Temperature and photoperiod changes affect cucumber sex expression by different epigenetic regulations

Abstract

Background: Cucumbers (Cucumis sativus) are known for their plasticity in sex expression. DNA methylation status determines gene activity but is susceptible to environmental condition changes. Thus, DNA methylation-based epigenetic regulation may at least partially account for the instability of cucumber sex expression. Do temperature and photoperiod that are the two most important environmental factors have equal effect on cucumber sex expression by similar epigenetic regulation mechanism? To answer this question, we did a two-factor experiment of temperature and photoperiod and generated methylome and transcriptome data from cucumber shoot apices.

Results: The seasonal change in the femaleness of a cucumber core germplasm collection was investigated over five consecutive years. As a result, 71.3% of the 359 cucumber accessions significantly decreased their femaleness in early autumn when compared with spring. High temperature and long-day photoperiod treatments, which mimic early autumn conditions, are both unfavorable for female flower formation, and temperature is the predominant factor. High temperatures and long-day treatments both predominantly resulted in hypermethylation compared to demethylation, and temperature effect was decisive. The targeted cytosines shared in high-temperature and long-day photoperiod treatment showed the same change in DNA methylation level. Moreover, differentially expressed TEs (DETs) and the predicted epiregulation sites were clustered across chromosomes, and importantly, these sites were reproducible among different treatments. Essentially, the photoperiod treatment preferentially and significantly influenced flower development processes, while temperature treatment produced stronger responses from phytohormone-pathway-related genes. Cucumber AGAMOUS was likely epicontrolled exclusively by photoperiod while CAULIFLOWER A and CsACO₃ were likely epicontrolled by both photoperiod and temperature.

Conclusions: Seasonal change of sex expression is a germplasm-wide phenomenon in cucumbers. High temperature and long-day photoperiod might have the same effect on the methylome via the same mechanism of gene-TE interaction but resulted in different epicontrol sites that account for different mechanisms between temperature- and photoperiod-dependent sex expression changes.

Keywords: Cucumber germplasm; DNA methylation; Photoperiod; Sex expression; Temperature.

Fig. 1

Seasonal change in cucumber femaleness indicated by the proportion of nodes with pistillate flowers (PNPF) value was observed during consecutive years using core germplasm collections. a The box plot shows the overall range and distribution of the normalized PNPF value. The bar in the box indicates median value. There is significant difference in the PNPF mean value of the entire germplasm between spring and early autumn in a t-test. b Number of cucumber accessions showing a significant seasonal change in the PNPF according to a t-test. ns, not significant (P > 0.05). c The distribution of 359 cucumber accessions with various PNPF value in spring and autumn. Instable gynoecious/subgynoecious lines as well as stable normal monoecious lines are exceptions in cucumber germplasm and therefore are highlighted by arrows. d The linear relationship between the PNPF in spring and early autumn

Fig. 2

Comparisons between transcriptomes. a Clustering analysis of the 12 transcriptomes by a dendrogram. Approximately unbiased (AU) P-value is calculated by multiscale bootstrap resampling and printed in red color. b Two-dimensional scatter diagram of the first and second principle component (PC) scores of the transcriptomes in the principle component analysis (PCA). c Venn diagram of differentially expressed genes (DEGs) between temperature condition changes (left), photoperiod condition changes (middle), and environment condition changes taking LS as a control (right). LL-HL, environment condition change from LL to HL. d GO-process enrichment analysis of DEGs that were significantly influenced by temperature and photoperiod in ANOVA. Left and right columns indicate the number of downregulated DEGs and upregulated DEGs, respectively

Fig. 3

Comparisons between methylomes. a The number (in Million), direction and degree of methylation changes at differentially methylated cytosines (DmCs) in different condition changes. The color from red to white, the extent of the methylation change from 1 to 0; the color from blue to white, the extent of the methylation change from − 1 to 0. LS-LL, environment condition change from LS to LL. b The general association of differential expression and DNA methylation in the two-factor treatment. K, thousand. c The general change extent in methylation level (ML) and mC/C percentage resulting from the treatments of LL, HS, and HL, taking LS as a control. C/CG/CHG/CHH ML, the increment rate of methylation level of cytosine/CG cytosine/CHG cytosine/CHH cytosine

Fig. 4

Common targeted DmCs by treatments of LL, HS, and HL, taking LS as a control. a Venn diagram of DmCs. The number of DmCs is indicated in thousand. b Correlation of the change in methylation level for common targeted DmCs. The total number of common DmCs in each two-dimensional coordinate is indicated. c The distribution of common targeted DmCs across chromosome 3. The black line indicates the distribution trend of mC. d The association of common targeted DmCs with genes and TEs. TE or gene, DmCs located in regions of genes and/or TEs; TE&gene, DmCs located in regions of genes and TEs at the same time. e The number of TEs for each TE type that was associated with common targeted DmCs

Fig. 5

Distributions of potential epiregulation of genes (red dots) and DETs (blue dots) across chromosomes. The gray lines indicate the distribution trend of mC. CsACO₃ gene, MADS-box genes, etc. that are proposed to be cucumber sex expression-related are shown with an arrow indicating the position. LS-LL, environment condition change from LS to LL. Boxes indicate the clustering of epiregulated regions. The nick indicates a window of 2 Mb on the chromosomes

Fig. 6

Transcription levels and promoter methylation state of MADS-box genes (CAULIFLOWER A and AGAMOUS) and CsACO₃ gene in the treatments. Each ring represents a cytosine and the number above the rings indicates the distance to TSS site. DMR, differentially methylated regions; FPKM, fragments per kilobase of transcript per million fragments