Transcriptome Analysis Reveals Candidate Genes Involved in Gibberellin-Induced Fruit Development in Rosa roxburghii

Gibberellins (GAs) play indispensable roles in the fruit development of horticultural plants. Unfortunately, the molecular basis behind GAs regulating fruit development in R. roxburghii remains obscure. Here, GA3 spraying to R. roxburghii ‘Guinong 5’ at full-bloom promoted fruit size and weight, prickle development, seed abortion, ascorbic acid accumulation, and reduction in total soluble sugar. RNA-Seq analysis was conducted to generate 45.75 Gb clean reads from GA3- and non-treated fruits at 120 days after pollination. We obtained 4275 unigenes belonging to differently expressed genes (DEGs). Gene ontology and the Kyoto Encyclopedia of Genes and Genomes displayed that carbon metabolism and oxidative phosphorylation were highly enriched. The increased critical genes of DEGs related to pentose phosphate, glycolysis/gluconeogenesis, and citrate cycle pathways might be essential for soluble sugar degradation. Analysis of DEGs implicated in ascorbate revealed the myoinositol pathway required to accumulate ascorbic acid. Finally, DEGs involved in endogenous phytohormones and transcription factors, including R2R3 MYB, bHLH, and WRKY, were determined. These findings indicated that GA3-trigged morphological alterations might be related to the primary metabolites, hormone signaling, and transcription factors, providing potential candidate genes that could be guided to enhance the fruit development of R. roxburghii in practical approaches.


Introduction
Fruit development of horticultural plants is a refined reproductive and consecutive process divided into three stages: fruit set, the cell division or expansion stage, and fruit maturation [1]. The transition from flowering to fruit set upon fertilization in angiosperm refers to the initiation of fruit development. It begins with a period of cell division, and its high cell division activities often rapidly take place within a few weeks inside the ovary after anthesis and mitotic cells can still be found [2]. The second phase lasts a long time, mainly determining the final fruit size and weight. The last period is fruit thorough maturation and the fruit size grows very slightly at this stage. However, multiple storage products and secondary metabolites accumulate, thus forming different intrinsic and extrinsic qualities of fruits that are accompanied and could be manipulated by several phytohormones [3].
Gibberellins (GAs) are phytohormones and play critical roles in all the physiological periods of horticultural fruit development [4]. Abundant GAs are perceived as signals by the ovary or other fruit tissues after fertilization, resulting in fruit cell expansion [2]. For example, enhancing expression levels of genes in the gibberellin synthesis positively regulates the fruit set of tomato plants (Solanum lycopersicum) [4,5]. Similarly, endogenous GA accumulation is consistent with high cell division levels and tomato fruit ex- Effects of GA3 on fruit weight, fruit shape index, seed number, prickl length, concentration of total soluble sugar, and L-ascorbic acid contents. Values were Mean ± S. D from at least twenty fruits. The asterisk indicates a significant difference. "*": p < 0.05; "**": p < 0.01

Transcriptome Sequencing and De Novo Assemble of R. roxburghii
To better explore the molecular mechanism underlying the physiological changes b GA3, we used RNA-seq to profile the transcriptomes of GA3-treated and non-treated fruit at 120 DAP. High-throughput sequencing generated a total of 45.75 Gb clean reads for si samples, and the percentage of Q30 bases ranged from 94.59% to 95.11% (Table S1). Th clean reads were de novo assembled into transcripts, and 32,895 unigenes were obtaine with an average length of 1294 bp (N50 = 1933). A length of 23,395 (71.11%) unigenes wa over 500 bp, while those of 15,437 (46.92%) unigenes were over 1 Kb, suggesting that th assembly quality of the R. roxburghii transcriptome was satisfactory (Figure 2A). A total o 26,294 unigenes were annotated according to at least one database. Among them, 93.85% of (24,677) unigenes were annotated based on the National Center for Biotechnology In formation (NCBI) non-redundant (Nr) database. In terms of the Nr annotations, 40.24% of the sequences had strong homology (E-value < 1 × 10 −150 ), 11.96% and 20.71% of th annotated sequences showed homology (1 × 10 −150 < E-value < 1 × 10 −100 ), and (1 × 10 −100 < E value < 1 × 10 −50 ) to available plant sequences, respectively ( Figure 2B). The similarity dis tribution was comparable, with 41.56% of the sequences having similarities higher tha 90%, while 48.66% of the sequences of R. roxburghii had similarities of 50-90% (Figure 2C  3 on fruit weight, fruit shape index, seed number, prickle length, concentration of total soluble sugar, and L-ascorbic acid contents. Values were Mean ± S. D. from at least twenty fruits. The asterisk indicates a significant difference. "*": p < 0.05; "**": p < 0.01.

Transcriptome-Scale Analysis of GA3-Responsive DEGs in R. roxburghii Fruits
Differentially expressed genes (DEGs) were analyzed using the Fragments Per Kilobase of the exon model per million mapped fragments (FPKM) to assess the degree of overlap between GA3-and non-treated fruits. A total of 4275 DEGs were detected, with 2782 up-regulated and 1513 down-regulated genes ( Figure 3A). Among all DEGs, 89.47% (3825) unigenes were annotated using at least one database (Table S2).
Many DEGs annotated within the biological process category of Gene ontology (GO) were analyzed ( Figure 3B). For instance, most genes were identified to play essential roles in metabolic, cellular, and single-organism processes, biological regulation, response to stimulus, and signaling. In addition, they were also found to be involved in other multiple biological processes, including developmental, multicellular organismal, and reproductive processes (Table S3). We also evaluated DEGs based on their cellular components and their molecular functions. Regarding the cellular component category, most DEGs fell within the cell part, membrane, organelle, macromolecular complex, and membrane-enclosed lumen. As for molecular function classification, the highest abundance of DEGs was related to catalytic activity and binding. At the same time, other fascinating groups, including transporter, structural molecule, and transcription factor activities, were clustered. Moreover, most DEGs also participated in molecular function regulator, electron carrier, antioxidant, and signal transducer activities. Clusters of Orthologous Groups of proteins (COG) assignment was performed to classify the functions of DEGs ( Figure 3C). GA3 treatment significantly regulated the expression of numerous genes encoding carbohydrate transport and metabolism, indicating their potential roles in the TSS accumulation in R. roxburghii fruits after GA3 application. Other COG functions triggered by GA3

Transcriptome-Scale Analysis of GA 3 -Responsive DEGs in R. roxburghii Fruits
Differentially expressed genes (DEGs) were analyzed using the Fragments Per Kilobase of the exon model per million mapped fragments (FPKM) to assess the degree of overlap between GA 3 -and non-treated fruits. A total of 4275 DEGs were detected, with 2782 up-regulated and 1513 down-regulated genes ( Figure 3A). Among all DEGs, 89.47% (3825) unigenes were annotated using at least one database (Table S2). were translation, ribosomal and biogenesis, energy production and conversion, and secondary metabolites biosynthesis. The Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis can identify the major biochemical and signal transduction pathways in which the DEGs were involved. The significant entries of DEGs were implicated in the carbon metabolism, oxidative phosphorylation, citrate cycle, glycolysis/gluconeogenesis, phenylpropanoid biosynthesis, and biosynthesis of amino acids ( Figure 4A). DEGs associated with ascorbate,  Many DEGs annotated within the biological process category of Gene ontology (GO) were analyzed ( Figure 3B). For instance, most genes were identified to play essential roles in metabolic, cellular, and single-organism processes, biological regulation, response to stimulus, and signaling. In addition, they were also found to be involved in other multiple biological processes, including developmental, multicellular organismal, and reproductive processes (Table S3). We also evaluated DEGs based on their cellular components and their molecular functions. Regarding the cellular component category, most DEGs fell within the cell part, membrane, organelle, macromolecular complex, and membraneenclosed lumen. As for molecular function classification, the highest abundance of DEGs was related to catalytic activity and binding. At the same time, other fascinating groups, including transporter, structural molecule, and transcription factor activities, were clustered. Moreover, most DEGs also participated in molecular function regulator, electron carrier, antioxidant, and signal transducer activities. Clusters of Orthologous Groups of proteins (COG) assignment was performed to classify the functions of DEGs ( Figure 3C). GA 3 treatment significantly regulated the expression of numerous genes encoding carbohydrate transport and metabolism, indicating their potential roles in the TSS accumulation in R. roxburghii fruits after GA 3 application. Other COG functions triggered by GA 3 were translation, ribosomal and biogenesis, energy production and conversion, and secondary metabolites biosynthesis.
The Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis can identify the major biochemical and signal transduction pathways in which the DEGs were involved. The significant entries of DEGs were implicated in the carbon metabolism, oxidative phosphorylation, citrate cycle, glycolysis/gluconeogenesis, phenylpropanoid biosynthesis, and biosynthesis of amino acids ( Figure 4A). DEGs associated with ascorbate, aldarate, tyrosine, tryptophan, starch, and sucrose metabolism were also enriched ( Figure 4B).
were translation, ribosomal and biogenesis, energy production and conversion, and secondary metabolites biosynthesis. The Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis can identify the major biochemical and signal transduction pathways in which the DEGs were involved. The significant entries of DEGs were implicated in the carbon metabolism, oxidative phosphorylation, citrate cycle, glycolysis/gluconeogenesis, phenylpropanoid biosynthesis, and biosynthesis of amino acids ( Figure 4A). DEGs associated with ascorbate, aldarate, tyrosine, tryptophan, starch, and sucrose metabolism were also enriched ( Figure  4B).

DEGs Related to Primary Metabolism after GA 3 Application in R. roxburghii Fruits
Considering that the TSS accumulation of R. roxburghii fruits decreased significantly after exogenous GA 3 treatment, we determined the DEGs concerning primary metabolism. Their pathways are roughly outlined in Figure 5. The KEGG enrichments suggested that 52, 44, and 15 DEGs were associated with glycolysis/gluconeogenesis, citrate cycle (TCA), and the pentose phosphate pathway, respectively. Regarding the glycolysis/gluconeogenesis pathway, the expression profiles of DEGs displayed two clusters. Among these, 32.69% of DEGs (17) were significantly enhanced in fruits after gibberellin application. For example, two unigene-encoded phosphoenolpyruvate carboxykinase were significantly raised. Moreover, we found that unigenes encoding pyruvate kinase, pyruvate dehydrogenase E1 component, fructose-bisphosphate aldolase, glyceraldehyde-3-phosphate dehydrogenase, dihydrolipoyl dehydrogenase, and glyceraldehyde 3-phosphate dehydrogenase were also substantially increased, suggesting that these DEGs might play critical roles in decomposing soluble sugar. (Table S4). Moreover, about half of DEGs related to the pentose phosphate pathway showed a similar tendency after GA 3 treatment. Among these, the expression levels of unigenes annotated as 6-phosphogluconate dehydrogenase and transaldolase were promoted. Likewise, 43.19% of DEGs involved in the TCA pathway were significantly enhanced, especially ATP citrate (pro-S)-lyase, malate dehydrogenase, and phosphoenolpyruvate carboxykinase. Hence, some candidate crucial genes in the primary metabolic pathway by exogenous GA 3 in R. roxburghii fruits were identified, indicating that they might play crucial roles in degrading the total soluble sugar of R. roxburghii fruits.
after exogenous GA3 treatment, we determined the DEGs concerning primary metabo-lism. Their pathways are roughly outlined in Figure 5. The KEGG enrichments suggested that 52, 44, and 15 DEGs were associated with glycolysis/gluconeogenesis, citrate cycle (TCA), and the pentose phosphate pathway, respectively. Regarding the glycolysis/gluconeogenesis pathway, the expression profiles of DEGs displayed two clusters. Among these, 32.69% of DEGs (17) were significantly enhanced in fruits after gibberellin application. For example, two unigene-encoded phosphoenolpyruvate carboxykinase were significantly raised. Moreover, we found that unigenes encoding pyruvate kinase, pyruvate dehydrogenase E1 component, fructose-bisphosphate aldolase, glyceraldehyde-3-phosphate dehydrogenase, dihydrolipoyl dehydrogenase, and glyceraldehyde 3-phosphate dehydrogenase were also substantially increased, suggesting that these DEGs might play critical roles in decomposing soluble sugar. (Table S4). Moreover, about half of DEGs related to the pentose phosphate pathway showed a similar tendency after GA3 treatment. Among these, the expression levels of unigenes annotated as 6-phosphogluconate dehydrogenase and transaldolase were promoted. Likewise, 43.19% of DEGs involved in the TCA pathway were significantly enhanced, especially ATP citrate (pro-S)-lyase, malate dehydrogenase, and phosphoenolpyruvate carboxykinase. Hence, some candidate crucial genes in the primary metabolic pathway by exogenous GA3 in R. roxburghii fruits were identified, indicating that they might play crucial roles in degrading the total soluble sugar of R. roxburghii fruits.

Ascorbate and Hormone-Related Signaling Pathways upon Exogenous GA3 Applications
The proposed L-ascorbic acid synthetic and recycling pathways are shown in Figure  6A. We identified and analyzed candidate DEGs implicated in ascorbate metabolism. The expression levels of DEGs encoding biosynthesized ascorbate were remarkably increased, mainly including UDP-glucose 6-dehydrogenase and inositol oxygenase ( Figure 6B). Lascorbic acid can be oxidized to ascorbate catalyzed by ascorbate oxidase and reversely by monodehydroascorbate reductase (NADH) activities. Despite the decrease in unigenes encoding these two enzymes, the reduction in unigenes annotated as L-ascorbate oxidase

Ascorbate and Hormone-Related Signaling Pathways upon Exogenous GA 3 Applications
The proposed L-ascorbic acid synthetic and recycling pathways are shown in Figure 6A. We identified and analyzed candidate DEGs implicated in ascorbate metabolism. The expression levels of DEGs encoding biosynthesized ascorbate were remarkably increased, mainly including UDP-glucose 6-dehydrogenase and inositol oxygenase ( Figure 6B). Lascorbic acid can be oxidized to ascorbate catalyzed by ascorbate oxidase and reversely by monodehydroascorbate reductase (NADH) activities. Despite the decrease in unigenes encoding these two enzymes, the reduction in unigenes annotated as L-ascorbate oxidase is more significant than the expression of monodehydroascorbate reductase, resulting in the accumulation of L-ascorbic acid in R. roxburghii fruits (Table S5).
To investigate the roles of other endogenous hormones in R. roxburghii fruits, genes associated with hormone signaling pathways were identified upon GA 3 treatment ( Figure 6C). Among them, DEGs were mainly involved in five principal plant hormones: auxin, cytokinins (CK), salicylic acid (SA), abscisic acid (ABA), and ethylene (Table S6). DEGs encoding AUXIN/Indole-3-acetic acid (AUX/IAA), Small auxin-up RNA (SAUR), and Gretchen Hagen 3 (GH3) were found, supporting the crosstalk of the GA 3 signaling pathway with the auxin signaling pathway in R. roxburghii. Given that AUX/IAA proteins repressed auxin signaling, the decrease in DEGs annotated as AUX/IAA suggested that GA 3 spraying enhanced the auxin signaling pathway. Likewise, histidinecontaining phosphotransferase protein 1-like (RrAHP) was stimulated to regulate cell division. By contrast, DEGs related to the signaling pathways of salicylic acid, abscisic acid, and ethylene were reduced, indicating that exogenous GA 3 may be conducive to alleviating the inhibitory effect of ABA and ethylene signal transduction that is commonly known to inhibit the growth of specific tissues and cells. These results indicated that exogenous GA 3 applications could affect other hormone-related pathways in R. roxburghii. associated with hormone signaling pathways were identified upon GA3 treatment ( Figure  6C). Among them, DEGs were mainly involved in five principal plant hormones: auxin, cytokinins (CK), salicylic acid (SA), abscisic acid (ABA), and ethylene (Table S6). DEGs encoding AUXIN/Indole-3-acetic acid (AUX/IAA), Small auxin-up RNA (SAUR), and Gretchen Hagen 3 (GH3) were found, supporting the crosstalk of the GA3 signaling pathway with the auxin signaling pathway in R. roxburghii. Given that AUX/IAA proteins repressed auxin signaling, the decrease in DEGs annotated as AUX/IAA suggested that GA3 spraying enhanced the auxin signaling pathway. Likewise, histidine-containing phosphotransferase protein 1-like (RrAHP) was stimulated to regulate cell division. By contrast, DEGs related to the signaling pathways of salicylic acid, abscisic acid, and ethylene were reduced, indicating that exogenous GA3 may be conducive to alleviating the inhibitory effect of ABA and ethylene signal transduction that is commonly known to inhibit the growth of specific tissues and cells. These results indicated that exogenous GA3 applications could affect other hormone-related pathways in R. roxburghii.

DEGs Related to Transcription Factors in R. roxburghii Fruits
The gene expression of transcription factors influenced by gibberellin in R. roxburghii fruits was investigated. A total of 577 candidate TFs were annotated in the transcriptome of R. roxburghii fruits. We obtained 24 TFs with significant differences (Table S7). TF members comprised the MYB family (MYB82 and R2R3 MYB), WRKY family (WRKY41 and WRKY75), bZIP family, MADS, NAC, and others. The expression levels of TGA3, C2H2, R2R3 MYB, and WRKY were significantly increased (Figure 7). These TFs might play significant roles in regulating the development of R. roxburghii fruit, providing information for studying the functions of TFs in the promoting effect of GA3 on the fruit quality of R. roxburghii.

DEGs Related to Transcription Factors in R. roxburghii Fruits
The gene expression of transcription factors influenced by gibberellin in R. roxburghii fruits was investigated. A total of 577 candidate TFs were annotated in the transcriptome of R. roxburghii fruits. We obtained 24 TFs with significant differences (Table S7). TF members comprised the MYB family (MYB82 and R2R3 MYB), WRKY family (WRKY41 and WRKY75), bZIP family, MADS, NAC, and others. The expression levels of TGA 3 , C2H2, R2R3 MYB, and WRKY were significantly increased (Figure 7). These TFs might play significant roles in regulating the development of R. roxburghii fruit, providing information for studying the functions of TFs in the promoting effect of GA 3 on the fruit quality of R. roxburghii.

Validation of DEGs by qRT-PCR
To verify the reliability of the transcriptome results, we randomly selected nine genes in the related pathways of R. roxburghii fruits for qRT-PCR validation, including unigenes involved in L-ascorbate acid, sugar, hormone signaling, and key transcriptional factors. Primers were listed (Table S8), and their expression levels were calculated using the 2 −∆∆Ct method. We further compared the expression data of the DEGs obtained by RNA-seq and qRT-PCR, and the results displayed a similar tendency ( Figure 8A). The correlation between RNA-Seq (FPKM) and qPCR (2 −∆∆Ct ) results for the nine DEGs was also calculated using log 2 fold variation measurements to produce a scatter plot. The qRT-PCR results of nine DEGs were significantly similar to the RNA-seq results (R 2 = 0.7598), indicating that our RNA-seq data were accurate and reproducible ( Figure 8B).

Validation of DEGs by qRT-PCR
To verify the reliability of the transcriptome results, we randomly selected nine genes in the related pathways of R. roxburghii fruits for qRT-PCR validation, including unigenes involved in L-ascorbate acid, sugar, hormone signaling, and key transcriptional factors. Primers were listed (Table S8), and their expression levels were calculated using the 2 −ΔΔCt method. We further compared the expression data of the DEGs obtained by RNA-seq and qRT-PCR, and the results displayed a similar tendency ( Figure 8A). The correlation between RNA-Seq (FPKM) and qPCR (2 −ΔΔCt ) results for the nine DEGs was also calculated using log2fold variation measurements to produce a scatter plot. The qRT-PCR results of nine DEGs were significantly similar to the RNA-seq results (R 2 = 0.7598), indicating that our RNA-seq data were accurate and reproducible ( Figure 8B).

Validation of DEGs by qRT-PCR
To verify the reliability of the transcriptome results, we randomly selected nine genes in the related pathways of R. roxburghii fruits for qRT-PCR validation, including unigenes involved in L-ascorbate acid, sugar, hormone signaling, and key transcriptional factors. Primers were listed (Table S8), and their expression levels were calculated using the 2 −ΔΔCt method. We further compared the expression data of the DEGs obtained by RNA-seq and qRT-PCR, and the results displayed a similar tendency ( Figure 8A). The correlation between RNA-Seq (FPKM) and qPCR (2 −ΔΔCt ) results for the nine DEGs was also calculated using log2fold variation measurements to produce a scatter plot. The qRT-PCR results of nine DEGs were significantly similar to the RNA-seq results (R 2 = 0.7598), indicating that our RNA-seq data were accurate and reproducible ( Figure 8B).

Exogenous Gibberellin Spraying Regulates Fruit Qualities of R. roxburghii
The GA application in R. roxburghii induced the apparent physiological divergences of the fruits, including promoting fruit size and weight, ascorbic acid accumulation, and prickle development. The increase in fruit size was consistent with other horticultural fruits, such as grapes, apples, and tomatoes [7]. Interestingly, the development of fruit prickle was also promoted by GA 3 , the same as the functions of increased wax layer thickness by GA 3 treatment in apples, which could reduce fruit water loss rate [8]. Meanwhile, the increase in ascorbic acid and prickle length accompanied by the reduction in TSS concentration was noticed, which might be aroused by the flux allocation between primary and secondary metabolism [24]. Moreover, we observed that the seed number was significantly decreased, possibly due to fertilization failure, sharing the same phenomenon of parthenocarpy (fertilized-independent fruit) that could be induced by gibberellin [4]. Similarly, overexpression of GA biosynthesis enzymes could significantly decrease the seed number of tomatoes [25]. Likewise, CRISPR/Cas9-generated knockout GA biosynthesis enzyme mutants increased the seed numbers of soybeans (Glycine max), suggesting that GA 3 application negatively regulated seed set [26].

GA 3 Facilitated the Decomposition of Total Soluble Sugars by increasing Crucial Genes in Primary Metabolism
In this current study, six DGE libraries were constructed using RNA-Seq and used to screen DEGs after GA 3 treatment. A total of 26,294 unique sequences were obtained, of which 24,677 unigenes were annotated. The results of the present study would be helpful to clone coding domain sequences and analyze gene families in R. roxburghii. Multiple genes were altered during fruit setting after GA 3 treatment through RNA-Seq analysis. The roles of many annotated genes showed the same plant responses with GAs in other species. For example, c241716.graph_c1 was annotated as an expansion, which was also increased in Malus domestica after the spray of gibberellin [27]. Moreover, GA 3 significantly decreases the transcript of genes encoding UDP-glycosyltransferase in R. roxburghii, as in F. vesca [28]. We analyzed GO and KEGG pathways and determined that 3825 DEGs participated in several pathways, yielding novel comprehensive insights into the GA 3 response in R. roxburghii fruits.
A few transcriptional events of GAs drive drastic shifts at the primary metabolites and developmental events [29]. We found that starch and sucrose metabolism pathways were significantly enriched, similar to the metabolic pathway that GAs induced during the fruit setting of triploid Loquat (Eriobotrya japonica) [21]. The RNA-seq data displayed that multiple DEGs involved in the glycolytic pathway, TCA, and pentose phosphate pathway were up-regulated, indicating that GAs accelerated the decomposing of storage sugars [30]. Spraying GA 3 reduced TSS contents in R. roxburghii fruits, mainly because of the enhanced DEGs in the primary metabolisms. A similar tendency was found that the sugar contents were also reduced in germinated seeds of Fraxinus hupehensis treated with GAs [30]. Most enzymes were encoded by gene families that displayed various expression levels in R. roxburghii. However, many enzymes might be increased after GA application by considering expressions of all genes in the same family [31]. For example, glyceraldehyde-3phosphate dehydrogenase (GAPDH), a key enzyme in the glycolytic pathway that catalyzes the conversion of D-glyceraldehyde 3-phosphate to 1,3-diphosphoglycerate, was annotated by four genes. The increase in one unigene (c266210.graph_c0) is more extensive than the sum of the other four genes (c252200.graph_c0, c252200.graph_c1, c265626.graph_c0, and c265626.graph_c1) (Table S4). Therefore, GA 3 facilitated the oxidative decomposition of soluble sugars by increasing crucial genes in primary metabolism.

The Myoinositol Pathway Was Required for the Accumulation of Ascorbic Acid in R. roxburghii
The abundance of ascorbic acid was influenced by the biosynthesis and recycling pathways [32]. GA 3 spraying could significantly enhance the concentration of ascorbic acid. Moreover, the RNA-seq determined that DEGs implicated in the biosynthesis of L-ascorbate were all impressively enhanced, especially UDP-glucose 6-dehydrogenase and inositol oxygenase, possibly suggesting that L-galactose and myoinositol pathways were essential for the accumulation of ascorbic acid. Meanwhile, exogenous GA 3 substantially decreased the expression of L-ascorbate oxidase, increasing the reduced content and redox state. The results would help the understanding of the molecular basis regulating ascorbic acid in R. roxburghii fruits.

The Altered Hormonal Signaling and TF Induced by GA 3 Were Responsible for the Fruit Development of R. roxburghii
Transcriptome analyses illustrated that other hormone signaling pathways showed enhanced sensitivity to the application of GA 3 . Five auxin-related DEGs were identified. c246759.graph_c0 encoding (AUX/IAA), which could bind to auxin response factors, was significantly down-regulated (Table S6) [33]. An early study also proved that silencing of tomato SlIAA9 led to the parthenocarpic fruit of Solanaceae [34]. Moreover, two genes encoding GH3 were decreased. Since GH3 was reported to catalyze the synthesis of indole-3acetic acid (IAA)-amino acid conjugates, the reduction in GH3 indicated a high level of free IAA concentration [35]. During the early periods of fruit expansion, the transcript of GH3 was at a lower level, according to the high auxin concentration at the same development stage in apples [36]. Hence, GAs may play a role in increasing auxin signaling to regulate the fruit development of R. roxburghii. Similarly, Illumina HiSeq high-throughput sequencing revealed that auxin-responsive genes, playing the same roles of c254917.graph_c0 in R. roxburghii, also increased in F. vesca when treated with 50 mg/L GA 3 [37] . In addition, histidine-containing phosphotransfer protein 1-like involved in the cytokinin signaling pathway was promoted after GA 3 treatment, indicating that GAs possibly induced fruit physiological changes via the cytokinin signal pathway [38]. ABA inhibits fruit growth in the early stages of strawberries, accompanied by low ABA levels in early development and a sharp increase during ripening [7]. Our study showed the decreased DEGs implicated in the ABA signaling pathway. Morever, some DEGs involved in other phytohormones were also determined. These results confirmed that auxin, ethylene, abscisic acid, and salicylic acid, which GAs can directly or indirectly trigger, might regulate fruit development, similar to the complex hormonal regulation in tomatoes [39].
Based on DGE analysis, the expression levels of the genes encoding several transcription factors were significantly changed during GA 3 treatment in R. roxburghii, including genes encoding NAC, MADS, bHLH, MYBs, and others. TFs play crucial roles in fruit qualities. For example, the development of R. roxburghii fruit prickles is determined by the complex of MYB-bHLH-WD40. Interestingly, the prickle length was significantly enhanced by GA application in our study. The same treatment also induced trichome formation by MYB (GLABROUS 1) in Arabidopsis [40]. RrGL1 (MYB), RrGL3/EGL3 (bHLH), and RrTTG1 (WD40) are responsible for prickle development [41,42]. The two former TFs are required for prickle elongation. In contrast, WD40 is necessary for inducing prickle initiation. Interestingly, The RNA-seq analyzed the transcripts of genes encoding R2R3 MYB and bHLH041 that were significantly up-regulated, showing that the conserved domains with RrGL1 and RrGL3/RrEGL3 might play similar effects on prickle development. However, RrTTG1 displayed no significant change in R. roxburghii, which explained why prickle only became longer but not numerous [17].

The Plant Materials
Seedlings of the Rosa roxburghii Tratt cultivar 'Guinong 5' were planted at the garden of Guizhou Normal University in Guiyang. The trees were 7 years old, in the adult phase. Flowers of R. roxburghii in full blossom were sprayed with 200 mg/L of GA 3 (Sigma-Aldrich, cat# Sigma G7645, St. Louis, MI, USA) daily for 7 days as treatment and distilled H 2 O as a control on 15 May 2022 ( Figure S1). We collected at least twenty fruits at 60 DAP and 120 DAP for morphology analysis. Some fresh fruits treated with GA 3 and control at 120 DAP were harvested from three separate seedings for measurement. The others for RNA extraction were snap-frozen immediately in nitrogen and stored in a −80 • C refrigerator. Transcriptome sequencing was performed with three biological replications.

Plant Biomass Yield and Biochemical Analysis of R. roxburghii
Physiological data of at least twenty fruits from three separate seedlings were recorded on the following traits: average fruit weight (g), fruit shape index (calculated as longitudinal diameter (cm)/horizontal diameter (cm)), seed number, and prickle length per each fruit. We weighed 100 g of fruits per treatment and homogenized it with 5000 rpm. Then, 2 g of the sample was collected, powdered with liquid nitrogen with 20 mL ethanol (80%, v/v), and sonicated for 15 min at 80 • C. After evaporation, the residue was dissolved with 2 mL of distilled water for HPLC analysis. The percentages (g/g) of total soluble sugar (TSS) and L-ascorbic acid concentration (g/g) were determined using high-performance liquid chromatography (HPLC) as described by Huang et al. [31].

RNA Extraction, Library Construction, and Sequencing of R. roxburghii
Total RNA was extracted from R. roxburghii fruits using the TRIzol reagent (Invitrogen, Carlsbad, CA, USA). RNA quality and quantity were determined using a NanoDrop spectrophotometer and Agilent 2100 Bioanalyzer for further transcriptome sequencing. RNA (1 µg) per sample was used for the preparations. Following the manufacturer's recommendations, sequencing libraries were generated using NEBNext ® Ultra™ RNA Library Prep Kit for Illumina ® (NEB, Ipswich, MA, USA). The library preparations were sequenced on an Illumina Hiseq 2000 platform. Clean data were obtained by removing reads containing adapter, ploy-N, and low-quality reads from raw data. All the RNAseq raw data have been uploaded and are available in Genbank under accession number PRJNA1003688.

De Novo Assembles and Unigenes Annotation of R. roxburghii
Transcriptome assembly was accomplished based on the left. fq and right. fq using Trinity with min_kmer_cov set to 2 by default and all other parameters set to default [43]. Gene functions were annotated and based on BLASTx with cutoff E-value ≤ 1 × 10 −5 and HMMER ≤ 1 × 10 −10 [44] using Nr, GO [45], COG [46], and KEGG [47] databases.

Differential Expression Analysis of Unigenes Expressed in R. roxburghii Fruits
The expression level was estimated with RNA-Seq by Expectation Maximization (RSEM). FPKM values indicate the abundance of the corresponding unigene [48]. DESeq R package (1.10.1) was adopted to compare the GA 3 treatment with the reference group [49]. The resulting p values were adjusted using Benjamini and Hochberg's approach for controlling the false discovery rate (FDR). An adjusted p-value < 0.001 and |log 2 (fold change)| ≥ 2 were set as the threshold for determining the DEGs. GO and KEGG pathway analyses were conducted to interpret the DEGs.

Real-Time Quantitative PCR
Total RNAs from the R. roxburghii fruits were isolated, and RT-PCR Kit ® (TaKaRa, Shiga, Japan) was performed using 2 µg of RNA and oligo dT-adaptor primer. Real-time quantitative PCR was performed in a LightCycler480 instrument (Roche, Basel, Switzerland). Each reaction contained 10 µL of SYBR green PCR master mix (TaKaRa, Shiga, Japan), 1.0 µL cDNA, 200 nM primers, and ddH 2 O up to the final volume of 20 µL. Amplification was performed at 95 • C for 5 min, followed by 40 cycles at 95 • C for 20 s, 58 • C for 30 s, and 72 • C for 1 min. The expression levels relative to the RrActin were estimated using the 2 −∆∆Ct method. All samples were performed for three biological replicates.

Statistical Analysis
The results were statistically evaluated by one-way analysis of variance (ANOVA) followed by Tukey's test using Statistical Program for Social Sciences (SPSS) program version 20.0 (SPSS Inc., Chicago, IL, USA). Statistical differences with p values under 0.05 are considered significant, using an asterisk that indicates a significant difference. "*": p < 0.05; "**": p < 0.01.

Conclusions
Our study reveals a complex network of genes regulated by GA 3 treatment in R. roxburghii fruits. The enriched GO terms and KEGG pathways highlight the multifaceted roles of GAs in promoting cell growth, accelerating fruit ripening, and modulating stress responses. These findings provide valuable insights into the molecular mechanisms of GA action in R. roxburghii fruits and provide a foundation for future studies to improve fruit quality and yield. Although significant DEGs have been identified regarding the roles of GAs in R. roxburghii, many problems remain to be studied. Efficient genetic transformation and plant regeneration systems are still lacking for R. roxburghii, limiting the functional validation of gene studies on relevant studies. It is worth constructing transgenetic systems in R. roxburghii in the future, which will undoubtedly promote further understanding of the molecular mechanisms underlying the effects of GAs on fruit development in R. roxburghii.

Supplementary Materials:
The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/plants12193425/s1, Figure S1. The experimental design map used in this study. Table S1. Statistics of the reads and RNA-seq mapping of R. roxburghii fruits; Table S2. All DEGs between GA 3 -and non-treated fruits and their annotations were listed; Table S3. Top GO enriched functions of DEGs between GA 3 -and non-treated fruits; Table S4. DEGs involved in the primary metabolism of R. roxburghii fruits; Table S5. DEGs associated with ascorbate metabolism of R. roxburghii fruits; Table S6. DEGs implicated in the endogenous hormone signaling pathway of R. roxburghii fruits; Table S7. The expression levels of DEGs encoding transcription factors following GA 3 treatment; Table S8. Primers used in qRT-PCR analysis.

Data Availability Statement:
The RNA-Seq of raw data is publicly available in the NCBI under number: PRJNA1003688 with the link: https://www.ncbi.nlm.nih.gov/sra/PRJNA1003688 accessed on 11 August 2023.

Conflicts of Interest:
The authors declare no conflict of interest.