Integrated Metabolomics and Transcriptomics Analysis Reveals New Insights into Triterpene Biosynthesis in Rosa rugosa

Rosa rugosa is highly regarded for its aesthetic and therapeutic qualities. In particular, R. rugosa’s flowers are known to produce essential oils containing a mixture of volatile terpenes, phenylpropanoids, and other compounds. Despite this, extensive research exists on volatile terpenes in flowers, while the knowledge of non-volatile terpenes in distinct tissues is still limited. Using UPLC–ESI–MS/MS, a comprehensive analysis of the terpene metabolites in five different tissues of R. rugosa was conducted. These metabolites accumulated in distinct tissues, and the majority of them were triterpenoids. Transcriptome data were collected from five tissues using RNA-seq. Transcriptomics and metabolomics were utilized to evaluate the triterpene biosynthesis pathway, resulting in new insights into its regulation and biosynthesis. The RrOSC10 was identified as a key enzyme in converting 2,3-oxidosqualene into α-amyrin, potentially contributing to the triterpene biosynthesis pathway. Furthermore, the expression of the RrOSC10 gene was upregulated by salinity for 0.5 h and 1 h, with subsequent downregulation at 2 h. This study lays a foundation for future research on the biosynthesis and accumulation of triterpenes in R. rugosa.


Introduction
Rosa rugosa Thunb is a flowering shrub species native to eastern Asia that belongs to the Rosaceae family.It is known for its numerous attractive features, adaptability, and traditional medicinal application, which have led to its worldwide recognition [1].R. rugosa has a wide range of applications, including the development of teas, juices, and wines [2].Notably, their significant content includes a variety of biologically active compounds, such as anthocyanins, terpenoids, procyanidins, and proanthocyanidins [3][4][5].Particularly, R. rugosa flowers have a pleasant aroma and are utilized to produce essential oil, which typically contains volatile mono-sesquiterpenes and phenylpropanoids/benzenoids [6].A comprehensive understanding of the terpenoid composition of R. rugosa is needed to highlight its medicinal importance [7].For instance, these plants have been utilized in traditional herbal medicine to treat various conditions, including stomach issues, pain, and inflammatory diseases [8].While numerous studies have focused on the regulation and biosynthesis of volatile terpenoids in flowers [9][10][11][12], this study shifts the focus from volatile to non-volatile terpenes, which are present in different plant tissues, to identify previously undiscovered compounds that play a pivotal role in the medicinal properties of plants [13,14].
The R. rugosa plant is renowned for its medicinal and edible properties, along with its production of a wide variety of secondary metabolites.Nevertheless, there is still a lack of understanding regarding its non-volatile terpenoids and the molecular process of triterpene biosynthesis in the plant.Non-volatile terpenoids, particularly triterpenoids, play an important role in the ecological interactions of plants and hold significant relevance in their medicinal applications, which have captivated human interest for centuries.For instance, triterpene compounds, such as tormentic acid and rosamultin, are mainly present in the roots of R. rugosa.These compounds exhibit anti-inflammatory and anti-HIV protease activities and protect cardiac muscle cells against oxidative stress and cellular apoptosis [31].The biosynthesis of rosamultin begins with "α-amyrin" and undergoes a three-step process catalyzed by cytochrome P450 monooxygenases (CYP450s) enzymes.Subsequently, a glycosylation reaction is carried out by uridine diphosphate glycosyltransferase (UGT) [30].CYP450s exhibit several biocatalytic activities, including oxidation, epoxidation, hydroxylation, and demethylation [32].In the triterpene biosynthesis, the CYP716A enzyme is capable of catalyzing α-amyrin to ursolic acid, lupeol to betulic acid, and β-amyrin to oleanolic acid [33,34].The CYP716C enzyme can convert ursolic acid to corosolic acid and oleanolic acid to maslinic acid [35].
This study aimed to address the lack of knowledge about non-volatile terpenoid biosynthesis in R. rugosa through two primary objectives.The first objective was to comprehensively determine the presence of non-volatile terpenoids in different tissues of R. rugosa.The ultra-performance liquid chromatography-electrospray ion source-mass spectrometry/mass spectrometry (UPLC-ESI-MS/MS) system was utilized to analyze the intricate chemical profiles of non-volatile terpenoids in R. rugosa.The second objective was to investigate the genetic basis of triterpene formation in R. rugosa by identifying the specific OSC genes responsible for triterpene formation using the integration of metabolomics and transcriptomics approaches.To the best of our knowledge, this is the first report integrating metabolomics and transcriptomics data in analyzing different rose tissues, elucidating the accumulation patterns of triterpenoid compounds, and uncovering the molecular mechanisms underlying their formation.

The Accumulation of Non-Volatile Terpenoids across Various Tissues of R. rugosa
Samples of petals (P), leaves (L), stamens (S), roots (R), and fruits (F) were collected to analyze the variation in the non-volatile terpenoids using the UPLC-ESI-MS/MS system (Figure 1) (Supplementary Figure S1).In this study, the accuracy of expression analyses was validated by the Pearson correlation coefficient (PCC), which showed a significant correlation among the three replicated tissues (Supplementary Figure S2).A total of 74 terpenoids were identified and classified into the following three groups: 6 monoterpenoid derivatives, 5 sesquiterpenoid derivatives, and 63 triterpenes and triterpene saponins (Supplementary Figure S3A) (Supplementary Table S1).A principal component analysis (PCA) showed that the biological replicates had similar metabolite profiles, and the first two principal components accounted for 41.23% (PC1) and 22.94% (PC2), respectively.The non-volatile terpenoids were found to accumulate in both the fruit and root tissues.Similarly, a high similarity was observed in the accumulation of non-volatile terpenoids in the petals, leaves, and stamen tissues (Supplementary Figure S3B).

The Accumulation of Non-Volatile Terpenoids across Various Tissues of R. rugosa
Samples of petals (P), leaves (L), stamens (S), roots (R), and fruits (F) were collected to analyze the variation in the non-volatile terpenoids using the UPLC-ESI-MS/MS system (Figure 1) (Supplementary Figure S1).In this study, the accuracy of expression analyses was validated by the Pearson correlation coefficient (PCC), which showed a significant correlation among the three replicated tissues (Supplementary Figure S2).A total of 74 terpenoids were identified and classified into the following three groups: 6 monoterpenoid derivatives, 5 sesquiterpenoid derivatives, and 63 triterpenes and triterpene saponins (Supplementary Figure S3A) (Supplementary Table S1).A principal component analysis (PCA) showed that the biological replicates had similar metabolite profiles, and the first two principal components accounted for 41.23% (PC1) and 22.94% (PC2), respectively.The non-volatile terpenoids were found to accumulate in both the fruit and root tissues.Similarly, a high similarity was observed in the accumulation of non-volatile terpenoids in the petals, leaves, and stamen tissues (Supplementary Figure S3B).The accumulation patterns of non-volatile terpenoids were analyzed through Kmeans and clustering heatmap analyses, leading to the identification of ten subclasses, labeled 1 to 10 (Figure 2) (Supplementary Figure S3C).The tissue-specific accumulation patterns of terpenoids were analyzed.Subclass 6 showed the highest levels of two nonvolatile terpenoids in petals, while subclass 2 exhibited the highest levels of six non-volatile terpenoids in stamens.Subclass 1 exhibited the highest levels of five non-volatile terpenoids in leaves, whereas the roots contained the highest levels of non-volatile terpenoids from subclasses 3, 5, 7, 8, and 10.The highest levels of non-volatile terpenoids from subclasses 4 and 9 were found in fruits.Analyzing the tissue-specific patterns of terpenoids helped to identify potential genes associated with terpenoid biosynthetic pathways through metabolite-gene correlation analysis.It is noteworthy that triterpenoids were predominantly synthesized in roots and fruits.The accumulation patterns of non-volatile terpenoids were analyzed through K-means and clustering heatmap analyses, leading to the identification of ten subclasses, labeled 1 to 10 (Figure 2) (Supplementary Figure S3C).The tissue-specific accumulation patterns of terpenoids were analyzed.Subclass 6 showed the highest levels of two non-volatile terpenoids in petals, while subclass 2 exhibited the highest levels of six non-volatile terpenoids in stamens.Subclass 1 exhibited the highest levels of five non-volatile terpenoids in leaves, whereas the roots contained the highest levels of non-volatile terpenoids from subclasses 3, 5, 7, 8, and 10.The highest levels of non-volatile terpenoids from subclasses 4 and 9 were found in fruits.Analyzing the tissue-specific patterns of terpenoids helped to identify potential genes associated with terpenoid biosynthetic pathways through metabolite-gene correlation analysis.It is noteworthy that triterpenoids were predominantly synthesized in roots and fruits.The analysis of differentially accumulated metabolites (DAMs) revealed a significant difference between the petals and roots.Specifically, 9 DAMs showed significant upregulation in the roots, while only one DAM (rosacorenol) displayed significant downregulation in the petals (Figure 3A).Interestingly, among these nine compounds, corosolic acid methyl ester, ursonic acid, 3-epiursolic acid, and rosamultin are likely derivatives of αamyrin.Similarly, 9 DAMs were significantly upregulated in fruits compared to petals, while only one DAM (rosacorenol) showed significant downregulation in petals (Figure 3B).These compounds, such as corosolic acid methyl ester and ursonic acid, were also upregulated in fruits, reflecting their levels in roots.The analysis of differentially accumulated metabolites (DAMs) revealed a significant difference between the petals and roots.Specifically, 9 DAMs showed significant upregulation in the roots, while only one DAM (rosacorenol) displayed significant downregulation in the petals (Figure 3A).Interestingly, among these nine compounds, corosolic acid methyl ester, ursonic acid, 3-epiursolic acid, and rosamultin are likely derivatives of α-amyrin.Similarly, 9 DAMs were significantly upregulated in fruits compared to petals, while only one DAM (rosacorenol) showed significant downregulation in petals (Figure 3B).These compounds, such as corosolic acid methyl ester and ursonic acid, were also upregulated in fruits, reflecting their levels in roots.

Transcriptome Profiling across Different Tissues of R. rugosa
Transcriptome data were obtained from the five different tissues of R. rugosa u the RNA-seq method to understand the biosynthetic pathways of terpenoids compre sively.These tissues are comprised of petals, leaves, stamens, roots, and fruit tissue average data generation was 12.23, 8.12, 9.64, 9.63, and 9.10 GB for the respective ti (Table 1).Moreover, the average number of clean reads was 64,953,322 bp, with a min error rate of 0.03%.The quality scores, Q20 (%) and Q30 (%), ranged from 97.71 to and 93.43 to 93.86, respectively.The GC content ranged from 46.52% to 47.46%.In the expression patterns of 29,758 genes are summarized in Supplementary Table S2.

Transcriptome Profiling across Different Tissues of R. rugosa
Transcriptome data were obtained from the five different tissues of R. rugosa using the RNA-seq method to understand the biosynthetic pathways of terpenoids comprehensively.These tissues are comprised of petals, leaves, stamens, roots, and fruit tissue.The average data generation was 12.23, 8.12, 9.64, 9.63, and 9.10 GB for the respective tissues (Table 1).Moreover, the average number of clean reads was 64,953,322 bp, with a minimal error rate of 0.03%.The quality scores, Q20 (%) and Q30 (%), ranged from 97.71 to 97.90 and 93.43 to 93.86, respectively.The GC content ranged from 46.52% to 47.46%.In total, the expression patterns of 29,758 genes are summarized in Supplementary Table S2.DEGs between each tissue were obtained.Non-volatile terpenoids accumulated significantly in the fruit and root tissues, while petals, leaves, and stamen tissues showed high similarity.A further analysis was conducted to compare the characteristics of the petals with the fruit and the petals with the roots.Two sets of comparisons revealed varying quantities of DEGs.Compared with the fruits, 11,650 DEGs were observed in the petals, with 6413 upregulated and 5237 downregulated (Figure 4A).Similarly, in a comparison of petals with roots, 12,289 DEGs were observed, with 7221 upregulated and 5068 downregulated (Figure 4B).In a further investigation into the functions of DEGs, gene ontology (GO) enrichment analysis was conducted to reveal the activation or deactivation of distinct gene functions.Several DEGs were enriched in carbon-oxygen lyase activity and isoprenoid metabolic processes when comparing petals to fruits (Figure 4C).When comparing petals to roots, the DEGs associated with carbon-oxygen lyase activity, hydrocarbon metabolic processes, isoprenoid metabolic processes, and terpene metabolic processes were found to Plants 2024, 13, 1600 6 of 15 be accumulated (Figure 4D).These DEGs are believed to play a role in terpenoid metabolite metabolism, aligning with their tissue specificity in R. rugosa.
parison of petals with roots, 12,289 DEGs were observed, with 7,221 upregulated and 5 downregulated (Figure 4B).In a further investigation into the functions of DEGs, ontology (GO) enrichment analysis was conducted to reveal the activation or deactiva of distinct gene functions.Several DEGs were enriched in carbon-oxygen lyase act and isoprenoid metabolic processes when comparing petals to fruits (Figure 4C).W comparing petals to roots, the DEGs associated with carbon-oxygen lyase activity, hy carbon metabolic processes, isoprenoid metabolic processes, and terpene metabolic cesses were found to be accumulated (Figure 4D).These DEGs are believed to play a in terpenoid metabolite metabolism, aligning with their tissue specificity in R. rugosa

Integrative Analysis of Transcriptome and Metabolome
The DEGs were selected for a metabolite-gene correlation analysis, as depicted in (Supplementary Figure S4).The integration of transcriptome and metabolome analysis resulted in the identification of functional genes in the triterpene metabolic pathway, as demonstrated in (Figure 5).The findings indicated that the roots and fruits were found to have a higher accumulation of triterpenoids, such as maslinic acid, betulinic acid, tormentic acid, and rosamultin.Additionally, key genes that showed a high expression level in roots and fruits were identified, indicating that they were potentially involved in triterpene biosynthesis.Specifically, four OSC genes (evm.TU.Chr1.999,evm.TU.Chr6.4899,evm.TU.Chr6.4912,and evm.TU.Chr6.4443)were identified to potentially facilitate the biosynthesis of triterpene scaffolds from 2,3-oxidosqualene.Additionally, three CYP716A genes (evm.TU.Chr1.1001,evm.TU.Chr3.2946,and evm.TU.Chr3.2942)and four CYP716C genes (evm.TU.Chr5.5528,evm.TU.Chr5.5530,evm.TU.Chr1.3428,and evm.TU.Chr1.3427)were found that are carboxylated at the C-28 position and hydroxylated at the C-2 position.The diverse accumulation patterns of these compounds might have resulted from variations in the gene expression patterns.

Characterization of RrOSC10
The OSC enzyme, due to its engagement in the first step of the triterpene biosynt pathway, was further characterized.A prior phylogenetic analysis indicated that o the 8 OSC genes, only two OSC genes, RrOSC6 (evm.TU.Chr6.4443)and RrO (evm.TU.Chr6.4912),could be potential candidate genes for α-amyrin [30].The pr sequence identity between RrOSC6 and RrOSC10 was 86%.To further study the cat activities of RrOSC6 and RrOSC10, the full-length coding regions were cloned into a expression vector, pESC-Leu, and subsequently transformed into a yeast, A

Characterization of RrOSC10
The OSC enzyme, due to its engagement in the first step of the triterpene biosynthesis pathway, was further characterized.A prior phylogenetic analysis indicated that out of the 8 OSC genes, only two OSC genes, RrOSC6 (evm.TU.Chr6.4443)and RrOSC10 (evm.TU.Chr6.4912),could be potential candidate genes for α-amyrin [30].The protein sequence identity between RrOSC6 and RrOSC10 was 86%.To further study the catalytic activities of RrOSC6 and RrOSC10, the full-length coding regions were cloned into a yeast expression vector, pESC-Leu, and subsequently transformed into a yeast, AM94, respectively.After cultivation, the products were extracted from the yeasts and identified using gas chromatography-mass spectrometry (GC-MS).The results showed that yeast cells expressing RrOSC10 accumulated α-amyrin as a single product (Figure 6A), whereas the yeast cells containing RrOSC6 did not produce any product.This was confirmed by comparing it with authentic standards.
respectively.After cultivation, the products were extracted from the yeasts and identi using gas chromatography-mass spectrometry (GC-MS).The results showed that y cells expressing RrOSC10 accumulated α-amyrin as a single product (Figure 6A), whe the yeast cells containing RrOSC6 did not produce any product.This was confirmed comparing it with authentic standards.
Terpenoids play a significant role in mediating the interaction between plants their environment [16,36].To investigate R. rugosa s response to salt stress, R. rugosa pla were treated with salt, and the expression of the RrOSC10 gene was analyzed using qPCR.The results indicated a significant increase in the mRNA levels of RrOSC10 in roots of R. rugosa at 0.5 h and 1 h, followed by a decrease at 2 h.This fluctuating patt suggested a potential defensive function of RrOSC10 in response to salt-induced st (Figure 6B).Terpenoids play a significant role in mediating the interaction between plants and their environment [16,36].To investigate R. rugosa's response to salt stress, R. rugosa plants were treated with salt, and the expression of the RrOSC10 gene was analyzed using RT-qPCR.The results indicated a significant increase in the mRNA levels of RrOSC10 in the roots of R. rugosa at 0.5 h and 1 h, followed by a decrease at 2 h.This fluctuating pattern suggested a potential defensive function of RrOSC10 in response to salt-induced stress (Figure 6B).

Discussion
Metabolomics is a powerful technique for comprehensively profiling and comparing the diverse secondary metabolites associated with each biological phenotype through highthroughput characterization [37].This approach enables the detection of many secondary metabolites or their derivatives under normal conditions or in response to various environmental stimuli [38].Our research revealed a diverse range of terpenoid compounds and provided new insights into the previously unexplored terpenoid composition in R. rugosa.This approach enhanced the understanding of R. rugosa's chemical profile and established a foundation for future investigations into the medicinal significance of these non-volatile terpenes.Integrating omics data analysis, such as transcriptomics and proteomics, is beneficial for a comprehensive understanding of the biochemical landscape [39,40].This study thoroughly investigated the non-volatile terpenoid profile in various tissues of R. rugosa, offering insightful observations on the distribution and diversity of these secondary metabolites in the plant.The terpenoids displayed distinct patterns, significantly varying their levels and compositions across diverse plant organs.In particular, the fruits and roots showed significant differences compared to the petal, leaf, and stamen tissues.Furthermore, the K-means cluster analysis suggested that specialized biosynthetic pathways or ecological roles are associated with specific tissue types.
Three groups of triterpenoids (ursane-, oleanane-, and lupane-type) were identified in the fruit of R. rugosa, indicating substantial variances in their biosynthesis and accumulation patterns [49].The fruit samples contained all three types of triterpenes, with a notably higher accumulation of compounds from subclades 4 and 9, suggesting a unique origin or biosynthetic pathway specific to fruits.Through NMR analysis, the methanol extract derived from the roots of R. rugosa enabled the identification of 13 triterpenoid saponins, which demonstrated an association with rat intestinal sucrase inhibitors [50].However, two triterpenoid saponins (rosamultin and 2,3,19-trihydroxyurs-12-en-23,28-dioic acid-28-O-glucoside) were detected in the roots in this study.The observed variations may be attributed to differences in measurement methods.Furthermore, no previous studies have investigated the triterpenoid profiles of petals, leaves, and stamen.Therefore, this study represents the first exploration of these profiles within these specific tissues, providing a foundation for understanding their diverse properties.
The combined omics data have successfully been used to investigate the complex metabolic network [51].In this study, the integrative analysis of triterpene metabolites and transcriptomics were obtained.Additionally, the expression patterns of all genes were analyzed, and these data will benefit future studies on R. rugosa.For example, this analysis can also quickly identify candidate modification enzymes like P450 and UGT.
R. rugosa plants are known for their wide adaptability to various environments, which may be associated with the presence of triterpenoids.In this study, the integration of metabolomics, transcriptomics, and biochemical approaches led to the identification of an α-amyrin synthase responsible for triterpene biosynthesis in R. rugosa.A previous study identified a lupeol synthase (RrOSC12: evm.TU.Chr1.999)from R. rugosa through in vitro analysis [30].The transcriptome data revealed that this gene was primarily expressed in the roots, whereas betulinic acid accumulation was higher in the stamens and fruits, suggesting an essential role of CYP716A enzymes in these tissues.The α-amyrin synthases have been identified in Eriobotrya japonica (EjAS) and Malus × domestica as catalysts that convert 2,3-oxidosqualene into α-amyrin as the main product, along with the minor products βamyrin and lupeol [33,52].A biochemical analysis has shown that α-amyrin is the primary product, although in relatively small quantities.This indicated a decrease in RrOSC10 enzymatic activity in the yeast system, emphasizing the need for further optimization to achieve comprehensive characterization.In contrast, RrOSC6 did not produce any product in the yeast system.This may be due to its ineffective expression in the yeast system.The tandem duplication of RrOSC6 and RrOSC10 on chromosome 2 indicates divergent evolution [26].These findings strongly suggested the involvement of these synthases in triterpene synthesis.Moreover, the findings revealed a notable increase in RrOSC10 gene expression in the roots of R. rugosa 0.5 h after salt treatment.It demonstrated that RrOSC10 played an active role in the initial phase of the plant's defense against induced salt stress.The subsequent decrease observed at 2 h post-treatment suggested that the plant likely regulated RrOSC10 expression through feedback after the initial response.This biochemical evidence has provoked further investigation into the potential additional functions of RrOSCs.The production of avenacins in roots by the oat β-amyrin synthase AsbAS1 provided resistance against various fungal pathogens [53].It would be interesting to investigate the potential role of RrOSC10 below the ground.The rice gene OsOSC12/OsPTS1 has a specific expression in anthers and is responsible for disrupting pollen coat formation.This results in humidity-sensitive male sterility [54].The dataset included measurements from stamen tissues, which provide a base for future research investigations of stamen-specific OSCs.Artemisia annua uses OSC2 and CYP716A14V2 enzymes to produce specific triterpenoids in its cuticle wax layer, which protect against both biotic and abiotic stress [55].Salvia mirzayanii plants exhibited elevated levels of oleanolate 3-β-d-glucuronoside-28-glucoside, taxol, and glycyrrhetinate terpenoids when subjected to salt stress [56].Elucidating the function of RrOCS10 and the integrative analysis of CYP450s and UGTs can help us better understanding the complex molecular mechanisms underlying triterpene biosynthesis in R. rugosa [57,58].

Materials and Reagents
The experimental R. rugosa plants were cultivated at Yangzhou University in Yangzhou, China at 32.391 • N, 119.419 • E. The samples used for this study were collected in May 2022 from the following five different parts: petals (P), leaves (L), stamens (S), roots (R), and fruits (F).The experiment utilized 15 samples of R. rugosa, each with three biological replicates.All samples were rapidly frozen in liquid nitrogen after collection and stored in at −80 • C. The compound α-amyrin was purchased from Sigma-Aldrich ® (www.sigmaaldrich.cn(accessed on 6 May 2023), Merck, Darmstadt, Germany).

Metabolite Extraction and Identification
After collecting the samples, they were submitted to Metware Biotechnology Co., Ltd.(Wuhan, China) for qualitative and quantitative analyses of the terpene compounds.The samples were freeze-dried, powdered, and weighed (50 mg per sample).Then, 1.2 mL solution of 70% methanol extract was added for rotational evaporation.Then, the mixture was filtered for 3 min at 12,000 rpm and analyzed using a UPLC-ESI-MS/MS system (UPLC, ExionLC™ AD, https://sciex.com.cn/(accessed on 10 March 2023); MS, Applied Biosystems 6500 Q TRAP, https://sciex.com.cn/(accessed on 10 March 2023)).Agilent SB-C18 1.8 µm, 2.1 mm × 100 mm, was used in the UPLC and ESI experimental conditions.The mobile phase consisted of ultrapure water (with 0.1% formic acid) as Phase A and acetonitrile (with 0.1% formic acid) as Phase B in the chromatographic column.An elution gradient was utilized.The process started with 5% B at 0.00 min, and the proportion of B increased linearly to 95% within 9.00 min and remained at 95% for 1 min.From 10.00 to 11.10 min, the proportion of B was decreased to 5% and kept constant until 14.00 min.The column's temperature was maintained at 40 • C, the flow rate was 0.35 mL/min, and 2 µL of the sample was injected.The electrospray ionization (ESI) temperature was adjusted to 500 • C. In positive ion mode, the ion spray voltage (IS) was 5500 V, while in negative ion mode, it was −4500 V. Gas source I (GSI), gas source II (GSII), and curtain gas (CUR) were set at 50, 60, and 25 psi, respectively.Collision-induced dissociation parameters were set to high.A triple quadrupole (QQQ) scan was carried out using the multiple reaction monitoring (MRM) mode.Nitrogen was used as the collision gas with medium settings.After certain voltage adjustments, each MRM ion pair was optimized by adjusting the collision energy (CE) and declustering potential (DP).A specific set of MRM ion pairs was monitored to capture metabolites during each elution period.

Screening for Differential Metabolites
The statistical function prcomp in R 4.2.3 (www.r-project.org(accessed on 20 March 2023)) was utilized to conduct a principal component analysis (PCA).The R software calculated the Pearson correlation coefficient (PCC) between samples.Differential metabolites between the two groups were identified using an absolute Log 2 FC (|Log 2 FC| ≥ 1.0).Subsequently, quantitative and qualitative analyses of the identified metabolites, along with sample grouping, and heatmaps were generated.The relative levels of the various metabolites in all comparison groups were normalized using unit variance scaling (UV) and then subjected to a K-means cluster analysis.

Transcriptome Analysis
The cDNA libraries were sequenced using the Illumina sequencing platform by Metware Biotechnology Co., Ltd. in Wuhan, China.RNA quality and integrity were assessed using the NanoPhotometer spectrophotometer (IMPLEN, Westlake Village, CA, USA) and Agilent 2100 bioanalyzer (Agilent Technologies, Santa Clara, CA, USA).Library construction was carried out using the NEBNext ® UltraTM RNA Library Prep Kit (NEB, Ipswich, MA, USA).Simultaneously, Oligo(dT) and magnetic beads enriched with mRNA containing a polyA tail.The M-MuLV reverse transcriptase system was utilized to synthesize the first cDNA strand, while the second cDNA strand was synthesized in the DNA polymerase I system using dNTPs.

Identification of Differentially Expressed Genes
Gene alignment was assessed utilizing FeatureCounts version 1.6.2, and the fragments per kilobase of exon model per million mapped fragments (FPKM) were calculated for each gene.A differential expression analysis between the two groups was executed using DESeq2 version 1.22.1, with adjustment of p values using the Benjamini and Hochberg method.The volcano plot illustrated the threshold for significant differential expression using |log2foldchange|.A GO enrichment analysis of differentially expressed genes (DEGs) in petals versus roots and petals versus fruit was performed utilizing GO terms GO:0010333 and GO:0042214.

Integrated Analysis of Transcriptome and Metabolome
The intricate relationship between metabolism and genes across different parts of the rose plant was investigated through a comprehensive analysis of transcriptome and metabolome data.OSCs were identified using the N-terminal domain (PF13249) and Cterminal domain (PF13243) with HMMER and aligned with the sequences [30].CYP716As were aligned with the typical sequence of Medicago truncatula CYP716A12 (Accession DQ335781), and CYP716Cs were aligned with CaCYP716C11 (accession KU878852.1)[59].The screened key genes were involved in the biosynthesis of terpenoid metabolites.

Enzymatic Assays in Yeast
The coding sequences of RrOSCs were initially sub-cloned into the pESC-leu yeast expression vector, followed by transformation into yeast strain AM94.To analyze triterpenoid products, yeast cultures expressing RrOSCs were collected, and the culture extracts were derivatized using trimethylsilylating agents.The derivatized samples were subsequently analyzed using gas chromatography-mass spectrometry (GC-MS) analysis (7890A-7697A).

Salt Treatment and Quantitative Real-Time PCR (RT-qPCR)
The roots of two-month-old R. rugosa seedlings were immersed in a 170 mM NaCl solution, while water was used as the control.Samples were extracted at intervals of 0.5, 1, and 2 h after treatment [60,61].Both NaCl-treated and control samples were subjected to RNA extraction using a kit, and cDNA was reverse-transcribed using the HiScript II Q Select RT SuperMix for qPCR kit (Vazyme, Nanjing, China).The RT-qPCR analysis utilized the SYBR Green qPCR Master Mix kit (Vazyme, Nanjing, China), and the resulting data were evaluated through a relative quantitative method before being visualized using GraphPad Prism 9 (https://www.graphpad.com/(accessed on 1 August 2023)).

Conclusions
This study demonstrates the complex mechanism underlying the biosynthesis and accumulation of triterpenoids in R. rugosa through the integrated analysis of metabolomics and transcriptomics.First, the dynamic accumulation of triterpenoids in different tissues was observed, particularly noting diverse triterpenoid profiles in the roots and fruits.Second, this combined analysis identified the ROSC10 gene encoding OSC, which plays a pivotal role in α-amyrin biosynthesis.The upregulation of RrOSC10 in response to salt stress suggested its involvement in salinity tolerance.This study increased our understanding of the plant's medicinal potential and established the foundation for future research to explore its beneficial properties for different purposes.

Supplementary Materials:
The following supporting information can be downloaded at https://www.mdpi.com/article/10.3390/plants13121600/s1: Figure S1: Mass spectrometry detection of total ion current plots; Figure S2: Sample-to-sample correlation analysis between different tissues; Figure S3: Classification of metabolites and tissue variations.Figure S4: Differential gene and metabolite-related clustering heatmap in petal versus root; Figure S5: Differential gene and metabolite-related clustering heatmap in petal versus fruit; Table S1: Total terpenoids were identified; Table S2: Expression patterns of 29,758 genes.

Figure 2 .
Figure 2. Metabolomic analysis of different tissues.Clustering heatmap of differential metabolites.Colors indicate distinct subclasses and demonstrate similar trends in metabolites.

Figure 2 .
Figure 2. Metabolomic analysis of different tissues.Clustering heatmap of differential metabolites.Colors indicate distinct subclasses and demonstrate similar trends in metabolites.

Figure 3 .
Figure 3. Differentially accumulated metabolites (DAMs) between tissues.(A) Top 10 metabolites in the comparison between petals and roots.(B) Top 10 metabolites in the comparison between petals and fruits.Red represents an increase in metabolite content, and green represents a decrease.

Figure 4 .
Figure 4. Differential expression gene (DEG) analysis between tissues.(A) Volcano plot displa DEGs of petal and fruit tissues.(B) Volcano plot presenting DEGs of petal and root tissues.T axis shows the fold change in gene expression, and the y-axis represents the significance level o DEG.(C) Directed acyclic graph (DAG) portraying gene ontology (GO) enrichment of DEGs petal and fruit tissues.(D) DAG illustrating GO enrichment of DEGs from petal and root tissu

Figure 4 .
Figure 4. Differential expression gene (DEG) analysis between tissues.(A) Volcano plot displaying DEGs of petal and fruit tissues.(B) Volcano plot presenting DEGs of petal and root tissues.The x-axis shows the fold change in gene expression, and the y-axis represents the significance level of the DEG.(C) Directed acyclic graph (DAG) portraying gene ontology (GO) enrichment of DEGs from petal and fruit tissues.(D) DAG illustrating GO enrichment of DEGs from petal and root tissues.

8 Figure 5 .
Figure 5. Changes in gene expression and metabolite accumulation levels of triterpene met pathway.Yellow markings indicate genes, while heatmaps display key metabolites and trans associated with triterpenoid biosynthesis in various tissues.

Figure 5 .
Figure 5. Changes in gene expression and metabolite accumulation levels of triterpene metabolic pathway.Yellow markings indicate genes, while heatmaps display key metabolites and transcripts associated with triterpenoid biosynthesis in various tissues.

Figure 6 .
Figure 6.Characterization of RrOSC10.(A) The standard compound α-amyrin is shown in the MS chromatograms.(B) The major triterpenes extracted from yeast expressed with RrOSC10 shown in the GC-MS chromatograms.(C) Gene expression analysis of RrOSC10 under salt tr ment.*, p < 0.05.

Figure 6 .
Figure 6.Characterization of RrOSC10.(A) The standard compound α-amyrin is shown in the GC-MS chromatograms.(B) The major triterpenes extracted from yeast expressed with RrOSC10 are shown in the GC-MS chromatograms.(C) Gene expression analysis of RrOSC10 under salt treatment.*, p < 0.05.

Table 1 .
Summary of RNA-Seq data of five issues of R. rugosa.

Table 1 .
Summary of RNA-Seq data of five issues of R. rugosa.