Introduction

Transcription factors are a class of DNA-binding proteins that regulate gene transcription by binding specifically to cis-acting elements in the promoter region of eukaryotic genes, through interactions with each other proteins [1]. GRAS gene family is present only in higher plants [2]. Its members have unique GRAS domains, and some of them also have DELLA protein structures. These domains are closely related to physiological processes such as plant growth, metabolism, and stress adaptation [2]. GRAS protein is named after the characteristic letters of the three members initially discovered: GAI (GIBBERELLIN INSENSITIVE) [3], RGA (REPRESSOR OF GAL-3) [4], and SCR (SCARECROW) [5]. The members of the GRAS protein family generally consist of 400 to 700 amino acid residues. The length and sequence of amino acids are highly complex due to their N-terminal structures, while the C-terminal amino acids are relatively conserved [6]. In general, the typical structural domains of the GRAS family include LHR I (Leucine heptad repeat I), VHIID (Val-His-Ile-Ile-Asp), LHR II, PFYRE (Pro-Phe-Tyr-Arg-Glu), and SAW (Ser-Ala-Trp) [7]. VHIID is considered the core region as it is highly conserved. It binds with two leucine heptad repeat regions to form LHR I - VHIID - LHR II complexes, regulating the binding activity with DNA and other proteins [6, 8]. Moreover, these two leucine-rich regions are composed of about 100 amino acid residues. In most cases, these two regions do not form a complete unit every seven residues, which distinguishes them from the Leucine zipper [9]. There is a hypothetical nuclear localization signal in the LHR I region at the C - terminus, therefore the SV40-type sequence could be recognized [10, 11]. Some LHR I motifs in the N-terminal of GRAS proteins contain a conserved LXXLL sequence (Leu-X-X-Leu, X represents any amino acid), which is common in most GRAS proteins [12,13,14,15]. However, the roles of the LXXLL sequences in mediating the interactions of plant GRAS proteins with their regulators and co-activators are still unclear. PFYRE motif, which contains a tyrosine phosphorylation site, is not as conserved as the VHIID region, but still exhibits high similarity and collinearity in most GRAS proteins. This motif typically consists of three parts: proline residue (P), phenylalanine residue (F), tyrosine residue (Y), arginine residue (R), and glutamate residue (E) [6, 11]. The functions of the PFYRE and SAW motifs are not fully understood yet, but they both have conserved amino acid residues or pairs, suggesting that these two structural motifs are important for protein function or stability [16]. N-terminal region of GRAS proteins is flexible and variable in length and sequence, forming inherently disordered regions (IDRs) that adopt specific molecular recognition features upon binding [9]. The diverse N-terminal interacts with different target proteins during expression, acting cooperatively and exhibiting protein specificity, which plays a key role in signal transduction pathways, depending on the different members or expression conditions [9, 16, 17].

According to the members of the GRAS family in the genomes of Arabidopsis and rice, this family can be divided into eight branches, including SCL3 (SCARECROW - LIKE3), SHR (SHORT ROOT), PAT1 (PHYTOCHROME A SIGNAL TRANSACTION), LISCL (Lilium longiflorum SCR like), DELLA, SCR (GAI - RGA - SCR), LAS (LATERAL SUPPRESSOR), and HAM (HAIRY MERISTEM) [18]. These subfamilies play their respective roles in plant growth, development, and metabolic regulation. Cenci and Rouard [7] also analyzed the GRAS transcription factors in various angiosperms, who found that there were other subfamilies such as DLT (Dwarf and Low Tillering, NSP1 (Nodulation Signaling Pathway 1), NSP2 besides the above eight subfamilies. Currently, the GRAS family has been reported to exist in over 50 plants, including Arabidopsis thaliana (n = 33) [19], Brachypodium distachyon (n = 48) [20], Brassica napus (n = 92) [21], Capsicum annuum (n = 50) [22], Chenopodium quinoa (n = 52) [23], Citrullus lanatus (n = 37) [24], Citrus sinensis (n = 50) [25], Fagopyrum tataricum (n = 47) [26], Glycine max (n = 117) [27], Hordeum vulgare (n = 62) [28], Jatropha curcas (n = 48) [29], Litchi chinensis (n = 48) [30], Malus domestica (n = 127) [31], Manihot esculenta (n = 77) [32], Medicago sativa (n = 51) [33], Oryza sativa (n = 57) [34], Phaseolus vulgaris (n = 55) [35], Ricinus communes (n = 48) [36], Setaria italica (n = 57) [37], Solanum lycopersicum (n = 54) [38], Sorghum bicolor (n = 81) [39], Triticum aestivum (n = 188) [40], Vitis vinifera (n = 52) [41], Zea mays (n = 86) [42], et al.

The GRAS family comprises diverse subfamilies with distinct structural and functional features. Members of different subfamilies may participate in various processes of plant growth, development and environmental adaptation [43,44,45,46,47]. The SCR is co-localized with SHR in the vascular bundle sheath cells of leaves and roots [48]. PAT1, SCL13, and SCL21 are members of the PAT1 subfamily and are implicated in regulating light signal transduction [47, 49]. DELLA is involved in the response to plant hormone signals, such as gibberellin, jasmonic acid, and auxin [50,51,52]. The protein phosphorylation and dephosphorylation processes that regulate GA signaling in plants are generally mediated by the proteasome-dependent destabilization of DELLA protein repressors, which modulate the response to endogenous gibberellins. Leaf elongation in seedlings that relies on the gibberellin pathway is governed by the proteasome-mediated derepression of DELLA [53]. LlDELLA1 facilitates flower and pod development in Lupinus luteus. Its expression level slightly declines from the flower bud stage to anther opening, but rapidly elevates during pollination, fertilization, podding, and early grain development [54]. LISCL is implicated in the meiosis of pollen and facilitates the formation of microspores in L. longiflorum [55]. HAM family members from various flowering plants sustain the indeterminacy of shoot meristem and facilitate the formation of re-axillary meristem [45, 56,57,58,59]. The loss-of-function of HAM leads to a defect in shoot apical meristem in Capsicum annuum [58]. PhHAM is specifically expressed in the vascular tissue of stem primordia in petunia, which plays a vital role in sustaining the activity of inter shoot apical meristem [59]. In Arabidopsis, DELLAs, SCL3, and IDDs constitute a “co-activator/co-repressor exchange regulation system” to fine-tune the feedback regulation of gibberellin [60]. Through the interactions and transcriptional networks among these proteins, they partake in various signaling pathways and physiological events in multiple aspects. DLT, OSH1, and OsOFP19 form functional complexes that play a pivotal role in brassinolide signaling and determining cell division patterns during plant growth and grain development in rice [61]. OsMOC1 is one of the key factors in determining the number of tillers in rice, which is essential for axillary meristem (AM) formation and bud growth [62]. Furthermore, salt, ultraviolet radiation, flooding, drought, and extreme temperatures can inflict irreversible damage to crop growth and development, ultimately impeding growth and diminishing yield [63]. Some studies have demonstrated that GRAS genes play a crucial regulatory role in plant responses to stress. NtGRAS1 partakes in the phosphorylation process of reactive oxygen species and nitric oxide stress induction in cells, thereby regulating the homeostasis of nutrient distribution within cells [64]. PeSCL7 is induced by drought and salt stress, which is repressed by gibberellic acid (GA) in poplar. The transgenic Arabidopsis plants over-expressing PeSCL7 exhibited enhanced tolerance to drought and salt treatment due to the increased activity of superoxide dismutase (SOD) and α-amylase (FAA) [65]. Compared with wild-type plants, OsGRAS23-overexpressing rice plants showed improved drought resistance and oxidative stress tolerance [66].

Rye (Secale cereale L.) is a member of the Secale genus in the Poaceae family and contains various nutrients for human consumption, including starch, vitamins, dietary fiber, protein, mineral elements, and phenolic compounds [67]. Rye has multiple applications in food, feed, bioenergy and alcohol production industries [68, 69] and exhibits probiotic activity that can lower the risk of cardiovascular and obesity diseases [70,71,72]. Rye is also a highly resilient crop that can withstand low temperatures, droughts, and poor soils [73]. As a diploid species in the Triticeae Dumortier, rye is of significant importance and closely related to barley and wheat [74]. Therefore, systematic gene mining and functional characterization of rye are essential for elucidating the physiological functions, evolutionary relationships, and genetic improvement of gramineous crops. In this study, we performed a comprehensive analysis of the ScGRAS family based on the recently published whole genome sequences of rye [75]. 67 GRAS genes were identified in S. cereale and assigned them to thirteen subfamilies. Further analysis was conducted on their gene structures, motif compositions, duplications, chromosome distributions, and phylogenetic relationships. We also characterized the expression patterns of selected ScGRAS members in different tissues and grain development stages, as well as under different stress and hormone induction. In addition, we investigated the paclobutrazol significantly reduced the plant height of rye, and promoted increase the weight of grains. Paclobutrazol may affect the filling process through the gibberellin pathway in rye.

Results

Identification of GRAS genes in S. Cereale

Based on their position on the rye chromosome, these GRAS members have been renamed ScGRAS1 to ScGRAS67 (Table S1). Their basic features including gene coding sequence (CDS), protein molecular weight (MW), isoelectric point (PI), and subcellular localization are systematically analyzed. Of the 67 ScGRAS proteins, ScGRAS57 was the smallest with 395 amino acids. The largest was ScGRAS50 with 1453 amino acids. Molecular weight of the proteins ranged from 41.47 kDa (ScGRAS57) to 163.58 kDa (ScGRAS50). The pI ranged from 4.75 (ScGRAS28) to 10.56 (ScGRAS14), with a median of 5.98. All the putative proteins encoded by the ScGRAS genes, contained the GRAS domain, which is necessary for their function as transcription factors. Based on the predicted subcellular localization, 28 ScGRASs were located in the nucleus, 16 in the chloroplast, 29 in the cytoplasmic, three (ScGRAS33, ScGRAS44, and ScGRAS47) in the mitochondria, two (ScGRAS35, and ScGRAS59) in the endoplasmic reticulum, two (ScGRAS3, and ScGRAS41) in the peroxisome (Table S1).

Phylogenetic analysis, and multiple sequence alignment of ScGRAS putative proteins

We constructed a phylogenetic tree encompassing S. cereale (67 ScGRASs), A. thaliana (33 AtGRASs), and O. sativa (46 OsGRASs) through the neighbor-joining method (Fig. 1, Table S1). Following the classification methodology proposed by Cenci et al. [7] and Tian et al. [34], the 146 GRAS proteins were categorized into thirteen distinct topological branches. Notably, LISCL exhibited the largest number of members (18 ScGRAS proteins), while OS43 (ScGRAS5), SCL4/7 (ScGRAS30), and DLT (ScGRAS44) possessed the fewest representatives (solely one ScGRAS protein each). The topology tree reveals a remarkable genetic affinity between certain ScGRAS proteins and numerous AtGRAS and OsGRAS proteins (bootstrap support ≥ 70), exemplified by ScGRAS5, ScGRAS8, and ScGRAS58. This suggests that these homologous proteins may share comparable gene structures and physiological functionalities.

Fig. 1
figure 1

Unrooted phylogenetic tree showing relationships among GRAS domains of Secale cereale (Sc), Arabidopsis thaliana (At) and Oryza sativa (Os). The phylogenetic tree was derived using the neighbor-joining method in MEGA7.0. The tree shows the 13 phylogenetic subfamilies. GRAS proteins from S. cereale are highlighted in red

Full size image

To elucidate the conserved amino acid residues within different subfamilies, a subset of AtGRASs, OsGRASs, and ScGRASs from 13 distinct subfamilies were randomly chosen for comprehensive multi-sequence comparisons (Figure S1, Table S1). The intricate conserved domains, namely LHR I, VHIID, LHR II, PFYRE, and SAW, displayed remarkable complexity among various subfamily members of rye GRAS proteins. The diversity inherent in these amino acids contributes to structural and functional divergences. Notably, the VHIID domain serves as the pivotal region for functionality, exhibiting highly similar amino acid configurations that are readily identifiable across different species. With the exception of GRAS33, GRAS34, GRAS35, and GRAS43, the conservation of His and Asp residues within the VHIID domain remained consistent. Additionally, certain non-polar hydrophobic amino acid residues exhibited potential cross-substitution, hypothesized to have minimal impact on peptide formation. It is worth to named select GRAS members demonstrated alternating residues, oscillating between Ile and Val within the VHLLD region. An observation worth noting is the presence of a highly disordered region at the N-terminus of ScGRAS proteins, showcasing discernible similarities across different subfamilies.

Gene structures, conserved motifs, and cis-acting elements analysis of ScGRAS genes

A comparative analysis of exon-intron structures reveals variations in the number and sequencing among the 67 ScGRAS genes, ranging from 1 to 5 exons (Fig. 2A and B, Tables S1). All ScGRAS genes contain the GRAS domain, with the majority (40, ~ 59.70%) lacking introns. Fifteen ScGRAS genes have one intron, while ScGRAS5, ScGRAS33, ScGRAS50, ScGRAS56, and ScGRAS63 possess two introns. ScGRAS23, ScGRAS43, and ScGRAS67 have up to four introns. Genes without introns exhibit compact structures and are widely distributed across all subfamilies, except for the Os43 subfamily, primarily in the LISCL subfamily. The DLT, LAS, DELLA, OS4, OS19, SCL4/7, SHR, and SCR subfamilies either lack introns or contain only one. PAT1 shows greater diversity in the number of exons, with five distinct exon types. Additionally, members within the same subfamily share similar gene structures, albeit with inconsistent exon and intron distributions.

Fig. 2
figure 2

Phylogenetic relationships, gene structure analysis, and motif distributions of S. cereale GRAS genes. A Phylogenetic tree was constructed using the neighbor-joining method with 1000 replicates for each node. B Exons and introns are indicated by yellow rectangles and grey lines, respectively. The green, yellow, and red rectangles represent the UTR, CDS, and GRAS conserved domains, respectively. C Amino acid motifs in the ScGRAS proteins (1–10) are represented by colored boxes. The black lines indicate relative protein lengths

Full size image

The motif analysis of the 67 ScGRAS proteins using MEME online software revealed ten conserved motifs (Fig. 2C, Table S2 ~ S4). These motifs exhibited varying distribution patterns among ScGRASs, with motifs 2, 3, and 4 being widespread, except in ScGRAS2, ScGRAS15, and ScGRAS43. Motifs 10, 6, 5, and 2 were often located in close proximity across most members. Generally, ScGRAS members within the same subfamily displayed similar motif compositions. The motifs 10, 6, 5, 2, 7, 3, and 4 were present in the DLT, LAS, LISCL, OS4, OS43, SCR, and SHR subfamilies. The DELLA, DLT, LISCL, OS4, OS43, SCL3, and SCR subfamilies shared motifs 3, 1, and 4. Certain subfamilies may lack specific motif compositions, such as the absence of motifs 8 and 9 in the Os19 subfamily. Additionally, specific motifs consistently occupy particular positions within the structures of these ScGRAS proteins. Motifs 10 and 6 consistently appear at the N-terminus of proteins in subfamilies DELLA, DLT, LAS, OS4, OS43, SCL3, SCL4/7, SCR, and SHR. Motif 2 is predominantly located at the beginning of OS19. Motif 4 is usually found near the C-terminus. Overall, the motif arrangement is generally similar within members of the same subfamily, supporting the classification observed in the phylogenetic trees. We further analyzed the conservation of specific amino acids in these motifs. Overall, some conserved amino acid sites have been identified (Figure S2, Table S3).

A total of 107 cis-regulatory elements, encompassing 46 distinct physiological functions, were identified (Table S5). These elements were classified into eight categories: development-related, light-responsive, site-binding, environmental stress-responsive, promoter-related, hormone-responsive, wound-responsive, and other elements. Among the promoter elements, light-responsive elements accounted for the largest proportion, including 25 cis-regulatory factors. Promoter-related elements, such as the TATA-box, were present in all ScGRAS genes. Sixteen hormone-responsive elements were identified, including those responsive to abscisic acid (AAGAA-motif, ABRE related), auxin (AuxRR-core, TGA-element, AuxRE, TGA-box), gibberellin (P-box, GARE-motif, TATC-box), MeJA (TGACG-motif, CGTCA-motif), and salicylic acid (TCA-element). Moreover, several cis-regulatory elements associated with anaerobic induction, drought, fungal elicitors, anoxic-specific inducibility, low-temperature, defense responses, and stress responsiveness were also discovered. Abscisic acid-responsive elements were present in nearly 98.51% of ScGRAS genes, while gibberellin-responsive elements existed in 61.19% of members, and auxin-responsive elements were found in approximately 44.78%. Twelve cis-acting elements were involved in the regulatory processes of different tissues (meristem, endosperm, root, leaf, and seed) during development in S. cereale. Consequently, ScGRAS genes are implicated not only in tissue development but also in responses to various abiotic stresses. It is worth pointing out that we have found that some cis acting elements may be unique to certain subfamilies. TGA-box was found to exist only in the DELLA subfamily (ScGRAS60 / ScGRAS61), which is an auxin responsive element, suggesting that the physiological functions of members of the DELLA subfamily may be complex. GATT-motif is only found in the HAM subfamily, which is a part of a light responsive element.

Chromosomal spread and gene duplication of ScGRAS genes

The 65 ScGRAS genes are unevenly distributed across chromosomes 1R to 7R (Fig. 3, Table S6). Additionally, two ScGRAS genes (ScGRAS66 and ScGRAS67) were located on unassigned chromosomes (Un) Chromosome 2R contained the highest number of ScGRAS genes (18 genes, ~ 26.87%), followed by 4R (17 genes, ~ 25.37%). The lowest numbers were observed on 1R and 7R (4 genes, ~ 5.97%). Chromosomes 6R, 3R, and 5R harbored 5 (~ 7.46%), 6 (~ 8.96%), and 11 (~ 16.42%) ScGRAS genes, respectively. Nine gene duplication events were detected within the GRAS gene family in S. cereale. Tandem repeat events were observed on chromosomes 2R, 3R, 4R, and 6R, particularly in ScGRAS36, ScGRAS37, ScGRAS38, and ScGRAS39. A region enriched with tandem repeats was identified, encompassing genes ScGRAS35 to ScGRAS40, all belonging to the LISCL subfamily. Three pairs of segmental duplications involving ScGRAS genes were detected (Fig. 4, Table S8). Five collateral homologs were identified in ScGRAS genes, accounting for 8.96% of the total, suggesting that these genes may have originated from segmental expansion events. In general, the typical domain of GRAS family is a VHIID motif flanked by two Leucine rich regions. The ‘VHIID’ motif represents several important amino acids. However, the core regions of these proteins are replaced by ‘LHIVD’. Except for ScGRAS35, the SAW motifs of other members are composed of three conserved amino acid residues: R (x4) E, W (x7) G, and W (x10) W structures. ScGRAS37/SCGRAS38 (86.2%) and ScGRAS39/SCGRAS40 (85.7%) had high similarity (Table S7). Chromosome 4R contained the most ScGRAS members (n = 3). In contrast to tandem duplication, two homologous expansion events involving four genes (ScGRAS32 / ScGRAS47, and ScGRAS33 / ScGRAS49) were discovered. These segmental duplications primarily involved the SCR and SCL3 subfamilies, while other groups exhibited greater conservation during evolution.

Fig. 3
figure 3

Schematic representation of the chromosomal distribution of the S. cereale GRAS genes. Vertical bars represent the chromosomes of S. cereale. The chromosome number is indicated to the left of each chromosome. The scale on the left represents chromosome length. Gene pairs with tandem repeat relationships are marked in red. The tandem gene pairs between pairs are connected by U-shaped lines

Full size image
Fig. 4
figure 4

Schematic representation of the chromosomal distribution and interchromosomal relationships of S. cereale GRAS genes. Colored lines indicate all synteny blocks in the S. cereale genome, and the red lines indicate duplicated GRAS gene pairs. The chromosome number is indicated at the bottom of each chromosome

Full size image

Synteny analysis of ScGRAS genes

A total of 52 ScGRAS genes showed homologous relationships with genes in A. thaliana (n = 3), O. sativa (n = 30), Z. mays (n = 34), Aegilops tauschii (n = 39), H. vulgare (n = 35), and T. aestivum (n = 49) (Fig. 5, Table S9). The number of collinear gene pairs between rye and other representative species (A. thaliana, O. sativa, Z. mays, A. tauschii, T. aestivum, and H. vulgare) were 4, 42, 54, 49, 42, and 137, respectively. Rye exhibited a relatively high proportion of GRAS gene orthologous pairs with A. tauschii and H. vulgare, accounting for 79.59% and 83.33%, respectively. Some homologous gene pairs between rye and Triticeae Dumortier plants were not identified in A. thaliana, O. sativa, and Z. mays. For example, ScGRAS5 had homologs AET1Gv20229700 / ARI1A01G110900 / HORVU1Hr1G020370, indicating possible expansion events specific to Triticeae Dumortier plants that differ from dicotyledonous plants like Arabidopsis and other monocotyledonous plants. Moreover, collateral homologous pairs were observed among dicotyledonous and monocotyledonous plants, with genes such as ScGRAS25, ScGRAS46, and ScGRAS64 suggesting ancestral origins before plant differentiation. Tajima-D neutrality testing was conducted on the 67 ScGRAS genes to better understand their targeted or balanced selection. The D value obtained was 7.49 (Table S10), significantly deviating from zero, indicating the involvement of the ScGRAS gene family in evolutionary neutral selection. Furthermore, we evaluated the Ka/Ks values within these subfamilies. This calculation will help estimate the selection pressure acting on these duplicated genes, advancing insights into three categories of selection: purifying, positive, and neutral. The results showed that most genes were subjected to purification selection (Table S11). This result also exists in most genes involved in repetitive events.

Fig. 5
figure 5

Synteny analyses of the GRAS genes between Secale cereale and six representative plant species (Triticum aestivum, Aegilops tauschii, Hordeum vulgare, Oryza sativa subsp. Indica, Zea mays, and Arabidopsi thaliana). Gray lines on the background indicate the collinear blocks in S. cereale and other plant genomes; red lines highlight the syntenic S. cereale GRAS gene pairs

Full size image

Evolutionary analysis of ScGRAS and GRAS genes of several different species

To analyze the genetic relationship between GRAS proteins in rye and six representative plants (A. thaliana, O. sativa, Z. mays, A. tauschii, T. aestivum, and H. vulgare), an unrooted NJ tree was constructed. Ten conserved motifs were identified in the sequences of 601 GRAS proteins from these plants using MEME online service software (Fig. 6 and S3, Table S2 ~ S4). Detailed genetic correspondences are provided in Tables S1 and S2. ScGRAS proteins tend to cluster with GRAS members of A. tauschii, T. aestivum, and H. vulgare. With a few exceptions such as ScGRAS14, ScGRAS16, ScGRAS43, and ScGRAS67, all other ScGRAS proteins contain motifs 2 and 3. The arrangements and structures of certain motifs exhibit specificity, differentiating genes from various subfamilies and forming distinct topological patterns. Motifs 1, 8, and 9 are absent in the subfamilies HAM and LAS. Members of the subfamily OS19 (ScGRAS14, ScGRAS15, and ScGRAS16) lack motifs 1, 7, 8, and 9. Overall, GRAS genes from Triticeae Dumortier plants and S. cereale that occupy the same topological branches share similar motif arrangements. Specific GRAS protein subfamilies in these plants often possess analogous motifs, indicating their evolutionary relationship. Motifs 8, 4, and 5 form a conserved structure and tend to cluster within the HAM and LAS subfamilies, while motifs 3, 7, 9, 8, 4, 1, and 5 tend to aggregate within the subfamilies DELLA, DLT, LISCL, OS4, OS43, PAT1, and SCL4/7.

Fig. 6
figure 6

Phylogenetic relationship and motif composition of the GRAS proteins from S. cereale with six different plant species (T. aestivum, A. tauschii, H. vulgare, O. sativa subsp. Indica, Z. mays, and A. thaliana). Outer panel: an unrooted phylogenetic tree constructed using Geneious R11 with the neighbor-joining method. Inner panel: distribution of conserved motifs in GRAS proteins. The differently colored boxes represent different motifs and their positions in each GRAS protein sequence. The sequence information for each motif is provided in Table S2

Full size image

Expression patterns of ScGRASs in several plant organs

To investigate the physiological functions of GRAS genes in rye, real-time PCR was employed to detect the expression levels of 19 members during the 21 DPA (days post-anthesis) of rye grains. Transcript accumulation in five organs (leaves, stems, roots, flowers, and grains) was assessed (Fig. 7A). Most ScGRAS members exhibited preferential expression in specific tissues. The highest expression was observed in roots for seven genes (ScGRAS8, ScGRAS18, ScGRAS24, ScGRAS25, ScGRAS60, ScGRA61, and ScGRAS65), in stems for five genes (ScGRAS15, ScGRAS46, ScGRAS47, ScGRAS48, and ScGRAS61), in leaves for two genes (ScGRAS29 and ScGRAS61), in flowers for two genes (ScGRAS5 and ScGRAS27), and in grains for eight genes (ScGRAS6, ScGRAS8, ScGRAS30, ScGRAS32, ScGRAS44, ScGRAS47, ScGRAS64, and ScGRAS65).

Fig. 7
figure 7

Expression patterns of selected 19 S. cereale GRAS genes. A Expression patterns of 19 S. cereale GRAS genes in the root, stem, leave, flower, and grain were examined via qRT-PCR. Relative expression level was shown as mean (± SE) from three independent experiments. B Expression patterns of 19 S. cereale GRAS genes were examined during different grain development stages: 7 DPA (early-filling stage), 14 DPA (mid-filling stage), 21 DPA (early-ripening stage), 28 DPA (mid-ripening stage), and 35 DPA (full-ripening stage). Lowercase letters above the bars indicate significant differences (α = 0.05, LSD) among the treatments

Full size image

Expression levels of most ScGRAS genes varied significantly at different stages of grain development. In general, gene expression was higher before the early ripening stage (21 DPA) compared to the mid-full filling stages. Nine genes (ScGRAS5, ScGRAS8, ScGRAS15, ScGRAS18, ScGRAS25, ScGRAS29, ScGRAS44, ScGRAS47) exhibited highest expression at 7 DPA, while four genes (ScGRAS24, ScGRAS25, ScGRAS32, and ScGRAS46) highest expression at 21 DPA (Fig. 7B). Except for ScGRAS30, ScGRAS48, and ScGRAS60, most genes exhibited stable expression levels in grains, with the lowest expression generally observed during the fully ripened stage (35 DPA).

Furthermore, certain ScGRAS members displayed coordinated expression patterns across multiple plant organs. The expression levels of some GRAS members exhibited significant positive correlations. For example, ScGRAS6, ScGRAS8, ScGRAS30, ScGRAS32, ScGRAS44, ScGRAS64, and ScGRAS65 were co-expressed in various plant organs (Figure S4), while ScGRAS5, ScGRAS8, ScGRAS15, ScGRAS18, ScGRAS29, ScGRAS30, ScGRAS44, ScGRAS47, and ScGRAS64 were co-expressed in grains (Figure S5). Notably, within the DELLA subfamily, the expression levels of ScGRAS24, ScGRAS29, and ScGRAS60 exhibited a significant positive correlation in different tissues.

Effects of grain developments and expression of DELLA subfamily genes after paclobutrazol and gibberellin treatments

Compared to the control (Mock), plant height in rye significantly decreased with paclobutrazol treatment, while grain filling was promoted (Fig. 8A). This effect was particularly noticeable during the later stages of grain development. As the grain-filling process advanced, endogenous gibberellin content gradually decreased in all groups, including the treatment and control groups. The gibberellin content in the paclobutrazol treatment group exhibited a rapid decline at 14 DPA and 21 DPA, stabilizing thereafter at 35 DPA. Interestingly, plant height in rye significantly increased during gibberellin treatment, particularly during the middle and late stages of grain filling, while the 1000-grain weight significantly decreased. These findings suggest that paclobutrazol primarily influences the filling process through the gibberellin pathway in rye.

Fig. 8
figure 8

Grain development of S. cereale under exogenous paclobutrazol and gibberellin treatment. A The plant height, 1000 grain weight, and gibberellin content during grain development. B Differences in the expression of DELLA subfamily genes under exogenous paclobutrazol and gibberellin treatment during grain development. Mock: the same amount of water treatment, Paclobutrazol: 250 mg/L paclobutrazol treatment. Gibberellin: 100 μm gibberellin treatment. Error bars were obtained from three measurements. We need information that asterisk described significant differences (α = 0.05/0.01/0.001, LSD) among the treatments. *, **, and *** indicate significant correlations at the 0.05, 0.01 and 0.001 levels, respectively

Full size image

Exogenous paclobutrazol and gibberellin treatments significantly influenced the expression of DELLA subfamily genes in rye (Fig. 8B). Expression levels of ScGRAS24, ScGRAS60, and ScGRAS61 demonstrated an initial increase followed by gradual decline, reaching their lowest values at 35 DPA. ScGRAS29, on the other hand, exhibited a steady decrease in expression. Moreover, most DELLA members showed significant down-regulation during the filling period following gibberellin treatment, indicating a potential antagonistic relationship. The expression of ScGRAS24 remained unchanged in the early stages of grain filling, highest expression at 28 DPA in the paclobutrazol treatment, suggesting its potential role in later filling stages. ScGRAS61 expression significantly increased at 7 DPA, reaching its highest expression at 14 DPA. Interestingly, ScGRAS60 expression was significantly up-regulated in nearly all induction treatments, indicating its sensitivity to paclobutrazol.

Expression patterns of ScGRAS genes in response to different treatments

Numerous ScGRAS members exhibited significant up-regulation or inhibition under various stress conditions due to specific treatments (Figure S6). The expression of ScGRAS6 and ScGRAS24 significantly increased in roots, stems, and leaves after one hour of cold stress. The expression of certain GRAS genes displayed dynamic patterns, with differential expression levels observed across organs or treatment durations. ScGRAS6, ScGRAS24, and ScGRAS60 were significantly up-regulated and subsequently down-regulated under heat stress. ScGRAS5 expression gradually increased in roots while decreasing in stems and leaves. Many ScGRAS genes showed contrasting expression patterns under different stress treatments. The expression of ScGRAS6, ScGRAS8, and ScGRAS24 was significantly up-regulated initially and then down-regulated in stems following UV-A, flooding, and heat treatments. Other genes exhibited distinct characteristics within specific tissues and exposure times. ScGRAS47 displayed significant responses to cold and NaCl treatments in roots and stems but exhibited no change in leaves. The correlations between the expression patterns of ScGRAS genes were observed (Figure S7). Most ScGRAS genes exhibited negative correlations, although certain genes demonstrated significant positive correlations, such as ScGRAS5, ScGRAS6, ScGRAS25, and ScGRAS47 (P < 0.05).

Expression patterns of GRAS members during different stages of grain development were analyzed under various treatments. All genes containing corresponding hormone-responsive elements in their promoter regions were detected (Figure S8, Table S12). Based on significant correlation connections used to construct a network, it becomes evident that the expression of some genes may be synergistic. Positive co-expression was observed among ScGRAS8, ScGRAS15, ScGRAS18, ScGRAS27, ScGRAS46, ScGRAS64, and ScGRAS65 under abscisic acid induction. Similarly, ScGRAS6, ScGRAS27, ScGRAS30, ScGRAS32, ScGRAS61, and ScGRAS64 showed positive co-expression under auxin induction. The expression patterns of DELLA family members did not consistently align with gibberellin and paclobutrazol induction, suggesting diverse functions. Although some co-expressed genes might interact, such as ScGRAS65 exhibiting positive correlation with ScGRAS24, ScGRAS29, and ScGRAS61 under gibberellin induction, these results underscore the complexity of physiological functions within different subfamilies of the GRAS family.

Discussion

ScGRAS gene structures and evolutionary analyses

The GRAS proteins in rye exhibit considerable structural diversity, particularly among the thirteen subfamilies, indicating that the physiological function of the GRAS gene family in rye is complex (Fig. 1 and S1, Table S1). The proportion of GRAS genes in the rye genome is approximately 0.15%, which is lower than that in other plants such as G. max (0.21%) [27], V. vinifera (0.17%) [41], H. vulgare (0.16%) [28], S. italica (0.16%) [37], S. bicolor (0.24%) [39], T. aestivum (0.17%) [40], Z. mays (0.22%) [42], but higher than that in A. thaliana (0.11%) [19]. Within the GRAS gene family of rye, there are thirteen subfamilies, including DELLA, DLT, HAM, LISCL, LAS, SCL3, SCL4/7, SCR, SHR, PAT1, OS4, OS43, and OS19 (Fig. 1, Table S1). It is speculated that these thirteen subfamilies may be present in most Gramineae plants and have fundamental physiological functions that are conserved throughout evolution [34]. Furthermore, the classification of the GRAS gene family may have become fixed in early higher plants and remained unchanged during plant evolution. However, the ancestral proteins within this family may continue to evolve, resulting in expansion and the emergence of new physiological functions in subsequent plant generations, depending on the specific plant species and environmental conditions [34]. These certain ScGRAS proteins (ScGRAS5, ScGRAS14, ScGRAS15, ScGRAS16, ScGRAS26, and ScGRAS27) have been classified into rice-specific subfamilies, indicating that the GRAS family may undergo further differentiation in monocotyledonous plants. Among the subfamilies, LISCL have the highest number of members (18, ~ 26.87%), while OS43 (ScGRAS5), SCL4/7 (ScGRAS30), and DLT (ScGRAS44) have the fewest members. Similarly to other plants such as Arabidopsis [19], rice [34], S. italica [37], S. bicolor [39], T. aestivum [40], and Z. mays [42], many subfamilies within the GRAS gene family of rye are likely to be conserved, whereas LISCL may exhibit greater differentiation ability. The differences in expansion among these subfamilies are speculated to be associated with the physiological functions of different proteins and their adaptation to the environment during evolution. However, more research is needed to determine whether the structural differences among these subfamilies are related to environmental adaptation. To further analyze the GRAS gene family in rye from different sources, we identified another important rye genome (Lo7) [67]. A total of 72 independent GRAS proteins were identified in the ‘Lo7’. Similarly, these genes were primarily classified into 13 typical subfamilies (Figure S9, Table S13). The GRAS proteins of ‘Weining’ was used to co construct the evolutionary tree, which was consistent with our original classification. To explain the differences and homology among these members, we constructed a comparative genome in the two rye (Figure S10, Table S14). Most genes were assigned to the corresponding chromosomes (1R ~ 7R), indicating the overall reliability of the results. However, we observed that there are still some genes that have not been defined as corresponding homologues. We speculate that this may be a difference in genome assembly.

Most of these GRAS genes in rye contain conserved domains, including LHR I, VHIID, LHR II, PFYRE, and SAW. As shown in Figure S1, the VHIID domain is considered the central region and contains highly conserved histidine and aspartic acid residues, which serve as the base and supporting sites of GRAS proteins [76,77,78]. There may be cross-substitution of non-polar hydrophobic amino acid residues, such as histidine (His), leucine (Leu), isoleucine (Ile), and valine (Val), within the core region. These substitutions are likely the result of genetic mutations, although they may not significantly alter the peptide chain structure [79]. Furthermore, some GRAS proteins belonging to the LISCL subfamily (ScGRAS34, ScGRAS35, and ScGRAS62) and PAT1 subfamily (ScGRAS43) do not contain conserved histidine and aspartic acid residues in the VHIID region. The structural differences of these genes may indicate further differentiation of GRAS proteins, as also observed in sorghum [39]. There are numerous variations in amino acid residues within the VHIID region of the LISCL and PAT1 subfamilies. It is speculated that the high activity of the LISCL and PAT1 subfamilies leads to structural differentiation in the domains, resulting in amino acid instability. This phenomenon may explain why these subfamilies have expanded and become the largest subfamily. Some conserved amino acid segments in the structural domain of ScGRAS43, the member of the PAT1 subfamily, have been lost, possibly due to chromosome fragment translocation or inversion [11, 80]. The acquisition and loss of structural domains are important driving forces for gene family expansion, as observed in other higher plants such as sorghum [37] and maize [42]. Inherently disordered regions, which can undergo conformational changes between order and disorder, are abundant in eukaryotic proteomes [37, 39]. These functional regions, which contain short molecular recognition features (MORFs) in the N-terminal structural domain of GRAS proteins, play crucial roles in cell signal transduction and transcriptional regulation. Therefore, GRAS proteins possess functional specificity [16]. Although the N-terminus of GRAS proteins exhibits high variability, some residues display similarities across different subfamilies. For example, the DELLA subfamily protein contains the DELL A structural domain at its N-terminus.

The introns of these ScGRAS genes were examined, and it was found that each gene contains between 1 and 5 exons (Fig. 2A and B). Approximately 59.7% of ScGRAS genes do not contain introns, which is higher than in rice (~ 55%) [34] and poplar (~ 54.7%) [65], but lower than in millet (~ 64.9%) [37], sorghum (~ 66.7%) [39], Arabidopsis (~ 67.6%) [19], and buckwheat (~ 87%) [26]. The gene structure of certain subfamily members may be compact, as some subfamilies such as DLT, LAS, and DELLA do not contain introns or have only one intron. Genes without introns are also observed in other gene families, including the small auxin-up RNA (SAUR) gene family [81], F-box families [82], and DEAD box RNA helicase [83]. Generally, genes without introns or with few introns tend to have lower expression levels in plants. However, it has been suggested that GRAS genes in plants may have originated directly from prokaryotes through horizontal gene transfer and duplication events [42]. Therefore, most GRAS members in plants may have compact gene structures [84]. Genes without introns can continuously encode proteins during transcription and translation, making them sensitive to the environment and capable of responding rapidly [85,86,87]. Furthermore, gene expression may not strongly depend on the density of introns in these genes, as evidenced by our research results [68]. Some highly expressed genes have introns of average length (Fig. 7 and S4), indicating that the expression level may depend on specific developmental processes or environmental stress [69]. For example, the expression of ScGRAS64 in leaves increases rapidly under cold, salt, and PEG stresses, suggesting that it may be a response to these abiotic stresses. Genes with compact structures may contribute to rapid responses to stress or tissue development. Ten conserved motifs were identified in ScGRAS proteins, which can be used to predict the function of unknown proteins within the same subfamily [39].

Tandemly repeated genes can rapidly expand or contract in response to environmental changes, maintaining a constant number of functionally related genes without increasing genetic complexity during evolution [88]. Segmental duplications are also common in animal and plant genomes, contributing to genetic diversity [37]. Thus, tandem repeats and segmental duplications play important roles in the expansion of gene families and genome evolution, enabling plants to adapt to their environment. For example, duplication events of OsSHR1 led to diversification, and the expression of OsSHR2 expanded in the endodermis and certain cortex cell layers, possibly acquiring additional functions in rice root development [89]. In our study, nine tandem repeat events involving thirteen ScGRAS genes were identified (Fig. 3, Table S6). Notably, a region of high-density tandem repeats was found on chromosome 4R, involving four members (ScGRAS36, ScGRAS37, ScGRAS38, and ScGRAS39) belonging to the LISCL subfamily. This may explain why LISCL is the largest subfamily in the ScGRAS family. Furthermore, three pairs of segmental duplications were observed in ScGRAS genes (Fig. 4, Table S8). Consistent with other plants such as Arabidopsis [19], rice [34], millet [37], and barley [62], all duplicated genes are within the same subfamily, indicating that duplication events do not occur between different subfamilies. However, tandem replication of ScGRAS genes may be a more significant contributor to the expansion of the GRAS gene family in S. cereale, which is distinct from S. italica [37] and may represent a unique evolutionary pattern in rye.

Expression patterns and function prediction of ScGRAS genes

The gene expression patterns were analyzed to preliminarily predict the physiological functions of these GRAS genes in rye. The expression of nineteen GRAS members was analyzed in different tissues and at different stages of grain filling (Fig. 7). Almost all ScGRAS genes exhibited significant differential expression (p < 0.05). ScGRAS25, encoding a member of the LISCL, displayed specific expression in roots and flowers, consistent with the homologous gene At2G29060 in Arabidopsis, which participates in root, flower, and seed development. Notably, ScGRAS18, encoding a member of the PAT1 subfamily, is specifically expressed in roots and flowers. PAT1 members primarily participate in the signal transduction of photoreceptor A, as demonstrated by the elongation of hypocotyls, closure of apical hooks, and folded cotyledons observed in the pat1 mutant under far-red light conditions in Arabidopsis [90]. ScGRAS46 and ScGRAS48, both members of the same subfamily, exhibited similar expression patterns, with high expression levels in stems. Few studies have been conducted on LISCL subfamily members in higher plants, but evidence suggests that they may play roles in transcriptional regulation. The LiSCL transcription factor plays a crucial role in meiosis during the meiotic process of L. longiflorum [55]. Similarly, PrSCL1 in Pinus radiata and CsSCL1 in Castanea sativa are mainly expressed in stems and roots, induced by exogenous auxin during cutting, and involved in early adventitious root formation [91]. Furthermore, the expression pattern of ScGRA44 was similar to that of GS6, a homologous gene belonging to the DLT in rice [73]. OsGS6, an important domestication gene, has been found to play a significant role in reducing the size of rice grains [92]. The expression patterns of DELLA family members may be complex. For example, GRAS24 is specifically expressed in roots, leaves, and grains, while GRAS29 exhibits high expression levels in leaves. Therefore, it is necessary to systematically analyze their expression characteristics in different tissues and at different stages of grain development. The expression patterns of many ScGRAS genes showed positive correlations, indicating potential synergistic effects in five plant organs (Figure S4). These findings provide insights into the function of the GRAS gene family in different tissues of rye, although further experiments are needed to verify their specific functions. Grain ripening is a critical process in rye, as it adapts to unfavorable climate and soil conditions and thrives in high-altitude, mountainous, and cold regions. The entire grain ripening process was divided into five representative stages, and the expression patterns of ScGRAS genes were analyzed to identify key candidate genes related to grain development. Most GRAS genes in rye are highly expressed before the early ripening stage (21 DPA), suggesting widespread involvement of the GRAS family in grain ripening. For instance, ScGRAS48 and ScGRAS60 are stably expressed in almost all stages.

DELLA proteins not only regulate the gibberellin response pathway but also function as central hubs in signaling pathways that integrate signals from various hormones, such as jasmonic acid, auxins, abscisic acid, and ethylene [50, 51]. Gibberellins are central regulators of plant growth and behavior, acting by degrading DELLA proteins. Peng, et al. [93] proposed that Rht-B1 / Rht-D1 and maize dwarf-8 are homologous to the gibberellin-insensitive (GAI) gene in Arabidopsis. DELLA proteins act as negative regulators in the gibberellin signaling pathway, inhibiting gene expression and plant growth. However, DELLA proteins can also be degraded by gibberellins, thereby eliminating their inhibitory effects. The degradation of DELLA proteins occurs through binding with the GA-GAI complex protein, leading to degradation and relieving inhibition by gibberellin. This balance between gibberellins and DELLA proteins forms a dynamic regulatory mechanism for gibberellin activity levels in plant growth and development [94, 95]. Furthermore, DELLA plays a crucial role in grain development. For example, the expression of DELLA genes in tomato and Arabidopsis induces parthenocarpy [96]. In rye grains, gibberellin can be detected throughout the entire developmental stage (Fig. 8). Therefore, it is hypothesized that young rye grains immediately produce gibberellin after fertilization to promote grain filling [39]. The endogenous gibberellin content in rye grains was analyzed and detected throughout the entire grain development stages, gradually decreasing as the grain ripens. In addition to ScGRAS60, ScDELLAs exhibit significantly higher expression levels during the early ripening stage (before 21 DPA) of grain development compared to the full ripening stage (35 DPA). These findings suggest that DELLA genes may play a role in the early to mid-stages of grain development. The plant growth regulator paclobutrazol, which regulates DELLA transcription and gibberellin biosynthesis, was used to treat rye plants [97]. Paclobutrazol treatment significantly reduced plant height and gibberellin content while increasing grain weight (Fig. 8B). It is speculated that under paclobutrazol treatment, more photosynthetic products are directed towards grain development rather than stem elongation [98]. Furthermore, exogenous gibberellin treatment had opposite effects on plant height and grain weight compared to paclobutrazol treatment. Paclobutrazol significantly inhibits gibberellin biosynthesis, especially during the early to middle filling stage (7 DPA and 14 DPA), potentially influencing the expression patterns of DELLA members due to the down-regulation of gibberellin. Almost all DELLAs exhibited suppressed expression levels under gibberellin treatment compared to the control group. After paclobutrazol treatment, the expression level of ScGRAS24 changed significantly at 28 DPA and 35 DPA, indicating potential sensitivity during the full ripening stage. Conversely, the expression level of ScGRAS61 significantly increased at 7 DPA and 14 DPA. ScGRAS29 displayed a unique expression pattern with fluctuating levels, possibly due to significant differences in amino acid structure and motif arrangement compared to other members. Notably, the expression level of ScGRAS60 significantly increased throughout grain development after paclobutrazol treatment and was more sensitive than other DELLA members. This suggests that ScGRAS60 may have potential value in breeding rye. Additionally, significant differences in the expression levels of four DELLA subfamily genes were observed during grain development after paclobutrazol treatment, indicating potential functional differentiation among different members of the DELLA. This is consistent with previous findings in Arabidopsis, where members of the DELLA exhibit differentiated functions while retaining some overlapping functions [99].

Rye possesses the ability to adapt to unfavorable climate and soil conditions, allowing it to thrive in high-altitude areas, mountainous regions, and cold environments. This adaptation may be regulated by a complex endogenous network and transcriptional signals that enable rye to respond to abiotic stresses [100]. However, the stress response of rye to complex abiotic stresses has not been systematically analyzed. To explore the physiological roles of GRAS in environmental adaptation in rye, the expression of nineteen GRAS members in response to six different abiotic stresses and three representative hormone treatments was analyzed in rye seedlings (Figures S6 and S8). Under cold stress, the expression levels of 11 ScGRAS genes in roots, 12 genes in leaves, and 13 genes in stems were significantly regulated, depending on the duration of the treatment. These responses may contribute to the adaptation of S. cereale to cold environments, which is consistent with its role as a cold-tolerant crop, as demonstrated in grape [101] and millet [37]. The member ScGRAS8 of the SHR subfamily showed rapid induction under UV-A treatment. In wheat, TaSCL14 is highly expressed in stems and roots in response to high light stress. Silencing TaSCL14 leads to decreased tolerance of wheat to high light stress, resulting in dark-induced leaf senescence and poor development [102]. SHOOT GRAVITROPISM 1 (At3G54220) and ScGRAS6 belong to the HAM and share similar basic sequence compositions. SGR1 promotes cell elongation and endodermis differentiation outside the meristematic tissue, which is crucial for root growth. Furthermore, the sgr1 mutant participates in the abscisic acid pathway and coordinates the oxidative stress response in plants by mediating the inhibition of cytokinin response in the meristematic tissue to promote root growth [103]. ScGRAS18 exhibits significant induction in roots under six different abiotic stresses. In rice, CIGR1 and CIGR2 are rapidly induced upon perception of N-acetylchitooligosaccharides elicitor, induced by exogenous gibberellins, which may play a key role(s) as transcriptional regulators in the early stages of defense signaling following fungal perception and pathogenesis [104]. GmGRAS37, which responds to drought, salt stress, abscisic acid, and brassinosteroids, enhances drought and salt stress resistance when overexpressed in soybean hairy roots [105]. In tomato, the HAM member SlGRAS40 is induced by D-mannitol and NaCl, playing a role in promoter- and auxin- and gibberellin-mediated signal transduction in response to abiotic stresses [106]. From the cluster tree, it can be seen that the HAM subfamily may have two highly differentiated sub-classes, similar to the HAM subfamily in sorghum we previously reported [39]. Therefore, two members of the HAM subfamily, ScGRAS6 and ScGRAS14, were selected for expression pattern analysis. Their expression patterns differed somewhat, with ScGRAS6 being highly expressed at 21 DPA and ScGRAS14 being highly expressed at 7 DPA. Under heat stress, the expression of ScGRAS6 increased and then decreased in stems, while the expression pattern of ScGRAS14 was opposite. These findings suggest that different branches of the GRAS subfamily may exhibit distinct responses to environmental stress. Additionally, some genes may participate in the response to abiotic stress through hormone regulation. For example, the AtDLT gene regulates brassinosteroid signaling by binding in the promoter of the BZR1gene, thereby regulating leaf curling and embryo sheath elongation [107]. Overall, different GRAS subfamilies have diverse biological functions, playing important roles in plant adaptation to abiotic stresses. This is supported by the correlation network (Table S12, Figure S8), which demonstrates that these GRAS transcription factors participate in a complex cross-regulatory network induced by stress and hormones.

Conclusion

In conclusion, this study identified 67 members of the GRAS gene family in rye and classified them into thirteen main subfamilies. Most ScGRAS genes do not contain introns, and their gene structures, conserved motifs, cis-acting elements, gene duplications, and expression patterns were analyzed. Overall, the gene structures of the same subfamily is always similar, including the number of exons, amino acid structures, and motif arrangements. Gene duplication events may have contributed to the emergence of certain ScGRAS genes, with tandem replication playing a more significant role in expanding the GRAS gene family compared to segmental duplication. Notably, a high-density tandem repeat region containing LISCL subfamily genes was discovered on chromosome 4R. The expression patterns of ScGRAS genes in different tissues and grain development stages were analyzed, and key candidate genes related to grain development were identified. Additionally, the relationship between DELLA genes, gibberellin content, and grain development was investigated. Furthermore, the expression patterns of ScGRAS genes under various abiotic stresses and hormone treatments were examined to shed light on their physiological functions in environmental adaptation. These findings provide valuable insights into the function and evolution of the GRAS gene family in rye.

Methods

Gene identification

The reference genome of rye was downloaded from the GenBank website of the National Center for Biotechnology Information, the accessed number was JADQCU000000000 [108]. Firstly, all of GRAS proteins of Arabidopsis and rice were used to search for candidate GRAS proteins from the rye genome via the blastp program [109]. Candidate genes were searched by blastp using a score value of ≥ 100 and e-value ≤ e − 10. Secondly, the Hidden Markov Model (HMM) file of the GRAS domain (PF03514) is downloaded from the Pfam protein family database (http://pfam.sanger.ac.uk/). Based on the HMM model in the HMMER 3.0 online software, the GRAS protein sequence in S. cereale was identified with a decision value of 0.01 (http://plants.ensembl.org/hmmer/index.html) [110]. Based on PFAM and SMART in thread sequencing, conserved motifs were found in the GRAS proteins in rye (http://smart.embl-heidelberg.de/) [111, 112]. Then, in the NCBI protein database, these ScGRAS proteins were used as the initial query for re-verification (https://blast.ncbi.nlm.nih.gov/Blast.cgi? PROGRAM = blastp&PAGE_TYPE = BlastSearch&LINK_LOC = blasthome). Finally, the ExPasy online program was used to identify the basic features of the GRAS genes in S. cereale, including sequence length, protein molecular weight, isoelectric points, and subcellular localization (http://web.expasy.org/protparam/). In addition, to further compare the similarity of these genes, we conducted pairwise sequence alignments on these proteins using the EMBOSS Need online website (https://www.ebi.ac.uk/Tools/psa/emboss_needle/).

GRAS gene structures and conserved motif analysis

Firstly, we conducted multiple sequence alignment projects on these GRAS proteins from rye to further analyze LHR I, VHIID, LHR II, PFYRE, and SAW domains within conserved domain intervals [113]. MEGA 7.0 and GeneDoc 2.7 software were used to manually adjust the conserved domain segments in the amino acid sequences of these GRAS proteins to elucidate their diversity and variability [39]. Specifically, these software tools were used for sequence alignment and manual modification of amino acid sequences to accurately identify conserved domains [114]. Gene Structure Display Server online software was used to analyze the exon-intron substructures of these ScGRAS genes [37]. Additionally, the MEME online program (http://meme.nbcr.net/) was used to analyze the conserved motifs and gene structure variations among these GRAS proteins [115]. The optimized parameters for the conserved motifs were as follows: a maximum of 10 motifs and the optimal width of residues ranging from 6 to 200 [37]. Visualize Gene Structure is completed using TBtools software (v1.120) [116]. Additionally, the PlantCARE online software was used to predict the physiological functions of cis-elements in the promoter regions (up-stream 2000 bp) of these 67 GRAS genes [39].

Chromosomal distribution and gene duplication

Firstly, based on the physical location of these genes in the annotation file, all ScGRAS genes have been designated as chromosomal details. Circos software was used to analyze these ScGRAS genes for chromosomal location information [117]. The presence of two or more gene members from the same family within the 200 kb chromosome region is defined as the presence of tandem repeats [37]. Multiple Collinear Scanning Toolkits (MCScanX) of TBtools software (v1.120) [116] was used with default parameters to analyze gene duplication events for these GRAS genes [110]. Homology of the GRAS genes between S. cereale and six other plants (T. aestivum, A. tauschii, H. vulgare, O. sativa ssp. Indica, Z. mays, and A. thaliana) was analyzed by using the project of dual synteny plotter in TBtools software. To further analyze the possible selection pressure in the GRAS genes of rye [118,119,120], the Ka/Ks values of all gene pairs in different subfamilies were calculated using the Simple Ka/Ks Calculator (NG) program of TBtools.

Phylogenetic analysis and classification of the ScGRAS family

According to the classification of AtGRAS and OsGRAS proteins, 67 GRAS proteins in S. cereale are divided into 13 main subfamilies. In MEGA 7.0, the Jukes-Cantor model is used to construct NJ (neighbor-joining method) trees. Bootstrap value of the constructed phylogenetic tree was set to 1000, and assigned with Geneious R11 with BLOSUM62 cost matrix. To elucidate the evolutionary relationships between these GRAS proteins in several plants, the synteny maps based on homologous genes from rye and six representative plants were constructed. Five monocotyledonous plants were selected, containing three Triticeae Dumortier plants (T. aestivum, A. tauschii, H. vulgare), one model plant (O. sativa), and one C4 plant (Z. mays). Meanwhile, the dicotyledonous model plant (A. thaliana) was also included in the comparison, which was obtained from the UniProt website [32, 33].

Plant materials, growth conditions, and abiotic stress in S. Cereale

S. cereale cv. Weining, a representative cultivated variety in Guizhou Province in southwest China was used. The cultivar was planted in a greenhouse at Chengdu University farm. At the early-ripening stage of rye, representative tissues were collected, including roots, stems, leaves, flowers, and grains. Additionally, to observe the expression levels of these representative genes in rye grain during the filling period, samples from five grain developmental stages were collected, i.e., 7 days (early-filling stage), 14 days (mid-filling stage), 21 days (early-ripening stage), 28 days (mid-ripening stage), and 35 days (full ripening stage). Many ScGRAS genes may be involved in the development of rye grains, thereby affecting the filling and nutritional structure of the grains. To determine these genes that may regulate the development of rye grains, the expression of these 19 ScGRAS members was evaluated during the five grain-filling stages after flowering. As far as possible, the selected members of different subfamilies exhibit significant differences in amino acid structures and distant clustering relationships. Except for the DELLA members, at least one member of different subfamilies was selected, depending on their topology and genetic structures. All the plants were grown under the same growth conditions, and these samples were collected from five plants. The collected samples were rapidly placed in liquid nitrogen and pre-cooled completely to fix their physiological status and stored at -80 °C until further use. Each sampling and stress treatment had three biological replicates. Meanwhile, these samples were performed by qRT-PCR with at least three technical repeats.

The plant RNA extraction kit (RNA Easy Fast Plant Tissue RNA Rapid Extraction Kit, DP452) was selected for total RNA extraction. In addition, to investigate the expression patterns of these ScGRAS genes under different abiotic stresses and hormones, seedlings of rye were subjected to abiotic stress treatment at the seedling stage (4 weeks after germination). All seedlings were planted in seedling trays, and each tray was added with 50 mL of solution to fully soak the roots of the plants. The treatment for six different abiotic stresses were UV-A radiation (70 μW/cm2, 67 V, 30 W), flooding (all plants), salt (5% NaCl), drought (10% PEG6000), high temperature (40℃), and low temperature (4℃). Each stress treatment was repeated three times, and samples of leaves, roots, and stems were taken at 0, 1, 4, and 12 hours for qRT-PCR analysis. Finally, considering that there were different hormone response elements in the promoter region of these genes, we conducted three different hormone treatments at the flowering stage: gibberellic acid (GA3, 100μM), auxin (indole-3-acetic acid, IAA, 100μM), and abscisic acid (ABA, 50μM). Paclobutrazol, a plant growth regulator, participates in the expression of members of the GRAS gene family [37, 39]. Therefore, it has also been considered as a candidate hormone. Whether there is a coordinated expression of these genes was observed. In addition, as a plant growth inhibitor, paclobutrazol regulates plant growth mainly by inhibiting biosynthesis of GAs by regulating DELLAs transcription [26]. In order to further investigate the relationship between DELLAs, GAs and grain development in rye, the materials of ‘Weining’ with similar growth status were selected and sprayed with 50 mL paclobutrazol (250 mg·L− 1) and gibberellin (100 μM) during the flowering period. Controls (mock) were sprayed with the same amount of water. Further analysis was conducted on the plant height, 1000-grain weight, gibberellin content, and gene expression level of the DELLA subfamily in control, paclobutrazol, and gibberellin-treated plants at 7, 14, 21, 28, and 35 DPA (days post-anthesis).

Endogenous GA analysis

Regarding the method of Fan et al. [37], the gibberellic acid (GA) content in rye grains was determined. Approximately 1 g of fresh tissue from the grain was collected and ground in liquid nitrogen. The ground powder was mixed with 50 mL of 80% ethanol and subjected to ultrasonic extraction three times for 1 h each time. Supernatant was concentrated at a low temperature, then mixed with water, and N-butanol was added to extract for 1 h. Finally, the n-butanol layer was dried under a stream of nitrogen (N2). Ten milligrams of the dried sample were accurately weighed and dissolved in 5 mL of methanol. The dissolved solution was filtered using a 0.22 μm microporous membrane, and LC / MS was used for content detection.

Total RNA extraction, cDNA reverse transcription, and qRT-PCR analysis

Fresh tissues of rye were extracted using a plant RNA extraction kit (RNA Easy Fast Plant Tissue RNA Rapid Extraction Kit, DP452) for total RNA extraction. Based on the primer sequences designed in Primer 5.0 software, the expression levels of different GRAS genes were detected (Table S15). ACTIN as an internal reference gene [121]. SYBR Premix ExTaqII (TaKaRa Bio) was used for standard expression detection, and experiments were performed with three replicates on a CFX96 real-time system (Bio-Rad). Real-time qPCR reaction included 40 cycles with parameter settings as follows: pre-denaturation at 95 ℃ for 30 s, denaturation at 95 ℃ for 5 s, annealing at 60 ℃ for 20 s, and extension at 72 ℃ for 20 s. All quantitative primers for genes were analyzed for their practicality through melting curves. The expression of these GRAS genes was analyzed using the 2− (ΔΔCt) method [122].

Statistical analysis

The least significant difference test (LSD) is further conducted using 0.05 and 0.01 significance levels to compare the means between the groups in JMP6.0 software (SAS Institute). Origin 2016 software (OriginLab Corporation, Northampton, Massachusetts, USA) has been employed to draw the histograms. Additionally, the Pearson correlation program was used to define the correlation coefficient of ScGRAS genes, and Sigmaplot 12.0 software (Systat Software, Inc, Point Richmond, CA) is utilized to calculate the correlation coefficient. A Pearson correlation matrix of the GRAS genes is generated using R2.11 (Bell Laboratories), and network analysis (CNA) of the correlation matrix is performed with the help of Cytoscape 2.7.0 software [123]. The correlation coefficient is defined as statistically significant at a P-value of less than 0.05.