Transcriptional landscape of highly lignified poplar stems at single-cell resolution

Abstract

Background: Plant secondary growth depends on the activity of the vascular cambium, which produces xylem and phloem. Wood derived from xylem is the most abundant form of biomass globally and has played key socio-economic and subsistence roles throughout human history. However, despite intensive study of vascular development, the full diversity of cell types and the gene networks engaged are still poorly understood.

Results: Here, we have applied an optimized protoplast isolation protocol and RNA sequencing to characterize the high-resolution single-cell transcriptional landscape of highly lignified poplar stems. We identify 20 putative cell clusters with a series of novel cluster-specific marker genes and find that these cells are highly heterogeneous based on the transcriptome. Analysis of these marker genes' expression dynamics enables reconstruction of the cell differentiation trajectories involved in phloem and xylem development. We find that different cell clusters exhibit distinct patterns of phytohormone responses and emphasize the use of our data to predict potential gene redundancy and identify candidate genes related to vascular development in trees.

Conclusions: These findings establish the transcriptional landscape of major cell types of poplar stems at single-cell resolution and provide a valuable resource for investigating basic principles of vascular cell specification and differentiation in trees.

Fig. 1

Protoplast isolation and identification of cell clusters in poplar stem. A Workflow of the scRNA-seq of poplar stems. Protoplasts isolated from bark and wood were loaded into a 10x Genomics Chromium Controller, separately. B UMAP visualization of 20 putative clusters derived from 6796 cells. Each dot denotes a single cell. Colors indicate corresponding cell clusters. C Violin plots showing expression of representative marker genes in each cell type. Clusters are indicated on the y-axis. Colors denote corresponding cell clusters. D Results of GO enrichment analysis of the cluster-enriched genes

Fig. 2

Differentiation trajectory of phloem cells. A A successive differentiation trajectory from phloem mother cells to SEs and CCs. Each dot indicates a single cell. The black arrow indicates the start of the trajectory. B Expression patterns of phloem-specific genes (AUX1, PIN1, APL, FTIP1, SEOR1, and CalS7). The colors represent expression levels of these genes in individual cells. C RNA in situ hybridization of APL with the sense probe as a negative control. Scale bars, 100 μm. D Heatmap showing expression of the branch-dependent genes over pseudo-time. Representative marker genes are shown on the right of the heatmap. Both sides of the heatmap are the end of pseudo-time. E Dot plot of GO enrichment analysis of the identified cell clusters. ER, endoplasmic reticulum. L-Glu, L-glutamate

Fig. 3

Differentiation trajectory of xylem cells. A The successive differentiation trajectory from xylem mother cells to mature xylem cells. Each dot indicates a single cell. The black arrow indicates the start of the trajectory. B Expression of representative marker genes (ACL5, PtrHB4, PtrHB7, and PtrHB8) at the beginning of the trajectory. The color bar indicates relative expression levels. C Heatmap showing expression levels of branch-dependent genes over pseudo-time. Representative marker genes are shown on the right of the heatmap. Both sides of the heatmap are the end of pseudo-time. D Bar chart showing results of GO enrichment analysis of the identified cell clusters. Expression of ABR1 (E) and ESK1a (F) at the trajectory and results of RNA in situ hybridization of them with the sense probe as a negative control. Scale bars, 100 μm. G Representative marker genes (PdDUF579-9, PtAP66, PtAP17, PtrMAN6, XCP1, and VND1) expressed in XSCW branch. The color bar indicates relative expression levels

Fig. 4

UMAP visualization of expression patterns of genes related to hormone response (auxin, cytokinin, gibberellin, brassinosteroid, strigolactone, and abscisic acid). The colors represent expression levels of these genes in individual cells

Fig. 5

scRNA-seq predicts potential gene redundancy. A Comparison of Jaccard index values between WGD-derived and random selected gene pairs. Asterisks indicate highly significant differences according to the Wilcoxon signed-rank test (P < 0.0001). B Boxplot showing that Jaccard index values of correlations between WGD-derived gene pairs decreased with increases in d_N/d_S. C Dot plot of GO enrichment analysis of paralogs within the top 10% of the Jaccard index. D Bar chart (top) and heatmap (bottom) showing the distribution of Jaccard index values of 5143 WGD-derived paralogs, and the distribution of the number of overlapping clusters, where both copies of a gene pair were expressed in more than 25% of the cells in that cluster. E Phylogenetic tree of CesA1/4/7/8 in P. alba var. pyramidalis and P. trichocarpa. F Frequency of Jaccard index between CesA7a and all other expressed genes. G Frequency of Jaccard index between CesA8a and all other expressed genes