Global transcriptome dissection of pollen-pistil interactions induced self-incompatibility in dragon fruit (Selenicereus spp.)

Abstract

Self-incompatibility (SI) is a major issue in dragon fruit (Selenicereus spp.) breeding and production. Therefore, a better understanding of the dragon fruit SI mechanism is needed to improve breeding efficiency and ultimate production costs. To reveal the underlying mechanisms of SI in dragon fruit, plant anatomy, de novo RNA sequencing-based transcriptomic analysis, and multiple bioinformatic approaches were used to analyze gene expression in the pistils of the self-pollinated and cross-pollinated dragon fruit flowers at different intervals of time after pollination. Using fluorescence microscopy, we observed that the pollen of 'Hongshuijing', a self-incompatible dragon fruit variety (S. monacanthus), germinated on its own stigma. However, the pollen tube elongation has ceased at 1/2 of the style, confirming that dragon fruit experiences gametophyte self-incompatibility (GSI). We found that the pollen tube elongation in vitro was inhibited by self-style glycoproteins in the SI variety, indicating that glycoproteins were involved in SI. That is to say the female S factor should be homologous of S-RNase or PrsS (P. rhoeas stigma S factor), both of which are glycoproteins and are the female S factors of the two known GSI mechanism respectively. Bioinformatics analyses indicated that among the 43,954 assembled unigenes from pistil, there were six S-RNase genes, while 158 F-box genes were identified from a pollen transcriptomic dataset. There were no P. rhoeas type S genes discovered. Thus, the identified S-RNase and F-box represent the candidate female and male S genes, respectively. Analysis of differentially expressed genes (DEGs) between the self and cross-pollinated pistils at different time intervals led to the identification of 6,353 genes. We then used a weighted gene co-expression network analysis (WGCNA) to find some non-S locus genes in SI responses in dragon fruit. Additionally, 13 transcription factors (TFs) (YABBY4, ANL2, ERF43, ARF2, BLH7, KNAT6, PIF3, two OBF1, two HY5 and two LHY/CCA) were identified to be involved in dragon fruit GSI. Thus, we uncovered candidate S and non-S genes and predicted more SI-related genes for a more detailed investigation of the molecular mechanism of dragon fruit SI. Our findings suggest that dragon fruit possesses a GSI system and involves some unique regulators. This study lays the groundwork for future research into SI mechanisms in dragon fruit and other plant species.

Keywords: Dragon fruit; Gametophyte self-incompatibility; Non-S genes; S genes; Self-incompatibility.

Figure 1. Pollen growth in pistil in vivo and effects of style glycoproteins on pollen tube elongation in vitro.

(A) Time line of pollen germination on stigmatic surface and elongation of pollen tubes in style using fluorescent microscopy. Middle panel depicts the pistil of Hongshuijing (scale bar = one cm), left panel corresponds to self-pollinated pistil sections (scale bar = 100 um), right panel corresponds to cross-pollinated pistil sections (scale bar = 100 um). Lines of different colors (blue, brown and gray) along with the left and right sides of the style represent where the pollen tube reached style at different time points. (B) Boxplots of pollen tube elongation rates at different time points after self-pollinated (SP) and cross-pollinated (CP). SP12/CP12, SP24/CP24, SP36/CP36 mean 12, 24, 36 hours after either self or cross pollinated, respectively. C and D: Pollen tube elongation of Hongshuijing (panel C) and Dahong (panel D) in vitro using different treatments. Control, medium without any extra protein added. Total soluble protein, medium supplemented with total soluble proteins derived from the style of Hongshuijing. Glycoprotein, medium supplemented with isolated glycoproteins from the style of Hongshuijing. Glycoproteins-free fractions, liquid medium supplemented with glycoproteins-free protein fractions from the pistil of Hongshuijing. The “*” or “**” on the figure means significant (p-value ≦ 0.05) or highly significant (p-value ≦ 0.01) different with T-test, respectively.

Figure 2. Validation of RNA-seq results by qPCR.

20 genes were used in this experiment. The results are displayed by line charts. RNA-seq data are depicted as blue lines, qPCR results are depicted as the red lines.

Figure 3. Distribution of the DEGs.

(A) DEGs across six pairwise comparisons of the self-pollinated samples (control vs SP12, control vs SP24, control vs SP36, SP12 vs SP24, SP12 vs SP36, and SP24 vs SP36). Yellow bars in they-axis represent the total number of DEGs in each parwise combination. Black bars in the x-axis represent the number of DEGs shared across pairwise combinations connected by the dark dots in the body of the plot. (B) DEGs across six pairwise comparisons of the cross-pollinated samples (control vs CP12, control vs CP24, control vs CP36, CP12 vs CP24, CP12 vs CP36, and CP24 vs CP36). Yellow bars in they-axis represent the total number of DEGs in each parwise combination. Black bars in the x-axis represent the number of DEGs shared across pairwise combinations connected by the dark dots in the body of the plot. (C) Common and special DEGs between pistils of self- and cross-pollinated at 12 h, 24 h, and 36 h after pollination. The 214 common DEGs in total in this plot is the rejection related gene set. (D) Common and special DEGs cross three pairwise comparisons (control vs SP12, control vs CP12, and SP12 vs CP12). The shared 982 DEGs in total in this plot are the recognition related gene set.

Figure 4. GO and KEGG enrichment of the overlapped 534 DEG and 132 DEGs.

For GO enrichment, the first circle: the GO term of the first 20 enriched terms. Different colors represent different ontologies (blue for biological process, yellow for molecular function); the second circle: the number of genes in the GO term of the background and Q value of the enrichment. The more genes the longer the bar, the smaller the Q value and the redder the color, most of the −log10 (Q value) is between 10 and 15; the third circle, enriched gene proportion bar chart in purple, the more genes the longer the bar; the fourth circle, enrich factor value of each GO term (the number of DEGs is divided by the number of genes in the background in the GO term, each grid of the background grid line represents 0.1). For KEGG enrichment, the first 20 enriched pathways were shown in the figure; the dot’s size and color represent the gene number and Q value enriched in each pathway.

Figure 5. Correlation of TF22 and the top 50 hub genes of the overlapped 534 DEG network.

This network consists of nodes (blue rhombuses, yellow square and brown arrows) as well as lines that represent genes and intergenic correlations. The red rhombuses and blue square represent the hub genes and other genes in network 534 DEGs, respectively. The green arrows represent the transcription factors of TF22. The 50 hub genes and 22 TFs share 2 genes.

Figure 6. Flowchart of the identificantion of key gene sets.