Melatonin Delays Postharvest Senescence through Suppressing the Inhibition of BrERF2/BrERF109 on Flavonoid Biosynthesis in Flowering Chinese Cabbage

Flowering Chinese cabbage is prone to withering, yellowing and deterioration after harvest. Melatonin plays a remarkable role in delaying leaf senescence and increasing flavonoid biosynthesis. However, the underlying molecular mechanisms of melatonin procrastinating postharvest senescence by regulating flavonoid biosynthesis remain largely unknown. In this study, melatonin could promote flavonoid accumulation and delay the postharvest senescence of flowering Chinese cabbage. Surprisingly, we observed that BrFLS1 and BrFLS3.2 were core contributors in flavonoid biosynthesis, and BrERF2 and BrERF109 were crucial ethylene response factors (ERFs) through the virus-induced gene silencing (VIGS) technique, which is involved in regulating the postharvest senescence under melatonin treatment. Furthermore, yeast one-hybrid (Y1H), dual luciferase (LUC), and β-glucuronidase (GUS) tissue staining experiments demonstrated that BrERF2/BrERF109 negatively regulated the transcripts of BrFLS1 and BrFLS3.2 by directly binding to their promoters, respectively. Silencing BrERF2/BrERF109 significantly upregulated the transcripts of BrFLS1 and BrFLS3.2, promoting flavonoid accumulation, and postponing the leaf senescence. Our results provided a new insight into the molecular regulatory network of melatonin delaying leaf senescence and initially ascertained that melatonin promoted flavonoid accumulation by suppressing the inhibition of BrERF2/BrERF109 on the transcripts of BrFLS1 and BrFLS3.2, which led to delaying the leaf senescence of postharvest flowering Chinese cabbage.


Introduction
Plant senescence is a complex and highly coordinated process, which obviously affects the quality of crops [1,2]. Leaf senescence is influenced by a variety of internal and external factors. For example, ethylene causes disruption of the antioxidant system, leading to the onset and development of leaf senescence, and even to the leaf necrosis and plant death [3][4][5]. Ethylene transcription factors (ERFs), as one of the largest plant transcription factors, play powerful roles in mediating the plant senescence [6,7]. For example, RhERF3 is strongly induced by ethylene, which directly interacts with 9-cis-epoxycarotenoid dioxygenase (RhNCED1) and activates its transcriptional activity, eventually speeding up the aging process of rose [8]. In Arabidopsis, AtERF34 directly activates the expressions of salt stress-responsive genes AtERD10 and AtRD29A, which negatively regulates the leaf senescence [9]. OsERF101 directly regulates the transcripts of chlorophyll degradation gene (OsNAP) and the core factor myelocytomatosis oncogenes (OsMYC2) of the jasmonic acid signaling pathway to promote the occurrence and progress of leaf senescence in rice [10]. Likewise, in tomato, SlERF.F5 involves in regulating jasmonic acid-induced leaf senescence by interacting with SlMYC2 [11]. However, whether and how ERFs are involved in leaf senescence of flowering Chinese cabbage (Brassica campestris L. ssp. Chinensis var. utilis Tsen et Lee) is still largely unclear.
Melatonin is widely found in plants and plays an essential role in delaying the aging of crops [12][13][14][15][16]. Melatonin biosynthesis begins with tryptophan, and at least six enzymes (Tryptophan decarboxylase, tryptamine 5-hydroxylase, Serotonin N-acetyltransferase, arylalkyl amine N-acetyltransferase, N-acetyl-serotonin methyltransferase, and hydroxyindole-O-methyltransferase) have been known to be involved in the biosynthesis [17]. The major genes encoding these enzymes of the entire melatonin biosynthetic pathway involved in delaying aging have been identified from various plants, including tryptophan decarboxylase, tryptophan hydroxylase, tryptamine 5-hydroxylase, serotonin N-acetyltransferase, N-acetylserotonin methyltransferase and caffeic acid O-methyltransferase [18]. Melatonin contributes to prolonging the shelf life of broccoli and Chinese cabbage, and maintains high nutrient content during late storage [19][20][21]. Melatonin also can reduce ethylene synthase activity, restrict ethylene production, resulting in postponing the senescence and maintaining the fruit quality of pear and mango [14,22]. Furthermore, melatonin slows down the fruit softening during later storage by decreasing the expressions of MdERF109 in apples, and AdERF4, AdERF74, and AdERF75 in kiwifruit [23]. These findings provide sufficient proof that melatonin plays pivotal roles in delaying senescence, and regulates the senescence process by affecting ethylene release and the expression of ERFs. It is well known that ERFs family contains multiple family members, however, it has not identified which members are the pivotal regulators in procrastinating the senescence of flowering Chinese cabbage mediated by melatonin.
Flavonoid, a class of secondary metabolites and phytonutrients widely existing in plants, plays important role in growth and development, antioxidation, and senescence of plants [24,25]. Previous studies have shown that flavonoid accumulation delays postharvest aging (broccoli) and prolongs storage time (tomato) [26,27]. The high level of flavonoid and the expressions of flavonoid synthesis-related genes are beneficial to delaying the leaf senescence in rice [28]. Flavonoids also cross-talk with melatonin to delay aging [29,30]. For example, melatonin delays the senescence by enhancing the flavonoid biosynthesis and activating the antioxidant capacity in kiwifruit [29]. Flavonol is the most abundant and extensive distribution of flavonoid, and directly affects the total amount of flavonoid. Flavonol synthase (FLS), as a key enzyme in flavonol biosynthesis, directly determines the flavonol content in plant [31][32][33]. However, the molecular regulation mechanism of the critical FLSs involved in hindering leaf senescence mediated by melatonin is still unclear.
Flowering Chinese cabbage is widely cultivated in South China [34]. It is very prone to senescence, wilt and decay after harvest, and only be stored for 2-3 days in the air at ambient temperature due to its higher water content [35]. Nowadays, only a limited number of studies have explored the postharvest senescence of flowering Chinese cabbage under melatonin treatment exhibiting that leaf senescence is associated with antioxidant system activities and abscisic acid biosynthesis [20,21]. The current studies show that delaying leaf senescence is accompanied by the flavonoid accumulation in leaves [28]. So far, the underlying molecular mechanisms of melatonin involved in retarding the postharvest senescence of flowering Chinese cabbage by mediating ERFs to regulate the flavonoid biosynthesis-related genes have not been explored. We aimed to identify which ERFs and flavonoid biosynthesis-related genes are the pivotal candidates in delaying the postharvest senescence of flowering Chinese cabbage mediated by melatonin.

Melatonin Delaying Leaf Senescence was Associated with Chlorophyll Degradation, Ethylene Biosynthesis, and Flavonoid Accumulation
Flowering Chinese cabbage is a kind of leaf vegetable, and leaf yellowing is a typical characteristic of plant senescence. In this study, exogenous melatonin treatment significantly hindered the postharvest senescence of flowering Chinese cabbage. The leaves of the control progressively turned yellow after 10 days of storage and had no commercial value at 20 days of storage. In contrast, melatonin treatment kept the leaves greener during storage, and the leaves did not turn yellow until 20 days of storage ( Figure 1A). Furthermore,

BrFLS1 and BrFLS3.2, Key Genes for Flavonoid Biosynthesis, Played a Positive Regulatory Role in Melatonin-Delayed Leaf Senescence
Flavonol synthase (FLS) is a key enzyme in flavonoid biosynthesis. To further confirm the roles of flavonoid in leaf senescence under melatonin treatment, we analyzed the transcripts of nine flavonoid biosynthesis-related genes (BrPAL3, BrC4H, Br4CL, BrFLS1, BrFLS2, BrFLS3. 1,BrFLS3.2,BrFLS3.3,and BrFLS4). Figure 2 demonstrated that the transcripts of BrPAL3, BrC4H, Br4CL, BrFLS1, BrFLS2, BrFLS3.1, BrFLS3.2, and BrFLS4 were higher in melatonin-treated than that in control during 20 days storage, especially, the expressions of BrFLS1 and BrFLS3.2 were most significantly up-regulated. However, there was no significant difference in the transcript of BrFLS3.3 between melatonin treatment and control. Moreover, the expression of BrFLS1 reached a peak on day 10 of storage, and the expression of BrFLS3.2 reached the highest level on day 10 or 20 of storage in melatonin treatment. These findings suggested that the flavonoid biosynthesis-related genes were involved in regulating leaf senescence, while BrFLS1 and BrFLS3.2 might be the key candidate genes to modulate flavonoid biosynthesis under melatonin treatment. Ethylene has been identified as a hormone capable of directly enhancing leaf senescence [4]. In this study, we observed that ethylene production increased significantly in the control group during storage, but melatonin treatment reduced the ethylene production by 14.61, 18.77, 11.62, and 10.24% at 5, 10, 15, and 20 days of storage, respectively ( Figure 1C). As expected, melatonin significantly suppressed the expressions of ethylene biosynthesisrelated genes such as BrSAMS2.1, BrSAMS2.2, BrACS5, BrACS10, BrACO2, and BrACO5 compared with the control ( Figure S4).
Flavonoid are essential for the anti-oxidant and anti-senescence process of plants [24]. To confirm whether melatonin can regulate the formation of flavonoid, the total flavonoid content in flowering Chinese cabbage during the 20 days of storage was analyzed. Our results displayed that exogenous melatonin treatment significantly affected the flavonoid biosynthesis during storage, especially in late storage. The total content of flavonoid in melatonin-treated leaves was higher than that in the control, as evident by the reduced loss of total flavonoid from 10 to 20 days of storage ( Figure 1D).
The aforementioned results demonstrated that exogenous melatonin delaying the senescence of postharvest flowering Chinese cabbage was related to inhibiting chlorophyll degradation and ethylene biosynthesis, and promoting flavonoid accumulation.  To further illustrate the regulating functions of the mentioned above nine BrFLSs on flavonoid biosynthesis in the leaf senescence, we created transgenic flowering Chinese cabbage lines of TRV-BrFLSs using VIGS technology. Six transient silencing transgenic lines (TRV-BrFLS1, TRV-BrFLS2, TRV-BrFLS3.1, TRV-BrFLS3.2, TRV-BrFLS3.3, and TRV-BrFLS4) were obtained for a further experiment, and the silencing efficiency was verified by qRT-PCR ( Figure 3A). We observed that the leaf senescence of flowering Chinese cabbage was dramatically accelerated after silencing BrFLSs ( Figures 3B, S5 and S6). At 20 days after silencing, the base leaves of the TRV-BrFLSs plants began to turn yellow, and the TRV-BrFLS1 and TRV-BrFLS3.2 plants exhibited more serious symptoms compared with the TRV2 plants ( Figure S5). At 30 days after silencing, most of the leaves in TRV-BrFLSs plants exhibited serious yellowing phenotype or even fall off, especially in TRV-BrFLS1 and TRV-BrFLS3.2 plants (Figures 3B and S6). Moreover, the contents of chlorophyll, and flavonoids in the TRV-BrFLSs plants were lower than that in the TRV2 plants ( Figures 3C,D and S6), especially, those of the TRV-BrFLS1 and TRV-BrFLS3.2 plants were significantly decreased, and the transcription levels of chlorophyll degradation genes (BrNYC1, BrHCAR, BrRCCR, BrCLH, BrMCS, BrPPH, and BrSGR2) were significantly up- To further illustrate the regulating functions of the mentioned above nine BrFLSs on flavonoid biosynthesis in the leaf senescence, we created transgenic flowering Chinese cabbage lines of TRV-BrFLSs using VIGS technology. Six transient silencing transgenic lines (TRV-BrFLS1, TRV-BrFLS2, TRV-BrFLS3.1, TRV-BrFLS3.2, TRV-BrFLS3.3, and TRV-BrFLS4) were obtained for a further experiment, and the silencing efficiency was verified by qRT-PCR ( Figure 3A). We observed that the leaf senescence of flowering Chinese cabbage was dramatically accelerated after silencing BrFLSs ( Figure 3B, Figures S5 and S6). At 20 days after silencing, the base leaves of the TRV-BrFLSs plants began to turn yellow, and the TRV-BrFLS1 and TRV-BrFLS3.2 plants exhibited more serious symptoms compared with the TRV2 plants ( Figure S5). At 30 days after silencing, most of the leaves in TRV-BrFLSs plants exhibited serious yellowing phenotype or even fall off, especially in TRV-BrFLS1 and TRV-BrFLS3.2 plants ( Figure 3B and Figure S6). Moreover, the contents of chlorophyll, and flavonoids in the TRV-BrFLSs plants were lower than that in the TRV2 plants ( Figure 3C,D and Figure S6), especially, those of the TRV-BrFLS1 and TRV-BrFLS3.2 plants were significantly decreased, and the transcription levels of chlorophyll degradation genes (BrNYC1, BrHCAR, BrRCCR, BrCLH, BrMCS, BrPPH, and BrSGR2) were significantly up-regulated in the TRV-BrFLS1 and TRV-BrFLS3.2 plants ( Figure S7). . The letters or asterisks above columns indicate a significant difference at * p < 0.05.

BrERF2 and BrERF109, the Transcriptional Suppressors, Were Involved in Leaf Senescence and Repressed by Melatonin
Given that melatonin could limit ethylene release, we speculated that melatonin delay senescence may be associated with some key ERFs. To confirm this hypothesis, we isolated nine ERFs (Bra024954, Bra022115, Bra017495, Bra035792, Bra018864, Bra038261, Bra001784, Bra011529, and Bra017656) from the Chinese cabbage Chiifu genome database. Furthermore, spraying melatonin to TRV-BrFLSs plants slowed down the leaf senescence ( Figure 3B and Figure S6) and partially suppressed the expressions of chlorophyll degradation-related genes (BrNYC1, BrHCAR, BrRCCR, BrCLH, BrMCS, BrPPH, and BrSGR2) ( Figure S7). In contrast, the chlorophyll content in TRV-BrFLS1 and TRV-BrFLS3.2 plants treated with melatonin significantly increased by 55.06 and 45.74%, respectively ( Figure 3C), and the flavonoid content significantly increased by 15.25 and 16.46%, respectively ( Figure 3D). The above results further demonstrated that BrFLS1 and BrFLS3.2 were the pivotal genes involved in delaying leaf senescence and enhancing flavonoid biosynthesis, while melatonin could procrastinate the postharvest senescence by up-regulating the expressions of BrFLS1 and BrFLS3.2 to promote the flavonoid biosynthesis.

BrERF2 and BrERF109, the Transcriptional Suppressors, Were Involved in Leaf Senescence and Repressed by Melatonin
Given that melatonin could limit ethylene release, we speculated that melatonin delay senescence may be associated with some key ERFs. To confirm this hypothesis, we isolated nine ERFs (Bra024954, Bra022115, Bra017495, Bra035792, Bra018864, Bra038261, Bra001784, Bra011529, and Bra017656) from the Chinese cabbage Chiifu genome database. Multi-sequence alignment of the transcription factors amino acids disclosed that each sequence contained a conserved AP2 domain ( Figure S8), which could bind to the GCCbox (GCCGCC) or DRE/CRT element (CCGAC), which are the typical characteristics of ERFs proteins. Phylogenetic tree analysis was performed between ERFs identified in Arabidopsis and Bra024954, Bra022115, Bra017495, Bra035792, Bra018864, Bra038261, Bra001784, Bra011529, and Bra017656 in the Chinese cabbage genome database, they were named as BrERF2-like1, BrERF2-like2, BrERF2, BrERF3-like1, BrERF3-like2, BrERF3-like3, BrERF3-like4, BrERF109-like, and BrERF109, respectively ( Figure 4A). In addition, we observed that the expressions of the nine BrERFs increased dramatically with the extension of storage time, and reached the peak at 20 days of storage. Nevertheless, melatonin could suppress the expressions of these genes compared to the control, especially, the BrERF2 and BrERF109 among the nine BrERFs were most significantly suppressed, which would be the focus of our extended investigation ( Figure 4B). In contrast, the transcripts of BrERF2-like2, BrERF3-like2, and BrERF3-like4 differed relatively little between the control and melatonin treatment ( Figure 4B). Interestingly, we obtained similar results after spraying melatonin on pre-harvest plants showing that the expressions of the nine BrERFs was markedly inhibited. More importantly, the expressions of BrERF2 and BrERF109 were most significantly repressed, followed by BrERF2-like1, BrERF2-like2, and BrERF109-like1, while BrERF3-like1, BrERF3-like2, BrERF3-like3, and BrERF3-like4 were almost unaffected by melatonin ( Figure 4C and Figure S9). These results further suggested that BrERFs accelerated the leaf senescence process, while BrERF2 and BrERF109 may be the crucial candidates, which were suppressed by melatonin.
To further verify the transcriptional activity of BrERF2 and BrERF109, we performed Y2H and LUC experiments. The Y2H assay displayed that that yeast strains containing BD-BrERF2, BD-BrERF109, negative control, or positive control could grow normally in SD-Trp medium, but only positive control could grow in SD-Trp/-His medium (containing X-α-gal) and appeared blue, the rest of the yeast cells could not survive ( Figure S10). These results implied that BrERF2 and BrERF109 had no transcriptional activation activity. Furthermore, the LUC assay exhibited that the transcriptional activities of BrERF2 and BrERF109 in tobacco cells were significantly lower than that of negative control ( Figure 4D,E). These findings demonstrated that BrERF2 and BrERF109 were transcriptional inhibitors.

BrERF2 and BrERF109 Directly Bound to the Promoters of BrFLSs and Suppressed Their Expression
Provided that the expressions of BrFLS1 and BrFLS3.2 were increased, but the ethylene production and the transcripts of BrERF2 and BrERF109 were markedly suppressed under melatonin treatment. We further analyzed whether BrERF2 and BrERF109 could directly bind to the promoters of BrFLS1 and BrFLS3.2. We observed that the promoter regions of BrFLS1 and BrFLS3.2 contained at least one of the DRE motifs, which were possible binding cis-elements of BrERFs ( Figure 5A). The Y1H assay further demonstrated that Br-ERF2/BrERF109 could directly bind to their promoters of BrFLS1 and BrFLS3.2, respectively ( Figure 5B). Moreover, we investigated the regulations of BrERF2/BrERF109 on BrFLS1 and BrFLS3.2 promoters by GUS staining experiment and LUC assay in N. benthamiana leaves. GUS staining analysis revealed that when Pro35S::BrERF2/BrERF109 was co-infiltrated with ProBrFLS1::GUS or ProBrFLS3.2::GUS, the blue phenomenon was significantly lighter in the tobacco than that in the controls (null, ProBrFLS1, and ProBrFLS3.2) ( Figure 5C,D). Likewise, the relative LUC/REN ratio indicated that promoter activity of BrFLS1 and BrFLS3.2 was significantly weakened in the presence of BrERF2 or BrERF109 ( Figure 5E  The empty pBD (set as 1) and pBD-VP16 vectors were used as the negative and positive controls, respectively. Each value indicates the mean ± SD of six biological replicates. The asterisks above columns indicate a significant difference at * p < 0.05.

BrERF2 and BrERF109 Directly Bound to the Promoters of BrFLSs and Suppressed Their Expression
Provided that the expressions of BrFLS1 and BrFLS3.2 were increased, but the ethylene production and the transcripts of BrERF2 and BrERF109 were markedly suppressed under melatonin treatment. We further analyzed whether BrERF2 and BrERF109 could directly bind to the promoters of BrFLS1 and BrFLS3.2. We observed that the promoter The empty pBD (set as 1) and pBD-VP16 vectors were used as the negative and positive controls, respectively. Each value indicates the mean ± SD of six biological replicates. The asterisks above columns indicate a significant difference at * p < 0.05. co-infiltrated with ProBrFLS1::GUS or ProBrFLS3.2::GUS, the blue phenomenon was significantly lighter in the tobacco than that in the controls (null, ProBrFLS1, and ProBrFLS3.2) (Figure 5C,D). Likewise, the relative LUC/REN ratio indicated that promoter activity of BrFLS1 and BrFLS3.2 was significantly weakened in the presence of BrERF2 or BrERF109 ( Figure 5E,F). These results indicated that both BrERF2 and BrERF109 could inhibit the transcripts of BrFLS1 and BrFLS3.2 promoters.

Silencing BrERF2 and BrERF109 Promoted Flavonoid Biosynthesis and Delayed Aging
In order to further determine that inhibiting BrERF2 and BrERF109 could promote the biosynthesis of flavonoid and delay leaf aging, we generated the TRV-BrERF2 and TRV-BrERF109 transgenic plants of flowering Chinese cabbage. The two transient silencing transgenic lines and control plants (TRV2) of flowering Chinese cabbage were obtained for subsequent investigations ( Figure 6A). As expected, the expressions of BrFLS1 and BrFLS3.2 in the TRV-BrERF2 and TRV-BrERF109 plants were dramatically higher than that in the TRV2 plants (control) as detected by qRT-PCR ( Figure 6B). Additionally, as shown in Figure 6C, the leaves of the TRV-BrERF2 and TRV-BrERF109 plants were greener than that of the control plants (TRV2). The content of chlorophyll in the transgenic plants was significantly higher than that in the control, and similar results were observed in the determination of flavonoid content ( Figure 6D,E). These findings attested that BrERF2 and BrERF109 were involved in promoting the leaf senescence of flowering Chinese cabbage, and inhibiting the expressions of BrERF2 and BrERF109 could promote the transcriptions of BrFLS1 and BrFLS3.2 and the accumulation of flavonoid, thus delaying leaf aging. Collectively, our results further testified that melatonin could enhance the flavonoid accumulation by suppressing the inhibition of BrERF2 and BrERF109 on the transcription of BrFLS1 and BrFLS3.2, which resulted in delaying the leaf senescence of flowering Chinese cabbage.

BrERF2 Interacted with BrERF109, Which Formed a BrERF2-BrERF109 Complex
As we have known, both BrERF2 and BrERF109 could inhibit the expression of target genes BrFLS1 and BrFLS3.2, respectively, but it was not clear that whether they could interact with each other. To test this hypothesis, we firstly performed Y2H screen to investigate whether BrERF2 could interact with BrERF109. Interestingly, the results showed that they could physically interact with each other in yeast cells ( Figure 7A). To further verify this interaction in plant cells, we observed that YFP fluorescence was only detected in the nucleus of tobacco leaf cells harboring BrERF2-nYFP and BrERF109-cYFP by bimo-

BrERF2 Interacted with BrERF109, Which Formed a BrERF2-BrERF109 Complex
As we have known, both BrERF2 and BrERF109 could inhibit the expression of target genes BrFLS1 and BrFLS3.2, respectively, but it was not clear that whether they could interact with each other. To test this hypothesis, we firstly performed Y2H screen to in-vestigate whether BrERF2 could interact with BrERF109. Interestingly, the results showed that they could physically interact with each other in yeast cells ( Figure 7A). To further verify this interaction in plant cells, we observed that YFP fluorescence was only detected in the nucleus of tobacco leaf cells harboring BrERF2-nYFP and BrERF109-cYFP by bimolecular fluorescence complementation (BiFC) experiments, but not in cells transformed with BrERF2-nYFP and cYFP, BrERF109-cYFP and nYFP or nYFP and cYFP ( Figure 7B). Moreover, subcellular localization analysis indicated that BrERF2 and BrERF109 were consistently localized in the nucleus ( Figure 7C). The aforementioned results suggested that BrERF2 and BrERF109 are nucleus-localized transcriptional suppressors. Unfortunately, although BrERF2 physically interacted with BrERF109, which formed a BrERF2-BrERF109 complex, they did not synergically inhibit the expressions of BrFLS1 and BrFLS3.2 ( Figure S11). BrERF109 complex, they did not synergically inhibit the expressions of BrFLS1 and BrFLS3.2 ( Figure S11).

Discussion
Flowering Chinese cabbage, which belongs to Brassica, is one of the largest cultivated leaf vegetables in China. However, because its edible parts are tender stems and young leaves, it is easy to lose water, wilt, yellow, and cause nutritional degradation after harvest, which seriously affects its postharvest quality [34,36]. Previous studies have indicated that plant senescence is associated with chlorophyll degradation and ethylene production, while melatonin can effectively procrastinate the senescence process in postharvest flowering Chinese cabbage through delaying chlorophyll breakdown, reducing expressions of senescence-related genes, and enhancing the antioxidant system [20,21,[37][38][39]. Other reports also showed melatonin delay the aging of pear fruit by repressing the expressions of ethylene biosynthesis-related genes PcACS1 and PcACO1 [29]. These findings were similar to our results exhibiting that melatonin could effectively attenuate the leaf yellowing of postharvest flowering Chinese cabbage, which by reducing the chlorophyll degradation, and suppressing the expressions of chlorophyll degradation-related genes and senescence-associated gene (BrNYC1, BrNOL, BrPAO, BrRCCR, BrHCAR,  BrMCS, BrPPH1, BrPPH2, BrSGR1, BrSGR2, BrSGR3, and BrSAG15) (Figures 1A,B, S2 and S3). These results further confirmed that melatonin was a potent anti-senescent compound that could be employed to extend the shelf-life and maintain the quality of postharvest horticultural products. Nevertheless, to date, the molecular regulatory network of mela-

Discussion
Flowering Chinese cabbage, which belongs to Brassica, is one of the largest cultivated leaf vegetables in China. However, because its edible parts are tender stems and young leaves, it is easy to lose water, wilt, yellow, and cause nutritional degradation after harvest, which seriously affects its postharvest quality [34,36]. Previous studies have indicated that plant senescence is associated with chlorophyll degradation and ethylene production, while melatonin can effectively procrastinate the senescence process in postharvest flowering Chinese cabbage through delaying chlorophyll breakdown, reducing expressions of senescence-related genes, and enhancing the antioxidant system [20,21,[37][38][39]. Other reports also showed melatonin delay the aging of pear fruit by repressing the expressions of ethylene biosynthesis-related genes PcACS1 and PcACO1 [29]. These findings were similar to our results exhibiting that melatonin could effectively attenuate the leaf yellowing of postharvest flowering Chinese cabbage, which by reducing the chlorophyll degradation, and suppressing the expressions of chlorophyll degradation-related genes and senescenceassociated gene (BrNYC1, BrNOL, BrPAO, BrRCCR, BrHCAR, BrMCS, BrPPH1, BrPPH2, BrSGR1, BrSGR2, BrSGR3, and BrSAG15) ( Figure 1A,B, Figures S2 and S3). These results further confirmed that melatonin was a potent anti-senescent compound that could be employed to extend the shelf-life and maintain the quality of postharvest horticultural products. Nevertheless, to date, the molecular regulatory network of melatonin delaying the senescence of flowering Chinese cabbage remains predominantly insufficient.
Flavonoid is an important active phytochemical and antioxidant that is widely found in plants benefiting to human health [25,36,40]. Our recent studies demonstrated that the flavonoid biosynthesis was closely associated with the senescence of postharvest flowering Chinese cabbage [34]. It was reported that melatonin can promote flavonoid biosynthesis during the leaf senescence of kiwifruit [29]. So far, although melatonin can regulate the flavonoid content and the expressions of flavonoid biosynthesis-related genes in the aging process, it is not clear which genes play key roles and in what way they regulate leaf aging [29]. Here, we found that melatonin not only increased flavonoid accumulation but also up-regulated the expressions of flavonoid biosynthesis-related genes (BrPAL3, BrC4H, Br4CL, BrFLS1, BrFLS2, BrFLS3.1, BrFLS3.2, and BrFLS4) during postharvest senescence, especially the BrFLS1 and BrFLS3.2 ( Figures 1D and 2). Thus, we speculated that BrFLS1 and BrFLS3.2 were key flavonoid biosynthesis genes and potential regulators in delaying the leaf senescence mediated by melatonin. To verify whether BrFLS1 and BrFLS3.2 were involved in the regulation of leaf senescence, the VIGS system was performed in flowering Chinese cabbage. We observed that silencing BrFLS1 and BrFLS3.2 resulted in severe leaf yellowing, a dramatic decrease of flavonoid and chlorophyll content, and increasing the transcripts of chlorophyll degradation genes and senescence marker gene (BrNYC1, BrNOL, BrPAO, BrRCCR, BrHCAR, BrMCS, BrPPH1, BrPPH2, BrSGR1, BrSGR2, BrSGR3, and BrSAG15). Interestingly, melatonin reversed the onset of this phenomenon in BrFLS1silenced and BrFLS3.2-silenced plants, thus delaying the leaf senescence (Figure 3 and Figure S6). In contrast, the BrFLS2, BrFLS3.1, BrFLS3.3, and BrFLS4 did not play key roles in promoting flavonoid biosynthesis and delaying the leaf senescence mediated by melatonin, and may be present functional redundancy (Figures S5 and S6). Taken together, BrFLSs were involved in delaying the leaf senescence, and melatonin could enhance the flavonoid accumulation to delay the leaf senescence by up-regulating the expressions of BrFLSs, while the BrFLS1 and BrFLS3.2 were core contributors in flavonoid biosynthesis pathway. Nonetheless, the upstream transcription factors of BrFLS1 and BrFLS3.2 and their regulation mechanism associated with senescence remain undisclosed.
Ethylene as a negative regulatory signal factor, plays a vital regulatory role in the complex senescence process [4,5], while melatonin antagonizes ethylene to delay the senescence process in horticultural crops [14]. Zhai et al. [22] showed that melatonin delays the senescence of pear fruit by repressing the ethylene production and the expressions of ethylene biosynthesis genes PcACS1 and PcACO1. Our present study had similar findings showing that melatonin could significantly inhibit the ethylene release and the expressions of ethylene biosynthesis-related genes (BrSAMS2.1, BrSAMS2.2, BrACS5, BrACS10, BrACO2, and BrACO5) ( Figure 1C and Figure S4). ERFs are ethylene-responsive transcription factors, which regulate a variety of plant-specific processes. Nowadays, the intention of ERFs has been raising in regulating secondary plant metabolism and plant senescence [7]. Recently, ERF2 plays important roles in seed germination, regulation of calcium deficiency in postharvest fruits, and biotic stresses [13,41,42]. ERF109 regulates lateral root formation, anthocyanin synthesis, and mediates plant responses to abiotic stresses [43][44][45]. Nevertheless, the roles of ERF2 and ERF109 have not yet been explored in the postharvest senescence. In the present study, we isolated nine BrERFs members (BrERF2-like1, BrERF2-like2, BrERF2, BrERF3-like1, BrERF3-like2, BrERF3-like3, BrERF3-like4, BrERF109-like, and BrERF109) of flowering Chinese cabbage ( Figure 4B), and found that the BrERFs negatively regulated leaf senescence, however, melatonin could antagonize their transcriptions, especially the BrERF2 and BrERF109, implying that they were key BrERFs involved in modulating the senescence of postharvest flowering Chinese cabbage under melatonin treatment.
Given that down-regulating BrERF2 and BrERF109, and up-regulating BrFLS1 and Br-FLS3.2 could delay the leaf senescence of postharvest flowering Chinese cabbage, however, available information about their interactions to regulate aging is still limited. To verify whether BrERF2 and BrERF109 were involved in the regulation of BrFLS1 and BrFLS3.2 to delay the leaf senescence, the combination experiments of Y1H, GUS straining assays, and LUC experiences were executed. In this work, we observed that BrERF2 and BrERF109 directly bound to the promoters of BrFLS1 and BrFLS3.2 in vitro and in vivo, respectively, and inhibited their expression ( Figure 5). However, these inhibitory effects were significantly attenuated by melatonin, which implied that melatonin could inhibit BrERF2/BrERF109 mediated flavonoid biosynthesis ( Figure 4B,C). Moreover, we were pleased to discover that BrFLS1 and BrFLS3.2 were significantly up-regulated in the BrERF2-silenced and Br-ERF109-silenced plants, accompanied by a significant increase in chlorophyll and flavonoid contents, maintaining greener leaves in the silenced plants ( Figure 6). Taken together, these findings further confirmed that melatonin could promote the flavonoid accumulation by suppressing the inhibition of BrERF2/BrERF109 on the expressions of BrFLS1 and BrFLS3.2, thus delaying the aging of postharvest flowering Chinese cabbage.
Transcriptional regulation is a dynamic network; numerous previous studies found that ERFs regulate the transcription of target genes by binding to their cis-acting elements [6]. A previous study showed that MdERF2 inhibits MdACS1 transcription, and directly interacts with MdERF3 to suppress the binding of MdERF3 with MdACS1 promoter, thus negatively affecting ethylene biosynthesis and fruit ripening in apples [46,47]. In soybean, GmDREB1 forms heterodimers with ERFs to regulate plant drought resistance [48]. In this study, we observed that BrERF2 physically interacted with BrERF109 and formed a complex ( Figure 7A,B), however, they did not further synergically inhibit the expressions of BrFLS1 and BrFLS3.2 ( Figure S11). Therefore, whether BrERF2 and BrERF109 have synergistic regulations of delaying aging through mediating flavonoid biosynthesis needs to be further explored.

Plant Materials and Treatments
The '60 days' variety of flowering Chinese cabbage were used in this study. The fresh samples were harvested from Huachuang Farm (Zhongshan, China). To determine the optimal concentration of melatonin that can delay the leaf senescence of postharvest flowering Chinese cabbage, a pre-experiment was conducted. It was demonstrated that 100 µmol L −1 of melatonin (Sigma-Aldrich, Saint Louis, MO, USA) was a more effective dose to postpone the leaf senescence of postharvest Chinese flowering cabbage than treating with 0, 50, and 200 µmol L −1 ( Figure S1). Accordingly, 100 µmol L −1 of melatonin was selected for the subsequent experiments. The experiments were divided into two groups. One group was sprayed with 100 µmol L −1 melatonin and another group was sprayed with distilled water (control). After treatment, the samples were put into a polyethylene bags (40 cm × 30 cm) and stored promptly at 4 • C. Each treatment group contained three biological replicates and 30 plants per replicate. Relevant physiological and molecular indexes were investigated at 0, 5, 10, 15, and 20 days of storage, respectively.
Tobacco was grown in an artificial light incubator of 23 • C (16 h light/8 h dark), and five-week-old plants were selected for Agrobacterium tumefaciens-mediated transient expression experiments.

Determination of Ethylene Content
Ethylene production of flowering Chinese cabbages during storage was measured according to the method of Zhou et al. [49]. In brief, ten plants of flowering Chinese cabbage were placed in a 10 L closed container with a gas sampling port, and the ethylene gas was collected with a syringe (1 mL). Then, the ethylene samples were infused into the headspace vial and the ethylene concentration was determined using a gas chromatograph (CP-3800, Varian, Palo Alto, CA, USA) equipped with a flame ionization detector (FID) and fitted with the chromatographic column (GDX-102, 3 m × 2 mm i.d., Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian, China). Chromatographic conditions were as follows: column temperature 85 • C; injector temperature 120 • C; detector temperature 200 • C; carrier gas nitrogen (N 2 ), N 2 , hydrogen and compressed air at flow rates of 20, 30, and 300 mL min −1 , respectively. The sample was injected at a flow rate of 1.0 mL min −1 and the determination was repeated three times. The concentration of ethylene was quantified using an external standard method and results were expressed as µmol kg −1 h −1 .

Determination of Chlorophyll and Total Flavonoid Content
Chlorophyll a, chlorophyll b, and total chlorophyll contents of flowering Chinese cabbage were performed by previously reported methods [50]. Briefly, 0.5 g of fresh leaves were immersed in a mixed solution of acetone and ethanol (v:v = 1:1) until the leaves turned white (after 24 h). Then, the values of OD 645 and OD 663 of the extract were calculated by spectrophotometer (Uv-1800, Shimadzu, Shanghai, China).
The total flavonoid content of flowering Chinese cabbage was determined according to the previous method and modified moderately [51]. In brief, 0.5 g of samples were ground into powder in liquid nitrogen, then the powder was transferred to 8 mL absolute ethanol and stood for 30 min. The extract was centrifuged at 10,000 g for 15 min (4 • C). Rutin was used as the standard substance. The supernatant was gathered for the determination of the total flavonoid content. The samples were measured for total flavonoid content at a wavelength of 510 nm using UV-2410PC spectrophotometer (Shimadzu Corporation, Kyoto, Japan).

Total RNA Extraction, Quantitative Real-Time PCR (qRT-PCR) and Sequence Analysis
Total RNA was isolated from the flowering Chinese cabbage leaves using a TRIzol kit (ComWin Biotech Co., Ltd., Beijing, China) following the manufacturer's protocol. The first-strand cDNA was synthesized from 1 µg of total RNA, after digestion of genomic DNA, using PrimeScript™ RT Master Mix (Takara Biomedical Technology Co. Ltd, Dalian, China). Then, cDNA was used as a template for quantitative real-time PCR (qRT-PCR) analysis. qPCR was carried out with the Q711-ChamQTM Universal SYBR ® qPCR Master Mix (Vazyme, Nanjing, China) on Bio-Rad CFX96 Real-Time PCR System. qPCR was executed using 0.4 µL (10 µM) of forward and reverse primers, 10 µL 2 × ChamQ Universal SYBR qPCR Master Mix and 2 µL of template cDNA in a total volume of 20 µL. qPCR reaction conditions were described as follows: 30 s at 95 • C, 40 cycles of 10 s at 95 • C and 30 s at 60 • C. The actin used as a reference gene, gene-specific primers designed according to cDNA sequences as shown in supplemental Table S1. Relative gene expression levels were computed as described by previous [52], three different biological samples and three technical replicates were used for all quantitative experiments.
The Chinese cabbage Chiifu genome database (http://brassicadb.org/brad/, accessed on 16 September 2022) was used in this study. Sequence analysis was performed by utilizing MEGA7.0 and GeneDoc software to construct phylogenetic trees and multiple alignments, respectively.

Transcriptional Assay
The full-length coding sequences of BrERF2 and BrERF109 were fused with GAL4 DNA-binding domain into pGBKT7 vector. Then, the recombinant pGBKT7-BrERF2 and pGBKT7-BrERF109 fusion plasmids, positive control (pGBKT7-53 + pGADT7-T), or negative control (pGBKT7) were transformed into yeast two-hybrid (Y2H) Gold strain using the Yeastmaker Yeast Transformation System 2 (Takara Biomedical Technology Co., Ltd., Dalian, China), respectively. The trans-activation activity of BrERF2 and BrERF109 were assayed on basis of the growth status and α-galactosidase activity of yeast cells grown on SD/-Trp media or SD/-Trp-His-Ade media. The experiment was performed in triplicate.
The full-length regions of BrERF2 and BrERF109 were individually fused to the GAL4 DNA-binding domain of the pBD vector and acted as effector proteins. The reporter vector was modified from the pGreenII 0800-LUC vector, which included firefly luciferase (LUC) driven by a 35S promoter with five repeats of the upstream activating sequence, and a 35S promoter internal control of sub-driven renilla luciferase (REN). Each pair of effector and reporter plasmids was transiently transformed into tobacco leaves by Agrobacterium infiltration. After 48 h, the activities of LUC and REN were evaluated in Luminoskan Ascent microplate luminometer (Thermo Fisher Scientific, Waltham, MA, USA) using a LUC assay kit (Promega Corporation, Fitchburg, WI, USA). Eventually, the transcriptional activity of BrERF2 and BrERF109 was reflected by the value of LUC/REN. Six replicates were used for each experiment.

Yeast One-Hybrid (Y1H) Assay
The coding regions of BrERF2 and BrERF109 were cloned as bait constructs into the pGADT7 vector. The BrFLS1 (1125 bp upstream of the predicted translation start site) and BrFLS3.2 (570 bp upstream of the predicted translation start site) promoter fragments were ligated into the pAbAi vector as prey constructs. First, the bait constructs of pAbAi-BrFLS1 and pAbAi-BrFLS3.2 were transformed into Y1H yeast cells and cultured on SD/-Ura medium. Second, these preys and baits were co-transformed into Y1H yeast cells and cultivated on the SD/-Leu (Aba) medium using the Yeastmaker Yeast Transformation System 2 according to the manufacturer's protocol.

β-Glucuronidase (GUS) Analysis
The CDS segments of BrERF2 and BrERF109 were cloned into the pRI101 vector utilizing specific primers to generate the effector construct. The BrFLS1 (1125 bp) and BrFLS3.2 (570 bp) promoter fragments were cloned into the pBI101 vector containing the 35S promoter as reporter constructs. The reporters and effectors were transformed into Agrobacterium strain GV3101, and the bacterial liquid was co-infiltrated into N. benthamiana. The GUS staining experiment was performed with reference to Ma et al. [45]. Permeabilization for each experiment was executed in three biological replicates.

Dual Luciferase (LUC) Assay
The full-length CDS sequences of BrERF2 and BrERF109 were inserted into pGreenII 62-SK vector as effectors, and the promoter fragments of BrFLS1 and BrFLS3.2 were inserted into pGreenII 0800-LUC vector as reporters. These recombinant plasmids were transformed into Agrobacterium, and the Agrobacterium solution containing effectors and reporters were infected into tobacco leaves. The empty vector PGreenII 62-SK was used as a negative control. LUC activity was measured using the LUC gene test kit (Promega). Six biological replicates were conducted.

Y2H Screening
The coding regions of BrERF2 and BrERF109 were amplified and cloned into pGADT7 and pGBKT7 vectors, respectively. The fusion plasmid was co-transformed into strain Y2H Gold according to the rules of Yeastmaker Yeast Transformation System 2 and grown on DDO (SD/-Leu/-Trp) medium for 3 days. The transformed colonies were spotted on the selective medium QDO (SD/-Leu/-Trp/-Ade/-His) containing AbA and X-α-gal, to assess the possible interactions according to their growth status and colony turning blue.

Bimolecular Fluorescence Complementation (BiFC) Assay
The CDS of BrERF2 was integrated into the p2YC vector using the BamHI and SalI sites. The CDS of BrERF109 was linked into the p2YN vector using the BamHI and SalI sites [54]. These recombinant plasmids were injected into Agrobacterium, then the bacterial fluid infiltrated into the N. benthamiana leaves. The injected tobacco was incubated in the dark for 48 h, and YFP fluorescence was observed with a fluorescence microscope (Zeiss Axioskop 2 plus). This instantaneous expression experiment was repeated at least three times and shown representative results.

Statistical Analysis
All data were expressed as mean ± SD. R v4.1.1 and IBM SPSS Statistics 24 (SPSS Inc., Chicago, IL, USA) were used for statistical analysis. A Tukey's test (p ≤ 0.05) was performed to evaluate the treatment effects. All figures were generated using R v4.1.1 and Origin 2021 (OriginLab Corp., Northampton, MA, USA).

Primers
Details of the primers involved in Table S1.

Conclusions
In this study, a new regulatory pathway of melatonin delaying postharvest senescence was observed. We found that flavonoid could delay the plant senescence of flowering Chinese cabbage. Suppressing BrERF2 and BrERF109, and enhancing BrFLS1 and BrFLS3.2 expressions could promote flavonoid biosynthesis. BrERF2 and BrERF109 directly targeted the promoters of BrFLS1 or BrFLS3.2 and restrained their transcripts. Melatonin could inhibit the transcripts of BrERF2/BrERF109 and attenuate the inhibitory effects of ERFs on the transcripts of FLSs, ultimately increasing the level of flavonoid. Taken together, our results provided a novel insight into the molecular regulatory mechanism of melatonin involved in delaying postharvest senescence and confirmed for the first time that melatonin could enhance flavonoid biosynthesis by suppressing the inhibition of BrERF2/BrERF109 on the transcript of BrFLS1 and BrFLS3.2, which resulted in promoting flavonoid accumulation and delaying the postharvest senescence of flowering Chinese cabbage. Our findings enrich the molecular regulatory network of melatonin involved in delaying plant senescence, and might have important practical implications for improving commercial product qualities and greatly assist in extending the shelf life of postharvest crops.