Circadian clock regulation in soybean senescence: a transcriptome analysis of early and late senescence types

Abstract

Background

Plant senescence is the process of physiological maturation of plants and is important for crop yield and quality. Senescence is controlled by several factors, such as temperature and photoperiod. However, the molecular basis by which these genes promote senescence in soybeans is not well understood. We identified senescence-related genes via transcriptome analysis of early-senescence (ES)- and late-senescence (LS)-type plants to elucidate the molecular mechanisms of senescence in soybeans.

Results

We obtained early-senescence (ES)- and late-senescence (LS)-type F7 plants from a cross between a hybrid (Glycine max × Glycine soja) and the Glycine max cultivar. The ES-type plants presented the reproductive (R2) growth stage at 50 days after sowing (DAS) and the R7 growth stage at 95 DAS, whereas the LS-type plants presented the beginning of the R1 and R6 growth stages at 50 and 95 DAS, respectively. To understand the molecular mechanisms underlying this senescence, we performed transcriptome analysis of leaves from 50 to 95 DAS of ES- and LS-type plants. A total of 2,414 and 2,471 genes at 50 and 95 DAS, respectively, were differentially expressed between ES-type and LS-type plants. Twenty-three candidate genes associated with the circadian clock, chlorophyll biosynthesis, phytohormones, and senescence-associated protein kinases were identified, and their expression levels were analyzed. In addition, to understand interaction between circadian clock and senescence, we analyzed expression patterns of seven circadian clock-related genes during the time period (light and dark condition): CIRCADIAN CLOCK ASSOCIATED 1 (CCA1), LATE ELONGATED HYPOCOTYL (LHY), CONSTANS-LIKE 9 (COL9), LUX ARRHYTHMO (LUX) EARLY FLOWERING 3 (ELF3), PSEUDO-RESPONSE REGULATOR5 (PRR5) and GIGANTEA (GI). The expression patterns of circadian clock-related genes were similar in the ES- and LS-type plants. However, the transcription levels of these genes were compared between ES- and LS-type plants, and the expression of these genes was greater than that in LS-type plants during the period when expression increased. Therefore, each set of candidate genes regulated senescence in each plant by regulating their expression level.

Conclusions

These findings provide novel insights into the regulation of senescence in soybean plants, which could lead to the development of new strategies to improve agriculture.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12864-024-11095-3.

Background

Senescence is a physiological maturation process that results in decreased photosynthetic effectiveness, yellowish color, and tissue death [1–4]. Senescence is regulated by leaf age, light, temperature, and stress factors [2, 5]. Thus, senescence has a significant effect on crop yields [2, 6–9]. Senescence is crucial for nutrient remobilization and reproductive success [9, 10]. Early senescence affects quality and yield [11], whereas late senescence prolongs the seed maturation period in crops [12, 13]. Although senescence is characterized by a visible yellow phenotype [14, 15], several highly organized regulatory mechanisms are involved [16–19]. Several mechanisms have been identified as key players in the regulation of senescence, such as the plant hormone response, development of chloroplasts, and the circadian clock system [19–23]. Abscisic acid (ABA), ethylene, jasmonic acid (JA), and brassinosteroid (BR) positively regulate senescence, whereas cytokinin, auxin, and gibberellin negatively regulate senescence [24–27]. STAY GREEN (SGR1) acts as a positive regulator of chlorophyll degradation in tomatoes and soybeans [28–30].

The circadian clock system has been widely studied in plants [31]. The circadian clock regulates several developmental processes, including flowering and senescence, and is stimulated by environmental cues such as light and temperature [22, 23, 32]. The CIRCADIAN CLOCK ASSOCIATED 1 (CCA1) gene acts as a key regulator of senescence [33]. CCA1 acts as a master regulator of reactive oxygen species (ROS) homeostasis through association with EARLY FLOWERING 3 (ELF3) and LUX ARRHYTHMO (LUX) [34, 35]. LATE-ELONGATED HYPOCOTYL (LHY), a homologous protein of CCA1, together regulates chlorophyll synthesis and photosynthesis metabolism [31, 36–39]. REVEILLE 1-RELATED (RVE1) is a MYB-like transcription factor that plays a significant role in facilitating circadian rhythms with PSEUDO-REPONSE REGULATOR5 (PRR5) [40, 41]. GIGANTEA (GI) regulates the timing of leaf senescence and flowering [42], and CONSTANS-LIKE (COL) plays an important role in the photoperiod pathway, circadian rhythm, and light signals for flowering in Arabidopsis [43–45]. The overexpression and knockout of circadian clock genes in plants resulted in changes in the aging process and development. For example, the overexpression of CCA1 resulted in late flowering and senescence [37, 46], and the PRR9 mutation delayed leaf senescence in Arabidopsis [47]. The rve8 knockout mutants have earlier flowering time than wild type plant [41]. The overexpression of COL induces senescence during dark light treatment in rice [48]. In Arabidopsis, the elf3 mutants show earlier senesce than wild type plant and overexpression of ELF3 plants delayed senescence [35]. However, the overexpression of OsELF3.1, a homologous gene in rice, induces early aging phenotype [49]. Recent studies have shown that flowering time provides flexibility in response to senescence at an appropriate time and that the circadian clock is related to the senescence system in various crop species [33, 35, 50–56].

In this study, phenotypic and transcriptomic analyses were performed on early-senescence (ES) and late-senescence (LS) plants obtained from crosses of wild and cultivated soybeans. To investigate the regulation of senescence, the transcriptomes of ES and LS plants at 50 and 95 DAS were analyzed at the flowering and leaf senescence stages, respectively. Our findings show that the circadian clock plays a significant role in the regulation of senescence by controlling the expression of genes involved in the flowering and maturation stages. We also discuss how chloroplast development, plant hormone responses, and protein kinase-related genes regulate senescence. These findings shed light on how the interaction of relevant genes mediates senescence in soybeans.

Methods

Plant materials

We found significantly different flowering and senescence lines derived from one line of F6 recombinant inbred line (RIL) populations crossed between Glycine max (G. max), a cultivated soybean, and a hybrid of G. max and Glycine soja (G. soja) [57]. Further, the phenotype of plants from F7 line was observed and the lines of different phenotype were sampled to RNA expression analysis. Flowering time and senescence were observed at 35, 50, 95, and 110 days after sowing (DAS). Reproductive stages of soybean development were observed: R1, beginning of flowering; R2, full bloom; R3, beginning pod, R4, Full pod; R5, beginning seed; R6, full seed; R7, beginning maturity; R8, full maturity [58]. The F7 seeds were subsequently harvested and grown in a greenhouse and growth room at the optimal temperature of 25 °C and 65 ± 5% humidity. Flowering time and senescence were observed in ES- and LS-type plants, especially at 50 and 95 DAS. Finally, from the F7 generation, two plants (2241 early-senescence and 2242 late-senescence) were selected for exploring the mechanism underlying senescence in soybeans. Leaves from early senescence (ES, 2241) and late senescence (LS, 2242) plants were sampled at 50 and 95 DAS and harvested immediately in frozen liquid nitrogen for RNA isolation.

35-day-old plants were subjected to light/dark treatments for both ES and LS plants. The light treatment consisted of exposure to light for 0, 6, and 12 h, followed by a period of darkness of 18 and 24 h. A growth chamber with three plates placed in each chamber was used to conduct the treatments. Each plate contained 5 plants spaced 20 cm apart. LED lights (Blue, 436 nm + Red, 658 nm) were used for light intensity treatments. The temperature was maintained at 25 °C throughout the light and dark periods, with a relative humidity of 60–70%. To analyze expression pattern of genes, the leaves of 35 DAS ES and LS plants at 0, 6, 12, 18 and 24 h were harvested and frozen in liquid nitrogen.

RNA extraction and reverse transcription‒quantitative polymerase chain reaction (RT‒qPCR)

Total RNA from the leaves of 50- and 95-DAS early- and late-senescence type plants, leaves of 35 DAS ES and LS plants under light and dark treatments were extracted via the GeneAll^® Ribospin™ Plant (GeneAll Biotechnology Co., Ltd., Seoul, Korea) according to the manufacturer’s protocol. Genomic DNA digestion was performed with DNase I (Sigma, St. Louis, MO, United States of America). The quality of isolated RNA was checked via a NanoDrop 2000 spectrophotometer-MULTISKAN Sky (Thermo Fisher Scientific, Inc., United States). Two and three biological replicates for each sample were prepared for RNA-sequencing.

For first-strand cDNA templates, cDNA was synthesized via the use of 2 µg of total RNA using a the CycleScript™ Reverse Transcriptase master mix (with oligo dT). RT-qPCR analysis was performed using the TB Green Premix Ex Taq II (TaKaRa, Japan) in a CronoSTAR™ Real-Time PCR system (Clotech, Japan). The PCR reactions were performed by initial denaturation at 95 °C for 1 min, followed by 45 cycles of 95 °C for 5 s, 60 °C for 10 s, and 72 °C for 10 s. The Actin11 (Act11) soybean gene was used as an internal control for normalization [59]. Three technical replicates were analyzed for quantitative experiments. The primer information is given in Additional file 10.

RNA-sequencing and differential gene expression (DEG) analysis

We constructed cDNA libraries 50 and 95 days after sowing (DAS) for both early-season (ES) and late-season (LS) crops, following the methods outlined in a previously published paper [57]. The raw data were obtained from the Illumina high-throughput sequencing platform and were preprocessed following the same protocol as described in a previously published research article [57]. The raw sequencing data were deposited into the NCBI SRA database with accession number GSE275263. The quality reads were filtered by eliminating low reads via Trimmomatic’s sliding window (4:20) with average quality (30) and read size of 36 bp. Reference genome Glycine max Wm82.a2.v1 was used to construct the index and alignment. StringTie v1.3.4d (https://ccb.jhu.edu/software/stringtie/index.shtml) software package was used to analyze the expression level. DESeq2 software was used to identify the DEGs with adjusted p value < 0.05 and fold change of 2. Transcript expression level was quantified via the fragments per kilobase of transcript per million mapped reads (FPKM) method. The heat map and Venn diagram of the DEGs were generated via in-house R script.

Functional annotations

The functions of the DEGs were annotated via Gene Ontology (GO) enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses. GO was performed via SoyBase (SoyBase.org). The Kyoto Encyclopedia of Genes and Genomes (KEGG) annotations were obtained from the KEGG database. BLASTp (e-value 1e-3) analysis was performed via the NCBI RefSeq (http://www.ncbi.nlm.nih.gov) plant protein sequence, TAIR, and UniProt databases. Gene Ontology analysis of biological process, cellular component, and molecular function terms was performed via InterProscan and BLAST. The KEGG pathway analysis was performed via BLAST2GO (https://www.blast2go.com/). A p value of ≤ 0.05 was used as a significant threshold for the GO and KEGG pathway enrichment analyses. The enrichment of DEGs in the KEGG pathways was carried out via clusterProfiler R.

Accession number

The sequence data from this study can be found in the SoyBase and NCBI databases under the following accession numbers: CCA1 (Glyma.16G017400), LHY (Glyma.19G260900), COL-9 (Glyma.14G190400), LUX (Glyma.02G162600), ELF3 (Glyma.04G050200), PRR5 (Glyma.13G135900), GI (Glyma.10G221500) and Act11 (Glyma.02G091900).

Results

Characterization of two recombinant inbred lines (RILs) of single seed descendants from F1

To understand the molecular mechanisms underlying senescence, the parent G. max (early senescence) and hybrid G. max X G. soja (late senescence) soybean were crossed to generate a recombinant inbred line (RIL). The F2–F5 generation plants, generated via single-seed descent from F1, were segregated by plant morphology (Additional file 1). Among the total RILs of F7, the two lines of the plant (2241: early senescence and 2242: late senescence) presented variation in senescence. To determine the senescence time of the plants, we analyzed the development stages [58] of early-senescence (ES)-type and late-senescence (LS)-type plants (Fig. 1). The beginning of flowering (R1 stage) and flowering at a node with completely unrolled leaves (R2 stage) occurred 35 and 50 days after sowing (DAS) in the ES plants (Fig. 1A and B). Unlike ES plants, LS-type plants showed a vegetative growth stage at 35 DAS and an initiated reproductive stage (R1) at 50 DAS (Fig. 1B, Additional file 2).

Fig. 1.

Morphology of early (ES)- and late-senescence (LS)-type plants. Phenotypic comparison between early (ES) and late senescence (LS) plants at 35 days (A), 50 days (B), 95 days (C and D) and 110 days (E and F) after sowing. (A) Representation of shoots (top) and nodes (bottom) at the young stage of ES and LS plants. (B-F) Senescence phenotypes of ES and LS plants: flower (B), leaf (C-E) and pod (F). The red arrow indicates the flowers. Scale bars = 10 cm (A and B) and 5 cm (C-F)

The ES-type plants showed the beginning of the R7 stage (leaf senescence) at 95 DAS (Fig. 1C and D), and many leaves turned yellow at 110 DAS (Fig. 1E), whereas the leaves of the LS-type plants still showed the R6 stage (mainly green) at 95 and 110 DAS (Fig. 1C, D and E). Furthermore, at 110 DAS, the ES pods were yellow and started to mature; however, at 110 DAS, the LS plants were still green (Fig. 1F). Based on the results, the growth stages clearly differed between EL- and LS-type plants, suggesting that the senescence of ES- and LS-type plants, which are genetically identical, is controlled by gene expression.

Transcriptome analysis and identification of DEGs

Library construction and sequencing were performed on 12 samples (three replicates of four types of leaves) of 50 and 95 DAS for ES- and LS-type plants, respectively. A total of 12 libraries were sequenced and analyzed. After low-quality reads were removed, the average number of reads per library exceeded 15 million. A total of 19,552,733,602 sequencing reads were generated from the raw data, and 17,488,036,692 reads were generated from the clean data (89.52% of the total reads) (Additional file 3). The RNA-sequencing reads were aligned with the reference map of the newly assembled Williams 82 (a2. v1) soybean genome (Additional file 4). Consequently, 95% of the sample reads were mapped to the reference genome. To identify the genes involved in senescence, we analyzed DEGs from the RNA-sequencing reads of ES- and LS-type plants. The DEGs were normalized by fragments per kilobase of transcript per million mapped reads (FPKM) and selected with a p value < 0.05 and | log2 (fold change) | ≥ 1. A total of 2414 DEGs were identified in 50 DAS ES- vs. LS-type plants, and 2471 DEGs were identified in 95 DAS ES- vs. LS-type plants. The expression of 946 genes and 1,468 genes was upregulated and downregulated, respectively, at 50 DAS in LS-type plants compared with 50 DAS in ES-type plants (Additional file 5). Similarly, the expression of 1317 genes and 1154 genes was upregulated and downregulated, respectively, at 95 DAS in LS-type plants compared with 95 DAS in ES-type plants (Additional file 5).

The heatmap revealed that the expression of DEGs was visually differentiated between 50 DAS and 95 DAS in ES- and LS-type plants (Fig. 2A). In addition, the expression patterns of DEGs were analyzed at 50 and 95 DAS in LS-type plants and compared with those in ES-type plants via a Venn diagram (Fig. 2B-E). Compared with that in ES-type plants, the expression of 219 genes in LS-type plants was upregulated at both 50 and 95 DAS (Fig. 2B). Compared with that in ES-type plants, the expression of 317 genes in LS-type plants was downregulated at both 50 and 95 DAS (Fig. 2C). Furthermore, 140 genes were upregulated at 50 DAS and downregulated at 95 DAS in LS-type plants (Fig. 2D). The expression of 244 genes was downregulated at 50 DAS and upregulated at 95 DAS in the LS-type plants (Fig. 2E). Taken together, these findings indicate that each senescence stage shares a distinct set of genes and that the expression of these DEGs may be involved in the occurrence and development of senescence variation among ES- and LS-type plants.

Fig. 2.

Transcriptome profiles at 50 and 95 days after sowing (DAS) of early senescence (ES)- and late senescence (LS)-type plants. (A) Hierarchical cluster analysis of differentially expressed genes (DEGs) between 50 and 95 DAS in ES and LS plants. The color key indicates gene expression FPKM values from − 2 to 2: blue = lowest and red = highest among the DEGs between ES and LS plants. (B-E) Venn diagram of up and downregulated DEGs in 50 and 95 DAS LS plants relative to ES plants. The upregulated (B) and downregulated (C) DEGs in both 50 and 95 DAS LS plants compared with those in ES plants. (D) Upregulated DEGs in LS plants at 50 DAS and downregulated DEGs in LS plants at 95 DAS with respect to ES plants. (E) Downregulated DEGs in LS plants at 50 DAS and upregulated DEGs in LS plants at 95 DAS. Red and blue circles indicate up and downregulated genes, respectively (P value < 0.05 and > log2-fold change)

DEG classification and functional annotation

Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses were performed to further investigate the functions of the DEGs. The identified DEGs were assigned to GO for biological processes, cellular components, and molecular functions. At 50 DAS, 867 DEGs associated with biological processes, 316 associated with cellular components, and 1,364 associated with molecular functions were annotated in ES- and LS-type plants. Similarly, at 95 DAS in ES- and LS-type plants, 824 DEGs were annotated in biological processes, 282 in cellular components, and 1,277 in molecular functions. The top ten significantly enriched GO terms for ES- and LS-type plants at 50 and 95 DAS (p < 0.05) were selected and categorized (Fig. 3A and B). Together, the significantly enriched GO terms were associated with protein phosphorylation, oxidation‒reduction process, transmembrane transport, regulation of transcription, DNA-template, proteolysis in biological processes, with integral components of the membrane, membrane, nucleus, intracellular, cytoplasm, and apoplast; and the cell wall in a cellular component, protein binding, protein kinase activity, ATP binding, catalytic activity, protein kinase activity, protein binding, nucleic acid binding, and oxidoreductase activity in molecular function (Fig. 3). Among the DEGs with KEGG pathway annotations, 814 DEGs were enriched in thiamine metabolism, purine metabolism, starch and sucrose metabolism, pentose and glucuronate interconversions, and porphyrin and chlorophyll metabolism (Additional file 6). These pathways are related to differences in the senescence of ES- and LS-type plants, indicating that several genes regulate soybean senescence by controlling biological processes.

Fig. 3.

Gene Ontology (GO) analysis of the DEGs at 50 and 95 days after sowing (DAS) in early senescence (ES)- and late senescence (LS)-type plants. GO analysis of enriched DEGs at 50 DAS (A) and 95 DAS (B), with three independent categories: biological process, cellular component, and molecular function. The x-axis represents the GO terms, and the y-axis represents the number of DEGs (P value < 0.05 and > log2-fold change)

Identification of genes involved in senescence

A comparison of the transcriptomes of ES- and LS-type plants at 50 and 95 DAS revealed the expression of twenty-three candidate genes: six phytohormone-responsive genes, four chlorophyll biosynthesis-related genes, six senescence-associated protein kinases (Additional file 7 and 8) and seven circadian clock-related genes (Table 1; Fig. 4), which are highly involved in flowering time and leaf senescence.

Table 1.

Significantly regulated expression of circadian clock-related genes at 50 and 95 days after sowing (DAS) in early senescence (ES)- and late senescence (LS)-type plants compared with that in ES-type plants

Category	Unigene	Symbol	Gene	50 days		95 days
				Log2FoldChange	p-value	Log2FoldChange	p-value
Circadian clock related genes	Glyma.16G017400	CCA1	CIRCADIAN CLOCK ASSOCIATED 1	9.119533	0.0045	4.16719	0.00023
	Glyma.19G260900	LHY	LATE ELONGATED HYPOCOTYL	8.14457	0.0026	3.80093	0.0023
	Glyma.14G190400	COL-9	CONSTANS-LIKE 9	8.843077	0.010499	-9.865114	0.007662
	Glyma.02G162600	LUX	LUX ARRHYTHMO	6.416708	0.00026	7.02376	0.0056
	Glyma.04G050200	ELF3	EARLY FLOWERING 3	9.222225	0.000364	-12.030404	0.000037
	Glyma.13G135900	PRR5	PSEUDO-RESPONSE REGULATOR5	-9.705565	0.00056	7.423427	0.0076
	Glyma.10G221500	GI	GIGANTEA	-8.704811	0.00044	6.23452	0.0098

Fig. 4.

The expression of circadian clock-related genes between 50 and 95 days after sowing (DAS) in early senescence (ES)- and late senescence (LS)-type plants. RT‒qPCR analysis of the expression levels of CCA1, LHY, COL-9, LUX, ELF3, PRR5 and GI at 50 and 95 DAS in ES and LS plants. Total RNA was prepared from the leaves of the indicated samples. The ES-type plants in the control group were 50 and 95 DAS, respectively. The data are the mean values of three biological replicates, and the error bars indicate the SDs. Asterisks indicate statistically significant differences between the corresponding samples and their controls (p < 0.01, student t-test). Act11 was used as the internal control, and the relative expression levels are shown as fold values

The expression of the phytohormone responsive genes auxin-responsive factor 1 (ARF1), cytokinin dehydrogenase 3 (CKX3), abscisic acid 8’-hydroxylase-1 (ABAH1), ethylene-responsive transcription factor (ERF118) and BRASSINOSTEROID INSENSITIVE receptor kinase 1 (BAK1), were downregulated at 50 and 95 DAS in LS-type plants compared with that in ES-type plants. In contrast, ETHYLENE INSENSITIVE 3 (EIN3) was upregulated at 50 DAS and downregulated at 95 DAS in LS-type plants compared with ES-type plants (Additional file 7 and 8). Plant hormones play pivotal roles in the regulation of plant senescence.

Chlorophyll biosynthesis genes such as cysteine desulfhydrase 1 (LCD1), solute carrier family 40 (S40), and chlorophyllide a oxygenase (CAO) were upregulated at 50 and 95 DAS in LS-type plants compared with ES-type plants, whereas STAY-GREEN (SGR) was downregulated at 50 and 95 DAS in LS-type plants compared with ES-type plants (Additional file 7 and 8). Senescence-associated protein kinase: inositol polyphosphate kinase (IPK) was found to be upregulated in both 50 DAS and 95 DAS LS-type plants, whereas asparagine synthetase 2 (ASNS2) was upregulated at 50 DAS and downregulated in 95 DAS LS-type plants compared with ES-type plants. Unlike ferredoxin-dependent glutamate synthase (Fd-GOGAT), sucrose nonfermenting 4-like protein serine/threonine protein kinase (SNRK), zinc finger CCCH domain-containing protein 41 (C3H41), and wall-associated receptor kinase-like 14 (WAK14) were downregulated at 50 DAS in LS-type plants and upregulated at 95 DAS in LS-type plants compared with ES-type plants. (Additional file 7 and 8). The circadian clock-related genes CIRCADIAN CLOCK ASSOCIATED 1 (CCA1) and LATE ELONGATED HYPOCOTYL (LHY) were upregulated in LS-type plants compared with ES-type plants (Table 1; Fig. 4). In other plants, the overexpression of CCA1 and LHY in Arabidopsis results in delayed flowering time and senescence [33, 46, 60, 61]. The expression of LUX was regulated at both 50 and 95 DAS in LS-type plants (Table 1; Fig. 4). Compared with those in ES plants, the expression levels of CONSTANS-LIKE 9 (COL9) and EARLY FLOWERING 3 (ELF3) were upregulated at 50 DAS and downregulated at 95 DAS in LS-type plants. Unlike COL9 and ELF3, the expression of PSEUDO-REPONSE REGULATOR5 (PRR5) and GIGANTEA (GI) was downregulated at 50 DAS and upregulated at 95 DAS in LS-type plant genes (Table 1; Fig. 4). Additionally, suppressing CO and GIGANTEA (GI) expression results in delayed senescence in Arabidopsis [44, 62]. These results indicate that senescence is controlled by specifically regulating the circadian clock system, phytohormone response, chlorophyll biosynthesis and senescence-associated protein kinase.

The expression patterns of circadian clock-related genes in ES- and LS-type plants under light/dark conditions

Previous studies have shown that the interaction between the expression of circadian clock genes and flowering time regulates the senescence of plants [23, 33]. The above results revealed differences in development stage between ES- and LS-type plants at 35 DAS (Fig. 1), and the expression patterns of circadian clock-related genes were differentially regulated at 50 DAS and 95 DAS in ES- and LS-type plants (Fig. 4). Therefore, we expected that the expression of circadian clock-related genes in ES and LS plants would be differentially regulated from that in young plants. To test this hypothesis, we extracted total RNA from ES- and LS-type plants (35 days old) under light (0, 6, or 12 h) and dark (18, 24 h) conditions and measured the transcript levels of these genes via RT‒qPCR (Additional file 9 and Fig. 5). The expression patterns of CCA1, LHY, COL9, LUX, ELF3, PRR5 and GI in ES- and LS-type plants were similar under dark/light conditions (Additional file 9). The expression of CCA1 and LHY decreased under light conditions and increased at the end of the dark period (24 h) (Fig. 5A and B). The expression of COL9 decreased at 6 h under light conditions but increased under dark conditions (Fig. 5C). The expression of LUX was induced at 12 h and highly expressed at 18 h under dark conditions (Fig. 5D). The expression of ELF3 increased at 12 h and 18 h under dark conditions (Fig. 5D). The expression of GI was induced from 6 h (light) and highly expressed at 18 h under dark conditions (Fig. 5G). Unlike the expression patterns of CCA1 and LHY, the expression of PRR5 increased under light conditions but decreased under dark conditions (Fig. 5F). Therefore, we compared the expression of these genes between ES- and LS-type plants under light/dark conditions. The expression of CCA1 and LHY was upregulated in LS at 18 h and 24 h (dark conditions) when their expression was high (Fig. 5A and B). Compared with those in ES-type plants, the expression levels of COL-9, ELF3 and GI were upregulated in LS-type plants under both light and dark conditions (Fig. 5C, E and G). Unlike the expression of CCA1 and LHY, the expression of PRR5 was upregulated under light conditions (0, 6, and 12 h) and was similar to that of the ES type under dark conditions (18 h and 24 h) (Fig. 5F). The expression of LUX was upregulated at 0 h, 6 h, 12 h, and 18 h in LS and was like that in the ES type at 24 h (Fig. 5D). These results indicate that the temporal expression of circadian clock-related genes is specifically regulated, which is important in mediating senescence in soybean.

Fig. 5.

Comparison of the expression patterns of circadian clock-related genes at 35 days after sowing (DAS) in early-senescence (ES)- and late-senescence (LS)-type plants under light and dark conditions. RT–qPCR analysis of the expression levels of CCA1 (A), LHY (B), COL-9 (C), LUX (D), ELF3 (E), PRR5 (F) and GI (G) at 35 DAS in ES and LS plants under light (yellow color) and dark (gray color) conditions. The following total RNA was extracted from the leaves of the indicated time samples: light (yellow) and dark (gray). The control plants were ES-type plants. The data are the mean values of three biological replicates, and the error bars indicate the SDs. Asterisks indicate statistically significant differences between the corresponding samples and their controls (p < 0.01, student t-test). Act11 was used as the internal control, and the relative expression levels are shown as fold values

Discussion

Flowering and maturation are essential developmental stages for crop yield [8, 63–66]. Recently, many studies have reported that senescence mechanisms, including circadian clock rhythm, phytohormone response and the chloroplast development system, have been discovered through the identification of genes [67–71]. However, the senescence process in soybean is still not fully understood. Therefore, studying gene expression and function with respect to phenotype between two plant relatives provides basic information for understanding the underlying mechanisms involved. In this study, transcriptome analysis was performed to determine the expression and mechanism of senescence between early-senescence (ES) and late-senescence (LS) types in soybean obtained from the F7 RIL population derived from a cross of wild and cultivated soybean (Additional file 1). Our results revealed that the ES type flowers to completely unrolled leaves and yellow leaves at 50 and 95 days after sowing (DAS), whereas the LS type begins flowering at rolled and still green leaves (Fig. 1 and Additional file 2). Based on these findings, we hypothesize that the leaf senescence stage determines whether a soybean is an early senescence (ES) or late senescence (LS) type. Therefore, the leaf transcriptome was investigated to understand the relationship between gene regulation and senescence.

The circadian clock controls growth and development, such as photosynthesis, flowering, chlorophyll metabolism and senescence [56, 72]. A previous study revealed that circadian clock genes are involved in development and senescence in plants [32]. The RNA-Seq data revealed that circadian clock genes are differentially expressed between early-senescence (ES)-type and late-senescence (LS)-type soybean (Table 1). The expression of CIRCADIAN CLOCK ASSOCIATED 1 (CCA1) and LATE ELONGATEDHYPOCOTYL (LHY), MYB-domain transcription factors, is induced at dawn [39]. They are reduced by the PSEUDO-RESPONSE REGULATOR (PRR) family (PRR9, PRR7, PRR5 and PRR1) in the morning and evening [32, 46, 73]. The overexpression of CCA1 and LHY in plants results in delayed senescence [60, 74, 75]. Conversely, a prr9 knockout mutant, in which PRR9 expression is suppressed, shows delayed flowering [76] and leaf senescence in Arabidopsis [47]. In our study, CCA1 and LHY were more highly expressed in the LS type than in the ES type (Table 1; Fig. 4). Unlike CCA1 and LHY, PRR5 is expressed at lower levels at the young stage and is more highly expressed at the reproductive stage in the LS type than in the ES type (Table 1; Fig. 4). These findings indicate that the temporal expression pattern of circadian clock-related genes plays a crucial role in the senescence process. In addition, the PRR family is sequentially expressed and represses CCA1 and LHY during the early morning [77]. During the night, the ELF3 and LUX complex represses the expression of the GI and PRR families [78, 79]. Our results revealed that the expression patterns of CCA1, LHY, COL9, LUX, ELF3, PRR5 and GI in ES- and LS-type plants were similar to previously reported results in other plants (Additional file 9) [32, 80]. In addition, when the temporal expression levels of these genes in ES and LS-type plants were compared, the expression of these genes was greater than that in LS-type plants during the period when expression increased (Fig. 4). Taken together, these results suggest that temporal and specific expression regulation of clock-related genes may play an important role in mediating senescence in soybean.

In the plant senescence process, leaf yellowing, which is controlled by developmental factors such as phytohormones and chloroplasts, is a representative phenomenon [20, 21, 81]. Among plant hormones, cytokinin, auxin and gibberellin are known to delay aging senescence, whereas ethylene, salicylic acid (SA), abscisic acid (ABA), and brassinosteroids (BRs) accelerate senescence [82–85]. We found that auxin-responsive factor 1 (ARF1), cytokinin dehydrogenase 3 (CKX3), ethylene-responsive transcription factor (ERF118), abscisic acid 8-hydroxylase-1 (ABAH1) and BRASSINOSTEROID INSENSITIVE receptor kinase 1 (BAK1) were downregulated at 50 and 95 DAS in LS plants compared with those in ES plants (Additional file 7 and 8). In other plants, the loss of CKX, ARF1, ARF2, ABAH1, ERF101, and BAK1 function resulted in a greenish leaf phenotype and delayed senescence compared with those of the control plants [86–91]. The chlorophyll biosynthesis-related genes cysteine desulfhydrase 1 (LCD1), solute carrier family 40 (S40), and chlorophyllide a oxygenase (CAO) were upregulated at 50 and 95 DAS in LS plants compared with those in ES plants (Additional file 7 and 8). Chlorophyll biosynthesis genes have been reported to delay senescence in chlorophyll-overexpressing plants [66, 92–95]. One of those genes, STAY-GREEN (SGR), was downregulated at 50 and 95 DAS in LS plants (Additional file 7 and 8) compared with that in ES plants, which resulted in chlorophyll degradation and delayed senescence [96, 97]. These results indicate that the down- and upregulation of genes are involved in a dynamic process contributing to variation in the two developmental stages through the regulation of senescence.

Conclusion

RNA-Seq was performed to investigate the molecular mechanism underlying senescence in both 50- and 95 days after sowing (DAS) early-senescence (ES) and late-senescence (LS) plants. Expression profiling analysis revealed 2,414 and 2,471 differentially expressed genes at 50 and 95 DAS in ES- and LS-type plants, respectively. Typically, genes related to circadian clock-related genes, phytohormone response genes, chlorophyll biosynthesis-related genes, and senescence-associated protein kinases were significantly enriched, indicating the crucial role of gene expression in the transition from 50 to 95 days in soybean ES- and LS-type plants. Furthermore, we observed variation in the expression of circadian related genes subjected to light and dark conditions during the vegetative stage. According to our findings, the expression of genes associated with senescence, as mentioned earlier, is regulated by the development stage, which in turn regulates senescence in soybeans. Additionally, understanding the identified genes will contribute to further molecular studies on the aging and development of soybeans, including senescence-assisted breeding, to improve crop yield and quality.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

DAS

Days after sowing

DEGs

Differentially expressed genes

CCA1

CIRCADIAN CLOCK ASSOCIATED 1

LHY

LATE ELONGATED HYPOCOTYL

COL9

CONSTANS-LIKE 9

ELF3

EARLY FLOWERING 3

PRR5

PSEUDO RESPONSE REGULATOR 5

ABA

Abscisic acid

JA

Jasmonic acid

BR

Brassinosteroid

SGR1

STAY GREEN

LUX

LUX ARRHYTHMO

RVE1

REVEILLE 1-RELATED

GI

GIGANTEA

LCD1

Cysteine desulfhydrase 1

LED

Light-emitting diode

S40

Solute carrier family 40

CAO

Chlorophyllide a oxygenase

IPK

Inositol polyphosphate kinase

ASNS2

Asparagine synthetase 2

Fd-GOGAT

Ferredoxin-dependent glutamate synthase

SNRK

Sucrose nonfermenting 4-like protein serine/threonine protein kinase

C3H41

Zinc finger CCCH domain-containing protein 41

WAK14

Wall-associated receptor kinase-like 14

NCBI

National Center for Biotechnology Information

FDR

False discovery rate

FPKM

Fragments per kilobase of transcript per million

GO

Gene Ontology

KEGG

Kyoto Encyclopedia of Genes and Genomes

RT‒qPCR

Reverse transcription‒quantitative polymerase chain reaction

Author contributions

I-YC and TU conceived and designed the study and edited the manuscript. PB contributed to data analysis and drafted the manuscript. SL (Sevin LE), KHM, N-IP, G-SL and SL (Seongkon Lee) prepared the samples and conducted the experiments on the substance content. All the authors have read and agreed to the published version of the manuscript.

Funding

This work was supported by the National Research Foundation of Korea (NRF) grant fund (NRF-2020R1I1A3052662 to I-YC), by a research grant from Kangwon National University in 2021 and by the “Cooperative Research Program for Agriculture Science and Technology Development (Project No. RS-2024-00401062 to TU)” Rural Development Administration, Republic of Korea.

Data availability

Data is provided within the manuscript or supplementary information files.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.