Background

Ovules carry the female gametophyte, which develops into a seed after fertilization. This is indispensable for the formation of seeds and the normal growth of angiosperms [1]. In addition to being consumed by humans, seeds also have a high economic value, as they are used for the production of animal feed and raw chemical materials, and the research and development of medicine [2]. Crop yield often depends on the quantity and quality of the seeds, while the quality of each seed depends largely on the number, size, and proper development of the ovules [3]. Although each ovule has the potential to be fertilized until it forms viable embryos, obstacles arising during sexual reproduction can prevent some ovules from develo**. The occurrence of ovule abortion may be affected by physiological, biochemical, embryological, environmental, and molecular factors, including male sterility [4], female sterility [5], poor pollination or fertilization [6], temperature [7], and water [8]. Consequently, understanding the mechanisms of ovule development is of great importance from both scientific and economic perspectives. The ovule development process is regulated by several factors, including starch metabolism and endogenous hormone regulation. Starch actively participates in the metabolic processes of embryonic development and regulates the dynamic balance of carbohydrate metabolism in plants. Moreover, starch is the most significant means of carbohydrate storage and therefore the primary energy store in some seeds, particularly in Castanea.

Starch metabolism is an essential part of seed development, and has a direct impact on grain yield and quality. Starch metabolism in plants is aided by a variety of complex and coordinated mechanisms governed by several enzymes [9, 10]. The key processes and pathways involved in starch metabolism in higher plants include starch synthesis and degradation. In storage tissues, sucrose is hydrolyzed into monomeric sugars by invertase (InVS) and sucrose synthase (SUSY) to form uridine diphosphate glucose (UDP-glucose) and fructose [11]. Subsequently, these products are transformed into glucose-1-phosphate (G-1-P) by UDP-glucose pyrophosphorylase. G-1-P forms adenosine diphosphate glucose (ADPG) through the action of ADP-glucose pyrophosphorylase (AGPase). ADPG is a direct precursor of starch synthesis and synthesizes amylose under the action of ADP-Glc pyrophosphorylase (AGPase) and granule-bound starch synthase (GBSS) [11, 12]. ADPG synergizes with AGPase, soluble starch synthase (SS), SBE (starch branching enzyme), and the starch debranching enzyme (DBE) to synthesize amylopectin. Synthesized starch is mainly degraded by alpha-amylases (AMY), beta-amylases, or through the phosphorylation pathway [13]. AMY hydrolyzes the alpha-(l,4) glycosidic bonds of amylose and amylopectin in the substrate molecule [14]. Beta-amylase cleaves α-(1,4) glycosidic bonds from the non-reducing ends of polysaccharides to generate maltose [18]. Gibberellins (GAs) regulate ovule development. In a previous study, treating flowers with an exogenous GA spray significantly increased the proportion of sweet cherries with nacreous ovules, prolonged the embryo sac life span, and increased the seed setting rate [19]. Abscisic acid (ABA) plays an important role in embryonic development, promoting the synthesis of storage proteins during seed maturity; consequently, a sharp increase in ABA is an important factor leading to embryo abortion [20]. Jasmonic acid (JA) regulates ovule senescence, and the JA signal response factor PbMYC2 can directly activate the transcription of senescence-related genes, thereby inducing the death of pear ovules. Salicylic acid (SA) also contributes to ovule abortion [40], have less starch accumulation in aborted ovules, possibly attributed to insufficient starch content that leads to ovary abortion. In the present study, the starch distribution in fertile ovules showed a sudden decrease at 49–56 DAP, while there was no starch distribution in the abortion ovules from 56 DAP. This contrasts with the regularity of carbohydrate changes in many plant ovules, such as those in rice, where the amount of starch in the grain increases along with the increasing sugar content from the early stages of growth [41]. It is speculated that this contradiction may be attributed to embryonic differentiation and development, accelerated cell division, starch utilization exceeding accumulation, and starch amount gradually decreasing or even disappearing [42]. It is also possible to convert starch into a form that is more stable for the development of stored ovules. A possible hypothesis supports the notion that in the critical stage of the formation of fertile ovules and abortive ovules, a large amount of storage material needs to be consumed, and starch is first used as the main nutrient for ovule growth and development [43]. Additionally, starch may be connected with the double fertilization and subsequent zygotic development of fertile ovules, which consume many nutrients, including starch [44].

The expression levels of essential enzymes involved in starch and sucrose metabolism were substantially linked with the changes in starch distribution in the C. henryi ovules. Ovule development is a complicated process that needs the coordination and cooperation of many critical factors, especially enzymes. All enzymes in the starch synthesis pathway play crucial roles. Three AGPase, three NPPS genes, one SBE gene, and one GBSS gene were highly expressed during the starch synthesis. The AGPase family showed the highest expression among the five samples. We believe that AGPase may be a key enzyme in starch synthesis. ADPG pyrophosphate acidulase isa key rate-limiting enzyme in the starch anabolic pathway and plays an influential role in the starch synthesis process of maize and various plants [45].

Starch can be broken down by various enzymes. In this study, BAM was an important starch-degrading enzyme in the single-seed setting process of C. henryi. Its expression level in abortive ovules is much higher than that of other gene families. In other plants, BAM has been confirmed to play an important role. Genome-wide analysis identified 9 BAM genes in Arabidopsis, 13 in maize, and 10 in rice. AtBAM4 acts as a chloroplast regulator in Arabidopsis, is potentially responsive to maltose concentration, and can fine-tune the rate of starch degradation [46, 47]. In potato, StBAM9 interacts with StBAM1 on the surface of the starch grains for further starch degradation [49]. In tobacco, the PrtBAM1 gene increases the amylase activity and promotes starch degradation [50].

Previous studies have shown that beta-amylase is the main degrading enzyme responsible for the hydrolysis of stored starch and the degradation of transition starch. Beta-amylase hydrolyzes starch to beta-limiting dextrin and beta-maltose [51, 52]. In the present study, the beta-amylase activity of fertile ovules decreased sharply at 56 DAP. It is speculated that the activity of the 49–56 DAP abortive ovules decreased rapidly due to the rapid decrease in the amount of starch and substrates for beta-amylase hydrolysis [53]. Interestingly, the beta-amylase activity was consistent with the starch distribution in fertile ovules, which gradually increased and decreased with the increasing and decreasing beta-amylase activity, respectively. Therefore, it is considered that critical enzymes, such as beta-amylase, are closely related to starch synthesis in fertile and abortive ovules, which are directly related to starch accumulation. However, the specific role of beta-amylase requires further exploration.

Hormonal regulation during single-seed formation

Endogenous hormones are an important class of plant growth regulators, and the interaction of a variety of endogenous hormones regulates the ovule development of plants. In this study, we mainly elucidate the regulatory mechanism of GAs, CKs, ABA, and auxin in the formation of single seeds of C. henryi. In addition, we further explain the relationship and interaction between single seed formation in C. henryi and the key genes involved in hormone biosynthesis and signal transduction (Fig. 9). Multiple studies have shown that GAs play an important role in ovule formation. In a previous study, grapes were sprayed with GA before flowering, and it was found that one of the ways through which GA halted seed development was by damaging the redox system in flowers or berries, which resulted in oxidative damage of seeds and affected the expression of genes related to embryo development [54]. Overexpression of GASA4 in Arabidopsis has a positive effect on seed size, weight, and yield [55]. In pea, mutations in the GA biosynthesis genes LH1 and LH2 greatly affect embryogenesis and lead to the production of sterile seeds [56]. Overexpression of the GA2OX gene in pea and Arabidopsis, which is responsible for inactivating active of GAs, resulted in seed abortion [57]. In this study, the content of gibberellin in fertile ovules was significantly higher than that in abortive ovules, and the content in fertile embryos of 63 DAP was as high as 467 ng/g. Among them, the content of GA4 is the highest, which may play a leading role in gibberellin that promotes the formation of single seed of C. henryi. The GA20X1 was significantly correlated with the gibberellin content during single seed formation in C. henryi. The GASA4 was highly expressed in aborted ovules.

Fig. 9
figure 9

A proposed model to explain endogenous hormone regulation during single seed formation in C. henryi. Genes in red font were upregulated in fertile ovules, and genes in black font were upregulated in abortive ovules

Full size image

Auxin is mainly involved in regulating cell elongation, differentiation, and division. Auxin has been shown to be involved in regulating the development of the embryo, endosperm, embryo, and seed coat before and after fertilization. The YUCCA gene family is a key gene in the process of IAA biosynthesis. In Arabidopsis, the auxin biosynthesis enzymes YUCCA10 and TARI are expressed in the endosperm [58]. PIN1 plays an important role in both ovule initiation and pistil development [59]. A strong loss-of-function allele of PIN1 induces flower formation without ovules and empty pistils with malformed styles [60]. Previous studies have shown that SAUR is involved in auxin-regulated cell expansion [61]. The expression level of the YUCCA gene family was not particularly high in the transcriptome of chestnut ovule. However, PIN1 was highly expressed in fertile ovules.

Increased pistil CK content can increase pistil morphology and ovule number. CK affects the expression of PIN1 by promoting the expression of SPL and AG genes, thereby affecting the development of pistil ovules [62]. In addition to CKX5, which is significantly expressed in abortive ovules, other CK genes are expressed in fertile and abortive ovules; however, their expression levels are not high. Interestingly, the content of auxin and cytocatenin reached the highest level at 49 DAP of fertile ovules, which may be in a critical period of growth and development, when cell division and elongation are accelerated.

ABA and GA play antagonistic roles in seed development [63]. Comparing the differential expression of genes at different stages of ovule development between seeded and seedless grapes, it was found that the genes related to ABA synthesis and transduction pathway are upregulated expression in 'seedless' grapes [64]. The content of ABA in aborted ovules was higher than that in fertile ovules, which was consistent with the expression levels of key differentially expressed genes involved in ABA synthesis. Similarly, the content of SA in abortive ovules was substantially higher than that in fertile ovules, and the results were consistent with the expression levels of key differentially expressed genes related to SA biosynthesis. SA may have the same effect as ABA, both of which promote the abortion of ovules and the formation of single seeds.

Conclusions

In this study, external morphological observations indicated that 49 DAP is a critical period for the formation of fertile and abortive ovules. We investigated the molecular mechanism involved in starch metabolism and the regulation of endogenous hormones in the ovules of C. henryi by a transcriptome analysis at three developmental stages. Critical genes related to starch metabolism were further confirmed by enzyme activity assays. We identified ten classes of 37 functional genes related to starch and sucrose metabolic pathways. Among these, beta-amylase may be a crucial starch-degrading enzyme, and the BAM gene may also be an important determinant of the regulation of starch degradation. The contents of auxin, CKs, GAs, and JA in fertile ovules were higher than those in abortive ovules. We also identified and mapped the DEGs associated with hormone synthesis and signal transduction during the formation of single seeds. In summary, we evaluated endogenous hormones and starch metabolism during single seed formation in C. henryi, providing a preliminary theoretical basis for explaining the reasons for single seed formation in C. henryi.

Methods

Plant materials and growth conditions

All methods were carried out in accordance with relevant guidelines. The Chinese chinquapin (Castanea henryi) cultivar ‘Huali 4’(** fruits were measured in the above six stages using a Vernier caliper (Mitutoyo 530, Japan). Twenty fruits were used to measure transverse and longitudinal diameters at each stage. Subsequently, the mean measures were calculated for each period. Using a stereomicroscope (Olympus SZX16, Japan), we photographed and observed the morphology and size of ovules and ovaries.

Observation of starch distribution

The distribution of starch in the ovules was observed using the paraffin section method combined with periodic acid-Schiff (PAS) staining. Previous research by our group found that the starch grains of C. henryi ovules were stained by PAS and their starch grains were purple-red (Supplementary Fig S1 A-C). An inverted microscope (Leica S/M432299, Germany) was used to observe the green fluorescence (460–550 nm). The imaging position of the starch grains was consistent with the position under non-fluorescence imaging, and its color formation effect was better than that of non-fluorescence imaging (Supplementary Fig S1 a–c). Thus, we used an inverted microscope with green fluorescence to photograph and observe ovule starch.

Measurements of hormone contents

In the present study, ovules with three representative stages (35 DAP, 49 DAP, 63 DAP) were selected for hormone extraction and determination following the methodology established in previous studies [65]. Phytohormones were extracted from approximately 0.5-mg (fresh weight) samples. The GAs, indoleacetic acid (IAA), ABA, c-Zeatin (cZ) and trans-zeatin riboside (tZ), isopentenyladenine (iP), and isopentenyladenosine (iPR) contents of the samples were determined using high performance liquid chromatography (HPLC). Hormone levels was measured using a high performance liquid chromatograph (LC-20AT) with a Hypersil BDS C18 chromatographic column. The flow rate was 1.0 mL·min−1.

RNA extraction and cDNA library construction

Total RNA was extracted from the ovules of the three developmental stages (35 DAP, 49 DAP, 63 DAP) of the C. henryi, namely ovule (O) at 35 DAP, fertile ovule (BO1), abortive ovule (SO1) at 49 DAP, fertile ovule (BO2), and aborted ovule (SO2) at 63 DAP. For the cDNA library construction and transcriptome sequencing, the total RNA was extracted using the Plant Total RNA Isolation Kit (Omega, China). We performed extraction with the Microto-Midi Total RNA Purification System according to the manufacturer’s instructions. RNA purity was checked using agarose gel electrophoresis and a NanoDrop ND1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA) [66]. The RNA integrity number (RIN) and rRNA ratio were used to assess RNA concentration and integrity using an Agilent 2100 Bioanalyzer (Santa Clara, CA, USA).

First, qualified samples were taken as total RNA and the fragmentation buffer was poured into mRNA containing oligo (dT) magnetic beads, to fragment the samples. Second, the fragmented mRNA was used as a template. We applied six base random primers to synthesize the first strand of cDNA, and a mixture buffer, dNTPs (A, U, G, C), RNase H, and DNA polymerase I were used to synthesize the second strand of the cDNA chain. Following this step, the cDNA was purified using magnetic beads and eluted using the EB buffer. The fragments were end-repaired and A-tailed [67]. Finally, by using PCR to purify and enrich double-stranded cDNA, a cDNA library was created. Completing the above steps indicated that the preparation of the entire library was complete.

Unigene assembly, annotation and analysis

On an Illumina NovaSeq 6000 Sequencing System, cDNA sequencing was completed using the paired-end sequencing method. We evaluated the measured raw read data. After the reads’ assembly, we obtained unigene samples of ovules, which were the basis for the gene structure annotation, functional annotation, and expression. Simultaneously, data generated through the RNA-Seq sequencing experiment were compared with the reference genome based on the species' reference genome and gene information. To determine the sequence orientation, we compared and annotated all detected unigene sequences with data samples that contained the NR, GO, KOG number, and KEGG databases [68,69,70].

Identification and analysis of differentially expressed genes

The differential expression analysis was performed using the EBSeq software [71]. To analyze the differences in gene expression between different samples, we introduced the FPKM values. The Poisson distribution (parameters: fold change 2 and adjusted p-value 0. 001) is used in the DEG seq method. The Q-values were calculated after the P-values were corrected. We identified genes with larger than twofold differences and a Q-value ≤ 0.05 and filtered them as significantly differentially expressed genes to increase DEGs accuracy [72, 73]. Because of the limitations of using the database for unigene annotation, a keyword search was performed to identify other differentially expressed genes related to starch metabolism.

Quantitative real time PCR (qPCR) analysis

To validate the RNA-seq results, nine genes were chosen. The HiScript II Q RT Super Mix for qPCR (+ g DNA wiper) Kit was used to generate the first cDNA after extracting total RNA from five samples subjected to RNA-seq, following the manufacturer’s instructions. Primer 5.0 (Premier, Canada) was used to create gene-specific primers, and the primer sequences are listed in Supplementary Table S4. In a total reaction volume of 20 L, ChamQ Universal SYBR qPCR Master Mix (10 L; Vazyme, Nan**g, China) was mixed with gene-specific primers, sterilized water, and synthesized cDNA. Quantitative PCR was performed using a CFX96 Real-Time System (Bio-Rad, Hercules, CA, USA) according to the manufacturer’s instructions. Because GAPC1 expression is stable throughout ovule development, the internal control gene GAPC1 was used as the reference gene. To acquire relative mRNA expression data, the data were analyzed using the 2−∆∆Ct method. The qRT-PCR analysis was performed with three biological and three technical replicates.

Determination of starch metabolism-related enzyme activities

Likewise, at six developmental stages (35, 42, 49, 56, 63, and 70 DAP), the ovules were removed from the fruits and immediately stored in liquid nitrogen. We collected various samples of ovules (1 g) at each stage, added the pre-cooled extraction buffer, and homogenized the samples and centrifuged them in an ice bath. The supernatant was collected as the enzyme extract, which was used for the assessment of the α-amylase, beta-amylase, adenosine glucose diphosphate pyrophosphorylase (AGP), and starch synthase (SUSY) activities. The activities of AMY and beta-amylase were homogenized by adding distilled water and extracted according to the method described by Duke et al. [74]. The activities of adenosine diphosphate glucose pyrophosphorylase (AGP) and starch synthase (SUSY) were determined using the Nakamura et al. Method [75]. The activity of each enzyme was measured and thrice.