Dissection of a Cell Atlas of Developing Somatic Embryos in Cotton by Single-cell RNA-sequencing and Spatial Transcriptomic and Metabolomic Analyses

doi:10.21203/rs.3.rs-2100867/v1

Download PDF

Research Article

Dissection of a Cell Atlas of Developing Somatic Embryos in Cotton by Single-cell RNA-sequencing and Spatial Transcriptomic and Metabolomic Analyses

https://doi.org/10.21203/rs.3.rs-2100867/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this older preprint version

Read the latest preprint version →

Somatic embryogenesis is a major pathway for the regeneration of cotton in tissue culture, and a major technique used for cotton breeding and improvement. The signaling pathway regulating the process of somatic embryogenesis in cotton are extremely complex and still poorly characterized. Little is known about the mechanisms regulating the changes that occur in gene expression and metabolites during somatic embryogenesis at a single-cell resolution. Here, we investigated the spatial and single-cell expression profiles of key genes and metabolic patterns of key metabolites by integrated single-cell RNA-sequencing, spatial transcriptomics, and spatial metabolomics. To evaluate the results of the above analyses, we characterized the potential roles of two representative marker genes AATP1 and DOX2 in the regulation of somatic embryo development. A publicly accessible web-based resource database (http://cottonsomaticembryo.cricaas.com.cn) was generated to facilitate future studies designed to analyze the expression patterns of marker genes at specific developmental stages.

Plant Molecular Biology and Genetics

Cotton

Cell atlas

Development

Single-cell

Spatial multi-omics

Somatic embryo

Cotton (Gossypium spp.) is one of the most important cash crops in the world ¹. However, genetic transformation of cotton relies on a complex, time-consuming and genotype-dependent process of somatic embryo regeneration ². Somatic embryogenesis (SE) is a developmental process where somatic cells dedifferentiate to totipotent embryonic stem cells (often via the formation of undifferentiated callus) that can give rise to somatic embryos under appropriate conditions ³. SE is the most studied plant regeneration process, but to date it is the least understood because the lack of information about the reprogramming of gene expression patterns involved in somatic embryo formation at the cellular level ⁴. Somatic embryogenesis in cotton is based on the process of indirect embryogenesis ⁵, in which cells of the hypocotyls or other types of explants are cultivated under in vitro conditions to generate undifferentiated callus tissue, and then callus cells are induced to form embryonic cells with the capacity to develop into an initial somatic embryo, called a globular embryo ⁶. Subsequently, globular embryos differentiate to form torpedo embryos in an orderly developmental process that includes the formation of well-developed vascular bundles, which are essential to maintain the shape of torpedo embryos ⁷. Cotyledons develop at the top of the torpedo embryo to form a cotyledonary embryo ⁸. On suitable medium, cotyledonary embryos develop to a mature stage capable of forming roots and buds ⁹. These materials are suitable for transplanting to develop into regenerated plants, thus completing the process of in vitro regeneration ¹⁰. Unlike many other plants, in vitro regeneration in cotton is still an extremely difficult and complex process ². A large number of genes are differentially expressed and interact with each other at different stages and tissues to accurately regulate the development of somatic embryos in cotton ¹¹. Several studies have been conducted to analysis global transcription changes during somatic embryogenesis in plants, including maize (Zea mays), Woody Species, Arabidopsis thaliana, soybean (Glycine max), etc. ^12-15. However, few studies have dissected this important developmental process at the single-cell level.

Single-cell RNA-sequencing (scRNA-seq) is an innovative technology that can be used to gain a deeper insight into the complex process of somatic embryo development at a single-cell resolution. Studies have been performed in Arabidopsis thaliana, rice (Oryza sativa), maize, as well as many other species ^16-20. The emerging spatial transcriptomics technology (ST) has also been used to further visualize gene expression patterns obtained from scRNA-seq data ²¹.

Recently, spatial metabolomics (SM) utilizes mass spectrometry imaging (MSI) technology to obtain information on the presence of a large number of known and unknown metabolites and their spatial distribution directly from biological tissues to create a metabolic imaging maps of the tissues ²². SM has been used to track metabolite movement caused by nematode infection in tomato (Lycopersicon esculentum) roots ²³, and to document the distribution of ginsenosides in Panax ginseng ²⁴. In this study, we used a newly-developed air flow-assisted desorption electrospray ionization mass spectrometry imaging (AFADESI-MSI) ²⁵ to examine the dynamic metabolic changes during the development of cotton somatic embryos.

A previous study reported that polyamines and H₂O₂ play a key role in the transformation of cotton embryogenic callus into somatic embryos ²⁶. Hormones, such as zeatin - a type of cytokinins ²⁷, are also important components for inducing direct somatic embryogenesis in cotton ²⁸. Therefore, in this study, a multi-omics analysis strategy was used to perform a detailed transcriptomic and metabolomic characterization of the four key stages of somatic embryo development in cotton (callus, globular embryo, torpedo embryo, and cotyledonary embryo) based on scRNA-seq, ST and SM analysis. The key genes and metabolites regulating somatic embryo development were screened and identified. Furthermore, we integrated the analysis of key metabolites, such as polyamines and hormones, and the analysis of genes related to the synthesis of these metabolites to achieve a multiplex analysis of gene-metabolite correlations.

Dynamics of spatial gene expression during the development of somatic embryos in cotton

We developed a research approach using ST to comprehensively assess the spatial expression patterns of genes during somatic embryo development in cotton (Fig. 1a). ST was used to conduct a comprehensive transcriptome analysis of different tissues of cotton across all the stages of somatic embryogenesis. First, frozen tissue sections of globular embryos, torpedo embryos, and cotyledonary embryos were made (Fig. 2a-c). The three stages of sampled embryos possessed three sections, and each spot on each section corresponds to 1-10 cells (Supplementary information, Fig. S1d). After being subjected to quantitative quality control, the number of high-quality spots was 801-1,217 per section, the average number of unique molecular identifiers (nUMIs) per spot was 11,857-16,401, the average number of expressed genes in each spot was 6,287-8,001 (Supplementary information, Fig. S1f-h, Table S1).

We used t-distributed Stochastic Neighbor Embedding (t-SNE) to divide the data into 13 cell clusters, and then mapped the spot cluster back to its original position in the tissue section (Fig. 2d and 2e). The 13 clusters were named based on their spatial location. We also constructed a heatmap of the 10 highest expressed genes in each cluster (Fig. 2f). For example, Gh_D05G159100, which is highly expressed in ‘pro-embryogenic cell’ (cluster 6), is a homologous gene of WUSCHEL-related homeobox 9 (WOX9) that plays a role in embryogenesis and meristems in plants ²⁹. In ‘cotyledon cell’ (cluster 9), Gh_A07G040500 displays specific expression pattern and shows high homology to axial regulator YABBY 5 (YAB5) ³⁰. In ‘cortex cell_12’ (cluster 12), Gh_D08G133100 was identified, which is homologous to YUCCA10 (YUC10) encoding a flavin monooxygenase that has been associated with embryogenesis and auxin synthesis in Arabidopsis ³¹.

The differentially expressed genes (DEGs) were identified in the 13 clusters (Supplementary information, Table S4). Gene ontology (GO) enrichment of the DEGs in the different cell clusters was conducted (Fig. 2g). Genes associated with the ‘regulation of post-embryonic development’ were highly expressed in ‘cotyledon cell’ (cluster 9), while genes of ‘defense response to other organism’ were found to be enriched in ‘cortex cell_11’ (cluster 11) (Fig. 2g). DOX2/Gh_A09G193100 and WOX9/Gh_D05G159100 were highly expressed in ‘pro-embryogenic cell’ (cluster 6), and this cluster of cells is involved in ‘shoot meristem development’ (Fig. 2f). These results suggest that genes involved in different GO functions may play a specific physiological function during the different stages of somatic embryogenesis.

Analysis of the spatial expression of genes regulating the initial development of somatic embryos in cotton

The initial induction of callus cells and the formation of embryogenic cells are the most critical stages for the development of cotton somatic embryos ³². scRNA-seq had been used to study different cell types of Arabidopsis callus ³³. Therefore we also conducted a ST analysis of the callus induced from cotton hypocotyl (Fig. 3). Six cell clusters were obtained (Fig. 3a). It was found that the homologue of vascular tissue marker gene aintegumenta (ANT) (Gh_A11G229400) and the homologue of stem cell formation-related gene of wuschel-related homeobox 13 (WOX13) (Gh_A02G209400) were expressed in cluster 1 ^{34, 35}, suggesting that cluster 1 contained vascular tissue cells from explants and callus founder cells. Combined with previous identification of callus cell types ³³, cluster 1 was identified as ‘explant vasculature and callus founder cells’. Similarly, the homologue of procambium developmental marker gene homeobox-leucine zipper protein ATHB-15 (ATHB-15) (Gh_D05G130700) was expressed in the ‘explant vasculature and callus founder cells _3’ (cluster 3) ³⁶, and the homologue of vascular cambium developmental marker genes wuschel related homeobox 4 (WOX4) (Gh_A05G188600) and the homologue of WOX13 were expressed in the ‘explant vasculature and callus founder cells _6’ (cluster 6) ^{34, 37}. The epidermal cell marker gene ethylene responsive element binding factor 4 (ERF4) (Gh_A10G019700) is expressed in the ‘outer cell layer of the explant’ (cluster 2) ³⁸. The homologue of embryonic root apical meristem marker gene like auxin resistant 2 (LAX2) (Gh_A01G254000) was expressed in the ‘middle cell layer of callus cells’ (cluster 4). The homologue of WOX4 was expressed in the ‘Inner cell layer of callus cells’ (cluster 5) ³⁷. The expression of marker genes for each cell cluster is presented in Fig. 3b. Further GO analyses indicated that the majority of DEGs in the cell clusters of ‘explant vasculature and callus founder cells_3’ and ‘inner cell layer of callus cells’ (clusters 3 and 5) are involved in the ‘response to hypoxia’ (Fig. 3d).

Analysis of the expression patterns of genes in multiple developmental stages of cotton somatic embryos

To further analyze gene expression in relation to tissue development in somatic embryos of cotton, we divided all the samples from the three developmental stages into 9 sections and determined the correlation of gene expression among the samples by Pearson correlation analysis (Supplementary information, Fig. S1c).

We then analyzed the GO enrichment terms and spatial expression of DEGs in different cell clusters within the nine sections, respectively (Supplementary information, Table S5-Table S13).

In section 1, the ‘non-embryogenic callus cell’ (cluster 1) contains the homologue of vascular tissue marker gene WOX13 ³⁹. ‘Globular embryo cell’ (cluster 2) expressed the homologue of early embryonic vascular marker gene homeobox gene 8 (ATHB-8) (Gh_D05G049400) and the homologue of marker gene baby boom (BBM) (Gh_D08G251600) that induces somatic embryogenesis ^{40, 41} (Supplementary information, Table S5). In section 2, ‘pro-embryogenic cells_3’ (cluster 3) expressed the homologue of embryonic epidermal cell marker gene protodermal factor 1 (PDF1) (Gh_D04G040400) and the homologous gene of WOX9 for embryonic pattern formation ⁴², and ‘pro-embryogenic cells_4’ (cluster 4) expressed the homologous marker gene of AP2-like ethylene-responsive transcription factor 6 (AIL6) (Gh_D10G104500) for somatic embryogenesis ⁴³, ‘pro-embryogenic cells_5’ (cluster 5) expressed the homologous marker gene of BBM associated with somatic embryogenesis ⁴⁰. The homologous marker gene of auxin response factor 5 (ARF5) (Gh_D05G188200) and ANT for embryonic vascular tissue formation were found in ‘globular embryo cell_1’ (cluster 1) ^{35, 44}, ‘globular embryo cell_2’ (cluster 2) expressed the homologous marker gene of floral homeotic protein apetala 2 (AP2) (Gh_A11G207700) for somatic embryogenesis ⁴⁵, and ‘globular embryo cell_6’ (cluster 6) also expressed the homologous marker gene of WOX9 and WOX13 ^{29, 39} (Supplementary information, Table S6). GO enrichment analysis for the DEGs in sections 3 to 9 were conducted and the results are presented in Supplementary information, Fig. S4 to Fig. S10.

Expression Analysis of Genes Regulating the Cambium Development of Somatic Embryos

The spatial division and differentiation of plant stem cells are the basis of plant growth and development ⁴⁶. The differentiation of embryogenic cells determines the beginning of embryonic development ⁴⁷. During embryonic development and maturation, differentiation starts from the shoot and root meristems at the base of the embryo, as well as in the cambium ^{48, 49}. The differentiation of cambial cells is essential for the formation of the cortex and the development of vascular tissue. Many well-known TF-encoding genes were highly expressed in these cambial clusters, including revoluta (REV) (Gh_D13G262600), short root (SHR) (Gh_D03G030200), MP/ARF5, and ATHB-15, which have been reported to be key genes associated with cambium differentiation and vascular development in Arabidopsis and rice ^50-53. Many new TFs in the HB-HD-ZIP gene family, such as homeobox-leucine zipper protein HAT2 (HAT2) (Gh_A08G191700), ATHB-13 (Gh_D02G129400), ATHB-20 (Gh_D09G022200), ATHB-6 (Gh_A02G158400) were also identified in the cambial clusters. In addition, the TFs newly discovered in the cambium, cotyledon primordium, and hypophysis are listed in Supplementary information, Table S15 and Table S16.

ScRNA-seq Analysis of Gene Expression During Somatic Embryo Development in Cotton

To distinguish the transcriptional heterogeneity of individual cells, we conducted scRNA-seq on protoplasts isolated from the samples used in the ST. Filtering of raw data was conducted using the linear model fitting curve and the violin map of the number of genes (nGene) per cell and the nUMIs (Supplementary information, Fig. S11). After quality control and screening, 8,842, 9,031 and 9,037 single-cell transcripts were identified in globular embryos, torpedo embryos and cotyledonary embryos, with an average of 2,339, 1,992 and 2,146 genes per cell, respectively (Supplementary information, Table S2).

A total of 11 cell clusters were identified after the t-SNE analysis (Fig. 4a). Heatmap of the expression of several highly expressed marker genes for each cell cluster was then constructed (Fig. 4c). A correlation analysis revealed that the scRNA-seq and ST data were strongly correlated (Supplementary information, Fig. S11e-g). The cell types were then annotated with several marker genes identified in the ST results. DEGs were subsequently identified in each cluster (Supplementary information, Table S14). For example, Gh_A05G087300, which was expressed in cluster 2, is a homologous gene of mitogen-activated protein kinase 3 (MPK3) that was shown to play an important role in epidermal cells and embryonic development in Arabidopsis ⁵⁴. Thus, cluster 2 was classified as ‘epidermal cells_2’. Genes with homology to auxin22 (AUX22) (Gh_D05G125900) and cytochrome P450 87A3 (CYP87A3) (Gh_Contig00374G000600) were highly expressed in ‘ground tissue cells_3’ (clusters 3) and ‘mesophyll cells’ (cluster 5), respectively (Fig. 4b). AUX22 encodes a TF that may be involved in the processes underlying the progression from embryogenic callus to somatic embryos morphogenesis ⁵⁵. CYP87A3 is an auxin-induced gene originally isolated from coleoptiles in the early stage of rice seedling growth ⁵⁶. Bcl-2-associated athanogene family molecular chaperone regulator 5 (BAG5) (Gh_D02G193400) was highly expressed in clusters 7 (Fig. 4b). In ‘apical cortex cells_7’ (cluster 7), Gh_D02G193400 plays a role in leaf development in cotton as its homolog BAG5 in Arabidopsis ⁵⁷. Therefore, based on the combined functions and location of the above genes, we annotated cluster 7 as ‘apical cortex cells_7’. Gh_A08G230200 is a homologous gene of LOB domain-containing protein 12 (LBD12) that is associated with the formation of shoot apical meristem in rice ⁵⁸. LBD12 is expressed in ‘apical cortex cells-1’, ‘epidermal cells-9’ and ‘cortex cells’ (clusters 1, 9 and 10).

Different with ST analysis, scRNA-seq allows studies at the high-resolution level of single cells ^{59, 60}. In this study, many similarities were found in the two datasets, such as common cell types like epidermal cells, vascular cells and cortex cells were identified by both ST and scRNA-seq; common gene expression, such as LBD12 is expressed in ‘cortex cell_12’ (cluster 12) in the ST, consistently, LBD12 also expressed in 'apical cortex cells-1', 'epidermal cells-9' and 'cortex cells' (clusters 1, 9 and 10) in the scRNA-seq data (Supplementary information, Fig. S12i, Table S14). Thus, we can combine the complementarity of ST and scRNA-seq to perform spatial and high-resolution studies of cell types in cotton somatic embryos.

Next we constructed a developmental trajectory for cell differentiation using a pseudo-time analysis (Fig. 5a and b). Different cell clusters were apparent at different branching positions on the pseudo-time trajectory (Fig. 5b), which can be used to make an initial determination of the relationship between these cells at different developmental stages ⁶¹. For example, ‘epidermal cells_2’ (cluster 2) and ‘ground tissue cells’ (cluster 3) exhibited relatively similar distribution patterns on their developmental trajectories, indicating that their developmental stages were similar (Fig. 5b). Then, we analyzed genes enriched at different stages of pseudo-time for different single cell clusters (Fig. 5c). The expression of these genes can be divided into three categories. In the first type, the expression of marker genes from ‘epidermal cells_11’ (cluster 11), ‘epidermal cells_9’ (cluster 9), ‘epidermal cells_8’ (cluster 8) and ‘cortex cells’ (cluster 10) was at their highest level in the early stage of pseudo-time, and then decreased gradually. The second type comprises marker genes from ‘vascular tissue cells’ (cluster 4), ‘mesophyll cells’ (cluster 5), and ‘epidermal cells_2’ (cluster 2), whose expression increased gradually with pseudo-time, and reached their highest level in late pseudo-time. The third type was comprised of DEGs from ‘apical cortex cells_1’ (cluster 1), ‘ground tissue cells’ (cluster 3), ‘epidermal cells_11’ (cluster 11), whose expression level was the highest in the middle stage of pseudo-time, and then gradually decreased.

Establishment of a Gene Expression Map for Somatic Embryo Development in Cotton

As previously mentioned, the initial onset of embryogenesis is a critical stage that transforms non-embryogenic callus cells into embryogenic cells. In the ST clusters, we found some highly expressed genes in ‘non-embryogenic callus’ (cluster 13), ‘pro-embryogenic cell’ (cluster 6) and ‘globular embryo cell_10’ (cluster 10) (Supplementary information, Fig. S12). More specifically, beta-D-xylosidase 1 (BXL1) (Gh_A09G261400) was specifically expressed in ‘non-embryogenic callus’ (cluster 13), and further ISH analysis indicated that BXL1 was primarily expressed in the cortex of globular embryos, but also in the epidermis of both torpedo and cotyledonary embryos (Supplementary information, Fig. S13a-f). Gh_A01G086600 of ‘pro-embryogenic cell’(cluster 6) is a homologous gene of indole-3-acetate O-methyltransferase 1 (IAMT1) that was reported to be involved in auxin metabolism in Arabidopsis ⁶², suggesting that Gh_A01G086600 may regulate the initial formation of globular embryos through auxin-related metabolic processes. Similarly, Gh_A12G174900 of ‘pro-embryogenic cell’ (cluster 6) is a homologous gene of receptor-like transmembrane kinase 1 (TMK1) that is involved in auxin signal transduction in Arabidopsis ⁶³ (Supplementary information, Fig. S12l), suggesting that Gh_A12G174900 may function in regulating the formation of globular embryos by participating in auxin signal transduction. Additionally, ISH analysis indicated that Gh_D11G232500/ABC transporter G family member 11 (ABCG11) was primarily expressed in the endodermis of globular and torpedo embryos, and mainly in the epidermis of cotyledonary embryos (Fig.s 6b-d and Supplementary information, Fig. S12g). In this regard, the Gh_D11G232500/ABCG11 reportedly encodes a transporter for epidermal lipid transport in Arabidopsis ⁶⁴.

The establishment and maintenance of embryonic epidermis is crucial for the normal development of embryos. For example, the epidermis of shoot organs of higher plants originates from the embryonic epidermis, and the epidermis also protects the embryo from biotic and abiotic stresses ^{65, 66}. Then we next analyzed the key genes expressed in ‘epidermal cell’ (cluster 3), ‘cortex cell_1’ and ‘cortex cell_12’ (clusters 1 and 12) of ST to identify genes that may regulate the development of cortical and epidermal cells (Fig. 2f). Cysteine-rich receptor-like protein kinase 29 (CRK29) belongs to a class of receptor-like kinases that encode cysteine-rich protein kinases ⁶⁷. In cotton, Gh_D10G017200/CRK29 was expressed in ‘epidermal cell’ (cluster 3) and the epidermis of early globular and torpedo embryos (Supplementary information, Fig. S12d). Additionally, ISH analysis indicated that Gh_A13G249900, a homolog of histidine-containing phosphotransmitter 3 (AHP3), was first expressed in the cortex of globular embryos, and was still distributed in the cortex of torpedo embryos and cotyledonary embryos (Supplementary information, Fig. S14g-m). AHP3 encodes a histidine phosphate transfer protein involved in the cytokinin signaling pathway in Arabidopsis ⁶⁸. ISH analysis indicated that LBD12 was primarily expressed in the cortex of globular embryos, torpedo embryos, and cotyledonary embryos (Supplementary information, Fig. S14n-s), which further verifies the reliability of the cell cluster classification.

Gh_A03G031900 is a homologous gene of heavy metal-associated isoprenylated plant protein 39 (HIPP39) that encodes a heavy metal-associated isoprenylated plant protein in rice ⁶⁹. This gene is specifically expressed in the embryonic cortex of cotyledonary embryos (Supplementary information, Fig. S12e).

AATP1 and DOX2 negatively regulate cotton somatic embryogenesis

Based on the analysis of ST, we identified that AAA-ATPASE1 (AATP1) was expressed in ‘globular embryo_10’, while DOX2 was expressed in ‘pro-embryogenic cells’ (cluster 6) (Fig. 2f), both cell clusters are key cell types associated with cotton somatic embryo regeneration ⁷⁰. Gh_D09G025900 has homology to AATP1 that has been demonstrated to play a role in inducing systemic signaling to increased resistance to pathogens in potato (Solanum tuberosum) tubers ⁷¹. DOX2 encodes an α-dioxygenase, which was reported to participate in α-linoleic acid metabolism and play a role in protecting tissues from oxidative damage and cell death in tomato ⁷². Given that AATP1 and DOX2 have been found to be highly expressed during the early stages of cotton somatic embryogenesis, we generated three overexpression (OE) and three DNA-edited lines (CRISPR/Cas, CAS) of AATP1 (Gh_D09G025900) and DOX2 (Gh_A09G193100) to determine their potential roles in cotton somatic embryogenesis (Supplementary information, Fig. S15). AATP1-CAS and DOX2-CAS transgenic calli (Cs) proliferated faster and had a larger volume at the end of the first 30 days of calli induction than wild-type (WT) calli (Supplementary information, Fig. S15a). By contrast, the calli proliferation rate of AATP1-CAS and DOX2-CAS transgenics, was lower than the WT control; and thus, their Cs volume was also lower than that of WT (Supplementary information, Fig. S15a). AATP1-CAS and DOX2-CAS Cs at the end of 90 days were friable and could be easily recognized as embryogenic. In contrast, AATP1-OE and DOX2-OE Cs proliferated more slowly during the initial stage of callus induction than the WT control and CAS Cs, and their callus volume was also relatively reduced. Consistently, the Cs proliferation rates in AATP1-OE and DOX2-OE Cs increased by the end of 60 days; however, no significant difference in Cs volume and weight, relative to the control, was observed (Supplementary information, Fig. S15). After 90 days of calli induction, the Cs of AATP1-OE and DOX2-OE transgenics were hard and had no signs of developing into embryogenic callus (ECs). The expression levels of AATP1 and DOX2 in overexpressing lines were found to be significantly increased compared with that of WT (Supplementary information, Fig. S15b). The Cs of AATP1-CAS and DOX2-CAS exhibited a higher induction and differentiation rates than that of WT control (Supplementary information, Fig. S15c and Fig. S15d).

Compared to AATP1-OE, DOX2-OE and control callus tissues at 30d, the cell morphology in AATP1-CAS and DOX2-CAS callus tissues was mainly round or oval, indicating that callus cells had begun to transform into embryogenic callus cells. In contrast, the cells of OE calli mainly exhibited an elongated shape, indicating that these cells were in a non-embryogenic cell state (Supplementary information, Fig. S15e). These results suggest that AATP1 and DOX2 negatively regulate callus initiation and embryogenic callus formation.

Mass Spectrometric Imaging Analysis of Metabolites During Somatic Embryo Development in Cotton

We next used AFADESI-MSI to analyze the metabolites in cotton embryo samples (Supplementary information, Fig. S16). We determined the mass charges and obtained the molecular contours in the tissue sections (Fig. 7a, Supplementary information, Fig. S16c-e). We divided the tissue sections to identify the marker metabolites related to somatic embryo development and made four groups of comparisons: (1) cortex vs. cotyledon, (2) cortex vs. vascular, (3) cotyledon vs. vascular, and (4) hypocotyl callus vs. hypocotyl.

We compared the metabolite information obtained in the different regions using unsupervised PCA (Supplementary information, Fig. S16f). We further used orthogonal projections to latent structures-discriminant analysis (OPLS-DA) to distinguish the overall differences in metabolites among the groups to identify the DPMs in the groups (Supplementary information, Fig. S16g). The metabolic pathway matching analysis of these selected DPMs was carried out based on the exact mass number, SmetDB database, and KEGG database (Fig. 7b). Among the DPMs detected during somatic embryo development (Supplementary information, Table S19 and Table S20), 74 metabolites were assigned to 26 metabolic pathways that are known to be involved in plant development (Fig. 7b). For example, ‘histidine metabolism’ has been reported to be necessary for embryonic development ⁷³, ‘pyrimidine and purine metabolism’ provides a sufficient level of purines and pyrimidines to support nucleotide biosynthesis during embryonic development ⁷⁴. In ‘arginine and proline metabolism’, arginine acts as a precursor for proline synthesis, and also plays a role in the synthesis of polyamines, which are essential for embryogenesis and organogenesis ^{75, 76} (Fig. 7b).

We compared the differences in annotated DPMs between cortex vs. cotyledon in group 1 (Supplementary information, Fig. S17a). These metabolites showed significant differences in abundance in different tissues (Supplementary information, Fig. S17c). For example, dihydrozeatin (C02029) and zeatin (C00371) were also highly abundant in cotyledons. Linoleic acid (C01595), (S)-β-aminoisobutyrate (C03284), and D-malic acid (C00497) were highly expressed in the cortex (Supplementary information, Table S19). In particular, the abundance of linoleic acid was particularly high in the cortex (Supplementary information, Fig. S17d).

We also compared the abundance of DPMs in the ‘cortex vs. vascular tissues” (Supplementary information, Fig. S18a). These metabolites participate in several metabolic pathways (Supplementary information, Fig. S18b). Pyroglutamic acid (C01879) is a cyclized derivative of L-glutamate and participates in ‘glutathione metabolism’, which in turn is a key component of the plant stress response network ⁷⁷. In the cortex, Succinic acid participates in many metabolic pathways, including the ‘citrate cycle (TCA cycle)’ and ‘butanoate metabolism’ (Supplementary information, Fig. S18a). The ‘TCA cycle’ and ‘butanoate metabolism’ are essential metabolic pathways in plants ^{78, 79}.

In the comparative analysis of DPMs between cotyledon and vascular tissues (Supplementary information, Fig. S19a), N-carbamoylputrescine (C00436), L-arginine (C00062) and indole-3-acetaldehyde (C00637) were highly abundant in cotyledons (Supplementary information, Table S19). The dihydrozeatin (C02029), zeatin (C00371) aND DTDP (C00363) were more abundant in hypocotyls (Supplementary information, Fig. S20a). These metabolites are involved in ‘zeatin biosynthesis’ and ‘purine metabolism’ (Supplementary information, Fig. S20b). On the other hand, L-aspartic acid (C00049), L-histidine (C00135) and galabiose (C00760) were more abundant in callus tissues (Supplementary information, Fig. S20d, Table S19). L-aspartic acid, L-histidine are utilized in ‘histidine metabolism’, which is essential for plant development, including embryonic development ⁷³.

Gene-Metabolite Joint Analysis During Somatic Embryo Development in Cotton

Next, a joint analysis of SM and ST data was conducted. Based on the ST data obtained on somatic embryos, a heatmap analysis of the expression of genes involved in the metabolic pathways of identified DPMs was conducted (Supplementary information, Fig. S21, Table S21). Results indicated that most genes related to the synthesis of polyamines (e.g. spermine and spermidine), including spermidine synthase 1 (SPDSYN1) (Gh_A03G164400) and spermine synthase (SPMS) (Gh_A08G106900), were highly expressed in globular embryos, which is consistent with higher abundance of these polyamines (Supplementary information, Fig. S21). And zeatin synthesis like CKX5/Gh_D06G005100 were highly expressed in cotyledonary embryos (Supplementary information, Table S21). The consistency of ST and SM data suggests that spermine and spermidine may represent marker metabolites of globular embryo development, while zeatin may be a marker metabolite of cotyledonary embryo development.

Further, KEGG analysis indicated that DOX2 participates in ‘alpha-linolenic acid metabolism’ as α-dioxygenase and plays a role in the generation of 2(R)-HPOT from alpha-linolenic acid (Supplementary information, Fig. S22a), which indicates that DOX2 acts as a negative regulator in the synthesis of alpha-linolenic acid. Combined with the SM data, we found that the accumulation of alpha-linolenic acid was elevated at the torpedo embryo stage (Supplementary information, Fig. S22b). While alpha-linolenic acid is a type of fatty acids (FA), studies have found that the reduction of FA synthesis may be one of the reasons for embryonic developmental defects ^{80, 81}. This suggests that DOX2 may affect cotton somatic embryogenesis by negatively regulating the synthesis of alpha-linolenic acid, thereby acting as a negative regulator.

Progress has been made in the study of cotton somatic embryogenesis in recent years ^{82, 83}. ST and scRNA-seq have been collectively used to study the process of mammalian cell regeneration, such as mouse gastrulation, in vitro ⁸⁴. We utilized ST, scRNA-seq, and SM methodologies to explore changes in the transcriptome and metabolome of specific tissues and cell types during the entire course of somatic embryo development. We obtained and annotated cell types and marker genes for somatic embryos at different developmental stages. Based on the spatial information on some representative marker genes expression in specific tissues and cells, we also identified the specific cell types produced by scRNA-seq, which overcomes the shortcoming of ST at the single cell resolution and helps us to analyze the cell heterogeneity within the same cell type. The expression feature of some representative marker genes in specific cell types was verified by ISH (Figures S13 and S14). These newly identified marker genes can be used as references in future studies.

Since the metabolites also function in feedback regulation of gene expression ^{85, 86}. Therefore, we utilized SM to analyze and identify DPMs in samples from different stages of somatic embryos (Fig. 7b and Supplementary information, Table S19). We found that linoleic acid and pyroglutamic acid were highly abundant in the cortex (Supplementary information, Table S19). Linoleic acid is involved in ‘linolenic acid metabolism’, a type of unsaturated fatty acid that is related to stress defense in plants ⁸⁷. These results together suggest that linoleic acid may contribute to the stress defense function of cortex ⁸⁸. Through in situ imaging of metabolites in embryonic tissues, we have found many key DCMs that may regulate the development of somatic embryos, which will bring a new perspective to the study of cotton somatic embryo regeneration from the perspective of metabolome.

We also conducted a joint network analysis of ST and SM data (Supplementary information, Fig. S21, Table S21). Among the examined metabolites, polyamines have been shown to play a role in the formation of somatic embryos ²⁶. Based on the data of both ST and SM, we found that DOX2 and AATP1 might be involved in the regulation of the initiation of cotton somatic embryos. And combined with the SM data, we found that DOX2 and alpha-linolenic acid were associated with normal embryonic development (Supplementary information, Fig. S22b) ^{80, 81}. Further functional analysis confirmed the potential roles of AATP1and DOX2 in the regulation of callus induction and differentiation and the formation of embryos (Supplementary information, Fig. S15). Finally, based on our ST and scRNA-seq data, we constructed a website (http://cottonsomaticembryo.cricaas.com.cn/) on gene spatial expression in cotton somatic embryos during their development (Supplementary information, Fig. S23). Taken together, by combining ST, scRNA-seq and SM technologies, we revealed the spatial gene expression map and key metabolites of the somatic embryos of cotton. These results demonstrated that the multi-omic technology will bring a more efficient strategy for solving the problems of cotton regeneration in vitro and enable us to analyze the process of plant development.

Spatial Transcriptomics

1. Section preparation and Hematoxylin -Eosin (H&E) staining

Take fresh materials (globular embryos, torpedo embryos, cotyledonary embryos, embryogenic callus, and non-embryogenic callus) were embeded directly with optimum cutting temperature (OCT) Compound. The embedded samples in a 1.5 ml centrifuge tube were incubated at −20℃ in the slicer for 20 min. RNA was extracted and RNA quality was assessed. Qualified samples with RIN > 7 were selected for HE staining. The slides were incubated at 37℃ for 5 min and then scanned for imaging.

2. Permeation of tissue slices

The scanning images were used to confirm that the section included the targeted region. Then, cotton globular embryos, torpedo embryos, cotyledonary embryos, and callus sections were permeated. Visium slides were placed in a Cassette, and 70 µl permease (2% cellulase R10, 0.4% Macerozyme R10, 1% pectinase, 1% hemicellulase, 0.4% snail enzyme) was added to each sample well. The Visium slides were patted to ensure that the permease covered the tissue evenly. The mRNA in cells was released and bound to the corresponding capture probe. After completion of the time allowed for permeation, the Visium Cassette was removed from the adapter of the PCR instrument along with the sealing strip.

3. Reverse transcription and library construction

After permeation was completed, the captured RNA was used as a template for reverse transcription of the sample to form cDNA. Then, 10 µl cDNA of each sample was used to perform a PCR reaction that included Adaptor Ligation on the prepared library.

4. Quality control of sequencing data and gene quantification

After sequencing, the raw reads generated in high-throughput sequencing were in fastq format. Official 10× Genomics Space Ranger software was used to process the Visium Spatial transcriptome sequencing data and bright field microscope section images, detect the captured tissue area in the chip, and distinguish reads from each spot based on spatial barcode information. The total number of spots, the number of reads in each spot, the number of detected genes, and the number of UMIs (unique molecular identifiers) were tabulated to evaluate the quality of the sequencing reads. STAR software was integrated with the cotton reference genome for comparative analysis, and a gene-spot matrix was generated for gene expression analysis.

5. Quantitative gene quality control and data preprocessing

The Seurat software package was used to conduct further quality control and processing of the obtained sample data based on the preliminary quality control results using Space Ranger. Sctransform was used to normalize the data, construct a regularized negative binomial model of gene expression, detect high variance features, and store the data in SCT for further analysis.

6. Dimensionality reduction and clustering analysis

PCA (principal component analysis) linear dimensionality reduction was performed using gene expression levels, and PCA (principal component analysis) results were visualized in two-dimensional space using tSNE (nonlinear dimensionality reduction) software. Dimensionality reduction and clustering analysis, enabled the data to be visualized and determine which genes were expressed, how much they were expressed, and the spatial location of the genes.

7. Differential gene and enrichment analysis

The Seurat software package was used to identify differential gene expression, and significant genes were identified based on a P-value < 0.05 and a multiple of difference greater than 2 (or 1.5). Gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses of differentially-expressed genes was performed using a hypergeometric distribution test.

Single cell transcriptomics

1. Preparation of single cell suspensions of cotton somatic embryos

We established a reagent formula and treatment protocol suitable for the dissociation of cotton protoplasts. The major improvements in reagents included a cracking fluid formulation (20 mM MES (pH 5.7), 3% (wt/vol) Cellulase R10, 1% (wt/vol) Macerozyme R10, 1% (wt/vol) Snailase, 1% (wt/vol) Cutinase, 0.4 M Mannitol, and 20 mM KCl) and improved treatment conditions (dark condition, 28℃, 20 RPM horizontal shaking). The samples were then soaked in the enzymolysis solution, and vacuum infiltrated for 15 minutes to help the enzymolysis solution thoroughly enter the tissue samples. The dissociated cells were filtered through a cell sieve to remove large tissue blocks and retain free cells.

2. Establishment of cDNA library

The scRNA-seq library was prepared using a Chromium Single Cell 3’ Gel Beads-in-emulsion (GEM) Library & Gel Bead Kit v3 according to the manufacturer’s instructions (10× Genomics, California, USA).

3. ScRNA-seq data preprocessing

The Cell Ranger pipeline (version 3.0.0) provided by 10× Genomics was used to demultiplex cellular barcodes and map reads to the cotton reference genome. Transcript quantification was determined via the STAR aligner. The unique molecular identifier (UMI) count matrix was processed using the R package Seurat (Version 2.3.4). To remove low-quality cells and multiple captures, we applied further criteria to filter out cells with UMI/gene numbers outside the limit of the mean value +/-2 standard deviations, assuming a Gaussian distribution of each cell's UMI/gene number. After the critical filtering process, the cells were retained for downstream analysis. Library size normalization was performed in Seurat on the filtered matrix to obtain normalized counts.

4. Dimensionality reduction and clustering analysis

Mutual nearest neighbors (MNN) and t-distributed Stochastic Neighbor Embedding (t-SNE) are the dimensionality reduction algorithms used for the single cell transcriptome. Based on the dimensionality reduction results of MNN, t-SNE was used to visualize single-cell clusters. Principal Components Analysis (PCA) and t-SNE algorithms were used for dimensionality reduction of the single cell transcriptome. Based on the dimensionality reduction results of PCA, t-SNE was used to visualize single-cell clusters. The SNN clustering algorithm was then used to obtain the optimal cell population.

5. Similarity analysis between cell populations

The Pearson correlation coefficient of average gene expression between cell populations was calculated to assess the degree of similarity between two cell populations.

6. Marker gene identification

The Bimod test was used to assess the difference between the designated cell population and all other cell populations, and to identify specific Marker genes for each cell population. VlnPlot and FeaturePlot functions were used to visualize the identified Marker genes.

7. Differential gene and enrichment analysis

The MAST difference test was used to identify differential gene expression, utilizing a P-value < 0.05 and a multiple of difference greater than 2 (or 1.5). Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis of significant genes was performed using a hypergeometric distribution test.

In situ hybridization

1.Preparation of materials

First, the tissues of cotton somatic embryos were placed into a glass bottle containing 4–5 ml freshly prepared 4% paraformaldehyde fixative (1× PBS 100 ml, 1M NaOH 0.5 ml, 4 g paraformaldehyde, pH 7.0). The fixative was then replaced and incubated overnight at 4℃ (not more than 16 h). The tissue samples were then dehydrated in a graded series of ethanol and xylene. Finally, the samples were embedded in paraffin using premade forms and stored at 4℃ until further use.

2. Tissue section

The embedded samples were sectioned on a microtome into continuous slices of 5 µm thickness.

3. In situ hybridization

The in situ hybridization was performed with the ISH-Kit with customized probes according to the manufacturer’s instructions (BOSTER BIOLOGICAL CHNOLOGY).

Cotton tissue culture

1. Construction of the CRISPR/Cas9 gene editing vectors and overexpression vectors

The basic vector pCambia2300 was used as the original vector for the construction of overexpression vectors. Gene was amplified and was driven by the cauliflower mosaic virus (CaMV) 35S promoter or other promoters. The neomycin phosphotransferase II gene NPT II that confers kanamycin resistance was used to select positive cells during tissue culture process. The pCambia2300 vector was used for construction of CRISPR/Cas9 gene editing vectors. NLS-Cas9-NLS was inserted to construct the pCambia2300-p35S-NLS-Cas9-NLS-tNOS (pCambia2300-Cas9) vector. The pAtU6-sgRNA-gRNA-U6t was amplified and inserted the upstream of 35S-NLS-Cas9-NLS using the restriction endonucleases Hind III and Sbf I.

2.Sterile cotton seedling culture

The cotton seeds were selected for sulfuric acid delinting. After delinting, the cotton seeds were put into an oven at 40℃ for drying. After 48 h, the delinted cotton seeds were cleaned in 75% alcohol. Seeds were sterilized by soaking them in 3% H₂O₂ for 20 h, followed by three rinses in sterile water. Seeds were then placed in a sterile seedling medium for 5 days (16 h light/8 h dark, 28℃ day/25℃ night, 65% relative humidity and PAR intensity 200 µmol m^− 2 s^− 1).

3.Preparation of medium

The basic medium used in this study was Murashige Skoog Base Salts with B5 vitamins medium (MS). The MS medium was then divided into callus induction medium, embryogenic callus induction medium, embryoid induction medium, and regeneration seedling induction medium.

4. Agrobacterium infection and co-culture

Etiolated seedlings of 5 days-old were collected and the hypocotyls were cut into 0.8 centimeter sections on an ultra-clean table. Approximately, 6–8 sections were cut from each etiolated seedling hypocotyl. The cut hypocotyl segments were soaked in the Agrobacterium solution containing 50 mg/L AS for 3–5 min, and then removed and placed in a dish with sterile filter paper. After drying for approximately 10 min in ultra-clean workbench, the cut hypocotyl segments were placed in co-culture medium containing 50 mg/L AS, and cultured in a light incubator for 48 h (dark, 23℃, 65% relative humidity).

5. Screening of transgenic callus

After two days of co-culture, the segments were cultured in callus induction medium with selected antibiotics to induce callus initiation and proliferation, and sub-cultured once a month and cultivated for a total of 2–3 times until the emergence of enlarge callus. The culture condition as follows: 16 h light/8 h dark, 28℃ day/25℃ night, 65% relative humidity and PAR intensity 200 µmol m^− 2 s^− 1. Callus with relatively large size was considered as positive callus.

6. Induction of embryogenic callus

Positive callus was cultured in embryogenic callus induction medium, sub-cultured once a month, and cultivated for a total of 2–3 times until the emergence of yellow-green, millet-sized, granular embryogenic callus. The culture condition as follows: 16 h light/8 h dark, 28℃ day/25℃ night, 65% relative humidity and PAR intensity 200 µmol m^− 2 s^− 1. Embryogenic callus is further propagated.

7. Induction of embryoids

The embryogenic callus was subcultured in embryoid induction medium, sub-cultured once every 2 weeks, and cultivated for a total of 2–3 times until the emergence of embryoids. The culture condition as follows: 16 h light/8 h dark, 28℃ day/25℃ night, 65% relative humidity and PAR intensity 200 µmol m^− 2 s^− 1.

8. Seedling regeneration

Developing normal somatic embryos were collected and subcultured in regeneration seedling induction medium, sub-cultured once every 2 weeks, and cultivated for a total of 2–3 times until regenerated seedlings appeared. The culture condition as follows: 16 h light/8 h dark, 28℃ day/25℃ night, 65% relative humidity and PAR intensity 200 µmol m^− 2 s⁻1.

9. Seedling planting

The positive regenerated seedlings were put into a nutrient solution after growing roots and then transplanted to the greenhouse for cultivation at 25℃ with a 14h light /10h dark photoperiod and a light intensity of 300 µmol m^− 2s^− 1.The seeds of T0 generation were self-crossed and harvested.

Cotton spatial metabolomics experiment

1.Section preparation and H&E staining

An appropriate amount of OCT compound was added to the bottom bracket of the embedding box, and the samples (globular embryos, torpedo embryos, cotyledonary embryos, embryogenic callus, and non-embryogenic callus) were placed in the embedding block. The samples were placed in the middle of the embedding block, additional OCT compound was added, and tweezers were used to remove any air bubbles surrounding the tissue section. The samples were then sectioned in a −20℃ freezing microtome. After preliminary trimming of the samples, sections were cut at a thickness of 30 µm. When the sections approached the target area, the cutting blade was replaced and 10 µm. sections were taken. The sections were placed on a Superfrost Plus positive charge anti-slip slide and flattened with the aid of an artist’s brush. One section was subjected to HE staining, and then stored at −80℃ for subsequent image analysis.

2. AFAI-MSI

Configure Spray solvent. In positive ion scanning mode, the spray solvent was acetonitrile: water = 80:20 (V/V, containing 0.1% formic acid). In negative ion scanning mode, the spray solvent composition was acetonitrile: water = 80:20 (V/V). The flow rate of the spray solvent was 5 µL/min. After the frozen tissue sections were taken out from the −80℃ freezer, they were quickly put into a vacuum dryer and dried for approximately 30 min at room temperature. The scanning parameters were set by the control program software of the mass spectrum imaging platform, and then the mass spectrum imaging data were collected by progressive scanning. The platform parameters used in the present study were set as follows: Vx was 0.2 mm /s, Dy was 0.1 mm, Dt was 7 s, X-axis scanning length of the scanning area was 10 mm, Y-axis scanning length was 10 mm. The mass spectrum scanning parameters were as follows. An Xcalibur data acquisition and processing system was used for data collection, and the data collection sequence was set according to sample size, step spacing and scanning speed.

3. Imaging analysis

All mass charges were analyzed by imaging to obtain metabolite profile information in tissue sections. The exact location of the specific metabolites was determined by H&E staining of adjacent tissue sections.

4. Data preprocessing

Mass spectrometry imaging data were processed with Cardinal, an open-source R-based software package for transformation, visualization and analysis. The raw data (.raw format) were converted to .imzML files using imzMLConverter. The Cardinal software package was used for background deduction, peak alignment, peak screening, normalization and smoothing processing. Based on the structural partition of the HE staining map, and combined with the result of K-Means clustering, a selection of features was selected for subsequent analysis. Peak detection and peak recognition were carried out in the selected area. A series of processing steps, such as peak alignment and missing value interpolation, were used to obtain quantitative data.

5. Comparative analysis

Metabolites were identified utilizing the SmetDB database. Multivariate statistical analysis was performed on the metabolome data, and unsupervised principal component analysis (PCA) was used to observe the overall distribution among samples and the stability of the whole analysis process. Supervised partial least squares analysis (PLS-DA) and orthogonal partial least squares analysis (OPLS-DA) were used to distinguish the overall differences in the metabolic profiles between groups and to identify the differentially abundant metabolites between groups. Variable importance of projection was obtained using the OPLS-DA model. Potential biomarkers were identified based on a VIP > 1.

Database Website

The Shiny program in R language was used to construct and interactive webpage database of cotton single cell transcriptomics data and spatial transcriptomics data. The design of the user interface (UI) uses the shinythemes package to set custom themes. Functions such as DimPlot, VlnPlot, FeaturePlot, and SpatialPlot in the Seurat package are used in the server to plot and display data and provide output content to the UI.

The webpage (http://cottonsomaticembryo.cricaas.com.cn) is divided into three main modules (three tabs are displayed at the top of the webpage): cotton single-cell transcriptomic data (scRNA-seq data), cotton Spatial Transcriptomics data (Spatial Transcriptomics data1), and cotton Spatial Transcriptomics data (Spatial Transcriptomics data2). Each label can be clicked to enter the webpage containing the detailed results display. The function parameter area is located on the left side of the webpage, and the result display area is located on the right side. Above the result display area is a cluster display plots/gene ViolinPlot/FeaturePlot. In the left function parameter area, one can set the size of point, dimensionality reduction method (t-SNE/UMAP), grouping information (cluster or sample) and select the specified gene to adjust the display result. A list of the top 10 identified marker genes appears below the result display area.

DATA AND SOFTWARE AVAILABILITY

Accession codes-All sequencing data generated in this work is available at the following web addresses: (https://dataview.ncbi.nlm.nih.gov/?search=; https://www.ncbi.nlm.nih.gov).

AUTHOR CONTRIBUTIONS

Conceptualization, XS and FL; Methodology, XY, ZL, JY (Jiachen Yuan), AQ, YW, and YC; Formal Analysis, XS, XY, ZL, JY (Jiachen Yuan) and AQ; Investigation, XG, XY, ZL, JY (Jiachen Yuan), AQ, WQ, YW, YL (Yao Lu),YC, YL, YZ (Yaping Zhou), JY (Jincheng Yang), XL, PW, and HL; Writing XS, XY and AQ; Writing – Review & Editing, XS, AQ, LHE, LSPT and FL; Supervision, XS and FL; Project Administration, XS and FL; Funding Acquisition, XS and FL.

COMPETING INTERESTS

The authors declare no conflict of interest.

ACKNOWLEDGEMENTS

This study was funded by support from the National Natural Science Foundation of China (31621005), National Key Research and Development Program (2021YFF1000102 to Z.Y.), National Natural Science Foundation of China (31670233), the Innovation Scientists and Technicians Troop Construction Projects of Henan Province, the Agricultural Science and the Technology Innovation Program of Chinese Academy of Agricultural Sciences.

Cui, Y. et al. Rapid Mining of Candidate Genes for Verticillium Wilt Resistance in Cotton Based on BSA-Seq Analysis. Front. Plant Sci. 12, 703011 (2021).
Kumar, S. et al. Comparative Transcriptomics of Non-Embryogenic and Embryogenic Callus in Semi-Recalcitrant and Non-Recalcitrant Upland Cotton Lines. Plants-Basel 10, 1775 (2021).
Elhiti et al. Molecular regulation of plant somatic embryogenesis. In Vitro Cell.Dev-Pl. 49, 631–642 (2013).
Salaün, C., Lepiniec, L. & Dubreucq, B. Genetic and Molecular Control of Somatic Embryogenesis. Plants-Basel 10, 1467 (2021).
Wu, X., Liu, C. & Zhang, C. Progress of Somatic Embryogenesis in Cotton. Chinese Bulletin of Botany 25, 469–475 (2008).
Souter, M. & Lindsey, K. Polarity and signaling in plant embryogenesis. J. Exp. Bot. 51, 971–983 (2000).
Shang, H., Liu, C., Zhang, C., Wu, Z. & Li, F. Progress in Mechanisms of Cotton Somatic Embryogenesis. Acta Botanica Boreali-Occidentalia Sinica 29, 637–642 (2009).
Hussain, S.S., Rao, A.Q., Husnain, T. & Riazuddin, S. Cotton somatic embryo morphology affects its conversion to plant. Biol. Plantarum 53, 307–311 (2009).
Guo, H. et al. Dynamic Transcriptome Analysis Reveals Uncharacterized Complex Regulatory Pathway Underlying Genotype-Recalcitrant Somatic Embryogenesis Transdifferentiation in Cotton. Genes 11, 519 (2020).
Xiao, Y. et al. Effects of GhWUS from upland cotton (Gossypium hirsutum L.) on somatic embryogenesis and shoot regeneration. Plant Sci. 270, 157–165 (2018).
Jin, F.Y. et al. Comparative transcriptome analysis between somatic embryos (SEs) and zygotic embryos in cotton: evidence for stress response functions in SE development. Plant Biotechnol. J. 12, 161–173 (2014).
Ding, M. et al. Transcriptomic analysis reveals somatic embryogenesis-associated signaling pathways and gene expression regulation in maize (Zea mays L.). Plant Mol. Biol. 104, 647–663 (2020).
Pais, M.S. Somatic Embryogenesis Induction in Woody Species: The Future After OMICs Data Assessment. Front. Plant Sci. 10, 240 (2019).
Magnani, E., Jiménez-Gómez, J.M., Soubigou-Taconnat, L., Lepiniec, L. & Fiume, E. Profiling the onset of somatic embryogenesis in Arabidopsis. BMC Genomics 18, 998 (2017).
Ji, L. et al. Genome-Wide Reinforcement of DNA Methylation Occurs during Somatic Embryogenesis in Soybean. Plant Cell 31, 2315–2331 (2019).
Liu, Z. et al. Global Dynamic Molecular Profiling of Stomatal Lineage Cell Development by Single-Cell RNA Sequencing. Mol. Plant 13, 1178–1193 (2020).
Satterlee, J.W., Strable, J. & Scanlon, M.J. Plant stem-cell organization and differentiation at single-cell resolution. Proc. Natl. Acad. Sci. USA 117, 33689–33699 (2020).
Farmer, A., Thibivilliers, S., Ryu, K.H., Schiefelbein, J. & Libault, M. Single-nucleus RNA and ATAC sequencing reveals the impact of chromatin accessibility on gene expression in Arabidopsis roots at the single-cell level. Mol. Plant 14, 372–383 (2021).
Denyer, T. et al. Spatiotemporal Developmental Trajectories in the Arabidopsis Root Revealed Using High-Throughput Single-Cell RNA Sequencing. Dev. Cell 48, 840–852.e845 (2019).
Xie, Y. et al. Single-Cell RNA Sequencing Efficiently Predicts Transcription Factor Targets in Plants. Front. Plant Sci. 11, 603302 (2020).
Giacomello, S. A new era for plant science: spatial single-cell transcriptomics. Curr. Opin. Plant Biol. 60, 102041 (2021).
Caprioli, R.M. Imaging mass spectrometry: Molecular microscopy for enabling a new age of discovery. Proteomics 14, 807–809 (2014).
Barbosa, E.A., Junior, M., Jr, C.B., Rocha, T.L. & Engler, J. Imaging Mass Spectrometry of Endogenous Polypeptides and Secondary Metabolites from Galls Induced by Root-Knot Nematodes in Tomato Roots. Mol. Plant-Microbe In. 31, 1048–1059 (2018).
Taira, S. et al. Mass Spectrometric Imaging of Ginsenosides Localization in Panax ginseng Root. A. J.Chinese Med. 38, 485–493 (2010).
He, J. et al. A Sensitive and Wide Coverage Ambient Mass Spectrometry Imaging Method for Functional Metabolites Based Molecular Histology. Adv. Sci. 5, 1800250 (2018).
Cheng, W.H. et al. Polyamine and Its Metabolite H2O2 Play a Key Role in the Conversion of Embryogenic Callus into Somatic Embryos in Upland Cotton (Gossypium hirsutum L.). Front. Plant Sci. 6, 1063 (2015).
Ha, S., Vankova, R., Yamaguchi-Shinozaki, K., Shinozaki, K. & Tran, L.S. Cytokinins: metabolism and function in plant adaptation to environmental stresses. Trends Plant Sci. 17, 172–179 (2012).
Zhang, B.H., Wang, Q.L., Liu, F., Wang, K.B. & Frazier, T.P. Highly efficient plant regeneration through somatic embryogenesis in 20 elite commercial cotton (Gossypium hirsutum L.) cultivars. Plant Omics 2, 259–268 (2009).
Tvorogova, V.E. et al. The WUSCHEL-related homeobox transcription factor MtWOX9-1 stimulates somatic embryogenesis in Medicago truncatula. Plant Cell Tiss. Org. 138, 517–527 (2019).
Robles et al. Differential contributions of ribosomal protein genes to Arabidopsis thaliana leaf development. Plant J. 65, 724–736 (2011).
Cheng, Y., Dai, X. & Zhao, Y. Auxin synthesized by the YUCCA flavin Monooxygenases is essential for embryogenesis and leaf formation in Arabidopsis. Plant Cell 19, 2430–2439 (2007).
Nic-Can, G.I. & De la Peña, C. Epigenetic Advances on Somatic Embryogenesis of Agronomical and Important Crops, in Epigenetics in Plants of Agronomic Importance: Fundamentals and Applications: Transcriptional Regulation and Chromatin Remodelling in Plants. (eds. R. Alvarez-Venegas, C. De la Peña & J.A. Casas-Mollano) 91–109 (Springer International Publishing, Cham; 2014).
Zhai, N. & Xu, L. Pluripotency acquisition in the middle cell layer of callus is required for organ regeneration. Nat. Plants 7, 1453–1460 (2021).
Sakakibara, K. et al. WOX13-like genes are required for reprogramming of leaf and protoplast cells into stem cells in the moss Physcomitrella patens. Development 141, 1660–1670 (2014).
Smetana, O. et al. High levels of auxin signalling define the stem-cell organizer of the vascular cambium. Nature 565, 485–489 (2019).
Baima & S. The Arabidopsis ATHB-8 HD-Zip Protein Acts as a Differentiation-Promoting Transcription Factor of the Vascular Meristems. Plant Physiol. 126, 643–655 (2001).
Ji, J. et al. WOX4 promotes procambial development. Plant Physiol. 152, 1346–1356 (2010).
Ding, A. et al. ERF4 and MYB52 transcription factors play antagonistic roles in regulating homogalacturonan de-methylesterification in Arabidopsis seed coat mucilage. Plant Cell 33, 381–403 (2021).
Sakakibara, K. et al. WOX13-like genes are required for reprogramming of leaf and protoplast cells into stem cells in the moss Physcomitrella patens. Development 141, 1660–1670 (2014).
Gulzar, B. et al. Genes, proteins and other networks regulating somatic embryogenesis in plants. J. Genet. Eng. Biotechnol. 18, 31 (2020).
Baima, S. et al. The expression of the Athb-8 homeobox gene is restricted to provascular cells in Arabidopsis thaliana. Development 121, 4171–4182 (1995).
Haecker, A. et al. Expression dynamics of WOX genes mark cell fate decisions during early embryonic patterning in Arabidopsis thaliana. Development 131, 657–668 (2004).
Wang, Y. et al. Transcriptomes analysis reveals novel insight into the molecular mechanisms of somatic embryogenesis in Hevea brasiliensis. BMC Genomics 22, 183 (2021).
Möller, B.K. et al. Auxin response cell-autonomously controls ground tissue initiation in the early Arabidopsis embryo. Proc. Natl. Acad. Sci. USA 114, E2533-e2539 (2017).
El Ouakfaoui, S. et al. Control of somatic embryogenesis and embryo development by AP2 transcription factors. Plant Mol. Biol. 74, 313–326 (2010).
Warghat, A.R., Thakur, K. & Sood, A. Plant stem cells: what we know and what is anticipated. Mol. Biol. Rep. 45, 2897–2905 (2018).
Capron, A., Chatfield, S., Provart, N. & Berleth, T. Embryogenesis: Pattern Formation from a Single Cell. Arabidopsis Book 7, e0126 (2011).
Scheres & Ben Stem-cell niches: nursery rhymes across kingdoms. Nat. Rev. Mol. Cell Biol. 8, 345–354 (2007).
Fukuda, H. & Ohashi-Ito, K. Vascular tissue development in plants. Curr. Top. Dev Biol. 131, 141–160 (2019).
Helariutta, Y. et al. The SHORT-ROOT gene controls radial patterning of the Arabidopsis root through radial signaling. Cell 101, 555–567 (2000).
Prigge, M.J. et al. Class III homeodomain-leucine zipper gene family members have overlapping, antagonistic, and distinct roles in Arabidopsis development. Plant Cell 17, 61–76 (2005).
Schlereth, A. et al. MONOPTEROS controls embryonic root initiation by regulating a mobile transcription factor. Nature 464, 913-U128 (2010).
Itoh, J.-I., Hibara, K.-I., Sato, Y. & Nagato, Y. Developmental role and auxin responsiveness of class III homeodomain leucine zipper gene family members in rice. Plant Physiol. 147, 1960–1975 (2008).
Wang, H.C., Ngwenyama, N., Liu, Y.D., Walker, J.C. & Zhang, S.Q. Stomatal development and patterning are regulated by environmentally responsive mitogen-activated protein kinases in Arabidopsis. Plant Cell 19, 63–73 (2007).
Qi, S. et al. Global Transcriptome and Coexpression Network Analyses Reveal New Insights Into Somatic Embryogenesis in Hybrid Sweetgum (Liquidambar styraciflua × Liquidambar formosana). Front. Plant Sci. 12, 751866 (2021).
Chaban, C., Waller, F., Furuya, M. & Nick, P. Auxin responsiveness of a novel cytochrome p450 in rice coleoptiles. Plant Physiol. 133, 2000–2009 (2003).
Li, L. et al. CaM/BAG5/Hsc70 signaling complex dynamically regulates leaf senescence. Sci. Rep. 6 (2016).
Ma, W. et al. The LBD12-1 Transcription Factor Suppresses Apical Meristem Size by Repressing Argonaute 10 Expression. Plant Physiol. 173, 801–811 (2017).
Oudenaarden, N.C.M.B.A.V. Spatially resolved transcriptomics and beyond. Nat. Rev. Genet. 16, 57–66 (2015).
Kolodziejczyk, A., Kim, J.K., Svensson, V., Marioni, J. & Teichmann, S. The technology and biology of single-cell RNA sequencing. Mol. Cell 58, 610–620 (2015).
Haghverdi, L., Büttner, M., Wolf, F.A., Buettner, F. & Theis, F.J. Diffusion pseudotime robustly reconstructs lineage branching. Nat. Methods 13, 845–848 (2016).
Takubo, E. et al. Role of Arabidopsis INDOLE-3-ACETIC ACID CARBOXYL METHYLTRANSFERASE 1 in auxin metabolism. Biochem. Bioph. Res.Co. 527, 1033–1038 (2020).
Cao, M. et al. TMK1-mediated auxin signalling regulates differential growth of the apical hook. Nature 568, 240-+ (2019).
McFarlane, H.E., Shin, J.J.H., Bird, D.A. & Samuels, A.L. Arabidopsis ABCG Transporters, Which Are Required for Export of Diverse Cuticular Lipids, Dimerize in Different Combinations. Plant Cell 22, 3066–3075 (2010).
Ogawa, E. et al. ATML1 and PDF2 Play a Redundant and Essential Role in Arabidopsis Embryo Development. Plant Cell Physiol. 56, 1183–1192 (2015).
Benková, E. et al. Local, efflux-dependent auxin gradients as a common module for plant organ formation. Cell 115, 591–602 (2003).
Yadeta, K.A. et al. A Cysteine-Rich Protein Kinase Associates with a Membrane Immune Complex and the Cysteine Residues Are Required for Cell Death. Plant Physiol. 173, 771–787 (2017).
Nishiyama, R. et al. Arabidopsis AHP2, AHP3, and AHP5 histidine phosphotransfer proteins function as redundant negative regulators of drought stress response. Proc. Natl. Acad. Sci. USA 110, 4840–4845 (2013).
Abreu-Neto, J., Turchetto‐Zolet, A., Oliveira, L., Zanettini, M. & Margis‐Pinheiro, M. Heavy metal‐associated isoprenylated plant protein (HIPP): characterization of a family of proteins exclusive to plants. FEBS J. 280 (2013).
Guillou, C. et al. Indirect somatic embryogenesis of Theobroma cacao L. in liquid medium and improvement of embryo-to-plantlet conversion rate. In Vitro Cell. Dev-Pl. 54, 377–391 (2018).
Conrath, U. et al. Enhanced resistance to Phytophthora infestans and Alternaria solani in leaves and tubers, respectively, of potato plants with decreased activity of the plastidic ATP/ADP transporter. Planta 217, 75–83 (2003).
Tirajoh, A., Aung, T.S.T., McKay, A.B. & Plant, A.L. Stress-responsive alpha-dioxygenase expression in tomato roots. J. Exp. Bot. 56, 713–723 (2005).
Stepansky, A. & Leustek, T. Histidine biosynthesis in plants. Amino Acids 30, 127–142 (2006).
Deltour, R. Nuclear activation during early germination of the higher plant embryo. J. Cell Sci. 75, 43–83 (1985).
Pathak, M.R., Teixeira da Silva, J.A. & Wani, S.H. Polyamines in response to abiotic stress tolerance through transgenic approaches. GM crops & food 5, 87–96 (2014).
Bagni, N. & Tassoni, A. Biosynthesis, oxidation and conjugation of aliphatic polyamines in higher plants. Amino Acids 20, 301–317 (2001).
Hicks, L.M. et al. Thiol-based regulation of redox-active glutamate-cysteine ligase from Arabidopsis thaliana. Plant Cell 19, 2653–2661 (2007).
Arkorful, E. et al. Metabolomic Analyses Provide New Insights into Signaling Mechanisms for Nutrient Uptake by Lateral Roots of Pruned Tea Plant (Camellia sinensis). J. Agric. Food Chem. 68, 7890–7903 (2020).
Han, T. et al. The roles of cadmium on growth of seedlings by analysing the composition of metabolites in pumpkin tissues. Ecotox. environ. safe. 226, 112817 (2021).
Bates, P.D. et al. Fatty acid synthesis is inhibited by inefficient utilization of unusual fatty acids for glycerolipid assembly. Proc. Natl. Acad. Sci. USA 111, 1204–1209 (2014).
Yin, Y. et al. Ultra-high α-linolenic acid accumulating developmental defective embryo was rescued by lysophosphatidic acid acyltransferase 2. Plant J. 103, 2151–2167 (2020).
Sun, Y., Zhang, X., Chao, H., Guo, X. & Nie, Y. Somatic embryogenesis and plant regeneration from different wild diploid cotton (Gossypium) species. Plant Cell Rep. 25, 289–296 (2006).
Jin, S. et al. Identification of a novel elite genotype for in vitro culture and genetic transformation of cotton. Biol. Plantarum 50, 519–524 (2006).
Brink, S., Alemany, A., Batenburg, V.V., Moris, N. & Oudenaarden, A.V. Single-cell and spatial transcriptomics reveal somitogenesis in gastruloids. Nature 579, 1–1 (2020).
Geng, Z. et al. Recent advances in genes involved in secondary metabolite synthesis, hyphal development, energy metabolism and pathogenicity in Fusarium graminearum (teleomorph Gibberella zeae). Biotechnol. Adv. 32, 390–402 (2014).
Bhambhani, S. et al. Transcriptome and metabolite analyses in Azadirachta indica: identification of genes involved in biosynthesis of bioactive triterpenoids. Sci. Rep. 7, 5043 (2017).
Abdelrahman, M. et al. Defective cytokinin signaling reprograms lipid and flavonoid gene-to-metabolite networks to mitigate high salinity in Arabidopsis. Proc. Natl. Acad. Sci. USA 118 (2021).
Javelle, M., Vernoud, V., Rogowsky, P.M. & Ingram, G.C. Epidermis: the formation and functions of a fundamental plant tissue. New phytol. 189, 17–39 (2011).

Supplementary Tables are not available with this version.

SupplementalDatacottonembryo20220925.pdf

Download PDF

Version 1

posted

You are reading this older preprint version

Read the latest preprint version →

Dissection of a Cell Atlas of Developing Somatic Embryos in Cotton by Single-cell RNA-sequencing and Spatial Transcriptomic and Metabolomic Analyses

Archived Versions:

Version 1

Abstract

Figures

Introduction

Results

Dynamics of spatial gene expression during the development of somatic embryos in cotton

Analysis of the spatial expression of genes regulating the initial development of somatic embryos in cotton

ScRNA-seq Analysis of Gene Expression During Somatic Embryo Development in Cotton

Establishment of a Gene Expression Map for Somatic Embryo Development in Cotton

Mass Spectrometric Imaging Analysis of Metabolites During Somatic Embryo Development in Cotton

Gene-Metabolite Joint Analysis During Somatic Embryo Development in Cotton

Discussion

Materials And Methods

In situ hybridization

Cotton spatial metabolomics experiment

2. AFAI-MSI

Database Website

Declarations

References

Supplementary Tables

Supplementary Files

Archived Versions:

Version 1