnRK2 (Sucrose non-fermenting-1-related protein kinase 2) type kinase family, could interact with ABF1 ( Figure  3A). Intriguingly, the transcription level of SAPK10 was also significantly elevated under cold treatment ( Figure S1). A bimolecular fluorescence complementation (BiFC) assay detected no YFP fluorescence signals when YN-ABF1 or YC-SAPK10 was expressed alone, and ABF1 and SAPK10 could interact to generate YFP fluorescence in the nuclei of tobacco leaf cells (Figure 4). In the GST pull-down assay, the purified His-SAPK10 was pulled down with GST-ABF1, but not with GST, and Coimmunoprecipitation (Co-IP) assay exhibited a similar result, ABF1-GFP was immunoprecipitated by SAPK10-FLAG, while GFP was not ( Figure 3B,C). Moreover, in vitro kinase assays detected a phosphorylation band of purified GST-ABF1 (p-GST-ABF1) when GST-ABF1 co-expressed with His-SAPK10 in E.coli, and the relative strength of p-GST-ABF1 band was reduced about 23% of that of GST-ABF1 band after CIAP treatment ( Figure 3D). Noteworthy, the band shift mobility of p-GST-ABF1 is significantly slower than that of GST-ABF1, further supporting the highly specific of the SAPK10-mediated ABF1 phosphorylation ( Figure 3D). Additionally, we also detected the auto-phosphorylation band of SAPK10-FLAG, which was consistent with previous work [28].

Disruption of SAPK10 Impairs Rice Chilling Tolerance
To further clarify the role of SAPK10 in regulating rice chilling tolerance, two independent homozygous mutant lines (sapk10-2 and sapk10-3) and two SAPK10 overexpression lines (OxSAPK10-1 and OxSAPK10-3) were used for further analysis, respectively ( Figure S3). The sapk10-2 and sapk10-3 mutant lines harbored an A insertion or an A/G insertion in the first exon and shifted the open reading frame of SAPK10 ( Figure S3). Interestingly, sapk10 mutant lines displayed a phenotype similar to abf1 under cold stress. sapk10 mutant lines also exhibited more severe wilting and crimping than the wild type, with the survival rates (less than 25%) lower than the wild type ( Figures 5 and 6A). Although the height of SAPK10 overexpression lines was shorter than the wild type under normal conditions as previous report [28], the survival rates (average 40%) of SAPK10 overexpression lines were higher compared with the wild type under cold treatment (Figures 5 and 6B). In accordance with the above observations, the electrolyte leakage of sapk10 increased more rapidly than in the wild type. In contrast, the electrolyte leakage was reduced in SAPK10 overexpression lines under cold treatment, whereas the proline contents of these transgenic lines showed an opposite tendency ( Figure 6C,D). These experiments showed that SAPK10 enhanced rice chilling adaptation.

Disruption of SAPK10 Impairs Rice Chilling Tolerance
To further clarify the role of SAPK10 in regulating rice chilling tolerance, two independent homozygous mutant lines (sapk10-2 and sapk10-3) and two SAPK10 overexpression lines (OxSAPK10-1 and OxSAPK10-3) were used for further analysis, respectively ( Figure S3). The sapk10-2 and sapk10-3 mutant lines harbored an A insertion or an A/G insertion in the first exon and shifted the open reading frame of SAPK10 ( Figure S3). Interestingly, sapk10 mutant lines displayed a phenotype similar to abf1 under cold stress. sapk10 mutant lines also exhibited more severe wilting and crimping than the wild type, with the survival rates (less than 25%) lower than the wild type ( Figures 5 and 6A). Although the height of SAPK10 overexpression lines was shorter than the wild type under normal conditions as previous report [28], the survival rates (average 40%) of SAPK10 overexpression lines were higher compared with the wild type under cold treatment ( Figures 5 and 6B). In accordance with the above observations, the electrolyte leakage of sapk10 increased more rapidly than in the wild type. In contrast, the electrolyte leakage was reduced in SAPK10 overexpression lines under cold treatment, whereas the proline contents of these transgenic lines showed an opposite tendency ( Figure 6C,D). These experiments showed that SAPK10 enhanced rice chilling adaptation. SnRK2 is a specific Ser/Thr protein kinase that plays essential roles in osmotic stress response and other abiotic stress processes [29]. In rice, there are ten SnRK2 family members, namely SAPK1-SAPK10, of which SAPK6 and SAPK8 act as crucial protein kinases in the low-temperature signal transduction pathway [30,31]. Under cold stress, SAPK6 interacts with and phosphorylates IPA1 at Serine 201 and 213 sites to stabilize the protein stability of IPA1, thereafter enhancing IPA1 accumulation and activating the expression level of CBF3, ultimately enhancing rice cold stress adaption [31].  report that SAPK8 phosphorylates a cyclic nucleotide-gated channel CNGC9, activating its channel activity, and promoting cold-induced calcium influx and cytoplasmic calcium elevation in rice [30]. In this study, we found a novel role of SAPK10 involved in positively regulating rice chilling tolerance, though it has been reported to either phosphorylate bZIP72 or bZIP20 to participate in abscisic acid and jasmonic acid synergistically inhibiting seed germination or enhancing drought and salt tolerance processes [28,32]. Interestingly, SAPK6, SAPK8, SAPK10, and ABF1 are all induced by abscisic acid (ABA) and act as positive regulators in the ABA signal pathway [25,29,33]. Since ABA biosynthesis is strengthened during cold stress and a series of ABA-responsive genes are induced under cold treatment, whereas the mutant lines interfering with ABA biosynthesis exhibit impaired cold adaptation, and application of exogenous ABA enhances chilling tolerance, suggesting that ABA functions as a vital phytohormone positively regulating cold stress response [34,35]. Whether the three protein kinases or other members of the SnRK2 family interact with ABF1 involved in the crosstalk between ABA and chilling tolerance would be an interesting topic to be further investigated. SnRK2 is a specific Ser/Thr protein kinase that plays essential roles in osmotic stress response and other abiotic stress processes [29]. In rice, there are ten SnRK2 family members, namely SAPK1-SAPK10, of which SAPK6 and SAPK8 act as crucial protein kinases in the low-temperature signal transduction pathway [30,31]. Under cold stress, SAPK6 interacts with and phosphorylates IPA1 at Serine 201 and 213 sites to stabilize the protein stability of IPA1, thereafter enhancing IPA1 accumulation and activating the expression level of CBF3, ultimately enhancing rice cold stress adaption [31].  report that SAPK8 phosphorylates a cyclic nucleotide-gated channel CNGC9, activating its channel activity, and promoting cold-induced calcium influx and cytoplasmic calcium elevation in rice [30]. In this study, we found a novel role of SAPK10 involved in positively regulating rice chilling tolerance, though it has been reported to either phosphorylate bZIP72 or bZIP20 to participate in abscisic acid and jasmonic acid synergistically inhibiting seed germination or enhancing drought and salt tolerance processes [28,32]. Interestingly, SAPK6, SAPK8, SAPK10, and ABF1 are all induced by abscisic acid (ABA) and act as positive regulators in the ABA signal pathway [25,29,33]. Since ABA biosynthesis is strengthened during cold stress and a series of ABA-responsive genes are induced under cold treatment, whereas the mutant lines interfering with ABA biosynthesis exhibit impaired cold adaptation, and application of exogenous ABA enhances chilling tolerance, suggesting that ABA functions as a vital phytohormone positively regulating cold stress response [34,35]. Whether the three protein kinases or other members of the SnRK2 family interact with ABF1 involved in the crosstalk between ABA and chilling tolerance would be an interesting topic to be further investigated.

abf1 Inhibits Seedling Cold Adaptation by Interfering with Trehalose Homeostasis
Previous reports have shown that trehalose improves chilling tolerance in several species [3,8,36], which triggered us to speculate that ABF1-mediated chilling alleviation may be partially based on trehalose accumulation. We first detect the endogenous trehalose level in abf1 mutant lines to verify this hypothesis. As anticipated, the abf1 seedlings had lower trehalose levels than the wild-type under cold treatment, while there was no difference under normal conditions ( Figure 7A). Furthermore, applying exogenous trehalose could enhance the chilling tolerance of abf1, relieving severe wilting and crimping phenomena ( Figure 8). Notably, the relative survival rates of abf1 were significantly higher than the wild type when exogenous trehalose was added ( Figure 7B,C). Accordingly, the endogenous trehalose contents in ABF1 over-expression lines were higher than the wildtype in response to chilling stress, though slightly higher than the wild-type under normal conditions ( Figure 7D). Thus, these results strongly suggested that ABF1 promotes trehalose contents to enhance seedlings chilling tolerance.
The underlying mechanisms of bZIP transcription factors regulating rice cold tolerance have been reported in the past decades.  have shown that one single nucleotide polymorphism difference in the coding region of bZIP73 causes the difference in low-temperature tolerance between Indica and Japonica rice. And Japonica type

abf1 Inhibits Seedling Cold Adaptation by Interfering with Trehalose Homeostasis
Previous reports have shown that trehalose improves chilling tolerance in several species [3,8,36], which triggered us to speculate that ABF1-mediated chilling alleviation may be partially based on trehalose accumulation. We first detect the endogenous trehalose level in abf1 mutant lines to verify this hypothesis. As anticipated, the abf1 seedlings had lower trehalose levels than the wild-type under cold treatment, while there was no difference under normal conditions ( Figure 7A). Furthermore, applying exogenous trehalose could enhance the chilling tolerance of abf1, relieving severe wilting and crimping phenomena ( Figure 8). Notably, the relative survival rates of abf1 were significantly higher than the wild type when exogenous trehalose was added ( Figure 7B,C). Accordingly, the endogenous trehalose contents in ABF1 over-expression lines were higher than the wildtype in response to chilling stress, though slightly higher than the wild-type under normal conditions ( Figure 7D). Thus, these results strongly suggested that ABF1 promotes trehalose contents to enhance seedlings chilling tolerance. stress, applying trehalose increases NO levels and stimulates H2O2 contents, resulting in enhanced antioxidant activity and reduced membrane lipid peroxide activity [4]. In addition, exogenous trehalose treatment increases the contents of endogenous spermidine, glutathione, and ascorbic acid, promotes the glutathione and ascorbic acid cycle, and finally reduces the ROS production to improve the cold tolerance of plants [36]. Hence, a possibility exists that trehalose may play a role in ABF1-mediated low-temperature tolerance by affecting the active oxygen homeostasis in vivo.  The underlying mechanisms of bZIP transcription factors regulating rice cold tolerance have been reported in the past decades.  have shown that one single nucleotide polymorphism difference in the coding region of bZIP73 causes the difference in low-temperature tolerance between Indica and Japonica rice. And Japonica type bZIP73 Jap interacts with bZIP71 to regulate abscisic acid (ABA) level and active oxygen balance, thus improving rice tolerance to low temperatures [13]. Moreover, the regulatory effect of bZIP73 Jap on low-temperature tolerance also occurs during the booting stage. Under natural cold stress conditions, the co-expression of bZIP73 Jap and bZIP71 transgenic lines significantly improves seed setting rate and grain yield, not only inhibiting ABA levels in anthers, but also promoting the transportation of soluble sugars from anthers to pollen [17]. This study found that trehalose, except ABA, active oxygen and soluble sugars, also participates in bZIPs-mediated cold tolerance processes. Under low-temperature stress, applying trehalose increases NO levels and stimulates H 2 O 2 contents, resulting in enhanced antioxidant activity and reduced membrane lipid peroxide activity [4]. In addition, exogenous trehalose treatment increases the contents of endogenous spermidine, glutathione, and ascorbic acid, promotes the glutathione and ascorbic acid cycle, and finally reduces the ROS production to improve the cold tolerance of plants [36]. Hence, a possibility exists that trehalose may play a role in ABF1-mediated low-temperature tolerance by affecting the active oxygen homeostasis in vivo.

Int. J. Mol. Sci. 2023, 24, x FOR PEER REVIEW
10 of 20 Figure 8. The chilling tolerance of abf1 mutant lines was enhanced by exogenous trehalose application. Phenotypes of wild type, abf1 mutant under chilling stress (4 °C) with or without 10 mM exogenous trehalose for 5 days. abf1 mutant lines are less tolerant to chilling stress with more severe wilting and crimping leaves than the wild type. Applying exogenous trehalose enhanced abf1 chilling tolerance with less wilting and crimping leaves. Bar = 3 cm.

ABF1 Binds to and Activates TPS2 Expression
Since ABF1 strengthens rice seedling adaptation by inducing trehalose accumulation, indicating that ABF1 may induce trehalose biosynthesis or suppress trehalose catabolism. This hypothesis triggered us to detect the expression level of a series of crucial enzymes, including TPS (Trehalose-6-phosphate synthase), TPP (Trehalose-6-phosphate phosphatase), and TRE (Trehalase) involved in the trehalose biosynthesis and catabolism pathway [6,7], in abf1 seedlings under cold treatment. As shown in Figure 9A, TPS2, TPP5 and TPP6 involved in trehalose biosynthesis were significantly reduced in abf1 mutant lines ( Figure  9A). Among these enzymes, TPS2 attracted our particular interest due to ABF1 binds the G-box (CACGTG) motif of TPS2 by EMSA (electrophoresis mobility shift assay), and the binding ability disappeared after the unlabeled probe was added ( Figure 9C). To our surprise, ABF1 did not interact with other TPP or TPS members, of which the promoters also harbored G-box cis-elements predicted by PlanCARE (Plant Cis-Acting Regulatory Elements) [37] (Figure S4). Yeast-one hybrid (Y1H) assay further confirmed that ABF1 could bind the promoter of TPS2 and activate the LacZ expression, though it displayed weak autoactivation activity ( Figure 9B). The binding pattern of ABF1 on the promoter of TPS2 was further validated by ChIP-qPCR (Chromatin immunoprecipitation-quantitative PCR) under chilling stress. ABF1 was notably enriched in the P3 region containing the G-box motif of the promoter of TPS2, nearly threefold higher than the other probes ( Figure 9D). Finally, a luciferase transient transcriptional activity assay (LUC) was used to detect the regulatory effect of ABF1 on TPS2. The luciferase activity of the proTPS2: firefly LUC reporter was significantly facilitated by pro35S: ABF1: tNOS, which was consistent with the expression pattern of TPS2 in abf1 ( Figure 9A,E). Intriguingly, the upregulation of TPS2 by ABF1 was remarkably enhanced when pro35S: SAPK10: tNOS was added, indicating that SAPK10-mediated phosphorylation on ABF1 promotes ABF1-mediated TPS2 expression ( Figure 9E). On the other hand, TPS2 was also induced in response to cold treatment, which was similar to the expression pattern of ABF1 and SAPK10 ( Figure S1). These results strongly implied that ABF1 specifically binds to TPS2 and activates the latter's expression. Phenotypes of wild type, abf1 mutant under chilling stress (4 • C) with or without 10 mM exogenous trehalose for 5 days. abf1 mutant lines are less tolerant to chilling stress with more severe wilting and crimping leaves than the wild type. Applying exogenous trehalose enhanced abf1 chilling tolerance with less wilting and crimping leaves. Bar = 3 cm.

ABF1 Binds to and Activates TPS2 Expression
Since ABF1 strengthens rice seedling adaptation by inducing trehalose accumulation, indicating that ABF1 may induce trehalose biosynthesis or suppress trehalose catabolism. This hypothesis triggered us to detect the expression level of a series of crucial enzymes, including TPS (Trehalose-6-phosphate synthase), TPP (Trehalose-6-phosphate phosphatase), and TRE (Trehalase) involved in the trehalose biosynthesis and catabolism pathway [6,7], in abf1 seedlings under cold treatment. As shown in Figure 9A, TPS2, TPP5 and TPP6 involved in trehalose biosynthesis were significantly reduced in abf1 mutant lines ( Figure 9A). Among these enzymes, TPS2 attracted our particular interest due to ABF1 binds the G-box (CACGTG) motif of TPS2 by EMSA (electrophoresis mobility shift assay), and the binding ability disappeared after the unlabeled probe was added ( Figure 9C). To our surprise, ABF1 did not interact with other TPP or TPS members, of which the promoters also harbored G-box cis-elements predicted by PlanCARE (Plant Cis-Acting Regulatory Elements) [37] ( Figure S4). Yeast-one hybrid (Y1H) assay further confirmed that ABF1 could bind the promoter of TPS2 and activate the LacZ expression, though it displayed weak autoactivation activity ( Figure 9B). The binding pattern of ABF1 on the promoter of TPS2 was further validated by ChIP-qPCR (Chromatin immunoprecipitation-quantitative PCR) under chilling stress. ABF1 was notably enriched in the P3 region containing the G-box motif of the promoter of TPS2, nearly threefold higher than the other probes ( Figure 9D). Finally, a luciferase transient transcriptional activity assay (LUC) was used to detect the regulatory effect of ABF1 on TPS2. The luciferase activity of the proTPS2: firefly LUC reporter was significantly facilitated by pro35S: ABF1: tNOS, which was consistent with the expression pattern of TPS2 in abf1 ( Figure 9A,E). Intriguingly, the upregulation of TPS2 by ABF1 was remarkably enhanced when pro35S: SAPK10: tNOS was added, indicating that SAPK10mediated phosphorylation on ABF1 promotes ABF1-mediated TPS2 expression ( Figure 9E). On the other hand, TPS2 was also induced in response to cold treatment, which was similar to the expression pattern of ABF1 and SAPK10 ( Figure S1). These results strongly implied that ABF1 specifically binds to TPS2 and activates the latter's expression. Moreover, we tested the chilling tolerance of tps2 mutant lines and TPS2 overexpression lines (OxTPS2-1 and OxTPS2-2) ( Figure S5). tps2 mutant lines harbored a T deletion and a T insertion in the first exon, and shifted the open reading frame of TPS2, respectively ( Figure S5). Moreover, the trehalose contents were substantially reduced in tps2 mutant Moreover, we tested the chilling tolerance of tps2 mutant lines and TPS2 overexpression lines (OxTPS2-1 and OxTPS2-2) ( Figure S5). tps2 mutant lines harbored a T deletion and a T insertion in the first exon, and shifted the open reading frame of TPS2, respectively ( Figure S5). Moreover, the trehalose contents were substantially reduced in tps2 mutant lines, and increased in TPS2 overexpression lines ( Figure S6). In accordance with this, tps2 mutant lines displayed more severe wilting and crimping, and lower survival rates than the wild type, while TPS2 overexpression lines exhibited the opposite tendency (Figures 10 and 11A,B). Accordingly, the electrolyte leakage of tps2 was higher, accompanied by reduced proline content, while TPS2 overexpression lines showed lower electrolyte leakage and higher proline concentration compared with the wild type in response to low-temperature stress ( Figure 11C,D). These findings demonstrated that TPS2 positively participates in rice seedling chilling tolerance, which was up-regulated by ABF1 to enhance trehalose biosynthesis. lines, and increased in TPS2 overexpression lines ( Figure S6). In accordance with this, tps2 mutant lines displayed more severe wilting and crimping, and lower survival rates than the wild type, while TPS2 overexpression lines exhibited the opposite tendency ( Figures  10 and 11A,B). Accordingly, the electrolyte leakage of tps2 was higher, accompanied by reduced proline content, while TPS2 overexpression lines showed lower electrolyte leakage and higher proline concentration compared with the wild type in response to lowtemperature stress ( Figure 11C,D). These findings demonstrated that TPS2 positively participates in rice seedling chilling tolerance, which was up-regulated by ABF1 to enhance trehalose biosynthesis.   As a non-reducing sugar found in bacteria and yeasts, trehalose serves as the source of carbon in higher plants. Trehalose regulates a series of physiological, biochemical, and molecular mechanisms in higher plants. Trehalose supplementation promotes root growth and water absorption, and thus improves plant growth under cold stress [11]. Trehalose supplementation also improves the efficiency of PS-II and the quantum yield of PS-II, while reducing the leakage of electrolytes to adapt to chilling stress [4]. On the other hand, trehalose stimulates the antioxidant defense system, such as CAT, APX, to maintain cellular integrity under cold treatment [38]. The functions and the underlying mechanisms of several TPS or TPP members involved in rice chilling tolerance have been reported [39]. TPS1 enhances rice's abiotic tolerance by increasing the trehalose and proline content and inducing the expression of stress-related genes [9]. Overexpression of TPS1 increased the tolerance of transgenic rice seedlings to low temperature, salt, and drought treatments, with the concentrations of trehalose and proline higher than the wild type [9].
MAPK3-bHLH002-TPP1 pathway positively regulates cold tolerance in rice. MAPK3 kinase phosphorylates bHLH002/ICE1 and inhibits the latter's ubiquitination by HOS1, then bHLH002 activates TPP1 expression and ultimately enhances rice cold tolerance [3]. Recently,  further report that TPP1 controls rice seed germination through crosstalk with the abscisic acid (ABA) catabolic pathway [40]. tpp1 mutant lines down-regulate ABA catabolism and inhibit seed germination. The transcription factor As a non-reducing sugar found in bacteria and yeasts, trehalose serves as the source of carbon in higher plants. Trehalose regulates a series of physiological, biochemical, and molecular mechanisms in higher plants. Trehalose supplementation promotes root growth and water absorption, and thus improves plant growth under cold stress [11]. Trehalose supplementation also improves the efficiency of PS-II and the quantum yield of PS-II, while reducing the leakage of electrolytes to adapt to chilling stress [4]. On the other hand, trehalose stimulates the antioxidant defense system, such as CAT, APX, to maintain cellular integrity under cold treatment [38]. The functions and the underlying mechanisms of several TPS or TPP members involved in rice chilling tolerance have been reported [39]. TPS1 enhances rice's abiotic tolerance by increasing the trehalose and proline content and inducing the expression of stress-related genes [9]. Overexpression of TPS1 increased the tolerance of transgenic rice seedlings to low temperature, salt, and drought treatments, with the concentrations of trehalose and proline higher than the wild type [9].
MAPK3-bHLH002-TPP1 pathway positively regulates cold tolerance in rice. MAPK3 kinase phosphorylates bHLH002/ICE1 and inhibits the latter's ubiquitination by HOS1, then bHLH002 activates TPP1 expression and ultimately enhances rice cold tolerance [3]. Recently,  further report that TPP1 controls rice seed germination through crosstalk with the abscisic acid (ABA) catabolic pathway [40]. tpp1 mutant lines down-regulate ABA catabolism and inhibit seed germination. The transcription factor GAMYB can activate the expression of TPP1, increase the trehalose content and induce the transcriptional abundance of the ABA catabolic genes to decrease ABA content in seeds [40]. In this study, we first report the function of cold-induced TPS2 in strengthening seedling cold adaption by enhancing trehalose and proline contents, similar to TPS1. A previous study has reported auxin and cytokinin regulate root system architecture under low-temperature stress, mainly through modulating root cell division, differentiation, and elongation. Moreover, cytokinin determines root architecture and function by fine-turning auxin transport and signaling [41], indicating the cross-talk of different phytohormones controlling the stem cells of the root in response to chilling stress. Recently, our lab found that ABAinduced SAPK10 can phosphorylate bZIP72 and enhance the latter's binding ability to the promoter of AOC, an essential gene in the jasmonic acid (JA) synthesis pathway, thereby inducing AOC expression and increasing endogenous JA levels, finally inhibiting seed germination [28]. Based on these regulators were all induced by cold treatment, the reduced trehalose and proline contents in the mutant lines under chilling conditions, and the interaction among the three regulators, we established a novel pathway 'SAPK10-ABF1-TPS2' in trehalose-mediated rice cold tolerance in this study ( Figure 12). Since ABA and trehalose enhance cold stress tolerance, SAPK10 and ABF1 are ABA signalling positive regulators, and TPS2 is the key enzyme involved in trehalose biosynthesis [25,29,35,39]. Further study will focus on the relationship between trehalose and ABA homeostasis in response to chilling stress, which would deepen our understanding of the regulatory mechanism acclimating to low temperatures. GAMYB can activate the expression of TPP1, increase the trehalose content and induce the transcriptional abundance of the ABA catabolic genes to decrease ABA content in seeds [40]. In this study, we first report the function of cold-induced TPS2 in strengthening seedling cold adaption by enhancing trehalose and proline contents, similar to TPS1. A previous study has reported auxin and cytokinin regulate root system architecture under low-temperature stress, mainly through modulating root cell division, differentiation, and elongation. Moreover, cytokinin determines root architecture and function by fine-turning auxin transport and signaling [41], indicating the cross-talk of different phytohormones controlling the stem cells of the root in response to chilling stress. Recently, our lab found that ABA-induced SAPK10 can phosphorylate bZIP72 and enhance the latter's binding ability to the promoter of AOC, an essential gene in the jasmonic acid (JA) synthesis pathway, thereby inducing AOC expression and increasing endogenous JA levels, finally inhibiting seed germination [28]. Based on these regulators were all induced by cold treatment, the reduced trehalose and proline contents in the mutant lines under chilling conditions, and the interaction among the three regulators, we established a novel pathway 'SAPK10-ABF1-TPS2′ in trehalose-mediated rice cold tolerance in this study ( Figure 12). Since ABA and trehalose enhance cold stress tolerance, SAPK10 and ABF1 are ABA signalling positive regulators, and TPS2 is the key enzyme involved in trehalose biosynthesis [25,29,35,39]. Further study will focus on the relationship between trehalose and ABA homeostasis in response to chilling stress, which would deepen our understanding of the regulatory mechanism acclimating to low temperatures.

Vector Construction and Plant Transformation
For overexpression of ABF1, SAPK10 and TPS2, the coding sequence (CDS) was fused into the binary vector PU1301 driven by the maize ubiquitin promoter using the KpnI and BclI sites, respectively. abf1, sapk10 and tps2 mutant lines were generated by CRISPR/Cas9 system as described previously [42]. Briefly, annealed double-strand oligos 20 nt in the length of the gDNA sequence were ligated into the pYLCRISPR/Cas9-MH vector. All the vectors were transformed into Nipponbare (Oryza sativa L. ssp. japonica cv Nipponbare) callus using the Agrobacterium-mediated transformation method [43]. The primers used were listed in Table S1.

Chilling Stress Treatment
Chilling stress treatment at the seedling stage was conducted as described previously [44] with a few modifications. Briefly, 7-day-old rice seedlings grown using rice nutrient solution [45] were transferred to a growth chamber at 4 • C (14 h: 10 h, light: dark, approximately 200 µmol m −2 s −1 photon density) for 5 days. Then the seedlings were restored for 7 days in a growth chamber at 28 ± 2 • C (14 h: 10 h light: dark, approximately 200 µmol m −2 s −1 photon density). For the short-time chilling stress treatment, samples were taken at 0, 1, 3, 6.12 h (light, 4 • C), respectively.

Yeast Hybrid Assay
Yeast two-hybrid assay was performed according to Matchmaker™ Gold Yeast Two-Hybrid System (Clontech, CA, USA). The CDS of ABF1 was cloned into the pGADT7 plasmid at the EcoRI and PstI sites. The CDS of SAPK10 was ligated to the pGBKT7 vector at the EcoRI and BamHI sites. The recombinant vectors were co-transformed into the yeast strain Y2H Gold (Clontech) and selected on a synthetic medium lacking leucine, tryptophan, histidine and Adenine with 0.04 mg mL −1 X-α-Gal and 100 ng mL −1 Aureobasidin A added.
Yeast one-hybrid assay was performed according to Matchmaker™ One-Hybrid System (Clontech, CA, USA). The CDS of ABF1 was cloned into the pB42AD plasmid at the EcoRI and XhoI sites. The 2 kb promoter of TPS2 was ligated to the pLacZ2µ vector at the KpnI and XhoI sites. The recombinant vectors were co-transformed into the yeast strain EGY48 and selected on a synthetic medium lacking tryptophan and uracil with 1% raffinose, 1× BU salts, 80 mg/L X-Gal, and 2% galactose added. The primers used were listed in Table S1.

BiFC Assay
For Bimolecular Fluorescence Complementation (BiFC) assay, the CDS of ABF1 was cloned into pDOE plasmid at the BspEI site, and the CDS of SAPK10 was fused to the pDOE at the BamHI site as described previously [46]. The recombinant vectors were then transformed into Agrobacterium strain EHA105, and infiltrated into tobacco leaves. Fluorescence signals were observed using a Zeiss LSM710 confocal laser-scanning microscope (Carl Zeiss AG, Jena, Germany) after 72 h of infiltration. The primers used were listed in Table S1.

Coimmunoprecipitation (Co-IP) Assays
The CDS of SAPK10 was digested with BamHI and cloned into the proUbi-FLAG vector, and the CDS of ABF1 was digested with KpnI and XbaI and cloned into pCAMBIA1300-35S-GFP vector, respectively. SAPK10-FLAG was transiently co-expressed with an empty green fluorescent protein (GFP) or ABF1-GFP in tobacco leaves by Agrobacterium infiltration. The total protein of cotransformed tobacco leaves was extracted by extraction buffer (25 mM Tris-HCl, pH 7.4, 150 mM NaCl,1 mM EDTA, 1% NonidetP-40, 5% glycerol, 1 mM PMSF, 20 µM MG132 and 1× Roche protease inhibitor cocktail (Roche, Basel, Switzerland)). After centrifugation twice (12,000× g for 10 min each time), the supernatant was incubated with anti-FLAG M2 magnetic beads (Sigma-Aldrich, St. Louis, MO, USA) at 4 • C for 2 h. Then the beads were washed five times with washing buffer (50 mM Tris, pH 7.5, 150 mM NaCl, 0.2% Triton X-100, 1 mM PMSF, and 1× Roche protease inhibitor cocktail). The immunoprecipitated proteins were suspended in 50 µL of 1× PBS and 10 µL of 6× SDS protein loading buffer, boiled for 5 min, and resolved on 10% acrylamide gels. Individual bands were detected using Supersignal West Pico Chemiluminescent Substrate and the ChemDoc TM Touch Imaging system. The dilution for anti-FLAG (Sigma-Aldrich, St. Louis, MO, USA) and anti-GFP (Yeasen, Shanghai, China) antibodies was 1: 5000. The primers used were listed in Table S1.

Electrophoretic Mobility Shift Assay (EMSA)
GST-ABF1 was expressed in Transetta (DE3) chemically competent cell (Transgen, Beijing, China), and purified using the GST-Sefinose TM Kit (Sangon Biotech, Shanghai, China). The probes in a length of 59 bp in the TPS2 promoter were commercially synthesized by Tsingke (Tsingke Biotech, Hangzhou, China) and labeled by an EMSA Probe Biotin Labeling Kit (Beyotime, Shanghai, China). Briefly, an equal amount of purified recombinant proteins was pre-incubated with EMSA/gel-shift binding buffer, and then incubated with 20 fmol labeled probes or nonlabeled competitive DNA probes. Then the incubated samples were separated by electrophoresis on 6% acrylamide gels. The labeled DNA probes were detected by the LightShift Chemiluminescent EMSA Kit (Thermo, Waltham, MA, USA). The primers used were listed in Table S1.

Luciferase Transient Transcriptional Activity Assay (LUC)
The CDS of ABF1 and SAPK10 were fused into 'None' vectors as effectors by using the BamHI and EcoRI sites, and the 2 kb promoter of TPS2 was ligated into 190fLUC vector as a reporter by using the HindIII site, respectively. The plasmids were co-transformed into rice protoplasts as described previously [47]. The luciferase activities were detected by the Dual-Luciferase Reporter Gene Assay Kit (Beyotime, Shanghai, China) using the Promega GLOMAX system (Promega, Madison, WI, USA) according to the manufacturer's instructions. The relative luciferase activity was defined as the ratio between fLUC and rLUC (fLUC/rLUC). AtUbi3:rLUC was adopted as the internal control. The primers used were listed in Table S1.

Chromatin Immunoprecipitation-Quantitative PCR (ChIP-qPCR)
The ChIP-qPCR was carried out as described previously [48]. Briefly, chromatin was isolated from 3 g crosslinked leaves of the wild-type under chilling stress at 4 • C for 12 h. Isolated chromatin was sonicated for DNA fragmentation ranging from 200 to 700 bp. Subsequently, sonicated DNA/protein complexes were immune-precipitated with polyclonal rabbit ABF1 antibody against the amino acid residues 149 to 162 (C-KGQEEAPDGSDGPR-COOH) of ABF1, which was commercially synthesized and affinity-purified by Genescript (Genescript, Shanghai, China). After reverse cross-linking and proteinase K treatment, the immunoprecipitated DNA was purified with phenol/chloroform. Then the immunoprecipitated and input DNA were used for quantitative polymerase chain reaction (qPCR), respectively. The primers used were listed in Table S1.

Trehalose Extraction and Quantification
Briefly, the leaves of 7-day-old seedlings (0.1 g) before or after chilling stress at 4 • C for 72 h were homogenized in 5 mL of 80% (v/v) hot ethanol for 20 min and centrifuged. Then the supernatants were collected for trehalose extraction and quantification using the Trehalose Content Kit (Grace Biotechnology, Suzhou, China) according to the manufacturer's instructions. The absorbance of trehalose was measured at 620 nm.

Proline Determination
Briefly, 0.1 g leaves of 7-day-old seedlings before or after chilling stress at 4 • C for 5 days were immersed in 5 mL 3% (v/v) hot sulfosalicylic acid solution for 30 min, during which the samples were constantly shaken and cooled to room temperature to prepare crude proline extract. A total of 1 mL crude proline extract was added to 2 mL reaction buffer containing 1 mL of 2.5% acidic ninhydrin (60 mL glacial acetic acid, 16.4 mL phosphoric acid, 36.6 mL deionized water, and 2.5 g ninhydrin) and 1 mL glacial acetic acid, then boiled for 30 min with continuous shaking before cooling to room temperature. Finally, 0.5 mL toluene was used to extract 0.75 mL reaction solution with vortexing for 1 min, then incubated for 30 min. The absorbance of proline was measured at 520 nm by a SpectraMax i3x Multi-Mode Microplate Reader (MOLECULAR DEVICES, Sunnyvate, CA, USA).

Membrane Ion Leakage Measurement
Membrane Ion Leakage was determined as described previously with a few modifications [49]. Leaves of 7-day-old seedlings weighed at 0.3 g before or after chilling stress at 4 • C for 5 days were immersed in 30 mL deionized water and shaken at 120 rpm for 4 h. The membrane ion leakage L 1 was measured using a DDS-307A conductivity meter (LeiCi, Hangzhou, China). Then, the leave samples were incubated in boiling water for 30 min. After cooling down to room temperature, the membrane ion leakage L 2 of the corresponding leaves was measured. The formula EL = L 1 /L 2 × 100% was used to calculate the ion leakage of leaves.

Conclusions
In this study, we characterized the function of ABF1 positively participating in rice chilling tolerance. ABF1 over-expression lines strengthened chilling tolerance with reduced ion leakage and increased proline contents, while abf1 mutant exhibited the opposite tendency. Furthermore, SnRK2 protein kinase SAPK10 directly interacts with and phosphorylates ABF1, and strengthens the DNA-binding ability of ABF1 to a member of Trehalose-6-phosphate synthase TPS2, thereby facilitating the transcription of TPS2 and endogenous trehalose biosynthesis under chilling stress. In a word, this study revealed a novel chilling tolerance strategy mediated by 'SAPK10-ABF1-TPS2' through fine-turning trehalose homeostasis.