A D-cysteine desulfhydrase, SlDCD2, participates in tomato fruit ripening by modulating ROS homoeostasis and ethylene biosynthesis

Abstract Hydrogen sulfide (H2S) is involved in multiple processes during plant growth and development. D-cysteine desulfhydrase (DCD) can produce H2S with D-cysteine as the substrate; however, the potential developmental roles of DCD have not been explored during the tomato lifecycle. In the present study, SlDCD2 showed increasing expression during fruit ripening. Compared with the control fruits, the silencing of SlDCD2 by pTRV2-SlDCD2 accelerated fruit ripening. A SlDCD2 gene-edited mutant was constructed by CRISPR/Cas9 transformation, and the mutant exhibited accelerated fruit ripening, decreased H2S release, higher total cysteine and ethylene contents, enhanced chlorophyll degradation and increased carotenoid accumulation. Additionally, the expression of multiple ripening-related genes, including NYC1, PAO, SGR1, PDS, PSY1, ACO1, ACS2, E4, CEL2, and EXP was enhanced during the dcd2 mutant tomato fruit ripening. Compared with the wild-type fruits, SlDCD2 mutation induced H2O2 and malondialdehyde (MDA) accumulation in fruits, which led to an imbalance in reactive oxygen species (ROS) metabolism. A correlation analysis indicated that H2O2 content was strongly positively correlated with carotenoids content, ethylene content and ripening-related gene expression and negatively correlated with the chlorophyll content. Additionally, the dcd2 mutant showed earlier leaf senescence, which may be due to disturbed ROS homeostasis. In short, our findings show that SlDCD2 is involved in H2S generation and that the reduction in endogenous H2S production in the dcd2 mutant causes accelerated fruit ripening and premature leaf senescence. Additionally, decreased H2S in the dcd2 mutant causes excessive H2O2 accumulation and increased ethylene release, suggesting a role of H2S and SlDCD2 in modulating ROS homeostasis and ethylene biosynthesis.


Introduction
Hydrogen sulfide (H 2 S) is emerging as a signaling molecule in all living organisms. Accumulating evidence over the past decades has shown that H 2 S participates in a vast number of physiological processes critical for plant growth and development, such as stomatal movement, f lowering, and responses to abiotic stresses [1][2][3][4]. Furthermore, multiple studies have revealed that H 2 S can delay leaf and f lower senescence and prevent the ripening of fruits, such as strawberry, tomato, and kiwifruit, by modulating the antioxidant system and antagonizing the ethylene pathway [5][6][7][8][9]. For example, H 2 S could maintain the freshness of cut f lowers by reducing the oxidative damage caused by excessive reactive oxygen species (ROS). Hu et al. [10] revealed that the senescence of mulberry fruit is delayed by exogenous H 2 S treatment and that the mechanism is related to the enhanced activity of antioxidant enzymes. Additionally, H 2 S counteracts the biological effect of ethylene and thereby alleviates the ripening and senescence of banana and tomato during postharvest storage [7,11]. Previous research has mainly revealed the mechanism through which exogenous H 2 S prevents fruit ripening, but the inf luence of endogenous H 2 S in fruit ripening still needs further investigation.
In higher plants, H 2 S emission has long been observed in pumpkin, cucumber, cantaloupe, corn, and others [12]. Currently, H 2 S can be generated during sulfur assimilation and cysteine decomposition. In the sulfate assimilation pathway, H 2 S is formed mainly through sulfite reductase (SiR), and sulfide is then integrated into the first organic sulfur-containing molecule cysteine by O-acetylserine thiol lyase (OAS-TL) [13]. Via another route, H 2 S is generated from L-cysteine by the catalysis of L-cysteine desulfhydrase (LCD) or from D-cysteine by D-cysteine desulfhydrase (DCD), but the two enzymes are not structurally related and are evolutionarily independent. DCD catalyzes the production of H 2 S from D-cysteine, which was first reported in Escherichia coli [14]. Two putative DCD homologous genes were later found in Arabidopsis, and these are called AtDCD1 (At1g48420) and AtDCD2 (At3g26115). DES1, a member of the OAS-TL family, is regarded as the main cysteine desulfhydrase in the cytosol of Arabidopsis cells [15]. In addition, β-cyanoalanine synthase (CAS) produces H 2 S during the detoxification of cyanide [16]. The common characteristic of these cysteine desulfhydrases is that they all use pyridoxal 5 -phosphate (PLP) as the cofactor. The mutation of DES1 or SiR as an endogenous H 2 S-generating enzyme results in early leaf senescence, which suggests that endogenous H 2 S may be a signal that inf luences the senescence process of plants [15,17,18].
Cysteine plays a critical role in plant primary and secondary metabolism through its incorporation into proteins and integration into sulfur-containing defense compounds [19]. L-cysteine, not D-cysteine, is the amino acid stereoisomer that is incorporated into peptides. D-amino acids have come to be regarded as physiological signaling molecules in mammals, particularly in the central nervous system [20]. For instance, endogenous D-cysteine in the mammalian brain acts as a negative impact factor of growth factor signaling during cortical development [20]. In plants, Dcysteine can be assembled into camalexin, which confers toxicity to several species of fungi [21]. DCD activity can be detected in multiple plants, such as Arabidopsis, Spinacia oleracea, Chlorella fusca, and Nicotiana tabacum, which suggests the presence of Dcysteine in plants. A serine/aspartate racemase has been identified as being involved in the generation of D-serine/aspartate in plants [22], but the potential for a cysteine racemase in plants has not yet been reported. The cysteine concentrations in the cytosol, which is the major site of cysteine synthesis, are estimated to be greater than 300 μM, whereas other cell compartments, such as the chloroplast and mitochondrion, each contain less than 10 μM cysteine [23]. The concentrations of free D-amino acids such as D-Asp, D-Asn, D-Glu, and D-Gln are approximately 0.2% to 8% relative to the corresponding L-amino acids in plants; however, cysteine data were not provided [24]. Therefore, research on Damino acids or their related metabolism in plants is an emerging field with many open questions.
In Arabidopsis, Cd-induced WRKY13 promotes the production of H 2 S by binding to the DCD (gene encoding DCD) promoter and thereby enhances plant tolerance to Cd [25]. In our previous study, we found that a SlLCD1 gene-edited mutant displays premature fruit ripening, which suggests that SlLCD1 is a negative impact factor of fruit ripening [26]. The role of SlLCD1 in natural and dark-induced leaf senescence was thus studied, and the results showed that SlLCD1 mutation causes earlier leaf senescence [9]. However, whether DCD participates in tomato fruit ripening and leaf senescence remains unclear.
Tomato is a typical climacteric fruit and an economically important vegetable worldwide [27]. Fruit ripening is associated with changes in metabolic pathways. Pigments, particularly chlorophyll and carotenoids, show substantial changes during tomato fruit ripening that lead to the change in fruit color from green to red [28]. In the ripening process, the degradation of chlorophyll begins with the conversion of chlorophyll b to chlorophyll a [29]. PAO (pheophorbide a monooxygenase) and PPH (pheophytin pheophorbide hydrolase) are then needed for subsequent chlorophyll degradation. In addition, SGR (STAY-GREEN) proteins can affect the degradation of chlorophyll by interacting with chlorophyll-degrading enzymes, and SGR1 in tomato can promote chlorophyll degradation. Tomato fruit ripening is alco accompanied by carotenoids biosynthesis; during this process, phytoene synthase (PSY) is the key rate-limiting enzyme, and phytoene dehydrogenase (PDS) is involved in subsequent steps leading to the formation of lycopene. Among plant hormones, ethylene elicits profound metabolic changes during tomato fruit ripening. During the ripening process, the ACC synthase-and ACC oxidase-encoding genes ACS1, ACS2, ACS4, ACO1, and ACO3 show increasing expression, which suggests their important roles in tomato fruit ripening. In a biological context, ROS, which are byproducts of natural aerobic metabolism, are composed of superoxide anion (O 2 •− ), hydrogen peroxide (H 2 O 2 ), and hydroxyl radicals (·OH), which have been previously associated with oxidative stress. However, accumulating evidence has revealed that ROS also operate as intracellular signaling molecules that participate in multiple physiological processes [30]. During peach fruit ripening, O 2 •− is required at the middle stage of fruit development, and H 2 O 2 acts as a potential signaling molecule to stimulate the late stage of fruit development [31]. In addition, H 2 O 2 is a signaling molecule that accumulates during the ripening of nectarine fruit and at the beginning of grape berry ripening [32,33]. Oxidative stress is also an integrative factor for triggering tomato fruit ripening, which suggests the signaling role of oxidants in initiating fruit ripening [34]. The interaction of H 2 S with other signals such as ethylene and ROS has attracted increasing attention. H 2 S alleviates oxidative stress by dynamically modulating antioxidant enzyme systems [35,36] and antagonizing the effect of ethylene; however, the mechanism through which endogenous H 2 S affect ROS homeostasis and ethylene biosynthesis during fruit ripening is largely unclear. Thus, in the present work, the function of SlDCD2 in tomato fruit ripening was investigated by virus-induced gene silencing (VIGS) and CRISPR/Cas9-mediated gene editing. In addition, the effect of SlDCD2 mutation on ROS homeostasis and ethylene biosynthesis was studied to reveal the role of SlDCD2 and H 2 S in modulating ROS and ethylene metabolism.

Phylogenetic analysis of DCDs and the expression profile and enzymatic activity of SlDCD2
To investigate the phylogenetic relationships between DCD proteins in tomato and other plant species, the gene encoding AtDCD1 (AT1G48420) in Arabidopsis was searched in the Phytozome v13 database, and the homologous proteins in Solanum lycopersicum, Arabidopsis thaliana, Vitis vinifera, and Glycine max were searched using the AtDCD1 protein sequence as the query. As shown in the phylogenetic tree in Fig. 1A, the identified DCDs could be classified into three groups. Subfamily I contained VvDCD1, SlDCD1, AtDCD1, and GmDCD1, and subfamily II contained VvDCD2, SlDCD2, AtDCD2, GmDCD2, and VvDCD3 and formed a single branch. The results indicated that SlDCD1/2 showed high homology with homologs in grapevine.
To explore the expression pattern of SlDCD2 in different tomato tissues, RNA from tomato roots, stems, leaves, f lowers, f lower buds, 1-cm fruits, mature green fruits, breaker fruits, and red fruits was reverse-transcribed into cDNA for RT-qPCR. As shown in Fig. 1B, the expression of SlDCD2 was relatively high in tomato leaves and increased gradually during fruit ripening, revealing the potential role of SlDCD2 in fruit ripening. To further investigate the potential role of SlDCD2 in fruit ripening, we determined the expression level of SlDCD2 in wild-type (WT) fruits at different ripening stages. The results showed that the relative expression of SlDCD2 increased significantly during fruit ripening, particularly at 41 DPA, 43 DPA, and 45 DPA (Fig. 1C). Therefore, SlDCD2 may play a more important role at the stage of fruit ripening.
To confirm the enzymatic activity of SlDCD2 in decomposing D-cysteine, we cloned the coding sequence of SlDCD2 into the pCOLD vector to express the N-terminal his-tagged SlDCD2 recombinant protein. SDS-PAGE and Western blot analyses indicated that the recombinant protein was highly expressed (Fig. 1D). Subsequently, the SlDCD2-His protein was purified by Ni-affinity chromatography (Fig. 1D). Using D-cysteine as the substrate, The SlDCD2 coding sequence was cloned into pCOLD, and protein expression in DE3 cells was induced by IPTG. SlDCD2 protein was then detected by Western blot using anti-His antibody. SlDCD2 protein was purified by nickel affinity chromatography using an Invitrogen Ni-NTA purification system. Lanes 1-4 present different imidazole elution fractions. The empty pCOLD vector was used as a negative control. M: protein molecular-weight markers in kDa. E Enzymatic activity of SlDCD2 in the presence of the substrate D/L-cysteine. The values are the means of three replicates ± their SDs based on the protein content. * * P < 0.01 as determined by Student's t test. Significant differences at the level of P < 0.05 are indicated with letters. the enzymatic activity of SlDCD2 was evaluated based on the H 2 S production rate. The enzymatic activity of the recombinant protein SlDCD2 was as high as 31.19 nmol • mg −1 (protein) • min −1 . We also used L-cysteine as a control to show the specificity of SlDCD2 enzyme activity. The enzyme activity observed with L-cysteine as the substrate was only 0.78 nmol • mg −1 (protein) • min −1 , which was not significantly different from that of the control group, and these findings confirmed the ability of SlDCD2 protein to catalyze the formation of H 2 S with D-cysteine and not L-cysteine as the substrate (Fig. 1E).

The transient silencing of SlDCD2 promotes tomato fruit ripening
The vector tobacco rattle virus (TRV)-SlDCD2 was used to transiently silence the expression of SlDCD2. As shown in Fig. 2A, after infection, SlDCD2-silenced fruit displayed yellow coloration on Day 22 and turned red on Day 25, whereas the control fruit was still green at Day 25. Chromaticity data showed that the fruit infected with TRV-SlDCD2 exhibited higher a * and b * values compared with the control fruit ( Fig. 2C and D), whereas no significant changes in lightness (L) were found (Fig. 2B). Tomato fruit ripening is accompanied by breakdown of the thylakoid structure, loss of photosynthetic capacity of the chloroplast, and carotenoids accumulation in the chromoplast of fruit cells. Thus, the contents of chlorophyll and carotenoids in SlDCD2silenced fruit were measured. Compared with the fruits of the control group, the SlDCD2-silenced fruits showed chlorophyll contents that were decreased by 16% and carotenoid contents that were increased by 74% at 20 days after infection. After 25 days of infection, the chlorophyll content in the SlDCD2-silenced fruit was decreased by 23%, and the carotenoids content was increased by 57%. Overall, the results showed that the contents of chlorophyll a (Fig. 2E), chlorophyll b (Fig. 2F), and total chlorophyll (Fig. 2G) in SlDCD2-silenced fruits were lower than those in control fruit, whereas the content of carotenoids (Fig. 2H) was higher in SlDCD2silenced fruit than in control tomato. In summary, the silencing of SlDCD2 accelerated chlorophyll degradation and carotenoids accumulation in the fruits.
The expression of SlDCD2 and ripening-associated genes was analyzed by RT-qPCR. The results shown in Fig. S1A (see online supplementary material) indicate that the expression of SlDCD2 in the SlDCD2-silenced fruit infected for 20 or 25 days was approximately 36% of that in the control fruit, suggesting that SlDCD2 was successfully silenced. To further verify the effect of SlDCD2 silencing on fruit ripening, we measured the expression of genes associated with fruit ripening and senescence, including the chlorophyll degradation genes NYC1, PAO, and SGR1, the carotenoids synthesis genes PSY1 and PDS, the ethylene synthesis genes ACO1 and ACS2, the ethylene response gene E4, and the cell wall metabolism gene CEL2, in the control and SlDCD2-silenced tomatoes. In SlDCD2silenced fruits infected for 25 days, the expression of NYC1 times that in the control group. Therefore, the results indicated that SlDCD2 silencing accelerated fruit ripening by improving the expression of ripening-associated genes and that SlDCD2 positively delayed the ripening process of tomato fruit.

Construction of an SlDCD2 gene-edited plant by CRISPR/Cas9
Two sgRNA targets of SlDCD2 were integrated into the CRISPR/ Cas9 vector, which was further transformed into the cultivar 'Micro Tom' by Agrobacterium-mediated transformation. For the genotyping of positive T2 plants, the gene fragments f lanking both sgRNA targets of SlDCD2 were amplified from genomic DNA of the dcd2 mutants. In addition, SlDCD2 CDS amplification products with cDNA from transformed tomato lines as the template were sequenced to verify that the CDS of SlDCD2 was mutated. Analysis of the plant genome ( As shown in Fig. 2J, the height of the WT was approximately 26 cm, whereas the height of the dcd2-1 mutant was only 22 cm, and that of the dcd2-5 mutant was approximately 23 cm, suggesting a minor effect of SlDCD2 deletion on plant growth. To further verify the mechanism through which SlDCD2 participates in fruit ripening, the ripening process in the WT and dcd2 fruits was monitored, and dcd2-1 and dcd2-5 mutants showed a similar early ripening phenotype. As shown in Fig. 2K, dcd2-1 and dcd2-5 fruits entered the breaker stage at 35 days post anthesis (DPA) and turned red at 39 DPA. However, the WT fruits began to turn yellow at 43 DPA and became completely red at 47 DPA. Based on the above-described result, the dcd2-1 and dcd2-5 fruits ripened 8 days earlier than the WT fruits, suggesting that the mutation of SlDCD2 accelerated the ripening of tomato fruit and that SlDCD2 delayed the ripening process of tomato fruit. D-cysteine is degraded by DCD2 to produce H 2 S, ammonia, and pyruvate. To rule out the effect of pyruvate and ammonia, the contents of which may be slightly reduced in dcd2 deletion mutants, WT tomato fruits at the white mature stage were soaked with H 2 O, 50 μM NaHS solution (H 2 S treatment group), 50 μM pyruvate solution, or 50 μM ammonium chloride solution for 8 h. The results presented in Fig. S4 (see online supplementary material) show that the fruits in the H 2 S group exhibited delayed fruit ripening compared with those in the H 2 O control group. Moreover, pyruvate and ammonium chloride treatment did not delay fruit ripening in comparison to the control, suggesting that H 2 S is the signal that delays fruit ripening and that the acceleration of fruit ripening in the dcd2 deletion mutant could be attributed to a decreased H 2 S content instead of the effects of pyruvate and ammonia.
To confirm the effect of SlDCD2 mutation on endogenous H 2 S production, we first measured the endogenous H 2 S content in dcd2-1/5 mutant leaves using lead acetate H 2 S detection strips. The results presented in Fig. 3A show that dcd2-1/5 mutant leaves produced less H 2 S with D-Cys as the substrate, and their gray intensity in the strips was significantly lower than that found for the WT leaves. We also verified the difference in H 2 S production with L-Cys and D-Cys as substrates using the methylene blue method. The rate of H 2 S production in WT leaves was approximately 1.5 times higher than that in dcd2 mutant leaves when D-Cys was used as the substrate. However, the difference between dcd2 mutants and WT was not significant when L-Cys was used as the substrate, and that between dcd2-1/5 mutant and WT leaves was not significant when L-Cys was used as the substrate (Fig. 3A), suggesting that D-cysteine rather than L-cysteine is the only substrate for SlDCD2. We then measured the endogenous H 2 S content in WT and dcd2-1 fruits at 36, 39, and 43 DPA using lead acetate H 2 S detection strips. The results showed that dcd2-1 fruits at 39 and 43 DPA produced less H 2 S with D-Cys as the substrate, and the gray intensity of the strips was significantly lower than that found for the WT (Fig. 3B). In addition, we determined the H 2 S production rate in the fruits using the methylene blue method. The results show that the production rate of H 2 S in the WT fruits at 39 and 43 DPA was approximately 1.3 and 1.2 times higher than that in the dcd2-1-mutant fruits at 39 and 43 DPA, respectively (Fig. 3B). When L-Cys was used as the substrate, the difference in the H 2 S content and production rate between WT and dcd2-1 fruits was not significant, suggesting that the mutation of SlDCD2 reduces H 2 S production during tomato fruit ripening and accelerates fruit ripening (Fig. 3B).
In addition, due to the ability of SlDCD2 to decompose Dcysteine, we propose that dcd2-1/5 mutations may cause the accumulation of total cysteine. It is difficult to discriminate L-cysteine from D-cysteine; therefore, we determined the content of total cysteine in tomato leaves. Consistently, Fig. 3C shows that the dcd2-1 mutation resulted in a significantly higher level of total cysteine than that in WT, implying that the mutation damaged the enzymatic function of SlDCD2 as a DCD.
Previous studies have indicated that H 2 S significantly alleviates fruit ripening and senescence by attenuating the effect of ethylene. The antagonistic effect of H 2 S and ethylene may play an important role in fruit senescence. To test this hypothesis, we measured the ethylene content of WT and dcd2-1 mutant fruits during ripening and found that the ethylene content in dcd2-1 fruit at 36, 39, and 43 DPA was significantly higher than that in WT (Fig. 3D). In particular, there was a higher increase in ethylene content in dcd2-1 fruits at 39 DPA than in WT fruits. Correlation analysis indicated a negative correlation between H 2 S produced by either L-cysteine or D-cysteine as a substrate and ethylene content. Specifically, the correlation between ethylene and H 2 S produced by L-cysteine was −0.65, and the correlation between ethylene and H 2 S produced by D-cysteine was −0.46 (Fig. 3F). Therefore, the mutation of SlDCD2 affects ethylene biosynthesis, which accelerates fruit ripening.
Subsequently, we treated WT fruits at 36 DPA with 1 g/L ethephon aqueous solution fumigation, and the expression of SlDCD2 under ethylene treatment was studied by RT-qPCR. The results indicated that ethylene treatment caused decreased expression of SlDCD2 at 3 h, 6 h, and 12 h compared with the control (Fig. 3E). At 24 h, SlDCD2 expression in fruits of both groups gradually returned to the pretreatment level (Fig. 3E). Therefore, we found that ethylene decreased the expression of SlDCD2.

Effect of SlDCD2 mutation on the chlorophyll and carotenoid contents and the expression of ripening-related genes during the fruit ripening process
In the dcd2 mutant, the mutation of SlDCD2 accelerated fruit ripening; therefore, we determined the chlorophyll and carotenoid contents in the dcd2-1 and dcd2-5 fruits at 36, 39, and 43 DPA. Fig. 4A-C show that the contents of chlorophyll a, chlorophyll b, and total chlorophyll decreased gradually from 36 to 43 DPA in the WT, dcd2-1, and dcd2-5 fruits, whereas the chlorophyll content was significantly lower in the dcd2 mutants than in the WT. Fig. 4D shows an increasing carotenoids content during fruit ripening in the WT and the dcd2-1 and dcd2-5 mutants, and the carotenoids content in the dcd2-1 fruits was approximately 3 and 3.4 times that in the WT fruits at 39 and 43 DPA, respectively. Overall, the two mutant lines dcd2-1 and dcd2-5 showed similar accelerated fruit ripening, and the results from the analysis of the metabolism of chlorophyll and carotenoids further imply a negative role of SlDCD2 in fruit ripening.
To further analyse the role of SlDCD2 in delaying fruit ripening, the expression patterns of ripening-associated genes in dcd2-1 and WT fruits were analysed by RT-qPCR. The expression of chlorophyll degradation-related genes including NYC1 (Fig. 4E), PAO (Fig. 4F), and SGR1 (Fig. 4G) in dcd2-1 fruits showed an increasing pattern from 36 to 43 DPA, and these expression levels were always higher than those in WT fruits. For instance, the expression of NYC1, PAO, and SGR1 in dcd2-1 fruits at 43 DPA was approximately 3-4 times that in WT fruits. The expression levels of PDS and PSY1, which encode key enzymes for carotenoids biosynthesis, were also analysed. PDS (Fig. 4H) and PSY1 (Fig. 4I) increased gradually in WT fruits during ripening, but their expression increased significantly in dcd2-1 fruits.
The ripening of tomato fruit, as a respiratory climacteric fruit, relies heavily on ethylene synthesis and response pathways. Fig. 4J shows that the expression of ACO1 increased in both the WT and dcd2-1 fruits during fruit ripening, whereas dcd2 mutation resulted in significantly higher expression of ACO1 at 36 and 39 DPA compared with that in the WT fruits. The expression of ACS2 (Fig. 4K) and E4 (Fig. 4L) increased during fruit ripening in the WT fruits, and significantly higher levels of ACS2 and E4 were observed in dcd2-1 fruits. The expression of CEL2 (Fig. 4M) and EXP (Fig. 4N), which are needed for cell wall metabolism, was also analysed. As shown in Fig. 6I and J, the expression of CEL2 and EXP in dcd2-1 fruits was approximately 12 and 10 times that in the WT fruits, respectively, at 39 DPA. Based on the abovedescribed results, the mutation of SlDCD2 increased the expression of ripening-associated genes during the tomato fruit ripening process, resulting in accelerated fruit ripening.

Effects of SlDCD2 mutation on ROS homeostasis
ROS in plants can be generated as byproducts during normal metabolic processes or as responses to stress conditions. ROS are also thought to be important impact factors of fruit ripening. The exogenous application of H 2 S could decrease ROS overaccumulation in multiple plant species; thus, we propose that SlDCD2 may delay fruit ripening by modulating ROS homeostasis. As shown in Fig. 5A, the H 2 O 2 level in both WT and dcd2-1 fruits increased significantly during fruit ripening, whereas the level in dcd2-1 fruits was significantly higher than that in WT fruits at 39 DPA.
The production of O 2 •− was not significantly different between WT and dcd2-1 fruits (Fig. 5B). As shown in Fig. 5C, MDA, a product of lipid peroxidation, did not show marked changes in WT fruits during ripening but tended to decrease in dcd2-1 fruits, but the content of MDA in dcd2-1 fruits was significantly higher than that in WT fruits at 36 and 39 DPA. The activities of antioxidative enzymes, including superoxide dismutase (SOD), catalase (CAT), ascorbate peroxidase (APX), and peroxidase (POD), were analysed. Fig. 5D shows that SOD activity in dcd2-1 fruits was always higher than that in WT fruits during fruit ripening. CAT and APX activity could help alleviate oxidative damage by decomposing H 2 O 2 to below toxic levels. CAT activity was increased in WT fruits at 36 and 39 DPA but decreased in dcd2-1 fruits at 39 DPA to a level that was about 75% of that in WT fruits (Fig. 5E). A decreasing trend of APX activity was observed in WT fruits during fruit ripening (Fig. 5F). Compared with the results obtained for the WT fruits, SlDCD2 mutation caused a significantly lower level of APX activity at 39 DPA and a significantly higher level at 43 DPA. POD activity in both WT and dcd2-1 fruits decreased gradually during fruit ripening, whereas the activity in dcd2-1 fruits was significantly higher than that in WT fruits (Fig. 5G). Moreover, the expression of genes encoding antioxidative enzymes, including SOD, CAT, APX, POD1, and POD2 was analysed. As shown in Fig. 5H, SOD expression was lower in dcd2-1 fruits at 36 and 43 DPA. As shown in Fig. 5I, a decreasing trend of CAT expression was observed in WT and dcd2-1 fruits, and SlDCD2 mutation caused significantly higher CAT expression at 36 DPA and significantly lower CAT expression at 43 DPA compared with the results observed in the WT fruits. As shown in Fig. 5J, K, and L, the expression patterns of APX, POD1, and POD2 were similar during fruit ripening. Their expression increased at 39 DPA and then decreased at 43 DPA, and the expression levels of APX, POD1, and POD2 were significantly lower in dcd2-1 fruits than in WT fruits.
Overall, SlDCD2 mutation induced excessive accumulation of H 2 O 2 and MDA, suggesting that the lower level of H 2 S in dcd2-1 fruits may lead to an imbalance in ROS metabolism and that excessive ROS may accelerate fruit ripening. As the primary enzyme for the decomposition of H 2 O 2 , CAT undergoes persulfidation in the leaves of Arabidopsis [37,38]. In the current study, decreased CAT activity was observed in dcd2-1 fruits; thus, we propose that an appropriate endogenous level of H 2 S may be needed for altering the activity of CAT. CAT extracted from tomato leaf tissue was treated with different levels of H 2 S, and the results in Fig. 5M indicate that 50 μM NaHS (H 2 S donor) could activate CAT. WT and dcd2-1 mutant fruits were also treated with 50 μM NaHS, and the results showed that CAT activity was enhanced (Fig. 5N). Subsequently, tomato fruits at 36 DPA were treated with 5 mmol/L H 2 O 2 to evaluate the potential interaction between H 2 O 2 and SlDCD2. The expression of SlDCD2 in H 2 O 2 -treated fruits after 3 h and 6 h of treatment was not significantly different from that in the control fruits, but at 12 and 24 h, the expression in the H 2 O 2 -treated fruits was significantly higher than that in the control fruits (Fig. 5O). Therefore, exogenous H 2 O 2 application to tomato fruits stimulated the expression of SlDCD2, which may lead to an increase in H 2 S production in the fruits and thereby alleviate H 2 O 2 -induced oxidative stress.
To reveal the potential associations among the parameters, then correlations among the chlorophyll a, chlorophyll b, total chlorophyll, carotenoids content, and H 2 O 2 contents, O 2 •− production rate, MDA content, and expression of ripening-related genes were analysed. As shown in Fig. 5P, the chlorophyll content was negatively correlated with the expression of ripening and

Effect of SlDCD2 mutation on ROS homeostasis in tomato leaves
To investigate the universal role of H 2 S generated by DCD in modulating ROS homeostasis and the MDA content, the activities of antioxidative enzymes and their gene expression levels in WT and dcd2-1 tomato leaves were determined. As shown in Fig. 6A in WT and dcd2-1 leaves did not differ significantly based on NBT staining (Fig. 6A), but a higher level of H 2 O 2 was found in dcd2 mutant leaves as revealed by DAB staining (Fig. 6B). The results  Fig. 6C suggested that SlDCD2 mutation resulted in increased H 2 O 2 in comparison to the levels found in WT plants, whereas the O 2 •− production rate did not significantly differ between WT and dcd2-1 tomato leaves (Fig. 6D). Additionally, SlDCD2 mutation decreased the activities of CAT (Fig. 6E), POD (Fig. 6G), and SOD (Fig. 6H) but did not significantly change the activity of APX (Fig. 6F). Furthermore, SlDCD2 mutation caused significantly lower gene expression of CAT (Fig. 6J), POD1 (Fig. 6L), and POD2 (Fig. 6M) and higher expression of APX (Fig. 6K) and did not significantly change the expression of SOD (Fig. 6I) compared with the levels found in the WT plants. Overall, the results indicate that SlDCD2 mutation also accelerated leaf senescence, which may be due to disturbed ROS homeostasis.

Discussion
In our previous study, the SlLCD1 gene-edited mutant displayed accelerated fruit ripening and earlier leaf senescence, suggesting that SlLCD1 is a negative impact factor of fruit ripening and leaf senescence [26]. However, the role of DCD in tomato fruit ripening and leaf senescence remains unclear. In the present work, the proteins in S. lycopersicum, A. thaliana, V. vinifera, and G. max that are homologous to AtDCD1 (AT1G48420) were obtained. SlDCD1/2 show high homology with homologs in grapevine, and SlDCDs are evolutionarily closer to VvDCDs. Previous studies have shown that the application of exogenous H 2 S effectively alleviates postharvest senescence in grape [39,40], but whether DCD inf luences grape ripening and senescence is unclear. Because SlDCD2 expression shows an increasing trend during fruit ripening, we propose that SlDCD2 may be involved in fruit ripening. First, the expression of SlDCD2 was transiently silenced by VIGS, and accelerated fruit ripening was observed in TRV2-SlDCD2 fruits compared with fruits carrying the empty vector. The expression of SlDCD2 in SlDCD2-slicenced fruits was approximately half that found in the control fruits, suggesting the efficiency of the gene silencing. Additionally, the chlorophyll content in the SlDCD2-silenced fruits degraded more rapidly than that in the control fruits, and greater carotenoids accumulation was detected in SlDCD2-silenced fruits. In addition, SlDCD2 silencing increased the expression of the chlorophyll degradation genes NYC1, PAO, and SGR1, the carotenoids synthesis genes PSY1 and PDS, the ethylene synthesis genes ACO1, ACS2, and E4, and the cell wall metabolism gene CEL2, suggesting the potential negative role of SlDCD2 in fruit ripening.
The role of SlDCD2 was confirmed by stable SlDCD2 gene editing of tomato lines using CRISPR/Cas9 technology. Both dcd2-1 and dcd2-5 showed decreased H 2 S release, accelerated fruit ripening and an increase in the content of cysteine. D-cysteine and L-cysteine were supplied as substrates for the dcd2-1 and dcd2-5 mutants, and the mutants only showed decreased H 2 S production when the substrate was D-cysteine. In addition, the substrate selectivity was confirmed by purified SlDCD2 protein, which only used D-cysteine as the substrate. However, AtDCD2 was found to catalyze the production of H 2 S from both D-cysteine and Lcysteine [19], suggesting that the function of DCD2 is not highly conserved in Arabidopsis and tomato. The levels of chlorophyll and carotenoids in the dcd2 mutant fruits were analysed, and the results indicated that SlDCD2 mutation caused accelerated chlorophyll degradation and premature carotenoids biosynthesis.
Previous studies indicate that H 2 S delays fruit ripening by antagonizing the effect of ethylene. The results show that H 2 S prolongs the postharvest ripening and senescence of banana during storage by antagonizing the effect of ethylene [11]. In the present study, the ethylene content in dcd2-1 fruits was significantly higher than that in WT fruits. The correlation analysis indicated a negative correlation between H 2 S produced with either L-cysteine or D-cysteine as the substrate and the ethylene content. In addition, exogenous ethylene treatment decreased the expression of SlDCD2 compared with that obtained with the control treatment. Therefore, mutation of SlDCD2 may induce enhanced ethylene biosynthesis and thereby accelerate fruit ripening, which further suggests the antagonizing effect between H 2 S and ethylene.
The effect of reduced H 2 S on the expression of ripeningassociated genes was also evaluated in the dcd2 mutant. The expression of NYC1, SGR1, and PAO was highly upregulated during dcd2 fruit ripening. In addition, the expression of the carotenoids biosynthesis genes PSY1 and PDS was upregulated in dcd2 mutant fruits. Tomato fruit ripening is also accompanied by increases in ethylene synthesis and cell wall degradation. The present study indicates that ACO1, ACS2, and E4 were all expressed at obviously higher levels in dcd2 than that in WT tomato fruits. Our data also show that the dcd2 mutant showed increased expression of CEL2 and EXP.
Exogenous H 2 S was found to delay the ripening and senescence of multiple plants by reducing excessive ROS [5][6][7]. The production of ROS during fruit ripening and senescence is natural and unavoidable due to imbalanced metabolism. However, how endogenous H 2 S affects ROS homeostasis remains unclear. Therefore, we determined the ROS content during fruit ripening. The data indicate that H 2 O 2 was significantly increased in both WT and dcd2-1 fruits during fruit ripening and that the content in dcd2-1 fruits was clearly higher than that in WT fruits. The production of O 2 •− did not significantly differ between WT and dcd2-1 plants. The dcd2-1 mutants also showed increased MDA contents compared with the WT at 36 and 39 DPA. In a previous study, Pilati et al. [33] observed a rapid accumulation of H 2 O 2 during grape ripening, which may be associated with grape softening and ripening. In addition, the H 2 O 2 levels linearly increased as the nectarine fruit developed and ripened on trees [32]. The promoting effect of H 2 O 2 on fruit ripening has also been observed in pears [41]. In the present study, H 2 O 2 was increased at 39 DPA in dcd2-1 fruits, which had just completed their transition from the breaker to the red stage. Consistently, previous studies observed elevations in the H 2 O 2 levels in tomatoes, which also exhibited changes in their skin color [42]. Therefore, fruit ripening appears to be positively associated with H 2 O 2 . The results indicate that H 2 O 2 is the ROS signal that participates in the tomato fruit ripening process. Additionally, the enzymatic activities of SOD, CAT, APX, and POD were analysed, and the results showed that SOD activity in dcd2-1 fruits was always higher than that in WT fruits during fruit ripening. CAT

Phylogenetic analysis
Putative DCD proteins in S. lycopersicum, A. thaliana, V. vinifera, and G. max were obtained with the BLASTP tool in the Phytozome v13 (https://phytozome.jgi.doe.gov/pz/portal.html#) database using AtDCD1 (AT1G48420) as the query. The amino acid sequences of SlDCD1 and SlDCD2 from S. lycopersicum; AtDCD1 and AtDCD2 from A. thaliana; VvDCD1, VvDCD2, and VvDCD3 from V. vinifera; and GmDCD1 and GmDCD2 from G. max were selected to construct a phylogenetic tree using the neighbor-joining method according to the parameters previously reported by Saitou et al. [47].

Expression, purification, and enzymatic activity assay of the recombinant protein SlDCD2 in E. coli
The 1329-bp coding sequence of SlDCD2 (accession number: Solyc01g008900) was amplified using the primers listed in Table S1 (see online supplementary material). The coding sequence of SlDCD2 was then ligated into the pCOLD vector, and the confirmed pCOLD-SlDCD2 plasmid was transformed into DE3 cells. Subsequently, the expression, collection, and purification of the SlDCD2 protein were performed using the methods described by Cheng et al. [48]. The expression and purification of the SlDCD2 protein were confirmed by SDS-PAGE and staining with Coomassie Blue. The detection of the SlDCD2 protein was verified by Western blot using anti-His antibody.
L/D-Cysteine was used as the substrate as previously described [49], and the enzymatic activity of DCD was determined at 670 nm. Standard curves were prepared using solutions of NaHS at different concentrations.

Transient silencing of SlDCD2 in tomato fruit
The fragment (with a length of 400 bp corresponding to nt 1-400) encoding SlDCD2 was amplified and inserted into the pTRV2 plasmid to yield recombinant pTRV2-SlDCD2. The fragment of the SlDCD2 gene was amplified with specific primers, which are listed in Table S1 (see online supplementary material). As described previously by Fantini et al. [50], Agrobacterium containing the pTRV1 vector and the corresponding pTRV2 vector was inoculated into pedicels of tomato plants after mixing at a ratio of 1:1, and infected tomato petioles infiltrated with the empty pTRV2 vector were used as controls. Tomato (S. lycopersicum, Micro Tom) plants were first incubated in the dark at 16 • C for 24 h and then incubated under the following conditions: 16-h day/8-h night cycle, 25 ± 2 • C/20 ± 2 • C day/night temperature, 65% relative humidity, and 250 μmol m −2 s −1 light intensity. The appearance of the tomato fruit color was assayed using a color difference meter (model WSC-100; Konica Minolta, Tokyo, Japan).

Generation and genotyping of the dcd2 mutant by CRISPR/Cas9
CRISPR/Cas9 mutagenesis of SlDCD2 in tomato was performed as previously described [51][52][53]. The primers for sgRNA are listed in Table S1 (see online supplementary material). For confirmation of the dcd2 mutant, we amplified a fragment of the sgRNA target sequence using genomic DNA from the dcd2 mutant. The amplification product was used for DNA sequencing, and the genotyping of tomato plants was analysed on the website DSDecodeM(http:// skl.scau.edu.cn/dsdecode/) [54].
mRNA from WT and dcd2 mutants were extracted and reverse transcribed into cDNA, and the cDNA was used as the template for amplification of the CDS of SlDCD2. The amplification products were subsequently sequenced. The sequencing peak maps were analysed using Chromas software.

Treatment of tomato fruit with pyruvate, NH 4 Cl, or NaHS
WT tomato fruits at the white mature stage were soaked with H 2 O, 50 μM NaHS solution (H 2 S treatment group), 50 μM pyruvate solution, or 50 μM ammonium chloride solution for 8 h. Subsequently, the fruits were dried and placed on wet sterile filter papers in Petri dishes. The Petri dishes, which contained 10 fruits, were stored at 23 • C for 4 days and photographed.

Treatment of tomato fruit using H 2 O 2 or ethylene
According to previous studies [11,55], tomato fruits at 36 DPA were fumigated with 1 g/L ethephon aqueous solution or sprayed with 5 mmol/L H 2 O 2 . Fruit samples collected after 0 h, 3 h, 6 h, 12 h, and 24 h of treatment were used for mRNA extraction and qPCR.

RNA extraction and RT-qPCR
Reactions were conducted using previously reported methods [9]. Tomato SlTubulin was used as an internal reference. The primers used for RT-qPCR analysis are listed in Table S1 (see online supplementary material).

Determination of the levels of chlorophyll, carotenoids, and cysteine in tomato fruit
A tomato fruit sample (0.5 g) without seeds was extracted in ethanol, and quantitative determination of the chlorophyll and carotenoid contents was conducted. The chlorophyll and carotenoids levels were measured and calculated based on the equations described by Wellburn [56].
A cysteine assay kit (Solarbio, Beijing, China) was used to determine the content of cysteine. Samples (0.2 g) of WT and dcd2 tomato leaves were used for cysteine determination according to the manufacturer's instructions.

Determination of the amount of H 2 S release in tomato leaves and gray intensity analysis
As mentioned previously [9], the release of H 2 S in 1.0 g of leaf or fruit samples was determined using zinc acetate test strips (Sigma, Darmstadt, Germany). The amount of H 2 S release is related to the color of the zinc acetate test strips. A gray intensity analysis of the zinc acetate test strips was performed using ImageJ software.
The release of H 2 S in 1.0 g of leaf or fruit samples was also determined using the methylene blue spectrophotometric method [57,58]. Leaf or fruit samples were ground to a powder in liquid nitrogen, homogenized in 5 mL of buffer (containing 100 mM potassium phosphate buffer pH 7.4, 10 mM Cys, and 2 mM pyridoxal 5 -phosphate) and then centrifuged to obtain the supernatant. After centrifugation, the released H 2 S was adsorbed with zinc acetate and further reacted with N,Ndimethyl-phenylenediamine (DPD) with FeCl 3 to form methylene blue, which was detected colorimetrically at 670 nm.

Determination of the ethylene content
The ethylene content was determined as described by Xie et al. [59] using the plant ethylene ELISA kit (ColorfulGene, Wuhan, China) and following the manufacturer's instructions. The content was expressed as ng/g of fresh weight (FW).

Antioxidant enzyme assay
The crude enzyme solution was prepared according to a previous study [60]. The activities of CAT, APX, SOD, and POD were measured and calculated spectrometrically [61][62][63][64]. An absorbance increase of 0.01 OD 470 nm min −1 was considered 1 U of POD activity, a decrease in absorbance of 0.01 at OD 240 nm min −1 was considered 1 U of CAT activity, the amount used to inhibit 50% of the photochemical reduction of NBT was considered 1 U of SOD activity, and a decrease in absorbance of 0.01 at OD 290 nm min −1 was considered 1 U of APX activity. The results are expressed on a FW basis as U·g −1 .

MDA content
As mentioned previously [65], 0.5 g of the plant sample was homogenized, incubated, and then centrifuged to collect the supernatant. The absorbance at 450, 532, and 600 nm was measured.

H 2 O 2 content
A 0.5-g sample of plant material was homogenized and centrifuged to collect the precipitate. The precipitate was then added to 1.5 mL of 2 M H 2 SO 4 . The absorbance of the mixture was measured at 412 nm, and the content of H 2 O 2 was calculated [66,67].

•−
The reaction buffer comprised 50 mM phosphate buffer (pH 7.8) containing 17 mM sulfanilic acid, 1 mM hydroxylamine hydrochloride, 7 mM 1-naphthylamine, and 50-μL sample extracts. The absorbance of the mixture was measured at 530 nm, and the production rate of O 2 •− was calculated using previously described formulas [60].

Detection of H 2 O 2 and O 2 •− in tomato leaves
The distribution of O 2 •− was detected as described previously [25]. Brief ly, WT and dcd2 leaves were vacuum infiltrated with 0.1 mg/mL NBT in 25 mM HEPES buffer (pH 7.6) for 1 min in darkness. Chlorophyll was removed using ethanol, and the leaves were photographed. WT and dcd2 mature leaves were stained with DAB according to a previously described method [25].

Statistical analysis
The data were calculated from three replicates in each experiment, and the experiments were repeated independently three times. Statistical significance was assayed by one-way analysis of variance using IBM SPSS Statistics (SPSS version 20.0; Armonk, NY, USA), and the results are expressed as the means ± SDs. Significant differences were calculated by a t test (P < 0.01 or P < 0.05). The different letters above the columns represent significant differences between two values (P < 0.05) at the same time-point.The correlations among the chlorophyll, chlorophyll a, chlorophyll b, and carotenoid contents, H 2 O 2 content, production of O 2 •− , MDA content and gene expression of NYC1, PAO, SGR1, PDS, PSY1, ACO1, ACS2, E4, CEL2, and EXP in the WT and dcd2-1 tomato fruits at 36, 39, and 43 DPA were analysed using the OmicShare platform (https://www.omicshare.com). The correlations among the H 2 S and ethylene content in the WT and dcd2-1 tomato fruits at 36, 39, and 43 DPA were analysed using the OECloud tools at https://cloud.oebiotech.cn.