Background

Tartary buckwheat (Fagopyrum tataricum Gaertn.) comprises twenty different species of extensively cultivated medicinal and edible crop species with excellent economic and nutritional values [1, 2]. Tartary buckwheat was domesticated in East Asia and is cultivated in Europe and North America [3]. Today, it is common in the Himalayan region and Southwest China, such as Sichuan Province. In 2016, global production was 2.3 million tons, with Russia and China accounting for 50% and 17% of the world’s total output, respectively (adopted from UN Food and Agriculture Organization, Corporate Statistical Database: http://www.fao.org/). Tartary buckwheat seeds are a rich source of four B vitamins, dietary fiber, protein and a variety of minerals, in which niacin, magnesium, manganese, phosphorus content is particularly high (adopted from USDA Food Composition Databases: https://ndb.nal.usda.gov/). Importantly, tartary buckwheat seeds are known for their high rutin content, which has been proven to be effective in preventing liver injury and especially inflammatory liver injury [4].

The development of most seeds can be divided into three stages: tissue differentiation, cell enlargement and mature dehydration. At the stage of tissue differentiation, a single-cell zygote forms a young embryo composed of cotyledons by cell division and differentiation [5, 6]. Seed development in higher plants is a very complex process involving many expression-dependent physiological changes and regulation of a large number of genes and phytohormones, and the development of tartary buckwheat seeds is no exception.

Although the regulatory networks of monocotyledonous plants (i.e. maize) seed development [12]. Based on the number of days postanthesis, the young fruit stage to maturity stage described here correspond to the globular embryo to the mature embryo in tartary buckwheat seeds.

Fig. 1
figure 1

Observation and measurement of the development of tartary buckwheat seeds. a Five stages of tartary buckwheat seeds development. Top row shows each longitudinal sections of tartary buckwheat seeds at different developmental stages. Second row shows individual seeds at different developmental stages. Third row shows schematic representation of developmental stages studied during tartary buckwheat seed ontogeny. 13 DPA: Tartary buckwheat seed at green fruit stage for transcriptome experiment; 19 DPA: Tartary buckwheat seed at discoloration stage for transcriptome experiment; 25 DPA: Tartary buckwheat seed at initial maturity stage for transcriptome experiment. DPA, days-post-anthesis. b The seed size and weight at different developmental stages. Error bars were obtained from five measurements. Small letter(s) above the bars indicate significant differences (α = 0.05, LSD) among treatments

Full size image

Phytohormone accumulation at different stages of seed development

In buckwheat seeds, the content of gibberellin (GA3) was at a high level during the young fruit stage and peaked at 13 DPA with a value of 109.76 μg g−1 fresh weight (FW). After a sharp reduction until the discoloration stage, the content of GA3 decreased to the lowest value of 25.28 μg g−1 FW. The second increase occurred before the initial maturity stage and reached a maximum of 117.34 μg g−1 FW at 19 DPA. Throughout the development process, the GA3 content appeared as a fluctuation curve (Fig. 2a). The level of indole-3-acetic acid (IAA) also appeared as a fluctuation curve. The IAA content was high at young fruit stage and decreased sharply at green fruit stage. The second increase occurred before the discoloration stage and reached a value of 20.17 μg g−1 FW at 19 DPA, after which it decreased again from this time to the initial maturity stage, with the lowest value of 5.56 μg g−1 FW. After a sharp reduction, it started to increase again to a content of 6.05 μg g−1 FW (Fig. 2b). The content of zeatin (ZT) was at a high level during the young fruit stage and reached a peak value of 9.08 μg g−1 FW, after which it decreased sharply before the initial maturity stage. The content of ZT decreased to the lowest value of 0.41 μg g−1 FW, but after a sharp reduction before the maturity stage, it started to rise again to a content of 0.98 μg g−1 FW. Throughout the development process, the ZT content remained at a low level (Fig. 2c). The abscisic acid (ABA) content increased sharply and reached a peak value of 124.93 μg g−1 FW at the maturity stage (Fig. 2d). The peak of these three hormones in the process of seed development appeared in the order of ZT, IAA, GA3, and ABA (Fig. 2e). Among the four phytohormones, their proportions were different at different stages of seed development (Fig. 2f).

Fig. 2
figure 2

Changes in the levels of GA3, IAA, ZT, and ABA during seed development in tartary buckwheat. a-d Changes of GA3, IAA, ZT, and ABA content during seed development in tartary buckwheat, respectively. e GA3 (red line), IAA (green line), ZT (black line), and ABA (blue line) content at different developmental stages. f Phytohormone ratio during seed development in tartary buckwheat. Error bars were obtained from three measurements for (a), (b), (c), (d) and (e). Small letter(s) above the bars indicate significant differences (α = 0.05, LSD) among treatments

Full size image

Nutrient accumulation at different stages of seed development

As shown in Fig. 3, the four following components were quantified, namely, rutin, starch, total proteins and soluble sugars. The content of rutin increased linearly after anthesis and notably reached a peak value of 12 055.78 μg g−1 dry weight (DW) at the mature stage (Fig. 3a). The content of starch rapidly increased during the transition to green fruit stage and reached a maximum of 167.40 mg g−1 DW, after which it decreased linearly to a content of 102.53 mg g−1 DW (Fig. 3b). The content of total protein gradually increased after anthesis, notably from the discoloration stage to the initial maturity stage, and reached a maximum of 4.64 mg g−1 FW at the initial maturity stage, after which it started to drop rapidly to a content of 4.40 mg g−1 FW (Fig. 3c). The soluble sugar linearly increased from the young fruit stage to the discoloration stage and reached a peak value of 0.21 mg g−1 DW at the discoloration stage. After a slight drop, it started to rise again to a content of 0.18 mg g−1 DW (Fig. 3d). Among the four nutrients, starch dominated with relatively high proportions at the different stages of seed development (Fig. 3e).

Fig. 3
figure 3

Changes in the levels of nutrients at different development stages of tartary buckwheat seed. a-d Changes of rutin, starch, total protein, and soluble sugar content during seed development in tartary buckwheat, respectively. e Nutrients ratio during seed development in Tartary buckwheat. The error bars were obtained from five measurements. Small letter(s) above the bars indicate significant differences (α = 0.05, LSD) among treatments

Full size image

Correlation between physiological and developmental changes in tartary buckwheat seeds

The linear relationship between seed maturity and physiological changes were analyzed. The results showed that total protein, ZT and ABA were significantly correlated with seed maturity, and the absolute values of the correlation coefficients were all greater than 0.94 (Fig. 4). Seed size (transverse diameter × longitudinal diameter) was positively related to starch, with a correlation coefficient value of 0.906. Longitudinal diameter was significantly related to starch, and the correlation coefficient was 0.972. Seed weight was positively related to soluble sugar, with a correlation coefficient value of 0.917 (Fig. 4). Total protein was significantly related to ABA and ZT, and the absolute values of the correlation coefficients were greater than 0.93. The total protein was positively related to ABA and was negatively related to ZT. ZT was negatively related to ABA, and the absolute value of the correlation coefficient was 0.922. However, rutin, GA3, IAA, seed size, seed weight and DPA were not significantly correlated (Fig. 4).

Fig. 4
figure 4

The correlation between physiological changes and seed maturity during seed development (7 DPA to 30 DPA). Digit: Pearson’s correlation coefficient; Red: positively correlated; Blue: negatively correlated. * and ** indicate significant correlation at 0.05 and 0.01 levels, respectively

Full size image

Transcriptomic profiles during seed development

To comprehensively assess the change in the transcript profile, we analyzed RNAs from three seed developmental stages (13, 19, and 25 DPA). The raw reads of 9 samples ranged from 51 to 63 million, with an error rate of approximately 0.02, resulting in 7.61–8.84 G of clean bases (Additional file 1: Table S1). In exon region, intron region and intergenic region, the reading rates of map** were 97.4, 0.68, and 1.93%, respectively (Fig. 5a). Approximately 91.94% of total reads were mapped to the reference genome. Among them, the unique map reads accounted for 86.0% of the total reads, and multiple map reads accounted for 5.94% (Fig. 5b, Additional file 1: Table S2). The overall quality of the RNA-seq data was evaluated by correlation analysis between samples and principal component analysis (PCA) of biological replicates (Fig. 5d). The correlation coefficients between all biological replicate samples were greater than 0.90 (Fig. 5c, Additional file 1: Table S3). Three biological replicates, especially at 19 DPA, were significantly clustered, indicating that the expression patterns in the sample had high similarity and that the sequencing data could be used to analyze the differential expression of the genes. The expression levels of the genes were estimated by the fragments per kilobase of transcripts per million mapped fragments (FPKM) [13]. The level of gene expression showed a similar trend among the three samples, and the expression level of most genes was relatively low (Fig. 5e).

Fig. 5
figure 5

Distribution and quality of reads in different samples. a Distribution of overall mapped reads for all samples in the different regions. b Overall results of paired-end reads (PEs) for all samples mapped to the reference genome. c Correlation between samples. d Principal component analysis (PCA) among biological replicates. e Gene expression distribution (FPKM) in three samples

Full size image

Analysis of differentially expressed genes during seed development

To understand gene expression during tartary buckwheat seed development (13, 19, and 25 DPA), we compared and analyzed the DEGs in two pairs of samples (13 DPA vs 19 DPA, 19 DPA vs 25 DPA). A total of 9 765 DEGs were divided into 6 groups (Fig. 6a). Similarly, they could be divided into two broad categories according to their expression patterns. The first category contained groups III, IV and VI, with upregulated genes. The second category contained groups I, II, and V, in which the genes showed downregulation patterns. In detail, groups III and IV showed a gradual upward trend during the seed development stage. The genes in group VI were upregulated at 25 DPA compared with those at 13 DPA and 19 DPA. Groups I, II and V showed a decreasing trend at the stage of seed development (Fig. 6b). We found significant differences in gene expression during seed development and detected 1 529 DEGs (791 downregulated and 738 upregulated) at 13 DPA vs 19 DPA and 5 998 DEGs (2 337 downregulated and 3 661 upregulated) at 19 DPA vs 25 DPA (Fig. 6c). Besides, the expression profiles of 20 genes in the 13DPA, 19DPA, and 25 DPA, based on RT-qPCR, were consistent with the RNA-seq results with a Pearson correlation coefficient of 0.31 (P < 0.05) (Additional file 1: Table S6). To characterize the expression pattern of genes during tartary buckwheat seed development, we analyzed the overlap** DEGs in different sample combinations and measured 665 DEGs in the three samples (Fig. 6d); in the comparisons of 13 DPA vs 19 DPA and 19 DPA vs 25 DPA, there were 864 and 5 244 DEGs, respectively (Fig. 6d).

Fig. 6
figure 6

DEGs in three samples. a Heat map of scaled FPKM values in three samples 13 DPA, 19 DPA, and 25 DPA. Red: high expression; Blue: low expression. b Expression profile of six clusters correspondance to the Hierarchical cluster result. c Number of up- and down-regulated DEGs in two sample pairs 13 DPA vs 19 DPA, and 19 DPA vs 25 DPA. d Venn diagram of the numbers of expressed genes in sample pairs 13 DPA vs 19 DPA, and 19 DPA) vs 25 DPA

Full size image

Expression patterns of DEGs related to phytohormones during seed development

Phytohormones affect gene expression and transcription levels, cellular division, and plant growth [14]. Therefore, we identified the DEGs related to phytohormones. The results showed that 277 genes were associated with four major phytohormones, including ABA (35 genes), IAA (185 genes), ZT (20 genes) and GA3 (37 genes) (Additional file 1: Table S4). These genes are related to all aspects of phytohormone homeostasis, including biosynthesis (33 genes), metabolism (15 genes), receptors (26 genes), responses (3 genes), signal transduction (178 genes) and transportation (22 genes) (Table 1). The number of them found in a complex regulatory network of seed development are presented on Fig. 7a. We detected 17 DEGs (9 downregulated DEGs and 8 upregulated DEGs) in 13 DPA vs 19 DPA and 57 DEGs (16 downregulated DEGs and 41 upregulated DEGs) in 19 DPA vs 25 DPA (Fig. 7i). In 13 DPA vs 19 DPA, the DEGs were mostly enriched in 4 KEGG pathways in the KEGG database (False discovery rate (FDR) < 0.05) (Fig. 7b). However, a Gene Ontology (GO) analysis indicated that DEGs were not significantly enriched (FDR < 0.05) (Fig. 7d). In 19 DPA vs 25 DPA, the DEGs were mostly enriched in 6 KEGG pathways in the KEGG database (FDR < 0.05) (Fig. 7c). A GO analysis indicated that 27 terms involving biological process (Fig. 7e), 9 terms involving cellular component (Fig. 7f), and 3 terms involving molecular function (Fig. 7h) were significantly enriched (FDR < 0.05) (Fig. 7e).

Table 1 Genes related to phytohormones during seed development of tartary buckwheat
Full size table
Fig. 7
figure 7

DEGs related to phytohormones in three samples. a Hierarchical cluster of the DEGs related to phytohormones in X2 (13 DPA), X3 (19 DPA), and X4 (25 DPA). Red: high expression; Blue: low expression. b Scatterplot of KEGG pathway enrichment in X2 (13 DPA) vs X3 (19 DPA) (FDR< 0.05). c Scatterplot of KEGG pathway enrichment in X3 (19 DPA) vs X4 (25 DPA) (FDR< 0.05). Rich factor is the ratio of the number of DEGs to the number of background genes in a KEGG pathway. d GO classification of DEGs in X2 (13 DPA) vs X3 (19 DPA) (FDR< 0.05). e GO classification of DEGs in X3 (19 DPA) vs X4 (25 DPA) (FDR< 0.05). The top 30 enriched GO classifications are listed. Stars above bars indicate the amounts of differentially expressed genes are significantly higher or lower than the amounts of genes in random samples from the GO classification of all genes. f DEGs related to phytohormones enriched biological processes in X3 (19 DPA) vs X4 (25 DPA). g DEGs related to phytohormones enriched cellular component in X3 (19 DPA) vs X4 (25 DPA). h DEGs related to phytohormones enriched molecular function in X3 (19 DPA) vs X4 (25 DPA). The different color frames indicate the extent of significance. Yellow: significant; Red: extremely significant. i Number of up- and down-regulated DEGs in two sample pairs X2 (13 DPA) vs X3 (19 DPA), and X3 (19 DPA) vs X4 (25 DPA)

Full size image

Expression patterns of DEGs related to flavonoids during seed development

Tartary buckwheat is rich in flavonoids, including rutin and other flavonoids [15]. As a result, 47 genes were annotated as related to flavonoids. Among these genes, the DEGs included 3 phenylalanine ammonia-lyase (PAL), 4 chalcone synthase (CHS), 2 chalcone isomerase (CHI), 8 flavonol synthase (FLS), 6 flavonoid 3',5'-hydroxylase, 14 UDP-glucose flavonoid, and 10 flavonoid 3'-monooxygenase (Additional file 1: Table S5). According to their level of expression by developmental stage, they have at least one highly expressed member at all developmental stages. In 13 DPA vs 19 DPA, the DEGs were mostly enriched in phenylalanine metabolism pathways (ath00360) in the KEGG database (FDR < 0.05) (Additional file 1: Figure S1b), including 1 DEG (Table 2). In 19 DPA vs 25 DPA, 4 DEGs were mostly enriched in flavonoid biosynthesis pathways (ath00941) in the KEGG database (FDR < 0.05) (Additional file 1: Figure S1c, Table 2). These DEGs showed a higher expression in the flavonoid biosynthesis pathway from 13 DPA to 19 DPA (Additional file 1: Figure S1a).

Table 2 Differentially expressed genes related to in flavonoid biosynthesis pathways during seed development of tartary buckwheat
Full size table

Expression patterns of DEGs related to starch during seed development

Starch is among the main nutrients in tartary buckwheat and dominates with relatively high proportions [16]. We identified 20 genes related to starch, including 5 soluble starch synthase, 9 starch synthase, and 6 granule-bound starch synthase genes, all of which belong to the glycosyltransferase 1 family (Table 3). In 13 DPA vs 19 DPA, no genes were differentially expressed (FDR < 0.05). In 19 DPA vs 25 DPA, 3 DEGs were involved in starch biosynthesis (Additional file 1: Figure S2).

Table 3 Genes related to starch synthesis (or metabolism) during seed development
Full size table

Expression patterns of DEGs related to storage proteins during seed development

On the basis of differential solubility, storage proteins include globulin, albumin, glutelin, and prolamin [17]. As such, of the 21 genes, 18 are DEGs related to storage proteins, including 16 related to globulin and 2 related to glutelin (Additional file 1: Figure S3a). In 13 DPA vs 19 DPA, a GO analysis showed that the DEGs were significantly enriched in the molecular function term “nutrient reservoir activity” (GO:0045735) (Additional file 1: Figure S3b, c). In 19 DPA vs 25 DPA, the GO analysis showed that the DEGs were significantly enriched in the molecular function terms including nutrient reservoir activity (GO:0045735) and acireductone dioxygenase [iron(II)-requiring] activity (GO:0010309) (Additional file 1: Figure S3d, e).

Discussion

Correlation analysis between embryo cell enlargement and phytohormones during the development of buckwheat seed

The role of auxin in cell enlargement and the development of plant organs have been fully confirmed [18, 19], and the seed is no exception. In our study, the content of IAA decreased gradually as a fluctuation curve during the development of buckwheat seed (Fig. 2b), and we identified 18 DEGs in transcriptome; these DEGs were enriched in the plant hormone signal transduction pathway (ath04075) and may be related to cell enlargement (Fig. 8m). Most of these DEGs were downregulated from 13 DPA to 25 DPA. FtPinG0007643300.01, FtPinG0003888700.01, and FtPinG0006908400.01 were matched to the amino acid/polyamine transporter 2 family, whose members are located on the plasma membrane (Fig. 8l) and are involved in proton-driven auxin influx. In addition, 7 DEGs were matched to the Aux/IAA family and located in the nucleus. Aux/IAA proteins are short-lived transcription factors and act as inhibitors of early auxin response genes at low concentrations of auxin [20]. Novel02020, FtPinG0000807700.01, and FtPinG0008443000.01 were matched to the ARF family and are located in the nucleus (Fig. 8l). Auxin response factors (ARFs) are transcription factors that specifically bind to the DNA sequence 5'-TGTCTC-3' found in auxin-responsive promoter elements (AuxREs) [21,22,23]. Five DEGs were matched to the ARG7 family whose members are located in the nucleus, and these DEGs may play a role in apical hook development [24]. According to our previous research results and related research, during the development of buckwheat seeds we hypothesize that the extent of seed embryo cell enlargement gradually decreased, and the final embryo cell size was no longer increased and tended to stabilize. To test this hypothesis, we compared the size of seed embryo cells from 7 DPA to 30 DPA. However, the extent of seed embryo cell enlargement gradually increased (Fig. 8k). In fact, the decrease in auxin concentration may not be enough to stop cell enlargement. Auxin plays a role in initiation or acts as a ‘on’ switch via secondary regulators. Only lowering the auxin concentration (that is, removing the ‘on’ switch) does not reduce other phytohormones. Considering this hypothesis, a second ‘on’ switch may be required to accurately determine the time frame for cell enlargement [25,26,27,28,29,30]. It was found that the content of ABA is different in embryogenic cell clusters, globular embryos, torpedo embryos, and cotyledon embryos [31]. During embryonic development, ABA affects starch biosynthesis and carbohydrate absorption and directly regulates or initiates the biosynthesis of protein [32], DNA, and mRNA [33], all of which have an effect on the enlargement of embryonic cells [34]. According to these reports, embryonic cells continuously biosynthesize and absorb nutrients, resulting in an enlarged size under the influence of ABA. This fact may reinforce the idea of ABA as the second switch for cell enlargement. Therefore, we speculate that the increase in and prolongation of ABA content from 7 DPA to 30 DPA may be related to the enlargement of seed embryo cells. In the end, via the content of ABA, we confirmed that the content increased during the development of buckwheat seeds, which is consistent with the increase in seed embryo cells (Fig. 8k).

Fig. 8
figure 8

Enlargement of embryo cell during tartary buckwheat seed development. a, c, e, g and i Microscopic longitudinal sections of Tartary buckwheat seeds at 7, 13, 19, 25 and 30 DPA, respectively. b, d, f, h and j are the enlarged view of the boxes in (a), (c), (e), (g) and (i), respectively. SC (Seed cover). EM (Embryo). SL (Seminal leaf). The arrow represents the cell being measured. Bars = 100μm. (k) Embryo cell size at 7, 13, 19, 25 and 30 DPA. Error bars were obtained from five measurements. Small letter(s) above the bars indicate significant differences (α = 0.05, LSD) among treatments. (l) Subcellular location of DEGs related to embryo cell enlargement. This section provides information on the location in the cell (Graphics by Christian Stolte). m A simplified representation of the plant hormone signal transduction pathway of cell enlargement (adopted from the KEGG PATHWAY Database: http://www.genome.jp/kegg/pathway.html) shows the following: auxin influx carrier (AUX1), transport inhibitor response 1 (TIR1), auxin-responsive protein IAA (AUX/IAA), auxin response factor (ARF), auxin responsive GH3 gene family (CH3), and SAUR family protein (SAUR). The expression value of each gene is colored in log10(FPKM) in three samples X2 (13 DPA), X3 (19 DPA), and X4 (25 DPA)

Full size image

Correlation analysis between dormancy and phytohormones during the development of buckwheat seed

Dormancy occurs during seed development, and seeds are scattered from the parent plant in a dormant state [35]. ABA is a positive regulator of seed dormancy induction. Transgenic tobacco plants expressing anti-ABA antibody lack ABA, and the seeds are not dormant [36]. Overexpression of ABA biosynthesis genes can increase the content of ABA in seeds, thus promoting seed dormancy or delayed germination [37]. The biosynthesis of GA does not correlate with the establishment of primary dormancy [38]. However, GA3 plays important roles in the seed development of tomato, pea (Pisum sativum) and some Brassica plants, including fertilization, assimilation, embryonic growth, prevention of seed abortion, and fruit growth [39, 40]. A model of seed dormancy and germination was regulated by ABA and GA3, and the complex phenomenon was explained by the response to environmental factors. The model was affected not only by many genes but also by plant hormones and environmental factors [41, 42]. According to this model, the biosynthesis of ABA and the signal of GA3 catabolism determine the dormancy of the seed, and the biosynthesis of GA3 and the signal of ABA catabolism determine the transformation of seeds to begin germinating. Our data show that the content of ABA was significantly related to seed development (Fig. 4). The content of ABA increased with the development of buckwheat seeds, especially during 13 DPA to 25 DPA (Fig. 2d), when a large amount of storage material accumulated (Fig. 3). Moreover, only 2 DEGs (FtPinG0002387400.01 and FtPinG0003614600.01) that were involved in the oxidative degradation of ABA were downregulated during 19 DPA to 25 DPA (Fig. 7a). We also found the peak ABA/GA ratio occurred at 19 DPA (Fig. 9a). Based on the results we described earlier and those of related studies, we hypothesize that a large number of genes related to buckwheat seed dormancy are expressed from 19 DPA to 25 DPA. To test this hypothesis, we identified 11 transcriptome DEGs that were enriched in plant hormone signal transduction pathways (ath04075) and that were related to seed dormancy from 19 DPA to 25 DPA (Fig. 9c). Among them, the upregulated FtPinG0005799600.01 and FtPinG0002311400.01 belong to the PYR/PYL/RCAR abscisic acid intracellular receptor family, whose members are located on the plasma membrane and in the nucleus [43] (Fig. 9b). FtPinG0005799600.01 was matched to the abscisic acid receptor (PYL4) gene, and FtPinG0002311400.01 was matched to the abscisic acid receptor (PYL10) gene; both are ABA receptors required for ABA-mediated responses, and when activated by ABA, they inhibit the activity of group-A protein type 2C phosphatases (PP2Cs) [Measurement of phytohormone contents

Fresh samples that were weighed to approximately 0.5 g were ground in liquid nitrogen. The powder was subsequently homogenized in 10 mL of 80% methanol, after which the solution was stirred overnight at 4°C. This suspension was then centrifuged at 12 000 rpm for 10 min under refrigeration (4°C). The supernatant was collected, after which 5 mL of 80% methanol was added to the residue. Similarly, the supernatant was collected after centrifugation. The pooled supernatant (~15 mL) was flash evaporated at 36°C until the methanol vaporized (~3 mL). A bottle was washed with 5 mL of ultrapure water, and then the water was combined with the residual liquid (~3 mL). The solution was decolorized with 15 mL of diethyl ether three times, after which the ether phase was discarded. The aqueous phase was collected and basified to pH 8.0 with 0.1 M Na2HPO3. The basified extract was kept in a shaker for 30 min with 50 mg of polyvinylpyrrolidone (PVP) at 4°C. In addition, the extract was centrifuged at 12 000 rpm for 10 min. The supernatant was subsequently collected and was acidified to pH 3.0 with 0.2 M citric acid. The resultant solution was partitioned three times against 5 mL of ethylacetate, after which the aqueous phase was discarded. The pooled ethylacetate phase (~15 mL) was flash evaporated at 36°C to near dryness. Afterward, the residue was dissolved in 1 mL of methanol [65,66,67].

The sample was filtered through a nylon 66 filter (25 mm diameter, 0.45 μm pore size) prior to injection into a high-performance liquid chromatograph. HPLC analysis was performed on an Agilent 1260 system using a C18-ODS (3.5 μm × 150 mm × 4.6 mm) column (Agilent, USA) and a UV/VIS detector. An injection volume of 10 μL, column temperature of 35°C, flow rate of 1 mL min-1, and run time of 10 min were maintained for all analyses. The system was calibrated with external standards of IAA, GA3, ABA, and ZT. For detection, separation was performed with a mixture of methanol and distilled water containing 0.6% acetic acid (V:V = 50:50) following isocratic elution. The elutant was scanned at 257 nm.

Measurement of rutin contents

The seeds of tartary buckwheat were dried at 60°C for 4 h, milled into a fine powder, and filtered through a 0.177 mm sieve. A total of 0.10 g of dry powder for each tissue was mixed with 1 mL of methanol and incubated at 60°C using an ultrasonic bath for 2 h. The mixture was then centrifuged at 4 000 r min-1 for 10 min, after which the supernatant was collected. The supernatant’s pH value was adjusted to 6 using 0.1 mol L-1 phosphoric acid and then filtered through 0.22 μm membrane filter. High-performance liquid chromatography was performed on an Agilent 1260 system using an Agilent C18 column (4.6 mm × 150 mm, 5 μm) (Agilent Technologies, USA) at 30°C. The mobile phase consisted of methanol:0.5% phosphoric acid (V/V) = 40:60 at flow rate of 1 mL min-1 with a total injection volume of 10 μL. The rutin quantity was estimated based on the linear calibration curve of standard rutin hydrate (98%) (Yuanye, China) at a detection wavelength of 257 nm. Three independent sample analyses were performed for each tissue [68].

Measurement of nutrient contents

The contents of total soluble sugars in buckwheat seeds at different developmental periods were measured by anthrone-H2SO4 colorimetry. Ground sample tissue (0.07 g) and 7 mL of distilled water were put in a 10 mL graduated test tube that was closed with a stopper. The mixture was then extracted in a water bath for 30 min at 100°C and centrifuged at 4 000 × g for 5 min, after which the supernatant was transferred to a 50 mL volumetric flask (extracted twice). Afterward, 100 μL of the extract was pipetted into the test tube, and 3 ml of anthrone reagent was added, after which the tube was placed in a boiling water bath for 10 min. The change in absorbance at 620 nm was then measured, after which the soluble sugar content was calculated according to the standard curve [69].

The total protein content was measured using the Coomassie brilliant blue G-250 method. A standard protein solution and a Coomassie brilliant blue G-250 solution were prepared, and fresh seeds (0.5 g) were ground in 10 mL of distilled water using a mortar and pestle. After the solution was centrifuged at 4000 × g for 10 min, the supernatant was transferred to a new tube to assay the total protein content, and the change in absorbance at 595 nm was measured [69].

The contents of starch in the seeds of buckwheat were measured by anthrone-H2SO4 colorimetry [70]. The solid residues that extract soluble sugars were used to test the starch content. First, 20 mL of hot water was added to the tube that contained solid residues. An extraction in a water bath (100°C) for 15 min then occurred, after which 2 mL of 9.2 mol L−1 perchloric acid was added; the water bath (100°C) extraction then continued for 15 min. After cooling and mixing the extract, the filtrate was brought to a constant volume in a 50 mL measuring flask. Afterward, 100 μL of the extract was pipetted into the test tube, and 3 mL of anthrone reagent was added. The extract was then placed in a boiling water bath for 10 min. The change in absorbance at 620 nm was then measured, and the starch content was calculated according to the standard curve [69].

Statistical analysis

All the data were analyzed by analysis of variance using the Origin Pro 2017 statistics program, and the means were compared by the least significant difference test (LSD) at the 0.05 level of significance. The Pearson's correlation coefficient between DPA, rutin, starch, total proteins, soluble sugars, ZT, GA3, IAA, ABA, transverse diameter, longitudinal diameter, transverse diameter × longitudinal diameter, and seed weight were analyzed by Origin Pro 2017 software.

Total RNA extraction, cDNA Library construction, and high-throughput sequencing

The total RNA samples were extracted using the TRIzol reagent (Invitrogen) and were then treated by RNase-free DNase I (Takara) to remove the genomic DNA. mRNA libraries were constructed according to the standard protocols provided by Illumina. The quality of mRNA, including purity, quantity and integrity, was tested using Nanodrop, Qubit, and Agilent 2100 systems. The mRNA was enriched using Dynabeads oligos (dTs) (Dynal; Invitrogen) and fragmented using fragmentation buffer. Double-stranded cDNAs were synthesized using both reverse transcriptase (Superscript II; Invitrogen) and random hexamer primers and were further purified using AMPure XP beads. Finally, the purified double-stranded cDNA samples were further enriched by PCR to construct the final cDNA libraries that were sequenced using a HiSeq 4000 (150 bp paired ends) by Novogene (China). All raw sequence read data were uploaded to the NCBI Sequence Read Archive (SRA, http://www.ncbi.nlm.nih.gov/Traces/sra) under accession number GSE111937.

Map sequencing results and differential expression analysis

Adaptor sequences and low-quality sequences were removed from the raw reads (Q < 20). The clean reads were aligned to the reference genome sequences of the tartary buckwheat (Pinku1) genome (http://www.mbkbase.org/Pinku1/) using TopHat (v2.0.12). The default parameters were used, allowing mismatches of no more than two bases. A reference-based assembly of all the reads was performed using the Cufflinks v2.1.1 reference annotation-based transcript (RABT) assembly method. The assembled transcript fragments were compared with the reference annotation to predict new genes and novel exons and to optimize gene structures using Cuffcompare.

Gene expression differences in the different sample pairs were detected using the DESeq package (v1.10.1). In this study, to investigate the genes involved in tartary buckwheat seed development, the two sample pairs (13 DPA vs 19 DPA and 19 DPA vs 25 DPA) were used to investigate the differentially expressed genes under different conditions. The P-value threshold was determined using the FDR for multiple tests [71]. The thresholds were set using an FDR ≤ 0.05 and the absolute value of log2 (fold change) with FPKM ≥ 1 to determine significant differences in gene expression. The FPKM was used to eliminate the influence of different gene lengths and sequencing discrepancies on the quantification of gene expression to ensure direct comparison of gene expression between different sample pairs [72].

Functional classification of differentially expressed genes

An analysis of functional enrichment, including GO, was performed to identify which DEGs were significantly enriched in GO terms. GO enrichment of the DEGs was conducted using the GOseq R package (release 2.12). GO terms with a corrected P-value < 0.05 were considered significantly enriched by the differentially expressed genes. The GO annotations were functionally classified using WEGO software for gene function distributions. KOBAS software (v2.0) was used to identify the statistical enrichment of the differentially expressed genes in the KEGG pathways. The pathways with an FDR value of less than 0.05 were considered those genes showing significantly differential expression.

Real-Time PCR confirmation of differentially expressed genes

Quantitative real-time PCR analysis was performed to confirm transcriptome results. The corresponding sequences of these genes were obtained from the tartary buckwheat (Pinku1) genome sequence database. The RT-qPCR primers were designed according to the transcript sequences of 20 genes using Primer3 software (http://frodo.wi.mit.edu/) (Additional file 1: Table S7). The FtH3 gene was used as the internal control. RT-qPCR experiments were replicated at least three times.

First-strand cDNA was synthesized from 1 mg of DNase I-treated RNA samples in a 40 μL reaction solution with random primers, using a PrimeScript RT Reagent Kit with gDNA Eraser (TaKaRa). Standard RT–qPCR was performed using SYBR Premix Ex Taq II (TaKaRa) on a CFX96 Real Time System (BioRad). Data were analyzed by the 2−(∆∆Ct) method to obtain relative mRNA expression data [73].