Analysis of epigenomic variability in grapevine and its relation with structural variation

Epigenetics encompasses a series of chromatin modifications that are potentially inheritable and can result into a change of gene expression without involving a change in the underlying DNA sequence. Epigenetics is involved in several fundamental mechanisms that regulate cell cycle in all eukaryotes including X chromosome inactivation, gene silencing, paramutation, parental imprinting, chromatin position effect, plant gametogenesis, flowering time, stress responses and light signaling. Within epigenetic modifications, DNA methylation is predominant and widespread in all eukaryotic kingdoms and consists in a reversible reaction which transfers a methyl group on a cytosine. In mammals methylation occurs in CG-rich regions known as CpG islands, whereas in plant methylation may occurs in CG, CHG and CHH contexts, where H may be A,C or T, with different mechanisms. Depending on the location, DNA methylation may have opposite effects: in heterochromatin it is generally associated to transcriptional inactivity but in the transcribed region of genes, methylation in the CG context is associated to medium-to-high transcriptional level. The silencing effect of DNA methylation represents a useful defense weapon against both retrovirus infection and transposable element (TE) insertions: indeed such sequence elements generally become highly methylated as a plant response to prevent further mobilization. Despite this mechanism, during evolution TEs colonized eukaryotic genomes, up to represent 75% of the total genome in some plant species. . TEs are a major factor underlying the tremendous intra-species genome variability that has been revealed thanks to the introduction of next generation sequencing (NGS) technology and the resequencing of several individuals of the same species. This led scientists to introduce, initially only for bacteria then for any organism, the concept of pan-genome, composed by a common genome shared by all the individuals of the species and a dispensable genome which is not essential for survival, but is the foundation for phenotypic variability. The dispensable genome consists of the entire set of structural variations (SVs) observed among individuals and is mainly represented by TEs. TE sequences are generally methylated and their methylation may spread in their flanking regions.Thus, when TEs accidentally insert nearby genes or regulating sequences, they may alter their epigenetic status creating epigenetic variants called epialleles. 4 Specific protocols of NGS, involving the use of bisulfite, which in the overall process converts unmethylated cytosines to thymines, allow to map all methylcytosines of a genome with single-base resolution. The aim of this work was to analyse the relationship between structural variations, mainly represented by TEs, and epigenetic variations in plants. Grapevine is a suitable model for this study because it is a perennial species and generally, it is vegetatively propagated in agriculture. This technique preserves the genome from recombination, thus it allows maintaining the genotypes stable across clonal generations and focusing on mere epigenetic variation. Moreover for grapevine a highly homozygous reference genome is already available as well as a set of grapevine-specific TEs annotated. Three biological replicates of leaf nuclear DNA of the cultivar Pinot Noir, which shares one haplotype with the sequenced genome reference, have been sequenced and analysed. To evaluate the spreading of internal TE methylation on the flanking regions, we considered hemizygous TEs in order to compare the same regions on homologous chromosome in presence or absence of TEs. Consistently with other species, grapevine TEs show high methylation in their sequence in both CG and CHG context whereas CHH context is extremely low methylated. Internal TE methylation is generally spread on their flanking regions. Within TEs, retrotranspons show a stronger impact on flanking regions compared to DNA-transposons, with different behaviors according to the differential genomic distribution of TE-groups: Ty3-Gypsy usually insert in highly methylated regions of pericentromeric chromatin, LINEs element are frequently found in highly CG methylated gene bodies, Ty1-Copia display more variable locations. Generally, where not saturated, retrotransposon insertions provoke an increase of methylation in both CG and CHG contexts, supported by statistical analyses. DNA methylation is also present in transcribed regions of grapevine genes, in particular in the CG context. A set of about 19000 genes was utilized to analyze gene body methylation (GBM) in grapevine. Similarly to other species, grapevine GBM displays an asymmetrical bell-shape profile, in which the 5’ is much less methylated than 3’. Surprisingly introns appear more methylated than exons, in contrast with other species such as Arabidopsis, humans and honey bee. Grapevine introns occupy a large part of the genome (36.7%) and are quite rich in TEs that represent 12.4% of their sequence. The moderate TE content of introns may partially explain their higher methylation compared to flanking exons. However, when excluding genes carrying TE from the analysis, methylation in both exons and introns is reduced but still present, confirming that GBM methylation is independent from TEs, although their insertion may increase it. 5 Analysis of gene expression showed that genes located in highly methylated regions, especially in the CHG context, show on average a lower expression rate and furthermore their expression tend to be more conserved within varieties. On the contrary, when methylation occurs in gene bodies, transcriptional activity is not reduced and it may be even higher. Gene expression may also be modulated by TEs: when these are located in gene flanking regions, gene expression rate is significantly lower than unaffected genes, whereas genes whose introns are enriched in TEs display significantly higher methylation and expression rates. Lastly, allele-specific expression analyses indicate that hemizygous TEs may affect the contribution of the two alleles to the expression rate of the gene. Taken together these data confirm that DNA methylation occurs in grapevine withpatterns comparable with other plant species, but with the peculiarity of highly methylated introns whose methylation is generally associated to moderate TE content and medium-to-high expression levels