ldentification of key genes involved in flavonoid and terpenoid biosynthesis and the pathway of triterpenoid biosynthesis in Passiflora edulis

XU Yi , HUANG Dong-mei, MA Fu-ning , YANG Liu, WU Bin XlNG Wen-ting SUN Pei-guang , CHEN Di XU Bing-qiang SONG Shun #

1 National Key Laboratory for Tropical Crop Breeding/Haikou Experimental Station, Chinese Academy of Tropical Agricultural Sciences/Key Laboratory of Genetic Improvement of Bananas, Sanya Research Institute of Chinese Academy of Tropical Agricultural Sciences, Sanya 572000, P.R.China

2 Hainan Yazhou Bay Seed Laboratory, Sanya 572000, P.R.China

3 Guangxi Crop Genetic Improvement and Biotechnology Laboratory, Guangxi Academy of Agricultural Sciences, Nanning 530007, P.R.China

Abstract Passion fruit (Passiflora edulis Sims) is a vine of the Passiflora genus in the Passifloraceae family.The extracted components include flavonoids and terpenoids, which have good anti-anxiety and anti-inflammatory effects in humans.In this study, we analyzed the transcriptomes of four tissues of the ‘Zixiang’ cultivar using RNA-Seq, which provided a dataset for functional gene mining.The de novo assembly of these reads generated 96 883 unigenes, among which 61 022 unigenes were annotated (62.99% yield).In addition to its edible value, another important application of passion fruit is its medicinal value.The flavonoids and terpenoids are mainly derivatives of luteolin, apigenin, cycloartane triterpenoid saponins and other active substances in leaf extracts.A series of candidate unigenes in the transcriptome data that are potentially involved in the flavonoid and terpenoid synthesis pathways were screened using homologybased BLAST and phylogenetic analysis.The results showed that the biosynthesis of triterpenoids in passion fruit comes from the branches of the mevalonate (MVA) and 2-C-methyl-D-erythritol 4-phosphate/1-deoxy-D-xylulose 5-phosphate (MEP/DOXP) pathways, which is different from the MVA pathway that is used in other fruit trees.Most of the candidate genes were found to be highly expressed in the leaves and/or flowers.Quantitative real-time PCR (qRT-PCR) verification was carried out and confirmed the reliability of the RNA-Seq data.Further amplification and functional analysis of these putative unigenes will provide additional insight into the biosynthesis of flavonoids and terpenoids in passion fruit.

Keywords: passion fruit, RNA-Seq, gene mining, flavonoids, terpenoids

1.lntroduction

Passion fruit (PassifloraedulisSims) is a vine of thePassifloragenus in the Passifloraceae family (Songet al.2022).Most passion fruit species are found in South America, Eastern Asia, Southern Asia and New Guinea (Ulmeret al.2004; Ortizet al.2011), including more than 60 edible species (Costaet al.2012).Today, passion fruit is cultivated outside its natural range for its beautiful flowers, delicious fruit and medicinal value (Martin and Nakasone 1970; Garcia-Ruizet al.2017).Purple passion fruit (P.edulisf.edulis) and its yellow relativeP.edulisf.flavicarpahave been introduced as commercial crops in tropical regions around the world.In China, more than 44 000 ha have been planted, and an annual output of about 590 300 t is produced.

‘Zixiang’, one of the purple passion fruit cultivars selected for this work, accounts for more than 50% of China’s planting area of passion fruit.Passion fruit and its byproducts are rich in various chemicals and phytonutrients, including polyphenols, dietary fiber, pectin, carotenoids and vitamins.The fruit pulp is a source of minerals, naturally rich in calcium, manganese, phosphorus, and potassium, contains 17 essential amino acids, is rich in vitamin C (50 mg 100 mL–1) (Jianget al.2017; Heet al.2020), and can be eaten directly or used in the industrial production of concentrated juice to enhance aromas.The bioactive substances in the seeds include high total phenolics, flavonoids and phenolic acids, which have been found to have anti-aging effects on the skin (Yepeset al.2021), as well as promising potential for protecting the placenta from ZIKV infection (Tanabeet al.2021).Myricetin is an important active component in the floral organs, and it plays a role in the production of nicotine withdrawal agents (Bedellet al.2019).The aqueous extract of passion fruit leaves is rich in bioactive polyphenols with antioxidant and anti-inflammatory properties (Johannyet al.2019), and the root system also contains significant amounts of flavonoids (Soareset al.2005).

Another important application of passion fruit is its medicinal value.It has been used in traditional folk medicines as a remedy for many neurogenic diseases in some American and European countries (Denget al.2010).Its unique extracted ingredients have been reported to exhibit anti-anxiety, anti-inflammatory, and other pharmacological activities.The main compounds reported for this species are flavonoids and terpenoids, which are mainly derivatives of luteolin and apigenin (Senaet al.2009; Wanget al.2022), isoorientin, vicenin-2 (Zucolottoet al.2009) and cycloartane triterpenoid saponins (Yoshikawaet al.2000a, b; Rottaet al.2019) in the leaf and seed extracts (Coletaet al.2006; Gadioliet al.2018; Rottaet al.2020).Passion fruit has been the subject of investigations and utilization for its nutritional and medical components.Products including the fruit juice, seed oil, fermented vinegar, fruit wine and medicine have been produced.Although various functional components have been developed and studied, the related genes and mechanisms still remain unclear.

To date, most molecular studies of passion fruit have focused on the evaluation of genetic diversity through molecular markers and determining the genetic relationships based on the bacterial artificial chromosome (BAC) library (Santoset al.2014), large fragment sequencing (Munhozet al.2018), and the chloroplast genome (Cauz-Santoset al.2017).In addition, RNA sequencing was carried out to identify the genes and signaling pathways involved in cold tolerance (Liuet al.2017), explore the nutrient adaptative mechanism in karst regions (Xuet al.2019), elucidate the mechanisms involved in ripening and rapid fruit senescence (Liet al.2020), and reveal the underlying mechanism of the color formation in passion fruit (Qiuet al.2020).However, there have been no reports on mining the function genes involved in passion fruit’s functional components.In this work, four tissues (leaves, flowers, roots, and seeds) were selected to represent the aboveground and underground parts.The contents of the terpenoids and flavonoids in the above-ground tissues and the expression of the key genes in their metabolic pathways were the main focus.RNA sequencing was performed on these four tissues of passion fruit to obtain the basic biological information for passion fruit and to explore the candidate genes involved in the biosynthesis of the functional components.

2.Materials and methods

2.1.Plant materials and extraction of total RNA

The growth cycle of theP.eduliscultivar ‘Zixiang’ purple passion fruit is 4–6 mon.The material used in this study was planted in the tropical area of Hainan Province, China, and the plants required about four months from planting the seedlings to fruit ripening.The plant material of the cultivar ‘Zixiang’ was used for transcriptome sequencing.Leaves, flowers, and roots were separated from plants after four months of growth, and the seeds were isolated from harvested standard fruits with the pulp removed.All the samples were surface sterilized with 3% H2O2for 10 min and then rinsed five times with distilled water.The leaves ofP.edulisf.edulis,P.edulisf.flavicarpa,P.foetida,P.laurifolia,P.cincinnata,P.alata, andP.setaceawere used for quantitative real-time PCR (qRT-PCR).All the samples were instantly frozen in liquid nitrogen and stored in a ?80°C freezer until further RNA isolation.All plant materials were provided by the Haikou Experimental Station, Chinese Academy of Tropical Agricultural Sciences (CATAS).Three biological replicates were set for the four tissues used for transcriptome sequencing and qRT-PCR, with eight different species of experimental materials for the qRT-PCR.

2.2.RNA-Seq library preparation, sequencing and data analysis

The amount of RNA used for sequencing was 3 μg.The raw data in the Fastq format were processed through inhouse perl scripts.In this step, clean data (clean reads) were obtained by removing reads containing an adapter, reads containing poly-N and low-quality reads from the raw data.At the same time, the GC-content, Q20, Q30 and sequence duplication level of the clean data were calculated.All the downstream analyses were based on clean data with high quality (Abolghasemiet al.2021).Denovoassembly was performed using the short reads assembling program Trinity (Grabherret al.2011), and we finally obtained transcriptome data with high integrity.BUSCO analysis was performed using BUSCO Software (version 5.2.2) to determine the quality of the raw transcriptome data (Beheraet al.2021).

2.3.Gene functional annotation

Gene functions were annotated based on the following databases: NR (NCBI non-redundant protein sequences), Pfam (Protein family, V 31.0), KOG (Clusters of orthologous groups for eukaryotic complete genomes, 2020 KOGs)/COG (Clusters of Orthologous Groups of proteins, 2020 COGs)/eggNOG (Evolutionary Genealogy of Genes: Non-supervised Orthologous Groups Database, V5.0), Swiss-Prot (a manually annotated and reviewed protein sequence database, 2022-05), KEGG (Kyoto Encyclopedia of Genes and Genomes, Release 103.0), and GO (Gene Ontology, 2022-05).

2.4.Differential expression analysis

Differential expression analysis of pairs of two conditions/groups was performed using the DESeq R package (1.10.1).DESeq provides statistical routines for determining differential expression in digital gene expression data.The resultingP-values were adjusted using the Benjamini and Hochberg’s approach (1995) for controlling the false discovery rate.Genes with an adjustedP-value<0.05 found by DESeq were considered to be differentially expressed.

2.5.Homology-based gene discovery and phylo- genetic analysis

The flavonoid biosynthesis pathway (ko00941 and ko00942) and terpenoid biosynthesis pathway (ko00900 and ko00909) in the KEGG Pathway Database (https://www.kegg.jp/kegg/pathway.html) were used as the reference metabolic pathways.Blastn was used to mine the candidate genes for the biosynthetic pathways of the functional components.MEGA X was applied to perform the phylogenetic analysis of the nucleotide sequences of the target genes using the Maximum Likelihood.

2.6.RNA-Seq reliability verification

To assess the accuracy of the sequencing data, eight differentially expressed genes (DEGs) were selected with an elongation factor 1-alpha (EF-1α) gene as the internal control (Wuet al.2020) to carry out qRT-PCR.The eight DEGs included four genes in the flavonoid biosynthesis pathway, i.e., CHS (chalcone synthase, c153569.graph_c0), CHI (chalcone isomerase, c154251.graph_c0), F3H (flavanone-3β-hydroxylase, c169417.graph_c0), and F3′5′H (flavonoid-3′,5′-hydroxylase, c193589.graph_c0), and four genes in terpenoid biosynthesis pathway, i.e., HMGR (hydroxymethylglutaryl-CoA reductase, c189795.graph_c0), DXR (1-Deoxy-D-xylulose-5-phosphate reductoisomerase, c188494.graph_c0), HDS ((E)-4-hydroxy-3-methylbut-2-enyl-diphosphate synthase, c181299.graph_c0), and SM (squalene monoxygenase, c193220.graph_c2).Gene-specific primers (Appendix A) were designed and analyzed using the online primer tool (http://www.primer3plus.com/primer3web/, version 4.0.0).The RNA from four tissue samples and eight leaf samples from the eight species ofPassifloragenus (P.edulisf.edulis,P.edulisf.flavicarpa,P.foetida,P.laurifolia,P.cincinnata,P.alata, andP.setacea) were used to reverse transcribe RNA into cDNA using the RT Reagent Kit (Thermo Fisher Scientific, Waltham, USA).The reaction solution of qRT-PCR was configured by the RT-PCR reagent (No.RR055A, TaKaRa, Beijing, China).The reaction equipment was a Roche Light Cycler?96 Real-Time PCR System (Roche, Basel, Switzerland).The relative gene expression values were analyzed using the 2–??Ctformula, and three biological replicates were used in this experiment.Data processing of the eight gene expression levels included the average calculation, and also created the differential expression histograms and legends.

3.Results

3.1.Sequencing, de novo assembly and functional annotation

The organization of the RNA library is shown in Fig.1-A with three biological repeats, and the sequence raw data of all expressed genes are shown in Appendix B.The summary of RNA-Seq data and read mapping are shown in Appendix C, with a Q30 base percentage of 88.56% and a mapped ratio of more than 75%.In total, 96 883 unigenes were obtained; the N50 of the unigenes was 2 000 nt, of which 24 801 unigenes were more than 1 kb in length (Appendix D).For the functional annotation of the unigenes, including comparisons with the NR, Swiss prot, KEGG, COG, KOG, GO and Pfam databases, a total of 61 022 unigene annotation results were obtained, and 22 947 were annotated with lengths ≥1 000 bp (Appendix E).The BUSCO results were 61 and 60.7% for eukaryotic dicotyledonous and all plant homologous pools, respectively, and 2.5 and 6.7% for the fragmented BUSCOs, respectively (Appendix F).The transcriptome data were qualified and considered to be appropriate for further analysis.

3.2.DEGs among the different tissues

The reliability of detections among the biological replicates was calculated based on the Pearson correlation coefficients (r) of the sequenced samples; most of the gene expression trends of the samples were relatively close, and the repeatability correlation was high (Appendix G).Gene expression was identified by the threshold of FPKM≥1.The number of expressed genes in the roots was 30 975, much higher than in the other three tissues (Fig.1-B).More than 73.4% of the genes had FPKM<25, while more than 93.6% had FPKM<100.In the root tissue, 68.6% of genes had FPKM values of 1≤FPKM<10, which was a higher percentage than in the other three tissues (each about 50%).Furthermore, 11 886 genes were expressed in all four tissues.In total, 993 (6.1%), 1 433 (8.3%), 15 771 (50.9%), and 1 011 (6.8%) were specifically expressed in the leaves, flowers, roots, and seeds, respectively.More than half of the genes expressed in the roots were tissue-specific genes, which was much higher than the proportions in the other tissues (Fig.2-C).

Fig.1 Transcript abundance measurements of four tissues.A, tissue samples of Passiflora edulis for RNA-seq, including leaf, flower, root and seed.B, number of FPKM statistics for four tissues.The number of total considered expressed genes (FPKM≥1) for each tissue is presented in brackets.C, number of DEGs in the pair-wise comparisons.D, Venn diagram of the expressed genes (FPKM≥1) for each tissue.

Pairwise comparison showed that the differences in the numbers of genes between the three above-ground parts (leaf, flower, seed) were relatively small, while the difference between the number of genes from the under-ground part (root) and the above-ground parts was relatively large.Moreover, since more than half of the genes expressed in the root were tissue-specific genes, the number of upregulated genes was 4–9 times higher than that of the downregulated genes compared with the aboveground parts.In addition, compared with the flowers and leaves, the number of the upregulated genes in the seeds was more than twice that of the downregulated genes, indicating that the differences in tissue site resulted in a significant differential distribution of gene expression (Fig.2-D; Appendix H).

3.3.Gene functional enrichment analysis of the DEGs in different tissues

GO classification of the DEGs and all genes was carried out.In the biological process (BP) GO terms, the most enriched terms included metallic process (GO: 0008152), cellular process (GO: 0009987), and single-organism process (GO: 0044699).Interestingly, the terpenoid biosynthetic process (GO: 0016114) and terpenoid metabolic process (GO: 0006721) were enriched with 136 and 150 genes, while 58 and 62 genes were significantly differentially expressed in the leafvs.the seed, respectively.Moreover, the flavonoid biosynthetic process (GO: 0009813) and flavonoid metabolic process (GO: 0009812) were enriched with 71 and 81 genes, while 22 and 25 genes were significantly differentially expressed in the leafvs.the seed, respectively.The cellular component (CC) GO terms were mainly enriched in the cell (GO: 0005623), cell parts (GO: 0044464) and organelle (GO: 0043226).The molecular function (MF) GO terms were mainly enriched in catalytic activity (GO: 0003824) and binding activity (GO: 0005488) (Fig.2-A; Appendix I).

Fig.2 Gene functional enrichment analysis of differentially expressed genes (DEGs) in the different tissues.A, the top enriched GO terms in biological process, cellular component, and molecular function of the DEGs and all genes in leaf vs.seed.B, classification of the top enriched KEGG pathways of DEGs in leaf vs.seed.C, KEGG pathway enrichment factor of the DEGs in leaf vs. seed.

We focused on the “metabolism” class in the KEGG pathway enrichment analysis.The main enriched metabolic pathways included the basic metabolism of biomacromolecules, such as “starch and sucrose metabolism”, “carbon metabolism”, “biosynthesis of amino acids”, “glycolysis/gluconeogenesis”, and “fatty acid metabolism”, which indicated the differences in the basic metabolism in the tissues.In pairwise comparisons, leafvs.flower DEGs showed significant enrichment in “pentose and glucuronate interconversions”, “fatty acid degradation”, and “phenylpropanoid biosynthesis”.The leaf and root DEGs showed enrichment in “toxidation”, “purine metabolism”, “cysteine and methionine metabolism”, and “glutathione metabolism”.The leaf and seed DEGs showed significant enrichment in “phenylpropanoid biosynthesis”, “amino sugar and nucleotide sugar metabolism”, “fatty acid metabolism”, and “alpha-linolenic acid metabolism” (Fig.2-B).The flower and root DEGs were enriched in “toxidation”, “pentose and glucuronate interconversions”, “cysteine and methionine metabolism” and “purine metabolism”.The flower and seed DEGs showed significant enrichment in “amino sugar and nucleotide sugar metabolism”, “phenylpropanoid biosynthesis”, “pentose and glucuronate interconversions”, and “purine metabolism”; while “oxidative phosphorylation”, “cysteine and methionine metabolism”, “purine metabolism” and “amino sugar and nucleotide sugar metabolism” were significantly enriched in the root and seed DEGs.Interestingly, “terpenoid backbone biosynthesis” and “flavonoid biosynthesis” were enriched with 117 and 37 genes in all samples, while 21 and 11 genes were significantly enriched in the leafvs.the seed, respectively.The KEGG pathway enrichment factor results showed that the “flavonoid biosynthesis” and “monoterpenoid biosynthesis” pathways were among the significant enrichment pathways (Fig.2-C; Appendices J and K).

3.4.Candidate gene mining of the functional component biosynthetic pathways

Terpenoids and flavonoids are functional components with high contents in passion fruit.To understand the expression of these functional component-related genes in the different tissues of passion fruit, expression heatmaps of the terpenoid and flavonoid biosynthetic pathways were drawn (Fig.3).The results show that most of the DEGs were highly expressed in the leaves and flowers.

In passion fruit, apigenin, kaempferol and their derivatives are the main functional and active flavonoid substances.In the flavonoid biosynthesis pathway, we found that the CYP73A (trans-cinnamate 4-monooxygenase), CHS, CHI, F3H, and F3′5′H gene families were the important gene families that were found to have tissue-specific expression (Fig.3-A).Two of the CHS and all of the CHI, F3H, and F3′5′H genes showed high expression in flowers, followed by the leaf samples, suggesting that these genes play an important role in the biosynthesis of functional and active flavonoids.In addition, the expression of the gene CCoAMT (caffeoyl-CoAO-methyltransferase), which is related to the synthesis of homoerodictyol, was also found, indicating that this secondary metabolite and its derivatives have biological activities.

In the terpenoid biosynthesis pathway, we identified seven important gene families, including the HMGS (hydroxymethylglutaryl-CoA synthase), HMGR, DXS (1-deoxy-D-xylulose-5-phosphate synthase), DXR, MDS (2-C-methyl-D-erythritol 2,4-cyclodiphosphate synthase), HDS, and SM gene families (Fig.3-B).Among these seven, the SM gene family had six members, suggesting that this gene family may have undergone gene amplification to enhance the plant’s ability to produce triterpenoids.The existence of the HMG, HMGR, DXS, DXR, MDS, and HDS gene families also indicated that both the mevalonate pathway (MVA) and the MEP/DOXP pathway are involved in passion fruit terpenoid biosynthesis.

3.5.Confirmation of the DEGs by qRT-PCR

According to the sequencing data, four genes related to the biosynthesis of terpenoids (HMGR, DXR, HDS, and SM) and four genes related to the biosynthesis of flavonoids (CHS, CHI, F3H, and F3′5′H) were selected for qRT-PCR analysis, in the leaf, flower, root and seed ofP.edulis, respectively.The expression pattern analysis in the four tissues showed that the qRT-PCR results were consistent with the transcriptome data (Fig.4-A; Appendix L).The four genes related to terpenoid biosynthesis were expressed at low levels or almost not at all in roots, while the CHS, CHI, and F3H were significantly highly expressed in the flowers, F3′5′H was significantly highly expressed in the leaves, and CHI and F3′5′H were highly expressed in the seeds.The four genes related to flavonoid biosynthesis were expressed at low levels or almost not at all in the seeds, but they were significantly highly expressed in the leaves and flowers, and the DXR and HDS were highly expressed in the roots.Therefore, these eight genes showed significant tissue-specific expression patterns.

The relative gene expression levels of the eight genes related to flavonoid and terpenoid biosynthesis in the leaf samples of eight species of thePassifloragenus showed variations (Fig.4-B; Appendix L).Among them, four species (P.edulisf.edulis,P.alata,P.edulisf.flavicarpaandP.setacea) attracted our attention, because the relative expression levels of the candidate genes related to flavonoid and terpenoid biosynthesis in the leaves of these four species were higher than those in the leaves of the other species.CHS is a key upstream gene in the flavonoid biosynthesis pathway (Fig.3-A), and it was highly expressed inP.edulisf.edulis,P.alataandP.edulisf.flavicarpa.F3H is an important gene in kaempferol biosynthesis, and it had high expression levels in all eight species.CHS and F3′5′H are related to the luteolin biosynthesis, and they were highly expressed inP.setacea.In the terpenoid biosynthesis pathway, SM is a key gene in triterpenoids biosynthesis (Fig.3-B), and its expression levels inP.edulisf.edulis,P.alata, andP.edulisf.flavicarpawere higher than those of the other species.HMGRis a key gene in the upstream of terpenoid biosynthesis and it had high expression levels inP.alata,P.edulisf.flavicarpaandP.setacea.The speciesP.alatais a medicinal plant, and all four genes (HMGR, DXR, HDS, and SM) in the terpenoid biosynthesis pathway were highly expressed inP.alata.

Fig.3 Expression heatmaps of the flavonoid (A) and terpenoid (B) biosynthetic pathways in passion fruit.An overview of the terpenoid and flavonoid biosynthetic pathways showed the gene number expansions in each step and their expression profiling in different samples of passion fruit.The heatmaps were drawn using log2-based FPKM-change fold values.Red is high, and blue is low.

Fig.4 Relative gene expression levels of eight genes related to the flavonoid and terpenoid biosynthetic pathways.A, relative gene expression in the four tissues of Passiflora edulis.B, relative gene expression in the leaf samples of eight species in the Passiflora genus.CHS, chalcone synthase; CHI, chalcone isomerase; F3H, flavanone-3β-hydroxylase; F3′5′H, flavonoid-3′,5′-hydroxylase; HMGR, hydroxymethylglutaryl-CoA reductase; DXR,1-Deoxy-D-xylulose-5-phosphatereductoisomerase; HDS, (E)-4-hydroxy-3-methylbut-2-enyl-diphosphate synthase; SM, squalene monoxygenase, c193220.graph_c2.

4.Discussion

Homologous gene alignment plays an important role in RNA-Seq functional gene annotation.Among the NR homologous species distribution,Populustrichocarpa(8%),Populuseuphratica(6.73%),Jatrophacurcas(6.43%) andRicinus communis(4.79%) are the four species with the highest annotation proportions (Fig.5-A).The phylogenetic analysis based on chloroplasts (Arayaet al.2017) showed that Passifloraceae is closely related to Salicaceae and Euphorbiaceae.This indicates that our results are consistent with previous research results.In addition, we chose squalene monoxygenase (SM, c193220.graph_c2) for the phylogenetic tree analysis, which is a significantly enriched key gene in the sesquiterpenoid and triterpenoid biosynthesis pathway (ko00909).The results also showed thatP.edulisis closely related toP.euphratica,P.tricocarpaandJ.curcas(Fig.5-B), which is consistent with the results of the previous research (Santoset al.2014; Munhozet al.2018), indicating that the annotation results of the unigene in this RNA-Seq data had high reliability.

Fig.5 Homologous analysis of Passiflora edulis.A, NR (NCBI non-redundant protein sequences) homologous species distribution.B, phylogenetic tree of squalene monooxygenase (c193220.graph_c2).The species and the protein ID for homology analysis are as follows: Populus euphratica XM_011033026, Populus trichocarpa XM_024604906, Jatropha curcas XM_012224069, Carica papaya XM_022034781, Citrus clementina XM_006426079, Pistacia vera XM_031393128, Gossypium arboreum XM_017747780, Gossypium hirsutum XM_016831745, Juglans regia XM_018960901, Quercus lobata XM_031118887, Quercus suber XM_024021389, Pyrusx bretschneideri XM_018648783, Malus domestica XM_029092559, Camellia sinensis XM_028243540, Medicago truncatula XM_013606367, Abrus precatorius XM_027498678, Ziziphus jujuba XM_016029604, Hevea brasiliensis XM_021831604, Ricinus communis XM_002509997, Gossypium raimondii XM_012589981, Durio zibethinus XM_022866698, Theobroma cacao XM_007018355.

Along with the edible fruit and high ornamental value of the flowers, many plants in thePassifloragenus have a long history of use in traditional folk medicines as a remedy for many neurogenic diseases (Denget al.2010).The extracts from the leaves, fruit pulp (Rottaet al.2019), and seeds of passion fruit have been reported to have sedative, antioxidant, anti-inflammatory, anxiolytic, and anti-depressant effects (Abourashedet al.2002; Xiaoet al.2016; Gadioliet al.2018).More than 110 phytochemical constituents have been found and identified from the different plant parts ofP.edulis, including phenols, flavonoid glycosides, alkaloids, triterpenoids and cyanogens, among which the flavonoids and triterpenoids hold the largest share (Coletaet al.2006; Heet al.2020).

Among the phytochemical constituents, flavonoids are a kind of secondary metabolites and important active substances in passion fruit extracts (Guoet al.2022).So far, 33 flavonoids have been identified in various parts ofP.edulis(Costaet al.2012; Xuet al.2013), and some of them present anxiolytic, antioxidant, anticancer, and antiinflammatory activities (Petryet al.2001).Vitexin is one of the active-ingredient flavonoids, and it has potential as an agent for nicotine cessation (Bedellet al.2019).

In this study, we found that CHS, CHI, F3H, and F3′5′H in the flavonoid biosynthesis pathway were expressed tissue-specifically, and most of the expression levels were high in the leaves and flowers.The high expression of these genes in the leaves indicates that the accumulation of these active substances is higher in the leaves than in other tissues.In the literature, most of the materials selected for evaluating the medicinal value of passion fruit were leaves, which are considered to be the main source of the functional substances, such as flavonoids and terpenes.Passion fruit is a tropical vine, and the plant grows rapidly.Therefore, it is a rich source of leaves, which can meet the requirements of functional substance extraction and convenience for processing and industrial utilization.Kaempferol is a flavonoid found in many edible plants and in plant products commonly used in traditional medicine; and numerous preclinical studies have shown that kaempferol and some glycosides of kaempferol have a wide range of pharmacological activities (Calderon-Montanoet al.2011).In this work, F3H, which is an important gene in kaempferol biosynthesis (Hammerbacheret al.2019), had high expression levels in all eight species; and this result indicates that kaempferol and/or its derivatives might be present in many species ofPassiflora.In addition, most of the flowers of passion fruit are bright in color, with purple and white para corollas.These candidate genes are not only involved in the synthesis of functional substances, but also in the synthesis of anthocyanins in the plant petals, which explains the high expression levels of some candidate genes in the flowers.

Terpenes are also an important functional compound in the passion fruit plant, especially triterpenoid saponins (Augustinet al.2011).Triterpenoids are compounds with a carbon skeleton based on six isoprene units, and they are derived biosynthetically from the acyclic C30 hydrocarbon, squalene (Ludwiczuket al.2017).The chemical composition of triterpene saponins includes the triterpene part of sapogenin and one or more connected sugar moieties, such as glycosyl, xylosyl, and glucuronic acid (Zhenget al.2014).Some triterpenes (cyclopassifloic acids A–H and 31-methoxyl-passifloic acid), and their related saponins (cyclopassiflosides I–XV) have been isolated from the leaves and stems ofP.edulis, and some of these cycloartane triterpenoids were reported to have antidepressant-like effects, so they can be used for the treatment of neurodegenerative disease (Bombardelliet al.1975; Yoshikawaet al.2000a, b; Xuet al.2016; Rottaet al.2019).A candidate gene in the triterpenoid biosynthesis pathway, squalene monoxygenase (SM, c193220.graph_c2), was found to be significantly enriched in this study, with high expression levels in the leaves and flowers (Fig.4-B).Squalene monoxygenase (SM, also known as SQLE or squalene epoxidase) oxidizes squalene to squalene-2,3-epoxide; and it is the intermediate branch point leading alternately to the sterol pathway or to triterpene saponins (Thimmappaet al.2014; da Silva Magedanset al.2020).SM determines the biosynthetic efficiency of squalene epoxide and affects the pharmacological activities.In passion fruit, the genes related to the synthesis of terpenoids are enriched, and the expression levels in thePassifloraspecies with high medicinal value (such asP.alata,P.edulisf.flavicarpa, andP.setacea) are higher than in the other species.P.alatawas officially recognized as a medicinal plant in the Brazilian Pharmacopoeia (Anvisa 2010), with several terpenoids, especially steroidal and triterpenoidal glucosides (Reginattoet al.2001); and the extract fromP.alataleaves has been considered as a substance for the complementary therapy of cancer patients (Ozarowskiet al.2018).At present, the economic value of the commercial breeding of passion fruit is still uncertain, and the evaluation of populations is needed to improve the functional component content, the yield, and the fruit quality traits (Chavarríaet al.2020).In this study, four genes related to terpenoid synthesis (HMGR, DXR, HDS, and SM) were all highly expressed inP.alata, which might be the reason for the strong aroma of this medicinal plant.Terpenoid synthase (TPS) is a key enzyme in terpene metabolism; and triterpenoids are terpenes composed of 30 carbon atoms.The TPS enzyme belongs to the upstream of the triterpenoid substance synthesis pathway, and the TPS anabolic pathway can synthesize monoterpenoids, sesquiterpenes, diterpenoids, triterpenoid, tetraterpene, and dozens of other terpenoids under the action of the TPS enzyme.Future research should focus on the involvement of SM in triterpenoid synthesis.

Overall, the transcriptome assembly was of high quality, and the genome coverage was sufficient for global analyses.Therefore, these data can be utilized for future gene expression analysis, primer design, marker assisted selection, source identification, and functional genomics inP.edulis.Further amplification and functional analysis of these putative unigenes will provide additional insight into the biosynthesis of flavonoids and terpenoids.

5.Conclusion

In addition to its edible value, passion fruit has a unique medicinal value due to its flavonoids and terpenoids.In this study, we analyzed the transcriptomes of four tissues of passion fruit ‘Zixiang’ using RNA-Seq, and the differential expression levels of different varieties and tissues and organs were obtained.Interestingly, unlike in other fruit trees, the MVA and MEP/DOXP metabolic pathways were jointly involved in the triterpenoid biosynthesis in passion fruit, which is one of the important findings of this study.Using homology-based BLAST and phylogenetic analysis, a series of candidate single genes in the transcriptome data that may be involved in the synthesis of flavonoids and terpenoids were screened, and found to be highly expressed in the leaves and/or the flowers.The functional analysis and verification of these hypothetical single genes will provide further insights into the biosynthetic pathways of flavonoids and terpenoids in passion fruit.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (32260737), the Project of Sanya Yazhou Bay Science and Technology City, China (SCKJJYRC-2022-84 and SCKJ-JYRC-2022-93), and the Hainan Provincial Natural Science Foundation of China (320QN305, 321MS091 and 320RC686).

Declaration of competing interest

The authors declare that they have no conflict of interest.

Appendicesassociated with this paper are available on https://doi.org/10.1016/j.jia.2023.03.005