A bHLH transcription factor, SlbHLH96, promotes drought tolerance in tomato

Abstract Drought stress caused by water deficit reduces plant productivity in many regions of the world. In plants, basic helix–loop–helix (bHLH) transcription factors regulate a wide range of cellular activities related to growth, development and stress response; however, the role of tomato SlbHLHs in drought stress responses remains elusive. Here, we used reverse genetics approaches to reveal the function of SlbHLH96, which is induced by drought and abscisic acid (ABA) treatment. We found that SlbHLH96 functions as a positive regulator of drought tolerance in tomato. Overexpression of SlbHLH96 in tomato improves drought tolerance by stimulating the expression of genes encoding antioxidants, ABA signaling molecules and stress-related proteins. In contrast, silencing of SlbHLH96 in tomato reduces drought tolerance. SlbHLH96 physically interacts with an ethylene-responsive factor, SlERF4, and silencing of SlERF4 in tomato also decreases drought tolerance. Furthermore, SlbHLH96 can repress the expression of the ABA catabolic gene, SlCYP707A2, through direct binding to its promoter. Our results uncover a novel mechanism of SlbHLH96-mediated drought tolerance in tomato plants, which can be exploited for breeding drought-resilient crops.


Introduction
Plant productivity is significantly limited by various environmental challenges, especially drought stress and soil salinity [1]. Drought is one of the most detrimental abiotic stress conditions for plant growth and development, and severely threatens sustainability in agriculture [2]. Drought inf luences many aspects of plant physiology and causes abnormal changes in cellular processes [3]. In particular, drought stress causes injuries to biological membranes, which significantly elevate ion leakage from plant cells [4][5][6]. Drought stress also induces the accumulation of excessive reactive oxygen species (ROS), which can cause oxidative damage [7]. Nonetheless, the ROS H 2 O 2 also acts as a signaling molecule and is involved in regulating stomatal closure, activities of ion channels, and specific stress responses [8]. Drought stress induces the biosynthesis and signaling of the phytohormone abscisic acid (ABA), which triggers a variety of adaptive responses in plants [9]. Under stress conditions, ABA increases the activity of enzymes such as superoxide dismutase (SOD), peroxidase (POD), catalase (CAT), which function in ROS scavenging [10]. In the ABA biosynthetic pathway, the 9-cisepoxycarotenoid dioxygenase (NCED) genes encode key enzymes involved in the speed-limiting step of ABA biosynthesis [11,12]. So far, three NCED genes have been isolated and analyzed in tomato [13][14][15]. For ABA catabolism, the CYP707A1, A2, A3, and A4 genes, encoding 8'-hydroxylases, play a pivotal role in ABA oxidation [16][17][18]. The principal ABA signaling pathway consists of the primary ABA receptor proteins, such as PYR/PYL/RCAR, protein phosphatases of type 2C (PP2Cs) from group A, and SNF1-related kinase 2 (SnRK2) [19][20][21][22].
The basic helix-loop-helix (bHLH) family is the second largest family of transcription factors in plants. Members of the bHLH transcription factor family contain two highly conserved and functionally different domains, such as the basic domain and the HLH domain. The basic domain, which is located at the Nterminal end of the bHLH structure, is responsible for binding to an E-box sequence present in the promoter regions of the target genes. The HLH domain, which is located at the C-terminal end of the bHLH structure, is important for protein-protein interactions. These protein complexes work at the E-box region to regulate their target genes' transcriptional activity to control a variety of developmental processes. Plant bHLH transcription factors are involved in a wide range of cellular activities related to plant growth and development. For example, the bHLH members regulate seed germination [23], f lowering time [24], fruit ripening [25], trichome formation [26], and root hair formation [27]. Furthermore, bHLH transcription factors have a role in plant responses to abiotic stressors such as drought, salt, and cold. Drought, salt, and osmotic stress responses are positively regulated by Arabidopsis AtbHLH122. Knockout of AtbHLH122 leads to increased sensitivity to salt and osmotic stress, whereas overexpression of AtbHLH122 improves plant performance under drought, salt, or osmotic stress conditions [28]. Drought tolerance is improved by overexpression of OsbHLH148 in rice plants. The possible mechanism is interaction of OsJAZ1 with OsbHLH148 to activate the jasmonate signaling pathway [29]. In Arabidopsis and cucumber seedlings, overexpression of CsbHLH041 improves salt and ABA tolerance [30]. Likewise, overexpression of SlbHLH22 increases drought and salt tolerance in tomato [31]. Ectopic expression of maize ZmbHLH55 in Arabidopsis improves salt stress tolerance, which is associated with higher ascorbic acid levels in the transgenic plants [32]. In apple, MdbHLH3 improves cold resistance by elevating anthocyanin accumulation via transcriptional regulation of the anthocyanin biosynthetic genes MdDFR and MdUFGT under cold conditions [33]. Furthermore, PtrbHLH regulates PtrCAT expression by direct binding to its promoter and overexpression of PtrbHLH in transgenic pummelo (Citrus grandis) improves cold tolerance [34].
Ethylene-responsive factors (ERFs) contain an AP2 DNAbinding domain, and this protein family is widely found in higher plants but is absent in mammals, fungi, and yeast [35][36][37][38]. Members of the ERF protein family are shown to play key roles in many abiotic stress responses in plants. For example, overexpression of the tomato ERF transcription factor SlTSRF1 in rice improves drought tolerance by upregulating the expression of stress-responsive genes [39]. In addition, overexpression of OsERF19 in rice plants enhances resistance to salt stress while causing an ABA hypersensitivity phenotype [40]. Overexpression of OsERF115 improves heat tolerance in rice plants at the vegetative stage [41]. Furthermore, overexpression of PagERF16 increases salt sensitivity in poplar [42]. In Arabidopsis, heterologous overexpression of SlERF84 increases drought and salt stress resistance [43]. Overexpression of SlERF5 in tomato plants shows similar effects [44].
Tomato (Solanum lycopersicum L.) is one of the world's most commonly grown and commercially significant vegetable crops [45]. Tomato growth, development, and productivity are severely affected by various abiotic stresses, such as salinity, drought, chilling, and high temperatures [46]. Therefore, improving abiotic stress tolerance is increasingly vital for sustainable tomato production. In this study, we used multiple genetics approaches and revealed that SlbHLH96 is vital for drought tolerance in tomato plants. Our results show that overexpression of SlbHLH96 in tomato improves drought tolerance, whereas silencing of SlbHLH96 in tomato reduces drought tolerance. Furthermore, we showed that SlbHLH96 physically interacts with SlERF4, and silencing of SlERF4 in tomato decreases drought tolerance. SlbHLH96 binds to the promoter of SlCYP707A2 to downregulate its expression to fine-tune the expression of ABA response-related genes.

Identification and characterization of SlbHLH96 gene in tomato
From RNA-seq experiments (accession numbers SAMN14996375-14996413), we found that SlbHLH96 is upregulated by drought treatment in tomato (Supplementary Data Figs S1 and S2A), suggesting its potential roles in drought stress responses. SlbHLH96 encodes a protein with 441 amino acid residues having a molecular weight of 48.74 kDa. The theoretical isoelectric point (pI) of this protein is 6.54, with an instability index of 53.40 and an aliphatic index of 64.81. Conserved domain analysis showed that SlbHLH96 possesses the typical structure of the bHLH transcription factors. Phylogenetic analysis suggested that SlbHLH96 is closely related to potato StbHLH117 (Supplementary Data Fig.  S2B). SlbHLH96 was highly expressed in leaf and f lower tissues while its expression was relatively low in root and fruit tissues (Supplementary Data Fig. S3). To investigate the subcellular localization of SlbHLH96, we transiently expressed the SlbHLH96-GFP fusion protein in tobacco leaves. Our results showed that GFP protein driven by the 35S promoter spread throughout the cell, whereas the SlbHLH96-GFP fusion protein was only observed in the nucleus (Supplementary Data Fig. S4).

SlbHLH96 expression is responsive to multiple abiotic stresses and hormone treatments
We examined the expression profile of SlbHLH96 under different abiotic stress and hormone treatments. SlbHLH96 expression was substantially induced by low water potential treatments imposed by infusion of polyethylene glycol (PEG; average molecular weight 8000) in the growth medium, and this is consistent with our RNA-seq results from drought-treated plants grown in soil ( Fig. 1A and Supplementary Data Fig. S2A). Similar expression patterns of SlbHLH96 were observed after ABA treatment (Fig. 1E). These results suggest that SlbHLH96 may function in drought stress responses in an ABA-dependent manner. The expression of SlbHLH96 appeared to be responsive to other abiotic stresses or hormones ( Fig. 1B-D and F-I). However, its expression levels under these conditions were much lower compared with those under drought or PEG treatment. These results indicate that SlbHLH96 may play a major role in drought stress responses through an ABA-dependent pathway.

Overexpression of SlbHLH96 in tomato improves drought tolerance
To investigate the biological significance of SlbHLH96 in drought tolerance, we produced tomato plants overexpressing SlbHLH96 in the 'Ailsa Craig' (AC) genetic background (wild type). The expression levels of SlbHLH96 in two independent T 2 homozygous transgenic lines were examined by qRT-PCR analysis and the results revealed that the transcript abundance of SlbHLH96 in the OE-SlbHLH96-2 and OE-SlbHLH96-17 plants was ∼60-fold and 55-fold that of the AC plants, respectively (Fig. 2B). We then examined the drought tolerance of the 30-day-old soil-grown SlbHLH96 overexpression lines and AC plants. Both genotypes were subjected to continuous drought treatment for 12 days. At the beginning of the experiment, the overexpression plants showed a phenotype similar to that of the AC plants ( Fig. 2A). However, after 5 days of drought the AC plants started to display a leaf wilting phenotype while the SlbHLH96 overexpression plants were essentially healthy. Although both genotypes became wilted at the end of 12 days of drought treatment, it was obvious that the AC plants displayed more severe drought-induced damage (such as leaves with drooping petioles) than the overexpression plants ( Fig. 2A). All the plants were then re-irrigated for recovery. After recovery for 7 days, ∼45-53% of the wilted SlbHLH96 overexpression plants survived, whereas <20% of the wilted AC plants survived ( Fig. 2A and C). In addition, the SlbHLH96 overexpression plants developed more vigorous root systems than the AC plants during the drought and the recovery period ( Fig. 2G-I). We also examined stomatal aperture to determine whether the improved drought stress tolerance in the SlbHLH96 overexpression plants is related to the difference in stomatal movement. We found that the SlbHLH96 overexpression plants had much narrower stomatal apertures than the AC plants under drought stress ( Fig. 2D and E). Consistent with this observation, detached leaves from the SlbHLH96 overexpression plants showed a slower water loss rate than leaves from the AC plants (Fig. 2F). These findings indicate that overexpression of SlbHLH96 in tomato improves drought tolerance at least partly by minimizing water loss.

SlbHLH96 regulates the expression of genes involved in ABA biosynthesis, catabolism, and signal transduction
The increased expression level of SlbHLH96 under ABA treatment prompted us to examine whether the expression of genes involved in ABA biosynthesis, catabolism, and downstream signal transduction pathway was altered in the SlbHLH96 overexpression plants under drought conditions. The expression of SlNCED1, which encodes a key enzyme in ABA biosynthesis, increased in the SlbHLH96 overexpression plants under both control and drought conditions (Fig. 4A), whereas the expression of SlCYP707A2, which encodes a major ABA 8'-hydroxylase essential for ABA catabolism, decreased in the SlbHLH96 overexpression plants under both control and drought conditions (Fig. 4B). In addition, we showed that the expression of one of the ABA receptors, SlPYL7, increased in the SlbHLH96 overexpression plants under both control and drought conditions while a substantial reduction in the expression of SlPP2C1 was found in the SlbHLH96 overexpression plants under both control and drought conditions ( Fig. 4C and E). Furthermore, the expression of SlPP2C4 decreased and the expression of SlSnRK2.6 increased in the SlbHLH96 overexpression plants under drought conditions ( Fig. 4D and F). In addition to the above changes, the reduced expression level of SlCYP707A2 in the SlbHLH96 overexpression plants suggests that SlbHLH96 might act as a negative regulator for ABA catabolism. We then analyzed ABA levels using LC-MS/MS in the SlbHLH96 overexpression and AC plants. We found that ABA levels were much higher in the SlbHLH96 overexpression plants than in the AC plants under drought stress (Fig. 4N), and a higher ABA content usually resulted in improved drought resistance. The results suggest that altered expression of genes involved in ABA biosynthesis, ABA catabolism, and ABA signaling may contribute to the increased drought tolerance of the SlbHLH96 overexpression plants.

Expression profiles of stress-related genes in SlbHLH96 overexpression plants under drought stress
To uncover the potential molecular mechanisms underlying the improved tolerance of the SlbHLH96 overexpression plants to drought stress, we investigated the transcript levels of stressrelated genes, including SlDREB1, SlDREB2A, SlAREB1, SlSOD,

Silencing of SlbHLH96 in tomato reduces drought tolerance
To further reveal the essentiality of SlbHLH96 in basal drought tolerance, the expression of SlbHLH96 was suppressed by virusinduced gene silencing (VIGS) in tomato. We observed that SlPDS-silenced plants showed a photo-bleached phenomenon ( Supplementary Data Fig. S5). The expression of SlbHLH96 in the TRV2:SlbHLH96 plants significantly decreased by 85% (Fig. 5A (Fig. 5C, E, and F). Furthermore, we measured the activities of SOD and POD and found that their activities were substantially decreased in the TRV2:SlbHLH96 plants under drought stress ( Fig. 5G and H). These results indicate that silencing of SlbHLH96 results in drought sensitivity in tomato plants.
We subsequently determined the expression levels of genes involved in ABA biosynthesis, catabolism, and signal transduction in the SlbHLH96-silenced and TRV2:00 control plants. The qRT-PCR analysis revealed that SlNCED1 expression was lower and SlCYP707A2 expression was significantly higher in the TRV2:SlbHLH96 plants under drought stress ( Fig. 6A and B). In addition, we observed a reduction in the expression of SlPYL7 and SlSnRK2.6 in the TRV2:SlbHLH96 plants under drought stress ( Fig. 6C and D). However, upregulated expression of SlPP2C1 and SlPP2C4 was detected in the TRV2:SlbHLH96 plants under drought stress ( Fig. 6E and F). Finally, we analyzed the expression of some stress-and antioxidant-related genes and found that their expression levels were lower in the SlbHLH96-silenced plants than in TRV2:00 control plants under drought stress (Fig. 6G-M).

SlbHLH96 interacts with SlERF4
To identify proteins that interact with SlbHLH96, a bioinformatics prediction was performed using STRING (https://cn.string-db. org/). This in silico analysis showed a possibility that SlbHLH96 could interact with SlERF4. SlERF4 is ubiquitously expressed in all tissues, with slightly less expression in unopened f lower buds, fully opened f lowers, and ripening fruits at the breaker stage (Supplementary Data Fig. S3). The transcriptional activation activity of SlbHLH96 was evaluated using a GAL4 activation system in yeast. Our results suggest that SlbHLH96 has self-activation activity in yeast, and the C-terminal segments of SlbHLH96 (SlbHLH96-C, SlbHLH96-C 1, and SlbHLH96-C 2), including the conserved bHLH domain, do not display the self-activation activity (Fig. 7A). It is possible that the amino acid residues at 101-199 from the N-terminal end of SlbHLH96 confer the self-activation activity because the SlbHLH96-C 3 segment still has self-activation activity compared with the SlbHLH96-C 2 segment. Truncated SlbHLH96 (SlbHLH96-C 2) was able to interact with SlERF4 in yeast (Fig. 7B). Bimolecular f luorescence complementation (BiFC) assays were performed to confirm the direct interaction between SlbHLH96 and SlERF4 in tobacco plants. Co-expression of SlbHLH96-cYFP and SlERF4-nYFP generated f luorescent signals in the nucleus, where both these two transcription factors are localized (Fig. 7C). The pull-down assay and split-luciferase assay also confirmed the interaction between full-length SlbHLH96 and SlERF4 ( Fig. 7D and E). In addition, the expression levels of SlERF4 were higher in the SlbHLH96 overexpression plants than in the AC plants under control conditions and drought treatment (Fig. 4J). In contrast, the expression of SlERF4 was reduced in the SlbHLH96silenced plants under control conditions and drought treatment ( Fig. 6J). These results suggest that SlbHLH96 may function as a positive regulator for SlERF4 expression.

Silencing of SlERF4 in tomato decreases tolerance to drought stress
A previous study showed that SlERF4 antisense plants exhibited salt stress-dependent growth inhibition [47]. However, the function of SlERF4 in the response to drought stress in tomato remains unknown. A particular 300-bp sequence of SlERF4 was selected to knock down SlERF4 following a VIGS protocol. Our qRT-PCR analysis revealed that the expression of SlERF4 was significantly reduced by VIGS in the TRV2:SlERF4 tomato plants (Fig. 8A). Compared with the control (TRV2:00) plants, SlERF4 knockdown (TRV2:SlERF4) plants were sensitive to drought stress simulated by 15% PEG8000 (Fig. 8B). The TRV2:SlERF4 plants showed a higher MDA content than the TRV2:00 plants (Fig. 8D). In addition, ROS assay showed that the TRV2:SlERF4 plants accumulated more O 2 ·− and H 2 O 2 than the TRV2:00 plants (Fig. 8C, E, and F). Consistent with this observation, SOD and POD activities were lower in the TRV2:SlERF4 plants (Fig. 8G-H). We subsequently observed that the transcript levels of some stress-and antioxidant-related genes were significantly lower in the SlERF4 knockdown plants under drought stress (Fig. 8I-R).

SlbHLH96 can repress SlCYP707A2 expression through direct binding to cis-elements in its promoter
A previous study showed that AtbHLH122 can bind to the G-box/ E-box in the AtCYP707A3 promoter [28]. SlCYP707A2 was identified as the closest homolog to AtCYP707A3 (75.05% similarity at the amino acid level). SlCYP707A2 is expressed at a relatively low abundance in all tissues in tomato ( Supplementary Data Fig.  S3). The consensus cis-elements (one G-box and three E-boxes) were found in the putative promoter region of SlCYP707A2 (Supplementary Data Fig. S6A). AtCYP707A3 and SlCYP707A2 share ∼40% sequence similarity at the DNA level in their putative promoter regions (Supplementary Data Fig. S6B). To examine whether SlbHLH96 can repress the transcription of SlCYP707A2, the dualluciferase reporter assay was performed in tobacco plants. The dual-luciferase assay revealed that SlbHLH96 can repress the activity of the SlCYP707A2 promoter. After mutating all three E-boxes and one G-box, SlbHLH96 could not repress the activity of the SlCYP707A2-mut promoter (Fig. 9A-C). Furthermore, SlbHLH96 was able to bind to the SlCYP707A2 promoter fragments that contained the cis-elements determined by yeast-one hybrid (Y1H) assays (Fig. 9D). The electrophoretic mobility shift assay (EMSA) further confirmed that SlbHLH96 could directly target the SlCYP707A2 promoter by binding to the E-box and G-box cis-elements (Fig. 9E). The signal was reduced when an unlabeled SlCYP707A2 probe was introduced to the system as a cold probe (Fig. 9E). Collectively, these results indicate that SlbHLH96 can repress SlCYP707A2 expression through direct binding to the cis-elements in its promoter. Furthermore, we determined whether SlERF4 could regulate the expression of SlCYP707A2. The dual-luciferase assay revealed that SlERF4 could not regulate the expression of SlCYP707A2, but the interaction between SlERF4 and SlbHLH96 enhanced the inhibitory effect of SlbHLH96 on the expression of SlCYP707A2 (Supplementary Data Fig. S7). Previous studies showed that bHLH proteins CsbHLH18 and PtrbHLH could bind and regulate antioxidant genes [34,48,49], but our results indicated that SlbHLH96 could not regulate the antioxidant enzyme genes SlCAT1 and SlPOD in tomato ( Supplementary  Data Fig. S8).

Discussion
In recent years, a significant reduction in crop productivity due to drought stress has emerged as a critical issue for the sustainability of global agriculture. Numerous investigations have demonstrated that overexpression of bHLH transcription factors can generate drought resistance in diverse plant species [28,29,31]. Nevertheless, few tomato bHLH proteins have been reported to play vital roles in drought responses. Herein, we characterized a bHLH transcription factor gene, SlbHLH96, which is responsive to drought stress and ABA treatment. Overexpression of SlbHLH96 enhanced drought resistance, while silencing of SlbHLH96 in tomato reduced drought tolerance, which was associated with ROS metabolism. The AP2/ERF transcription factor family includes DREB proteins as a subfamily. DREB genes have been implicated in drought stress responses in a variety of plant species [50][51][52]. SlAREB1 is a bZIP transcription factor that belongs to the AREB/ABF subfamily, and it confers drought and salt stress tolerance in tomato [53]. In the current study, the expression of stress-related genes (SlDREB1, SlDREB2A, and SlAREB1) increased significantly in the SlbHLH96 overexpression plants under drought stress, while the expression of these stress-related genes decreased significantly in the SlbHLH96-silenced plants. ABA is sensed by the PYL ABA receptor proteins [20,21]. In this study, SlPYL7 expression increased in the SlbHLH96 overexpression plants under both control and drought conditions while downregulated expression of SlPYL7 was detected in the SlbHLH96-silenced plants under drought stress. In Arabidopsis, AtPYL9 promotes drought resistance and leaf senescence [54]. In comparison with wild-type plants, SlPYL9 overexpression lines showed increased drought tolerance, but SlPYL9-RNAi lines showed weak tolerance [55]. Overexpression of cotton PYL10, PYL12, and PYL26 independently in Arabidopsis improves tolerance to drought stress [56]. ZmPYL8 or ZmPYL9 overexpression in Arabidopsis increases drought resistance [57]. In this study, downregulated expression levels of SlPP2C1 and SlPP2C4 were found in the SlbHLH96 overexpression plants under drought conditions, while upregulated expression levels of SlPP2C1 and SlPP2C4 were detected in the SlbHLH96-silenced plants under drought stress. SlPP2C3 overexpression plants were found to be more drought-sensitive than wild-type plants, while SlPP2C3-RNAi plants showed a considerable increase in drought tolerance [58]. Compared with wild-type plants, SlPP2C1-RNAi transgenic lines showed improved drought tolerance [59]. In the case of ABA signaling, OsPP2C9 has a positive effect on plant growth but a detrimental effect on drought tolerance [60]. Wheat PP2C-a10 decreased drought tolerance of transgenic Arabidopsis [61].
In Arabidopsis, overexpression of ZmPP2C-A6 reduced drought tolerance [62]. In this study, increased SlSnRK2.6 expression was found mainly under drought stress in the SlbHLH96 overexpression plants, while downregulated expression of SlSnRK2.6 was detected in the SlbHLH96-silenced plants under drought stress. In transgenic Arabidopsis, overexpression of cucumber CsSnRK2.5 improves drought tolerance [63]. Overexpression of MpSnRK2.10 confers resistance to drought in apple [64]. Drought tolerance is severely diminished in the Arabidopsis srk2d/e/i triple mutant [65,66]. A previous bioinformatics prediction showed that SlbHLH96 is a non-G-box-binding protein [67]. Although SlbHLH132 is predicted as a non-DNA-binding protein, the EMSA result showed that SlbHLH132 is a G-box cis-element DNA-binding protein [68]. By Y1H, EMSA, and dual-luciferase analyses, we demonstrated that SlbHLH96 directly binds to cis-elements (E-box and G-box) in the SlCYP707A2 promoter region to downregulate its transcription. The increased level of endogenous ABA in the SlbHLH96 overexpression plants might be caused by the direct repression of SlCYP707A2 transcription by SlbHLH96. Improved ABA-inducible gene expression and increased drought tolerance are both seen in the atcyp707a3 mutant [17]. In sweet cherry, when PacCYP707A1 was silenced, fruits were more resistant to drought stress than control fruits [69].
Multiple functions of tomato SlERF4 have been reported. Compared with the wild type, SlERF4 knockdown tomato plants displayed a salt stress-sensitive phenotype [47]. SlERF4 is desumoylated by the Xanthomonas type III effector XopD, which suppresses ethylene responses and enhances pathogen growth. During Xcv infection, SlERF4 is essential for the activation of XopD-repressed genes [70]. Overexpression of ERF4-SRDX (chimeric dominant repressor version) causes a significant delay in ripening as well as increased climacteric ethylene production [71]. SlERF4 regulates the expression of SlIAA27, which controls ethylene and auxin signaling [72]. SlERF5 overexpression in tomato plants led to enhanced salt and drought stress resistance [44]. The evolutionary relationship between SlERF4 and SlERF5 is very close. SlERF4 has been functionally characterized under salt stress, disease resistance, fruit ripening, and auxin signaling. However, the function of SlERF4 in drought stress remains unclear. In this study, we demonstrated that SlbHLH96 physically interacts with SlERF4. The SlERF4 knockdown plants showed a higher MDA content than the control plants. Notably, MDA is a primary indicator of the peroxidation of membrane polyunsaturated fatty acids. Moreover, SOD and POD activities were higher in the control plants. Transcript levels of some stress-related genes and antioxidant-related genes were significantly lower in the SlERF4 Based on the results of this study, we proposed a working model for the function of SlbHLH96 under drought stress (Fig. 10). Brief ly, drought stress induces SlbHLH96 expression. SlbHLH96 directly binds to cis-elements in the SlCYP707A2 promoter and downregulates its transcription, leading to increased levels of endogenous ABA, which, in turn, regulates the expression of ABA response-related genes. Furthermore, SlbHLH96 physically interacts with SlERF4, and the SlbHLH96-SlERF4 complex controls the expression of genes encoding antioxidants and stress-related genes. The study unveils novel mechanisms by which SlbHLH96 confers drought tolerance to tomato plants, thus providing important clues for breeding drought-resistant crops.

Plant growth conditions
The tomato cultivar 'Alisa Craig' (AC) was used in this study and it also served as the transgene recipient. The plants were cultivated in growth chambers under a 16-h day (at 25 • C), 8-h night (at 22 • C) cycle and 80% relative humidity.

Abiotic stress and hormone treatments for gene expression analysis
Surface-sterilized AC seeds were planted on 1 / 2 MS (Murashige-Skoog) medium plates for germination. Seedlings of identical size were moved to 1 / 2 MS medium plates after 7 days. For treatment with cold stress, the medium plates were transferred to an illuminating incubator at 4 • C and sampled at 0, 1, 3, 6, 12, and 24 h. For heat treatment, the plates were placed in an illuminating incubator at 42 • C and sampled at 0, 2, 4, 6, 12, and 24 h. For NaCl treatment, seedlings of similar size were transferred to 1 / 2 MS medium plates supplemented with 200 mM NaCl and incubated for 0, 1, 3, 6, 12, or 24 h for sampling. For PEG treatment, seedlings of similar size were transferred to 1 / 2 MS medium plates infused with different concentrations of PEG (average molecular weight 8000) solutions to achieve low water potentials from −0.25 to −1.7 MPa and incubated for 12 h. To detect the expression of SlbHLH96 in response to exogenous hormones, treatments were performed as follows: 7-day-old seedlings with similar size grown on 1 / 2 MS medium were transferred to 1 / 2 MS medium plates supplemented with 0, 10 μM ABA, 10 μM IAA (indole-3-acetic acid), 10 μM GA 3 (gibberellic acid 3), 10 μM SA (salicylic acid), or 10 μM JA (jasmonic acid), and incubated for 0, 1, 3, 6, 12, and 24 h.

Subcellular localization of SlbHLH96
The full-length complete coding sequence (CDS) of SlbHLH96 without the stop codon was constructed into a 35S promoterdriven pCAMBIA2300-GFP vector, resulting in the 35S-SlbHLH96-GFP plasmid. Leaves of Nicotiana benthamiana plants were infiltrated with the Agrobacterium strain GV3101 harboring the 35S-SlbHLH96-GFP plasmid or the empty vector of pCAM-BIA2300-GFP (35S-GFP; as a control). Co-transformation of a red f luorescent protein (RFP) coupled with the nucleus marker mCherry made it possible to observe nuclei. After 48 hours, the f luorescence signals from the GFP protein expressed in the epidermal cells were observed with a BX53 (Olympus, Japan).

Transcriptional activation analysis in yeast
The CDS and truncation of SlbHLH96 were inserted into the pGBKT7 vector. The plasmids were inserted into Y2H-Gold, and were then grown on SD/−Leu and SD/−Leu/−Trp/−His medium at 30 • C for 3 days.

Tomato transformation
The full-length CDS of SlbHLH96 was amplified by PCR from the first-strand tomato cDNA synthesized with the SlbHLH96specific primer. Then, the SlbHLH96 CDS was constructed into the plant expression vector pBI121. Finally, the recombinant vector was introduced into tomato cultivar AC by tissue culture-based Agrobacterium-mediated stable transformation (strain GV3101).

RNA extraction and qRT-PCR analysis
Total RNA was isolated from AC tomato leaves using TRIzol (Tiangen, China). The cDNA was synthesized from the total RNA using the M-MLV Reverse Transcriptase kit (Vazyme, China). qRT-PCR reactions were performed with Tip Green SuperMix (Trans-Gen, China). The relative expression was calculated using the 2 − Ct method. The SlACTIN7 gene was used as a reference gene. Primer sequences in this study are listed in Supplementary Data Table 1.

Methods for physiological measurements
NBT and DAB staining assays were performed as previously described [73]. O 2 ·− content and H 2 O 2 content were determined using Solarbio detection kits (Solarbio, China). The relative electrolytic leakage was measured to assess injuries to biological membranes as described previously [74]. The MDA content and proline content were measured as previously described [74,75]. The activities of SOD and POD were measured as previously described [74].

Bimolecular fluorescence complementation assay
The full-length CDS of SlbHLH96 was cloned into pSPYCE vector to fuse with half of a YFP protein (SlbHLH96-cYFP). The full-length CDS of SlERF4 was cloned into a pSPYNE vector to fuse with half of a YFP protein (SlERF4-nYFP). The recombinant plasmids were transformed into GV3101, which were then used to co-infiltrate N. benthamiana leaves. After 48 hours, f luorescence was observed with the BX53 (Olympus, Japan).

Yeast two-hybrid assay
The full-length SlERF4 and truncation of SlbHLH96 were introduced into the pGADT7 and pGBKT7 vectors, respectively. The plasmids were introduced into yeast strain AH109 and grown on −Leu/−Trp/−His/−Ade medium (Coolaber, China).

GST pull-down
Full-length SlbHLH96 and SlERF4 were inserted into the pMAL-c5X and pET42a vectors, respectively. The fusion proteins were purified with Amylose resin (NEB, USA) and Glutathione resin (GenScript, China), respectively. The GST pull-down assays were performed according to the MagneGST™ protein purification system User Manual (Promega, USA). The proteins were detected by western blotting with anti-MBP antibody and anti-GST antibody.

Split-luciferase assay
Full-length SlbHLH96 and SlERF4 were cloned into the pCAM-BIA1300-cLuc and pCAMBIA1300-nLuc vectors, respectively. The recombinant plasmids were transformed into Agrobacterium tumefaciens strain GV3101, and were then used to co-infiltrate N. benthamiana leaves. After 3 days, f luorescence was detected by a camera system (Lumazone Pylon 2048B, Princeton, USA).

Dual-luciferase assay
The CDS of SlbHLH96 and SlERF4 was cloned into the pGreen62-SK vector. The promoters of SlCYP707A2, SlCAT1, and SlPOD were introduced into the pGreen0800-LUC vector, respectively. The recombinant vectors were transformed into GV3101 (pSoup-19) and infiltrated into 4-week-old N. benthamiana leaves. The Dual-Luciferase ® kit (Promega, USA) was used for dual-luciferase assays.

Yeast one-hybrid assay
The promoter regions (P1 with one G-box and one E-box; P2 with 2 E-boxes) of SlCYP707A2 were inserted to pHis2 and transformed into the Y187 yeast strain. The recombined yeast strain was transformed with the SlbHLH96-pGADT7 plasmid and the empty pGADT7 plasmid, respectively. The interactions between SlbHLH96 and SlCYP707A2 promoter regions were indicated by the growth of the colony on SD/−Leu/−Trp/−His in the presence of 3-AT.

Electrophoretic mobility shift assay
The CDS of SlbHLH96 was cloned into pMAL-c5X to fuse with MBP. The SlbHLH96-MBP fusion protein was induced in Escherichia coli BL21 (DE3). EMSA was conducted as previously described [78].

Statistical analysis
Data are reported as the means ± standard deviation. Statistical significance was determined by one-way ANOVA (Tukey's test) using SPSS (version 26.0, USA). Variations were considered significant if P < .05. In some cases, significant differences in mean values, determined by Student's t-test, are indicated by asterisk(s) ( * P < .05; * * P < .01).