Skip to main content
Download PDF

Spatiotemporal transcriptomics reveals key gene regulation for grain yield and quality in wheat

Genome Biology volume 26, Article number: 93 (2025) Cite this article

Abstract

Background

Cereal grain size and quality are critical agronomic traits in crop production. Wheat grain development is governed by intricate regulatory networks that require precise spatiotemporal coordination of gene expression to establish functional compartments in different cell types.

Results

Here, we perform a spatial transcriptomics study covering the early stages of wheat grain development, from 4 to 12 days after pollination. We classify the grain into 10 distinct cell types and identify 192 marker genes associated with them. WGCNA analysis reveals that highly expressed genes in different cell types exhibit distinct enrichment patterns, significantly influencing grain development and filling. Through co-expression and motif analyses, we identify a specific group of genes that may regulate wheat grain development, including TaABI3-B1, a transcription factor specifically expressed in the embryo and surrounding endosperm, which negatively affects embryo and grain size.

Conclusions

This study presents a comprehensive spatiotemporal transcriptional dataset for understanding wheat grain development. Additionally, it identifies key genetic resources with potential applications for improving wheat yield.

Peer Review reports

Background

Wheat (Triticum aestivum L.) is one of the top three cereal crops and is cultivated across approximately 230 million hectares globally. Enhancing wheat yield carries substantial implications for global food and nutrition security. The wheat grain comprises three primary components: the diploid embryo, the triploid endosperm, and the pericarp, also known as the seed coat [1]. Gene expression in the seed coat, endosperm, and embryo exhibits spatiotemporal changes, forming functional compartments. These tissues collaboratively regulate nutrient transport and grain development in wheat. However, current cell classifications of wheat grain are predominantly based on morphological observations and only a limited number of specifically expressed genes have been documented [2].

Genetics has helped to identify crucial genes involved in grain development [3, 4]. Multi-omics analysis techniques were used to comprehensively understand the different composition of wheat grain [5, 6]. All these studies typically lack spatial gene information, limiting the understanding of the specific functions of each cell population. The wheat grain consists of multiple distinct tissue layers. The bran includes the pericarp, testa, hyaline, and aleurone layers. Beneath the aleurone, there are 2–3 layers of sub-aleurone cells, followed by elongated prismatic cells oriented towards the grain center [7]. Previous studies typically lack spatial gene information, limiting the understanding of the specific functions of each cell population. A significant gap remains in our understanding of the complex interaction between the different layers in wheat grain and multiple tightly regulated pathways during wheat grain development in distinct layers. For instance, at 4 days after pollination (dap), early embryogenesis begins with the formation of a totipotent zygote, while the endosperm undergoes cellularization, and starch granules start accumulating in the pericarp. By dap 8, the embryo and endosperm undergo rapid differentiation and expansion. By dap 12, nutrient breakdown starts, facilitating the transfer of nutrients to both the endosperm and the growing embryo. Research is particularly needed to clarify stage- and cell-type-specific gene expression and crosstalk within and among the embryo, endosperm, and the whole grain.

Spatial transcriptomics offers single-cell level transcriptional information, enabling a detailed exploration of tissue-specific gene expression [8,9,10]. Spatial transcriptomics has been extensively used in plant developmental studies because it allows for the precise mapping of gene expression within the spatial context of developing tissues. In recent years, significant discoveries have been made in organogenesis, cell type identification, and cell fate lineage reconstruction through spatial transcriptomics [11], including maize inflorescence [12] and kernel [13] development, barley grain germination [14], peanut cell heterogeneity [15], soybean nodule maturation [16], tomato callus regeneration [17], orchid flower organogenesis [18], and dragon fruit senescence [9]. This application allows for the construction of comprehensive gene expression maps and identification where specific genes are active and how their expression patterns change over time [19].

Wheat grain yield and quality are regulated by a complex interplay of multiple genes. Genome-wide association studies (GWAS) have identified numerous QTLs associated with grain shape traits such as grain length, grain width, and thousand-grain weight. These QTLs are located within genes including ABC transporters [20], cytochrome P450 [21], and thioredoxin M-type proteins [22, 23], all of which are involved in seed development and ultimately influence yield and quality. Several key genes have been identified as crucial for these traits, especially some cell type specifically expressed transcription factors. For example, an aleurone layer and pericarp highly expressed gene, TaPGS1, influences wheat endosperm structure and regulates grain size and weight by binding to the promoter of TaFl3 [24]. Another endosperm-specific transcription factor, TaNAC019, interacts with the glutenin regulatory factor TaGAMyb, suppressing the expression of the key starch synthesis genes TaAGPS1-A1 and TaAGPS1-B1, thereby negatively regulating starch synthesis in the endosperm [25]. TaNF-Y indirectly enhanced the expression of starch biosynthesis genes by repressing TaNAC019, sucrose synthase 2 (TaSuS2), and other genes involved in starch metabolism. Conversely, TaNF-Y suppressed the expression of Gliadin-γ− 700 (TaGli-γ− 700) and low molecular weight- 400 (TaLMW- 400) proteins [26]. TaABI19 binds to the promoter of the TaPBF gene, enhancing its expression and affecting starch synthesis and grain development [27]. In maize, the scutellum-expressed transcription factor VIVIPAROUS- 1 (VP1) regulates genes related to globulin and nutrient metabolism pathways, facilitating the transfer of proteins from the endosperm to the embryo, which in turn affects grain size and quality [28].

Here, we employ spatiotemporal transcriptomic analysis to construct a detailed expression map for wheat grain at dap 4, dap 8, and dap 12. A series of candidate genes and gene mutants specific to various cell types (endosperm, aleurone, endosperm cavity fluid, and pericarp) were identified as regulators of wheat grain development. A key factor controlling embryo and endosperm development was further validated using allelic variations and transgenic materials. This study offers a valuable resource for understanding wheat grain development and provides new insights for enhancing yield.

Results

Generating spatial transcriptome profiles of early development stages in wheat grain

We conducted spatial transcriptomics in Jimai 22, a variety widely cultivated in China for its high yield (Fig. 1A). Three developmental stages were selected for cryosection: the pro-embryo stage (dap 4), the transition stage (dap 8), and the differentiation stage (dap 12) (Fig. 1B).

Fig. 1
figure 1

Spatially resolved transcriptome analysis of wheat grains. A A workflow for sampling and sequencing of wheat grains on a BMKMANU S1000. B Developing grains at dap 4, dap 8, and dap 12. C Spatial visualization of the unbiased spot clustering for dap 4, dap 8, and dap 12 sections. Merged bright field images and spatial clusters of the other three sections. The tissue/cell-type identity of each cluster was assigned based on the location of each cluster. D Heatmap and spatial distribution map of total expression counts in the spot of each sample. E UMAP of spatial spots from dap 4, dap 8, and dap 12 wheat sections. Dots correspond to individual spots on the BMKMANU S1000 chip; colors indicate cell type annotation for each spot

Full size image

We prepared 10-μm-thick sections and mounted them onto the sequencing areas of the BMKMANU S1000 Gene Expression chip (6.8 mm × 6.8 mm), which could fully cover the longitudinal and cross sections of the whole grain at the observed stages, facilitating observation of gene expression across the entire grains (Additional file 1: Fig. S1).

The cDNA library was sequenced using the BMKMANU S1000 platform (Additional file 1: Fig. S2). Unique Molecular Identifiers (UMIs) were used to quantify RNA molecules across sample sections. Our results show that the number of tissue-covered spots across all sections ranged from 9934 to 15,107, with a resolution of 27 μm, and the median UMI counts per spot varied from 330 to 1786. Notably, a total of 15,107 spots were identified, with an average of 1410 genes per spot (refer to Additional file 2: Table S1 and Additional file 1: Fig. S3 A, B). Moreover, the gene number sequenced in each cell displayed a significant correlation with the number of identified UMI (mRNA) (see Additional file 1: Fig. S3 C).

Ten functional cell types were finally delineated with both gene expression similarity and anatomical information from cryosections and toluidine blue (TB)-stained images (Fig. 1C, Additional file 1: Fig. S4), comprising (1) two maternal regions: inner pericarp and outer pericarp; (2) seed coat region: testa; (3) embryo region; and (4) six endosperm regions: cavity fluid, aleurone layer, sub-aleurone, central cells of starchy endosperm, transfer cell surrounding endosperm (TCSE), and embryo-surrounding region (ESR). Specifically, we identified six distinct groups within the aleurone layer, five groups within the sub-aleurone, and nine groups within the endosperm cavity fluid (Additional file 1: Fig. S5).

The gene expression abundance across different tissues was statistically analyzed using BSTMatrix software. The number of counts (nCounts) in spots was visualized through a heatmap (Fig. 1D). According to the heatmap (Fig. 1D), gene activity decreased in the outer layers, such as the pericarp and testa, as grain development progressed, but increased in the embryo and ESR. This suggests that gene expression around the embryo was gradually activated during seed development. In particular, at dap 4, the mRNA counts were greater in the aleurone layer, and cell boundaries could be observed there. On the contrary, mRNA transcription was less vigorous in endosperm where cells fused. Therefore, cells in the aleurone layer begin to differentiate before the formation of mature endosperm layer (Additional file 1: Fig. S6). This differentiation leads to the establishment of a specialized cell layer covering the entire surface of the endosperm, except in regions with transfer cells.

For illustrating the developmental relationship among cell types during wheat grain filling, dimension reduction for solely expression of highly variable genes were carried out. Distinct spots were embedded through uniform manifold approximation and projection (UMAP) clustering analysis of different chips. Various cell types demonstrated consistent clustering in embedding coordinates throughout different developmental stages (Fig. 1E). The embryo cells were adjacent to ESR, indicating the similar expression states between them. The endosperm subtypes were located in the margin of the map, while testa and pericarp cells were clustered together in the center. As the grain develops, the degradation of the pericarp and testa begins, and the endosperm gradually starts to proliferate and differentiate. At dap 8, the number of endosperm cells significantly increases, with an even more pronounced increase at dap 12, as indicated by the rising proportion of endosperm cells (Additional file 3: Table S2).

In summary, we generated a comprehensive spatial transcriptomic dataset for wheat. This dataset encompasses 80,000 spatially resolved gene expression profiles across three distinct developmental stages (Additional file 4: Table S3, http://omicsplant.cn/WheatDB/index.html), providing a dynamic perspective on the three-dimensional developmental processes of wheat grains.

Spatial transcriptomic visualization and marker gene identification

We further classified twelve cell clusters through non-supervised dimensionality reduction and clustering of spatial transcriptomic data (Fig. 2A). The identified clusters corresponded well to respective cell types using UMAP (see Fig. 2A, B), but with more obvious boundaries among cell types to avoid artificial annotation errors by visually marking the corresponding cell types based on visible positions. Subsequently, clustering differential expression analysis was conducted between these twelve clusters and the previously defined ten cell types, revealing a concordance correspondence between them (see Fig. 2B).

Fig. 2
figure 2

Identification and validation of tissue-specific marker genes using in situ hybridization. A UMAP representation of the twelve unsupervised clusters. B Cell identity correspondence between annotated cell types and unsupervised clusters. Y-axis indicated annotation of cell types in grain. X-axis indicated seruat clusters. Color: scaled (overlapped cell proportions). C Bubble plot showing transcript enrichment (average expression and percentage) of representative cell type-specific marker genes in the ten cell types. Spot colors correspond to the same tissues in the bubble plot. Genes highlighted in red were validated by in situ hybridization. D In situ hybridization of selected marker genes (i) TraesCS5B02G531100; (ii) TraesCS2B02G347200; (iii) TraesCS7 A02G183600; (iv) TraesCS7 A02G261100; (v) TraesCS7B02G160000 confirmed localization of tissue type-specific transcripts at dap 12. For each in situ hybridization, top panels show spatial visualization and bottom panels with the antisense probes. Scale bar = 100 μm. Aleurone layer (AL), sub-aleurone (Sub-AL), cavity fluid (CF)

Full size image

Embryo and endosperm and seed coat have been precisely separated by cell type markers [29]. To verify the consistency of our data with published data and achieve better-refined classification, we compared our spatial transcriptomic data to the published embryo, endosperm and pericarp transcriptome data (Additional file 1: Fig. S7, Additional file 5: Table S4). Notably, the embryo data in spatial transcriptomics was well confirmed by previous results, while endosperm and pericarp could be further divided into new subtypes. For example, among the 72 endosperm marker genes identified in the previous report [29], 68 were detected in our spatial transcriptome dataset, of which 50 can be used as peripheral endosperm and central endosperm marker genes. Among them, alpha-amylase (TraesCS2B02G004100) is a marker for the TCSE, cytochrome P450 family gene (TraesCS6D02G164800) for the central endosperm, and oxygen-dependent coproporphyrinogen-III oxidase (TraesCS4 A02G363000) for the starchy endosperm (Additional file 1: Fig. S8, Additional file 5: Table S4). Meanwhile, we analyzed the expression profiles of previously reported regulatory genes in grain development (Additional file 6: Table S5) and observed that most of them are actively expressed in the spatial transcriptome, allowing for precise localization of them to specific cell types.

For each identified cell type, specific marker genes were determined based on their tissue-specific expression and functional significance (refer to Fig. 2C, Additional file 1: Fig. S9, and Additional file 7: Table S6). Furthermore, we have compiled a list of marker genes for eight cell types from the endosperm and pericarp, which are presented in Table 1. To validate the mRNA distribution, we conducted in situ hybridization experiments on five genes in different cell types. Consistent with spatial transcriptomic data (Fig. 2D, Additional file 1: Fig. S10), the marker genes demonstrated specific expression localized to the aleurone layer, sub-aleurone, and cavity fluid.

Table 1 Summary of marker genes for ten cell types identified through spatial transcriptomics
Full size table

Transcriptomic differences between grain cell types

To gain deeper insight into the spatial and temporal dynamics of grain filling and grain development, we divided the grain into five key components: seed coat, endosperm, embryo, aleurone layer, and cavity fluid (Fig. 3A) for thorough analysis. We created co-expression networks for ten grain cell types across three developmental stages and then calculated the average expression of each gene within each cell type at every time point. A total of 11,365 eligible genes were grouped into 21 modules (Fig. 3B, Additional file 8: Table S7). Most of the genes (including genes in ME1, 2, 5, 6, 9, 10, and 18) were globally highly expressed in dap4, genes in ME3, 15, and 20 were transcribed vigorously in later stages. Genes in the ME2/19, with higher expression levels in embryo cells, are enriched in nucleocytoplasmic transport and amino acid biosynthesis pathways, according to Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis (Fig. 3C). The aleurone layer higher-expressed gene (ME1/11/12/15) were primarily enriched in protein processing in the endoplasmic reticulum. The endosperm higher-expressed gene (ME3/7/9/13/14/16) were primarily enriched in starch and sucrose metabolism pathways. In addition, cavity fluid (ME5/8/18) and seed coat (ME0/6/17) highly expressed genes were mainly involved in photosynthesis (Fig. 3C) In summary, gene expression across different cell types showed high heterogeneity, indicating that specific biological processes took place in distinct tissues during seed development. According to the KEGG analysis of the representative modules from WGCNA, we classified the differentially expressed genes in specific cell types and presented the relatively specific representative genes such as sucrose synthase (SS) genes, GDSL esterase/lipase LTL1-like genes in Fig. 3D. We utilized KN9204 mutants to validate the functions of representative cell type-specific genes (Fig. 3E) [30]. The aleurone layer mutant taltl1-d1 and the TCSE mutant ltp-a1 both exhibited a significant reduction in grain size (Fig. 3E, F; Additional file 1: Fig. S11, 12; Additional file 9: Table S8). Frozen section analysis revealed that aleurone layer cells in the taltl1-d1 mutants were significantly smaller, whereas the ltp-a1 mutants exhibited an increased number of transfer cells and an expansion of the endosperm cavity (Fig. 3G–I). Further analysis using scanning electron microscopy (SEM) highlighted structural differences in the prismatic endosperm cells, including reduced cell length and increased width (Fig. 3G, J, K). Collectively, these findings evident that mutations in genes with high expression levels in grain-specific cell types significantly impact grain development and filling.

Fig. 3
figure 3

Spatiotemporal co-expression networks for ten grain cell types during development. A Model for wheat grain gene expression. B Spatiotemporal co-expression networks for ten grain cell types during grain development. The columns represent different patterns of co-expression modules and the rows represent development times. C KEGG enrichment for co-expression genes. The enriched KEGG categories were determined using the one-sided version of Fisher’s exact test, followed by the Benjamin-Hochberg correction to obtain adjusted p values for multiple testing. D Dynamic expression of spatio-marker during grain development. Relative expression = Number of cells with gene expression/Total number of cells. E The taltl1-d1 mutant and the taltp1-a1 mutant exhibited a significant reduction in grain size. Scale bar = 1 cm. F Grain surface of mature seeds from WT, taltl1-d1 mutant, and taltp1-a1 mutant. Scale bar = 1 mm. G The top row displays panoramic views of TB staining in mature seeds from WT, taltl1-d1 mutant, and taltp1-a1 mutant. The middle row provides magnified images of the regions outlined by the red solid boxes in the top row. The bottom row displays fractured cross-sections of grains captured via scanning electron microscopy (SEM). Orange dashed boxes highlight aleurone layer cells, while red dashed boxes indicate prismatic cells, orange solid boxes indicate TCSE. Scale bar = 100 μm. H Quantification of AL cell number from WT, taltl1-d1 mutant, and taltp1-a1 mutant. Student’s t test was used to determine the significant difference. I Quantification of AL cell area is from WT, taltl1-d1 mutant, and taltp1-a1 mutant. Student’s t test was used to determine the significant difference. *, p ≤ 0.05; **, p ≤ 0.01. J Quantification of prismatic cells length from WT, taltl1-d1 mutant, and taltp1-a1 mutant. Student’s t‐test was used to determine the significant difference. *, p ≤ 0.05; **, p ≤ 0.01. K Quantification of prismatic cells width from WT, taltl1-d1 mutant, and taltp1-a1 mutant. Student’s t test was used to determine the significant difference. *, p ≤ 0.05; **, p ≤ 0.01

Full size image

Transcriptome dynamics at the transition stage during endosperm differentiation

Endosperm development encompasses intricate processes, including cell division, enlargement, and the accumulation of storage compounds such as starch and protein. Understanding the trajectory of wheat endosperm development is essential for unraveling the molecular mechanisms and regulatory networks that control grain filling and seed development. To investigate the shared progenitor cells and the cell fates necessary for their differentiation, we analyzed all endosperm and embryo cell types at various stages. At dap 4, most cells of the endosperm cavity fluid, aleurone layer, and ESR were mixed (Additional file 1: Fig. S13). After dap 8, they formed a Y-shaped trajectory, with embryo cells at the origin and two distinct paths towards two endpoints: central endosperm (including the starchy endosperm and cavity fluid) and peripheral endosperm (including the embryo, aleurone layer, and ESR) (Fig. 4A,B). The trajectory of cell-type changes was associated with dynamic changes in gene expression, determining cell fates. Thus, we analyzed the top 3000 highly variable genes along the pseudotime of development and categorized them into four clusters (Fig. 4C). KEGG analyses suggest that the initial endosperm cavity fluid, aleurone layer, and ESR cells underwent cell division without significant specialization, and gene expression was enriched in ribosomal, DNA replication, and amino acid synthesis processes (Fig. 4D). The aleurone layer and ESR gradually differentiated towards the distal end of the upper branch, with gene expression enriched in carbon fixation, and monosaccharide and amino acid metabolism and transport (cluster 3 and cluster 4), suggesting their role in nutrient transport. On the lower branch, the endosperm cavity fluid and central endosperm transitioned gradually from the distal end to the central and prismatic endosperm. The gene expression profiles in this path showed enrichment in sucrose and starch synthesis, galactose metabolism, and plant hormone biosynthesis (cluster 1 and cluster 2).

Fig. 4
figure 4

Developmental trajectories of wheat early endosperm. A Visualization of embryo/endosperm cells via UMAP colored by development stages (left) and respective cell types (right). B Visualization of the development trajectory of embryo/endosperm cells colored by development stages (left) and respective cell types (right). C Expression of the top 3000 highly variable genes along pseudotime. D KEGG enrichment for differentially expressed genes. The enriched KEGG categories were determined using the one-sided version of Fisher’s exact test, followed by the Benjamin-Hochberg correction to obtain adjusted p values for multiple testing. E Heatmap of the expression of cluster 1 and cluster 2

Full size image

We analyzed gene expression related to carbon fixation, glycolysis, and sugar and starch synthesis in cluster 1 and cluster 2 (Fig. 4E). These genes were highly expressed in the central and prismatic endosperm at dap 12. The enolase-like isoform (TraesCS5B02G012300) had the highest expression in the prismatic endosperm, indicating its key role in starch synthesis there. Phosphoglycerate kinase (TraesCS6B02G187500) and other sugar synthesis genes were highly expressed in the central starchy endosperm at dap 8 and dap 12, suggesting their primary role in nutrient storage in these cells.

The analysis confirmed that the endosperm undergoes dual differentiation pathways, delineated for nutrient transport and storage functions, respectively. These pathways collectively orchestrate grain-filling processes, thereby facilitating enhanced nutrient storage within the grain.

B3 domain-containing transcription factor TaABI3-B1 identified through spatiotemporal transcriptomics contributes to grain development

Transcription factors (TF) play crucial roles in the development of wheat grains by regulating gene expression during various stages. In total, 4617 TF genes were identified, including 1535 genes from the A subgenome, 1520 genes from the B subgenome, and 1526 genes from the D subgenome. TF genes showed significant differences across cell types at different stages (Additional file 1: Fig. S14). For example, 534 TFs were highly expressed in the embryo, which is the highest number observed across all cell types. The expressed TF families AP2, ARF, B3, HD-ZIP, NAC, and WRKY showed embryo-specific enrichment, whereas B3, BZIP, DOF, MYB, NAC, and WRKY TFs were enriched in the endosperm, and BHLH, BZIP, GRAS, and MADS TFs were enriched in the pericarp, consistent with previous reports (Additional file 10: Table S9). We selected six representative gene families for PCA analysis. The expression of AP2, ARF, and WRKY family genes could separate the embryo from other cell types in PCA plot, indicating a unique expression pattern in embryo cells. The expression of HD-ZIP family genes could separate endosperm from other cell types in PCA plot, indicating a unique expression pattern in endosperm cells. The expression of MADS family genes could separate AL and seedcoat from other cell types in PCA plot, indicating a unique expression pattern in seedcoat cells. But all of them showed little clustering difference in tissues at different stages (Additional file 1: Fig. S15 A), whereas B3 family displayed distinct clustering at dap 4, dap 8, and dap 12, especially in embryo and endosperm, indicating its core role in development.

Bread wheat is a hexaploid (AABBDD) with genetic information acquired from three distinct diploid species. It has been reported that bias in paralog expression (A:B:D) varies between tissues, and this expression unbalance represents the first steps toward neo or subfunctionalization of wheat paralogs [31]. To illustrate transcriptional and functional differences among paralogs in distinct cell types, we conducted a subgenome preference analysis focusing on overall gene expression, with particular attention to TFs. No significant subgenome-specific preference was observed (Additional file 1: Fig. S16 and Additional file 1: Fig. S17). Additionally, we conducted a preferential analysis of the B3 TF family and found a B subgenome preference in the endosperm and aleurone layer at dap 8. However, in the embryo, there was no obvious preference, except for one gene, TaABI3-B1 (TraesCS3B02G452200), which exhibited higher expression from the B subgenome (Additional file 1: Fig. S15B). The biased expression of homologs during embryo and endosperm development was associated with differences in chromatin accessibility across the A, B, and D subgenomes [32]. We combined the published genome-wide profiling of ATAC-seq data in embryo and endosperm [33, 34] and our spatial transcriptomics data of different tissues better to identify epigenetic influences during embryo and endosperm development. The results showed that TaABI3-B1 shows bias towards the B subgenome in both embryo and endosperm. We also analyzed its preference in ATAC and found no significant difference (Additional file 1: Fig. S15B), indicating that TaABI3-B1 expression is mainly transcriptionally regulated rather than epigenetically modified.

By combining the developmental trajectory data and spatially varying genes in the embryo and endosperm, we identified TaABI3-B1, a key regulatory gene in the embryo and endosperm (Fig. 5A). In situ hybridization showed that TaABI3-B1 was specifically expressed in early developmental embryos and ESR (Fig. 5B). Mutants of taabi3-b1 displayed notable enlargement in embryo and grain size (Fig. 5C, D). To further elucidate TaABI3 s role in wheat embryo and endosperm development, we generated three knockdown lines with knockdown of all three homoeologs in the Fielder background. Statistical analysis of T3 generation transgenic lines with amiRNA knockdown revealed significant grain length, width, and hundred-grain weight increases (Fig. 5D, E; Additional file 11: Table S10).

Fig. 5
figure 5

TaABI3-B1 regulates the grain size of wheat. A Transcript distribution of TaABI3-B1. Color bar indicates the normalized UMI counts. B Spatiotemporal expression pattern of TaABI3-B1 in dap 12 embryo as indicated by in situ hybridization. Sense probe served as the negative control. Scale bar = 100 μm. C Observation of mature embryo. Scale bars = 1 mm. D amiR-ABI3 s and taabi3-b1 mutants increase the grain size. Scale bars = 1 mm. E Quantification of grain agronomic traits related traits between the WT plants and amiR-ABI3 s lines. Student’s t test was used to determine the difference significance between amiR-ABI3 s and WT. *, p ≤ 0.05, **, p ≤ 0.01, Data represent mean ± SD (n = 15 biological replicates). F Representative cross sections of mature grains. Scale bars = 100 μm. The red box shows the aleurone layer cell. The experiment was independently repeated three times. G amiR-ABI3 s transgenetic lines and taabi3-b1 mutants significantly increased the aleurone layer cell size. *, p ≤ 0.05, **, p ≤ 0.01. H RT-qPCR assay confirming the relative expression of TaABI3 s in amiR-ABI3 s and WT plants. N = 3

Full size image

Additionally, quality traits showed higher protein content, reduced total starch content, and decreased grain hardness (Fig. 5E), all without negatively impacting normal plant growth. The phenotypes observed in transgenic lines were consistent with those of mutants. Cross-sections of mutants and transgenic lines all showed enlargement of aleurone layer cells (Fig. 5F, G). The expression levels of amiR-TaABI3 s were down than WT (Fig. 5H). This suggests that downregulating TaABI3 s expression can enhance wheat yield components.

Genetic variations in TaABI3-B1 contribute to grain weight and quality

To explore the correlation between natural variations in TaABI3-B1 and grain-related traits (grain length, width, weight, and quality), we assessed polymorphisms in the coding region, 2-kb promoter regions, and 2-kb downstream region of TaABI3-B1 in re-sequencing Watkins collection, which consists of 1056 hexaploid wheat (827 landraces, 229 modern cultivar) that represent global wheat diversity [35]. Fourteen SNPs were found in the promoter of TaABI3-B1, as well as 4 SNPs in exons, and 19 SNPs in 3’ downstream region. Based on SNPs, the 1056 wheat accessions were divided into three groups: 615 accessions with TaABI3-B1-Hap1, 150 accessions with TaABI3-B1-Hap2, and 124 accessions with TaABI3-B1-Hap3 (Fig. 6A). Accessions with TaABI3-B1-Hap1/3 displayed higher grain weight, grain length, and grain width and identified as a high-yield haplotype (Fig. 6B), while TaABI3-B1-Hap2 exhibited greater hardness.

Fig. 6
figure 6

Haplotype analysis of TaABI3-B1 and breeding selection of elite allele. A Schematic showing the polymorphism for each haplotype of TaABI3-B1. The coordinate is related to the transcription start site (TSS). B Violin plot indicating the comparison of grain size-related traits among wheat accession with different haplotypes of TaABI3-B1. Turkey’s honestly significant difference (HSD) multiple tests were used to determine the statistical significance between the three groups. C The percentages of accessions carrying different haplotypes of TaABI3-B1 during the different breeding processes in China. D The percentage of accessions carrying different haplotypes of TaABI3-B1 in each ecological zones of China. The size of pie charts in the geographical map shows the number of accessions, with percentages of the three alleles in different colors (Hap1, purple; Hap2, pink; Hap3, green)

Full size image

To determine the selection of the high-yield haplotype TaABI3-B1-Hap1/3 in China’s wheat breeding process, a detailed analysis was conducted on the Hap frequency within a mini-core collection (MCC) population. We found that the TaABI3-B1-Hap2, was significantly more prevalent in modern wheat cultivars than in traditional landraces and cultivars introduced from other countries (Fig. 6C). Furthermore, the study highlighted variations in allele distributions across different major agricultural ecological zones in China. Specifically, the frequency of the Hap1/3 was higher in Zones I-V than in other regions (Fig. 6D). This geographic distribution indicates that specific ecological conditions might have influenced the selection and prevalence of different alleles, adapting wheat cultivars to diverse environmental conditions.

Discussion

Wheat grain development has been intensively studied, including the molecular mechanisms underlying embryo and endosperm development, dynamic regulatory pathways during embryogenesis, and nutrient mobilization and storage within the grain. Many techniques, such as Laser Capture Microdissection (LCM) and X-ray Micro-Tomography, have been employed in the study of grain development [29, 36, 37]. However, there remains a notable gap in research focusing on the comprehensive analysis of whole-seed morphology at the single-cell level in wheat.

In this study, we generated a high-resolution spatial transcriptomics atlas aimed at elucidating gene expression patterns during the early developmental phase. Through this approach, we subdivided the grain into ten discrete cell types. We identified some cell type-specific genes, consistent with previously identified endosperm markers from various cereals, such as TaSS1, a homolog of maize Shrunken 1 (SH1) [38]. TaRKIN1, a homolog of rice Sucrose non-fermenting- 1-related protein kinas 1b (SnRK1b) [39], and TaSBE1, a homolog of maize and barley Starch branching enzyme [13]. We also identified some genes that showed different expression patterns and whose temporal expression dynamics differed across species, such as TaAL9, first detected high expressed at dap 4 at central starchy endosperm and prismatic endosperm, while HvAL9 in barley was expressed exclusively dap 8 in the aleurone, which indicates that the timing of cell type formation and distinctive gene function can differ across cereals. The aleurone layer, sub-aleurone, central cells of starchy endosperm, prismatic cells of starchy endosperm, and ESR markers were already expressed at dap 4, consistent with previously reported data in barley endosperm, which shows specificity of AL, ETC, and ESR markers were already expressed at dap 4 and dap 8 [40, 41]. This indicates endosperm differentiation is initiated before cellularization.

Previously reported genes involved in grain size regulation and grain quality were found to exhibit cell type-specific expression in our spatial data. For instance, TaATG8a showed high expression in the pericarp at dap 4, influencing grain size by regulating the rate of pericarp degradation [42]. TaPHS was robustly expressed at dap 8 and dap 12 in the central endosperm and prismatic cells, which indicates the spatiotemporal specific expression of TaPHS1 is crucial for the initiation of B-type granules [43]. TaMADS29 showed high expression in the cavity fluid at dap 4, suggesting an influence on nutrient transportation into the endosperm and on filling of developing grains [44]. In our spatiotemporal transcriptomics, we have discovered several high-confidence novel candidates with potential regulatory roles. Additionally, natural variations at these gene loci with GWAS analysis [45, 46] and the use of a wheat mutant library to validate gene function would greatly help the study of wheat grain development and grain yield and quality improvement.

Studies on cultivated barley and other cereals like rice and maize reveal that increasing embryo size often results in a decrease in endosperm size, such as giant embryo (GE) in rice and maize [47], big embryo 1 (BIGE1) in maize [48], and prolamin binding factor (PBF) in barley have been identified [49]. Here we identified that downregulated TaABI3-B1 could increase both embryo size and endosperm size. This enlargement of the endosperm is associated with higher protein levels, lower total starch content, and softer grain texture (Fig. 5E). Notably, a haplotype of TaABI3-B1 exhibiting higher grain width, lower grain weight, and grain length has been selected during the breeding process in China, especially from 1958 to 1999 (Fig. 6C). But in China’s primary wheat-producing regions, the Hap1/3 with higher grain weight and grain length continues to be a predominant cultivated variety, maintaining a dominant position (Fig. 6D). The geographic distribution suggests that ecological conditions influenced the selection of different alleles, adapting wheat cultivars to diverse environments. Additionally, data on relative embryo sizes in modern and legacy wheat along with gene information can be used in breeding programs to improve grain quality.

In conclusion, we utilized spatial transcriptomics to identify a large number of genes with specific spatiotemporal expressions. Numerous genes display spatially and temporally precise expression profiles within embryonic tissues. Notably, certain transcription factors exhibit predominant expression within distinct cell types or localized regions of the embryo, indicative of their involvement in modulating cell fate determination and tissue patterning processes [47]. However, applying spatial transcriptomics in wheat grain has its limitations, particularly when it comes to genes with low expression abundance at the early stage in embryo development. Many genes associated with zygotic activation were mainly observed at dap 2 and dap 4, including cell cycle and cytokine signaling genes, but sharply decreased at dap 8 and dap12 [29, 33]. Detecting low expressed genes remains challenging. Thus, integrating spatial transcriptomics data with complementary techniques such as single-cell RNA sequencing or immunohistochemistry can help overcome some of these limitations and provide a more comprehensive understanding of gene expression patterns within tissues.

Conclusions

In this study, we constructed a spatial transcriptomic atlas of wheat grains and identified a large number of genes with spatiotemporally specific expression patterns (http://omicsplant.cn/WheatDB). Many genes exhibited precise spatial and temporal expression profiles in embryonic tissues. We found that certain transcription factors are differentially expressed across distinct cell types or developmental stages, suggesting their involvement in key processes of grain differentiation. Furthermore, through co-expression network and cell differentiation trajectory analyses, we identified a key transcription factor, TaABI3-B1, which regulates grain differentiation. TaABI3-B1 is specifically expressed in the early-stage embryo and surrounding endosperm, and functional validation confirmed its role as a negative regulator of both grain and embryo size. In summary, our study provides a visualized atlas of early grain gene expression, offering valuable insights for precise molecular breeding of wheat grains in the future.

Methods

Plant growth and tissue preparation

Seeds of winter wheat (Jimai 22 and KN9204) were germinated and subsequently transferred to soil, where they were grown at 4 °C in the dark for 45 days. Then they were transferred to a greenhouse (20 °C–22 °C, 16-h light/8-h dark). Grains were harvested at dap 4, dap 8, and dap 12.

The EMS mutants from the tetraploid variety “Kronos” were produced by Dr. Jorge Dubcovsky’s group at the University of California, Davis [50]. The EMS mutants from the tetraploid variety “KN9204” were produced by Dr. Jun Xiao’s group at the Institute of Genetics and Developmental Biology, Chinese Academy of Sciences [30], and were backcrossed two times before analysis.

K ronos mutants (spring wheat) and amiRNA-ABI3 s T3 generation lines were grown in the experimental field of the Peking University Institute of Advanced Agricultural Sciences, Shandong, China (119° 18′ E, 36° 16′ N). For the field experiment, each line was planted in three replicates. Each replicate consisted of 10 rows, each 1.5 m long, with a row-to-row spacing of 0.20 m.

For agronomic trait measurements, grains were harvested from the spike, and 150 grains were randomly selected for analysis. A normal distribution analysis of the selected grains confirmed that grain sizes followed a normal distribution, ensuring the representation of grains from both the apical and basal regions. Grain width, grain length, and hundred-grain weight were measured using a Wanshen SC-G seed detector (Hangzhou Wanshen Detection Technology Co., Ltd.). Grain length and width were measured for more than 70 grains. Protein content and total starch content were measured using a near-infrared spectrum instrument (DA7250) with over 150 grains. Grain hardness was assessed using a Single Kernel Characterization System (SKCS4100) by crushing 300 kernels. Agronomic traits, including grain length, grain width, hundred-grain weight, protein content, and total starch content, were measured with 15 representative plants. Grain hardness was measured with three biological replicates.

Grain fixation, staining, and imaging

Wheat grains at dap 4, dap 8, and dap 12 were prepared: soaked in a 75% optimal cutting temperature compound (OCT; Sakura Finetek Europe B.V.) solution under vacuum for 5 min, then embedded in OCT and stored at − 80 ℃. Once equilibrated to − 20 ℃, samples were sliced into 10-μm-thick sections. Quality-approved samples were mounted onto BMKMANU S1000 chips, treated with methanol fixation, and stained with TB for imaging.

Library construction, sequencing, and expression atlas analyses

Tissue sectioning, TB staining, imaging, and initial permeabilization were conducted following the BMKMANU S1000 Tissue Optimization Kit user guide (BMKMANU, ST03003). Each capture area of the gene expression slide (6.8 × 6.8 mm2) contains 2,200,000 barcoded spots that are 2.5 μm in diameter (4.8 μm center to center between spots) providing an average 4 to 20 spots under a cell. Before performing the complete protocol, Tissue Optimization Kit was performed according to the manufacturer’s instructions, permeabilization (8 min). Sequence libraries were then processed according to the manufacturer’s instructions (BMKMANU, Library Construction Kit, ST03002 - 34). The Illumina library was sequenced using the Illumina NovaSeq platform. Raw sequencing data were mapped to the wheat reference genome (IWGSC RefSeq v1.1) using STAR v2.5.3a [51] with default parameters.

For each sample, FASTQ files and manually aligned histology images were analyzed with BSTMatrix 2.0. Then the result was mapped to the genome by using STAR genome aligner version v.2.5.1b. Processed data were imported into R via Seurat V.4.0.1. Spatial spots featuring more than 30% of mitochondrial genes and less than 300 genes were filtered out. Genes with counts in less than 5 spatial spots were discarded. Spots featuring folded were removed. Raw counts were normalized with the SCTransform function of Seurat using the “assay = spatial” parameter.

The image, adjusted by BSTViewer V1.42, along with the corresponding level 4 (27 μm) matrix, was used for downstream analysis. Quality control and normalization of gene expression matrixes for respective samples were performed using Seurat V 4.3.0.1 [52] with parameters set to min.cell = 5, min.features = 100. Batch effect correction was carried out with the IntegrateData function in the Seurat package with 3000 anchors. The scaled data were further used for dimension reduction and clustering with 30 principal components and a resolution of 0.5. The cell identities defined by spatial information were used for annotation of each cluster. Specific expression genes in respective clusters were identified with FindAllMarkers with logfc.threshold = 0.1, only.pos = TRUE, min.pct = 0.01. The expression level for each gene for cell types of the four samples was identified by two methods: 1, mean = sum of normalized UMI counts/number of cells, and 2, expression proportion = number of cells with expression/number of cells * 100. The resulting expression matrixes were analyzed using the WGCNA (v.1.69) pipeline with the default filtration process for the identification of spatially co-expressed gene modules [53].

Construction of embryo and endosperm cell trajectories

The Seurat objects of embryo and endosperm cells were extracted for the reconstruction of cell atlases with the same pipeline described above. Three thousand highly variable genes were identified. The expression matrix was further analyzed and ordered by the identified highly variable genes using monocle2 v2.14.0 [54]. Dimension reduction was performed with DDRTree methods. Differential expression genes along the pseudotime were identified and the top 3000 genes were used for another round of dimension reduction.

Marker gene identification and RNA in situ hybridization

After analyzing the gene expression of cell populations, the top 20 genes in each cell population were specifically expressed in tissues and gene functions, and some genes were used as marker genes of the cell population. After electronic in situ hybridization, these genes were randomly selected to express tissue-specific and functionally related genes for in situ hybridization validation. Using gene-specific fragments as templates, DIG-labeled probes were synthesized. Developing grains were fixed with formalin-aceto-alcohol solution (50%), dehydrated in a series of ethanol concentrations and cleared in histoclear, embedded in paraplast (Sigma, P3558), and sectioned to a thickness of 7 μm. The following steps of RNA hybridization, immunologic detection, and signal capture with the hybridized probes were compiled as described [55].

Scanning electron microscopy

The grains were broken transversely, mounted on aluminum stubs, and coated with gold. A Talos L120 C TEM (TTALOSL120 C) scanning electron microscope was used to observe the samples.

KEGG enrichment analysis

The enrichment level of differentially expressed genes (DEGs) in the KEGG pathways (http://www.genome.jp/kegg/) was determined using the KOBAS (3.0) software with a threshold of FDR ≤ 0.05 considered for defining significantly enriched pathways.

Creation of transgenic plants of TaABI3s

To knock down TaABI3 s in wheat cv. Fielder, oligos for amiRNA 5′-AAAATCGGTACCGCATGCTT- 3′ located in the third exon were used. The amiRNA vector was constructed following the methodology described in reference [56]. The amiR vectors were introduced into hexaploid wheat (cv. Fielder) via Agrobacterium-mediated transformation as described [57]. PCR, herbicide (glufosinate) spraying, and a QuickStix Kit for bialaphos resistance (bar) were used to verify positive transgenic plants across the T0, T1, T2, and T3 generations.

Haplotype analysis of TaABI3-B1

Natural variation retrieved from the coding region, 2-kb promoter regions, and 2-kb downstream region sequencing project of the re-sequencing Watkins collection, which consists of 1056 hexaploid wheat landraces that represent global wheat diversity [28], were used to assess the allelic variation of TaABI3-B1. The polymorphism with missing rate < 0.5, min allele frequency > 0.05, and heterozygosity < 0.5 were retained for further haplotype analysis using Haploview 4.2, and the differences of the grain size phenotypes corresponding to different haplotypes were tested. The Chinese wheat mini-core collection [58], composed of 287 representative selected varieties for the Chinese national collection, were used to assess the haplotype frequency in each breeding process of China and among the major Chinese agro-ecological zones.

Data availability

The raw sequencing data of spatial transcriptomics generated in this study have been deposited in the NCBI (https://www.ncbi.nlm.nih.gov/) Sequence Read Archive under the BioSample of SAMN41774528 [59]. The expression levels of all genes in this study have been deposited in Figshare (https://figshare.com/articles/dataset/___/28683071) [60]. RNA-seq data of embryo in diploid wheat was download from Gene Expression Omnibus (GEO) under accession number GSE129695 [61]. RNA-seq and ATAC-seq generated during embryo development was download from Genome Sequence Archive (GSA) under accession number PRJCA008382 [62]. RNA-seq and ATAC-seq generated during endosperm development was download from Genome Sequence Archive (GSA) under accession number PRJCA022666 [63]. All scripts used in this work are available at Github (https://github.com/XiaohuiLi1220/Wheat_grain_SRT) [64] and Zenodo (https://zenodo.org/records/15110459) [65].

References

  1. Chaudhury AM, Koltunow A, Payne T, Luo M, Tucker MR, Dennis ES, Peacock WJ. Control of early seed development. Annu Rev Cell Dev Biol. 2001;17:677–99.

    Article  CAS  PubMed  Google Scholar 

  2. Zhong Y, Chen Y, Pan M, Wang H, Sun J, Chen Y, Cai J, Zhou Q, Wang X, Jiang D. Insights into the functional components in wheat grain: spatial pattern, underlying mechanism and cultivation regulation. Plants. 2023;12: 2192.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  3. Rathan ND, Krishna H, Ellur RK, Sehgal D, Govindan V, Ahlawat AK, Krishnappa G, Jaiswal JP, Singh JB, Sv S, et al. Genome-wide association study identifies loci and candidate genes for grain micronutrients and quality traits in wheat (Triticum aestivum L.). Sci Rep. 2022;12:7037.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  4. Tekeu H, Ngonkeu ELM, Bélanger S, Djocgoué PF, Abed A, Torkamaneh D, Boyle B, Tsimi PM, Tadesse W, Jean M, Belzile F. GWAS identifies an ortholog of the rice D11 gene as a candidate gene for grain size in an international collection of hexaploid wheat. Sci Rep. 2021;11:19483.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  5. Khakimov B, Jespersen BM, Engelsen SB. Comprehensive and comparative metabolomic profiling of wheat, barley, oat and rye using gas chromatography-mass spectrometry and advanced chemometrics. Foods. 2014;3:569–85.

    Article  PubMed  PubMed Central  Google Scholar 

  6. Zhang S, Ghatak A, Bazargani MM, Bajaj P, Varshney RK, Chaturvedi P, Jiang D, Weckwerth W. Spatial distribution of proteins and metabolites in developing wheat grain and their differential regulatory response during the grain filling process. Plant J. 2021;107:669–87.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  7. Shewry PR, Wan Y, Hawkesford MJ, Tosi P. Spatial distribution of functional components in the starchy endosperm of wheat grains. J Cereal Sci. 2020;91:102869.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  8. Bressan D, Battistoni G, Hannon GJ. The dawn of spatial omics. Sci Am Assoc Adv Sci. 2023;381:4964.

    Google Scholar 

  9. Li X, Li B, Gu S, Pang X, Mason P, Yuan J, Jia J, Sun J, Zhao C, Henry R. Single-cell and spatial RNA sequencing reveal the spatiotemporal trajectories of fruit senescence. Nat Commun. 2024;15:3108.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  10. Rhaman MS, Ali M, Ye W, Li B. Opportunities and challenges in advancing plant research with single-cell omics. Genomics Proteomics Bioinform. 2024;22:qzae026.

    Article  Google Scholar 

  11. Giolai M, Verweij W, Lister A, Heavens D, MacAulay I, Clark MD. Spatially resolved transcriptomics reveals plant host responses to pathogens. Plant Methods. 2019;15:114.

    Article  PubMed  PubMed Central  Google Scholar 

  12. Wang Y, Luo Y, Guo X, Li Y, Yan J, Shao W, Wei W, Wei X, Yang T, Chen J, et al. A spatial transcriptome map of the developing maize ear. Nature Plants. 2024;10:815–27.

    Article  CAS  PubMed  Google Scholar 

  13. Fu Y, Xiao W, Tian L, Guo L, Ma G, Ji C, Huang Y, Wang H, Wu X, Yang T, et al. Spatial transcriptomics uncover sucrose post-phloem transport during maize kernel development. Nat Commun. 2023;14:7191.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  14. Peirats-Llobet M, Yi C, Liew Lim C, Berkowitz O, Narsai R, Lewsey Mathew G, Whelan J. Spatially resolved transcriptomic analysis of the germinating barley grain. Nucleic Acids Res. 2023;51:7798–819.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  15. Liu Y, Li C, Han Y, Li R, Cui F, Zhang H, Su X, Liu X, Xu G, Wan S, Li G. Spatial transcriptome analysis on peanut tissues shed light on cell heterogeneity of the peg. Plant Biotechnol J. 2022;20:1648–50.

    Article  PubMed  PubMed Central  Google Scholar 

  16. Liu Z, Kong X, Long Y, Liu S, Zhang H, Jia J, Cui W, Zhang Z, Song X, Qiu L, et al. Integrated single-nucleus and spatial transcriptomics captures transitional states in soybean nodule maturation. Nature Plants. 2023;9:515–24.

    Article  CAS  PubMed  Google Scholar 

  17. Song X, Guo P, Xia K, Wang M, Liu Y, Chen L, Zhang J, Xu M, Liu N, Yue Z, et al. Spatial transcriptomics reveals light-induced chlorenchyma cells involved in promoting shoot regeneration in tomato callus. Proc Natl Acad Sci. 2023;120:e2310163120.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  18. Liu C, Leng J, Li Y, Ge T, Li J, Chen Y, Guo C, Qi J. A spatiotemporal atlas of organogenesis in the development of orchid flowers. Nucleic Acids Res. 2022;50:9724–37.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  19. Li R, Chen X, Yang X. Navigating the landscapes of spatial transcriptomics: how computational methods guide the way. Wiley Interdiscip Rev RNA. 2024;15:e1839.

    Article  PubMed  Google Scholar 

  20. Bhati KK, Alok A, Kumar A, Kaur J, Tiwari S, Pandey AK. Silencing of ABCC13 transporter in wheat reveals its involvement in grain development, phytic acid accumulation and lateral root formation. J Exp Bot. 2016;67:4379–89.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  21. Ma M, Wang Q, Li Z, Cheng H, Li Z, Liu X, Song W, Appels R, Zhao H. Expression of TaCYP78A3, a gene encoding cytochrome P450 CYP78A3 protein in wheat (Triticum aestivum L.), affects seed size. Plant J. 2015;83:312–25.

    Article  CAS  PubMed  Google Scholar 

  22. Balmer Y, Vensel WH, Cai N, Manieri W, Schürmann P, Hurkman WJ, Buchanan BB. A complete ferredoxin/thioredoxin system regulates fundamental processes in amyloplasts. Proc Natl Acad Sci. 2006;103:2988–93.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  23. Kumari J, Lakhwani D, Jakhar P, Sharma S, Tiwari S, Mittal S, Avashthi H, Shekhawat N, Singh K, Mishra KK, et al. Association mapping reveals novel genes and genomic regions controlling grain size architecture in mini core accessions of Indian National Genebank wheat germplasm collection. Front Plant Sci. 2023;14:1148658.

    Article  PubMed  PubMed Central  Google Scholar 

  24. Guo X, Fu Y, Lee YRJ, Chern M, Li M, Cheng M, Dong H, Yuan Z, Gui L, Yin J, et al. The PGS1 basic helix-loop-helix protein regulates Fl3 to impact seed growth and grain yield in cereals. Plant Biotechnol J. 2022;20:1311–26.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  25. Liu Y, Hou J, Wang X, Li T, Majeed U, Hao C, Zhang X. The NAC transcription factor NAC019-A1 is a negative regulator of starch synthesis in wheat developing endosperm. J Exp Bot. 2020;71:5794–807.

    Article  CAS  PubMed  Google Scholar 

  26. Chen J, Zhao L, Li H, Yang C, Lin X, Lin Y, Zhang H, Zhang M, Bie X, Zhao P, et al. Nuclear factor-Y-polycomb repressive complex2 dynamically orchestrates starch and seed storage protein biosynthesis in wheat. Plant Cell. 2024;36:4786–803.

    Article  PubMed  Google Scholar 

  27. Liu Y, Wang X, Hao C, Irshad A, Li T, Liu H, Hou J, Zhang X. TaABI19 positively regulates grain development in wheat. J Integr Agric. 2023;22:41–51.

    Article  CAS  Google Scholar 

  28. Zheng X, Li Q, Li C, An D, Xiao Q, Wang W, Wua Y. Intra-kernel reallocation of proteins in maize depends on VP1-mediated scutellum development and nutrient assimilation. Plant Cell. 2019;31:2613–35.

    CAS  PubMed  PubMed Central  Google Scholar 

  29. Xiang D, Quilichini TD, Liu Z, Gao P, Pan Y, Li Q, Nilsen KT, Venglat P, Esteban E, Pasha A, et al. The transcriptional landscape of polyploid wheats and their diploid ancestors during embryogenesis and grain development. Plant Cell. 2019;31:2888–911.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  30. Wang D, Li Y, Wang H, Xu Y, Yang Y, Zhou Y, Chen Z, Zhou Y, Gui L, Guo Y, et al. Boosting wheat functional genomics via an indexed EMS mutant library of KN9204. Plant Commun. 2023;4:100593.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  31. Wang M, Li Z, Zhang Y, Zhang Y, Xie Y, Ye L, Zhuang Y, Lin K, Zhao F, Guo J, et al. An atlas of wheat epigenetic regulatory elements reveals subgenome divergence in the regulation of development and stress responses. Plant Cell. 2021;33:865–81.

    Article  PubMed  PubMed Central  Google Scholar 

  32. Pei H, Li Y, Liu Y, Liu P, Zhang J, Ren X, Lu Z. Chromatin accessibility landscapes revealed the subgenome-divergent regulation networks during wheat grain development. Abiotech. 2023;4:8–19.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  33. Zhao L, Yang Y, Chen J, Lin X, Zhang H, Wang H, Wang H, Bie X, Jiang J, Feng X, et al. Dynamic chromatin regulatory programs during embryogenesis of hexaploid wheat. Genome Biol. 2023;24:7.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  34. Zhao L, Chen J, Zhang Z, Wu W, Lin X, Gao M, Yang Y, Zhao P, Xu S, Yang C, et al. Deciphering the transcriptional regulatory network governing starch and storage protein biosynthesis in wheat for breeding improvement. Adv Sci. 2024;11:e2401383.

    Article  Google Scholar 

  35. Song B, Ning W, Wei D, Jiang M, Zhu K, Wang X, Edwards D, Odeny DA, Cheng S. Plant genome resequencing and population genomics: current status and future prospects. Mol Plant. 2023;16:1252–68.

    Article  CAS  PubMed  Google Scholar 

  36. Boudichevskaia A, Ruban A, Thiel J, Fiebig A, Houben A. Tissue-specific transcriptome analysis reveals candidate transcripts associated with the process of programmed B chromosome elimination in Aegilops speltoides. Int J Mol Sci. 2020;21:7596.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  37. Legland D, Le TDQ, Alvarado C, Girousse C, Chateigner-Boutin AL. New growth-related features of wheat grain pericarp revealed by synchrotron-Based X-ray micro-tomography and 3D reconstruction. Plants. 2023;12:1038.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  38. Chen C, Li T, Zhu S, Liu Z, Shi Z, Zheng X, Chen R, Huang J, Shen Y, Luo S, et al. Characterization of imprinted genes in rice reveals conservation of regulation and imprinting with other plant species. Plant Physiol. 2018;177:1754–71.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  39. Kanegae H, Miyoshi K, Hirose T, Tsuchimoto S, Mori M, Nagato Y, Takano M. Expressions of rice sucrose non-fermenting-1 related protein kinase 1 genes are differently regulated during the caryopsis development. Plant Physiol Biochem. 2005;43:669–79.

    Article  CAS  PubMed  Google Scholar 

  40. Hertig C, Rutten T, Melzer M, Schippers JHM, Thiel J. Dissection of developmental programs and regulatory modules directing endosperm transfer cell and aleurone identity in the syncytial endosperm of barley. Plants. 2023;12:1594.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  41. Kovacik M, Nowicka A, Zwyrtková J, Strejčková B, Vardanega I, Esteban E, Pasha A, Kaduchová K, Krautsova M, Červenková M, et al. The transcriptome landscape of developing barley seeds. Plant Cell. 2024;36:2512–30.

    Article  PubMed  Google Scholar 

  42. Li YB, Yan M, Cui DZ, Huang C, Sui XX, Guo FZ, Fan QQ, Chu XS. Programmed degradation of pericarp cells in wheat grains depends on autophagy. Front Genet. 2021;12:784545.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  43. Kamble NU, Makhamadjonov F, Fahy B, Martins C, Saalbach G, Seung D. Initiation of B-type starch granules in wheat endosperm requires the plastidial α-glucan phosphorylase PHS1. Plant Cell. 2023;35:4091–110.

    Article  PubMed  PubMed Central  Google Scholar 

  44. Liu G, Zhang R, Li S, Ullah R, Yang F, Wang Z, Guo W, You M, Li B, Xie C, et al. TaMADS29 interacts with TaNF-YB1 to synergistically regulate early grain development in bread wheat. Sci China Life Sci. 2023;66:1647–64.

    Article  CAS  PubMed  Google Scholar 

  45. Khan H, Krishnappa G, Kumar S, Mishra CN, Krishna H, Devate NB, Rathan ND, Parkash O, Yadav SS, Srivastava P, et al. Genome-wide association study for grain yield and component traits in bread wheat (Triticum aestivum L.). Front Genet. 2022;13:982589.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  46. Yang Y, Amo A, Wei D, Chai Y, Zheng J, Qiao P, Cui C, Lu S, Chen L, Hu Y-G. Large-scale integration of meta-QTL and genome-wide association study discovers the genomic regions and candidate genes for yield and yield-related traits in bread wheat. Theor Appl Genet. 2021;134:3083–109.

    Article  CAS  PubMed  Google Scholar 

  47. Zhang P, Allen WB, Nagasawa N, Ching AS, Heppard EP, Li H, Hao X, Li X, Yang X, Yan J, et al. A transposable element insertion within ZmGE2 gene is associated with increase in embryo to endosperm ratio in maize. Theor Appl Genet. 2012;125:1463–71.

    Article  CAS  PubMed  Google Scholar 

  48. Suzuki M, Sato Y, Wu S, Kang BH, McCarty DR. Conserved functions of the MATE transporter BIG EMBRYO1 in regulation of lateral organ size and initiation rate. Plant Cell. 2015;27:2288–300.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  49. Trafford K, Orman-Ligeza B, Borrill P, Chia T, Chirico M, Dolezel J, Drea S, Karafiátov M, Schatlowski N, Solomon C, et al. LYS3 encodes a prolamin-box-binding transcription factor that controls embryo growth in barley and wheat. J Cereal Sci. 2020;93:102965.

    Article  PubMed  PubMed Central  Google Scholar 

  50. Krasileva KV, Vasquez-Gross HA, Howell T, Bailey P, Paraiso F, Clissold L, Simmonds J, Ramirez-Gonzalez RH, Wang X, Borrill P, Fosker C, et al. Uncovering hidden variation in polyploid wheat. Proc Natl Acad Sci. 2017;114:E913–21.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  51. Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, Batut P, Chaisson M, Gingeras TR. STAR: ultrafast universal RNA-seq aligner. Bioinformatics. 2012;29:15–21.

    Article  PubMed  PubMed Central  Google Scholar 

  52. Butler A, Hoffman P, Smibert P, Papalexi E, Satija R. Integrating single-cell transcriptomic data across different conditions, technologies, and species. Nat Biotechnol. 2018;36:411–20.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  53. Langfelder P, Horvath S. WGCNA: an R package for weighted correlation network analysis. BMC Bioinformatics. 2008;9: 559.

    Article  PubMed  PubMed Central  Google Scholar 

  54. Qiu X, Mao Q, Tang Y, Wang L, Chawla R, Pliner HA, Trapnell C. Reversed graph embedding resolves complex single-cell trajectories. Nat Methods. 2017;14:979–82.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  55. Meng WJ, Cheng ZJ, Sang YL, Zhang MM, Rong XF, Wang ZW, Tang YY, Zhang XS. Type-B Arabidopsis response regulators specify the shoot stem cell niche by dual regulation of WUSCHEL. Plant Cell. 2017;29:1357–72.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  56. Park W, Zhai J, Lee J-Y. Highly efficient gene silencing using perfect complementary artificial miRNA targeting AP1 or heteromeric artificial miRNA targeting AP1 and CAL genes. Plant Cell Rep. 2009;28:469–80.

    Article  CAS  PubMed  Google Scholar 

  57. Ishida Y, Tsunashima M, Hiei Y, Komari T. Wheat (Triticum aestivum L.) transformation using immature embryos. Methods Mol Biol. 2015;1223:189–98.

    Article  CAS  PubMed  Google Scholar 

  58. Li A, Hao C, Wang Z, Geng S, Jia M, Wang F, Han X, Kong X, Yin L, Tao S, et al. Wheat breeding history reveals synergistic selection of pleiotropic genomic sites for plant architecture and grain yield. Mol Plant. 2022;15:504–19.

    Article  CAS  PubMed  Google Scholar 

  59. Li XH, Wan YM, Wang DZ, Li XG, Wu JJ, Xiao J, Chen KM, Han X, Chen Y. Spatiotemporal transcriptomics reveals key gene regulation for grain yield and quality in wheat. Datasets. Biosample. 2025. https://www.ncbi.nlm.nih.gov/biosample/SAMN41774528/.

  60. Li XH, Wan YM, Wang DZ, Li XG, Wu JJ, Xiao J, Chen KM, Han X, Chen Y. Spatiotemporal transcriptomics reveals key gene regulation for grain yield and quality in wheat. Figshare. 2025. https://figshare.com/articles/dataset/___/28683071.

  61. Xiang D, Quilichini TD, Liu Z, Gao P, Pan Y, Li Q, et al. The transcriptional landscape of polyploid wheats and their diploid ancestors during embryogenesis and grain development. datasets. Gene Expression Omnibus. 2019. https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE129695.

  62. Zhao L, Yang Y, Chen J, Lin X, Zhang H, Wang H, Wang H, Bie X, Jiang J, Feng X, et al. Dynamic chromatin regulatory programs during embryogenesis of hexaploid wheat. Datasets. Genome Sequence Archive. 2022. https://ngdc.cncb.ac.cn/gsa/browse/CRA006151.

  63. Zhao L, Chen J, Zhang Z, Wu W, Lin X, Gao M, Yang Y, Zhao P, Xu S, Yang C, et al. Deciphering the transcriptional regulatory network governing starch and storage protein biosynthesis in wheat for breeding improvement. Datasets. Genome Sequence Archive. 2024. https://ngdc.cncb.ac.cn/gsa/browse/CRA014813.

  64. Li XH, Wan YM, Wang DZ, Li XG, Wu JJ, Xiao J, Chen KM, Han X, Chen Y. Spatiotemporal transcriptomics reveals key gene regulation for grain yield and quality in wheat. Github. 2025. https://github.com/XiaohuiLi1220/Wheat_grain_SRT.

  65. Li XH, Wan YM, Wang DZ, Li XG, Wu JJ, Xiao J, Chen KM, Han X, Chen Y. Spatiotemporal transcriptomics reveals key gene regulation for grain yield and quality in wheat. Zenodo. 2025. https://zenodo.org/records/15110459.

Download references

Acknowledgements

We thank Dr. Daolin Fu of Shandong Agriculture University for providing Kronos mutants which were originally produced by Dr. Jorge Dubcovsky’s lab at the UC Davis. We thank Dr. Sheila McCormick from UC Berkeley and USDA Plant Gene Expression Center (PGEC) for valuable suggestions. We thank Dr. Bosheng Li and Dr. Zhiliang Yue from Peking University Institute of Advanced Agricultural Sciences (PKU-IAAS) for high quality frozen section imaging.

Peer review information

Eduard Akhunov and Wenjing She were the primary editors of this article and managed its editorial process and peer review in collaboration with the rest of the editorial team. The peer-review history is available in the online version of this article.

Funding

This research is supported by the Shandong Provincial Natural Science Foundation (SYS202206 and ZR2021MC041 ZR2021MC190) and the Taishan Scholars Program (tsqn202103162).

Author information

Authors and Affiliations

  1. State Key Laboratory of Wheat Improvement, Peking University Institute of Advanced Agricultural Sciences, Shandong Laboratory of Advanced Agricultural Sciences in Weifang, Shandong, 261325, China

    Xiaohui Li, Yiman Wan, Xue Han & Yuan Chen

  2. Key Laboratory of Plant Cell and Chromosome Engineering, Institute of Genetics and Developmental Biology, Chinese Academy of Sciences, Beijing, 100101, China

    Dongzhi Wang & Jun Xiao

  3. College of Life Sciences, Shandong Agricultural University, Taian, Shandong, 271018, China

    Xingguo Li

  4. State Key Laboratory of Wheat Improvement, College of Agronomy, Shandong Agricultural University, Taian, Shandong, 271018, China

    Jiajie Wu

  5. College of Life Sciences, National Key Laboratory of Crop Improvement for Stress Tolerance and Production, Northwest Agriculture and Forestry University, Yangling, Shaanxi, 712100, China

    Xiaohui Li & Kunming Chen

  6. College of Plant Protection, Shandong Agricultural University, Taian, Shandong, 271018, China

    Yiman Wan

  7. College of Advanced Agricultural Sciences, University of Chinese Academy of Sciences, Beijing, 100049, China

    Jun Xiao

  8. Centre of Excellence for Plant and Microbial Science (CEPAMS), JIC-CAS, Beijing, 100101, China

    Jun Xiao

Authors
  1. Xiaohui Li
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Yiman Wan
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Dongzhi Wang
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Xingguo Li
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Jiajie Wu
    View author publications

    You can also search for this author inPubMed Google Scholar

  6. Jun Xiao
    View author publications

    You can also search for this author inPubMed Google Scholar

  7. Kunming Chen
    View author publications

    You can also search for this author inPubMed Google Scholar

  8. Xue Han
    View author publications

    You can also search for this author inPubMed Google Scholar

  9. Yuan Chen
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

Y.C. conceived and supervised the study. X.L. performed tissue sectioning, spatial transcriptomes sequencing experiments and analyzed the data; X.H. conducted bioinformatic analysis and presented the data. Y.W generated transgenic lines, investigated phenotypes, and performed molecular experiments; X.L. and Y.W. did in situ hybridization; D.W. did the haploid and selection analysis; J.W. provided Kronos mutants; J.X. provided KN9204 mutants; X.L.and Y.C. wrote the manuscript; J.W., J.X., X.L. and K.C. revised the manuscript; All authors discussed the results and commented on the manuscript. All authors read and approved the final manuscript.

Corresponding authors

Correspondence to Xue Han or Yuan Chen.

Ethics declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

13059_2025_3569_MOESM1_ESM.pdf

Additional file 1. Fig. S1 Overview of spatial transcriptomics experiment on wheat development grains. Fig. S2 The analysis pipeline for spatial transcriptomic study. Fig. S3 The density of expressed genes and transcripts in a spot. Fig. S4 Spatial visualization of the unbiased spot clustering for dap 4, dap 8 and dap 12 wheat sections. Merged bright field images and spatial clusters of the other three sections. The white circle showed cluster 11. Fig. S5 The defined clusters after dimensional reduction. Fig. S6 Light microscopy of grain at dap 4, dap 8 and dap 12. Fig. S7 Heat map shows the expression of known markers that are consistent with the previous studies. Fig. S8 Representatives of spatial visualization using known markers that are consistent with previous studies. Fig. S9 Representatives of spatial visualization using newly defined markers. Fig. S10 The sense controls of the marker genes. Fig. S11 RNA-seq analysis of the expression of TaLTL1-D1 and TaLTP1-A1. Fig. S12 Schematic diagram of mutants in KN9204 resulting in truncated proteins (gray box). Scales indicate the number of amino acid (aa) residues. Fig. S13 Visualization of grain development along with pseudotime. Fig. S14. Heat map showing the expression of TFs in ten grain cell types across three developmental stages. Fig. S15 Diverse expression patterns of TFs across cell types and stages. Fig. S16 Balanced homoeologs in gene expression show unbalanced expression patterns in ten grain cell types across three developmental stages defined by stRNA-seq data. The ternary plot shows the expression pattern of detected genes in stRNA-seq. Fig. S17 Balanced homeologs in TFs expression show unbalanced expression patterns in ten grain cell types across three developmental stages defined by stRNA-seq data. The ternary plot shows expression patterns of detected genes in stRNA-seq.

Additional file 2. Table S1. The statistics of sequencing data.

Additional file 3. Table S2. Cell types vs stages.

Additional file 4. Table S3. Summary of all gene expression in grain spatiotemporal transcriptomics.

13059_2025_3569_MOESM5_ESM.xlsx

Additional file 5. Table S4. Summary of marker genes for ten cell types identified from RNA-seq data in laser microdissection studies.

Additional file 6. Table S5. List of key genes known to regulate grain size in wheat.

Additional file 7. Table S6. Molecular marker genes in different cell types.

Additional file 8. Table S7. WGCNA module genes.

Additional file 9. Table S8. The mutation information and grain data of the KN9204 mutants.

Additional file 10. Table S9. All TF genes and TF families in grain spatiotemporal transcriptomics.

13059_2025_3569_MOESM11_ESM.xlsx

Additional file 11. Table S10. Grain length and width of amiR-ABI3 s transgenic lines under field and greenhouse growth conditions.

Additional file 12. Table S11. Primers used in this study.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Li, X., Wan, Y., Wang, D. et al. Spatiotemporal transcriptomics reveals key gene regulation for grain yield and quality in wheat. Genome Biol 26, 93 (2025). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13059-025-03569-8

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13059-025-03569-8

Genome Biology

ISSN: 1474-760X

Contact us