Integrated analysis of transcriptomic and proteomic data reveals novel regulators of soybean (Glycine max) hypocotyl development

Plant material

For the study of proteome and transcriptome change during hypocotyl development, the soybean (Glycine max L.) cultivar ‘Williams 82’ was used. For the comparison of proteomes of cultivars with longer and shorter hypocotyls, cultivar ‘Liu Tiao Qing’ (named as SJZ1045 in this study), ‘Lv 75’ (SJZ190), ‘Ban Ye Sheng’ (SJZ292) and ‘PI496 A’ (SJZ1234) were used. For the ectopic expression of soybean genes in Arabidopsis thaliana L., the ecotype Col-0 was used.

All soybean seeds were germinated in vermiculite at 25°C, under dark conditions unless specified. Soybean seeds used in this study were harvested in 2018, from the same field at Shijiazhuang, Hebei Province, China, and stored at 5°C until they were germinated.

To generate transgenic lines overexpressing GmSTS, GmPHYA1 and GmPIP1-6, the full-length coding sequences (CDS) of these genes were amplified and cloned into the pCambia1300-eCFP vector using a homologous recombination-based method. Transgenic Arabidopsis lines were generated using the floral dip method. Positive transgenic lines were screened on Murashige and Skoog (MS) medium with 50 μg/mL of hygromycin, followed by real-time quantitative PCR (qPCR) for confirmation of overexpression with gene-specific primers (data not shown). The T1 transgenic seedlings were used for subsequent experiments.

Proteomics assay

Proteomes of hypocotyl samples were investigated using tandem mass tags (TMT)-based quantitative proteomic assay, performed by PTM Biolabs (Hangzhou, China). Briefly, for each sample, total protein from three biological replicates was extracted using the protamine sulfate (PS) precipitation method with trichloroacetic acid (TCA)/acetone precipitation and digested using a filter-aided sample preparation (FASP) approach. The digested peptides were collected and labelled with TMT. After fractionation and separation, the peptides were analysed with liquid chromatography tandem mass spectrometry. Data from the mass spectrometry were analysed using the MaxQuant software (ver.1.5.2.8). The relative amount of each protein was quantified using the label-free quantification (LFQ) method.

Protein extraction and quantification

For storage protein extraction, soybean hypocotyls were homogenised in liquid nitrogen with pre-cooled mortar and pestle. The homogenate was mixed with 10 times of diethyl ether and kept at room temperature for 8 h. The supernatant was then removed and the precipitate was kept and air-dried. For every 500 mg precipitate, 5 mL 0.2 M Tris-HCl buffer (pH 7.8) was added, and the mixed solution was blended at 4°C for 9 h and then centrifuged at 17 800g for 30 min. The supernatant was collected as seed storage protein solution and subjected to electrophoresis. The protein concentration was determined by the Bradford method using BSA as the standard.

Hydrogen peroxide quantification

Content of hydrogen peroxide in soybean hypocotyls was measured using a hydrogen peroxide content assay kit (Solarbio, Beijing, China) following the manufacturer’s instruction. Briefly, soybean hypocotyls were homogenised in acetone on ice and then centrifuged. The supernatant was collected, mixed with the reaction buffer and centrifuged again. The pellet was kept and dissolved in the dissolution buffer. The amount of hydrogen peroxide was calculated by measuring the absorbance at 415 nm of the solution.

Sugar extraction and quantification

Total soluble sugar, including sucrose and maltose that can be hydrolysed to reduced monosaccharides and starch that can be partially hydrolysed to glucose, was extracted and quantified using a glycogen content assay kit (Boxbio, Beijing, China) following the manufacturer’s instruction. Briefly, soybean hypocotyls were ground into powder, added with the extraction buffer and treated with boiling water bath. The suspension was centrifuged and the supernatant was collected to measure the sugar content using a standard curve method provided by the manufacturer.

Bioinformatics analysis

The R package clusterProfiler was used to perform the Gene Ontology (GO) and Kyoto Encyclopaedia of Genes and Genomes (KEGG) enrichment analysis. Fisher’s exact test was used to determine whether there are significant differences between samples. New PLACE (https://www.dna.affrc.go.jp/PLACE) was used to perform the prediction of cis-acting regulatory elements, using the 2.5 kb upstream sequence from ATG start codon of target genes.

RNA extraction and qPCR assay

Total RNA was extracted from soybean hypocotyls using a plant RNAprep Kit (TianGen, Beijing, China). RNA concentration and purity were checked with the Qubit RNA Assay Kit using Qubit 2.0 Fluorometer (Life Technologies, CA, USA). First-strand complementary DNA (cDNA) was synthesised with 0.5–1 μg total RNA using a Reverse Transcription kit (Vazyme, Nanjing, China) and used as the template for qPCR, which was performed using a CFX96 Real-Time PCR System (Bio-Rad, CA, USA). The relative expression levels of target genes were determined using the standard-curve method as described before (Shen and Chen 2022). The reference gene was a soybean ACTIN gene (GLYMA_02G091900).

Translation efficiency (TE) analysis of proteins expressed in soybean hypocotyls during development

The seed germination and subsequent seedling development of soybean are accompanied by dramatic changes in transcriptome and proteome. In order to study the regulation mechanism of soybean hypocotyl development, we sampled hypocotyls from soybean dry seeds (undeveloped hypocotyls) and 48-h germinated seeds (developing hypocotyls) and sequenced the transcriptomes and proteomes. We obtained the expression data of 57 147 transcripts and 11 269 proteins, respectively (see Supplementary Fig. S1). By comparing the expression level (LFQ intensity) of each protein and its corresponding transcript abundance (fragments per kilobase of transcript per million, FPKM), we found that there was a weak positive correlation between protein and transcript abundance in undeveloped and developing hypocotyls (R value was 0.42 and 0.16, respectively) (Fig. 1a, b). This result is consistent with the earlier research (Galland and Rajjou 2015), indicating that expression of many hypocotyl proteins might be regulated by factors other than transcription, such as protein stability, protein degradation, mRNA stability and protein transport.

Fig. 1.

Protein and mRNA expression during soybean hypocotyl development. (a, b) Scatter plot of protein-vs-transcript abundance in undeveloped (a) and developing (b) hypocotyls. R denotes Pearson’s correlation coefficient. FPKM, fragments per kilobase of transcript per million; LFQ, label-free quantitation intensity. (c, d) Gene Ontology cell component (GO-CC) and Kyoto Encyclopaedia of Genes and Genomes (KEGG) analysis results of genes with extreme high TE value in undeveloped (c) and developing (d) hypocotyls.

TE is the rate at which mRNA is translated into protein in cells, referred to as the rate of protein production per mRNA. Modulating TE allows the cell to fine-tune the spatiotemporal expression of genes. Several studies have shown that many proteins expressed in plant seeds are with abnormally higher or lower TE than the average level, which may play important roles in seed germination and subsequent seedling development (Galland and Rajjou 2015). By calculating the TE value of each protein in our study (TE = log₂(LFQ intensity/FPKM)), we found that the overall TE is higher in developing hypocotyls (median value 12.75) than undeveloped ones (median value 11.79), indicating that proteins are more regulated by post transcription mechanisms along with hypocotyl development in general. To further study the function of proteins with abnormal TE, we selected 500 proteins with highest TE (TE > 17.5) in undeveloped and developing hypocotyls (TE > 20), respectively, where 90% of the proteins are with a TE value less than 15.6 in both situations (Fig. 1a, b).

Proteins with extreme high TE in soybean hypocotyls

The GO analysis results showed that in the undeveloped hypocotyls, most of the proteins with highest TE are storage vacuole components, including many seed storage proteins (SSPs) (Fig. 1c, Table S1). It is not surprising as SSPs are synthesised and deposited into storage vacuoles during seed maturation and stably accumulated to high levels for the preparation of germination. Consistent with this result, KEGG analysis showed that many proteins with high TE in undeveloped hypocotyls participate in the metabolism of nutrients, including amino acid, starch, sugar and fat acid (Fig. 1c). Correspondingly, in developing hypocotyls, more abnormal proteins are involved in the synthesis of amino acid, ribose and cofactors (Fig. 1d), indicating that hypocotyl development relies on the energy and metabolic intermediates derived from the mobilisation of stored nutrients, including SSPs.

Proteins with highest TE in developing hypocotyls include 10 XTHs, four pectate lyases (PELs), three pectin acetylesterases (PAEs), six pectin esterases (PEs) and four polygalacturonase (PGs) (Table S2). These enzymes were previously reported to be involved in the hydrolysis and reorganisation of cellulose molecules in cell wall; thus, allowing the expansion and growth of hypocotyl cells (Osato et al. 2006; Derbyshire et al. 2007; Pelletier et al. 2010; Miedes et al. 2013). Interestingly, by comparing the protein levels of them in soybean cultivars with different length of hypocotyls (Fig. S1, S2), we found that the XTHs were with higher protein levels in varieties with longer hypocotyls (Fig. S3a), which is consistent with their potential role in promoting hypocotyl elongation. A similar trend was observed for the PELs and PAEs, while it was opposite for the PEs (Fig. S3b), suggesting that these PAEs and PELs are positive regulators of hypocotyl elongation and PEs are negative ones.

Interestingly, there were many oxidoreductases with extreme high TE in undeveloped and developing hypocotyls. For example, five lipoxygenases and four superoxide dismutases (SODs) were found to be with highest TE in undeveloped hypocotyls (Table S1) and 34 peroxidases with similar pattern were found in developing hypocotyls (Table S2). These enzymes could catalyse the formation of reactive oxygen species (ROS), including hydrogen peroxide (H₂O₂), or use H₂O₂ as the electron acceptor to catalyse downstream oxidative reactions (Quan et al. 2008; Petrov and Van Breusegem 2012). Previously reports have shown that intracellular H₂O₂ can break off polysaccharides to loosen the cell wall structure and play a vital role in the cell elongation growth (Schopfer 1996). Other results suggested that H₂O₂ could selectively oxidises proteins and mRNAs during post-germination seedling establishment (Wojtyla et al. 2016). These results lead us to hypothesise that H₂O₂ play a role in promoting soybean hypocotyl elongation. In our study, H₂O₂ was shown to be maintained a high level during the development of soybean hypocotyls (Fig. 2a) and the overall H₂O₂ content and peroxidase protein levels is higher in varieties with longer hypocotyls (Figs 2b and S3c). Most importantly, exogenous application of H₂O₂ significantly promoted the hypocotyl elongation while increasing the endogenous H₂O₂ content (Fig. 2c, d). All these results showed that H₂O₂ is indeed a positive regulator of soybean hypocotyl elongation and the enzymes related to its production and metabolism may play important roles in hypocotyl development.

Fig. 2.

H₂O₂ is a promoting factor of soybean hypocotyl elongation. (a) Endogenous H₂O₂ concentration (μmol/g) in soybean hypocotyls dissected from dry seeds (ds) and seedlings germinated for 3, 6, 9, 12 and 24 h (hours after germination, HAG). (b) H₂O₂ concentration in 5-day-old hypocotyls of different varieties (with longer and shorter hypocotyls, respectively). (c) Exogenous application of H₂O₂ promoted the elongation of hypocotyls. The hypocotyl length was measured 3 days after germination on water agar (0.9%) medium without or with different levels of H₂O₂. (d) Exogenous H₂O₂ increased the endogenous H₂O₂ content in 3-day-old hypocotyls. Data are presented as mean ± s.d. with n = 3 in (a), n ≥ 5 in (b) and n ≥ 10 in (c). Different lower case letters indicate significant difference in (b). *P < 0.05 relative to mock (without treatment) in (c).

Proteins with dramatic TE change during hypocotyl development

Many proteins were shown to be with dramatically changed TE during hypocotyl elongation in our experiments (Figs 3a and 4a), which may be associated with soybean hypocotyl development. We therefore selected the proteins with significant increase (TE48-TE0 > 5) (Fig. 3a) and decrease (TE48-TE0 < −5) (Fig. 4a) of TE for further study, where the TE value of 95% of the proteins changed between −3.3 and 3.3. It is worth noting that during the development of hypocotyls, the number of proteins with increased TE value was much bigger than those with decreased TE, reflecting the exuberant protein synthesis activity during hypocotyl development.

Fig. 3.

Proteins with rapidly decreased TE during hypocotyl development. (a) A large number of proteins in hypocotyls are with rapidly decreased TE along with germination. (b) GO analysis result of these proteins. (c) Expression heat map of these proteins in 4-day-old hypocotyls of different varieties. (d) Expression of selected proteins in 4-day-old hypocotyls of different varieties. Data are from label-free quantitative proteomic data and presented as mean ± s.d. (n = 3). *P < 0.05 relative to SJZ1045. The protein/gene names corresponding to the abbreviations are in Table S3. SHs, short hypocotyls; LHs, long hypocotyls.

Fig. 4.

Proteins with rapidly increased TE during hypocotyl development. (a) A large number of proteins in hypocotyls are with rapidly increased TE along with germination. (b) GO analysis result of these proteins. (c) Expression heat map of these proteins in 4-day-old hypocotyls of different varieties. (d) Expression of selected proteins in 4-day-old hypocotyls of different varieties. Data are from label-free quantitative proteomic data and presented as mean ± s.d. (n = 3). *P < 0.05 relative to SJZ1045. The protein/gene names corresponding to the abbreviations are in Table S4. SHs, short hypocotyls; LHs, long hypocotyls.

Proteins with rapidly decreased TE during hypocotyl development

Proteins with decreased TE are more likely to play a role during the early stage of seed germination (imbibition and nutrition digestion). The decrease of TE is usually due to protein degradation and/or translation stagnation. In fact, the abundances of proteins with significant decreased TE were shown to descend rapidly during hypocotyl development (Fig. S4a), while the corresponding transcript levels went down synchronously (Fig. S4b), indicating that these proteins are more likely to be degraded along with hypocotyl development. GO analysis showed that these proteins are mainly involved in response to abiotic stimulus and ribosome biogenesis/RNA binding (Fig. 3b), implying that they may be accumulated in undeveloped hypocotyl for the stress response and the preparation for translation initiation upon germination. Besides, many seed maturation proteins (late embryogenesis abundant (LEA) proteins, dehydrins, etc.) (Table S3) were found to be included in this category, which were reported to be involved in responses to stress such as desiccation during seed development and early germination (Battaglia and Covarrubias 2013). Interestingly, most of these seed maturation proteins were with higher expression in cultivars with shorter hypocotyls (Fig. 3c, d). Previous studies have shown that dehydrins and LEA proteins play important roles in plant response to abiotic stresses (Battaglia and Covarrubias 2013; Sun et al. 2021). Thus, the increased expression of these proteins in shorter hypocotyl samples may reflect a higher level of stress response, which may consume additional energy and inhibit hypocotyl elongation.

Proteins with reduced TE also include a lot of SSPs, which were thought to be accumulated during embryogenesis and rapidly mobilised upon seed germination to fuel organ development (Golombek et al. 2001; Müntz et al. 2001). As mentioned above, many of SSPs are with extreme high TE in undeveloped hypocotyls (Table S1). It is not surprising that abundances of these proteins were decreased rapidly upon seed germination, as well their transcripts, which partly explains the decrease of TE levels (Fig. S4). Strikingly, the abundances of residual SSPs were much higher in developing hypocotyls of cultivars with longer hypocotyls than shorter ones, indicating that the mobilisation efficiency of SSPs is positively related to the hypocotyl elongation (Fig. 3d). Consistent with this result, many enzymes related to protein degradation were shown to be with significant increased TE during hypocotyl development (see below), while a potential negative regulator of storage nutrient metabolism (GLYMA_19G217700, encoding a stachyose synthase), was shown to be with significant decreased TE (Table S3). Ectopic expression of this soybean stachyose synthase (STS) in Arabidopsis led to shorter hypocotyls (Fig. 5a), confirming that the mobilisation efficiency of seed storage substances regulates hypocotyl elongation.

Fig. 5.

The hypocotyl phenotype of Arabidopsis seedlings overexpressing (a) GmSTS, (b) GmPHYA1 and (c) GmPIP1-6. In each figure, the left are representative photos of seedling and the right are statistical results. Data are presented as mean ± s.d. (n ≥ 10). *P < 0.05 relative to the wild-type (WT).

Proteins with elevated TE during hypocotyl development

Most proteins with dramatically increased TE were shown to be with increased protein abundances along with hypocotyl development (Fig. 4a), while the corresponding transcript levels did not change as much (data not shown), indicating that the increase of TE levels may be attributed to increased protein productivity and/or less protein degradation. GO analysis showed that many of these proteins are involved in cell wall organisation and sugar/amino acid metabolic processes (Fig. 4b), which may contribute to the rapid growth at the late stage of germination. The rapidly increased production rate of these proteins ensured the fast hypocotyl growth upon germination.

Proteins with increased TE showed diverse differences between varieties with longer and shorter hypocotyls (Fig. 4c). Some were shown to be expressed higher in varieties with shorter hypocotyls, including UDP glycosyltransferase (UGTs) encoded by GLYMA_19G187400 and GLYMA_08G244700, and multiple stress or disease responsive proteins, such as thaumatin-like proteins encoded by GLYMA_08G053600 and GLYMA_05G245800, a stress response protein encoded by GLYMA_13G318600, a malate synthase encoded by GLYMA_17G128000, a glycine-rich protein encoded by GLYMA_13G244800 and an immune related protein encoded by GLYMA_19G176700 (Fig. 4c, Table S4). Previous studies have shown that UGTs play key roles in modulating auxin dependent hypocotyl elongation via buffering auxin homeostasis in Arabidopsis (Mateo-Bonmatí et al. 2021); therefore, the phenotype of shorter hypocotyls may be related to the excessive expression of these UGTs. For stress or disease resistance proteins, as hypocotyls are vulnerable structures compared to seeds, their increased expressions may be required to protect hypocotyls from adverse environment and pathogens. However, considering that the stress and disease resistance processes are energy consuming and might compete with growth and development, it is easy to understand that these proteins are expressed more highly in shorter hypocotyl varieties.

Other proteins were expressed higher in cultivars with shorter hypocotyls, including three aspartic proteases, two plasma membrane intrinsic proteins (PIPs) and two PHYAs (Fig. 4c, d, Table S4). It is worth noting that for these aspartic proteases, PIPs, PHYs and their homologues, most of them were with elevated TE during hypocotyl development (for aspartic proteases, 17 out of 22; for PIPs, 7 out of 8; for phytochromes, 3 out of 3), indicating that these proteins may play important roles in hypocotyl elongation. Among them, aspartic proteases are involved in the degradation of SSPs and participate in regulating seed germination as reported in Arabidopsis (Shen et al. 2018). In our results, we found that the protein and transcript abundances of these aspartic proteases were quickly increased during hypocotyl development, especially for two of them, GLYMA_06G093600 and GLYMA_18G090900. Our data showed that their transcription and translation were both started over from zero upon seed germination and may be responsible for the mobilisation of SSPs during hypocotyl development. Consistent with this assumption, the cultivars with longer hypocotyls contained more carbohydrates as the product of protein mobilisation in the developing hypocotyls (Fig. S5b), although there is no significant difference of SSPs in undeveloped hypocotyls from all the varieties (Fig. S5a). Thus, the protein mobilisation efficiency in the cultivars with longer hypocotyls may be more sufficient to promote hypocotyl elongation.

For the two PHYA homologues (PHYA1 and PHYA2) with increased TE (Fig. S6a), it was shown that their protein and mRNA abundances were increased in the wake of hypocotyl development (Fig. S6b), suggesting that they were actively transcribed and translated upon seed germination. Our results showed that these two soybean PHYAs were with higher protein abundances in the cultivars with shorter hypocotyls (Fig. 4c, d, Table S4), indicating their inhibitory role in hypocotyl development, which is consistent with the results in Arabidopsis (Yang et al. 2009; Sheerin et al. 2015). Not surprisingly, ectopic expression of PHYA1 in Arabidopsis caused shorter hypocotyls (Fig. 5b). Therefore, these two PHYAs are negative regulators of hypocotyl elongation.

For the two PIPs (PIP2-14 and PIP1-6) with rapidly elevated TE (Fig. S6), they were with higher expression in the varieties with longer hypocotyls (Fig. 4c, d, Table S4). PIPs belong to plant major intrinsic protein (MIP) family and act as water channels through plasma membrane, playing important roles in plant water homeostasis. Plant PIPs were reported to be involved in multiple biological processes, including nutrient transport, seed germination and the lateral root emergence (Péret et al. 2012; Liu et al. 2013; Lu et al. 2018). There are 22 genes encoding PIPs in soybean genome, including eight PIP1 and 14 PIP2. Interestingly, there are four other PIPs (PIP2-8, PIP2-5, PIP2-1 and PIP1-5) with higher expression in the cultivars with longer hypocotyls (Fig. 4c, d, Table S4), indicating that PIPs may participate in soybean hypocotyl development by regulating water transport during cell elongation. To confirm this, we overexpressed PIP1-6 in Arabidopsis, and found that the hypocotyls of transgenic plant seedlings were significantly longer than those of wild-type (Fig. 5c). Therefore, these soybean PIPs may play as positive regulators of hypocotyl elongation.

The regulation of gene expression is mostly carried out through transcription factors binding to cis elements on promoters. By analysing the upstream promoters of genes encoding the above potential regulators of hypocotyl development, we found that most of promoters contain at least one of the elements responsive to light (e.g. G-box, a specific binding site for PIF transcription factors) and phytohormones (e.g. GARE motif, a cis acting element in response to GA; AuxRR core, a binding site for ARFs) (Fig. 6a and S7), suggesting that environmental factors and hormone signals may regulate hypocotyl elongation through these regulators (Fig. 6a). To confirm this, we measured the expression of two PIPs in hypocotyls with light and auxin/GA treatment. We found that light could indeed inhibit the expression of PIP2-8 and PIP2-14 in hypocotyls (Fig. 6b, c), and exogenous auxin could induce the expression of these two genes (Fig. 6b), but the exogenous GA didn’t show significant effect (Fig. 6c).

Fig. 6.

The potential regulators of hypocotyl elongation are under regulation of light and phytohormones. (a) The promoter regions (2.5 kb) of these potential regulators contain multiple light, auxin and GA responsive elements. a: PIPs, b: PGs, c: aspartic proteases, d: XTHs, e: PHYAs. (b) Expression of two GmPIPs in the hypocotyl after 1-day treatment of light or 2 h of exogenous 10 μM IAA. (c) Expression of two GmPIPs in the hypocotyl after 1-day treatment of light or 2 h of exogenous 5 μM GA₃. *P < 0.05 relative to the negative control (without treatment).