Allele-specific DNA methylation and gene expression during shoot organogenesis in tissue culture of hybrid poplar

Abstract

Plant tissue regeneration is critical for genetic transformation and genome editing techniques. During the regeneration process, changes in epigenetic modifications accompany the cell fate transition. However, how allele-specific DNA methylation in two haplotypes contributes to the transcriptional dynamics during regeneration remains elusive. Here we applied an inter-species hybrid poplar (Populus alba × P. glandulosa cv. 84 K) as a system to characterize the DNA methylation landscape during de novo shoot organogenesis at allele level. Both direct and indirect shoot organogenesis showed a reduction in genome-wide DNA methylation. At gene level, non-expressed genes were hypermethylated in comparison with expressed genes. Among the genes exhibiting significant correlations between levels of DNA methylation and gene expression, the expression patterns of 75% of genes were negatively correlated with DNA methylation in the CG context, whereas the correlation patterns in the CHH context were the reverse. The allele-biased DNA methylation was consistent during shoot organogenesis, with fewer than one-thousandth of allele-specific methylation regions shifted. Analysis of allele-specific expression revealed that there were only 1909 genes showing phase-dependent allele-biased expression in the regeneration process, among which the allele pairs with greater differences in transcription factor binding sites at promoter regions exhibited greater differences in allele expression. Our results indicated a relatively independent transcriptional regulation in two subgenomes during shoot organogenesis, which was contributed by cis-acting genomic and epigenomic variations.

Figure 1

Characterization of the DNA methylome of 84 K poplar during the process of direct and indirect shoot organogenesis. a Illustration of samples collected for genome-wide bisulfite sequencing. Leaves from 40-day-old seedlings were used as leaf explant (LE); shoots grown from LEs at 15 DAC (DP₁) and 30 DAC (DP₂) on SIM during direct organogenesis, as well as callus or shoot samples at 14 DAC (IP₁) on CIM and14 DAC (IP₂), 28 DAC (IP₃) on SIM during indirect organogenesis were collected. Scale bar = 0.2 cm. b Gene density and methylation levels on 19 chromosomes of the two subgenomes of P. alba and P. glandulosa in LE samples. c Distribution ratio of average methylation levels of mCG, mCG, mCHG, and mCHH in samples from different phases of shoot regeneration. P values were calculated by one-way ANOVA: *P < 0.05; **P < 0.01; ***P < 0.001. d Difference analysis of methylation levels between samples at different phases in the CG, CHG, and CHH contexts. Lower-case letters represent a significant difference between samples (P < 0.05). e Characterization of the DNA methylation pattern along genes and TEs.

Figure 2

Identification of DMRs between samples from different phases during shoot regeneration. a, b Methylation changes in each DMR during direct and indirect shoot organogenesis processes, respectively. The upper panel displays DMR with increased methylation, while the lower panel displays DMR with decreased methylation, and among them the black circle indicates that the methylation level of DMRs transitioned to high (>80%) or low (<20%) levels. c Relative fraction of DMR distribution in genic regions. d Time-resolved analysis of DNA methylation levels of all identified DMRs in CG, CHG, and CHH contexts. The number on the left represents the P value of each temporal expression profile, while the number on the right represents the number of DMRs in the profile. e Functional analysis of DMGs using the KEGG database.

Figure 3

Association between DNA methylation and gene expression. a Gene expression patterns related to DNA methylation establishment, maintenance, and demethylation during shoot regeneration of 84 K poplar. Colour scale indicates gene expression levels. b Comparison of methylation levels between five classes of genes based on expression levels for gene body and upstream and downstream (flanking 2-kb regions). Small letters represent a significant difference in methylation levels among different groups of genes (P < 0.05). c Pearson correlation between gene expression and methylation levels of 80 bins located in gene body and upstream and downstream regions (|r| > 0.6 and P < 0.05, significant correlation). The number of genes whose methylation level was closely related to the level of their expression was counted in CG, CHG, and CHH contexts.

Figure 4

Allele-specific expression (ASE) during shoot regeneration. a, b PCA score plots of allele expression in samples from each phase of direct and indirect shoot organogenesis, each point representing an independent biological replicate. c Venn diagram displaying the numbers of ASEGs and the frequency statistics of their occurrence in each phase sample. d Venn diagrams displaying the numbers of phase-dependent ASEGs in samples at regenerate phases compared with the explant samples. e, f Allelic expression patterns of universal master regulators of callus formation and shoot regeneration, respectively. Colour scale indicates gene expression levels.

Figure 5

Cis-regulatory analysis of ASE. a Numbers of ASMRs and ASMGs in CG, CHG, and CHH contexts. b Venn diagram displaying the numbers of identified alleles, ASMGs, and phase-dependent ASEGs. c Statistical analysis was conducted on the percentage of 27 954 ASMGs and 1871 phase-dependent ASEGs containing different numbers of ASMRs. d Correlation analysis between TFBSs cosine similarity of alleles or phase-dependent ASEGs and their expression levels in five groups. Asterisks represent a significant difference in correlation coefficient of each group between alleles and phase-dependent ASEGs (***P < 0.001).