Comparative Transcriptomic Analysis of the Development of Sepal Morphology in Tomato (Solanum Lycopersicum L.)

Sepal is an important component of the tomato flower and fruit that typically protects the flower in bud and functions as a support for petals and fruits. Moreover, sepal appearance influences the commercial property of tomato nowadays. However, the phenotype information and development mechanism of the natural variation of sepal morphology in the tomato is still largely unexplored. To study the developmental mechanism and to determine key genes related to downward sepal in the tomato, we compared the transcriptomes of sepals between downward sepal (dsp) mutation and the wild-type by RNA sequencing and found that the differentially expressed genes were dominantly related to cell expansion, auxin, gibberellins and cytokinin. dsp mutation affected cell size and auxin, and gibberellins and cytokinin contents in sepals. The results showed that cell enlargement or abnormal cell expansion in the adaxial part of sepals in dsp. As reported, auxin, gibberellins and cytokinin were important factors for cell expansion. Hence, dsp mutation regulated cell expansion to control sepal morphology, and auxin, gibberellins and cytokinin may mediate this process. One ARF gene and nine SAUR genes were dramatically upregulated in the sepal of the dsp mutant, whereas seven AUX/IAA genes were significantly downregulated in the sepal of dsp mutant. Further bioinformatic analyses implied that seven AUX/IAA genes might function as negative regulators, while one ARF gene and nine SAUR genes might serve as positive regulators of auxin signal transduction, thereby contributing to cell expansion in dsp sepal. Thus, our data suggest that 17 auxin-responsive genes are involved in downward sepal formation in the tomato. This study provides valuable information for dissecting the molecular mechanism of sepal morphology control in the tomato.


Introduction
The tomato (Solanum lycopersicum L.) is an important commercial crop and model for studying the floral organ development of angiosperms. After flowering is completed, tomato sepals are persistently protect young fruits and improve the quality of the appearance of mature fruits. However, as living standards increase, many people started to consider the quality and appearance of tomato sepals.
Healthy and flat sepals have become an important standard for measuring the quality of tomato fruits, enhancing visual esthetics and reflecting fruit freshness. Therefore, the molecular mechanism of the regulation of sepal morphology regulation should be investigated.
Sepals affect the flower development by coordinating cell division, cell differentiation and cell expansion with other parts of the flower whorl. The morphology and size of the sepals have been associated with the yield and quality of the fruit. The larger sepal size tightly associates with the protection of flower whorl and better fruit quality [1]. In SlMBP21-RNAi tomato, the sepals are longer and fruit sets are improved [2]. Among the green parts of the flower, sepal has the greatest ability to photosynthesis, follow by the receptacle [3]. The contents of Chl and the activity of ribulose-1, 5-bisphosphate carboxylase/oxygenase, the key photosynthetic enzyme, are both increased in longer sepals, and photosynthesis is enhanced in longer sepals, which may contribute by improving the fruit set [2]. Conclusively, sepal morphology is closely associated with fruit development.
Cell size affects sepal morphology. In Arabidopsis sepals, the loss of single-cell variability in an ftsh4 mutant leads to the destruction of the entire sepal shape. This result indicates that changes in individual cell shape and size are important factors influencing the final organ morphology [4]. Other studies have shown that cell growth rate also results in differences in sepal morphology. The growth rate between cells in sepals significantly differs, the relative growth rate of cells in different regions of Arabidopsis sepals is about 0% to 5% (cell growth size/h) [5]. Although the cell growth rate during sepal development is variable, all cells achieve the maximum relative growth rate almost at the same time during development [5]. However, cell growth is not synchronized, and the time needed to reach the maximum growth rate varies between cells [5].
The morphological characteristics and growth of plant organs are regulated by hormones, including auxin, cytokinin, gibberellin, ethylene, abscisic acid (ABA), jasmonic acid (JA), brassinosteroids, stearolactone, and many peptides [6,7]. Auxin and gibberellin affect petal expansion and flowering [8,9]. The petal growth of Arabidopsis thaliana is regulated by AUXIN RESPONSE FACTOR8 (ARF8), and the petals of arf8 mutants are significantly larger than those of wild-type (WT) petals because of the increased number of cells and their expansion [8]. The miRNA319a mutant in Arabidopsis has narrow and short petals, and this trait is regulated by TCP transcription factors, which regulate auxin function [10][11][12][13]. A large number of signals (e.g., hormones, mechanical signals and polar fields) are distributed in various regions of plant organs, thereby coordinating the developmental behavior of multiple cells; as a result, different growth regions form. For example, TCPs promote the cessation of cell division and slow down the growth rate from the sepal apex to the base, resulting in a mechanical conflict that leads to sepal apex formation. The functional redundancy of these signals reinforces them, thereby enhancing the stability of organ morphology.
Despite a large diversity in flower morphology among Solanum species, studies on the genetic basis of the variations in these organs are limited. In the tomato, MACROCALYX (MC), TAGL1, SlFYFL, SlMBP21 and, SlCMB1, are involved in the regulation of sepal development [14][15][16][17][18][19]. The T-DNA insertion of Arlequin (Alq) and the overexpression of the MADS-box gene TAGL1 result in the conversion of sepals into fleshy, fruitlike organs [15]. The nucleotide sequences of TAGL1 (syn. ALQ) show a high similarity to those of Arabidopsis D-class genes SHATTERPROOF1 (SHP1 and AGL1) [15]. The overexpression of another MADS-box gene, namely SlFYFL, gives rise to longer sepals [16]. The MADS-box gene SlMBP21 regulates the development of sepals and the floral abscission zone [2,17,18]. Moreover, the SEPALLATA (E-function) MADS-box gene SlCMB1 participates in the development of inflorescence architecture and also regulates sepal size in tomato plants [19].
In this study, high-throughput sequencing was conducted to detect differentially expressed genes (DEGs) during sepal development in the tomato. One wild-type (WT) accession with typical flat-spread sepal and dsp mutant with downward sepal were used to identify sepal shape-related candidate genes by comparing sepal transcriptomes at key stages, and to reveal the pathways and related genes possibly involved in sepal development. This work could offer valuable information as a basis for further studying the tomato flower shape. It would also provide insights into the molecular mechanism underlying flower organ development in the tomato for future breeding programs.

Phenotypic Development Features and Diversity of Tomato Sepals
One sepal morphology mutant was verified. Although the WT tomato plant was characterized by typical flat-spread sepals, the dsp mutant, which was generated through natural mutation, exhibited an unusual "downward sepal" phenotype ( Figure 1a). The morphological development of sepals from the flower bud stage to the fruit maturity stage in WT and dsp was subdivided into seven stages (Figure 1a). At stage 1, i.e., unopened stage, sepals combined with embracing the three inner whorls of floral organs. At stage 2, i.e., slightly opened stage, sepals began to separate, and an inflated corolla emerged at the top of the calyx. At stage 3, i.e., fully opened stage, sepals and petals were completely separated, and the calyx and corolla were fully expanded. At stage 4, i.e., reclosed stage, the calyx and corolla began to close again, and the corolla started to wilt. At stage 5, i.e., reopened stage, the calyx of WT began to open again and bore fruit, but the dsp calyx was not completely separated. At stage 6, i.e., morphological differentiation stage, the WT sepals were fully opened again, but the dsp sepals were still closed and wrapped around the fruit. At stage 7, i.e., final morphology, the morphological characteristics of sepals were fully developed, the sepals of WT were fully expanded and flat, whereas the dsp sepals were oriented downward. Notably, the differences in the sepal morphology between the WT and dsp mutant were apparent even at stage 4 of sepal development (Figure 1a,b). The upturned degree and rolling index of the sepals from >10 independent WT and dsp plants at different stages were measured. The upturned degree of the sepals of the dsp plants was significantly lower than that of the WT plants at stages 4 to 7. The rolling index of the sepals of the dsp plants was significantly greater than that of the WT plants at stages 5 to 7 (Figure 1b). The WT and dsp plants showed the same morphological differentiation of sepals from stages 1 to 3. Sepal morphology was significantly diverse during stages 4 to 7.

Identification of Differentially Expressed Genes in Sepals from WT and dsp Mutant Plants
Genome-wide expression analysis was conducted to compare the transcriptome profiles of the sepals between WT and dsp mutant plants through a differential gene expression (DGE) approach and to identify the genes involved in the downward sepals of tomato [20]. Phenotypic changes occurred at stage 4 ( Figure 1a), so the sepals at stage 3 (0 days after flowering [DAF]; length of~8 mm) and stage 4 (15 DAF; length of~11 mm) were chosen for RNA-seq analyses. After the RNA was sequenced, 49.11 million and 48.89 million reads were obtained from the WT and dsp plants at stage 3, respectively; of these reads, 91.76% and 90.74% could be mapped to the annotated tomato genome, respectively ( Figure 2a). Meanwhile, 55.56 million and 52.56 million reads were obtained from the WT and dsp plants at stage 4, respectively. Of these reads, 68.12% and 65.39% could be mapped to the annotated tomato genome, respectively (Figure 2a). The RNA data displayed good correlations between the two biological replicates and were used for further analysis ( Figure S1).
After the transcriptomes were compared with those of WT plants, 3066 and 2459 DEGs were detected in the dsp plants at stages 3 and 4, respectively, and a total of 4729 DEGs were identified at two stages (Figurer 2b,e). Of the DEGs at stage 3, 1355 were upregulated and 1711 were downregulated (Figure 2b,f). However, 1236 DEGs were upregulated and 1223 DEGs were downregulated in the dsp plants at stage 4 ( Figure 2b,g).
The top five genes in the upregulated DEGs at stage 3 were two unknown protein-coding genes (Solyc08g044230. 1,19065

Identification of Differentially Expressed Genes in Sepals from WT and dsp Mutant Plants
Genome-wide expression analysis was conducted to compare the transcriptome profiles of the sepals between WT and dsp mutant plants through a differential gene expression (DGE) approach and to identify the genes involved in the downward sepals of tomato [20]. Phenotypic changes occurred at stage 4 (Figure 1a), so the sepals at stage 3 (0 days after flowering [DAF]; length of ~8 mm) and stage 4 (15 DAF; length of ~11 mm) were chosen for RNA-seq analyses. After the RNA was

Annotation of DEGs in Sepals from WT and dsp Mutant Plants
A total of 4729 DEGs were classified into three categories based on gene ontology (GO) assignments: biological processes (BP), molecular functions (MF), and cellular components (CC). Further GO enrichment analysis resulted in the identification of 205 GO terms (114 BP, 23 CC, and 68 MF; false discovery rate [FDR] <0.05; Table S2). For MF, the top five enriched GO terms were "transferase activity" (738 genes), "oxidoreductase activity" (428 genes), "kinase activity" (307 genes) "phosphotransferase activity" (294 genes), and "transporter activity" (294 genes; Figure 3a; Table S2). For BP, the DEGs were primarily enriched in response to stimuli, lipid metabolism, and organic acid metabolism ( Figure 3a; Table S2). Changes in CC occurred primarily in the membrane and cell wall   Table S2). For MF, the top five enriched GO terms were "transferase activity" (738 genes), "oxidoreductase activity" (428 genes), "kinase activity" (307 genes) "phosphotransferase activity" (294 genes), and "transporter activity" (294 genes; Figure 3a; Table S2). For BP, the DEGs were primarily enriched in response to stimuli, lipid metabolism, and organic acid metabolism ( Figure 3a; Table S2). Changes in CC occurred primarily in the membrane and cell wall ( Figure 3a; Table S2). MF analysis revealed that pectinesterase activity, pectinesterase inhibitor activity, and pectate lyase activity significantly changed, and pectate lyase activity significantly changed according to MF analysis, and they were related to cell wall modification. (Figure 3a; Table S2). The DEGs that function in cell wall organization and biogenesis were enriched. These results suggested that cell wall biogenesis and components in downward sepals were influenced.  Table S2). MF analysis revealed that pectinesterase activity, pectinesterase inhibitor activity, and pectate lyase activity significantly changed, and pectate lyase activity significantly changed according to MF analysis, and they were related to cell wall modification. (Figure 3a; Table  S2). The DEGs that function in cell wall organization and biogenesis were enriched. These results suggested that cell wall biogenesis and components in downward sepals were influenced.  Furthermore, 4729 DEGs were mapped into 177 Kyoto Encyclopedia of Genes and Genomes (KEGG) signaling pathways. Of these DEGs, 20 were significantly enriched (p < 0.05; Figure 3b; Table S3). The top five enriched KEGG pathways were "brassinosteroid biosynthesis (rich factor = 0.41)," "butanoate metabolism (rich factor = 0.39)", "taurine and hypotaurine metabolism (rich factor = 0.38)", "diterpenoid biosynthesis (rich factor = 0.33)", and "valine, leucine, and isoleucine biosynthesis (rich factor = 0.33)," as shown in Figure 3b and Table S3. KEGG analysis showed that most of the DEGs (71 genes) were enriched in "plant hormone signal transduction". A number of DEGs were enriched in the following pathways: 15 genes in "diterpenoid biosynthesis," 7 genes in "brassinosteroid biosynthesis," and 16 genes in "zeatin biosynthesis". In addition, many DEGs that were involved in the pathways of cell wall metabolism and "cutin, suberine, and wax biosynthesis" (12 genes) were altered in the dsp plants. These results indicated that dsp affected sepal morphology by controlling the genes related to cell wall modification and plant hormone regulation.

Coexpression Pattern of DEGs in Sepals from WT and dsp Mutant Plants
Hierarchical clustering was applied to 4729 DEGs between WT and dsp sepals at stages 3 and 4 to further examine the transcriptomic response to downward sepal. Five large clusters were obtained, and the DEGs were clustered into five clusters by using the k-means (k = 5) clustering algorithm. These clusters were then visualized with a heat map and a cluster trend chart (Figure 4a,b). The results revealed the general patterns of transcriptomic profiles during sepal development.

Regulation of Cell Expansion by dsp to Control Sepal Morphology
According to the results of DEG analysis, the expression levels of genes related to cell expansion significantly changed in dsp sepals (Table 1; Table S4). Among the DEGs, 16 were XTH genes, 1 was xyloglucan galactosyltransferase gene, 24 were pectinesterase genes, 1 was pectin methyltransferase gene, 10 were pectin lyase protein genes, 5 were pectin acetylesterase genes, 4 were pectinesterase inhibitor genes, 10 were pectin methylesterase inhibitor genes, 14 were cellulose synthase genes, 1was The expression levels of the genes in clusters A and C of the dsp plants decreased at stage 3, whereas the expression levels of the genes in cluster D of the dsp plants continuously decreased from stage 3 to stage 4 ( Figure 4b). The expression levels of the genes in cluster C did not significantly differ in the same plant sepals at different stages ( Figure 4b). By contrast, the expression levels of the genes from clusters B and E of the dsp plants decreased at stage 3. The five clusters were subjected to GO annotation enrichment analysis to gain further insights into the transcriptomic changes in the sepals of the dsp plants. In clusters A and C, the genes involved in cell wall organization and biogenesis were enriched (Figure 4c), suggesting that changes in the cell wall were one of the reasons for the altered sepal morphology in the dsp plants. Conversely, the gene functioning in hormone metabolism and biogenesis were enriched in clusters B and E. This result indicated that the differences in the sepal morphology of the dsp plants were attributed to alterations in hormonal levels, particularly auxin, gibberellin and cytokinin.

Regulation of Cell Expansion by dsp to Control Sepal Morphology
According to the results of DEG analysis, the expression levels of genes related to cell expansion significantly changed in dsp sepals (Table 1; Table S4). Among the DEGs, 16 were XTH genes, 1 was xyloglucan galactosyltransferase gene, 24 were pectinesterase genes, 1 was pectin methyltransferase gene, 10 were pectin lyase protein genes, 5 were pectin acetylesterase genes, 4 were pectinesterase inhibitor genes, 10 were pectin methylesterase inhibitor genes, 14 were cellulose synthase genes, 1 was a glycine-rich cell wall structural protein 1 precursor gene, and 4 were cell wall protein genes; these genes were upregulated or downregulated from 287.36-fold to 2.00-fold (Table 1; Table S4). They were also involved in cell wall metabolism, which is important for cell expansion and cell growth rates. In addition, the expression of 14 expansion (-like) protein genes and 12 extension (-like) protein genes remarkably differed (from 196.85-fold to 2.16-fold; Table 1 and Table S4). These genes were required for cell expansion, cell size, and organ shape development. The cell expansion and cell growth rate determine the size and shape of organs. The DEGs related to cell expansion remarkably varied in dsp sepals (Table 1 and Table S4). Hence, the cells in the middle part of the sepal in dsp and WT plants were observed through histological analysis. In the sections from stages 4 to 7, the cell size in the adaxial part of the sepals of dsp plants was larger than that in the sepals of the WT plants ( Figure 5A-K). The cell number per unit view clearly decreased, whereas the cell area significantly increased (p < 0.05) in the sepals of dsp plants at stages 4 and 5 ( Figure 5B,C,F,G,K). The cell area in the sepals of the dsp plants at stages 6 and 7 significantly increased (p < 0.01; Figure 5D,H-K). Thus, cell enlargement or abnormal cell expansion in the adaxial part of sepals might contribute to downward sepals in dsp plants.
The cell expansion and cell growth rate determine the size and shape of organs. The DEGs related to cell expansion remarkably varied in dsp sepals (Table 1; Table S4). Hence, the cells in the middle part of the sepal in dsp and WT plants were observed through histological analysis. In the sections from stages 4 to 7, the cell size in the adaxial part of the sepals of dsp plants was larger than that in the sepals of the WT plants ( Figure 5A-K). The cell number per unit view clearly decreased, whereas the cell area significantly increased (p < 0.05) in the sepals of dsp plants at stages 4 and 5 ( Figure 5B,C,F,G,K). The cell area in the sepals of the dsp plants at stages 6 and 7 significantly increased (p < 0.01; Figure 5D,H-K). Thus, cell enlargement or abnormal cell expansion in the adaxial part of sepals might contribute to downward sepals in dsp plants.
Consequently, the alteration of the expression of these cell expansion-related genes induced cell enlargement, which might contribute to downward sepals in dsp plants. Therefore, dsp might regulate cell expansion to control sepal morphology.

Effects of dsp on Auxin, Cytokinin, and Gibberellin Levels in Sepals
Hormones serve as crucial regulators of organ morphology development. The duration and rate of cell proliferation are positively controlled by auxin and cytokinin. The transition from cell division to expansion is correlated with gibberellin. Phytohormones play a critical role in integrating developmental signals to control organ morphology. The DEGs related to auxin, cytokinin, and gibberellin remarkably differed in dsp sepals ( Figure 6; Table 1). The majority of these genes were related to auxin, and some of them were associated with cytokinin and gibberellin ( Figure 6). Int. J. Mol. Sci. 2020, 21, x FOR PEER REVIEW 12 of 25 Consequently, the alteration of the expression of these cell expansion-related genes induced cell enlargement, which might contribute to downward sepals in dsp plants. Therefore, dsp might regulate cell expansion to control sepal morphology. Hormones serve as crucial regulators of organ morphology development. The duration and rate of cell proliferation are positively controlled by auxin and cytokinin. The transition from cell division to expansion is correlated with gibberellin. Phytohormones play a critical role in integrating developmental signals to control organ morphology. The DEGs related to auxin, cytokinin, and gibberellin remarkably differed in dsp sepals ( Figure 6; Table 1). The majority of these genes were related to auxin, and some of them were associated with cytokinin and gibberellin ( Figure 6). Among DEGs, 53 were linked to auxin, i.e., 13 of them were upregulated and 23 were downregulated at stage 3. At stage 4, 9 were upregulated and 16 were downregulated ( Figure 6; Table  1; Table S4). The expression levels of the following genes were altered from 404.79-fold to 2.04-fold: 31 auxin-responsive protein-coding genes (GH3 genes, SAUR genes, ARF genes, IAA3, etc.), four auxin efflux carrier protein-coding genes, nine auxin-induced protein-coding genes, three auxin transporter-like protein-coding genes, two auxin-binding protein-coding genes (ABP19a and ABP19a-like), one auxin-repressed protein-coding genes, one auxin-regulated protein-coding gene, and one small auxin-up protein 58 gene ( Figure 6; Table S4). In addition, the content of IAA increased 1.69-fold and 1.44-fold in the dsp sepals at stages 3 and 4, respectively, compared with that in the WT sepals ( Figure 7a). (Figure 7a). Consequently, dsp influenced the efflux, signaling, and content of auxin in sepals. Among DEGs, 53 were linked to auxin, i.e., 13 of them were upregulated and 23 were downregulated at stage 3. At stage 4, 9 were upregulated and 16 were downregulated ( Figure 6; Table 1; Table S4). The expression levels of the following genes were altered from 404.79-fold to 2.04-fold: 31 auxin-responsive protein-coding genes (GH3 genes, SAUR genes, ARF genes, IAA3, etc.), four auxin efflux carrier protein-coding genes, nine auxin-induced protein-coding genes, three auxin transporter-like protein-coding genes, two auxin-binding protein-coding genes (ABP19a and ABP19a-like), one auxin-repressed protein-coding genes, one auxin-regulated protein-coding gene, and one small auxin-up protein 58 gene ( Figure 6; Table S4). In addition, the content of IAA increased 1.69-fold and 1.44-fold in the dsp sepals at stages 3 and 4, respectively, compared with that in the WT sepals ( Figure 7a). (Figure 7a). Consequently, dsp influenced the efflux, signaling, and content of auxin in sepals.
Among DEGs, 17 were cytokinin synthesis-related factors, and two were cytokinin degradation-related factors. Of these genes, nine were upregulated and nine were downregulated at stage 3; at stage 4, two were upregulated and one was downregulated ( Figure 6; Table 1 and  Table S4). Two cytokinin riboside 5"-monophosphate phosphoribohydrolase (SlLOG) genes were upregulated from 4.01-fold to 2.94-fold at stage 3, one isopentenyltransferase 2 (SlIPT) gene was downregulated up to 5.39-fold, and one LOG gene was upregulated up to 3.06-fold at stage 4. Two cytokinin oxidase/dehydrogenase (SlCKX) genes were upregulated from 6.54-fold to 3.33-fold at stages 3 and 4 (Table 1, Figure 6). The first step of CK biosynthesis in plants is the N-prenylation of adenosine 5 -phosphates via dimethylallyl diphosphate, resulting in the biosynthesis of CK nucleotides. [21,22]. This step is catalyzed by IPT. The next step is the phosphoribohydroxylation of CK nucleotides to synthesize biologically active CK nucleobases [23]. This step is catalyzed by CK riboside 5 -monophosphate phosphoribohydrolase (LOG). Cytokinin conjugation occurs mainly through cytokinin oxidase (CKX) enzymes. Furthermore, two zeatin O-glucosyltransferase genes were upregulated from 4.75-fold to 4.14-fold, whereas four zeatin O-glucosyltransferase genes were downregulated from 5.01-fold to 2.03-fold) ( Table 1 and Table S4). Three zeatin O-xylosyltransferase genes were upregulated from 11.02-fold to 3.17-fold, whereas four zeatin O-xylosyltransferase genes were downregulated from 5.96-fold to 2.64-fold) ( Table 1 and Table S4). Cytokinins with a hydroxyl group on the side chain can undergo O-glycosylations. Zeatin O-glucosyltransferase and zeatin O-xylosyltransferase, which are zisZOG genes, play important roles in regulating the levels of cis-zeatin and maintaining appropriate levels of active cytokinins because biosynthetic enzymes work slowly [24]. Zeatin O-glycosyl derivatives are resistant to the cytokinin degrading enzyme CKX and are considered to be storage forms because they can be cleaved by b-glucosidase [22,25]. In our study, the contents of cytokinin in the dsp sepals increased 1.16-, 2.35-and 1.34-fold at stages 3, 4, and 5, respectively, compared with those in the WT sepals (Figure 7). In our data, the upregulation of SlCKX might be responsible for the increase in the cytokinin levels in the sepals of dsp plants. The upregulation of SlLOG might accelerate cytokinin synthesis and result in an increase in cytokinin levels. These results indicated that the increased expression of cytokinin biosynthesis-related genes (SlLOGs), the increased expression of SlCKXs, and the increased cytokinin content contributed to the expansion of sepal cells in dsp. Among DEGs, 17 were cytokinin synthesis-related factors, and two were cytokinin degradationrelated factors. Of these genes, nine were upregulated and nine were downregulated at stage 3; at stage 4, two were upregulated and one was downregulated ( Figure 6; Table 1 and Table S4). Two cytokinin riboside 5"-monophosphate phosphoribohydrolase (SlLOG) genes were upregulated from 4.01-fold to 2.94-fold at stage 3, one isopentenyltransferase 2 (SlIPT) gene was downregulated up to 5.39-fold, and one LOG gene was upregulated up to 3.06-fold at stage 4. Two cytokinin oxidase/dehydrogenase (SlCKX) genes were upregulated from 6.54-fold to 3.33-fold at stages 3 and 4 (Table 1, Figure 6). The first step of CK biosynthesis in plants is the N-prenylation of adenosine 5ʹphosphates via dimethylallyl diphosphate, resulting in the biosynthesis of CK nucleotides. [21,22]. This step is catalyzed by IPT. The next step is the phosphoribohydroxylation of CK nucleotides to synthesize biologically active CK nucleobases [23]. This step is catalyzed by CK riboside 5′monophosphate phosphoribohydrolase (LOG). Cytokinin conjugation occurs mainly through GAs are essential for cell elongation and other plant growth and developmental processes [26][27][28]. GA20oxs (GA20-oxidases) are key GA biosynthesis regulators that determine the GA content of plant species [29]. Among the DEGs in our study, seven were gibberellin synthesis-related genes and 11 were gibberellin response-related genes. Of these genes, eight were upregulated and seven were downregulated at stage 3 while six were upregulated and three were downregulated at stage 4 ( Figure 6; Table 1 and Table S4). The two GA20ox genes in dsp were downregulated from 10.87-fold to 4.31-fold at stage 3. One of them was downregulated to 2.92-fold at stage 4. Two GA2ox genes, which are responsible for the deactivation of GAs, were significantly upregulated from 4.98-fold to 3.81-fold at stage 3 of the dsp plant (Table 1 and Table S4). Furthermore, the contents of gibberellin in the dsp sepals of the plant decreased by 1.35-and 1.85-fold at stages 4 and 5, respectively, compared with those in the WT sepals (Figure 7). These results indicated that the decreased expression of GA biosynthesis-related genes (GA20oxs) and the increased expression of GA2oxs contributed to a decrease in the GA content.

Possible Function of 17 Auxin-Responsive Genes as Regulators in Downward Sepal Formation in the Tomato
The detected DEGs were mapped to reference canonical pathways in the KEGG to further identify the key gene regulatory pathways responsible for the formation of downward sepals in the tomato [30]. Among the enriched KEGG pathways, the "auxin signal transduction" pathway, which was putatively associated with the downward sepal phenotype in the dsp mutant, was also enriched.

Discussion
The histological analysis showed that adaxial cells in the dsp sepals were larger than those in the WT sepals ( Figure 6). Cell expansion is an important developmental force of organ morphology development at the cellular level. The transcriptome data indicated that cell expansion-related genes, such as XTH, pectinesterase genes, pectin lyase, cellulose synthases, expansins, and extensins, significantly changed (Table 1 and Table S4). Plant cell walls are complex structures composed of cellulose, xyloglucan, pectic polysaccharides, and structural proteins. To adapt to expansive forces and plant growth, plant cells selectively loosen their cell wall [40]. In the dsp sepals, seven XTH genes were upregulated at stage 4, and these genes were responsible for loosening the cell wall and promoting cell expansion and floral organ growth during flower opening [41,42]. Xyloglucan endotransglycosylase/hydrolases (XTHs) are encoded by a large multigene family, which cut xyloglucan and join the new reducing end to the non-reducing end of another xyloglucan (a transglucosylation) or to water (a hydrolysis) [43]. XTH itself cannot induce wall relaxation or creep, but can synergistically enhance wall extension, which might be termed an indirect or secondary loosening agent [43]. Moreover, 34 genes related to cell wall degradation were expressed differently in dsp sepals. For example, pectate lyase and pectinesterase break down pectins and participate in cell wall metabolism [44][45][46]. Ten pectin methylesterase (PME) inhibitor genes were expressed differently in dsp sepals, and they are implicated in the demethylation of pectin, resulting in the relaxation of the cell wall so that cells can enlarge [47]. Twelve DEGs related to cellulose synthesis were downregulated at stages 3 and 4 and involved in depositing cellulose to either primary or secondary walls in Arabidopsis [48]. In addition, 26 expansin/expansin-like and extensin/extensin-like genes were expressed differently in dsp sepals. They were considered two of the most important regulators of cell wall expansion and loosening during plant cell growth [49][50][51]. The α-expansins induced wall creep and wall relaxation, which mediated acid-induced extension of plant cell walls without mechanically weakening the cell walls [52][53][54]. In our study, 16 XTHs, 14 expansin genes and 11 PMEs/PME inhibitor genes were expressed differently in dsp sepals (Table S4). These results suggested that dsp positively regulated cell expansion through cell wall loosening, thus inducing the alteration of sepal morphology. Organ development is regulated by different phytohormones, which manipulate appropriate cell division and cell expansion. As an important hormone in plant development, auxin plays important role in cell expansion [55][56][57][58]. In our work, the IAA content increased in dsp sepals, compared with those in WT sepals ( Figure 7A). Moreover, the expression of auxin-related genes involved in response and efflux, especially AUX/IAA and SAUR family genes, significantly varied ( Figure 8; Table S4). SMALL AUXIN UP-RNA (SAUR) genes constitute the largest family of auxin-induced genes, and 79 members are found in Arabidopsis [34]. The overexpression of stabilized SAUR19 proteins confers numerous phenotypes indicative of increased cell expansion, including increases in hypocotyl length and leaf size, and altered tropic growth [59][60][61] Similar findings were obtained with plants expressing SAUR63 proteins [62], suggesting that SAURs may positive effectors of cell expansion [59,62,63]. SAUR gene expression is upregulated by treatments/conditions that promote growth (e.g., IAA, brassinosteroids) and downregulated by factors that repress growth (e.g., abscisic acid, jasmonic acid, abiotic stress) [64,65]. Auxin is known to stimulate the activity of plasma membrane H + -ATPase proton pumps, which promote proton efflux to acidify the apoplast and facilitate the uptake of solutes and water to drive plant cell expansion [39,[66][67][68][69][70][71]. This process is regulated by auxin-inducible SAUR proteins [39,70]. The ensuing decrease in apoplastic pH alters the activity of cell wall-modifying proteins, including expansins [72], xyloglucan endotransglycosylase/hydrolases (XTHs) [73], and pectin methylesterases (PMEs) [74], leading to changes in wall extensibility. Auxin promotes the reorientation of microtubules from random to transverse, and suppresses the peroxidase activity in the cell wall. Consequently, cell wall extensibility is also promoted [75]. Therefore, dsp likely affected cell expansion by regulating auxin response, which induced the alteration of sepal morphology.
Transcriptomic analysis demonstrated that the genes related to cytokinin were expressed differentially; correspondingly, the cytokinin production level clearly increased in the dsp sepals compared with that in the WT sepals ( Figure 7B). Furthermore, 19 DEGs were related to cytokinin: 17 biosynthesis-related genes (11.02-fold, the maximum fold change) and two degradation-related genes (6.54-fold, the maximum fold change; Figure 6C; Table S4). For example, two cytokinin oxidase/dehydrogenase (CKX) genes were upregulated 2.71-fold to 1.74-fold (Table S4). The overexpression of CKX1 or CKX2 in Arabidopsis and other species causes the elongation of the primary root and increases root branching [76][77][78], whereas the overexpression of AtCKX7 results in an opposite phenotype [79]. Therefore, each CKX determined a specific developmental and physiological function. Cytokinin stimulated cell expansion in plants [80,81] and induced fourα-expansin subfamily members, namely, EXPA1, EXPA10, EXPA14, and EXPA15, in Arabidopsis root [82]. As a result, the cell wall loosened, and the cells expanded. In angiosperms, the ancient tRNA-IPTs and CKXs preferred the cZtype cytokinins as substrates, play a housekeeping role to maintain basic cellular functions. On the other hand, the nonancient ATP/ADP-IPTs and CKXs preferred the iP-and tZ-type cytokinins as bstrates, contribute more to the regulation of organ development and abiotic stress responses [83]. Five zeatin O-glucosyltransferase and zeatin O-xylosyltransferase (cisZOG) genes were upregulated 3.46 fold to 1.67-fold. cisZOG is involved in zeatin biosynthesis [24]. These results implied that cytokinin biosynthesis and signaling in dsp sepals were affected, and these processes might promote cell expansion and lead to downward sepals.
Gibberellin promotes cell expansion [84]. In Petunia and Mimulus, stamen removal leads to a reduced number of petals likely because stamens produce gibberellins [85,86]. In addition, GA-deficient Arabidopsis mutants exhibit a reduction in petal elongation, but gibberellins promote cell elongation by inhibiting the function of DELLA protein. The DELLA protein RGA, RGL1 and RGL2 in Arabidopsis inhibit the growth of petals in gibberellin-deficient plants [87]. Our results revealed that the content of gibberellins decreased, whereas the cell size increased in the dsp sepals, suggesting that gibberellins antagonized CK in a wide range of developmental events, including cell differentiation, shoot and root elongation, and meristem maintenance [88,89]. Active CK and recessive GA signals not only induce the expression of GA2ox, a GA-deactivating enzyme but also promote the expression of the cytokinin-biosynthesis gene ISOPENTENYL TRANSFERASE7. Consequently, cytokinin signaling is elevated [89,90]. Therefore, dsp promoted downward sepals through the regulation of adaxial cell size in sepals, and this process might be mediated by auxin, cytokinin and gibberellin.

Plant Materials
The dsp mutant of tomato was generated in the background of the inbred line TI1101 through natural mutation and stabilized via six generations of selfing prior to this study. The seeds of dsp mutant and WT were germinated on wet filter paper in a Petri dish at 28 • C in dark for 3 days. Then the resulting seedlings were grown in a growth chamber under a 16h/8h (light/dark) photoperiod with 25 • C/16 • C temperatures, respectively. Upon four true-leaf stage, plants were transferred to a greenhouse in the experimental field of the Northwest A&F University. Pest control and water management were carried out according to standard practices. All the materials were grown in a plastic greenhouse in the Northwest A&F University (Shaanxi, China).

Measurement of Sepal Morphology at Different Developmental Stages
Nine ripened fruits with whole calyx per accession were used as repeats in the phenotypic analysis. Each sepal was evaluated for 2 traits: (1) Sepal Upturned Degree (SEUD, • ) = 90 • + α = 90 • + arctan (H/L 1 ), degree of upwarp or down-wrap of sepal. Put a whole calyx on a L-square ruler, then keep the calyx base level on one side of L-square ruler and the sepal to be measured leaning against another side of L-square ruler. H represents the vertical distance between the furthest point from sepal to stalk in the horizontal line and the base line, L 1 is the longest distance from sepal to stalk ( Figure 9). (2) Sepal Rolling Index (SERI, %) = (SEL − L 2 )/SEL × 100%, SEL (Sepal Length, mm) is the length of sepal in flat condition, L 2 is the distance from apex to base point ( Figure 9). wrap of sepal. Put a whole calyx on a L-square ruler, then keep the calyx base level on one side of Lsquare ruler and the sepal to be measured leaning against another side of L-square ruler. H represents the vertical distance between the furthest point from sepal to stalk in the horizontal line and the base line, L1 is the longest distance from sepal to stalk (Figure 9).

DGE (Differentially Gene Expression) Library Construction and Sequencing
Sepal at stage 3 (0 DAF, length around 8mm) and stage 4 (15 DAF, length around 11 mm) were collected from WT and dsp mutant at the same time on the same day. Samples were immediately frozen in liquid nitrogen and stored at −80°C for RNA-Seq analyses. Total RNA was isolated using the TRIzol ® Plant RNA Purification Reagent according to the manufacturer's instructions (Invitrogen, Carlsbad, CA, USA) and genomic DNA was removed using DNase I (TaKara, Dalian, China). RNA quality was determined by 2100 Bioanalyser (Agilent, Silicon Valley, CA, USA) and quantified using

DGE (Differentially Gene Expression) Library Construction and Sequencing
Sepal at stage 3 (0 DAF, length around 8mm) and stage 4 (15 DAF, length around 11 mm) were collected from WT and dsp mutant at the same time on the same day. Samples were immediately frozen in liquid nitrogen and stored at −80 • C for RNA-Seq analyses. Total RNA was isolated using the TRIzol ® Plant RNA Purification Reagent according to the manufacturer's instructions (Invitrogen, Carlsbad, CA, USA) and genomic DNA was removed using DNase I (TaKara, Dalian, China). RNA quality was determined by 2100 Bioanalyser (Agilent, Silicon Valley, CA, USA) and quantified using the ND-2000 (NanoDrop Technologies, Wilmington, DE, USA). Only high-quality RNA sample (OD260/280 = 1.8-2.2, OD260/230 ≥ 2.0, RIN ≥ 6.5, 28S:18S ≥ 1.0, >10 µg) was used to construct sequencing library.
RNA-seq transcriptome library was prepared following TruSeq TM RNA sample preparation Kit from Illumina (San Diego, CA, USA) using 5 µg of total RNA. Shortly, messenger RNA was isolated according to polyA selection method by oligo(dT) beads and then fragmented by fragmentation buffer firstly. Secondly double-stranded cDNA was synthesized using a SuperScript double-stranded cDNA synthesis kit (Invitrogen, Carlsbad, CA, USA) with random hexamer primers (Illumina, San Diego, CA, USA). Then the synthesized cDNA was subjected to end-repair, phosphorylation and "A" base addition according to Illumina's library construction protocol. Libraries were size selected for cDNA target fragments of 200-300 bp on 2% Low Range Ultra Agarose followed by PCR amplified using Phusion DNA polymerase (NEB) for 15 PCR cycles. After quantified by TBS380, the paired-end RNA-seq sequencing library was sequenced with the Illumina HiSeq xten (2 × 150 bp read length).
The raw paired end reads were trimmed and quality controlled by SeqPrep (https://github.com/ jstjohn/SeqPrep) and Sickle (https://github.com/najoshi/sickle) with default parameters. Then clean reads were separately aligned to reference genome with orientation mode using TopHat (http://tophat. cbcb.umd.edu/,version2.0.0) [91] software. The mapping criteria of bowtie was as follows: sequencing reads should be uniquely matched to the genome allowing up to 2 mismatches, without insertions or deletions. Then, the region of the gene was expanded following depths of sites and the operon was obtained. In addition, the whole genome was split into multiple 15k bp windows that share 5k bp. New transcribed regions were defined as more than 2 consecutive windows without the overlapped region of the gene, where at least 2 reads mapped per window in the same orientation.

Bioinformatics Analysis of DGE Data
To identify DEGs (differential expression genes) between two different samples, the expression level of each transcript was calculated according to the fragments per kilobase of exon per million mapped reads (FRKM) method. RSEM (http://deweylab.biostat.wisc.edu/rsem/) [92] was used to quantify gene abundances. R statistical package software EdgeR (Empirical analysis of Digital Gene Expression in R, http://www.bioconductor.org/packages/2.12/bioc/html/edgeR.html) [93] was utilized for differential expression analysis. In addition, functional-enrichment analysis including GO and KEGG were performed to identify which DEGs were significantly enriched in GO terms and metabolic pathways at Bonferroni-corrected p-value ≤ 0.05 compared with the whole-transcriptome background. GO functional enrichment and KEGG pathway analysis were carried out by Goatools (https://github.com/tanghaibao/Goatools) and KOBAS (http://kobas.cbi.pku.edu.cn/home.do) [94].

Paraffin Sectioning and Electron Microscopy Experiment
The sepal cells were observed by paraffin sections stained with safranin-fast green. The sepals of WT and dsp at stage 3 to stage 7 were harvested and immediately fixed in formalin/glacial acetic acid (FAA). The samples were then dehydrated in a xylene and alcohol series (75,85,90,95 and 100%) and then embedded in paraffin wax. Four-micrometer-thick sections were cut and stained with safranin (1%)-fast green (0.5%) for histological examination. Sepal cells underwent electron microscopy using a general laboratory biology electron microscope

Extraction, Purification and Quantification of the Phytohormones in Sepals
Stage 3, stage 4, stage 5 and, stage 6 sepals for phytohormones IAA, zeatin and GA quantification. The method for extraction, purification, and quantification of phytohormones was modified from the description of Wang [95]. ELISA kits used for estimation of the hormonal levels came from China Agricultural University (Beijing, China).
SPSS software was used for statistical analysis. The dsp mutants were compared with the WT using Tukey's test at p < 0.05.

Conclusions
Conclusively, in the first part of the study, we observed the seven stages of the sepal morphological development of dsp (downward sepal) mutant and WT. The WT and dsp plants showed to be significantly diverse during stages 4 to 7. To understand the molecular mechanisms of downward sepal development, a transcriptome analysis of dsp mutant and WT plants at stages 3 and 4 was performed. We observed that dsp affected sepal morphology by controlling genes related to cell wall biogenesis and modification and plant hormone regulation, particularly auxin, gibberellin, and cytokinin. Among these genes, seven AUX/IAA genes functioned as negative regulators, while one ARF gene and nine SAUR genes served as positive regulators of auxin signal transduction. According to cell morphology observation between dsp and WT, we found that cell enlargement or abnormal cell expansion in the adaxial part of sepals might contribute to downward sepals in dsp. Meanwhile, dsp mutant led to increase in auxin and cytokinin, and a decrease in gibberellin. In conclusion, dsp promoted downward sepals through the regulation of adaxial cell size in sepals, and this process might be mediated by auxin, cytokinin and gibberellin.