Introduction

Due to their sessile lifestyle, plants have to deal with various biotic and abiotic stresses during their life cycle. In particular, drought is one of the most destructive environmental stresses, which adversely affect plant growth, development, and agricultural productivity. In recent years, changing climate aggravates the frequency and severity of drought stress (Jury and Vaux 2005; Pachauri et al. 2015). To survive under drought, plants have evolved comprehensive mechanisms to achieve physical adaptations such as changes in leaf morphology, deep root system, stomatal closure, and cutinization of the leaf surface (Burgess and Huang 2016; Xue et al. 2017).

In all land plants, the epidermis of aerial organs is sealed with a hydrophobic surface layer, commonly known as the cuticle (Yeats and Rose 2013). It consists of two major components: the polymer cutin and cuticular waxes. The cutin is polymer of fatty acid derivatives linked by polyesters (Beisson et al. 2012), while the cuticular waxes comprise mostly of very-long-chain fatty acids (VLCFAs) and their derivatives, including alkanes, fatty acids, ketones, primary alcohols, aldehydes, secondary alcohols, wax esters, and triterpenes. Wax biosynthesis begins with the de novo synthesis of C16 or C18 VLCFA, which is converted to the wax components through several biosynthetic pathways (Kunst and Samuels 2009). A number of genes such as long-chain acyl-CoA synthase (LACS), β-ketoacyl-CoA synthetase (KCS), β-ketoacyl-CoA reductases (KCRs), fatty acyl-CoA reductase (FAR), and very-long-chain aldehyde decarbonylase (eceriferums, CER) are thought to be involved in the biosynthesis of waxes (Kunst and Samuels 2009). Additionally, the cuticular waxes are either deposited within the cutin matrix or onto the surface often forming complex crystal microstructures such as platelets, ribbons, rodlets, plates, threads, crusts, and several others (Jeffree 2008).

As the outermost hydrophobic layer, cuticular waxes serve as an essential waterproof barrier to protect plants from desiccation by limiting non-stomatal transpirational water loss (Jenks et al. 2001; Cameron 2005; Isaacson et al. 2009). It has been shown in the fruits of tomato and sweet cherry, the resistance of the epicuticular wax layer contributes approximately 70% to total cuticular resistance to water diffusion (Knoche et al. 2000; Leide et al. 2007). Generally, it is considered that the wax amount or the cuticle thickness is positively associated with the resistance to water loss and subsequent drought tolerance. However, increasing studies have demonstrated that cuticular water permeability (CWP) is inversely correlated with the proportion of nonpolar wax components, such as alkanes, but not the total wax amount (Buschhaus and Jetter 2012; Parsons et al. 2012). Additionally, the CWP is also determined by the packing and organization of wax within the cuticle architecture (Riederer and Schreiber 1995).

The biosynthesis of cuticular waxes can be regulated by multiple environmental cues such as water deficit and endogenous factors such as plant hormones and transcription factors (Yeats and Rose 2013). As a key component of hormone signaling networks that regulate plant tolerance to drought, ABA can not only promote stomatal closure, but also regulate wax biosynthesis to prevent water loss under water deficit conditions (Kosma et al. 2009). Numerous studies have shown that melatonin acts as a key growth regulator that greatly influences plant growth, development, and responses to environmental stresses, such as drought (Arnao and Hernández-Ruiz 2015; Ahammed et al. 2019, 2020a, b; Hasan et al. 2019). Similar to ABA, melatonin has been recently demonstrated to induce cuticular wax biosynthesis in tomato plants under drought (Ding et al. 2018); however, the mechanism is still unknown.

Watermelon (Citrullus lanatus (Thunb.) Matsum. & Nakai) is a delicious juicy fruits, widely grown as an important vegetable crop around the world. According to FAOSTAT (https://www.fao.org/), the cultivated area of watermelon in 2016 was approximately 3.51 million hectares worldwide, the annual global production of watermelon is about 117.20 million tons, and the gross production value of watermelon in 2014 was $38.56 billion. Despite being originated from Africa, watermelon plants show high water demand and low irrigation water use efficiency, especially during the fruit development stage, and are sensitive to drought (Rashidi and Gholami 2008). Water deficit has become a major abiotic stress that severely limits watermelon productivity (Zhang et al. 2011). Similar to the studies on Arabidopsis, Nicotiana glauca, and Sorghum bicolor (Cameron et al. 2006; Kim et al. 2007; Kosma et al. 2009; Lu et al. 2012), cuticular wax biosynthesis and accumulation play a vital role in watermelon tolerance to drought (Li et al. 2019). However, the cuticular wax components and structure in watermelon leaves and their response and regulation under drought are still unknown. In this study, three watermelon germplasms, with similar stomatal response but different tolerance to drought stress, were used to assay the responses of cuticular wax compositions and structure under unwatered conditions. In addition, we analyzed the effects of melatonin and ABA on the wax compositions and deposition under drought in watermelon. The study unveiled a critical role of alkanes (C29 and C31) in wax compositions and their contributions to melatonin and ABA-induced drought tolerance in watermelon.

Materials and Methods

Plant Materials, Experimental Conditions, and Treatments

Three germplasms of watermelon with different levels of drought tolerance were used in the present study, including watermelon inbred lines M20, M08, and J5F with high, moderate, and low drought tolerance, respectively. The seeds were provided by the Watermelon and Melon Research Group at Northwest A&F University, Yangling, Shaanxi, China. Germinated seeds of these germplasms were sown into the plastic pots (10 × 8.5 × 7 cm3) filled with an amount of 450 g growth substrate (per pot), containing an equal volume of sand and a commercial compost. The watermelon seedlings were then grown in growth chambers with the following environmental conditions: 25/18 °C (day/night), a constant relative humidity of 70–80%, a 12-h photoperiod, and a photosynthetic photon flux density of 600 μmol m−2 s−1. The plants were watered daily and supplied with half-strength Hoagland’s nutrient solution every 5 days.

Experiment I

When seedlings developed four true leaves, one half of seedlings from each genotype were challenged with drought by withholding water supply for 8 days. The control plants (CK) received water at 75 (± 5) % field capacity based on soil weight every evening. Each treatment included three replicates and each replicate included 12 plants. For biochemical assay, the second fully expanded leaves below the growing points were sampled, immediately frozen in liquid nitrogen and kept in a freezer at − 80 °C.

Experiment II

At the four-leaf stage, the leaves of J5F seedlings were sprayed with 100 μM ABA or 150 μM melatonin for 3 times (one time every 2 days). Seedlings treated with distilled water served as control. The melatonin and ABA solutions were dissolved in ethanol and then diluted with distilled water at a 1:10,000 ratio (v/v). Twelve hours after the 3rd application of melatonin or ABA, the watermelon plants were deprived from water supply for 6 days. Then, the second fully expanded leaves under the growing points were sampled, immediately frozen in liquid nitrogen and kept in a freezer at − 80 °C until analysis.

Leaf Water Content and Relative Electric Conductivity Analysis

Leaf water status was revealed by measuring the relative water content (RWC) of leaves following the method of Barrs and Weatherley (1962). The relative electric conductivity (REC) was determined according to the method of Zhou and Leul (1998). Briefly, samples (1.0 g) were collected from the second topmost fully expanded leaves by tearing 0.5-cm circles and submerged them in 50-mL test tubes filled with deionized water (25 mL). The test tubes were then vacuumed, and the leaf samples were vibrated for 20 min. The conductivity of solution was measured with a conductivity meter (DDS-2307) to record the first reading (C1) followed by boiling for 10 min. When the solution was cooled to room temperature, the solution conductivity (C2) was measured again. REC was determined by using following formula: REC = C1/C2 × 100%.

Leaf Gas Exchange and Non-invasive Chlorophyll Fluorescence Analysis

The gas exchange was measured in the second topmost fully expanded leaves using a Li-Cor-6400 Portable Photosynthesis System (LI-COR Inc., Lincoln, NE, USA) at the growth temperature, 600 μmol m−2 s−1 photosynthetic photon flux density (PPFD), 500 mL min−1 cuvette air flow rate, and 400 μmol mol−1 CO2 concentration. The net photosynthetic rate (Pn), stomatal conductance (Gs), and transpiration rate (Tr) were recorded simultaneously. Chlorophyll fluorescence was determined with a Portable Chlorophyll Fluorometer (PAM2500; Heinz Walz, Effeltrich, Germany) following dark adaptation for 30 min at room temperature. The maximum photochemical efficiency of photosystem II, PSII (Fv/Fm) was analyzed according to Maxwell and Johnson (2000).

Cuticular Wax Load and Chemical Composition Analysis

The total wax load and chemical composition analysis were performed following the method of Wang et al. (2015). The harvested watermelon leaves were immersed in chloroform for 60 s to extract waxes and n-tetracosane (C24) was incorporated as an internal standard. Extracted wax samples were evaporated under gaseous nitrogen (N2) and derivatized using a mixture of bis-N, N-(trimethylsilyl) trifluoroacetamide (Sigma-Aldrich) and pyridine (Fluka). Afterward, the solvent was dried again using N2 gas and the waxes were re-dissolved in chloroform. The samples were analyzed with temperature-controlled capillary gas chromatograph equipped with a mass spectrometer (GCMS-QP2010, Shimadzu, Kyoto, Japan). The components were separated on an Rxi-5 ms column (30 m × 0.25 mm, 0.25 μm film thickness) using helium (He) as the carrier gas. The initial temperature of wax separation was 50 °C for 2 min, elevated by 20 °C min–1 to 200 °C, kept at 200 °C for 2 min, and then increased at 2 °C min–1 to 320 °C, at this point the temperature was held for 15 min. The wax components were quantified against the C24 internal standard.

Visualization of Wax Deposition by Scanning Electron Microscopy Analysis

For epicuticular wax crystal observation, the second fully expanded leaves under the growing points were air-dried at room temperature. The dried samples were sputter-coated with gold film. Coated surfaces were observed using a scanning electron microscope (Hitachi S4800, Hitachi, Tokyo, Japan) (Wang et al. 2015).

Quantification of Abscisic Acid

Abscisic acid (ABA) was extracted and determined according to a previously described protocol (Xia et al. 2014). For the extraction of ABA, frozen leaf tissues (0.5 g) were homogenized in 4 mL of 80% methanol (v/v) containing 1 mM 2,6-di-t-butyl-p-cresol. The homogenate was incubated at 4 °C in the dark for 4 h. After being centrifuged, the obtained supernatants were filtered using a Sep-Pak C18 cartridge and dried under the flow of N2 gas. The dried samples were re-dissolved in 5 mL of elution buffer (pH 8.1) containing 1 mM MgCl2, 10% (v/v) methanol in 50 mM Tris, 150 mM NaCl. ABA was determined with an immunoassay kit (China Agricultural University, Beijing, China) following the instructions of manufacturer. The absorbance was recorded on a M200 pro Multimode Plate Reader (Tecan, Männedorf, Switzerland).

Total RNA Extraction and Gene Expression Analysis

Total RNA from leaves was extracted using an RNA extraction kit (Axgen, Union City, CA, USA) following the supplier’s instructions. Residual DNA was removed using a DNase Mini Kit (Qiagen, Hilden, Germany). An aliquot of total RNA (1 µg) was reverse transcribed using a ReverTra Ace qPCR RT Kit (Toyobo, Osaka, Japan) according to the supplier’s recommendations. The primers for qRT-PCR were designed based on their cDNA sequences (supplemental Table S1). According to Kong et al. (2014), the β-actin (Cla007792) gene was used as the internal reference. qRT-PCR was performed using an iCycler iQ™ Multicolor PCR Detection System (Bio-Rad, Hercules, CA, USA). PCR products were amplified using SYBR® Premix ExTaq TM II (2 ×) kit 170 (Takara, Tokyo, Japan). The PCR conditions consisted of denaturation at 95 °C for 3 min; followed by 40 cycles of 30 s at 95 °C, 58 °C for 30 s, and 72 °C for 30 s. The relative expression of genes was obtained as described previously (Livak and Schmittgen 2001).

Statistical Analysis

The experiment was carried out in a complete randomized design with 3 replicates. Experimental data were subject to analysis of variance (ANOVA) and significant differences between treatments were determined at P < 0.05 using Tukey’s test.

Results

Responses of M20, M08, and J5F Under Drought Stress

Drought stress caused leaf wilting and significant decrease in leaf relative water content (RWC), stomatal conductance (Gs), and transpiration rate (Tr) in all three watermelon genotypes (Fig. 1). The RWC was significantly reduced by 21.1%, 40.8%, and 53.5% in M20, M08, and J5F, respectively, after stopping water supply for 8 days. M20 showed the greatest tolerance against drought based on its minimal degree of wilting and the highest RWC, followed by M08 and J5F. There was no significant difference in Tr among M20, M08, and J5F under well-watered conditions. The lowest Tr was observed in M20 followed by that in M08 and J5F under drought conditions. However, the three genotypes showed no significant difference in Gs under both normal and drought conditions.

Fig. 1
figure 1

Effects of drought on a plant phenotypes, b leaf relative water content (RWC), c stomatal conductance (Gs), and d transpiration rate (Tr) in three watermelon genotypes (M20, M08, and J5F). At the four-leaf stage, watermelon seedlings were either deprived from water supply (drought stress, DS) or well-watered (as control, CK) for 8 days. Data are presented as the means of three replicates (± SD). Means designated by the same letters do not significantly differ at P < 0.05

Full size image

Identification of Cuticular Wax Components on Watermelon Leaves

According to the GC–MS qualitative analysis, the wax components on the leaf epidermis of three watermelon materials were roughly the same. In total, 28 compounds were identified, including 11 alkanes, 6 primary alcohols, 5 aldehydes, 3 fatty acids, and 3 triterpenes (Fig. 2). The chain lengths of alkanes varied from C25 to C35. The even-numbered chain lengths of primary alcohols, aldehydes, and fatty acids ranged from C22 to C32, from C24 to C32, and from C26 to C30, respectively. C31, C28, C32, and C30 chains were the dominant chain lengths of alkanes, primary alcohols, aldehydes, and fatty acids, respectively. The α-aromatic resin alcohols were the major triterpenoids, followed by β- and δ-aromatic resin alcohols.

Fig. 2
figure 2

GC–MS analysis of the chemical compositions of cuticular wax extracts from watermelon leaves. Each wax constituent is designated by carbon (C) chain length and is labeled by chemical class. Alk alkanes, Alc primary alcohols, Ald aldehydes, FA fatty acids, Tri triterpenoid

Full size image

Responses of Wax Compositions and Deposition Under Drought Stress

M20 showed higher total wax contents than M08 and J5F under normal conditions. Wax biosynthesis in leaves was obviously increased by drought in all three watermelon genotypes (Fig. 3a). After the imposition of drought, the total wax load on leaf was 6.82, 7.03, and 5.37 μg cm−2 in M20, M08, and J5F, respectively, higher than that in their respective control. Consistently, most of the genes involved in wax biosynthesis, including 3-ketoacyl-CoA synthase 1 (KCS1, Cla021116), fatty acyl coA reductase 3 (FAR3, Cla008654), and very-long-chain aldehyde decarbonylase 1 (CER1, Cla006996), were induced by drought in all three genotypes, with some exceptions. M20 showed the least transcripts of KCS1 and FAR3 but the most transcripts of CER1 under drought, when compared to M08 and J5F. Among different wax constituents, alkanes with chain lengths of C29 and C31 showed the greatest increase amplitude after drought (Fig. 3b). The contents of C29 and C31alkanes were 1.19 and 1.55 μg cm−2, respectively, in M20 under drought, which were higher than that in M08 and J5F.

Fig. 3
figure 3

Responses of the a total wax load, wax biosynthetic gene expressions, and b relative wax compositions under drought. Seedlings received the same treatments as described in Fig. 1. Each wax constituent is designated by carbon chain length, and is labeled by chemical class along the x-axis. Data showed the means of three replicates (± SD). Means designated with the same letters do not significantly differ at P < 0.05. CK control, DS drought stress, KCS 3-ketoacyl-CoA synthase, FAR Fatty acyl coA reductase, CER Very-long-chain aldehyde decarbonylase

Full size image

The cuticular waxes are either deposited as filler within the cutin matrix or coated the surface as wax crystals with complex three-dimensional microstructures. The morphology of epicuticular waxes on the upper and lower epidermis of watermelon leaves was observed using a scanning electron microscopy (SEM). A small number of wax crystals were observed on the upper epidermis of all three watermelon genotypes under well-watered conditions (Fig. 4a). After exposure to water deficit stress, more platelet-like wax crystals were accumulated on the upper epidermis of M08 and especially J5F, but few wax crystals were observed in M20. Few wax crystals appeared on the lower epidermis of all three watermelon genotypes under both normal and drought conditions (Fig. 4b).

Fig. 4
figure 4

Responses of the wax crystals on the a upper and b lower epidermis of leaves of three watermelon germplasms under drought. Seedlings received the same treatments as described in Fig. 1. The wax crystals were observed under a scanning electron microscope. Scale bar = 5 μm

Full size image

The Effects of Melatonin and ABA on Wax Compositions and Deposition

Drought induced leaf wilting and decreased Fv/Fm (Fig. 5a, b). However, melatonin or ABA pretreatment alleviated drought-caused leaf wilting and the reduction of Fv/Fm. The total wax load of leaves was increased by drought stress and such increase was promoted by both melatonin and ABA pretreatment. Total wax load of leaves treated with melatonin and ABA was 36.6% and 49.8% higher than that of control leaves, respectively, after exposure to drought (Fig. 5c). Among wax compositions, alkanes with chain lengths of C29 and C31 showed the greater increase amplitude by melatonin and ABA under deficit water supply, than the other compositions. Wax crystals on leaves were increased by drought; however, such increase was attenuated by ABA and melatonin pretreatment (Fig. 6).

Fig. 5
figure 5

The effects of abscisic acid (ABA) and melatonin (MT) on a plant phenotypes, b Fv/Fm, and c wax compositions under drought. At the four-leaf stage, the leaves of J5F seedlings were sprayed with 100 μM ABA or 150 μM melatonin for 3 times (one time every 2 days). After melatonin or ABA pretreatment, plants were exposed to drought for 6 days. Each wax constituent is designated by carbon chain length, and is labeled by chemical class along the x-axis. Data showed the means of three replicates (± SD). Means designated with the same letters do not significantly differ at P < 0.05. CK control, DS drought stress

Full size image
Fig. 6
figure 6

The effects of ABA and melatonin on the wax crystals under drought stress. Seedlings received the same treatments as described in Fig. 1. The wax crystals were observed by using a Hitachi S4800 scanning electron microscope. Scale bar = 5 μm

Full size image

The Effects of Melatonin on ABA Signaling Under Drought

To understand the effects of melatonin on ABA signaling under drought, we first analyzed the dynamic changes in ABA contents in leaves. Melatonin pretreatment increased the content of ABA under normal conditions as indicated at 0 days of drought stress (Fig. 7a). After exposure to drought stress, the ABA contents in control and melatonin-pretreated plants sharply increased from 1st and 2nd day, respectively. Finally, ABA levels in melatonin-treated plants were lower on the 2nd day but higher from the 4th day than that in control under drought. Consistently, the transcripts of ABA biosynthetic gene 9-cis-epoxycarotenoid dioxygenase 6 (NCED6, Cla002942) and ABA responsive gene abscisic acid responsive element binding factor 2 (ABF2, Cla022580) were up-regulated, but the transcripts of ABA metabolic gene abscisic acid 8′-hydroxylase 1 (CYP707A1, Cla005457) were down-regulated by melatonin after water withholding for 4 days (Fig. 7b, c, d).

Fig. 7
figure 7

The effects of melatonin on the a dynamic changes in ABA accumulation and the expressions of ABA b response-, c biosynthesis-, and d catabolism-related genes under drought stress. Seedlings were treated as described in Fig. 5. Data showed the means of three replicates (± SD). Means designated with the same letters do not significantly differ at P < 0.05. CK control, DS drought stress, ABA abscisic acid, MT melatonin, ABF2 abscisic acid responsive element binding factor 2, NCED6 9-cis-epoxycarotenoid dioxygenase 6, CYP707A1 abscisic acid 8′-hydroxylase 1

Full size image

Discussion

In the current study, three watermelon germplasms M20, M08, and J5F were used to analyze the response of cuticular wax compositions and deposition on leaves under drought. In terms of leaf phenotypes and Fv/Fm, M20 showed the highest drought tolerance, followed by M08 and J5F (Fig. 1). Water can be evaporated through stomatal transpiration and non-stomatal transpiration from plant leaves. To investigate the roles of cuticular waxes on leaves in plant tolerance to drought, the stomatal transpiration should be excluded as much as possible (Kerstiens 1996). The stomatal conductance in the three watermelon germplasms was similar under both normal and drought conditions, suggesting that these germplasms were suitable to analyze the functions of cuticular waxes. The lower transpiration rate of M20 under drought stress might be attributed, at least in part, to the changes in cuticular wax properties.

Identification of Wax Compositions

The cuticular wax compositions exhibit a high degree of variability among different species as well as in different tissues and organs (Samuels et al. 2008). For instance, the very-long-chain alkanes and triterpenoids were predominant constituents of waxes on tomato leaves (Ding et al. 2018), while the wheat blade cuticular waxes predominantly consist of primary alcohols and alkanes (Wang et al. 2015). Here, a total of 28 compounds were identified on watermelon leaves, including 11 alkanes, 6 primary alcohols, 5 aldehydes, 3 fatty acids, and 3 triterpenes (Fig. 2). The alkanes (C29 to C34) and primary alcohols (> C28) were the most dominant constituents, while the amount of triterpenoids was very low and negligible.

Roles of Wax Compositions and Deposition in Drought

The water permeability of the cuticular waxes is not necessarily correlated with its total coverage but its compositions and spatial arrangement (Kerstiens 2006). Generally, the nonpolar components, such as alkanes, are a more effective water barrier than the nonaliphatic wax compounds such as triterpenoids (Yeats and Rose 2013). A reduction of wax monomers especially alkanes with chain lengths of C29 is attributed to a decline in drought tolerance (Panikashvili et al. 2007). We found that a greater increase of alkanes with chain lengths of C29 and C31 in M20 than that in M08 and J5F after exposure to drought (Fig. 3b). Moreover, the primary alcohols with chain lengths of C28 to C32 decreased in M20 but increased in M08 and J5F under drought (Fig. 3b). CER1, KCS1, and FAR3 are responsible for alkane synthesis, fatty acid elongation, and primary alcohol synthesis, respectively (Bourdenx et al. 2011; Bernard and Joubès 2013). Consistently, more transcripts of CER1 but fewer transcripts of KCS1 and FAR3 were observed in M20 under water deficit conditions (Fig. 3a). Therefore, the distinct shifts and the profound accumulation of alkanes with chain lengths of C29 and C31, at least partly, contributed to the higher tolerance of M20 to drought.

The cuticular waxes embedded into the cutin matrix (intracuticular wax) are more effective to waterproof the membranes than the wax crystals deposited on cutin surface (epicuticular wax) (Bargel et al. 2006; Koch and Ensikat 2008). The total amount of waxes was increased but few wax crystals were observed on the upper epidermis of M20 leaves under drought (Fig. 4a), suggesting that more intracuticular waxes was formed and contributed to the increased tolerance of M20 to drought. It has been reported that the morphologies of epicuticular waxes are determined by wax compositions (Jeffree et al. 1976; Zhang et al. 2013). Therefore, the increase of alkanes (C29 and C31) and the decrease of primary alcohols (C28 to C32) might contribute to the wax deposition in M20 leaves under drought.

Roles of Melatonin and ABA in Regulating Wax Compositions and Deposition

A growing number of studies have reported that enhanced drought tolerance by both melatonin and ABA, two important plant growth regulators, is associated with their regulation of wax accumulation. Exogenous melatonin increases the cutin thickness and n-alkane content on tomato leaves following exposure to water deficit (Ding et al. 2018). The results of our recent study also indicate that melatonin and ABA induced the total amounts of waxes and the expression of wax biosynthetic genes in watermelon under drought (Li et al. 2019). In the present study, we demonstrated that melatonin- and ABA-increased wax accumulation under drought was mainly attributed to the increase of alkanes with chain lengths of C29 and C31 (Fig. 5c). Moreover, melatonin and ABA attenuated drought-induced accumulation of wax crystals on upper epidermis (Fig. 6), suggesting that the enhanced drought tolerance was also associated with the regulation of wax arrangement.

Melatonin and ABA function synergistically or antagonistically to regulate multiple processes such as seed germination, stomatal movement, and defense against environmental stresses in plants (Zhang et al. 2014; Li et al. 2015; Fu et al. 2017). Our previous study shows that melatonin increases ABA levels under drought conditions (Li et al. 2019), but this conclusion is contrary to the study of Li et al. (2015), in which melatonin decreased the content of ABA in apple trees under drought. The results of these two studies only reflect the effects of melatonin on ABA accumulation at a certain time point after drought. Here, we further analyzed the dynamic changes in ABA levels as influenced by melatonin under prolonged drought conditions. The results showed that melatonin inhibited the increase of ABA content after drought for 2 days, but promoted ABA accumulation from the 4th day after the imposition of drought treatment, accompanied by the up-regulation and down-regulation of ABA biosynthesis- and degradation-related genes, respectively (Fig. 7). Accordingly, melatonin and ABA might function antagonistically to regulate stomatal movement to balance gas exchange and transpirational water loss under mild drought (Li et al. 2015), but synergistically induce alkanes (C29 and C31) accumulation to limit non-stomatal water loss under severe drought conditions and ABA might act downstream of melatonin.

Conclusions

In the present study, a total of 28 compounds of cuticular waxes were identified on watermelon leaves and alkanes (C29, C31, and C32) and primary alcohols (C28, C30, and C32) were the most dominant constituents. The increases of alkanes with chain lengths of C29 and C31 and the more intracuticular wax formation within the cutin matrix contributed to the increased tolerance to drought. Melatonin and ABA, which can be increased by drought, function synergistically to regulate wax response to limit non-stomatal water loss under severe but not mild drought in watermelon (Fig. 8).

Fig. 8
figure 8

A working model depicting how cuticular wax is regulated to confer watermelon tolerance to drought. When watermelon leaves suffer water deficit, melatonin and ABA are increased and then induce the biosynthesis of alkanes (C29 and C31), which confer intracuticular wax accumulation and drought tolerance

Full size image