Positive Interaction between H2O2 and Ca2+ Mediates Melatonin-Induced CBF Pathway and Cold Tolerance in Watermelon (Citrullus lanatus L.)

Cold stress is a major environmental factor that detrimentally affects plant growth and development. Melatonin has been shown to confer plant tolerance to cold stress through activating the C-REPEAT BINDING FACTOR (CBF) pathway; however, the underlying modes that enable this function remain obscure. In this study, we investigated the role of H2O2 and Ca2+ signaling in the melatonin-induced CBF pathway and cold tolerance in watermelon (Citrullus lanatus L.) through pharmacological, physiological, and genetic approaches. According to the results, melatonin induced H2O2 accumulation, which was associated with the upregulation of respiratory burst oxidase homolog D (ClRBOHD) during the early response to cold stress in watermelon. Besides, melatonin and H2O2 induced the accumulation of cytoplasmic free Ca2+ ([Ca2+]cyt) in response to cold. This was associated with the upregulation of cyclic nucleotide-gated ion channel 2 (ClCNGC2) in watermelon. However, blocking of Ca2+ influx channels abolished melatonin- or H2O2-induced CBF pathway and cold tolerance. Ca2+ also induced ClRBOHD expression and H2O2 accumulation in early response to cold stress in watermelon. Inhibition of H2O2 production in watermelon by RBOH inhibitor or in Arabidopsis by AtRBOHD knockout compromised melatonin-induced [Ca2+]cyt accumulation and melatonin- or Ca2+-induced CBF pathway and cold tolerance. Overall, these findings indicate that melatonin induces RBOHD-dependent H2O2 generation in early response to cold stress. Increased H2O2 promotes [Ca2+]cyt accumulation, which in turn induces H2O2 accumulation via RBOHD, forming a reciprocal positive-regulatory loop that mediates melatonin-induced CBF pathway and subsequent cold tolerance.


Introduction
Plants are sessile organisms and, therefore, must withstand multiple environmental stresses throughout their life cycle. As a major environmental constraint, cold stress causes adverse effects on plant growth and development, threatening agricultural production worldwide [1]. Based on temperature and various physiological mechanisms, cold stress in plants is classified as chilling stress (temperatures below optimum but above 0 • C) and freezing stress (<0 • C) [2]. Watermelon (Citrullus lanatus L.) is an economically important crop globally. Its origin can be traced to tropical and subtropical regions of Africa. Watermelon production is threatened by its susceptibility to low temperatures [3]. Chilling temperatures adversely affect watermelon seedlings, including reduction of photosynthetic ability, oxidative damage, membrane dysfunction, and hormonal imbalance, leading to 3 of 19 miculite (3:1). The seedlings were cultivated in a growth chamber under photoperiod of 12/12 h (day/night), temperature of 28/18 • C (day/night), and a photosynthetic photo flux density (PPFD) of 600 µmol m −2 s −1 . Arabidopsis seeds of Atrbohd (SALK_120299) mutant with Columbia-0 (Col-0) genetic background were received from the Arabidopsis Biological Resource Center (https://www.arabidopsis.org/, accessed date 16 May 2019) [29]. After surface sterilization with 75% ethanol and 3% sodium hypochlorite, Arabidopsis seeds were sown on half-strength Murashige-Skoog (MS) medium containing 0.8% agar and 1.0% sucrose and cultured at 22 • C in a growth chamber under 16/8 h photoperiod.

Experimental Treatment
To evaluate the influences of melatonin, H 2 O 2 , or Ca 2+ on watermelon response to cold stress, the seedlings were sprayed with 150 µM melatonin [24], H 2 O 2 (0.04, 0.2, 1, 5 mM), 20 mM CaCl 2 [30,31], or distilled water (as the control). Melatonin (Sigma-Aldrich, St. Louis, MO, USA) was dissolved in ethanol and then diluted with distilled water at a ratio of 1:10,000 [ethanol:water; v:v]. After 12 h, the watermelon seedlings were exposed to chilling treatment at 4 • C for 48 h. The leaves were sampled at 3, 6, 12, and 24 h time points during chilling exposure to analyze the expression of CBF pathway genes and at 48 h for cold tolerance assay. After the initial experiments, 1 mM H 2 O 2 was selected for the subsequent experiments.
To determine the role of H 2 O 2 in melatonin-or Ca 2+ -induced cold tolerance, watermelon seedlings were pretreated with 100 µM diphenyleneiodonium (DPI, an inhibitor of NADPH oxidases, which catalyzes H 2 O 2 production) [32] 2 h before melatonin or Ca 2+ application. After 12 h, the seedlings were transferred to 4 • C for 48 h. To investigate the role of Ca 2+ in melatonin-or H 2 O 2 -induced cold tolerance, the watermelon seedlings were sprayed with 10 mM lanthanum chloride (LaCl 3 , a Ca 2+ channel blocker) [33] 2 h before melatonin or H 2 O 2 treatment. After 12 h, the seedlings were transferred to 4 • C for 48 h.
To explore the function of RBOHD in melatonin-or Ca 2+ -induced freezing tolerance in Arabidopsis, three-week-old wild-type or Atrbohd mutant Arabidopsis plants were pretreated with 10 µM melatonin [34] or 1 mM CaCl 2 [35]. After 12 h, the seedlings were subjected to freezing at −10 • C for 1 h and then recovered at 22 • C for 5 days [36]. The proportion of plants with green leaves was recorded to determine the survival rate. To determine the role of RBOHD in melatonin-or Ca 2+ -induced CBF pathway in response to cold, the Arabidopsis seedlings pretreated with melatonin or CaCl 2 were exposed to 4 • C for 24 h. Leaves were sampled at 3, 6, 12, and 24 h after 4 • C treatment.
Protoplasts from watermelon or Arabidopsis leaves were extracted and used to examine the effects of melatonin or H 2 O 2 on the accumulation of cytosolic free calcium ([Ca 2+ ] cyt ) in response to chilling stress. The protoplasts were incubated with Fluo-4 acetoxymethyl (AM) ester (a Ca 2+ -sensitive fluorescent dye) at 37 • C in the dark. After 30 min, the protoplasts loaded with Fluo-4/AM were treated with melatonin (10 µM), H 2 O 2 (100 µM), or a combination of melatonin and DPI (10 µM) and then placed at 4 • C for 5 min.

Cold Tolerance Assay
The maximum photosystem II quantum yield (Fv/Fm) was determined on the upper second fully expanded leaves of watermelon seedlings after 30 min of dark adaptation using a FluorCam fluorescence imaging system (SN-FC800-240; Photon Systems Instruments; Brno, Czech Republic) [37]. The relative electrical conductivity (REC) was determined as described by Zhou and Leul [38]. The level of lipid peroxidation in plant cells was assessed by determining malondialdehyde (MDA) contents using a 2-thiobarbituric acid (TBA) reaction [39].

Hydrogen Peroxide Content Assay
Hydrogen peroxide (H 2 O 2 ) content was determined as described by Willekens et al. [40]. In brief, leaf samples (0.5 g each) were ground in 5 mL of ice-cold 1 M HClO 4 . Next, the homogenate was centrifuged at 6000× g for 5 min at 4 • C and neutralized to pH 6.0-7.0 with 4 M KOH. After that, the homogenate was further centrifuged at 12,000× g for 5 min at 4 • C, and the supernatant was loaded on an AG1-X8 prepacked column (Bio-Red, Hercules, CA, USA) and eluted with 4 mL double-distilled water. The sample extract (800 µL) was added to the reaction mixture of 400 µL 100 mM potassium acetate buffer (pH 4.4) containing 4 mM 2,2 -azino-di (3-ethylbenzthiazoline-6-sulfonic acid), 400 µL deionized water, and 0.25 U of horseradish peroxidase (HRP). The absorbance at OD 412 was recorded to calculate H 2 O 2 content.

Protoplast Isolation and Measurement of [Ca 2+ ] cyt
Protoplasts were isolated as described previously with modifications [41,42]. Briefly, leaves from three-week-old watermelon or four-week-old Arabidopsis seedlings were cut into 0.5 mm wide strips. Watermelon leaf strips were digested with 10 mL enzyme solution containing 1.5% cellulose R10 and 0.3% macerozyme R10, whereas Arabidopsis leaf strips were digested with enzyme solution containing 1.0% cellulose R10 and 0.2% macerozyme R10. The digestion was performed for 3-4 h in the dark, after which the enzyme solutions were diluted with 10 mL W5 solution (pH 5.7) containing 2 mM MES, 154 mM NaCl, 125 mM CaCl 2 , and 5 mM KCl, then filtered through a 75 µm nylon mesh. The protoplasts were collected after centrifugation at 100× g for 5 min at 4 • C and re-suspended in 10 mL W5 solution.

RNA Extraction and qRT-PCR Assay
Total RNA was extracted from watermelon or Arabidopsis leaves using an RNA simple Total RNA kit (TIANGEN, Beijing, China). After extraction, the total RNA samples were treated with gDNase, then reverse-transcribed (1 µg per sample) to cDNA using a FastKing RT kit (TIANGEN, Beijing, China). The qRT-PCR assay was performed on a StepOnePlus TM Real-Time PCR System (Applied Biosystems, Carlsbad, CA, USA) using SYBR ® Premix ExTaq TM II (2×) kit (Takara, Tokyo, Japan). The gene-specific primers used for the qRT-PCR are listed in Table S1. The qRT-PCR amplification was conducted under the conditions reported by Li et al. [29]. β-actin or AtActin2 served as the internal control genes for the normalization of gene expression [11,44]. Relative gene expression was calculated as described by Livak and Schmittgen [45].

Phylogenetic Analysis
Arabidopsis AtRBOH and watermelon ClRBOHD protein sequences were downloaded from Arabidopsis Information Resource (https://www.arabidopsis.org/, accessed date 16 May 2019) and Cucurbit Genomics Database (CuGenDB, http://cucurbitgenomics.org/, accessed date 23 July 2020), respectively. Multiple sequence alignment of these RBOH protein sequences was performed via ClustalW with default parameters [46]. Phylogenetic analysis was conducted using the MEGA7.0.21 software. Based on the result of multiple sequence alignment, a phylogenetic tree was constructed using the Neighbor-Joining method and the parameters were Jones-Taylor-Thornton (JTT) matrix-based model and 1000 bootstraps [47].

Statistical Analysis
The experiments were performed in a completely randomized design. Each experiment was repeated thrice, and each replicate included at least 18 plants. The differences among treatments were determined via one-way ANOVA using SPSS statistics 19 (SPSS Inc., Chicago, IL, USA), followed by Tukey's test at p < 0.05.

The Requirement of H 2 O 2 for Melatonin-Induced CBF-Responsive Pathway and Chilling Tolerance in Watermelon
The effect of melatonin on H 2 O 2 accumulation was first examined to determine whether H 2 O 2 is necessary for melatonin function in watermelon response to chilling stress. The results showed no significant differences in H 2 O 2 accumulation between melatoninpretreated and control plants under optimum growth conditions ( Figure 1A). Chilling exposure induced an H 2 O 2 burst from 6 h; however, such induction was accelerated by melatonin. Pretreatment with melatonin triggered an H 2 O 2 burst from 1 h after chilling treatment. For instance, after chilling exposure at 3 and 6 h, H 2 O 2 contents in melatoninpretreated plants increased by 15.9% and 19.0%, respectively, compared to the control plants. However, H 2 O 2 contents in melatonin-pretreated plants were less than that in the control plants at 12 and 48 h after chilling treatment. The changes in ClRBOHD transcript levels showed similar trends with H 2 O 2 in response to melatonin or/and chilling stress. For example, the transcript levels of ClRBOHD in melatonin-pretreated plants were 0.5 and 5.2 fold higher than that in the control plants at 3 and 6 h, respectively, after chilling exposure.
Application of H 2 O 2 at optimum concentrations (0.04-5 mM) alleviated chillinginduced increases in REC and MDA. The most effective H 2 O 2 concentration was 1 mM ( Figure 1B). Notably, H 2 O 2 concentrations higher or lower than 1 mM attenuated the positive effect of H 2 O 2 on chilling tolerance. Similarly, melatonin application enhanced watermelon defense against chilling stress, as exhibited by the alleviation of leaf wilting, an increase in Fv/Fm, and a decrease in MDA content ( Figure 1C-E). However, pretreatment with DPI (an inhibitor of H 2 O 2 production, 100 µM) completely abolished the melatonininduced chilling tolerance. The CBF-responsive pathway plays a central role in plant defense against cold stress [9]. Chilling stress induced the expression of ClCBF1 and its regulons, including cold-responsive gene 47 (COR47), early responsive to dehydration 10 (ERD10), and cold induced gene 1 (KIN1). Importantly, transcript levels of the genes mentioned above were significantly higher in the melatonin-pretreated plants than in the control plants after chilling treatment. However, pretreatment with DPI prevented the melatonin-induced increases in the transcripts of CBF-responsive pathway genes.

Involvement of Ca 2+ Signal in Melatonin-and H 2 O 2 -Induced Chilling Tolerance in Watermelon
To investigate whether Ca 2+ signal is involved in melatonin-or H 2 O 2 -mediated cold tolerance, the [Ca 2+ ] cyt accumulation in watermelon protoplasts was first measured using the Fluo-4 AM ester as a fluorescent indicator of Ca 2+ . According to the results, chilling exposure induced the accumulation of [Ca 2+ ] cyt in watermelon protoplasts ( Figure 2). Notably, both melatonin and H 2 O 2 pretreatment significantly increased the cold-induced accumulation of [Ca 2+ ] cyt . For example, the fluorescence intensity of [Ca 2+ ] cyt in protoplasts pretreated with melatonin and H 2 O 2 increased by 46% and 32%, respectively, compared with the control protoplasts after chilling stress. However, DPI application prevented the melatonin-induced [Ca 2+ ] cyt accumulation under chilling stress. Cyclic nucleotide-gated channel 2 (CNGC2) encodes a plasma membrane cation channel that directs extracellular Ca 2+ into the cytosol [48,49]. In line with [Ca 2+ ] cyt accumulation, both melatonin and H 2 O 2 promoted the cold-induced ClCNGC2 upregulation; however, the effect was blocked by DPI pretreatment. These results demonstrate that H 2 O 2 participates in melatonin-induced [Ca 2+ ] cyt accumulation under chilling stress. . For (C-F), the seedlings were sprayed with 100 µ M diphenyleneiodonium (DPI, an inhibitor of H2O2 production) 2 h before MT treatment. After 12 h, the seedlings were subjected to 4 °C. Relative expression of genes in CK plants was set as 1.0. Data show the means of three replicates ± standard deviation (SD). The different letters denote significant difference at p < 0.05 according to Turkey's test.

Involvement of Ca 2+ Signal in Melatonin-and H2O2-Induced Chilling Tolerance in Watermelon
To investigate whether Ca 2+ signal is involved in melatonin-or H2O2-mediated cold tolerance, the [Ca 2+ ]cyt accumulation in watermelon protoplasts was first measured using the Fluo-4 AM ester as a fluorescent indicator of Ca 2+ . According to the results, chilling exposure induced the accumulation of [Ca 2+ ]cyt in watermelon protoplasts ( Figure 2). Notably, both melatonin and H2O2 pretreatment significantly increased the cold-induced ac- (F) The expression of ClCBF1 and its regulons, including cold-responsive gene 47 (ClCOR47), early responsive to dehydration 10 (ClERD10), and cold induced gene 1 (ClKIN1). For (C-F), the seedlings were sprayed with 100 µM diphenyleneiodonium (DPI, an inhibitor of H 2 O 2 production) 2 h before MT treatment. After 12 h, the seedlings were subjected to 4 • C. Relative expression of genes in CK plants was set as 1.0. Data show the means of three replicates ± standard deviation (SD). The different letters denote significant difference at p < 0.05 according to Turkey's test.
melatonin-induced [Ca 2+ ]cyt accumulation under chilling stress. Cyclic nucleotide-gated channel 2 (CNGC2) encodes a plasma membrane cation channel that directs extracellular Ca 2+ into the cytosol [48,49]. In line with [Ca 2+ ]cyt accumulation, both melatonin and H2O2 promoted the cold-induced ClCNGC2 upregulation; however, the effect was blocked by DPI pretreatment. These results demonstrate that H2O2 participates in melatonin-induced [Ca 2+ ]cyt accumulation under chilling stress. The relative expression of cyclic nucleotide-gated channel 2 (ClCNGC2) at 6 h after chilling exposure. The seedling leaves were sprayed with 150 µ M melatonin (MT) or 1 mM H2O2. After 12 h, the seedlings were transferred to 4 °C. To inhibit H2O2 production, the seedlings were sprayed with DPI at 100 µ M 2 h before melatonin application. Data are reported as means ± standard deviation (SD, n = 3). The different letters denote significant difference at p < 0.05 according to Turkey's test. CK, control.

Role of H 2 O 2 in Ca 2+ Signal-Induced Chilling Tolerance in Watermelon
To investigate the function of Ca 2+ in melatonin-induced H 2 O 2 accumulation in response to chilling stress, the effects of CaCl 2 and LaCl 3 on H 2 O 2 accumulation and ClRBOHD expression were analyzed. Like melatonin, CaCl 2 application accelerated the cold-induced upregulation of ClRBOHD and subsequent H 2 O 2 burst (Figure 4). After chilling exposure, the levels of H 2 O 2 and ClRBOHD transcripts in CaCl 2 -pretreated plants increased by 0.5 and 10.3 fold, respectively, at 6 h, while those in control plants were induced at 12 h. However, LaCl 3 pretreatment completely blocked melatonin-induced H 2 O 2 accumulation and ClRBOHD upregulation under chilling stress, suggesting that Ca 2+ signal mediates melatonin-induced ClRBOHD expression and H 2 O 2 accumulation in response to chilling stress.

Role of H2O2 in Ca 2+ Signal-Induced Chilling Tolerance in Watermelon
To investigate the function of Ca 2+ in melatonin-induced H2O2 accumulation i sponse to chilling stress, the effects of CaCl2 and LaCl3 on H2O2 accumulation and BOHD expression were analyzed. Like melatonin, CaCl2 application accelerated the induced upregulation of ClRBOHD and subsequent H2O2 burst (Figure 4). After chi exposure, the levels of H2O2 and ClRBOHD transcripts in CaCl2-pretreated plant creased by 0.5 and 10.3 fold, respectively, at 6 h, while those in control plants were ind at 12 h. However, LaCl3 pretreatment completely blocked melatonin-induced H2O2 a mulation and ClRBOHD upregulation under chilling stress, suggesting that Ca 2+ s mediates melatonin-induced ClRBOHD expression and H2O2 accumulation in respon chilling stress. Like melatonin and H2O2, CaCl2 alleviated cold-induced leaf wilting, decre Fv/Fm, and increased MDA content; however, these effects were abolished by DPI t ment ( Figure 5A-C). Moreover, CaCl2 promoted cold-induced upregulation of CBF p way genes, including ClCBF1, ClCOR47, ClERD10, and ClKIN1, but these effects blocked by DPI treatment ( Figure 5D). These findings suggest that H2O2 plays a vital f tion in Ca 2+ signal-induced CBF-responsive pathway and subsequent chilling toleran To block Ca 2+ influx, the leaves were pretreated with LaCl 3 at 10 mM 2 h before melatonin application. Seedlings sprayed with distilled water under temperatures of 28/18 • C (day/night) were set as control (CK). Data show the means of three replicates ± standard deviation (SD). The different letters denote significant difference at p < 0.05 according to Turkey's test.
Like melatonin and H 2 O 2 , CaCl 2 alleviated cold-induced leaf wilting, decreased Fv/Fm, and increased MDA content; however, these effects were abolished by DPI treatment ( Figure 5A-C). Moreover, CaCl 2 promoted cold-induced upregulation of CBF pathway genes, including ClCBF1, ClCOR47, ClERD10, and ClKIN1, but these effects were blocked by DPI treatment ( Figure 5D). These findings suggest that H 2 O 2 plays a vital function in Ca 2+ signal-induced CBF-responsive pathway and subsequent chilling tolerance.  ClKIN1). The plant leaves were sprayed with 20 mM CaCl2 for 12 h and then exposed to cold stress (CS) at 4 °C. To inhibit H2O2 production, the seedlings were pretreated with 100 µ M DPI 2 h before CaCl2 application. Seedlings sprayed with distilled water under temperatures of 28/18 °C (day/night) were set as control (CK). Relative expression of genes in CK plants was set as 1.0. Data show the means of three replicates ± standard deviation (SD). The different letters denote significant difference at p < 0.05 according to Turkey's test.  ClKIN1). The plant leaves were sprayed with 20 mM CaCl 2 for 12 h and then exposed to cold stress (CS) at 4 • C. To inhibit H 2 O 2 production, the seedlings were pretreated with 100 µM DPI 2 h before CaCl 2 application. Seedlings sprayed with distilled water under temperatures of 28/18 • C (day/night) were set as control (CK). Relative expression of genes in CK plants was set as 1.0. Data show the means of three replicates ± standard deviation (SD). The different letters denote significant difference at p < 0.05 according to Turkey's test.

Involvement of RBOHD in Melatonin-Induced [Ca 2+ ]cyt Accumulation and Freezing Tolerance in Arabidopsis
Phylogenetic analysis showed that watermelon ClRBOHD has high homology to Arabidopsis AtRBOHD with 79.3% similarity ( Figure 6A). ClRBOHD and AtRBOHD contain the same critical conserved domains, including NADPH_Ox domain (PF08414), EFh (IPR002048), FAD_binding_8 (PF08022), and NAD_binding_6 (PF08030) ( Figure 6B). The NADPH_Ox domain, which is found in respiratory burst NADPH oxidase proteins, shares 75.5% similarity to AtRBOHD and ClRBOHD. Thus, the Arabidopsis loss-of-function mutant Atrbohd was used to study the role of RBOHD in melatonin-induced [Ca 2+ ]cyt accumulation and cold tolerance. Atrbohd mutant protoplasts showed 38.6% and 50.8% less accumulation of [Ca 2+ ]cyt than wild-type (WT) protoplasts under normal and chilling conditions, respectively (Figure 7). Similar to watermelon protoplasts, melatonin significantly induced [Ca 2+ ]cyt accumulation in WT protoplasts under normal and chilling conditions. However, these effects were abolished by AtRBOHD mutation. Moreover, AtRBOHD knockout significantly increased Arabidopsis sensitivity to freezing at −10 °C and downregulated the expression of AtCBF1 and its targets, including AtCOR47, AtERD10, and AtKIN1 under chilling stress of 4 °C (Figure 8). After freezing exposure, the survival rate of Atrbohd mutant was 80.7% lower than the WT plants, suggesting that AtRBOHD plays an essential role in Arabidopsis Atrbohd mutant protoplasts showed 38.6% and 50.8% less accumulation of [Ca 2+ ] cyt than wild-type (WT) protoplasts under normal and chilling conditions, respectively (Figure 7). Similar to watermelon protoplasts, melatonin significantly induced [Ca 2+ ] cyt accumulation in WT protoplasts under normal and chilling conditions. However, these effects were abolished by AtRBOHD mutation. Moreover, AtRBOHD knockout significantly increased Arabidopsis sensitivity to freezing at −10 • C and downregulated the expression of AtCBF1 and its targets, including AtCOR47, AtERD10, and AtKIN1 under chilling stress of 4 • C (Figure 8). After freezing exposure, the survival rate of Atrbohd mutant was 80.7% lower than the WT plants, suggesting that AtRBOHD plays an essential role in Arabidopsis response to freezing. Both melatonin and Ca 2+ application dramatically induced the freezing tolerance and enhanced the expression of CBF pathway genes in WT Arabidopsis; however, such effects were attenuated in Atrbohd mutants. Overall, these results indicate AtRBOHD mediates melatonin-induced [Ca 2+ ] cyt accumulation and melatonin/Ca 2+ -induced freezing tolerance.     (Ca 2+ ). After 12 h, the seedlings were subjected to freezing at −10 °C for 1 h and then recovered at 22 °C for 5 days or exposed to chilling at 4 °C for 24 h. Relative expression of genes in untreated WT plants was set as 1.0. Data show the means of three replicates ± standard deviation (SD). The different letters denote significant difference at p < 0.05 according to Turkey's test.

Discussion
The CBF transcriptional regulatory cascade plays a central role in the regulation of cold stress response in plants. Consistent with the previous studies [25,27,34], melatonin application enhanced watermelon tolerance to chilling stress, and its action was closely correlated with the activated CBF-responsive pathway in this study (Figure 1). Furthermore, our results verified that the interaction between H2O2 and Ca 2+ signals mediates melatonin-induced CBF-responsive pathway and subsequent cold tolerance. The relative expression of C-repeat binding factor 1 (AtCBF1) and its regulons, including cold-responsive gene 47 (AtCOR47), early responsive to dehydration 10 (AtERD10), and cold induced gene 1 (AtKIN1) under cold stress at 4 • C. Three-week-old wild-type (WT) and Atrbohd mutant seedlings grown on the 1 2 MS medium were treated with 10 µM melatonin (MT) or 1 mM CaCl 2 (Ca 2+ ). After 12 h, the seedlings were subjected to freezing at −10 • C for 1 h and then recovered at 22 • C for 5 days or exposed to chilling at 4 • C for 24 h. Relative expression of genes in untreated WT plants was set as 1.0. Data show the means of three replicates ± standard deviation (SD). The different letters denote significant difference at p < 0.05 according to Turkey's test.

Discussion
The CBF transcriptional regulatory cascade plays a central role in the regulation of cold stress response in plants. Consistent with the previous studies [25,27,34], melatonin application enhanced watermelon tolerance to chilling stress, and its action was closely correlated with the activated CBF-responsive pathway in this study ( Figure 1). Furthermore, our results verified that the interaction between H 2 O 2 and Ca 2+ signals mediates melatonininduced CBF-responsive pathway and subsequent cold tolerance.

RBOHD-Dependent H 2 O 2 Is Required for Melatonin-Induced CBF-Responsive Pathway and Cold Tolerance
As an essential signaling molecule, RBOH-generated H 2 O 2 plays a primary role in regulating plant response to various abiotic stimuli, such as cold stress [51][52][53]. For instance, cold acclimation enhanced RBOH1 transcript levels and apoplastic H 2 O 2 accumulation in tomato, while RBOH1 silencing suppressed acclimation-induced cold tolerance [54]. Arabidopsis has 10 RBOH genes, AtRBOHA to AtRBOHJ. Among them, AtRBOHD encoded protein potentially regulates ROS-derived responses and plays a fundamental role in stress tolerance [51,[55][56][57]. Plant RBOH genes are highly conserved [58]. Phylogenetic analysis and protein sequence alignment performed herein revealed that Arabidopsis AtRBOHD and watermelon ClRBOHD are homologous genes and their proteins share 79.3% similarity ( Figure 6). Thus, we speculated that ClRBOHD might exhibit a similar function of producing H 2 O 2 like AtRBOHD in Arabidopsis. In this study, chilling exposure induced H 2 O 2 accumulation accompanied by ClRBOHD upregulation. Furthermore, exogenous application of H 2 O 2 induced the expression of CBF-responsive pathway genes and chilling tolerance in watermelon. However, AtRBOHD knockout in Arabidopsis reduced the expression of CBF-responsive pathway genes and freezing tolerance. These results demonstrate that RBOHD-dependent H 2 O 2 regulates the CBF-responsive pathway and cold tolerance in plants.
Melatonin plays primary functions in reducing ROS accumulation and alleviating stress-induced oxidative stress in plants [59]. Consistently, our results showed that melatonin alleviated chilling-induced H 2 O 2 accumulation and lipid peroxidation after watermelon seedlings were exposed to 4 • C for 48 h (Figure 1). Many studies have shown that H 2 O 2 plays an essential signaling role in melatonin-mediated regulation of various physiological processes, including stomatal closure, seed germination, lateral root formation, and response to environmental stimulus [19,29,60,61]. A recent report revealed that H 2 O 2 mediates melatonin-induced cold tolerance in grafted watermelon plants [24]. In this study, melatonin induced H 2 O 2 accumulation and ClRBOHD expression in early response (within 6 h) to chilling stress in watermelon ( Figure 1A). Furthermore, exogenous application of H 2 O 2 enhanced watermelon tolerance to chilling stress. However, DPI application in watermelon or AtRBOHD knockout in Arabidopsis inhibited H 2 O 2 accumulation, suppressing the melatonin-induced expression of CBF pathway genes and subsequent cold stress tolerance (Figures 1, 3 and 8). These findings demonstrate that RBOHD-dependent H 2 O 2 is required for melatonin-induced CBF-responsive pathway and the subsequent cold stress tolerance.

Ca 2+ Mediates Melatonin-Induced CBF-Responsive Pathway and Cold Tolerance
Like H 2 O 2 , Ca 2+ , as an important second messenger, plays a critical role in regulation of responses to various abiotic stimulus in plants [62]. Low temperature induces a rapid and transient Ca 2+ influx in plant cells by activating Ca 2+ channels, such as cyclic nucleotide-gated ion channels (CNGCs) [8,63,64]. Various Ca 2+ sensors decode the Ca 2+ signal evoked by Ca 2+ transient changes, subsequently regulating the expression of CBFs and cold responsive (COR) genes [8,65]. Several studies have confirmed that Ca 2+ signaling is involved in melatonin-induced stomatal closure, seed germination, and salt tolerance in plants [19,29,66]. However, the relationship between Ca 2+ and melatonin in response to cold stress remains unclear. In this study, melatonin promoted [Ca 2+ ] cyt accumulation under cold stress, accompanied by ClCNGC2 upregulation (Figure 2). Arabidopsis CNGC2, a homologous gene of ClCNGC2, plays a critical role in contributing to Ca 2+ entry into cytosol [48,49]. Here, exogenous CaCl 2 induced the expression of CBF-responsive pathway genes and chilling tolerance in both watermelon and Arabidopsis. Meanwhile, blocking Ca 2+ influx channels by LaCl 3 counteracted melatonin-induced expression of CBF-responsive pathway genes and chilling tolerance in watermelon (Figures 3 and 8). These results reveal that Ca 2+ signaling is essential for melatonin-mediated CBF-responsive pathway and subsequent cold tolerance.

H 2 O 2 and Ca 2+ Function Together in a Self-Amplifying Feedback Loop in Melatonin-Induced CBF-Responsive Pathway and Cold Tolerance
As two crucial signal molecules, the interactions between H 2 O 2 and Ca 2+ signal have been well documented in various physiological actions, especially in defense against abiotic stressors [53,67]. For instance, stress-triggered Ca 2+ signal can phosphorylate and activate RBOHD via Ca 2+ -dependent protein kinase to generate H 2 O 2 , which in turn elicits a Ca 2+ signal through receptor-like kinases, such as GUARD CELL HYDROGEN PEROXIDE RESISTANT1, forming a self-propagating mutual activation loop between Ca 2+ and H 2 O 2 signals [62,67]. Consistent with the previous findings, we found that H 2 O 2 and Ca 2+ interact, forming a reciprocal positive-regulatory loop that mediates melatonin-induced CBF pathway and cold tolerance (Figure 9). The following evidence supports this conclusion:  reveal that Ca 2+ signaling is essential for melatonin-mediated CBF-responsive pathway and subsequent cold tolerance.

H2O2 and Ca 2+ Function Together in a Self-Amplifying Feedback Loop in Melatonin-Induced CBF-Responsive Pathway and Cold Tolerance
As two crucial signal molecules, the interactions between H2O2 and Ca 2+ signal have been well documented in various physiological actions, especially in defense against abiotic stressors [53,67]. For instance, stress-triggered Ca 2+ signal can phosphorylate and activate RBOHD via Ca 2+ -dependent protein kinase to generate H2O2, which in turn elicits a Ca 2+ signal through receptor-like kinases, such as GUARD CELL HYDROGEN PEROX-IDE RESISTANT1, forming a self-propagating mutual activation loop between Ca 2+ and H2O2 signals [62,67]. Consistent with the previous findings, we found that H2O2 and Ca 2+ interact, forming a reciprocal positive-regulatory loop that mediates melatonin-induced CBF pathway and cold tolerance (Figure 9). The following evidence supports this conclusion: (1) H2O2 promoted [Ca 2+ ]cyt accumulation, while H2O2 deficiency in watermelon and Arabidopsis prevented melatonin-induced [Ca 2+ ]cyt accumulation in response to cold stress (Figures 2 and 7); (2) CaCl2 stimulated H2O2 accumulation and upregulated ClRBOHD expression, while the blocking of Ca 2+ influx by LaCl3 inhibited melatonin-induced increases in H2O2 accumulation and ClRBOHD transcripts (  During the early response to cold stress, melatonin increases the accumulation of H2O2 and cytoplasmic free Ca 2+ ([Ca 2+ ]cyt) by upregulating the expressions of respiratory burst oxidase homolog (RBOH) D and cyclic nucleotide-gated channel (CNGC) 2, respectively. Increased H2O2 further promotes the accumulation of [Ca 2+ ]cyt, which in turn elevates H2O2 accumulation, forming a reciprocal positive-regulatory loop that mediates melatonin-induced CBF pathway and subsequent cold tolerance.

Conclusions
At present, the signaling mechanisms underlying melatonin-induced cold tolerance in watermelon are still elusive. This study reveals an intricate signaling cascade of melatonininduced cold stress tolerance in watermelon. Melatonin induces H 2 O 2 accumulation by upregulating ClRBOHD expression in early response to cold stress. Increased H 2 O 2 induces ClCNGC2 expression and [Ca 2+ ] cyt accumulation, which boosts H 2 O 2 accumulation by triggering the ClRBOHD expression, forming a reciprocal positive-regulatory loop that induces the expression of CBF pathway genes and subsequent cold tolerance. This is the first study to investigate the interplay between H 2 O 2 and Ca 2+ signaling in melatoninmediated cold tolerance to the best of our knowledge. However, other components of melatonin signaling in plant response to cold stress need to be further explored in the future.