Transcriptomic and metabolomic study of the biosynthetic pathways of bioactive components in Amomum tsaoko fruits

Abstract

Amomum tsaoko is a significant medicinal and edible plant with documented efficacy in the treatment of various diseases. Additionally, it is a crucial food additive and spice. 1,8-cineole and curcumin are the main bioactive compounds of A. tsaoko, and research on these compounds has mainly focused on their chemical composition and pharmacological activity, with relatively less exploration of synthetic pathways and identification of key genes. This study employed transcriptome sequencing and metabolomic analysis of A. tsaoko at five different developmental stages (May fruit – September fruit) to assess the accumulation patterns of terpenoid and curcuminoid compounds and to explore the key genes and transcription factors (TFs) involved in their synthesis pathways. The results showed that three genes encoding 1-deoxy-D-xylulose-5-phosphate synthase (DXS), hydroxymethylglutaryl-CoA synthase (HMGCS) and phosphomevalonate kinase (mvaK2) and TFs such as AP2-ERF, bHLH, WRKY were screened for involvement in terpenoid biosynthesis. In addition, three genes encoding trans-cinnamate 4-monooxygenase (C4H), curcumin synthase (CURS) and TFs such as MYB, bHLH, bZIP were screened for involvement in curcuminoid biosynthesis. This study provides a theoretical foundation for further research into the biosynthesis of active components in A. tsaoko, establishing a basis for in-depth investigations into the mechanisms underlying its medicinal quality formation. Additionally, it offers guidance for the utilisation of its aromatic components and natural pigments.

Introduction

Amomum tsaoko Crevost et Lemarie is a perennial herbaceous plant belonging to the genus Amomum, family Zingiberaceae. It has been documented in various traditional Chinese medicine texts and has a cultivation history spanning several thousand years in China, with primary distribution in the Yunnan, Guangxi, and Guizhou provinces. It is noteworthy that Yunnan accounts for 90% of the total production in the country [1]. The dried mature fruit is known as Tsaoko Fructus (Caoguo in Chinese) and is included in the 2020 edition of the Pharmacopoeia of the People’s Republic of China [2]. Traditionally, it has been used to alleviate abdominal pain, diarrhea, hemorrhoids, throat infections, and malaria [3]. Modern pharmacological studies have revealed its diverse biological activities, including antimicrobial, antioxidant, anti-inflammatory, and antitumor effects [4, 5]. Furthermore, some formulations containing A. tsaoko have been developed for the treatment of Coronavirus Disease 2019 (COVID-19) [6] and Severe Acute Respiratory Syndrome (SARS) [7].

As a commercially important medicinal and edible plant, A. tsaoko is rich in active ingredients, including terpenoids, flavonoids, diarylheptanoids, and organic acids [8]. Among these, terpenoids, as the most structurally diverse class of plant natural products, play a pivotal role in plant growth and development [9], plant adaptation to environmental stress [10], and the treatment of human diseases [11]. Terpenoids compounds are produced from two structural units: isopentenyl pyrophosphate (IPP) and its double bond isomer dimethylallyl pyrophosphate (DMAPP). The synthesis of terpenoids is a complex process involving a series of key enzymes, including HMGCS, DXS, geranyl diphosphate synthase (GPPS), and terpene synthases (TPSs). HMGCS and DXS are both key rate-limiting enzymes, and GPPS catalyzes the formation of geranyl diphosphate (GPP) from IPP and DMAPP molecules, providing a carbon skeleton for monoterpenes [12]. TPSs can catalyze the formation of monoterpenes, sesquiterpenes, and diterpenes from GPP, farnesyl diphosphate (FPP), and geranylgeranyl diphosphate (GGPP). Based on the different synthetic products, TPSs can be divided into monoterpene synthases, sesquiterpene synthases, and diterpene synthases [11]. The type and quantity of terpenoid products are determined by terpenoid biosynthesis, which are encoded by genes that are precisely regulated [13].

Among the structurally and typologically diverse compounds of A. tsaoko, curcumin is also a well-known active molecule among the diarylheptanoids [14]. Its pharmacological activities, including anti-inflammatory and antioxidant properties, have been the subject of considerable research interest within the medical community [15, 16]. Furthermore, curcumin is employed as an anticancer and antimutagenic agent, and has been categorised as a third-generation chemopreventive drug for cancer by the National Cancer Institute [17, 18]. Moreover, as one of the natural pigments with a high safety, it has been used to enhance the color and flavor of foods [19]. The biosynthesis of curcumin is a complex multistep process, and its biosynthesis shares a common upstream synthetic pathway with that of phenylpropanes and flavonoids. It starts from phenylalanine and is catalyzed by phenylalanine lyase (PAL), 4-coumaroyl-CoA ligase (4CL), and diketide-CoA synthase (DCS), curcumin synthase (CURS) [18]. Among them, DCS catalyzes the formation of feruloyl diketone CoA from feruloyl CoA and malonyl CoA. CURS catalyzes the formation of curcumin from feruloyl CoA and feruloyl diketone CoA [20].

Transcriptome sequencing is the high-throughput sequencing of mRNA from a certain species, and the sequencing results can reflect the gene expression under specific conditions and time points [21]. Metabolomics refers to the qualitative and quantitative analysis of metabolites in organisms under specific conditions [22]. The correlation analysis between transcriptomics and metabolomics can explore the relationship between differential genes and differential metabolites, and conduct corresponding biological function analysis of both. It can deeply explore metabolic pathways and identify key genes and metabolites, and systematically analyze the regulatory mechanisms of plant growth and development [23].

As a significant medicinal and edible plant, A. tsaoko holds considerable commercial value in the seasoning market due to its distinctive aroma and flavour, extensive culinary applications, medicinal properties, high nutrient content, and extensive pharmacological activities. Previous studies have primarily examined the chemical constituents and pharmacological activities of A. tsaoko [3]. With the advent of modern molecular biology techniques and the omics era, research on the biosynthesis of terpenoids and curcuminoids has progressed to the genetic level. Integrating biosynthetic pathways with molecular regulatory mechanisms through systematic studies provides a more profound and comprehensive insight into the biosynthesis of terpenoids and curcuminoids. This study used transcriptome sequencing and metabolomics analysis to investigate the accumulation patterns and regulatory networks of terpenoids and curcuminoids across five developmental stages of A. tsaoko fruit. The research findings will facilitate further studies, such as breeding engineering plants with high levels of bioactive components through gene editing, while also laying the foundation for the selection of excellent varieties of A. tsaoko and in-depth analysis of the mechanism of its pharmacological quality formation.

Results

Metabolite profiling of A. tsaoko fruit at different developmental stages

The fruit gradually swells in the first one or two months and remains the same size in the last three months. The color of the fruit is mainly reddish brown, and the changes throughout the process are not significant (Fig. 1A). To understand the molecular mechanisms underlying the differences in metabolites during the development of A. tsaoko fruit, this study conducted metabolomic analysis on fruit at five different developmental stages. Based on the UPLC-MS/MS detection platform and a self-built database, a total of 1,879 metabolites were detected and categorized into 13 classes. Phenolic acids were the most abundant (323 metabolites), followed by flavonoids (250 metabolites), amino acids and derivatives (214 metabolites), terpenoids (200 metabolites), among other classes (Fig. 1B). The OPLS-DA showed that the samples within each group were clustered together across the five developmental stages, while the samples from different groups were distinctly separated (Fig. 1C). This indicates that the metabolite profiles within each developmental stage of A. tsaoko fruit are highly reproducible, while significant differences are evident between developmental stages.

Fig. 1

Metabolomics analysis of A. tsaoko. A The appearance of A. tsaoko at different stages during fruit development. B Classification of metabolites in A. tsaoko. C OPLS-DA score chart. D The number of total, upregulated, and downregulated DAMs in the comparison group of adjacent stages. E Dynamics of metabolite during development stages in A. tsaoko

A total of 1,432 differentially accumulated metabolites (DAMs) were identified through pairwise comparisons of different stages of fruit development. It is noteworthy that among the comparison groups of adjacent stages, the July fruit (JF7) vs. June fruit (JF) group exhibited the highest number of DAMs (Fig. 1D), indicating that the metabolites from JF to JF7 are more active. To gain further insight into the accumulation patterns of different metabolites during fruit development, a K-means clustering analysis was conducted on all the DAMs (Fig. 1E, Tab. S2). The DAMs were classified into eight principal categories. The largest number of metabolites were classified within Class 7, which included 300 metabolites whose accumulation exhibited a gradual decrease throughout the course of fruit development. The majority of the metabolites in this class were lipids, flavonoids, organic acids, and amino acids. Subsequently, Class 3 comprised a total of 243 DAMs. These metabolites initially accumulated at higher levels during the May fruit (MF) and JF stages and then gradually decreased as the fruit develops. They included phenolic acids, flavonoids, amino acids, and other compounds. Furthermore, Class 8 contained a total of 183 DAMs, that showed a gradual increase in accumulation throughout the course of fruit development. These included alkaloids, lipids, and phenolic acids, among others. The accumulation of DAMs in Class 6 also showed a gradual increase from the MF to the August fruit (AF) stage, after which it stabilized at the September fruit (SF) stage. This class included the majority of terpenoids, phenolic acids, and some flavonoids. Moreover, some terpenoid and phenolic acid metabolites were classified as belonging to Classes 1 and 4, which demonstrated the greatest accumulation at the JF7 stage.

KEGG enrichment analysis of DAMs in the adjacent stage comparisons showed that the first two groups were enriched in more pathways than the last two groups. Furthermore, the DAMs of the first three groups were predominantly enriched in pathways related to terpenoid and curcuminoid biosynthesis, such as stilbenoid, diarylheptanoid and gingerol biosynthesis, phenylpropanoid biosynthesis and diterpenoid biosynthesis, among others (Fig. S1).

RNA-seq analysis and DEGs identification

A high-throughput sequencing platform was employed to perform transcriptome sequencing of A. tsaoko at five different stages, resulting in the construction of 15 cDNA libraries. Following the removal of adapters and low-quality reads, the clean read count per sample ranged from 41.3 to 43.39 million, a total yield of 96.07 Gb of high-quality bases was obtained (Tab. S3). For each sample, the number of clean bases exceeded 6 Gb, with a Q30 base percentage above 90%. The ratio of reads uniquely mapped to the A. tsaoko genome (unpublished data) relative to the total read count ranged from 81.29% to 94.18%. Gene expression levels were determined using fragments per kilobase of exon per million fragments mapped (FPKM) values. To investigate the changes in transcriptional levels during A. tsaoko fruit development, an analysis of DEGs was performed utilising RNA-Seq data from five developmental stages. A total of 14,205 DEGs were identified in this study. Among the pairwise comparisons of adjacent stages, the JF7 vs. JF group showed the highest number of DEGs (Fig. 2A), a finding that is consistent with the results of the metabolome analysis.

Fig. 2

Transcriptomics analysis of A. tsaoko. A The number of total, upregulated, and downregulated DEGs in the comparison group of adjacent stages. B KEGG enrichment of DEGs in each comparison group. C Dendrogram showing co-expression modules (clusters) identified by WGCNA across fruit developmental and ripening stages of A. tsaoko. D Heat map showing module-metabolite correlations of A. tsaoko. Each column corresponds to a module indicated by different colors. The gray module contained genes that could not be classified into any other module. Each row corresponds to a metabolite. Red color indicates a positive correlation between the cluster and the tissue. Blue color indicates a negative correlation

GO enrichment analysis was used to identify and functionally classify DEGs from four adjacent stage comparison groups. The enriched GO terms in the JF vs. MF group and JF7 vs. JF group were largely similar and primarily included “heme binding” and “extracellular region”. Notably, the first three adjacent stage comparisons were enriched in “diterpenoid biosynthetic process” and “terpene synthase activity”, which are associated with terpenoid synthesis. In addition, the SF vs. AF group was enriched in two GO terms related to abscisic acid: “abscisic acid-activated signaling pathway” and “abscisic acid binding” (Fig. S2). The KEGG enrichment analysis (Fig. 2B) showed that the DEGs in the four pairwise comparisons between adjacent stages were consistently enriched in pathways related to terpenoid and curcuminoid synthesis, such as “terpenoid backbone biosynthesis”, “diterpenoid biosynthesis”, “stilbenoid, diarylheptanoid, and gingerol biosynthesis”, and “phenylpropanoid biosynthesis”, which mirrored the metabolomic results.

Dynamic changes in terpenoid biosynthesis pathways throughout the development of A. tsaoko fruit

Terpenoid compounds are the primary constituents of the volatile oil extracted from A. tsaoko fruits. To further understand the changes in terpenoid biosynthesis pathways during fruit development, we analyzed the DAMs and DEGs related to terpenoid backbone biosynthesis, monoterpenoid biosynthesis, and diterpenoid biosynthesis, as annotated by KEGG. Based on this analysis, detailed pathway diagrams were generated. A total of six metabolites and 39 genes were annotated to various terpenoid biosynthesis pathways (Fig. 3A, Tab. S4). The accumulation of 1,8-cineole, abietic acid, and isopimaric acid gradually increased during fruit development and ripening, whereas the accumulation of geniposidic acid steadily decreased. Additionally, ferruginol showed the lowest accumulation during the MF stage, peaked during the JF stage, and then gradually decreased. The accumulation of pisiferic acid reached its highest level during the JF7 stage. Among the thirty-nine genes, half showed their highest expression levels during the SF stage, including genes encoding key enzymes such as HMGCS, HMGCR, DXR, and ispH. Some genes showed peak expression during the MF stage, including those encoding enzymes such as DXS and FDPS. Furthermore, a small number of genes showed peak expression during the JF stage, including those encoding enzymes such as DXS and idi. Only three genes showed their highest expression levels during the JF7 stage, consisting of two genes encoding the DXS enzyme and one gene encoding the ACAT enzyme.

Fig. 3

Analysis of the metabolic pathways of terpenoid biosynthesis and differences in A. tsaoko (A) Putative DEGs in the terpenoid biosynthesis pathway and their expression levels. Heat maps represent the expression of genes at different stages. Genes in red boxes represent that they have been screened in the WGCNA and are highly correlated with terpenoid biosynthesis. B Terpenoid putative transcriptional regulatory network in A. tsaoko. Blue prisms represent TFs, orange triangles represent genes, and purple ellipses represent metabolites

To further investigate the molecular mechanisms underlying terpenoid biosynthesis during A. tsaoko fruit development, we performed WGCNA on the 14,025 DEGs derived from pairwise comparisons of various developmental stages. A total of 21 modules showed similar gene expression patterns (Fig. 2C, Tab. S5). To generate a regulatory network related to terpenoid biosynthesis, we initially selected modules that were highly correlated with terpenoid compounds and further screened for genes and TFs associated with terpenoid biosynthesis pathways within these modules. By integrating expression correlations, we constructed a transcriptional regulatory network that is specifically related to terpenoid biosynthesis in A. tsaoko.

The module-metabolite correlation plot indicated that the accumulation of 1,8-cineole was highly positively correlated with the MEtan and MEmidnightblue modules, while the accumulation of abietic acid was highly positively correlated with the MEgreenyellow module. Additionally, the accumulation of pisiferic acid showed a strong positive correlation with the MEpink module (Fig. 2D). Further screening revealed several genes and TFs associated with terpenoid biosynthetic pathways within these modules. By combining expression correlations, a total of three genes (Ats18G00267700, Ats14G00212560, and Ats14G00197850) were identified as having a highly correlated expression with the accumulation of 1,8-cineole and abietic acid (Fig. 3B). Furthermore, the expression of 36 TFs was found to be significantly correlated with the accumulation of both 1,8-cineole and abietic acid, including members of the AP2-ERF, bHLH, WRKY, MYB, NAC, and bZIP transcription factor families.

Dynamic changes in curcuminoid biosynthesis pathways throughout the development of A. tsaoko fruit

Curcuminoids, a class of diarylheptanes isolated from medicinal plants such as Zingiberaceae, are the main characterizing compounds in A. tsaoko. In this study, a total of three metabolites and 24 genes were annotated to the curcuminoid biosynthesis pathway (Fig. 4A, Tab. S6). Curcumin showed higher accumulation during the MF and JF periods, while the accumulation of bisdemethoxycurcumin (BDMC) and demethoxycurcumin (DMC) peaked during the JF7 period. Among the 24 genes, three genes encoding the enzymes PAL, C4H, and DCS showed the highest expression during the SF period, while the remaining genes showed elevated expression levels during the MF or JF periods. Further WGCNA revealed a strong correlation between the accumulation of curcumin and the MEred module (Fig. 2D). By combining the expression correlations, a total of three genes (Ats12G00174610, Ats14G00202690, and Ats06G00094720) and 47 TFs from families such as MYB, bHLH, and bZIP were found to be highly correlated with the accumulation of curcumin (Fig. 4B).

Fig. 4

Analysis of the curcuminoid biosynthesis in A. tsaoko. A Putative DEGs in the Curcuminoid biosynthesis pathway and their expression levels. Heat maps represent the expression of genes at different stages. Genes in red boxes represent that they have been screened in the WGCNA and are highly correlated with curcuminoid biosynthesis. B Curcuminoid putative transcriptional regulatory network in A. tsaoko. Blue prisms represent TFs, orange triangles represent genes, and purple ellipses represent metabolites

Verification of RNA sequencing data by qRT-PCR

To verify the accuracy of the transcriptome data, qRT-PCR was employed to examine the expression levels of eight candidate genes: DXR (Ats12G00174660), ispE (Ats14G00199250), ispH (Ats10G00146940), gcpE (Ats21G000302260), idi (Ats10G00153630), HMGCR (Ats24G00339410), NES1 (Ats04G0065680), and GERD (Ats24G00336830), with three biological replicates for each gene. The expression levels of all eight candidate genes varied across the MF, JF, JF7, AF, and SF stages of A. tsaoko and exhibited expression patterns similar to those observed in the transcriptome data (Fig. 5). Consequently, the results of our RNA sequencing and qRT-PCR analyses are highly reliable and can be used for in-depth research on the transcriptional regulation of metabolites associated with A. tsaoko fruits.

Fig. 5

qRT-PCR validation of key genes in A. tsaoko. The bar chart represents the results of qRT-PCR, while the line chart represents the results of RNA-seq

Discussion

A. tsaoko has been widely used in the fields of food and folk medicine. On the one hand, it is a key food additive and spice for removing odors and improving taste. On the other hand, as a traditional Chinese medicine, it has significant efficacy and high nutritional value in treating various diseases. Essential oil is the most important active ingredient in A. tsaoko, and its content determines the quality of fruit. 1,8-cineole, which belongs to monoterpenes, is the main component of A. tsaoko essential oil [24]. Additionally, curcumin, as a well-known active molecule in the diarylheptanoid compounds, is regarded as a natural antioxidant with anti-tumor activity and has received much attention in related research and applications [25]. This further highlights the value of A. tsaoko fruits in different fields, from food flavor adjustment to medicinal efficacy exploration, demonstrating its unique position and potential, and becoming an important object in many research and application scenarios. During fruit development and ripening, changes in flavor and appearance result from the accumulation of metabolites. The metabolic and molecular mechanisms underlying these changes require detailed analysis through transcriptional and metabolic studies.

To investigate the accumulation of metabolites at different developmental stages of A. tsaoko fruits, this study used UPLC-MS/MS technology to perform qualitative and quantitative analyses. Consistent with previous studies, the major metabolites identified in A. tsaoko fruit included terpenoids, phenolic acids, and flavonoids [3]. The accumulation levels of the different metabolites varied across developmental stages, with approximately one-third of terpenoid metabolites showing a gradual increase during fruit development, consistent with findings from previous studies [26]. Notably, the JF7 vs. JF group showed the highest number of both DAMs and DEGs in the adjacent stage comparisons, suggesting a more pronounced metabolite and gene regulation between the JF and JF7 stages. GO and KEGG analyses revealed that both DAMs and DEGs were enriched in pathways related to terpenoid and curcuminoid biosynthesis, indicating significant metabolic changes in these compounds during fruit development and ripening. Furthermore, GO analysis revealed that DEGs from the last two developmental stages were enriched in abscisic acid-related terms. As a non-climacteric fruit, the ripening of A. tsaoko may be primarily influenced by abscisic acid [27].

The medicinal and edible value of A. tsaoko fruits is mainly due to their high content of essential oils, especially terpenoid compounds. Previous studies have investigated the biosynthesis mechanisms of terpenoids in aromatic model plants such as Amomum villosum [28] and kiwifruit [29]. However, little is known about the elucidation of the synthetic pathways and the identification of key enzyme genes for terpenoids and curcuminoids in A. tsaoko. In this study, a total of six metabolites and 39 genes were mapped to several terpenoid pathways. The accumulation of 1,8-cineole gradually increased during fruit ripening. As the major important monoterpene component in essential oils [30], the content of 1,8-cineole is a crucial criterion for evaluating the quality of A. tsaoko as both a medicinal and edible resource. It possesses carminative, sedative, antibacterial, antiviral, antiparasitic, and diaphoretic properties [1, 31, 32]. The accumulation of abietic acid also gradually increased during fruit ripening. As an antimicrobial agent, abietic acid derivatives have been shown to control persistent plant bacterial diseases by inhibiting virulence factors in rice [33]. The WGCNA analysis in this study further identified three genes (Ats18G00267700, Ats14G00212560, and Ats14G00197850) encoding the enzymes DXS, HMGCS, and mvaK2, whose expression was highly correlated with the accumulation of 1,8-cineole and abietic acid. The DXS and HMGCS enzymes are involved in the rate-limiting steps of terpenoid skeleton biosynthesis and subsequent modification. Previous research on the assembly and analysis of A. tsaoko genomic data suggested that DXS may be the genetic basis for essential oil formation [34]. In grapes, the expression of five VvDXS genes and three VvHMGS genes was shown to be positively correlated with monoterpene biosynthesis [35]. Overexpression of DXS in Lavandula latifolia also increased the production of essential oil (monoterpenes) [36], further confirming that the expression of DXS in this study may indeed be related to the accumulation of 1,8-cineole. Additionally, the up-regulation of the mvaK2 gene in Morchella eximia was also associated with an increase in the biosynthesis of the key terpenoid precursor isopentenyl pyrophosphate [37], which is consistent with our study. Furthermore, the terpenoid biosynthesis can be promoted by modulating the expression of terpenoid synthase genes through TFs. Our study identified 36 TFs whose expression was highly correlated with 1,8-cineole and abietic acid accumulation, including transcription factor families such as AP2-ERF, bHLH, WRKY, MYB, NAC, and bZIP. Previous research also identified that 20 TF families in A. tsaoko, including MYB, WRKY, NAC, bZIP, and bHLH, may be involved in the regulatory network of terpenoid compounds [26]. Moreover, the regulatory effect of TFs on key genes may be achieved by binding to their promoters. For example, LiMYB, LibZIP, and LiERF in Lavandula × media may positively regulate gene transcription by activating the promoter of the 1,8-cineole synthase genes [38]. AtWRKY18/40 in Arabidopsis thaliana also regulates the biosynthesis of diterpenoid by binding to promoter sequences of DXS and DXR genes [39]. Therefore, transcription factors such as ERF, bHLH, WRKY, and MYB in this study may also activate regulation by binding to the promoters of key genes, thereby promoting the accumulation of terpenoids such as 1,8-cineole. Additionally, epigenetics plays a key role in regulating terpenoid biosynthesis, as seen in tea leaves where hypermethylation reduces monoterpene-related gene expression [40], suggesting that terpenoid biosynthesis in A. tsaoko fruits may also be influenced by epigenetic factors.

Curcuminoids, including curcumin, DMC, and BDMC, are phenylpropanoid derivatives. In our study, a total of three metabolites and 24 genes were annotated to the diarylheptanoid biosynthesis pathway. Among them, curcumin showed greater accumulation at the MF and JF stages, while DMC and BDMC had the highest accumulation levels at the JF7 stage. Curcumin has anti-inflammatory, antioxidant, antiparasitic properties [41] and is a potential COVID-19 adjuvant [42]. DMC has anti-inflammatory, neuroprotective and vasodilatory properties [43]. BDMC, which is more stable than curcumin and DMC [44], inhibits cancer cell growth and induces apoptosis [45]. Most of the genes encoding PAL, C4H, DCS, and CURS enzymes showed the highest expression levels during the MF and JF stages, consistent with the accumulation trend of curcumin. Further WGCNA revealed that curcumin accumulation was highly correlated with the expression of three genes (Ats12G00174610, Ats14G00202690, and Ats06G00094720) encoding C4H and CURS enzymes. CURS is a key enzyme responsible in the curcuminoid biosynthesis. Previous studies have shown that the CURS gene is involved in the synthesis of diarylheptanoids through co-expression network analysis and phylogenetic analysis [46]. The CURS gene family also showed remarkable expansion in curcumin-rich Zingiberaceae plants such as turmeric and ginger [47]. In the study of Zingiber officinale genome, it has also been speculated that some DCS/CURS may be responsible for the synthesis of curcumin compounds [48]. These are consistent with the speculative results of this study. Additionally, 47 TFs such as MYB, bHLH, bZIP were screened and their expression was highly correlated with curcumin accumulation. Studies on ginger have also found that many TFs, including DOF, CPP, NLP, bZIP, C3H, and MYB, are involved in the regulation of genes related to curcumin synthesis [48]. The phenylpropanoid biosynthesis pathway, which is upstream of curcuminoid biosynthesis, is also regulated by the MYB-bHLH-WD40 transcription factor complex [49]. So, the several transcription factor families screened in this study may regulate the curcumin biosynthesis.

Conclusion

In conclusion, to investigate the molecular mechanisms of differential accumulation of terpenoids and curcuminoids during A. tsaoko fruit development, this study performed transcriptomic and metabolomic analyses on A. tsaoko fruits from five different developmental stages, and comprehensively investigated their metabolites and related gene regulatory pathways. The biosynthetic pathways of terpenoids and curcuminoids were mapped, and several key genes and transcriptional regulatory networks promoting their accumulation were revealed. Although this study did not conduct further experimental validation on the selected key genes and TFs, its conclusions were drawn from abundant RNA-seq data, rigorous bioinformatics analysis, and a wealth of research publications. This study delves into the formation mechanism of the pharmacological quality of A. tsaoko fruit, providing scientific guidance for the application of A. tsaoko spices and natural pigment components, and enhancing our understanding of their specific metabolite biosynthesis. The biosynthetic pathways of terpenoids and curcumin compounds revealed in this study provide a scientific basis for strategies to increase the production of other valuable plant metabolites. These achievements are not only of great significance for the development of new drugs in the pharmaceutical field, but also for the improvement of crop nutritional and economic value in the agricultural field, promoting the sustainable development of related industries.

Materials and methods

Plant material and sampling

The samples were collected in Xichou, Yunnan, China (34°47′ N, 109°2′ E). It was identified as Amomum tsaoko Crevost et Lemarie by Professor Guodong Li of Yunnan University of Chinese Medicine. The voucher specimens (5,326,230,012) were deposited in the Herbarium of Yunnan University of Chinese Medicine (YUNCM). The collection site has the characteristics of no extreme cold in winter, relatively mild in summer, and high precipitation. The annual average rainfall is 1,294 mm, the annual average temperature is 15.9 °C, and the relative humidity is 82%. The fruiting period begins in May and continues until September, with the harvesting occurring from July to September. The fruit size gradually expands in the first month and remains almost unchanged in the last four months. The color of the fruit is mainly reddish brown, with little change throughout the entire process. The fruiting period is from May to September. So, five plants with the same growth conditions were selected for investigation, and fruits at different maturity stages were collected from May to September (samples were collected monthly). Each sample group was represented by three biological replicates, which were subsequently wrapped in tin foil and labelled before being frozen in liquid nitrogen and stored in a -80 ℃ freezer for future metabolomic and transcriptomic analyses.

Metabolite extraction and derivatization and UPLC–ESI–MS/MS conditions

Using the technique of vacuum freeze-drying, biological samples are placed in a freeze-dryer model Scientz-100F, then grinding (30 Hz, 1.5 min) the samples to powder form by using a grinder (MM 400, Retsch). Next, 50 mg of the powdered sample is weighed using an electronic balance (MS105DΜ). To this is added 1,200 μL of a prechilled 70% methanolic aqueous solution maintained at -20 °C and containing an internal standard. For samples with a mass less than 50 mg, the extractant is adjusted proportionally to 1,200 μL per 50 mg of sample. The mixture is vortexed for 30 s every 30 min for a total of 6 times. After centrifugation at 12,000 rpm for 3 min, the supernatant is removed. The sample is then filtered through a 0.22 μm microporous filter and stored in an injection vial for subsequent UPLC-MS/MS analysis. The analysis of the sample extracts was conducted on an ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-ESI–MS/MS) setup. The chromatographic separation was achieved using an Agilent SB-C18 column, with dimensions of 2.1 mm by 100 mm and a particle size of 1.8 µm. The mobile phase was a mixture of 0.1% formic acid in water (solvent A) and in acetonitrile (solvent B). The gradient elution started at 95% A/5% B, shifted to 5% A/95% B over 9 min, held for 1 min, then returned to 95% A/5% B in 1.1 min and held for 2.9 min. The flow rate was 0.35 mL/min, the column temperature was 40 °C, and the injection volume was 2 μL. The outlet of the system was connected to an ESI-QTRAP-MS. ESI settings included a source temperature of 500 °C, ion spray voltages of 5,500 V (positive) and -4,500 V (negative), and GSI, GSII, and CUR gases at 50, 60, and 25 psi. High CAD and intermediate nitrogen were used for MRM scans, with custom DP and CE for each transition to track specific metabolite elution times.

Metabolite data processing

Metabolite identification was facilitated by comparison of secondary spectral data with the internal MetWare database (MWDB). Quantitative analysis of metabolites was performed in multiple reaction monitoring (MRM) mode using a triple quadrupole mass spectrometer. Mass spectrometry data were processed using Analyst software (version: 1.6.3). Total ion current (TIC) normalization method was employed in this data analysis. This study used the MetaboAnalystR package in R software to conduct Orthogonal Partial Least Squares Discriminant Analysis (OPLS-DA) modeling, and conducted 200 random permutation and combination experiments on the data to verify the reliability of the model. Based on the OPLS-DA model, the Variable Importance in Projection (VIP) was used to preliminarily screen metabolites with differences in developmental stages. Metabolites with VIP ≥ 1 or fold change (FC) ≥ 2 and ≤ 0.5 are considered to have differences. The identified metabolites were then annotated with reference to the Kyoto Encyclopedia of Genes and Genomes (KEGG) Compound database. These annotated metabolites were mapped to the KEGG Pathway database to identify pathways with significant metabolic regulation. The significance of these pathways was evaluated by metabolite set enrichment analysis (MSEA), with p-values derived from the hypergeometric test determining their statistical significance.

RNA extraction, cDNA preparation, and illumina sequencing

Total RNA was extracted from A. tsaoko fruits at five developmental stages, each with three independent biological replicates. RNA integrity was verified by agarose gel electrophoresis, RNA purity was assessed using NanoDrop, and accurate RNA quantification was performed using Qubit. After sample quality control, mRNA was enriched using oligo (dT) magnetic beads. Fragmentation reagent was added to the enriched mRNA to induce fragmentation. The fragmented mRNA was then used to synthesize first- and second-strand cDNA according to standard protocols. Subsequent steps, including end repair, A-tailing, sequencing adapter ligation, purification, PCR amplification, and product circularization, were performed to complete library preparation. After quality control of the library, the libraries were subjected to sequencing with the sequencing platform DNBSEQ (BGI, Shenzhen, China), and the products were considered raw reads. Sequence data were submitted to the National Center of Biotechnology Information (NCBI) databases under accession number “PRJNA1175636”.

Raw data filtering and mapping

The raw data were filtered using fastp (version: 0.21.0) to generate clean reads for subsequent analyses [50]. The quality of the filtered data was assessed using fastqc (version: 0.11.9) [51]. The filtered transcriptome sequences were then aligned to the A. tsaoko reference genome (unpublished data, and its assembly and annotation process is described in detail in Supplementary Material 1) using Star (version: 2.7.9a) [52], and gene expression was subsequently analyzed.

Differentially expressed genes (DEGs) and functional analysis

RSEM [53] was used to quantify transcript read counts per sample, which were converted to FPKM (Fragments Per Kilobase per Million bases) values to determine gene expression levels. DESeq2 (version: 1.26.0) [54] was then used for differential expression analysis, filtering DEGs with p-value < 0.05 and |log2FoldChange|> 1. Next, clusterProfiler (version: 3.14.3) [55] was used for GO [56] and KEGG [57] enrichment analysis of the DEGs.

Weighted Gene Co-Expression Network Analysis (WGCNA)

WGCNA was applied to the A. tsaoko fruit developmental transcriptome to detect highly coordinated gene clusters. Genes exceeding an FPKM threshold of 10 were used to construct a weighted co-expression network with modules identified by WGCNA (version: 1.6.6) in R (version: 3.4.4). The network construction deviates from the default value. Based on the scale-free topology criteria and consideration of network connectivity the soft threshold power is set to 11 (Fig. S3), unsigned TOMtype is used, the minimum module size is 30, and the merged cutting height is 0.2. The transcriptional regulatory networks were constructed by integrating Pearson correlation coefficients (PCC > 0.8) among structural genes, TFs, and metabolites. Subsequently, these networks were visualized using Cytoscape software (version: 3.9.1, USA).

Quantitative RT-PCR validation of differential gene expression

Based on the FPKM values from the transcriptome data, eight DEGs in A. tsaoko were selected for qRT-PCR experiments. Total RNA was extracted from fruits at five developmental stages of A. tsaoko using the RNAprep Pure Plant Plus Kit (Tiangen, China), with three biological replicates for each stage. Primers for the eight genes and the internal reference gene (TUA) were designed using Primer Premier 5.0 (Tab. S1). cDNA was synthesized using FastKing gDNA Dispelling RT SuperMix (Tiangen, China). The amplification protocol was as follows: incubation at 37 ℃ for 2 min to remove DNA contamination; incubation at 55 ℃ for 15 min, followed by 85 ℃ for 5 min to terminate the reaction. The melting curve program used to determine the specificity of the reaction was as follows: 95 ℃ for 2 min, followed by 40 cycles of 95 ℃ for 5 s and 60 ℃ for 30 s. The relative gene expression was calculated using the 2^−△△CT method.