Transcriptome and Metabolome Jointly Revealed the Regulation and Pathway of Flower and Pod Abscission Caused by Shading in Soybean (Glycine max L.)

Abstract

Flowers and pod abscission significantly reduces soybean yield. This study aims to identify the main signaling pathways and key candidate genes in soybean leaves that affect flower and pod abscission under shade conditions. This information will be useful for the localization and cloning of genes related to abscission. Two soybean cultivars with different abscission rates (Liaodou 32 and Shennongdou 28) were used in this experiment. The soybean leaves were subjected to 50% shading treatment and the transcriptome and metabolome was sequenced during the light-sensitive period. The effects of weak light at different growth stages on the metabolic pathways of soybean leaves and organ abscission were investigated by analyzing plant phenotype and physiological changes. The results showed that both two cultivars triggered the same molecular mechanism and similar metabolite accumulation mode by shading, but the regulations of the two cultivars were different. The key candidate genes identified for soybean flower and pod abscission caused by shading were DIV, MYB06, MYB44, MY1R1, MYBS3, WRKY6, WRK53, WRK70, WRK40, DOF14, CDF3, CDF2, GATA5, DREB3 and ERF3; the differentially expressed genes that caused the variation between the cultivars were SRM1, MYB16, WRK24, COL16, MYB61 and TRB1. The main metabolic pathways involved in soybean flower and pod abscission caused by shading were secondary metabolite biosynthesis pathway, metabolic pathway, cofactor biosynthesis pathway, phenylpropanoid biosynthesis pathway, flavonoid biosynthesis pathway, fatty acid biosynthesis pathway and amino acid biosynthesis pathway; the DEMs that caused the differences among the cultivars were carbon metabolism, glutathione metabolism, biotin metabolism, nucleotide metabolism, purine metabolism, plant hormone signaling, flavone and flavonol biosynthesis, lysine degradation, arginine and proline metabolism, amino sugars and nucleotide sugars metabolism, etc. In conclusion, shading treatment in the light-sensitive period of soybean changed the physiological response and gene expression level of leaves, inhibited carbohydrate metabolism and transport and biosynthesis of secondary metabolites, and thereby leading to increased competition and hormonal disruption, which promoted the abscission of soybean flowers/pods and reduced grain yield.

1. Introduction

Organ abscission is a key process of plant growth and development, which is regulated by environmental factors, physiological and biochemical metabolism, gene expression and other factors [1,2,3,4]. Light is one of the important environmental factors regulating the development of plant organs. In the process of plant development, leaves always sense changes in light and make corresponding adjustments [5]. Abnormal loss of crop organs in response to environmental dynamics, especially the abscission of flowers, fruits and seeds, is a direct factor that causes economic losses and is also one of the important factors that restrict the high yield of crops. During the growth and development of soybean, a large number of flowers will be produced, but the abscission rate is 30–80%, and the abscission of flowers and pods is the main factor limiting the increase in soybean yield, restricting its genetic potential and effective cultivation control measures.

In actual production, dense planting of soybeans often leads to a weak light environment inside the population, which inhibits the photosynthesis of leaves. When the photosynthetic rate was below a certain threshold, plants chose whether to shed part of the flower and pod organs to reduce competition and with sufficient energy to ensure full ripening of seeds [6]. Guinn (1974) [7] studied and found that low light promoted the shedding of cotton buds and young fruits. Studies from Byers et al. [8,9] showed that weak light stress led to insufficient carbon storage, increased competition between vegetative organs and reproductive organs, and promoted flower and fruit abortion and shedding. How does the leaf transmit the perceived weak light signal to other tissues and organs, and how does the short time weak light signal cause the organ to fall off? The leaf is the main organ for sensing and receiving light signals. A weak light signal affected the accumulation of photosynthetic products, physiological metabolism and transcription functions of leaves, and produced a series of signal factors to regulate the growth and development of the whole plant. Jiang et al. [10] found through shading treatment of sorghum leaf that the weak light generated by shading significantly reduced the net photosynthetic capacity and stomatal conductance of leaf. Weak light mainly affected the photosynthetic capacity by inhibiting the photosynthetic electron transport and reducing the activity of carbon assimilation enzyme, so that the transport and metabolism of carbohydrates were blocked. Studies have confirmed that plants subjected to weak light stress in a specific period increased the competition between vegetative organs and reproductive organs, and promoted flower abortion and shedding. Meanwhile, weak light leads to changes in the osmotic potential inside and outside the cell, imbalance of active oxygen metabolism and membrane lipid peroxidation and a series of reactions producing carbon starvation and shedding signals. The leaf transmitted the perceived shedding signals to mature exzonal cells and started the shedding process. Li et al. [11] found that maize leaves first sensed the stress signal, and then a series of physiological and biochemical reactions occurred and regulated the expression of related genes, thus generating tolerance.

Shedding of plants under stress was regarded as a self-protection mechanism of plants. There was a critical period when weak light affected organ shedding (light sensitive period). In the light sensitive period, a short period of low light stress will lead to disproportionate shedding of organs. The 6–9 stages of tomato flower development were the most sensitive stages to low light, shading of apples for a short time after flowering lead to the shedding of reproductive organs [12]. Giulia et al. [13] found in their research on induced early apple fruit shedding that the contents of isoprene and abscisic acid increased. Under low light conditions, hormone balance was disturbed, auxin transport was blocked, and abscisic acid, as a sensor of nutrient deficiency intensity, cooperated with other hormones and secondary signals, resulting in a rapid abscisic reaction. Transcriptomic sequencing showed that the genes involved in shedding were MADS-box, AP2/ERF, ARF, AUX/IAA, bZIP, WRKY, MYB and some ethylene target genes (ACC, PP2Cs, MAT1) and other transcription factors, most of these transcription factors were enriched in secondary metabolite pathways, including flavonoids, terpenes, amino acids and their derivatives, sphingolipids and other secondary metabolites.

Previous studies mostly focused on the organ abscission influence of light intensity or shade duration, but there were few reports on the effects of weak light at different growth stages on the internal metabolic pathways of crop leaves and organ abscission. Based on our previous studies on flower and pod abscission rate of different soybean cultivars, Liao Dou 32 and Shennong Dou 28, which were selected for different pod abscission rate, were selected as test materials in this paper. The objectives of this study were: (a) To confirm the abscission of flowers and pod of different soybean cultivars under shade conditions; (b) To analyze the effects of shading on physiology, biochemistry, transcription factors and metabolites of soybean leaves at different growth stages; (c) To screen the main signaling pathways and key candidate genes related to flower and pod abscission.

2. Materials and Methods

2.1. Experiment Design

The experiment site was selected in the pot experiment field (41°82′ N, 123°57′ E) of the Scientific Experimental Base of Shenyang Agricultural University, Shenyang, Liaoning Province. Two soybean varieties, Liaodou 32 (32, maturity 128 days, more flowers and less abscission) and Shennongdou 28 (28, maturity 130 days, more flowers and more abscission), were selected by previous experiments with similar maturity and total number of flowers with different flowers and pod abscission rates. Seeds of uniform size were selected and sown into a plastic bucket containing 15 kg of soil (diameter × height 30 × 35 cm). The nutrient content of the soil was organic matter 16.77 g/kg, total nitrogen 0.77 g/kg, available phosphorus 0.02 g/kg, available potassium 0.14 g/kg and pH 7.33. After emergence, one plant was kept in each pot when the true leaves were fully developed.

In order to better simulate the shading state between plants in actual production, according to the results of previous studies and preliminary tests, choose a 50% shade net. Shading (S) was performed at two leaf stage (V2), four leaf stage (V4) and first flowering stage (R1). Each variety was treated with 20 pots at a time and the same number of controls (C) were set up for a total of 240 pots. Measurements were taken and samples were taken on the 5th day of treatment and repeated three times. The fully unfolded leaves on the upper part of the plants treated with S and C were removed, the veins were removed, and the leaves were quickly frozen in liquid nitrogen and brought back, and stored in the refrigerator at −80 °C for subsequent index determination.

2.2. Test Method

2.2.1. Measurement of Dry Matter

In the three growth stages of V2, V4 and R1, 3 soybean plants were randomly selected to measure the dry matter mass of different tissues and organs in each growth stage by drying weighing method.

2.2.2. Measurement of Flower and Pod Abscission Rate and Agronomic Traits

Three soybean plants with the same growth were selected, and the plants were surrounded by white gauze before the beginning of flowering, with the bottom tightening in an inverted triangle shape, for collecting the dropped flowers and pods. The seeds were tested at maturity, and the number of main stem pods, the number of branch pods, and the number of branches were recorded. The sum of the number of pods per plant and the total number of flowers and pods shed was the total number of soybean flowering, and the related abscission ratio was calculated.

2.2.3. Transcriptome and Metabolome Analysis

Metabolome Analysis

Metabolites in soybean leaves with important biological significance and statistically significant differences under stress were detected and screened using ultra-high-performance liquid chromatography tandem mass spectrometry (3 biological replicates). After vacuum freeze-drying (Scientz-100F), soybean leaves were ground (MM400, Retsch, 30 Hz, 15 min) into powder, weighed 50 mg and added to 70% methanol extract solution of 1200 μL precooled at −20 °C, and Vortex 30 s/30 min, 6 times in total, centrifuged at 12,000 rpm for 3 min to extract the supernatant, filtered with a microporous filter membrane (0.22 μm pore size) and stored in the sample vial. Ultra Performance Liquid Chromatography (UPLC) and tandem mass spectrometry (MS/MS) were used for analysis. The software Analyst 1.6.3 was used to process the mass spectrum data, the metabolites were qualitatively identified according to the secondary spectrum information of the self-built database, and the metabolites were quantitatively analyzed and calibrated using the multi-reaction monitoring mode of the triple four-bar mass spectrometry. Metabolites were identified from the calibrated metabolome data and analyzed for quality control of the sample data, and metabolites with differences were screened out, and related functional prediction and analysis were conducted for these metabolites.

Transcriptome Analysis

Total RNA from soybean leaves was extracted using TRIzol (Invitrogen, Carlsbad, CA, USA), mRNA was enriched by Oligo (dT) magnetic beads, and double-stranded cDNA was synthesized using mRNA fragments as a template and purified by AMPure XP beads. Then, the end repair and A-tail connection were performed, and the cDNA library was constructed by PCA enrichment. Qubit 2.0 was used for preliminary quantification, Agilent 2100 was used for library insert size detection, Q-PCR method was used for accurate quantification, and the Illumina platform was used for sequencing. The sequencing data was filtered and sequence matched to obtain mapped data and perform variable splicing analysis. New gene discovery and gene structure optimization were performed for equal structural level analysis. Expression level analysis such as differential expression analysis, functional annotation and functional enrichment of differential expression genes were performed according to gene expression quantity.

2.2.4. Real-Time Fluorescence Quantitative PCR Detection

As with RNA-seq test samples, target genes were quantified using platinum ^®Taq DNA polymerase (Thermo Fisher Scientific, Waltham, MA, USA) enzymes and specific primers. The plasmid containing the target gene was diluted 10-fold to construct a standard curve, and real-time fluorescence quantification was performed on a PCR apparatus (7500, ABI, Thermo Fisher Scientific, Waltham, MA, USA). RT-qPCR analysis was performed using a 15 μL reaction system [14], including 3 μL 100 ng DNA, 1.2 μL 10 μM PCR positive primer, 1.2 μL 10 uM reverse primer, 7.5 μL SYBR qPCR mixture and 2.1 μL ddH₂O. All expression level data obtained by RT-qPCR had 3 biological replicates.

2.3. Data Analysis

Data were processed by Excel 2019, SPSS 22.0 statistical software for difference significance analysis, and Origin 2023 software for plotting.

3. Result

3.1. Effect of Shading on the Shedding Rate of Soybean Flower and Pod

Statistics are shown in

Table 1. Effects of shading treatment on flower pod shedding of different soybean varieties.

Cultivar	Shade Treatment	Pods Per Plant	Abscissed Flowers	Abscissed Pods	Total Abscission	Total Flowers and Pods	Total Abscission Rate (%)
32	CK	166 ± 3.61 Aab	13.33 ± 0.58 Cc	13.33 ± 1.53 Dd	26.67 ± 1.53 Dd	193.67 ± 3.51 Bb	14 ± 0.01 Fh
	V2	147.67 ± 2.08 Bc	14 ± 1.00 Cc	14 ± 1.00 Dd	28 ± 1.00 Dd	175.67 ± 1.53 Cc	16 ± 0.01 Fg
	V4	103.67 ± 2.52 De	15 ± 2.00 Cc	16 ± 1.00 Dd	31 ± 1.73 Dd	134.67 ± 2.89 Ee	23 ± 0.01 Ef
	R1	115 ± 2.65 Cd	33.33 ± 2.52 Bb	20.67 ± 1.53 Cc	54 ± 2.65 Cc	169 ± 2.65 CDcd	32 ± 0.01 Cc
28	CK	169.33 ± 2.08 Aa	32.33 ± 1.53 Bb	33.33 ± 1.15 Bb	65.67 ± 0.58 Bb	235 ± 2.00 Aa	28 ± 0.00 De
	V2	163.33 ± 2.08 Ab	34.33 ± 0.58 Bb	35 ± 1.73 Bb	69.33 ± 2.08 Bb	232.67 ± 2.08 Aa	30 ± 0.01 Dd
	V4	94.33 ± 2.52 Ef	37 ± 1.00 Bb	34 ± 1.00 Bb	71 ± 1.73 Bb	165.33 ± 1.15 Dd	43 ± 0.01 Bb
	R1	105.33 ± 3.51 de	83.67 ± 5.86 Aa	49.67 ± 3.06 Aa	133.33 ± 6.66 Aa	238.67 ± 10.07 Aa	56 ± 0.01 Aa

Note: V2, V4 and R1 indicate soybean two-leaf stage, four-leaf stage and initial flowering stage, respectively. The data in the table are repeated three times. Lowercase letters represent significant difference (p < 0.05), uppercase letters represent extremely significant difference (p < 0.01).

, the total flower and pod abscission rate of plants under shade treatment in V2 stage was only 0.01% different from CK. The total flower and pod abscission rate of shade treatment in R1 stage was significantly higher than that of other treatments. The total flower and pod abscission rate in V4 treatment was significantly higher than that in V2 and CK treatment, and the pod number per plant and total flower and pod numbers in V4 treatment were significantly lower than those in other treatments.

3.2. Plant Dry Matter Accumulation

As shown in Figure 1, there was no significant difference in dry matter accumulation between different treatments at V2 stage, and the dry matter accumulation in shade treatment in V4 and R1 stages was significantly lower than that in CK. There was no significant difference in dry matter accumulation between the two varieties under shade treatment at V4 stage, and the dry matter accumulation of Liaodou 32 under shade treatment at R1 stage was significantly higher than Shennong Dou 28. In summary, according to Table 1 and Figure 1, it can be determined that the V4 period is the light-sensitive period of soybean leaves, and the transcriptome and metabolome are used to further analyze the pathway of shading regulation of flower and pod shedding in soybean leaves at the V4 stage.

3.3. Effects of Shading Treatment on Transcriptome and Metabolome of Soybean Leaves

3.3.1. Differential Gene Analysis of Leaf Transcriptome of Soybean under Shade Treatment

In order to analyze the regulatory mechanism of flower and pod shedding genes in response to shade treatment in soybean leaves during light sensitive period, four cDNA libraries (C28-VS-S28, C28-VS-C32, C32-VS-S32 and S28-VS-S32) were pairwise compared to identify the differential genes (DEGs) among the shade treatment samples, the number of fragments read per kilobases per million (FPKM) for all DEGs is calculated. Under the control condition, the DEGs value between the two varieties reflected a difference of genetic background between the two varieties. In C28-VS-S28, 545 genes were activated, and in C32-VS-S32, only 291 genes were activated. Only 500 genes were activated in the C28-VS-C32 comparison. When transferred to shade treatment conditions, the number of DEGs increased between varieties, with 1680 genes activated in the S28-VS-S32 comparison and only 11 genes in common across the 4 treatments (Figure 2A). By comparing varieties 28 and 32 under shade treatment and control treatment, it was found that there were 1791 differentially expressed genes in C28-VS-C32 under control treatment (808 up-regulated and 983 down-regulated), and 3319 differentially expressed genes in S28-VS-S32 under shade treatment (1326 up-regulated and 1993 down-regulated). There were 1058 differentially expressed genes (696 up-regulated and 362 down-regulated) in the C28-VS-S28 comparison, and 629 differentially expressed genes (218 up-regulated and 411 down-regulated) in the C32-VS-S32 comparison (Figure 2B).

An amount of 2530 DEGs were identified under C28-VS-C32 treatment, involving 69 transcription factor (TF) families. A total of 2546 DEGs were identified under S28-VS-S32 treatment, 2575 DEGs under C28-VS-S28 treatment, and 2609 DEGs under C32-VS-S32 treatment, all involving 70 TF families. The number of DEGs was up-regulated in shade treatment and low-light-insensitive varieties 32. According to the gene expression of the top 20 TF families, the top 5 TF families of variety 28 and 32 were MYB (190; 185), bHLH (153; 157), C2C2 (129; 156), HB (127; 132) and AP2/ERF (125; 122), the largest TF families of the two varieties DEGs are C2C2 and WRKY, and the TF families of variety 28 higher than 32 are MYB and AP2/ERF (Figure 2C).

In order to better understand the function of DEGs, the 20 DEGs pathways with the most significant enrichment in each treatment were graphically displayed (Figure 3). In the C28-VS-S28 process, the main enrichment is in fatty acid biosynthesis (ko00061), metabolic pathway (ko01100), secondary metabolite biosynthesis (ko01110), fatty acid metabolism (ko01212), ascorbic acid and uronic acid metabolism (ko00053), glutathione metabolism (ko00480), biotin metabolism (ko00780) propionate metabolism (ko00640), pyruvate metabolism (ko00620), ribosome (ko03010), glycolysis/gluconogenesis (ko00010), and cofactor biosynthesis (ko01240) pathways (Figure 3A). Under C32-VS-S32 processing, Metabolic pathways (ko01100), biosynthesis of flavonoids and flavonols (ko00944), interconversion of pentose and glucuronate (ko00040), glycolide metabolism (ko00561), biosynthesis of secondary metabolites (ko01110) and anthocyanin biosynthesis (ko00942) are the main enrichment pathways (Figure 3B). The comparison of DEGs enrichment pathways of the same varieties under the same shade treatment showed that varieties 28 and 32 were significantly enriched in metabolic pathways, biosynthesis of secondary metabolites, glucose metabolism and antioxidant pathways. These results indicated that varieties 28 and 32 activated the same molecular mechanism under shade treatment.

In the S28-VS-S32 comparison, it is mainly concentrated in plant–pathogen interaction (ko04626), MAPK signaling pathway (ko04016), plant hormone signaling (ko04075), pentose and glucuronate interconversion (ko00040), sphingosphaerolipid biosynthesis (ko00604), galactose metabolism (ko00052), and other aggregate pathways of sugar degradation (ko00511), diterpenoid biosynthesis (ko00904), sphingoid metabolism (ko00600), glycosaminoglycan degradation (ko00531), and secondary metabolite biosynthesis (ko01110) pathways (Figure 3C). In the C28-VS-C32 comparison, phytohormone signaling (ko04075), plant–pathogen interaction (ko04626), mutual conversion of pentose and glucuronate (ko00040), dterpenoid biosynthesis (ko00904), MAPK signaling pathway (ko04016), glycosphingoid biosynthesis (ko00604), other glycan degradation (ko00511), fatty acid elongation (ko00062), and glycosaminoglycan degradation (ko00531) were the main enrichment pathways (Figure 3D). By comparing the enrichment pathways of DEGs between different varieties under the same treatment, it was found that biosynthesis of secondary metabolites and sphingolipid metabolism were the main enrichment pathways affecting the differences between the two varieties.

The cluster analysis of metabolites and samples showed that the metabolites of soybean leaves under different treatments were significantly different. Under shade treatment, the accumulation of differential metabolites (DEMs) in variety 28 was significantly higher than that in control (Figure 4A). In order to compare the metabolites related to shedding in the leaves of varieties 28 and 32 under shade treatment, a total of 1888 metabolites in soybean leaves were detected by UPLC-MS/MS, which were divided into 12 categories (Figure 4B), with flavonoids (27.5%) accounting for the largest proportion, followed by amino acids and derivatives (13.62%), phenolic acids (12.67%), other (8.59%), terpenoids (8.37%), lipids (8.16%) and alkaloids (5.72%). The unsupervised pattern recognition analysis of 12 samples using metabolite principal component analysis (Figure 4C) can also clearly distinguish the data points of varieties and treated samples, indicating that there are significant differences in metabolites of cultivars 28 and 32 treated samples.

A total of 408 DEMs were detected in C28-VS-S28, of which 175 were up and 233 were down (Figure 5A,D). A total of 189 DEMs were detected in C32-VS-S32, with only 46 up-regulated and 143 down-regulated (Figure 5B,D). These DEMs can be divided into 12 classes (Figure 5D), and the top 5 of the 28 cultivar are flavonoids (43.9%), amino acids and their derivatives (15%), phenolic acids (9.1%), alkaloids (5.6%), and organic acids (5.6%). Flavonoids can be further divided into 26 flavonols, 21 flavonoids, 5 dihydroflavones and anthocyanins, 3 isoflavones and other flavonoids, 2 chalcones and dihydroflavonols, and 1 flavanols. Among the 32 varieties, the top 5 categories were flavonoids (43.4%), phenolic acids (15.9%), organic acids (13.2%), amino acids and their derivatives (9.5%) and other categories (4.8%). These flavonoids are further divided into 15 flavonoids, 12 flavonols, 3 isoflavones, 2 each of dihydroflavones and anthocyanins, and 1 each of chalcones and other flavonoids. Multiple comparative analyses showed that only 12 DEMs of species 28 and 32 were identical (Figure 5C). By analyzing the results of DEMs in different treatments of the same varieties showed that the two varieties had similar metabolite accumulation patterns, with the major differences being flavonoids. Shading resulted in downregulation of most metabolic intermediates of varieties 28 and 32, suggesting that shading inhibited normal metabolic processes.

A total of 531 DEMs were detected in C28-VS-C32, of which 344 were up and 187 were down (Figure 6A,D). A total of 550 DEMs were detected in S28-VS-S32, of which 271 were up-regulated and 279 down-regulated (Figure 6B,D). These DEMs can be divided into 12 classes (Figure 6D). The top six classes in C28-VS-C32 are flavonoids (59.7%), amino acids and their derivatives (10%), phenolic acids (8.1%), terpenoids (6.5%), lignans and coumarins (3.4%), and others (3.3%). Flavonoids were further divided into 167 flavonoids, 164 flavonols, 62 isoflavones, 51 dihydroflavones, 39 anthocyanins, 13 chalones and dihydroflavonols each, 11 other flavonoids, and 2 flavanols. The top six categories in S28-VS-S32 were flavonoids (58.6%), amino acids and their derivatives (10.1%), phenolic acids (9.2%), terpenoids (5.3%), lignans and coumarins (3.7%) and organic acids (3.7%), respectively. These flavonoids are further divided into 97 flavonoids, 41 flavonols, 38 isoflavones, 35 dihydroflavones, 24 other flavonoids and anthocyanins each, 14 chalcones, 7 dihydro flavonols, and 3 flavanols. Multiple comparative analyses showed that there were 274 DEMs in common between varieties 28 and 32 (Figure 6C). The results of the comparison of DEMs among different varieties of the same treatment showed that flavonoids, amino acids and their derivatives, phenolic acids, and terpenoids were the main DEMs of the two varieties. Shading treatment significantly down-regulated the numbers of most metabolites, and up-regulated the numbers of organic acid metabolites. These results indicated that the down-regulated metabolites were the key factors leading to the difference between the two varieties and the control, and the up-regulated organic acid compounds might be the main factors causing the difference between the two varieties under shade treatment.

3.3.2. Joint Analysis of the Transcriptome and Metabolome of Soybean Leaves under Shade Treatment

The pathways with the top 25 p-values in the number of KEGG pathways co-enriched by the transcriptome and metabolome were plotted as bar charts (Figure 7) to identify the major physiological, biochemical, and signal transduction pathways that DEGs and DEMs are involved in together. In the C28-VS-S28 treatment, DEGs and DEMs were mainly involved in metabolic pathways (Meta: 21, Gene: 225), biosynthesis of secondary metabolites (Meta: 12, Gene: 136), cofactor biosynthesis (Meta: 6, Gene: 30), and amino acid biosynthesis (Meta: 7, Gene: 19), phenylpropanoid biosynthesis (Meta: 5, Gene: 17) and flavonoid biosynthesis (Meta: 6, Gene: 12) pathway enrichment (Figure 7A). In C32-VS-S32 treatment, DEGs and DEMs are mainly involved in metabolic pathways (Meta: 10, Gene: 103), biosynthesis of secondary metabolites (Meta: 5, Gene: 53), plant hormone signal transduction (Meta: 1, Gene: 20), and biosynthesis of cofactors (Meta: 1, Gene: 14), phenpropyl biosynthesis (Meta: 2, Gene: 5) and flavonoid biosynthesis (Meta: 2, Gene: 4) pathway enrichment (Figure 7B). In C28-VS-C32 treatment, DEGs and DEMs are mainly involved in secondary metabolite biosynthesis (Meta: 28, Gene: 150), plant hormone signaling (Meta: 1, Gene: 118), and mutual conversion of pentose and glucuronate (Meta: 1, Gene: 30), starch and sucrose metabolism (Meta: 1, Gene: 28), isoflavone biosynthesis (Meta: 14, Gene: 10) and flavonoid biosynthesis (Meta: 14, Gene: 8) pathway enrichment (Figure 7C). In the S28-VS-S32 treatment, DEGs and DEMs were mainly involved in metabolic pathways (Meta: 29, Gene: 462), biosynthesis of secondary metabolites (Meta: 29, Gene: 280), biosynthesis of cofactors (Meta: 4, Gene: 44), and biosynthesis of phenylpropyl (Meta: 4, Gene: 41), isoflavone biosynthesis (Meta: 17, Gene: 25), flavonoid biosynthesis (Meta: 15, Gene: 15), and flavonol biosynthesis (Meta: 13, Gene: 8) pathway enrichment (Figure 7D). Under normal light conditions, glucose metabolism was the main pathway to produce differences between varieties 28 and 32. Under shade treatment, the main pathway leading to differences between varieties was amino acid biosynthesis and plant hormone signal transduction. Metabolic pathway and secondary metabolite biosynthesis pathway were the key pathways affecting light sensitivity in the leaf.

4. Discussion

In actual production, dense planting of soybeans often leads to a low light environment with in the population, which increases the rate of flower and pod shedding. Soybeans are a typical short-day crop with an extremely sensitive photoperiod, and the impact of low light signal on plants varies greatly in different growth periods, and low light signal will lead to a disproportionate increase in flower and pod shedding during the critical photosensitive period. The specific photosensitive period of soybean is not clear [15,16]. Therefore, it is of great significance to determine the key period of low light signal affecting flower and pod abscission for improving the pod rate per plant of soybean. In this study, Liao Dou 32 and Shen Nong Dou 28, which have different flower and pod abscission rates at similar growth periods, were selected for detection. The low light signals from shading affected plant phenotypes, dry matter accumulation, transcription factors, and metabolites of the two varieties. Combining the transcriptome and metabolome, differences in transcription factors and metabolites of cultivars 28 and 32 were analyzed to determine the metabolic pathways that DEGs and DEMs are involved in, and screen out major metabolites and key candidate genes affecting flower and pod shedding. The results showed that the two varieties activated the same molecular mechanism and similar metabolite accumulation mode under shade treatment, but the regulation modes of the two varieties were different.

Weak light inhibited the carbohydrate pathway in plants [17,18], and the level of carbohydrate in plants was significantly reduced under low light, which is caused by the inhibition of photosynthesis. Carbohydrate reduction caused by low light will also promote the competition between vegetative organs and reproductive organs, affecting flower bud differentiation, flowering, pollination, fruit setting and fruit development [19,20]. Hieke et al. [21] found that the developing fruit would lead to shedding due to nutritional stress caused by low light formed by shade. Domingos et al. [6] showed that shading treatment severely inhibited photosynthesis and carbohydrate metabolic pathways in grapes, and the significant reduction in carbohydrates promoted the shedding of flower organs. In this study, shading resulted in a significant decrease in plant dry matter accumulation and the lowest pod rate per plant at V4 stage.

Weak light induced abscission was a complex regulatory process, which is regulated by multiple genes and multiple pathways. In this study, the TFs with the largest differential expression caused by shading were MYB, bHLH, C2C2, HB, AP2/ERF and WRKY family members (Figure 2), and DEGs were mainly enriched in metabolic pathways, biosynthesis of secondary metabolites, glucose metabolism, antioxidant and sphingoid metabolism pathways (Figure 3). Studies have shown that these TF families are closely related to plant stress resistance [22,23]. WRKY family transcription factors are mainly involved in plant biological and abiotic stresses and regulate hormone signal transduction, and WRKY6 regulates ABA signal transduction by down-regulating RAV1 [24,25]. In this study, the two varieties with high levels of expression genes were WRKY6, WRK53, WRK70, WRK40 and WRK17, and the expression level of WRK24 gene in variety 28 was higher than that of 32, indicating that these WRKY transcription factors play a special role in low light stress, the WRK24 may be one of the key genes in the formation of varietal differences under shade treatment. Most MYB and bZIP family members regulate downstream processes related to ABA [26], and MYB also interacts with photoperiochrome Phy B to regulate stomatal closure in plant leaves [27], and phytochrome affects plant response to stress by regulating the anabolic or signaling pathways of jasmonic acid. In this study, the MYB family of two varieties had the largest number of differentially expressed genes, among which DIV and MYB06 were highly expressed genes, and SRM1 and MYB16 gene expression levels of variety 28 were significantly higher than those of 32, and MYB61 and TRB1 gene expression levels of variety 32 were significantly higher than those of 28. These results indicated that DIV and MYB06 played important regulatory roles in shading response, and SRM1, MYB16, MYB61 and TRB1 genes were some of the main factors leading to differences between varieties under low light conditions. In addition, this study also identified that the expression levels of AP2/ERF family members were significantly changed under the influence of shade, and these transcription factors have been shown to positively regulate plant stress resistance in other plants [28,29]. Further study of these genes will help to further understand the regulatory mechanism of low light sensitivity in soybean.

Metabolites are the key substances to maintain the whole life cycle of plants, and low light stress can significantly affect the secondary metabolites of plants [30], when the grape is shed by shading treatment, secondary metabolism is found to be inhibited [6]. In this study, the main DEMs in shaded soybean leaves were flavonoids, amino acids and their derivatives, phenolic acids and terpenoids, and the main DEMs among varieties were alkaloids and organic acids. Flavonoids are the main components of plant secondary metabolism, participate in self-protection and antioxidant defense systems, and can improve plant resistance to biological and abiotic stresses [31,32]. Nakabayashi reported that flavonoids have strong free radical scavenging activity, which can slow down the oxidation and drought stress of Arabidopsis [33]. Terpenoids are the most abundant natural compounds in plant secondary metabolites [34], and can play an important role as special signal molecules in various biological and abiotic stresses and signal transmission between plants [35,36,37,38]. Other studies have pointed out that within a certain range of light intensity, the contents of terpenoids and amino acids are significantly positively correlated with light intensity [39]. In addition to the basic nutritional function of protein synthesis, amino acids were also precursors of the synthesis of many signal molecules, directly involved in a variety of physiological activities and defense mechanisms [40]. Guo et al. [41] found that the key metabolites were amino acids, sugars and organic acids in their studies on different drought tolerant wheat varieties. Organic acids are the intermediate products of the tricarboxylic acid cycle and glyoxylic acid cycle in mitochondria [42]. In addition to participating in photosynthesis and respiration, organic acids can also be used as a metabolically active solute as a key component in the physiological metabolism of various plants [43]. When subjected to environmental stress, plants will produce a large amount of organic acids to form a resistance mechanism against environmental stress [44]. Alkaloids can help plants resist the invasion of microorganisms and viruses, but their tolerance to abiotic stress needs further study.

In order to explore the molecular mechanisms of plant response to weak light, many studies have identified key candidate genes and metabolites and pathways that may play important roles in plant resilience. In this study, the key candidate genes were identified as DIV, MYB06, MYB44, MY1R1, MYBS3, WRKY6, WRK53, WRK70, WRK40, DOF14, CDF3, CDF2, GATA5, DREB3, ERF3 and SRM1, MYB16, TRB1, MYB61, WRK24, and COL16. The main metabolites are flavonoids, amino acids and their derivatives, phenolic acids and terpenoids, alkaloids and organic acids. These genes and metabolites jointly regulate shading response through metabolic pathways, biosynthesis of secondary metabolites, biosynthesis of amino acids and hormone signal transduction in plants. Mehrotra et al. [45,46] found that plants regulate the expression of genes through transcription factors such as MYB, WRKY, bZIP and AP2/ERF, induce the synthesis of protective substances such as plant hormone signal transduction and secondary metabolites, and activate stress-responsive genes to enhance plant stress resistance. For example, by synthesizing proline and soluble sugar, the osmotic potential and cell membrane stability of plants under low temperature stress were regulated, and reactive oxygen species were removed [47]. Studies have shown that overexpression of WRKY30 increased the accumulation of soluble sugar and proline and the activity of ROS scavenging enzyme to enhance drought tolerance of plants [48]. Jiang et al. [49,50] observed that overexpression of CsMYB5a promoted the expression of three structural genes LAR, ANS and ANR in flavonoid metabolism, and R2R3-MYB, bHLH and WD40 could form a terpolymer complex MBW, which jointly regulated the transcription intensity of multiple structural genes downstream of flavonoid synthesis and positively regulated anthocyanin accumulation. Transcription factors such as MYB, WRKY, AP2/ERF and bHLH not only regulated the accumulation level of flavonoid metabolites, but also were key transcription factors affecting the metabolism of terpenoids [51,52].

To sum up, it was preliminarily concluded that the pattern of weak light causing flower pod abscission of the two cultivars may be as follows: Weak light resulted in the inhibition of carbohydrate metabolism and transport, and promoted the competition of photoassimilate between vegetative organs and reproductive organs. Carbohydrate metabolism determined the biosynthesis of secondary metabolites, while inhibiting hormone transport, resulting in disruption of hormone balance. Assimilate competition and hormone disturbance directly induced the shedding signal, and led to the initiation of the abscission process, the weakening of secondary metabolic biosynthesis, the downregulation of antioxidant and stress resistance, and the further occurrence of shedding.

5. Conclusions

In this study, plant phenotypes, dry matter accumulation, leaf transcription factors, and major metabolites of varieties 28 and 32 were measured under shade treatment. The changes in gene and metabolite expression in the two cultivars were analyzed to identify expression and metabolic pathways that both DEGs and DEMs participate in and screen key candidate genes and major metabolites. The results showed that the two cultivars activated the same molecular mechanism and similar metabolite accumulation mode under shade treatment, but the regulation modes of the two cultivars were different.

The identified common DEGs were DIV, MYB06, MYB44, MY1R1, MYBS3, WRKY6, WRK53, WRK70, WRK40, DOF14, CDF3, CDF2, GATA5, DREB3, and ERF3. The main DEGs of variety 28 are SRM1, MYB16, WRK24 and COL16; variety 32 are MYB61 and TRB1. The identified common metabolic pathways were secondary metabolite biosynthesis, metabolic pathway, cofactor biosynthesis, phenylpropanoid biosynthesis, flavonoid biosynthesis, fatty acid biosynthesis and amino acid biosynthesis. The main DEMs of CV.28 are carbon metabolism, glutathione metabolism, biotin metabolism, nucleotide metabolism, and purine metabolism. CV.32 mainly focused on plant hormone signal transduction, biosynthesis of flavonoids and flavonols, lysine degradation, metabolism of arginine and proline, metabolism of amino sugars and nucleotide sugars.

Based on the above results, through the comparison of the two cultivars, the related genes and metabolic pathways affecting soybean flower and pod abscission could be preliminarily identified. The results of this study will be helpful for the localization and cloning of genes related to flower and pod abscission, as well as the study of related metabolic mechanisms.