Single-cell and spatial RNA sequencing reveal the spatiotemporal trajectories of fruit senescence

Abstract

The senescence of fruit is a complex physiological process, with various cell types within the pericarp, making it highly challenging to elucidate their individual roles in fruit senescence. In this study, a single-cell expression atlas of the pericarp of pitaya (Hylocereus undatus) is constructed, revealing exocarp and mesocarp cells undergoing the most significant changes during the fruit senescence process. Pseudotime analysis establishes cellular differentiation and gene expression trajectories during senescence. Early-stage oxidative stress imbalance is followed by the activation of resistance in exocarp cells, subsequently senescence-associated proteins accumulate in the mesocarp cells at late-stage senescence. The central role of the early response factor HuCMB1 is unveiled in the senescence regulatory network. This study provides a spatiotemporal perspective for a deeper understanding of the dynamic senescence process in plants.

Fig. 1. Generation of a H. undatus pericarp cell atlas.

a Flowchart of experiments in this study. Different colors represent Barcode and UMI (Unique Molecular Identifier) sequences, where barcodes are used for cell differentiation, and UMIs are used for transcript differentiation. In 10× Genomics reagents, there are a total of 4 million Barcode variations, and UMIs consist of 10 nucleotides, allowing for up to 1,048,576 unique combinations. b UMAP visualization of 13 cell clusters in CK and Post group of H. undatus pericarp samples. Each dot denotes a single cell. Colors denote corresponding cell clusters. c Expression patterns of representative cluster-specific marker genes on UMAP. Dot diameter indicates the proportion of cluster cells expressing a given gene. The color scale represents the gene expression levels, with red indicating high expression and blue indicating low expression.

Fig. 2. Reconstruction of a cellular atlas for mature pericarp of H. undatus using spatial transcriptomics.

a Workflow for sampling and sequencing H. undatus pericarp on the 10× Visium platform. b Illustrations of cell types discovered on glass slides, overlaid on corresponding H&E-stained images. Clusters are named based on the spatial positioning of cell types. c Spatial localization maps (left) and violin plots (right) showing the expression of marker genes for different cell types. The violin plot showed six data nodes for each set of data, arranged from largest to smallest, namely the maximum value (upper edge), the upper quartile, the median, the lower quartile, and the minimum value (lower edge). The sample sizes (n number) are shown on each panel. d Association analysis between spatial transcriptomics and single-cell transcriptomics using four algorithms: SciBet, SingleR, RCTD, and CARD. e UMAP plot displaying 13 clusters of single cells classified into 5 cell types after spatial transcriptomics identification.

Fig. 3. Subcluster analysis for exocarp and endocarp cells.

a Pie charts depicting the percentage composition of various cell types in CK and Post samples. b Optical microscope images of CK and Post samples. Three independent experiments were repeated with similar results. c Bar charts illustrating the statistical proportions of CK and Post sample cells within each subcluster of cells. d UMAP plots for cells of various subclusters. e Sankey diagram showing the distribution of clusters in components and these cells gathered in components in CK and Post samples. Colors according to Fig. 1b. f Heatmap and UMAP of top 1 marker genes of 5 subclusters from the exocarp and endocarp cells, with the top1 gene ID emphasized. The UMAP plot in Fig. 3f illustrated the expression localization of the top 1 gene of subclusters 0, 1, 2, 3, 4 from the heatmap in each subcluster, consistent with the distribution of CK and Post in Fig. 3c. g RNA FISH indicated that the predominant location of HuSAG12 was in the mesocarp. Three independent experiments were repeated with similar results. Components of EX, ME, and VB were labeled in the light field image. HuSAG12 probes were labeled with FAM (green). Nuclei were stained with DAPI (blue). Scale bar: 40 μm.

Fig. 4. Senescent trajectories of exocarp cells and mesocarp cells.

a Gene set scoring results plot showcasing cells highly correlated with target genes in CK and Post samples. b Latent time showed the internal clock of cells. Different colors of latent time represent different differentiation times, with darker shades of red indicating earlier times and darker shades of blue indicating later times. c RNA velocity analysis mapped three cellular states on the pseudotime plot. The colors represent different cell states. The direction of the black arrows represents the potential trajectories of the cells, and the length of the arrows represents the strength of the trends. d Clustering of differentially expressed genes along a pseudotime progression of EX cells and ME cells. e Visualization of the gene expression patterns of top genes in the clusters of Fig. 4d mapped onto the pseudotime trajectory. Pseudotime mapping of each gene with expression curves below. The color of each point represents different cellular states, and the horizontal axis represents time progression from left to right. The figure illustrates the gene’s expression changes across three different states of cells over time. f, g Changes in endogenous superoxide anion and flavonoid concentrations within pericarp during the post-harvest storage in CK and Post samples. Data are presented as mean values ± SD. The style of connecting is spline. The area under curve is filled. h Accumulation of endogenous H₂O₂ within the pericarp of CK and Post samples. Three independent experiments were repeated with similar results. Arrows indicate the deposition of cerium peroxide (Ce[OH]₂OOH and Ce[OH]₃OOH) formed after CeCl₃ staining, representing the deposition of H₂O₂. Scale bar: 20 μm. i Clustering of 529 transcription factors-encoding genes along a pseudotime progression. j The Gene Regulatory Network (GRN) was inferred from the dynamic expression of top 100 genes at pseudo-temporal branching points and 529 transcription factors integrated dynamically expressed across senescence differentiation pseudotime with a parameter cutoff of 2.0. Solid and dotted lines represent positive and negative regulation, respectively. Node size corresponds to the predicted connectivity. Nodes from clusters obtained via MCODE are labeled with different colors. Nodes that are specifically upregulated during senescence in the EX and ME sections were also depicted as inverted triangles and triangles. Nodes co-expressed in EX and ME were represented as circles. k Hierarchical layout of 10 hubs in Fig. 5J. The nodes were ranked and colored by cytoHubba.

Fig. 5. Silencing of HuCMB1 led to faster senescence of H. undatus.

a Expression of four senescence-related genes at different components of pericarp and time points of senescence. n = 3 biologically independent samples. Data are presented as mean values ± SD. A paired two-tailed t-test was used for all statistical analyses. No adjustments were made for multiple comparisons.* represents p < 0.05, and ** represents p < 0.01. b Schematic showing the domain structure of HuCMB1 gene. c–e Changes in fruit phenotype (c), weight loss rate (d), and flavonoid levels in the exocarp (e) after silencing the HuCMB1 gene. n = 5 biologically independent samples used in (c–e). d, e The style of connect is spline. The area under the curve is filled. f RT-qPCR was used to analyze the expression of HuCMB1, HuERD6-2, HuSAG12, and HuMED32 in H. undatus after 9 days of storage under HuCMB1 gene silencing. d–f n = 3 biologically independent replicates for each experiment. Data are presented as mean values ± SD. No adjustments were made for multiple comparisons.

Fig. 6. Hypothesis of senescence regulatory trajectories in the exocarp and mesocarp cells.

Our findings indicated that the senescence process of the fruit exhibited a distinct spatiotemporal transition. The activation of ROS signals in the mesocarp occurred initially, followed by a significant upregulation of resistance genes in the exocarp cells. In the mid to late stages of fruit senescence, senescence-related genes like HuSAG12 showed high expression in the mesocarp. The early responsive transcription factor HuCMB1 held a central position in the senescence regulatory network created using the SCODE algorithm. EX represents the exocarp; ME represents the mesocarp; EN(F) represents the endocarp and endocarp fibers. SRP represents senescence-related proteins. SOD stands for superoxide dismutase. AQP denotes aquaporin proteins. Pol II represents RNA polymerase II. TF refers to transcription factor.