Abstract Camellia oleifera Abel., recognized as a significant oil plant, is of immense potential health and economic value. Due to the self-incompatibility of C. oleifera, pollination relies on cross-pollination from other flowers. Additionally, the asynchronous flowering periods of individual plants result in low fruit set and yield, which limits the broader cultivation and utilization of this species. The study investigated the dynamic changes and regulatory patterns of different tissues within flower buds from flower bud development to flowering, employing a multi-faceted approach that included cellular dissection, analysis of hormone content, and transcriptome analysis. This study demonstrates that ABA and SA, rather than GA[3], IAA, ZT, acts as potentially effective endogenous agents to promote flowering in the later stages of flower development, which is a critical period for the maturation of pollen and embryo sacs; while ZT plays a more significant role in the early stages of flower bud development. Transcriptome analysis indicated that C.oleifera primarily regulates the late stages of flower bud development via regulating genes involved in starch and sucrose metabolism in petals, monoterpene synthesis and ABC transporters in pistils and stamens. WGCNA identified four key modules associated with the development of stamens and pistils in the late stage of C.oleifera flower buds, and also screened out key core genes, including CoBMY7/8, CoTPP6/10, and CoG8H7/11, which are involved in the regulation of flowering time. These findings enhance our understanding of the developmental changes in stamens, pistils, and petals during the flower bud development of C. oleifera. Moreover, they provide a foundation for manipulating flowering time and improving fruit set by regulating the expression of key genes in future studies. Supplementary Information The online version contains supplementary material available at 10.1186/s12870-025-06201-w. Keywords: Camellia oleifera, Flower bud development, Monoterpene synthesis, ABC transporters, Sucrose metabolism Introduction Camellia oleifera Abel, along with palm, coconut, and olive, is recognized globally as one of the four major woody oil plant. It is an essential woody edible oil tree species in Southern China [[38]1]. C. oleifera has a lengthy history of cultivation in China, boasting various varieties with unsaturated fatty acid content as high as 90%, renowned for its capacity to soften blood vessels and reduce blood lipids and pressure [[39]2, [40]3]. As of 2023, the national C. oleifera plantation area has reached 4.5 × 10^6 hectares, with an average oil yield exceeding 40 kg per mu. However, 70% of this area consists of low-yielding plantations, which significantly limits the overall growth and development of the C. oleifera industry [[41]4–[42]6]. Factors such as lax control of C. oleifera cultivars, insufficient application of cultivation management techniques, and outdated industrial infrastructure have led to low-yielding plant varieties [[43]7]. C. oleifera is a late-stage self-incompatible plant, initiating flower bud differentiation in mid-May, with the peak flowering period occurring from late October to late November [[44]8]. Based on the peak flowering period, they can be categorized into early flowering (‘Autumn Flower No. 1’ in mid-to-late October), medium flowering (‘Winter Flower No. 6’ in early to mid-November), and late flowering (‘Winter Flower No. 8’ after late November) types [[45]9]. However, the occurrence of inconsistent flowering periods disrupts normal pollination. Additionally, the unfavorable temperature conditions and increased precipitation in the southern region during autumn and winter (post-November) adversely affect the flowering time of C. oleifera, hindering pollen germination and disrupting insect pollination activities. These factors collectively increase the difficulty of pollination, leading to a lower fruit set and the phenomenon of ‘more flowers, fewer fruits [[46]10]. Consequently, enhancing the fruit-setting rate and yield per unit area of C. oleifera has become an urgent issue for the development of China’s C. oleifera industry [[47]11]. The flowering period of C. oleifera is a pivotal factor affecting its pollination rate and fruit-setting rate. Cellular anatomical analysis of common C. oleifera varieties, such as ‘Huashuo’ (full flowering period in November) and the superior line ‘Chunhua No. 2’ (full flowering period in March), reveals minimal differences in the early stage of flower bud differentiation [[48]11]. However, the flowering time exhibits a significant disparity of several months. This suggests that the slow development period of floral organs in the late stage of flower bud differentiation is a key determinant of different flowering periods [[49]11]. Research findings indicate that the application of 6-BA and GA[3] during the initial stages of flower bud development in C. oleifera can influence the process. Specifically, the application of 50 mg/L GA[3] accelerates the onset of flowering, albeit prolonging the duration to full bloom by 3–4 days. Furthermore, a comprehensive study conducted by the research team involved the application of five plant growth regulators GA[3], GA[4 + 7], 6-BA, TDZ, and cyanamide at varying concentrations on C. oleifera flower buds during the later stages of floral organ differentiation. This investigation revealed diverse impacts of these growth regulators on the regulation of flowering time in C. oleifera. Notably, the application of 0.5% hydrogen cyanamide exhibited the most pronounced effect, advancing the flowering period by 12 days [[50]11]. These studies suggest that the developmental period of flower organs in the late stage of flower bud differentiation is a critical phase for regulating the flowering period. This study focuses on the crucial role of the floral organ development stage in the late differentiation stage, regulating the flowering stage of C. oleifera. The research involves flower bud dissection (petals, stamens, and pistils), flower bud anatomical analysis, determination of plant hormone content, alongside transcriptome sequencing and other experiments. Through an in-depth analysis of the key genes and mechanisms of petals, stamens, and pistils during the slow development period of floral organs in the late stage of flower bud differentiation. This research aims to provide technical support for improving the fruit-setting rate and yield per unit area of C. oleifera. It lays the foundation for artificial regulators to control the flowering period of C. oleifera and offers guidance for transforming low-yield forests in China. Materials and methods Plant materials The C.oleifera cultivar ‘Huashuo’ was sourced from the C.oleifera experimental base at Jiangxi Academy of Forestry National Camellia Seedling Base. Sampling began on day 0 of the late stage of the morphological differentiation of C.oleifera flowers (from mid-September), defined as the formation of pistil and stamen primordia and the early development of male and female gametophytes, and continued on days 2, 4, 8, 16, and 32. Three replicates were set for samples of different days, and a total of 18 groups of samples were collected. The 0-day sample refers to the C.oleifera where flower bud differentiation commences subsequent to the completion of stamen and pistil differentiation. Ninety flower buds with optimal growth and uniform development were collected from the middle and upper parts of the plant crown. Thirty of these buds were washed with distilled water and placed immediately in FAA (formalin, 70% alcohol, glacial acetic acid mixture) fixative for subsequent paraffin sectioning or storage at 4 °C. The remaining 60 buds were rapidly frozen in liquid nitrogen, dissected on dry ice into petals, pistils, and stamens, ground into powder in liquid nitrogen, and promptly stored at -80 °C for later use in the determination of endogenous hormone content and RNA extraction. Anatomical analysis of flower buds Tissues used in the anatomical analysis were prepared according to a previously described method with some modifications [[51]12]. Key steps included sectioning, dewaxing, staining, mounting, and drying before observing and photographing the paraffin sections of flower buds using an Olympus microscope. Determination of hormone content in different flower tissues Plant hormones such as indole-3-acetic acid (IAA), zeatin (ZT), and abscisic acid (ABA) contents, gibberellic acid 3 (GA3), and salicylic acid (SA) were determined using liquid chromatography–electrospray ionization–tandem mass spectrometry apparatus (6410; Agilent, Santa Clara, CA, USA), referring to a previously described method [[52]13]. RNA preparation, transcriptome sequencing, and gene functional annotation Total RNA was extracted from petals, pistils, and stamens at six stages of floral organ development (0, 2, 4, 8, 16, and 32 days) using the RNAprep Pure Plant kit (Tiangen, Beijing, China). RNA concentration, purity, and integrity were assessed, and the samples met the required standards. Sequencing libraries were then prepared using a library preparation kit (Shanghai Yasen Biotechnology), and library quality was evaluated on an Agilent Bioanalyzer 2100 system. The libraries were sequenced on the Illumina NovaSeq platform, generating 150 bp paired-end reads. Valid data were aligned to the reference genome, and further analysis and annotation were performed using Hisat2 software. Gene functions were annotated through sequence alignment with various databases, including Nr (NCBI non-redundant protein sequences), Pfam (protein family), KOG/COG (Clusters of Orthologous Groups of proteins), Swiss-Prot (a manually annotated and reviewed protein sequence database), KO (KEGG Orthology), and GO (Gene Ontology). Gene expression levels were quantified using the following formula: fragments per kilobase of transcript per million mapped reads (FPKM) = cDNA fragments / [mapped fragments (millions) * transcript length (kb)]. Differential expression screening, annotation and enrichment analysis Differential expression between the two groups was analyzed using DESeq2, and the P values were corrected using the Benjamini–Hochberg method to control the false discovery rate. Genes with a corrected P-value < 0.01 and a fold change ≥ 2, as determined by DESeq2, were considered differentially expressed genes (DEGs). qRT-PCR was conducted using GAPDH and TIF3H1 as the internal reference genes to validate the alterations in key genes within the three floral organs [[53]14–[54]16]. A GO enrichment analysis of the DEGs was conducted using the cluster-Profiler package, based on a Wallenius non-central hypergeometric distribution. A Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis was performed using the KOBAS database and cluster Profiler software. The sequences of the DEGs were aligned (BLASTX) with the genomes of related species to obtain predicted protein interactions, which were then visualized in Cytoscape. Weight gene co-expression network analysis (WGCNA) The key genes involved in regulating the flowering period were also explored by constructing a co-expression network using the WGCNA (v1.51) software package in R. All genes with FPKM values < 1 in the samples were filtered out. A minimum module size of 30 and a module similarity threshold of 0.25 were defined based on automatic network construction and block level module detection. The characteristic gene values of each module were calculated, and their correlations with various physiological indicators were tested. The gene co-expression network was visualized using Cytoscape [[55]17]. Results Slicing observation of organ development in C. oleifera flowers The developmental trajectory of the stamens and pistils in C.oleifera was meticulously observed over six distinct periods (Fig. [56]1A and B). The stamens underwent several key developmental stages, starting with the emergence of young anthers at the flower bud’s apex on day 0, followed by the initiation of germination and the formation of microspore mother cells on day 8. By day 16, these cells had separated into single entities, marking the completion of pollen maturation, and by day 32, the microspores had transformed into binucleate pollen (Fig. [57]1C). Concurrently, the pistils exhibited significant developmental changes. On day 0, a closed ovary developed around the carpel primordium. By day 2, the epidermal cells at the base of the nucellus underwent rapid division, contributing to the nucellus’s formation and stabilization. On day 16, the pistils were characterized by the emergence of an egg cell and two auxiliary cells, along with the formation of three antipodal cells at the chalazal end (Fig. [58]1C). Fig. 1. [59]Fig. 1 [60]Open in a new tab C. oleifera flowers (A), anatomical tissues (B), and slices (C) showing the slow development stage of flower organs (days 0–32). C-a, 1: ovary, 2: ovule primordia, 3: compartments, 4: filaments, 5: anthers (an embryonic pollen sac can be seen in the corner). C-b, 1–3: young anthers (in 1 and 2, multiple spore progenitor cells have already formed), 4: ovules, 5: bead stalks, 6: nucellus, 7: sub-rooms. C-c, 1: drug separates vascular bundles, 2: the thin-walled tissue of the septum, 3 and 4: primary sporogenic cells, 5 and 6: ovules, 7: megaspore mother cells, 8: megaspore mother cells undergoing meiosis. C-d, 1 and 2: anthers, 3: isolated vascular bundles, 4: microspore mother cells, 5: ovules, 6: megaspore mother cells. C-e, 1 and 2: microspore mother cells (meiosis), 3: pollen chambers, 4: cyst walls (after maturation of the pollen grains, the cyst walls of adjacent pollen sacs rupture, thus connecting the adjacent pollen sacs), 5: 8-cell-nuclei embryo sac. C-f, 1: microspore, 2: 8-cell-nuclei embryo sac. bar represents 50 μm Changes in hormone content in the petals, pistils, and stamens of C. oleifera flowers The contents of ZT, GA[3], IAA, ABA, and SA in petals, pistils, and stamens were measured at different time points (Fig. [61]2; Table [62]S1). The results showed a higher content of ZT in stamens (6.15–23.91 ng/g) than in petals and pistils. The trend was largely consistent, with the highest ZT content measured at 8 days. The GA[3] content in stamens was highest on day 0 (3758.25 ng/g) and decreased significantly, by 1/3 − 1/4 of that amount, until day 32. However, there was less variation in the GA[3] content in the petals and pistils, with the maximum reached at day 8 in both. The content of IAA in stamens (301.73 ng/g), petals (972.52 ng/g), and pistils (227.51 ng/g) peaked at day 2, 4, and 16, respectively, and was significantly higher in petals than in stamens and pistils. The ABA and SA contents in stamens did not change significantly between days 0 and 16, but increased significantly, by ~ 4-fold, by day 32 (ABA: 175.5 ng/g; SA:233.53 ng/g). The SA content in petals followed the same pattern. Fig. 2. [63]Fig. 2 [64]Open in a new tab Changes in the absolute contents of the plant hormones zeaxanthin (ZT), gibberellin (GA[3]), auxin (IAA), abscisic acid (ABA), and salicylic acid (SA) in the petals, pistils, and stamens of C. oleifera during the late stage of flower development. Duncan’s test was conducted using SPSS software DEGs identified by transcriptome analysis in the petals, pistils, and stamens of C. oleifera flowers A transcriptome analysis was performed of the petals, pistils, and stamens of C.oleifera on days 0, 2, 4, 8, 16, and 32 to investigate the mechanism of flower organ differentiation in regulating flowering. The data were divided into five groups (2 d vs. 0 d, 4 d vs. 2 d, 4 d vs. 8 d, 16 d vs. 8 d, and 32 d vs. 16 d) for petals, pistils, and stamens, for a total of 15 sets of comparative data. The results of a principal component analysis performed on the data of petals (Fig. [65]3A), pistils (Fig. [66]3B), and stamens (Fig. [67]3C) showed very similar gene expression patterns among the three biological replicates at each time point and that the data were well separated from each other. The number of DEGs in each of the 15 groups is shown in Table [68]1; Fig. [69]3D. In petals and pistils, the 32 d vs. 16 d group had the highest number of DEGs [9,308 (3,887 up-regulated, 5,421 down-regulated) and 12,640 (5,867 up-regulated, 6,773 down-regulated), respectively]. In the 16 d vs. 8 d group, fewer DEGs were detected [1,307 (443 up-regulated, 864 down-regulated) and 3,521 (2,131 up-regulated, 1,390 down-regulated), respectively]. In stamens, the highest number of DEGs was in the 2 d vs. 0 d group [8,740 (4,176 up-regulated and 4,564 down-regulated)]; the lowest number was in the 16 d vs. 8 d group [1,474 (688 up-regulated and 786 down-regulated)]. Among the five groups, the most common DEGs in petals, pistils, and stamens occurred in the 2 d vs. 0 d group (Fig. [70]3D), with 1,588 up-regulated and 1,733 down-regulated DEGs. The least common DEGs were in the 16 d vs. 8 d group (37 up-regulated and 117 down-regulated) (Fig. [71]3D). In petals, pistils, and stamens, there were 114, 279 and 186 common DEGs in the five groups, respectively (Fig. [72]3E). Fig. 3. [73]Fig. 3 [74]Open in a new tab Transcriptome PCA analysis (A-C), differential gene distribution (D), and veen analysis(E) of petals, pistils, and stamens of C. oleifera in the late stage of flower development Table 1. Numbers of DEGs among fifteen sample combinations in different flower organs Flower organs Sample combinations Total Up-regulated Down-regulated Petal 2 d vs. 0 d 7069 3375 3694 4 d vs. 2 d 5188 2039 3149 8 d vs. 4 d 3369 2322 1047 16 d vs. 8 d 1307 443 864 32 d vs. 16 d 9308 3887 5421 Pistil 2 d vs. 0 d 9599 4560 5039 4 d vs. 2 d 6824 3277 3547 8 d vs. 4 d 4667 3049 1618 16 d vs. 8 d 3521 2131 1390 32 d vs. 16 d 12,640 5867 6773 Stamen 2 d vs. 0 d 8740 4176 4564 4 d vs. 2 d 5979 2763 3216 8 d vs. 4 d 5151 3022 2129 16 d vs. 8 d 1474 688 786 32 d vs. 16 d 7447 3373 4074 [75]Open in a new tab A KEGG enrichment analysis of the DEGs of the 15 different groups of petals, pistils, and stamens was conducted to determine the metabolic pathways significantly associated with the DEGs (Table [76]S2). In petals, phenylpropanoid biosynthesis (ko00940) and starch and sucrose metabolism (ko00500) were the most significantly enriched pathways based on the number of DEGs (Table [77]S2.1). In pistils, the DEGs were enriched in ABC transporters (ko02010) and monoterpenoid biosynthesis (ko00902) (Table [78]S2.2). The DEGs in stamens were enriched in phenylpropanoid biosynthesis (ko00940), ABC transporters (ko02010), stilbenoid, diarylheptanoid and gingerol biosynthesis (ko00945), and monoterpenoid biosynthesis (ko00902), with phenylpropanoid biosynthesis (ko00940) and ABC transporters (ko02010) being the most significantly enriched pathways (Table [79]S2.3). Analysis of DEGs involved in starch and sucrose metabolism in petals As well as providing the energy and carbon framework required for flower growth and development, sugar also acts as a signaling molecule to regulate flowering [[80]18]. During the late differentiation stage of C.oleifera flowers, DEGs in the petals were found to be significantly enriched in starch and sugar metabolism pathways. Sucrose is hydrolyzed by acid beta-fructofuranosidase (FOS) in the extracellular matrix to generate glucose and fructose, which are then transported into cells through monosaccharide transporters. Sucrose is also hydrolyzed by cytoplasmic FOS to likewise obtain glucose and fructose, and can be cleaved by cytoplasmic sucrose synthase (SuSy) to obtain fructose and uridine diphosphate glucose UDPG) (Fig. [81]4A). In C.oleifera, CoFOS3/7 was up-regulated in petals by 5.3- and 4.96-fold on day 32 and day 16, respectively. CoSuSy19 was up-regulated by 4.04-fold in the 4 d vs. 2 d group, while CoSuSy48 was up-regulated by 4.06- and 4.39-fold in the 2 d vs. 0 d and 8 d vs. 4 d groups, respectively (Fig. [82]4B). Glu-1-P is transformed into adenosine diphosphate glucose (ADPG) through the action of ADPG pyrophosphorylase (AGPase) (Fig. [83]4A). Subsequently, ADPG serves as a substrate in the synthesis of starch, a process facilitated by the enzymes GBBS1 (granule-bound starch synthase), STS (starch synthase), and GBE (1,4-alpha-glucan-branching enzyme). The starch is subsequently converted into dextrin by the enzyme beta-amylase (BMY) (Fig. [84]4A). In this study, the expression of CoGBSS1.1/1.5 and CoGBE1/2 in the starch synthesis pathway was down-regulated in the 32 d vs. 16 d group, while CoBMY8/19, encoding the enzyme that converts starch to dextrin, was significantly up-regulated, indicating a decrease in starch content (Fig. [85]4B). The genes encoding key enzymes (CoTPSe9/17 and CoTPP6/8/10) in the synthesis of trehalose (Tre) were up-regulated, indicating the accumulation of Tre in the later stages of flower development. Significant differential expression was detected by qRT-PCR (Fig. [86]S1). Our findings suggest that CoTPSe9/17 plays a crucial role in regulating flowering time in C.oleifera, particularly within the context of the sucrose and starch metabolism pathways. Fig. 4. [87]Fig. 4 [88]Open in a new tab The expression profiles of genes in starch and sucrose metabolism pathway. (A) A model based on the starch and sucrose metabolism pathway in plants [[89]19, [90]20]. The green ellipse represents key enzymes in this metabolic pathway, while the blue ellipse represents transcription factors related to HXK. (B) Comparison of key gene expression patterns related to starch and sucrose metabolism pathway in petal of C.oleifera. Blue and red respectively indicate a relative increase and decrease in expression levels (Log[2]fold change). The ID and expression level for these genes are shown in Table [91]S3. Enzyme and chemical names are abbreviated as follows; SE/CC, sieve element/companion cell; Suc: sucrose; Suc-6P: sucrose 6-phosphate; Fru: fructose; Fru-6P: fructose 6-phophate; Hex-P: hexose phosphates; Glu: glucose; Glu-1P: glucose 1-phosphate; UDP-G: UDP-glucose; ADP-G: adenosine diphosphate glucose; Tre-6P: trehalose 6-phophate; Tre: trehalose; AP2: APETALA2-type TF; bZIP: basic leucine zipper; EIN3: ethylene insensitive 3; B3: B3 domain protein; CCT: CCT domain protein; TPSe: trehalose-6-P synthase; TPP: trehalose-6-P phosphatase; snRK1: SNF1-related protein kinase; SuSy: sucrose synthase; SPS: sucrose phosphate synthase; SPP: sucrose phosphate phosphatase; UGP2: UDP-Glucose Pyrophosphorylase 2; AGPase: Glucose-1-phosphate adenylyltransferase large subunit; GBSS1: Granule-bound starch synthase; STS: starch synthase; GBE: 1,4-alpha-glucan-branching enzyme; BMY: Beta-amylase; FOS: acid beta-fructofuranosidase; HXK: hexokinase; FRK: fructokinase; GPI: glucose-6-phosphate isomerase; FT: fructosyltransferases Analysis of DEGs involved in monoterpene synthesis in pistils and stamens In plants, the synthesis of terpenoids, a diverse class of organic compounds, involves two primary pathways: the mevalonate (MVA) pathway, which occurs in the cytosol, and the methylerythritol 4-phosphate (MEP) pathway, found in the plastids, primarily responsible for the production of monoterpenes. Both pathways necessitate the action of terpenoid synthases (TPSs) to form the final terpenoid products (Fig. [92]5A) [[93]21]. Monoterpenes are synthesized through the MEP pathway. Previous studies have shown that terpenoid biosynthetic pathways are induced in the late stages of Lonicera japonica flower development [[94]22]. As our study showed that monopenoid biosynthesis is significantly enriched in the pistils and stamens of C.oleifera (Table [95]S2), the expression patterns of key genes in the MEP and MVA pathways were investigated in these organs (Table [96]S4). In the MVA pathway, CoHMGR1 and CoHMGR2 were up-regulated in pistils in the 32 d vs. 16 d group, by 1.7- and 2.0 -fold, respectively (Table [97]S4), while CoHMGR3 and CoHMGS1 were up-regulated by 2- and 1.4-fold in stamens, respectively, also in the 32 d vs. 16 d group (Table [98]S3). Fig. 5. [99]Fig. 5 [100]Open in a new tab Expression profiles of genes in the monoterpenoid synthesis pathway. (A) A model based on the monoterpenoids synthesis pathway in plants [[101]24], with small ellipses representing key catalytic enzymes in this pathway. (B, C) Comparison of the expression patterns of key genes related to monoterpene synthesis in the pistils and stamens of C.oleifera. Blue (red) indicates a relative increase (decrease) in expression levels (log[2]-fold change). The IDs and expression levels of these genes are shown in Table [102]S4. HMGS: hydroxymethylglutaryl-CoA synthase; HMGR: hydroxymethyl glutaryl-CoA reductase; MK: mevalonate kinase; PMK: phosphomevalonate kinase; IPI: isopentenyl pyrophosphate isomerase; DXS: 1-deoxy-D-xylulose 5-phosphate synthase; DXR: 1-deoxy-D-xylulose 5-phosphate reductoisomerase; MDH: mannitol dehydrogenase; IO: iridoid oxidase; NES: (3 S,6E)-nerolidol synthase; G8H: geraniol 8-hydroxylase-like; BIS: beta-bisabolene synthase; NDH: (+)-neomenthol dehydrogenase; TPS: terpene synthases; MVAPP: mevalonate-5-diphosphate; DMAPP: dimethylallyl pyrophosphate; GPP: geranyl pyrophosphate; FPP: farnesyl pyrophosphate; DXP: 1-deoxy-D-xylulose 5-phosphate; MEP: 2-C-methyl-D-erythitol 4- phosphate; IPP: isopentenyl diphosphate; 8-HYG: 8-Hydroxygeraniol; 8-OXG: 8-oxogeranial The interconversion between IPP (isopentenyl diphosphate) and DMAPP (dimethylallyl diphosphate) is a reversible enzymatic reaction (Fig. [103]5A) facilitated by isopentenyl pyrophosphate isomerase, which dynamically modulates the cellular levels of both compounds. The short-chain isoprene transferase GPPS catalyzes the sequential condensation of IPP and DMAPP to yield geranyl diphosphate (GPP) [[104]23]. The expression levels of CoGPPS1/3 and CoGPPS2 were up-regulated by 3-fold in pistils in the 32 d vs. 16 d and 8 d vs. 4 d groups, respectively (Table [105]S4). GPP is used to synthesize monoterpenes such as iridotrial, 8-oxolinalool, neomenthol, and α-terpineol (Fig. [106]5A). In pistils, CoTPS2/3/8/11/13/14 expression was significantly up-regulated in the early stage (2 d vs. 0 d) and significantly down-regulated in the later stage (32 d vs. 16 d), while CoG8H1/2/3/6/7/11/12/13/14 was continuously up-regulated, by 2- to 6-fold (Fig. [107]5B). CoNES1/2/3, CoTPS4/9, and CoBIS were significantly up-regulated in the early stage (2 d vs. 0 d), by 8.5-fold in pistils (Fig. [108]5B), as was the expression of COBIS. CoG811 and CoTPS7 were significantly up-regulated in the 32 d vs. 16 d group, but there were no changes in the other groups (Fig. [109]5B). The expression of certain DEGs was validated through qRT-PCR. (Fig. [110]S1). These findings suggest that critical genes within the terpenoid biosynthesis pathway are implicated in the regulation of pistil and stamen development during the later stages of C.oleifera flower bud maturation. Analysis of DEGs involved in ABC transporters in pistils and stamens ABC transporter protein (ATP-binding cassette transporter), integral to material transport across various stages of plant growth and development, play roles including hormone transport, regulation of heavy metal ion efflux, response to environmental stress, and the transport of secondary metabolites (Fig. [111]6A). To elucidate the role of ABC transporters in floral organ development, this study observed the level of ABC transporters significant enrichment in stamens and pistils during the late differentiation phase of C.oleifera flowers. In pistils, CoABCG2.2 and CoABCG 2.8 were up-regulated by 7.33- and 7.8-fold, respectively, in the 32 d vs. 16 d group (Fig. [112]6B). The levels of CoABCC1.3, CoABCB1.2/1.9/1.20, and CoABCG2.2/2.8/2.24 in stamens were up-regulated 3- to 5.5-fold, respectively, in the 32 d vs. 16 d group, and by ~ 2-fold in the 2 d vs. 0 d group, with no significant changes at other stages (Fig. [113]6C). CoABCA3.4/3.5/3.6, CoABCB1.27/1.5, CoABCC1.6, and CoABCG2.3 in pistils were initially up-regulated and then continuously down-regulated during the late floral differentiation stage (Fig. [114]6B), as also observed for the expression of CoABCA3.5/3.6, CoABCB1.6/1.12/1.21, CoABCC1.5/1.6/2.1, and CoABCG2.3/2.16 in stamens (Fig. [115]6C). Furthermore, the significant differential expression of these genes was validated through qRT-PCR (Fig. [116]S1). ABCG transporters are known to influence the transport and synthesis of essential substrates necessary for pollen development [[117]25, [118]26]. Additionally, ABCG plays a role in regulating the secretion of cuticles on the surfaces of petals and stamens, which in turn affects their elongation [[119]27]. Furthermore, ABCB is involved in the transport of auxin during anther development, a process that is crucial for pollen development [[120]28]. Previous research and our findings indicate that CoABCG2.2/2.8 and CoABCB1.6/1.12/1.21 are critical genes that regulate the flowering time of C.oleifera. Fig. 6. [121]Fig. 6 [122]Open in a new tab The expression profiles of ABC transporter genes. (A) A model based on ABC transporters in plants [[123]29]. The orange and yellow rectangles represent gene names and gene functions, respectively, and the green circles are substances transported by ABC proteins. (B, C) Comparison of the expression patterns of key ABC transporter genes in the pistil and stamen of C.oleifera. Blue (red) indicates a relative increase (decrease) in expression levels (log[2]-fold change). The IDs, expression levels, and annotations of these genes are shown in Table [124]S5 Co-expression network analysis A WGCNA was conducted using 11,513 genes (FPKM > 1 at all sequencing points) and the FPKM values from different sampling points to further identify the genes regulating C.oleifera flowering. Genes with the same expression pattern were clustered into the same module, and a clustering tree diagram was generated (Fig. [125]7A). The resulting 17 co-expressed modules (M1-M17) contained 30 (M5) to 3,585 (M11) genes (Fig. [126]7B). Among these modules, M1 and M2 genes in stamens and petals, respectively, had the highest expression levels at 32 days. In pistils, M3 and M4 were highly expressed at day 16 and day 32, respectively. Three modules (M15-M17) were highly expressed in petals; the most significant correlations were those of M15 and M16. Four modules (M6-M9) were highly expressed in pistils, with the most significant correlations being those of M6 and M7. Four modules (M13, M14, M16, and M17) were highly expressed in stamens; the most significant correlations were those of M13 and M14. Fig. 7. [127]Fig. 7 [128]Open in a new tab Cluster analysis and construction of WGCNA co-expression modules based on RNA-Seq data and flower development stages. (A) A clustering tree diagram of genes with topological overlap-based dissimilarity and specified module colors. (B) Associations between modules and examples thereof. The cell represents the correlation coefficient between intrinsic genes of the module and the sample; the p-value is shown in parentheses below the correlation. All cells are color-coded by correlation. (C) Different key co-expressed gene networks in C. oleifera. The node size indicates the extent of the connection with other genes. Enzyme annotations, IDs, and expression levels in petals, pistils, and stamens are listed in Table [129]S7 The co-expression networks of the aforementioned genes, including those involved in monoterpene synthesis, ABC transporters, and starch and sugar metabolism, were visualized to reveal the key genes involved in regulating the flowering period during the late stage of C.oleifera flower differentiation (Fig. [130]7C; Table [131]S6). The co-expression network contained 61 interrelated genes, among which CoBMY7, CoBMY8, CoG8H7, CoG8H11, CoTPP6, CoTPP10, CoABCG2.24, and CoABCC1.9 were identified as the most highly connected hub genes. CoBMY7/8 and CoTPP6/10 encode beta-amylase and Tre-6-P synthase, key enzymes in the synthesis of dextrin and trehalose, respectively. CoG8H7/11 encodes G8H, which catalyzes the synthesis of 8-oxolinalool. Therefore, these genes may mediate communication between major key metabolites and flowering time. The other 53 DEGs were clustered into two modules: turquoise (52 DEGs) and grey (1 DEG) (Table [132]S7). Their gene IDs and expression levels in petals, pistils, and stamens are shown in Table [133]S8. CoNDA1 was up-regulated 4.4-, 10.1-, and 8.5-fold, respectively, in the 32 d vs. 16 d group. CoPLP2 (12.9-fold), CoCYP79D16 (13.4-fold), and CoCYP79A68 (12.5-fold) were also significantly up-regulated at this stage. Discussion Hormonal signaling plays a pivotal role in flowering. For instance, SA regulates flowering time by modulating the transcription of CONSTANS, FLOWERING LOCUS C, FLOWERING LOCUS T, and CONSTANS 1 [[134]30]. Interactions among various phytohormones such as ethylene (ET), IAA, cytokinins, and ABA are significant for floral induction in A thaliana [[135]31], apple [[136]32], pear [[137]33] and Pharbitis nil [[138]34, [139]35]. In a previous study, a transcriptome analysis revealed that the application of exogenous hydrogen cyanamide induced earlier flowering (by 1 day) of C.oleifera, by regulating IAA metabolism and ET-induced MAPK (mitogen-activated protein kinase) signal transduction [[140]11]. In our study, the levels of ZT, GA[3], IAA, SA, and ABA changed significantly during the late differentiation of C.oleifera (Fig. [141]2). However, in the transcriptome analysis, genes in hormone signaling pathways were not notably enriched. This discrepancy suggests that changes in hormone content through the activation of hormone signaling pathways are not the primary control mechanism of C.oleifera flowering during the later stages of C.oleifera flower differentiation. In tomato (Solanum lycopersicum), the activation of a gene encoding sucrose-phosphate synthase allows tomatoes to store more sucrose during flowering. The overexpression of this gene in transgenic plants leads to earlier flowering than in wild-type plants [[142]36]. Inhibiting AGPase in potato (Solanum tuberosum) using antisense RNA led to a 10-fold reduction in starch levels and a significant increase in soluble sugar in the leaves. As a result, the transgenic plants flowered 2–4 weeks earlier, but their growth and development did not differ significantly from that of the wild type [[143]37]. Similarly, in C.oleifera, the genes CoAGPase/4/5/7, CoGBBS1, and CoGBE are down-regulated in the late stage of flower differentiation, such that starch synthesis and soluble sugar buildup are prevented. At the same time, expression of the BMY gene, responsible for starch conversion, is up-regulated, and existing starch is thus consumed. However, this mechanism for regulating the flowering of C.oleifera requires further experimental verification. Tre-6P, a sugar signal molecule found in minimal quantities in plants, is produced from UDPG and Glu-6P through Tre-6P synthase (TPSe) activity and subsequently converted to Tre by Tre-6P phosphatase [[144]38]. Reducing TPS1 expression was shown to delay flowering and to promote sucrose accumulation, regardless of day length. The significance of TPS1 in Arabidopsis flowering has been reported [[145]18]. In rice and tobacco, overexpressing TPS1 enhances drought tolerance but does not influence flowering time [[146]39]. CoTPSe9/17 is up-regulated during C.oleifera differentiation but its role in flowering regulation is unclear. The up-regulation of CoTPP6/8/10 promotes Tre accumulation, possibly linked to an increase in petal biomass. In rice, the tissue-specific expression of OsTPP1 in florets and developing grains induced by the OsMADS6 promoter can enhance yield [[147]40]. The sucrose signal in petals is transmitted to the pistils and stamens, activating the synthesis of monoterpenes including iridotrial, 8-oxolinalool, neomenthol, and α-terpineol (Fig. [148]5). These monoterpene ethers and alcohols have antibacterial, anticancer, and insecticidal effects, among other effects [[149]41]. The enzyme G8H catalyzes the synthesis of 8-oxolinalool from linalool, and its activity differs significantly during the development of C. oleifera flowers. G8H, previously known as geraniol 10-hydroxylase, is a membrane-bound cytochrome-P450 monooxygenase (CYP) belonging to the CYP76 family [[150]42]. Its role in regulating the production of secologanin, which is involved in the biosynthesis of monoterpene indole alkaloids and cyclic ether monoterpenes, has been examined [[151]42]. However, whether G8H is involved in flowering regulation is unknown. Our study showed significant inhibition of CoTPS expression in both pistils and stamens at 32 days, indicating that a decrease in the α-terpineol content during C. oleifera flower development is related to the flowering process, although this relationship remains to be verified experimentally. ABC transporters are represented in plants by eight subfamilies, distinguished by their structural composition and primary biological functions [[152]43]. For instance, the ABCA subfamily is associated with sterol transport, and the ABCB subfamily with detoxification and auxin transport [[153]44, [154]45]. Recent studies of Arabidopsis and rice have underscored the pivotal role of ABC family members in pollen growth and development. ABC proteins also significantly influence C. oleifera development, particularly through the CoABCB and CoABCG subfamilies (Fig. [155]6). In Arabidopsis, the auxin transporters AtABCB1 and AtABCB19 participate in anther development and are crucial for early stamen development, tapetum development, and meiotic cells [[156]28]. Deletion of the AtABCG26 gene in A. thaliana disrupts normal pollen wall development, leading to abnormal pollen formation and the eventual sterility of male plants [[157]25]. The involvement of AtABCG1, AtABCG16, AtABCB1, AtABCB19, AtABCG9, and At ABCG31 in Arabidopsis pollen development has also been reported. Hence, further exploration of the regulatory role of ABC transporters in C. oleifera development is warranted. Conclusion In summary, the results of this study suggest that ABA and SA function as promoters of flowering in C. oleifera, while the development and maturation of flower buds are closely associated with ZT. Transcriptome analysis revealed that genes involved in sucrose metabolism, monoterpene biosynthesis, and ABC transporters play critical roles in the development and maturation of various floral organs in C. oleifera. WGCNA clustered 11,513 genes into 17 modules, identifying four modules linked to the late-stage development of stamens and pistils in flower buds. Further experimental validation is required to confirm the key roles of these genes in regulating the flowering time of C. oleifera. Another promising avenue for future research is the in-depth analysis of the regulatory hormones and maturation pathways in different floral tissues, particularly stamens and pistils, with the goal of identifying effective treatments to induce flower bud maturation and mitigate the impact of climate change on the flowering time of perennial woody plants. Electronic supplementary material Below is the link to the electronic supplementary material. [158]Supplementary Material 1^ (260.5KB, docx) [159]Supplementary Material 2^ (99.1KB, xlsx) Author contributions H. S. and Z. D. contributed to the experimental design, performed plant phenotypic measurements, and analyzed the data. Y. W. and J. C. were involved in sample preparation. M. L. conceived the study, acquired funding, and revised the manuscript. L. J. revised the manuscript. X. W. and H. H. provided revision to the manuscript. All authors reviewed and approved the final submission. Funding This work is supported by National Natural Science Foundation of China (32101486), and Jiangxi Province Key R&D Project (20232BBF60021). Data availability The RNA-seq data presented in the study are deposited in the SRA repository in BioProject: PRJNA1140287 (https://www.ncbi.nlm.nih.gov/bioproject/PRJNA1140287), with accession numbers ranging from SRR29999737 to SRR29999790. Declarations Ethical approval and consent to participate The authors declare that the experimental research conducted on the plants described in this paper adheres to institutional, national, and international guidelines. The collection of fruits was carried out with full permission from Jiangxi Academy of Forestry National Camellia Seedling Base (115°28′E, 29°5′N). Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Huiyun Song and Zhihao Duan have contributed equally to this work. Contributor Information Liang Jin, Email: jinliang079@163.com. Mengfei Lin, Email: linmengfeilixi@163.com. References