Wild-type Malus domestica cv. ‘Royal Gala’ tissue-cultured plants were initially grown on an MS medium supplemented with 0.2 mg L⁻¹ IAA and 0.3 mg L⁻¹ 6-BA for 4 weeks. They were then transferred to a rooting medium (MS + 0.5 mg IBA and 0.5 mg IAA) for about 60 days. After rooting, the tissue-cultured plants were transplanted to small black plastic pots with good drainage (12 cm × 12 cm) that were filled with soil and sand, and were transferred to a culture room maintained at 23°C and a 14-h photoperiod supplemented with fluorescent light (60 μM m⁻² s⁻¹). One month after acclimating to controlled conditions, plants of similar growth and size were selected for exogenous melatonin treatments. The control was, supplied with thirty milliliters water every 3 days. Treatments received thirty milliliters melatonin solution at concentrations of 0, 100, 500, 1,000, and 5,000 μM as described before (Wang et al., 2013), applied to the soil in each pot every 3 days for 14 days. Afterward, plant heights were measured, and the fourth to seventh mature leaves from the base of the stem (fully mature leaves) were harvested and stored at −80°C.

We obtained MdFRK2-OE transgenic apple lines (L1, L4, and L9) with the cauliflower mosaic virus 35S promoter (CaMV 35S) promoter (Supplementary Figure S1). Compared to untransformed WT lines, MdFRK2 expression increased by 9.2-, 13.0-, and 13.2-fold in lines L1, L4, and L9, respectively. To test whether MdFRK2 is involved in melatonin-mediated variations in sugar metabolism and growth inhibition, MdFRK2-OE tissue-cultured plants (L1, L4, and L9) were selected to receive exogenous melatonin treatments. The growth and culture conditions of the transgenic plants are the same as those of WT described above. In a previous study, 1,000 μM of melatonin initially had an inhibitory effect on the growth of WT plants, so we selected 1,000 μM as the most appropriate concentration for the treatment of the MdFRK2-OE lines. Healthy and uniform plants of WT and transgenic lines (L1, L4, and L9) were selected to receive melatonin application. Regarding the WT and transgenic lines (L1, L4, and L9), control group was irrigated with standard water, and another group was irrigated with 1,000 μM of melatonin dissolved in the irrigation water. The treatment method was similar to that mentioned above. We recorded the day on which the melatonin treatment was started as the first day, and the plant height was measured every 7 days.

Tobacco plants (Nicotiana tabacum L., ‘Samsun NN’) were grown at 25°C, and the 16-h photoperiod was supplemented with lamps at 120 μM m⁻² s⁻¹. After 1 month, 35 plants were separated into seven groups (one group consisted of at least four plants as one biological replicate). Five groups were irrigated with different concentrations of exogenous melatonin (10, 100, 500, 1,000, or 5,000 μM) while another two groups were well-watered as controls.

Soluble sugars were obtained and derivatized according to the method of Yang et al. (2018). Each sample (0.1 g) was extracted in 1.4 ml of 75% methanol, and ribitol was added as an internal standard. After fractionation of non-polar metabolites in chloroform, 5 μL of the polar phase was transferred to 2.0 ml Eppendorf vials to measure the metabolites in each sample, such as sorbitol, sucrose, glucose, and fructose. The metabolites were dried under vacuum without heating and then derivatized with methoxyamine hydrochloride and N-methyl-N-trimethylsilyl-trifluoroacetamide sequentially. Afterward, the metabolites were analyzed with a Shimadzu GCMS-2010SE device (Shimadzu Corporation, Kyoto, Japan). The metabolites were identified by comparing their fragmentation patterns with those from a mass spectral library generated on our GC/MS system and those from an annotated quadrupole GC/MS spectral library downloaded from the Golm Metabolome Database¹. The quantifications were based on standard curves generated for each metabolite and internal standard.

The total RNA was extracted from each sample using an RNAprep plant kit (Tiangen, Beijing, China), according to the manufacturer’s instructions. The RNA was then reverse-transcribed into cDNA with a PrimeScript^TM RT reagent kit and gDNA Eraser to avoid possible genomic DNA contamination (TaKaRa, Dalian, China). The primer sequences used for quantitative real-time polymerase chain reaction (qRT-PCR) assays are presented in Supplementary Table S1. Mdactin (CN938023) was chosen as an internal control for data normalization. qRT-PCR was performed on an ABI 7300 Real-Time PCR instrument (Thermo Fisher Scientific) using a SYBR Green Premix Ex Taq Kit (TaKaRa, Kyoto, Japan). The PCR conditions were as follows: predenaturing at 94°C for 5 min followed by 40 cycles of 95°C for 15 s and 58°C for 60 s. For each sample, total RNA was extracted from three biological replicates. The data were analyzed using the ddCT method.

The enzymes in the mature leaves of WT and transgenic lines were extracted as described by Li et al. (2012), with some modifications. Briefly, a 0.5 g sample was homogenized in 2 ml of 200 mM Hepes-KOH (pH 8.0) buffer, containing 5 mM MgCl₂, 2 mM EDTA, 2.5 mM dithiothreitol (DTT), 2 mM benzamidine, 2 mM ε-aminocaproic acid, 0.1 mM 4-(2-aminoethyl)-benzenesulfonyl fluoride (AEBSF), 1% bovine serum albumin (BSA) (w/v), 2% glycerol (v/v), 0.05% Triton X-100 (v/v) and 2% polyvinylpyrrolidone (PVP) (w/v). The extract was then centrifuged at 16,000 g for 20 min at 4°C, after which 1 ml of the supernatant was immediately desalted in a Sephadex G25 PD-10 column (GE Healthcare, United Kingdom) and subsequently equilibrated with the above mentioned extraction buffer containing a concentration of 50 mM Hepes-KOH (pH 7.4) but without Triton X-100, DTT, BSA, and PVP.

We determined the activity of SPS as described by Li et al. (2016). The assay mixture (200 μl) contained 50 mM Hepes-KOH (pH 7.4), 4 mM MgCl₂, 1 mM EDTA, 4 mM F6P, 3 mM UDP-glucose, 20 mM G6P and 100 μl of the sample. The reaction was carried out at 27°C for 30 min, then it was stopped by boiling in water for 3 min. The blank was run for each assay by adding denatured extracts. After centrifugation for 2 min at 12,000 g, 75 μl of the reaction mixture was used for UDP measurements in a spectrophotometric assay. The assay mixture (1.0 mL) contained 50 mM Hepes-KOH (pH 7.0), 0.8 mM phosphoenolpyruvate, 0.3 mM NADH, 5 mM MgCl₂, 14 U of lactate dehydrogenase, and 4 U of pyruvate kinase (to start the reaction).

Neutral invertase was incubated at 37°C for 60 min in a 200 μl assay mixture that contained 100 mM phosphate-citrate buffer (pH 7.2), 100 mM sucrose, and 50 μl of the desalted extract or denatured extract (as a blank). The assay was stopped by boiling for 3 min before adding 0.75 M Tris-HCl buffer (pH 8.5). The blanks contained the same mixture, but the extract was boiled for 5 min before being mixed. The level of glucose produced from sucrose was determined by the enzyme-coupling method (Li et al., 2016).

Sorbitol dehydrogenase activity was assayed in a 1.0 ml reaction mixture that contained 1 mM NAD⁺, 300 mM sorbitol, and 0.2 ml of desalted extract in 100 mM Tris-HCl (pH 9.6), and NADH production was determined spectrophotometrically at 340 nm.

Hexokinase (HXK) and FRK activities were assayed via a continuous spectrophotometric assay as described by Li et al. (2016), with minor modifications. For HXK, the 0.5 ml assay mixture contained 50 mM Tris-HCl (pH 8.0), 2.5 mM ATP, 4 mM MgCl₂, 1 U of G6P dehydrogenase, 0.33 mM NAD⁺, 1 mM glucose and 25 ml of desalted extract. For FRK, one unit of phosphoglucoisomerase was also added, and 0.4 mM fructose was used instead of glucose.

Based on the genome contig sequence of MDC019777.147, we designed primers to clone the promoter of MdFRK2 (Supplementary Table S1). As the promoter sequence, the upstream 1.8 kb region of MdFRK2 mRNA was cloned from the genomic DNA of ‘Royal Gala’ apple. To examine the transient gene expression in tobacco, the MdFRK2 promoter (PMdFRK2) was introduced into a pC0390GUS vector to generate PMdFRK2-GUS. The empty pC0390GUS vector served as a negative control. Agrobacterium-mediated transient transformation was conducted with samples of fully expanded tobacco leaves in accordance with a previously described protocol (Sparkes et al., 2006).

At 48 h after infiltration, the tissues were irrigated with 25 mL of different concentrations of exogenous melatonin (0, 10, 100, 500, 1,000, or 5,000 μM). After 12 h of irrigation, leaves were stained for 2 h using a GUS histochemical assay as previously described (Xu et al., 2011), while two leaves from each plant were harvested to analyze the GUS activity.

The data were expressed as the means ± standard deviations (SDs) of three replicate samples per measurement. All statistical analyses were performed using IBM SPSS Statistics 21 software (SPSS, Inc., Chicago, IL, United States), and graphs were generated with Sigma Plot 12.5. The data were analyzed using one-way ANOVA and Duncan’s multiple range test, where differences were considered significant at P < 0.05.

To investigate the concentration-dependent effects of melatonin on the growth of apple plants via phenotypic analysis, 2-month-old apple plants grown in soil were irrigated with different concentrations of melatonin solution (0, 100, 500, 1,000, and 5,000 μM) for 14 days (see section “Materials and Methods”). Compared with the control treatment, the low-concentration melatonin treatments (0, 100, 500 μM) caused no significant phenotypic changes; however, the high-concentration melatonin treatments (1,000 and 5,000 μM) inhibited apple plants growth after 14 days of treatment, as reflected by the shoot height and growth rate. Compared to the control, the plant growth rates in response to the 5,000 μM melatonin treatment were reduced by 50.03% (Figure 1A,C). Additionally, when treated with high concentrations of melatonin (1,000 and 5,000 μM), the young leaves and shoot tips of plants clearly appeared red in color (Figure 1B). These observations implied that high concentrations of exogenous melatonin inhibited stem growth and induced leaf anthocyanin accumulations.

As reported previously (Wang et al., 2013; Zhao et al., 2015a), the concentrations of sucrose, glucose, and fructose in the melatonin treatment group were higher than those in the control (water) treatment group (Figure 2). The melatonin concentrations were positively correlated with the sugar concentrations (sucrose, glucose, and fructose) (Figure 2). Compared to the control, the concentrations of fructose, glucose, and sucrose in mature leaves treated with 1,000 μM of melatonin increased 2.47-, 2.02-, and 1.95-fold, respectively. However, sorbitol concentrations did not significantly differ between the treated groups and the control group.

To understand how melatonin regulates carbohydrates in apple mature leaves, the expression of key genes and the activities of enzymes involved in sugar metabolism were subsequently assessed. The transcript levels of MdFRK2 clearly decreased as the melatonin concentrations increased (Figure 3A), and FRK activity showed a similar trend (Figure 3B). NINV activity was significantly upregulated after melatonin treatment, although the expression of MdNINV1 was inhibited by melatonin (Figure 4A,B). The activity of SPS, the major enzyme participating in the sucrose biosynthesis pathway, was also upregulated after melatonin treatment, especially in response to the 1,000 and 5,000 μM treatments (Figure 4B). SDH, the key enzyme regulating sorbitol degradation, was downregulated after melatonin application. HXK activity raised gradually with the increase of melatonin concentration (Figure 4B).

To examine the hypothesis that the expression of MdFRK2 is inhibited by melatonin, a β-glucuronidase (GUS) reporter gene was fused downstream from the MdFRK2 promoter (MdFRK2P), and then the resulting MdFRK2P::GUS construct was introduced into tobacco by Agrobacterium-mediated transient transformation (Figure 5A). The transgenic plants were treated with different concentrations of melatonin (0, 10, 100, 500, 1,000, and 5,000 μM), after which histochemical staining was performed. As shown in Figure 5B, the transgenic plants treated with low concentrations of melatonin showed a strong blue precipitate, while high concentrations of melatonin showed a slight blue precipitate (Figure 5B). To accurately measure the level of GUS activity, quantitative GUS activity assays were performed. The highest activity occurred in the transgenic lines treated with 10 μM of melatonin (Figure 5C), but the MdFRK2 promoter activity decreased as exogenous melatonin concentrations in the transgenic lines increased. These results suggested that the MdFRK2 promoter activity could be inhibited by melatonin.

MdFRK2 was expected to be involved in melatonin-mediated sugar metabolism in apple plants. To further determine the role of MdFRK2 in regulating sugar concentrations under melatonin treatment, we overexpressed MdFRK2 in apple plants with the 35S promoter. Compared with those in untransformed WT lines, MdFRK2 mRNA levels and FRK enzyme activity in the leaves of three transgenic apple lines, lines L1, L4, and L9, significantly increased (Supplementary Figure S1, Yang et al., 2018). 14 days after applying 1,000 μM of exogenous melatonin, we measured the MdFRK2 expression and sugar concentrations of these transgenic and WT apple plants (Figure 6, 7). As expected, compared with those in WT plants treated with control solutions, the transcript levels of MdFRK2 in WT plants treated with 1,000 μM of exogenous melatonin were reduced to 51%. However, compared with those in the three transgenic lines treated with water, MdFRK2 expression treated with 1,000 μM of exogenous melatonin decreased by 8% for L1, 19% for L4, and 14% for L9. Similarly, enzyme analysis exhibited evident decrease in the FRK activity in melatonin-treated WT plants as compared to the WT controls (Figure 6). Exogenous melatonin application resulted in significantly high concentrations of fructose, sucrose, and glucose in the WT lines (Figure 7). However, in the mature leaves of the MdFRK2-overexpression (OE) lines (L1, L4, and L9), exogenous melatonin treatment led to a slight increase in fructose, glucose, and sucrose concentrations (Figure 7).

There were no obvious alterations in growth performance between WT and MdFRK2-OE (L1, L4, and L9) lines (Yang et al., 2018). In the presence of exogenous melatonin, the transgenic lines (L1, L4, and L9) did not show marked growth inhibition nor did their shoot tips became red after 45 days of 1,000 μM of exogenous melatonin treatment (Figure 8B–D). However, the treated WT plants appeared significantly stunted at approximately 21 days after melatonin treatment (Figure 8A).

Melatonin, acting as a plant growth and development regulator as well as a biostimulant, has shown great potential in plant physiology (Arnao and Hernándezruiz, 2015). For example, melatonin can counteract abiotic stress (Hardeland et al., 2012; Park et al., 2013; Liang et al., 2015; Wang et al., 2015). Exogenous melatonin-induced sugar accumulations (including sucrose, glucose, and fructose) have been reported in different plant species (Sarropoulou et al., 2012; Wang et al., 2015; Zhao et al., 2015a,b). Additionally, melatonin applications have also been shown to inhibit FRK2 expression in maize plants (Zhao et al., 2015a). In accordance with the results of previous studies, we observed that, compared with applications of control solutions, applications of exogenous melatonin inhibited the expression of MdFRK2 (Figure 3) and caused higher concentrations of sucrose, fructose, and glucose in cells (Figure 2).

Exogenous melatonin inhibited the expression of MdFRK2, which was confirmed by our experiments. First, MdFRK2 transcription was inhibited by different concentrations of melatonin: the higher the melatonin concentration was, the lower the expression of MdFRK2 and the activity of FRK (Figure 3). Second, in the presence of exogenous melatonin, GUS activity in transgenic tobacco that contained the MdFRK2 promoter-driven reporter gene decreased as the melatonin concentration increased (Figure 5). Third, compared with those in the FRK2-OE lines (L1, L4, and L9), which overexpressed the MdFRK2 gene under the control of the 35S promoter, MdFRK2 expression and FRK activity in the WT lines was clearly inhibited in response to 1,000 μM of melatonin (Figure 6). Melatonin consistently increased fructose accumulations more in the WT lines than in the FRK2-OE lines (L1, L4, and L9) (Figure 7). However, expression of MdFRK2 was also inhibited in the OE lines, suggesting that melatonin inhibit the own MdFRK2 promoter. There should exist a yet to be determined transcription factor that is responsible for melatonin-mediated MdFRK2 expression. Although one melatonin receptor has been identified in Arabidopsis thaliana (Wei et al., 2018), its signaling pathway is unknown in plant cells. The identification of the involved transcription factor will be helpful to understand the mechanism of melatonin in regulating sugar metabolism.

Fructokinase is crucial for fructose metabolism and negatively regulates the accumulation of fructose in plants (Pego and Smeekens, 2000). Expression patterns and functional analysis of FRK isoforms in tomato, rice, maize, and Populus suggest that FRK2 orthologs are the major FRK-encoding genes and the dominant fructose-phosphorylating enzymes in sink tissues (Kanayama et al., 1997, 1998; Pego and Smeekens, 2000; Jiang et al., 2003). MdFRK2, the gene that codes for the main fructose-phosphorylating enzyme in apple plants (Li et al., 2018), controls FRK activity and fructose concentrations in mature apple leaves (Yang et al., 2018). In this study, MdFRK2 expression was inhibited by exogenous melatonin applications at both low and high concentrations (Figure 3, 5), leading to a decrease in FRK activity and an increase in fructose levels (Figure 2, 3). These results are in agreement with those of a previous report with respect to decreases in FRK2 expression (Zhao et al., 2015a) and increases in fructose levels upon short-term exposure to a range of melatonin concentrations (10–100 μM) in maize seedlings (Zhao et al., 2015a). Taken together, these observations indicate a general role for melatonin in limiting FRK2 expression and increasing fructose levels.

Sucrose metabolism, including its synthesis and degradation, is induced by exogenous melatonin at the gene expression and enzyme activity levels. These phenomena are reflected by the increased activities of SPS and invertase, especially under high concentrations of melatonin (100 μM) (Zhao et al., 2015a). Our results further confirm previous findings, as the SPS and NINV activity significantly increased (Figure 4). Additionally, high concentrations of melatonin were found to counteract sucrose phloem loading, as indicated by the reduced expression of the tie-dyed2 (Tdy2) and Sxd1 genes (Zhao et al., 2015a). As the closest homolog of MdFRK2, LeFRK2 has been demonstrated to be required for phloem development and sugar transport (German et al., 2003). In tomato, suppression of LeFRK2 reduced the length and width of sieve elements and resulted in their low levels of callose deposition, which restricted the downward movement of sucrose and greatly elevated soluble sugar levels in tomato source leaves (German et al., 2003). Therefore, it is possible that exogenous melatonin inhibited the expression of MdFRK2 and then caused phloem interruption, thus leading to an abundance of sucrose in the leaves (Figure 2). The significantly increased sucrose levels may stimulate sweet immunity responses through sucrose sensing and signaling (Bolouri Moghaddam and Van den Ende, 2013).

Sugars fulfill many essential functions in plant cells (Pego and Smeekens, 2000). Plants have a system that controls sugar homeostasis to help avoid the negative effects associated with fluctuations in sugar concentrations. Excess fructose in the cytosol is harmful to plant growth and development. On the one hand, high concentrations of fructose can lead to stress reactions such as senescence. On the other hand, the inhibition of fructose phosphorylation can limit carbon flux to glycolysis and other pathways via F6P. Under high concentrations of fructose, apple plants have at least two alternative approaches to limit the production of fructose in cells: one involves sorbitol conversion to fructose via SDH catalysis (Yamada et al., 1998), and the other involves sucrose cleavage by SUSY or invertase (Pego and Smeekens, 2000; Ruan, 2014). Our results indicated that the decreased fructose concentrations in the leaves after low melatonin applications were partly limited by decreased sorbitol dehydrogenization, reflected by decreased SDH expression and activity (Figure 4). Correspondingly, the NINV activity increased (Figure 4), causing an increase in glucose levels. Interestingly, we observed opposite changes of SDH, SUSY, and NINV activities after melatonin treatment (Figure 4) compared to those in MdFRK2 overexpressing lines (Yang et al., 2018). These results indicated that carbohydrate metabolism changes in response to melatonin were mediated by MdFRK2.

Compared to the control solutions and low-concentration melatonin treatments, shoot growth in response to high-concentration melatonin applications was significantly inhibited (Figure 1, 8). The observed growth inhibition effect was similar to that in previous reports on growth inhibition of maize seedlings in response to high concentrations of exogenous melatonin treatments (Zhao et al., 2015a). However, in the present study, it is interesting that the negative effect of growth was significantly alleviated in the MdFRK2-OE lines in response to the 1,000 μM melatonin treatment, as reflected by the similar shoot height between the MdFRK2-OE and WT lines (Figure 8A). Moreover, expression of MdFRK2 and activity of FRK were suppressed slightly less in the treated MdFRK2-OE plants than in WT plants after 14 days of melatonin treatment (Figure 6). These phenomena indicate that MdFRK2-OE counteracts the effects of the high concentration melatonin-mediated growth inhibition of MdFRK2-OE apple plants, suggesting that the reduced expression of MdFRK2 is an important cause of the stunted phenotype of WT plants (Figure 8).

FRK2 plays important physiological roles in regulating plant growth and development. For example, RNAi-MdFRK2 tomato plants displayed inhibited stem and root growth and reduced numbers of seeds per fruit (Odanaka et al., 2002). The young leaves of FRK2-antisense tomato plants showed inhibited growth and wilting because of the suppression of a particular area of active xylem (German et al., 2003). Antisense suppression of the FRK2 ortholog in potato (StFRK1) inhibited aerial growth and resulted in decreased tuber yields, possibly due to reduced phloem transport (Cho and Yoo, 2011). Reduced FRK2 activity led to reduced length and width of sieve elements, thus leading to inhibited stem water conductivity and restricted downward movement of sucrose (Damari-Weissler et al., 2009). As MdFRK2 expression is extremely high in actively growing sink tissues such as shoot tips in apple plants (Li et al., 2012), modulated expression of MdFRK2 in shoot tips might negatively affect stem development, potentially causing growth inhibition after melatonin treatment.

Previous studies showed that hexokinase1 (AtHXK1) overexpression inhibits growth (Jang et al., 1997; Dai et al., 1999), and the stunted growth occurs only when AtHXK1 expressed in photosynthetic tissues (Dai et al., 1999). Therefore, the expression of MdHXK1 was monitored in our study. We found that the MdHXK1 transcripts were upregulated in mature leaves, especially under high concentration of melatonin application (Figure 4). Although our results cannot provide the exact mechanism by which this occurs, these data allow us to speculate that the increased HXK1 expression may be also involved in the inhibitory effect (Figure 1B).

It has been demonstrated that suppression of the FRK2 orthologs resulted in increased sugar concentrations (German et al., 2003; Davies et al., 2005; Roach et al., 2012). Our data showed MdFRK2 represents a link in the cross-talk between increased sugar concentrations and the melatonin application. Current results also support that exogenous melatonin may mimic DAMPs, counteracting both abiotic and biotic stresses through elevated leaf sweetening (Versluys et al., 2017). Spraying exogenous melatonin on leafy vegetables may be very promising in terms of agronomical applications. Especially melatonin/fructan mixtures hold great promise, as natural and sustainable alternatives for toxic agrochemicals, to be used on leafy vegetables (Versluys et al., 2017). Further investigations of the role of melatonin in leafy vegetables will be done to support or refute these theories.