Background

Growth-regulating factors (GRFs) are plant-specific transcription factors which regulate plant growth, development and abiotic stress response [1,2,3,4,5]. The first GRF, named OsGRF1, was identified from deepwater rice (Oryza sativa); expression of OsGRF1 was induced by gibberellin (GA), and it mediated stem elongation in a GA-dependent manner [6]. Numerous studies demonstrated that there are two conserved domains, QLQ and WRC, in the N-terminal region of GRF proteins. The QLQ (Gln, Leu, Gln) domain serves as a protein-protein interaction feature which could interact with the GRF-interacting factor (GIF) [7], while the plant-specific WRC (Trp, Arg, Cys) domain comprises a C3H motif for DNA binding and a nuclear localization signal (NLS) [8]. Compared with the conserved nature of the amino acid residue sequence in the N-terminal region, the C-terminal region of the GRFs is variable, with some investigations indicating that the C-terminal region possessed transactivation activity [7,8,9]. In addition, some less-conserved motifs, such as TQL and FFD, are usually present in the C- terminal region of GRFs [10].

As they are development-related transcription factors, it is not strange that GRFs mediate the shape and size of leaves by regulating cell proliferation [11, 12]. Overexpression of AtGRF1, AtGRF2 and AtGRF5 resulted in larger leaves than in wild-type (WT) Arabidopsis, while the leaves of grf mutants, such as grf3–1, grf5–1, grf1–1/grf2, grf2/grf3 and grf1/2/3, were much smaller than the WT [1, 2, 11, 12]. Furthermore, GRFs also regulate root growth, floral development and seed size [9, 13,14,15]. In addition to GRF, the GIFs, are also involved in the control of leaf size and architecture. Compared to the WT, the leaves of gif1 were narrower and smaller [7, 16]. Biochemical analysis showed that GRF and GIF combine to form a transcriptional complex in vivo to modulate cell proliferation and ultimately to control leaf size [5, 16]. Furthermore, it has been shown that the microRNA miR396 directly inhibits the expression of GRFs through post-transcriptional regulation [11, 14]. Constitutive overexpression of Arabidopsis miR396a and miR396b, or heterologous expression of ptc-miR396c (from Populus trichocarpa) and ath-miR396a (from Arabidopsis thaliana) in tobacco (Nicotiana tabacum) significantly reduce mRNA levels of GRFs and lead to narrower and smaller leaves, which mimic Arabidopsis grf1/2/3 [17,18,19,20]. Thus, the transcription levels of GRFs are regulated strictly and quantitatively by the miRNA-GRF-GIF cascade.

Light is an important environmental factor that plays critical roles in plant growth and development, and ultimately determines crop yield [21]. However, in maize (Zea mays)-soybean (Glycine max) relay strip intercropping systems, the light environment in the soybean canopy changes due to it being lower than the maize canopy, so that soybean was under shade stress (resulting from reductions in light quantity and in the red: far-red light ratio) [22, 23]. Furthermore, dense-planting patterns in crops also lead to shade stress among neighboring seedlings [24]. Previous studies had demonstrated that soybean morphological traits changed markedly under shade conditions, resulting in increased plant height, decreased yield, and reduced root length [23, 25, 26]. Notably, leaf expansion is also suppressed when soybean responds to shade stress [27, 28]. However, the specific regulatory mechanisms underlying leaf development under shade conditions are still largely unknown, especially with respect to the GRF family-mediated pathways.

In the present study, 22 GmGRFs were dissected from the soybean genome. Their sequence characteristics, chromosome distribution, phylogenetic relationships, gene structures, conserved motif compositions and synteny were then systematically characterized. Based on these findings, the expression profiles of GmGRFs in various vegetative and reproductive tissues were documented. Furthermore, we also analyzed the expression patterns of GmGRFs under shade stress conditions. These results will not only help us to better understand the functions of the GmGRF family, but will also provide a foundation for improving crops, especially soybean, through genetic modification.

Results

Identification of GmGRFs

Based on the Hidden Markov Model (HMM) of WRC and QLQ domains, a total of 22 GmGRFs were identified from soybean genome and named as GmGRF1 – GmGRF22 according to their locations on chromosomes (Table 1). The length of the coding sequences (CDS) of GmGRFs varied from 927 bp (GmGRF10) to 1830 bp (GmGRF19). Accordingly, GmGRF10, with 308 amino acid residues, was the smallest GmGRF, whereas the largest GmGRF was GmGRF19 (609 amino acid residues). The theoretical molecular weight (MW) of these putative GmGRFs ranged from 34.58 to 66.90 kDa and the isoelectric point (pI) ranged from 6.35 (GmGRF13) to 9.10 (GmGRF1) (Table 1).

Table 1 Characterization of the GmGRF family in soybean
Full size table

GmGRFs were unevenly distributed across the chromosomes. Twenty GmGRFs were distributed on 14 chromosomes and two GmGRFs were located on scaffold_28 (Fig.1). Of the 14 chromosomes, chromosomes 11 and 17 contained the largest number of GmGRFs, each with three genes. Both chromosomes 1 and 9 carried two GmGRFs each, while only one GmGRF was observed on each of chromosomes 3, 4, 6, 7, 10, 12, 13, 15, 16 and 19 (Fig.1).

Fig. 1
figure 1

Distribution of GmGRFs on soybean chromosomes or scaffold

Full size image

Sequence and phylogenetic analysis of GmGRFs

All GmGRFs contained conserved QLQ and WRC domains in their N-terminal regions (Additional file 1: Figure S1). To gain an insight into the evolutionary relationships among GRFs from soybean (22), rice (12) and Arabidopsis (9), MEGA 7.0 software was used to construct a neighbor-joining phylogenetic tree. As shown in Fig. 2, a total of 43 GRFs from different species were clustered into six subgroups (I–VI). Five of the six subgroups contained GmGRFs, whereas subgroup VI had only AtGRF and OsGRF members. Among the six subgroups, subgroups IV and VI were relatively small, with only four GRFs each. By contrast, subgroups I and V contained the largest number of GRFs (ten each), followed by subgroups II (nine), III (six). The phylogenetic tree suggested that the GmGRFs showed a closer relationship with AtGRFs than with OsGRFs, which may be partly because both soybean and Arabidopsis are dicotyledonous plants.

Fig. 2
figure 2

Phylogenetic analysis of GRFs from Glycine max (Gm), Oryza sativa (Os) and Arabidopsis thaliana (At). Clustal W was used to align 43 GRFs, namely nine AtGRF, 12 OsGRF, and 22 GmGRF, while MEGA 7.0 software was employed to construct a neighbor-joining phylogenetic tree with 1000 bootstrap replications

Full size image

Structural analysis of GmGRFs

To further explore the evolutionary relationship among GmGRFs, we constructed a phylogenetic tree and analyzed the gene structures and motif characteristics of GmGRFs (Fig. 3). According to Fig. 2, GmGRFs could be clustered into five groups. It was obvious that the 22 GmGRFs contained two to four introns (six genes with two introns, twelve with three introns, and four with four introns) (Fig. 3c). The conserved structure of GmGRFs was similar to those from other plant species, in which most genes contained three introns [29,30,31]. The lengths of individual GmGRFs were variable in intron length could partly reflect the length of different genes. For instance, the longest gene, GmGRF2, with a size of 5.7 kb, was due mainly to the fact that it contained a total intron length of 4.8 kb.

Fig. 3
figure 3

Phylogenetic analysis, gene structures and conserved motifs of GmGRFs. a The phylogenetic relationship of GmGRFs. A neighbor-joining phylogenetic tree was constructed by MEGA7.0 software with the Poisson model and 1000 bootstrap replications. b Conserved motif arrangements of GmGRFs. Ten conserved motifs labeled with different colors were found in the GmGRF sequences using the MEME program. Among them, motif 1 and motif 2 are the QLQ and WRC conserved domains. c Exon-intron organizations of GmGRFs. The green boxes represent 5’or 3′ untranslated regions, yellow boxes represent the coding sequences, and black lines represent the introns. The lengths of the exons and introns can be determined by the scale at the bottom

Full size image

The Multiple EM for Motif Elicitation (MEME) web server was employed to identify the conserved motifs of GmGRFs (Fig. 3b and Additional file 2: Table S1). All GmGRFs contained motif 1 and motif 2 which were annotated as the GRF specific domains, QLQ and WRC, in their N-terminal regions. Each GmGRF has between four and ten conserved motifs, with the GRFs of subgroup V (GmGRF18, GmGRF19, GmGRF21 and GmGRF22) containing the largest number of motifs. The GmGRFs belonging to the same subgroup have a similar motif composition (e.g. GmGRF4 and GmGRF5, GmGRF18 and GmGRF21). In addition, some motifs appeared in only certain specific subgroups. For example, motif 3 is unique to subgroup I and II, while motifs 7, 8, 9 and 10 are specific to subgroup V. Overall, gene structures and motif characteristics strongly supported the phylogenetic relationships of GmGRFs.

Synteny analysis of GmGRFs

Gene duplication plays an important role in increasing the numbers of genes and their subsequent evolution. For instance, over 90% of regulatory and developmental genes in the Arabidopsis genome had been duplicated [32]. In order to analyze the duplication events of GmGRFs, Multiple Collinearity Scan toolkit X (MCScanX) software was employed. Ultimately, we found that 19 of the 22 GmGRFs (86.36%) were distributed in the duplication regions, suggesting that these genes were generated by large-scale duplication events, whole-genome duplication (WGD) or segmental duplication (Fig. 4 and Additional file 3: Table S2). Additionally, according to a previous methodology [33], GmGRF18-GmGRF19 and GmGRF21-GmGRF22 belonging to chromosome 17 and scaffold_28, respectively, were identified as tandem duplication genes. Further, we calculated the nonsynonymous substitution rate (Ka) and synonymous substitution rate (Ks) of these duplicated gene pairs. The results showed that the Ka/Ks ratios of most GmGRF pairs were less than 1, indicating that these GmGRFs had undergone purifying selection processes.

Fig. 4
figure 4

The syntenic relationships among GmGRFs. The scaffold_28 was not shown in the Circos map

Full size image

Expression profiles of GmGRFs

Based on the analysis of the results from Figs. 2, 3 and 4, we selected eight GmGRFs belonging to different subgroups for GA response and tissue-specific expression analysis. After GA3 treatment, the expression levels of all selected eight GmGRFs were down-regulated (Additional file 4: Figure S2). Then we investigated the expression patterns of these GmGRFs in seven soybean tissues (roots, flowers, stems, pods, leaves, shoot apical meristem, and developing seeds). As shown in Fig. 5, although all GmGRFs were expressed in all seven tissues, the transcription levels of the different genes varied greatly among the different tissues. In general, GmGRFs were highly expressed in growing tissues, such as developing seeds, flowers and shoot apical meristems. Of the GmGRFs, the expression levels of GmGRF1, GmGRF6 and GmGRF18 were highest in developing seeds (Fig. 5a, c and g). Among these three genes, GmGRF6 was more preferentially expressed in developing seeds, indicating 100-fold higher than that in root. However, in flowers, the transcription levels of GmGRF5, GmGRF11 and GmGRF20 were highest (Fig. 5b, e and h). Furthermore, GmGRF9 was abundantly expressed in roots, while GmGRF17 showed the highest transcription level in shoot apical meristems (Fig. 5d and f). The results demonstrated multiple potential functions of GmGRFs in regulating growth and development of distinct soybean tissues.

Fig. 5
figure 5

Expression profiles of GmGRFs in seven tissues. Expression levels of eight selected GmGRFs were examined by qRT-PCR. The housekeeping GmTubulin was used as an endogenous reference gene. R, root; S, stem; L, leaf; F, flower; P, pod; SM, shoot apical meristem; DS, developing seed. Error bars represent standard errors

Full size image

Expression profiles of GmGRFs in response to shade stress

Soybean plant morphology changed greatly in response to shade stress, dubbed as shade avoidance syndrome (SAS), including decreased leaf area and weight, and excessive elongation of stems (Fig. 6 and Additional file 5: Figure S3). Previous studies have demonstrated that GRFs play important roles in regulating leaf size and plant response to abiotic stresses, including abscisic acid (ABA) and osmotic stresses [4, 34,35,36]. Therefore, in order to further explore the roles of GmGRFs under shade conditions, we measured the transcription levels of eight GmGRFs by qRT-PCR. The results showed that the transcription of all the GmGRFs tested was influenced by shade. Expression of almost all GmGRFs was significantly down-regulated under shade stress, albeit to different degrees (Fig. 7). Among these genes, the transcription levels of GmGRF9 and GmGRF17 were more strongly suppressed, while GmGRF20 showed the weakest transcription repression. On the other hand, the expression level of GmGRF5 was up-regulated under shade stress conditions (Fig. 7b). Based on the expression patterns of GmGRFs, GmGRF5, GmGRF9 and GmGRF17 may play important roles in regulating leaf growth under shade conditions.

Fig. 6
figure 6

Soybean leaf area and weight decreased under shade conditions. a Representative photographs of soybean leaves under white light and shade conditions. Bar = 10 mm. Leaf fresh and dry weight (b, c), leaf area (d) and leaf mass per area (LMA; e) were analyzed after shade stress. Ten soybean leaves were measured under each condition. Error bars represent standard errors. The asterisk (*) indicates the significant difference at P < 0.05 by Student’s t-test analysis. L, shade; S, shade

Full size image
Fig. 7
figure 7

Time course of expression patterns of GmGRFs under shade conditions. The housekeeping GmTubulin was used as an internal control. Error bars represent standard errors. The asterisk (*) indicates the significant difference at P < 0.05 by Student’s t-test analysis

Full size image

Discussion

As members of plant-specific gene family, GRFs play important roles in plant growth and development, especially in regulating organ size [5, 11, 15]. Recent studies in rice have shown that OsGRF4 can improve rice yield by promoting cell division and nitrogen uptake efficiency [37,38,39]. Therefore, it is necessary to understand the mechanisms regulating levels of GRFs during plant development and response to stress. To date, GRF family has been identified in various plants (Additional file 6: Table S3), such as Arabidopsis, rice, Zea mays, Brassica napus, Citrus sinensis, Camellia sinensis, Populus trichocarpa and Pyrus×bretschneideri [1, 8, 40,41,42,43]. However, little information on the function of GRFs in soybean is available.

In order to better understand the characteristics and functions of GmGRFs, we studied members of the GmGRF family by bioinformatics assay, qRT-PCR assay, and plant morphology analysis. In the current study, a total of 22 GmGRFs was identified by searching soybean genome database, and all putative GmGRFs were shown to contain QLQ and WRC domains (Table 1 and Additional file 1: Figure S1). Phylogenetic analysis showed that 22 GmGRFs could be clustered into five subgroups according to their evolutionary relationships (Fig. 1). The exon-intron organizations and motif arrangements were consistent with the phylogenetic analysis, while the structure of GmGRFs belonging to different subgroups displayed lower identities.

Numerous studies have shown that gene duplication can not only increase the number of GRFs, but is also an avenue for generating novel genes, a phenomenon which is conducive to enabling the plant to adapt to various environments [44,45,46,47]. Indeed, the expansion of the GRF family occurred mainly through gene duplication, especially large-scale duplications (WGD or segmental duplications) [5, 30, 43]. Consistently, in the present study, most GmGRFs were distributed in duplication blocks, indicating that WGD or segmental duplications had played major roles in the expansion of the GmGRF family. In addition, soybean contains more GRFs than does another legume, Medicago truncatula, a model plant (Additional file 6: Table S3 and Additional file 7: Table S4). This may mainly be due to two WGD events which occurred during the evolution of the soybean genome (58 and 13 million years ago), while the Medicago genome experienced only a single WGD event at 58 million years ago [48, 49]. It is noted that this phenomenon also exists in other soybean gene families, such as the homeodomain-leucine zipper and homeobox gene families [50, 51]. Altogether, these results suggested that large-scale duplications were universal during the expansion process of the GmGRF family.

Previous research had reported that GRFs are pivotal regulators of plant growth and development [5, 34]. For example, overexpression of AtGRF5 leads to leaf enlargement, while the leaves of the Arabidopsis grf5 mutant are narrower than those of the WT [12]. Generally, the expression levels of GRFs in actively growing tissues was higher than that in mature tissues [29]. Indeed, several studies had revealed that, with the aging of organs, the transcription levels of GRFs decreased [19, 41]. In the present study, we found that GmGRFs are highly expressed in shoot apical meristems, developing seeds and flowers (Fig. 5). Interestingly, among all eight selected GmGRFs, GmGRF1, GmGRF6 and GmGRF18 were highly expressed in developing seeds, suggesting that these genes may play important roles in seed development. In addition to developing seeds, GRFs are also important in flower development and root growth [9, 52,53,54]. Correspondingly, we found that some GmGRFs, such as GmGRF9, GmGRF11and GmGRF20, were expressed abundantly in flowers and roots. The gene expression analysis indicated that GmGRFs may play important roles in the growth and development of soybean tissues.

Light not only provides energy for plants, but also regulates plant growth and development [55, 56]. Plants can sense changes in the light environment through photoreceptor systems, such as phytochromes, cryptochromes and phototropins [57,58,59]. As sessile photoautotrophs, plants need to compete with neighbors for light, nutrients and other resources. For example, when exposed to shade stress from neighboring plants, plants adapted to open ranges (e.g. Arabidopsis, rice and soybean) can develop shade avoidance response, leading to shade avoidance syndrome [60,61,62]. Shade avoidance response is conducive to survival of the plant under shade conditions [61, 63]. In the context of agricultural production, however, shade avoidance can cause a decline in crop yield [63]. In maize-soybean intercropping systems, for example, soybean yield decreased significantly, because of the shade stress arising from the taller neighboring maize plants [23]. In the present investigation, the transcription of almost all GmGRF genes tested decreased under shade treatment (Fig. 7), suggesting that GmGRFs possess the veiled functions with regard to plant shade response.

The leaf is the primary tissue of photosynthesis and its size directly affects photosynthetic efficiency [64, 65]. Numerous studies have shown that leaf area is regulated by many genes related to cell division and expansion [36]. Among them, GRFs can regulate leaf area by controlling cell proliferation [65]. In addition, UV-B radiation inhibits leaf growth by decreasing the expression of GRFs in Arabidopsis and maize [66, 67]. Morphological analysis showed that soybean leaf size, leaf area and dry and fresh weight decreased significantly under shade stress (Fig. 6), and, in line with this, the expression levels of almost all GmGRF genes were down-regulated under shade conditions (Fig. 7). Given the positive regulatory effect of GRFs on plant cell proliferation in diverse plant species [68], these results were consistent (Figs. 6 and 7). Interestingly, under shade conditions, the expression levels of most GmGRFs were down-regulated while GmGRF5 transcription was up-regulated under these conditions. Therefore, further study of the function of GmGRF5 under shade stress may be helpful in increasing the leaf area of intercropped soybean, thereby potentially increasing the yield of soybean.

Conclusions

In this study, we systematically analyzed the basic characteristics and functions of GmGRFs. A total of 22 GmGRFs were identified from the soybean genome with the help of the HMM of two conserved domains specific to GRF, namely QLQ and WRC. These GmGRFs were distributed on 14 chromosomes and one scaffold, and could be clustered into five subgroups according to their phylogenetic relationships. By further analysis, we found that GmGRFs belonging to the same subgroup had similar gene structures and motif compositions. Gene duplication analysis suggested that both large-scale and tandem duplications, especially the former, contributed to the expansion of the GmGRF family. The result of qRT-PCR studies showed that GmGRFs were involved in the growth and development of soybean, as well as in shade response. Importantly, we identified some potentially useful genes, such as GmGRF1, GmGRF5 and GmGRF6 by analyzing the GmGRF family. These results will provide a foundation for future research on the GRFs in soybean.

Methods

Plant materials and shade treatment

Soybean cultivar Nandou-12, a prevailing cultivar in southwestern China, was employed in this study. The seeds of Nandou-12 were originally obtained from Nanchong Institute of Agricultural Sciences, Sichuan Province, China. Soybean plants were grown in a phytotron with 65% relative humidity, a 12-h light (25 °C)/12-h dark (20 °C) photoperiod and a light intensity of 380 μmol m− 2 s− 1. In order to analyze the expression patterns of GmGRFs in different tissues of soybean, we collected roots (16 d after sowing), stems (16 d after sowing), leaves (16 d after sowing), shoot apical meristems (16 d after sowing), flowers (36 d after sowing), pods (44 d after sowing) and developing seeds (60 d after sowing). For the GA treatment, six-day soybean seedlings were sprayed with 100 μM GA3 and the hypocotyls of soybean seedlings were harvested after 0, 3, 6 and 9 h. Furthermore, to explore the transcription profiles of GmGRFs under shade stress, black nylon net and far-red light-emitting diode (LED) were employed to adjust the light environment (light intensity and quality) [26]. 10-day soybean seedlings were transferred to the shade environment with 113 μmol m− 2 s− 1 of photosynthetically active radiation (PAR) and 0.4 of red: far-red light ratio (R/FR). The first compound leaves of soybean seedlings were collected at 0, 3, 6 and 9 h after shade treatment. All samples were frozen immediately in liquid nitrogen and then stored at − 80 °C for further analysis.

Seven days after shade treatment, ten soybean seedlings from different pots were used for analyzing morphological characteristics. Leaf area of first compound leaves was measured by ImageJ software. Seeding height, stem diameter, leaf and root fresh weight were also determined. Subsequently, leaves, stems and roots were exposed to 105 °C for 0.5 h and then dried at 80 °C until constant weight. Finally, dry weight, root-shoot ratio and leaf mass per area (LMA) were also calculated.

Identification of GmGRFs

Soybean protein sequences and genome annotation were downloaded from Phytozome database (http://www.phytozome.net/). The HMM of WRC (PF08879) and QLQ (PF08880) domains were obtained from PFAM database (http://pfam.xfam.org/) and used to predict GmGRFs with HMMER software. Then, these protein sequences (E-value ≤1e− 10) were used to construct soybean-specific HMM for identifying GmGRFs [69]. Finally, predicted proteins were considered as GmGRFs only if they contained QLQ and WRC conserved domains verified by SMART software (http://smart.embl-heidelberg.de/) and PFAM databases. All putative GmGRFs were drawn on chromosomes using MapGene2Chrom web v2 (http://mg2c.iask.in/mg2c_v2.0/).

Sequence and phylogenetic analysis

The exon-intron distribution patterns of GmGRFs were analyzed using Dual Systeny Plotter software (https://github.com/CJ-Chen/TBtools) [70, 71]. To predict MW and pI of GmGRFs, the ExPASy proteomics server (https://web.expasy.org/protparam/) was used [72]. Conserved motifs of GmGRFs were predicted by MEME online program (http://meme-suite.org/tools/meme) [73]. Multiple sequence alignments of GmGRFs were analyzed by Clustal W [74]. MEGA 7.0 software was employed to construct phylogenetic trees using the neighbor-joining method with Poisson model, pairwise deletion, and 1000 bootstrap replications [75].

Gene duplication and evolution analysis

MCScanX program with the default parameters was used to analyze the duplication events of GmGRFs [76]. According to the results of MCScanX, the nonsynonymous substitution rate (Ka) and synonymous substitution rate (Ks) of duplicated genes were calculated by KaKs_Calculator 2.0 software [77].

Gene expression analysis

Total RNA preparation, first-strand cDNA synthesis and the qRT-PCR assay were performed as previously described [78]. According to the manufacturer’s protocol, total RNA was treated with DNase I and then 2 μg total RNA was reverse-transcribed using Moloney murine leukemia virus reverse transcriptase (200 units per reaction; Promega Corporation). The qRT-PCR was performed using Vazyme™ AceQ qPCR SYBR Green Master mix on a QuantStudio 6 Flex Real-Time PCR System (Thermo Fisher Scientific, USA) [79]. The soybean housekeeping GmTubulin was used as an endogenous reference gene and each reaction had three repetitions. 10 μL reaction mixture included 5 μL Vazyme™ AceQ qPCR SYBR Green Master mix, 0.2 μL forward primer, 0.2 μL reverse primer, 1 μL cDNA template and 3.6 μL Dnase-free ddH2O. The qRT-PCR reaction procedure was set as follows: 95 °C for 30s, and then 40 cycles of 95 °C for 15 s, 60 °C for 30 s. The expression levels of GmGRFs were calculated by the comparative CT method [80]. Sequences of the primers for qRT-PCR were listed in Additional file 8: Table S5.

Statistical analysis

In this study, the Student’s t-test in Microsoft Excel software was used to analyze the differences between different treatments. P < 0.05 was considered as statistical significance.