Atypical DNA methylation, sRNA-size distribution, and female gametogenesis in Utricularia gibba

Abstract

The most studied DNA methylation pathway in plants is the RNA Directed DNA Methylation (RdDM), a conserved mechanism that involves the role of noncoding RNAs to control the expansion of the noncoding genome. Genome-wide DNA methylation levels have been reported to correlate with genome size. However, little is known about the catalog of noncoding RNAs and the impact on DNA methylation in small plant genomes with reduced noncoding regions. Because of the small length of intergenic regions in the compact genome of the carnivorous plant Utricularia gibba, we investigated its repertoire of noncoding RNA and DNA methylation landscape. Here, we report that, compared to other angiosperms, U. gibba has an unusual distribution of small RNAs and reduced global DNA methylation levels. DNA methylation was determined using a novel strategy based on long-read DNA sequencing with the Pacific Bioscience platform and confirmed by whole-genome bisulfite sequencing. Moreover, some key genes involved in the RdDM pathway may not represented by compensatory paralogs or comprise truncated proteins, for example, U. gibba DICER-LIKE 3 (DCL3), encoding a DICER endonuclease that produces 24-nt small-interfering RNAs, has lost key domains required for complete function. Our results unveil that a truncated DCL3 correlates with a decreased proportion of 24-nt small-interfering RNAs, low DNA methylation levels, and developmental abnormalities during female gametogenesis in U. gibba. Alterations in female gametogenesis are reminiscent of RdDM mutant phenotypes in Arabidopsis thaliana. It would be interesting to further study the biological implications of the DCL3 truncation in U. gibba, as it could represent an initial step in the evolution of RdDM pathway in compact genomes.

Figure 1

Architecture and annotation of lncRNAs in U. gibba and other plants. (a) lncRNAs size distribution in U. gibba is shown in a barplot grouping each 100-nt in size (X axis) and by percentage of the total (Y axis). (b) Structure of lncRNAs in selected plants (Additional file 1: Dataset S1). The box plot groups the proportion of exon number (X axis) by percentage (Y axis) among lncRNAs identified from various plant genomes. (c) Genomic annotation of U. gibba lncRNAs. The proportion is from 4295 putative lncRNAs loci identified. Red represents the proportion of lncRNAs mapped to body gene regions; blue shows lncRNAs located in intergenic regions; green indicates those that overlap gene body and intergenic regions. (d) Coding and noncoding transcripts expression. The violin plot represents the total of mRNAs, lncRNAs and lincRNAs identified in this study with correspondent log2(tpm).

Figure 2

Noncoding RNA landscape in Utricularia gibba. Representation of the U. gibba genome in a circus plot where the outside blocks represent complete chromosomes (dark gray) and Unitigs or contigs (gray) larger than 1 Mb in size. The density plots were calculated in 10Kb windows. Red histogram shows the gene density. The Purple histogram displays transposable element density. lincRNA density is represented in the green histogram, and small RNA density for sRNAs 20-nt to 24-nt in size are shown in orange.

Figure 3

Annotation of sRNAs in U. gibba and distribution of small RNA size abundance in plants. (a) Top, a phylogenetic tree of plants with representative species for green algae, early branching land plants, gymnosperms, and angiosperms. Bottom, a stacked plot of small RNA abundances (20-nt to 24-nt in size) from various small RNA-seq studies. Y-axis shown the percentage of reads for each size group. (b) Total locus predictions from ShortStack for miRNAs and siRNAs at different sizes. (c) Numbers of reads mapped per loci at different sizes from ShortStack. (d) The largest complete chromosome (Unitig_0) representation of gene density, TE density, 21-nt small RNAs loci density distribution and 24-nt small RNAs loci distribution. Vca: Volvox carteri; Cre: Chlamydomonas reinhardtii; Cco: Chara coraline; Mpo: Marchantia polymorpha; Ppa: Physcomitrella patens; Mqu: Marsilea quadrifolia; Pab: Picea abies; Ginkgo biloba; Cru: Cycas rumphii; Atr: Amborella trichopoda; Sbi: Sorghum bicolor; Zma: Zea mays; Osa: Oryza sativa; Tae: Triticum aestivum; Ath: Arabidopsis thaliana; Ghi: Gossypium hirsutum; Ptr: Populus trichocarpa; Han: Helianthus annuus; Ugi: Utricularia gibba; Can: Capsicum annuum; Sly: Solanum lycopersicum; Lsi: Lagnaria siceraria; Cmo: Cucurbita moschata; Cpe: Cucurbita pepo; Cla: Citrullus lanatus; Mdo: Malus domestica; Ppe: Prunus persica; Mtr: Medicago truncatula; Gma: Glycine max; Pvu: Phaseolus vulgaris.

Figure 4

Truncated DCL3 and female gametophyte development in the ovule of U. gibba. (a) Phylogenetic analysis of DCL3 in some angiosperms including U. gibba. A typical DCL3 protein has the DEAD, Helicase, Dicer dimerization, PAZ, and two Ribonuclease III domains, while UgDCL3 only contains the PAZ and Ribonuclease III domains. (b) Multiple sequence alignment of DCL3 proteins in blocks of 150 aa showing the missing domains in UgDCL3. (c) Developing ovules showing a single pre-meiotic precursor corresponding to the MMC. (d) Twin pre-meiotic precursors prior to megasporogenesis. (e) Six differentiated cells (asterisk) corresponding to gametic precursors prior to megasporogenesis. (f) Functional megaspore (FM) and two degenerated megaspores (asterisk) following megasporogenesis. (g) Developing ovule showing a single 8-nuclear female gametophyte. (h) Developing ovule showing a female gametophyte (FG) and two ectopic gametic precursors (arrows) at the chalazal region. i Developing ovule showing two independent female gametophytes (dashed). Scale bars: 12.5 µm in (c) to (f); and 20 µm in (g) and (i).

Figure 5

Genome-wide methylation in Utricularia gibba. (a) Distribution of methylation (5mC) representation only in the complete chromosomes in U. gibba for location of pericentromeric regions. Here, the blue light red shows the methylation density in BS-Seq replicates and the red histogram represent the methylation distribution in SMRT-Seq with normalized data in Y-axis representing the 5mC density in windows of 100kb (Scale 1Mb). (b) DNA methylation level for each context at the genome level. X-axis shown the methylation level according to Schultz et al., 2012. CG methylation in blue, CHG methylation in green, and CHH methylation in yellow. (c) Methylation levels for CG, CHG and CHH context. (d) Distribution of DNA methylation around all genes. Methylation representation in all contexts (CG, CHG and CHH) 2-kb upstream and downstream. (e) TE DNA methylation representation in all contexts (CG, CHG and CHH) 2-kb upstream and downstream.