Epigenetic processes involve chromatin remodelling and imprinting. Epigenetic marks are set in place and maintained through DNA methylation, modifications of proteins such as histones and remodelling partners, transcription activating factors and miRNAs (Kim et al. 2009). Epigenetics is important in normal development and differentiation, adaptation to stress, ageing, cancer and life style related diseases (Laird and Jaenisch 1996; Jones and Baylin 2007). Epigenetics has been implied in the cross talks between myriads of processes with crucial regulatory roles (Allis and Jenuwein 2016).

DNA methylation is a postreplicative modification of DNA carried out by a DNA methyltransferase using SAM as a methyl donor. It can alter DNA protein interactions, DNA conformation and chromatin structure thereby influencing activity and functional states. It involves distinct DNA methyltransferases namely, DNMT1, DNMT2, DNMT3A and DNMT3B (figure 1) (Jurkowska and Jeltsch 2016). DNA methylation patterns vary across development and differentiation in cells and tissues (Barlow and Bartolomei 2014). Studies with identical monozygotic twins and altered nutritional and physiological states have demonstrated that these are characterized by specific but diverse patterns of DNA methylation (Cooney et al. 2002; Fraga et al. 2005). DNA methylation has been implicated in several functions during development, throughout the life processes and in ageing. Although genetic changes are thought to be the primary drivers of many cancers, in more than 50% cancers, DNA methylation patterns can be correlated with origin, aggressiveness and progression of cancer (Jones and Baylin 2007).

Fig. 1
figure 1

(a) Summary of functions of DNA methylation in various organisms. DNA methylation plays a role in various processes throughout the development in plants, invertebrates and vertebrates. (b) De novo and maintenance methylation.

Full size image

Microbes

A primary function of DNA methylation in microbes is preservation of the genetic material and genome integrity. In microorganisms, DNA cytosine and adenine methylation are involved in the restriction modification systems (Sánchez-Romero et al. 2015) which are important in preserving genetic identity of an organism. It works to prevent incoming foreign DNA from getting internalized in the bacteria. DNA adenine methylation plays a role in conferring strand discrimination in mismatch repair (Radman and Wagner 1986). This elegant process ensures that the mismatches in the newly replicated strand of the DNA are selectively repaired. DNA methylation influences gene expression (Low and Casadesús 2008) and is known to regulate virulence and phase variation (Marinus and Casadesus 2009). Deinococcus radiodurans, the world’s most radiation and desiccation resistant microbe, harbours methylation machinery which comprises of an adenine methyltransferase which plays a role in metabolic rewiring (Prasad et al. 2005; Shaiwale et al. 2015; Patil et al. 2017) and an unusual cytosine DNA methyltransferase (Patil et al. 2017). When as many as 230 diverse bacterial and archaeal genomes were analysed, evidence of DNA methylation was found in 215 microbes (93% of those sequenced) and over 600 DNA enzymes (methyltransferases) were annotated (Blow et al. 2016). While several DNA methylation enzymes are part of the restriction modification systems (consistent with their known role in defence against viruses), a substantial number of them appeared to be involved in genome regulation, and have a more crucial and perhaps critical role in prokaryotic physiology and biology (Blow et al. 2016). The role in restriction modification systems are confined to microbes. As the organisms have evolved, the functions of DNA methylation have undergone changes, with methylation assuming more regulatory role in eukaryotes in epigenetic processes involved in chromatin structure and adaptation.

DNA methylation in mammals

In mammals, DNA methylation is involved in regulating gene expression, chromatin structure and imprinting (Deobagkar et al. 1990). Mutations in genes encoding DNA methyltransferases demonstrate embryonic lethality (Li et al. 1992). Several roles played by DNA methylation have been reviewed previously (Herman et al. 1995). In mammals, DNA methylation plays a crucial role in development (Li et al. 1992; Okano et al. 1999; Smith and Meissner 2013), cancer initiation and progression (Jones and Baylin 2007) and shows transgenerational transmission. DNA cytosine methylation pattern shows an alteration in response to nutrition, environmental cues (Jones and Baylin 2007) and upon exposure to stress and toxicants (Deobagkar et al. 2012). Recently, novel autoregulatory aspects of DNA methyltransferases have been revealed which involve N-terminal domain and post-translational modifications (Jetsch and Juroskava 2016). The mechanisms of mammalian X chromosome inactivation have been extensively studied and are reviewed in the following section as a ‘case study’ to elaborate the complex interplay in the epigenetic regulation.

X chromosome inactivation

X chromosome inactivation in female mammals is an example of differential regulation of homologous chromosomes. Mary Lyon (Rastan 2015) suggested the hypothesis of X chromosome inactivation to explain the presence of the Barr body in female mammals (Rastan 2015). In eutherian mammals and marsupials, one of the two X chromosomes undergoes inactivation. Imprinting and epigenetic processes have been implicated in X chromosome inactivation (Prothero et al. 2009; Migeon 2017). In marsupials, the paternal X is inactivated, and in mice, in extraembryonic tissue, the paternal X chromosome is inactivated while in the embryo proper, the X inactivation is random. The inactive X chromosome (Xi) is late replicating, heterochromatic and transcriptionally silenced (Gartler and Riggs 1983). It has been well documented that in mammals, one X chromosome is marked as active while the other gets inactivated in all the cells. Mammals have evolved an elaborate mechanism for establishing and maintaining the inactive X chromosome. X chromosome inactivation has been suggested to be an important process in sex determination (Chandra 1985). For a particular X chromosome to undergo inactivation, presence of X chromosome inactivation centre (XIC) or X chromosome control element (XCE) is important. X inactivation specific transcript (XIST) and its antisense transcript, Tsix (Plath et al. 2002) are transcribed from XIC. XIST is transcribed exclusively from the inactive X chromosome, coats the inactive X chromosome and has been shown to be an important component of the X inactivation machinery. X activation specific transcript (XACT), a long noncoding RNA (lncRNA) is seen to localize to active X chromosome in human pluripotent stem cells (Vallot et al. 2013). Other partners, namely noncoding RNAs RepA, JpX, FtX and PcG polycomb group of proteins, are shown to interact with Xist in establishing the X chromosome inactivation (Zhao et al. 2008; Tian et al. 2010; Soma et al. 2014). In mouse blastocyst, Xist is regulated by transacting protein factors, namely Oct4, Sox2, Nanog and Rex1 (Silva et al. 2009; Gontan et al. 2012). Rnf12/RLIM (Jonkers et al. 2009) which encode the ubiquitinase enzyme are also involved in the silencing of the X chromosome. In addition to silencing, Xist RNA has multiple roles in XCI such as spatial reorganization of the X chromosome and chromatin remodelling. As a consequence of Xist coating, a subnuclear ‘domain’ is created thereby silencing genes during X chromosome inactivation (Chaumeil et al. 2006). The Xist lncRNA is involved in multiple steps during X inactivation, including coating of the presumptive inactive X (Xi) chromosome, exclusion of RNA polymerase, reorganization of chromatin into inactive domains, methylation of DNA, localization of the inactive X to the nuclear periphery and packing of the chromatin (Lu et al. 2017). Another X-linked lncRNA Firre is involved in regulation of chromatin conformation, (CTCF) binding and methylation of histone H3K27me3 and helps in positioning of inactive X chromosome in the nucleus (Yang et al. 2015). Methylation of histone H3 at Lys-9 has been recognized as an early chromatin mark on the inactive X chromosome (Heard et al. 2001). Differential expression of miRNAs has been reported in cells with varying number of inactive X chromosomes (Rajpathak and Deobagkar 2017a). In XO Turner cases, these differentially expressed miRNAs appear to be participating in the epigenetic processes and are involved in various patho-physiological conditions observed in a Turner patient. These include aneuploidy, obesity, cancer, type-I diabetes, renal senescence, neural development and differentiation (Rajpathak and Deobagkar 2017b). Five differentially expressed lncRNAs in human X monosomy cells have been reported (Rajpathak et al. 2014). Further analysis (using DIANA LncBase V2) led to the identification of four (miR-10b-5p, miR-125a-5p, miR-4325 and miR-615-5p) miRNAs which can interact with lncRNA Xist and are differentially expressed in 45, XO cells. It has been suggested that some lncRNAs can act as molecular ‘sponges’ of miRNAs and titrate away the active miRNA thereby influencing the expression of genes.

DNA cytosine methylation is involved in X chromosome inactivation and early reports have shown that treatment with 5 azaC, a known demethylating agent, could reactivate the inactive X chromosome (Mohandas et al. 1980). DNMT1 knockout mice embryos also showed reactivation of X-linked genes (Sado et al. 2000). SmcHD1 (Blewitt et al. 2008) and alpha thalassaemia / mental retardation X-linked protein (Baumann and De La Fuente 2009) have been implicated in the maintenance of X chromosome inactivation. By employing photo acoustic spectroscopy the amount of DNA cytosine methylation was shown to increase linearly with the number of inactive X chromosomes in the human fibroblast cells with varying number of X chromosomes (namely XO, XX and XXX chromosomes) (Achwal et al. 1984; Deobagkar and Chandra 2003). There are reports of methylation of CpG islands on the inactive X chromosome (Pfeifer et al. 1990; Hellman and Chess 2007). The inactive X chromosome has more methylation in the repeats, transposons and LINE elements along with genic regions (Patil et al. 2014). A comprehensive map of tissue-specific pattern of gene expression for the X-linked genes has been recently reported (Tukiainen et al. 2016).

X chromosome has evolved sequence patterns which facilitate setting up of inactivation by the remodelling machinery. Thus sequences that are responsible for the inactivated state to spread may be enriched along the X chromosome and could be present as ‘way stations’ on the X chromosome (Riggs 1990). Such sequences could potentially be docking sites for a molecule like Xist that coats the inactive X chromosome and could be responsible for the maintenance and spread of inactivation. There have been studies using X autosome translocations and transgenes inserted into the X chromosome that report partial silencing or even escape from inactivation of autosomal regions (Lee and Jaenisch 1997). Along the X chromosome, motifs from L1 or LINE1 are enriched in sequences that undergo inactivation whereas they are absent from regions that escape inactivation (Wang et al. 2006). We have designed a sensitive high throughput microarray-based immunochemical approach to detect methylation of a gene or a region (Kelkar and Deobagkar 2009, 2010; Deobagkar et al. 2012; Rajpathak and Deobagkar 2014; Rajpathak et al. 2014) and examined the human diploid fibroblast cell lines with varying number of inactive X chromosomes, namely 45, X (no Xi), 46, XX (one Xi) and 47, XXX (two Xi). This has helped in generating the methylation map of the active and the inactive X chromosome in human (Rajpathak and Deobagkar 2014). X chromosome has been distributed into various strata based on the evolutionary origin of the sequences on the chromosome. When the locations of consistently methylated genes were examined most of them belonged to the S1, namely the earliest evolutionary strata, implicating that regions of X chromosome which were acquired early in evolution had more stable and consistent methylation pattern. It appears that sequences on the X chromosome have evolved differently from other chromosomes, so that the process of inactivation can identify these domains and thereby maintain stable inactive states (Kelkar and Deobagkar 2009; Kelkar et al. 2009). As many as 25% genes from the inactive X chromosome show partial or complete escape from inactivation and most of these genes localize on the recently acquired regions (Cotton et al. 2013, 2014; Disteche and Berletch 2015). Aneuploidy (missing the second X or Y chromosome) led to the misregulation of the epigenetic machinery and this altered methylation (BMP2, IGF1) correlated well with phenotypes of the XO Turner syndrome (Rajpathak and Deobagkar 2017a, b). These included not only the genes involved in setting epigenetic marks but also genes encoding phenotypes related to bone remodelling, growth, thyroid metabolism, glucose and sugar metabolism and sex differentiation, e.g. ovarian development. Since misregulation of epigenetic machinery appears to be important in establishing phenotypes of Turner syndrome, epigenetic, in particular methylation modulators, can have therapeutic potential for treatment of phenotypes of Turner syndrome. Novel methylation inhibitors have been designed which will find applications in the treatment of cases XO Turner, cancers and many other lifestyle diseases (Gawade et al. 2016; Joshi et al. 2016). Mutations or modulation in the methyltransferase machinery or its regulation have also been implicated in many human diseases (Hamidi et al. 2015).

Previously it has been reported that DNMT1-depletion led to global DNA hypomethylation and this destabilized the cells leading to aneuploidy (Barra et al. 2012). Both DNA methylation and expression of DNA methyltransferase 1 increased with the increase in the number of inactive X chromosomes in cells with increasing number of inactive X chromosomes (namely aneuploid 45, XO, 46, XX and 47, XXX) (Rajpathak and Deobagkar 2017a, b).

There are several interesting features of the epigenetic machinery and its regulation in relation to the X chromosome inactivation with respect to the mechanism of silencing, chromosome architecture, imprinting which will remain the subject of further analysis (Bonora and Disteche 2017).

DNA methylation in insects with particular reference to Drosophila

Until recently, insect DNA methylation has not been a subject of extensive research. Honey bees have evolved an intricate caste system and possess DNA methylation along with the entire enzymatic machinery, namely the DNMT1, DNMT3A and DNMT3B. Although recently discovered, the role of DNA methylation is implied to be important in gene expression and social behaviour in this eusocial insect (Elango et al. 2009). DNA methylation is reported to be within the coding sequence and may have some role in the caste system consisting of workers and the queen (Foret et al. 2012). It has been suggested that DNA methylation in honey bees affects life span (Cardoso-Júnior et al. 2017).

Fig. 2
figure 2

Comparison between D. melanogaster and mammalian methylation systems.

Full size image

Insects have therefore emerged as an interesting model to delineate novel roles for DNA methylation. In carpenter bees, which is a sub social insect, DNA methylation is targeted to its exons (Rehan et al. 2016). Queen pheromones have been shown to modulate DNA methyltransferase activity in bees and ant workers (Holman et al. 2016). It has been hypothesized that in Hymenoptera (bees, ants, wasps and sawflies), the evolution of eusocial division of labour is associated with DNA methylation in the genomes of Hymenoptera (bees, ants, wasps and sawflies). However, this has been questioned in a recent study (Glastad et al. 2017).

Drosophila is a model organism for genetic analysis of development and has been utilized extensively as a model system for human diseases, neural differentiation and evolving many basic concepts including the homeotic genes, developmental pattern formation, remodelling machinery. However, Drosophila is strikingly different with respect to DNA methylation machinery. It has been a topic of much debate from the 1980s with reports either claiming presence (Achwal et al. 1983, 1984; Gowher et al. 2000; Lyko et al. 2000; Phalke et al. 2009) or absence of DNA cytosine methylation (Urieli-Shoval et al. 1982; Patel and Gopinathan 1987; Raddatz et al. 2013). DNA methylation in Drosophila has remained enigmatic for several years. Drosophila along with other Dipterans, lacks DNA methyltransferases DNMT1/3A/3B and possesses only DNMT2 (Tang et al. 2003). DNMT2 was identified as a DNA/RNA methyltransferase (https://www.brenda-enzymes.org), particularly with reference to lower eukaryotes. DNMT2 knockout flies have reduced life span, compromised immune function and sensitivity to stress (Durdevic et al. 2013). Recently, two independent LC-MS/MS based studies have reported the presence of 5mC in the genome of adult D. melanogaster (between 0.01 and 0.034% of cytosine) (Capuano et al. 2014; Rasmussen et al. 2016). Presence of DNA cytosine methylation was confirmed in specific DNA sequences in Drosophila genome by selective enrichment of methylated DNA followed by bisulphite sequencing in stage 5 embryos (Takayama et al. 2014). Rasmussen et al. (2016) quantified 5mC in adult D. melanogaster and reported a lower 5mC level in fruit flies than in honey bees. Using methylation microarray-based technique; the changes in pattern of DNA methylation during embryonic, pupal and adult stages of D. melanogaster were identified (Panikar et al. 2017). Using a novel microarray approach, the presence of an active DNA methyltransferase was demonstrated in protein extract from S2 cells (Pannikar et al. 2004, 2015). Interestingly, presence of another epigenetic modification 6 methyladenine (6mA) was detected earlier in the genome of D. melanogaster (Achwal et al. 1983). Recently, by employing UHPLC-MRM-MS/MS technique, the presence of 6mA was reported in the embryonic stage of fruit flies (Zhang et al. 2015). This study also suggested that DNA 6mA demethylase, Drosophila TET homolog (DMAD), catalyses 6mA demethylation in vitro. Although 6mA has been reported to be present in Drosophila, there is no report on a putative DNA methyltransferase which can generate 6mA. In general, there are very few reports on presence of 6mA in eukaryotes (Heyn and Esteller 2015) and the enzymatic machinery remains to be identified or studied.

When the presence of methylation was analysed in the DNMT2 null mutants, 5mC was detected in the genomic DNA, albeit with an altered methylation pattern (Takayama et al. 2014). From this evidence, it appears that although dNMT2 may participate in the DNA methylation in Drosophila or may modulate the pattern, DNA methylation is present in the genomic DNA even in a DNMT2 null mutant. The search for a functional DNA methyltransferase in Drosophila genome and proteome needs to be carried out. It is interesting to note that Drosophila has been reported to have nonCpG methylation that is methylation in CpA or CpT dinucleotides (asymmetric) (Chatterjee et al. 2004; Takayama et al. 2014; Epigenetic regulation, stress and adaptation in Drosophila development, Deshmukh 2018, Ph.D. thesis). It will be interesting to unravel how this may be inherited across cell replication and development and differentiation. It can thus be concluded that although Drosophila has been employed as a model system for development, cancer, apoptosis etc., it shows distinct differences with respect to an important aspect of the epigenetic machinery (figure 2) and appears to manage very well without the DNMT1, 3A and 3B type of methylation. Our studies on methylation in Drosophila have led to the demonstration of changes in lipid metabolism along with a distinct suppression of immune function in both cellular and humoral arms associated with ageing in DNMT2 mutant flies (Epigenetic regulation of pathogenic stress and innate immunity genes, Abhyankar 2018, Ph.D. thesis). This could be due to the role of DNMT2 protein as an RNA methyltransferase or altered methylation in Drosophila. Patterns of methylation vary during development and life cycle stages (Epigenetic regulation, stress and adaptation in Drosophila development, Deshmukh 2018, Ph.D. thesis; Panikar et al. 2017). It will be very interesting to explore how the fruit fly compensates for the lack of methylation machinery. The DNA cytosine methylation present in Drosophila is sparse, asymmetric and has not been assigned any biological role. Drosophila appears to manage with the little methylation it possesses. Has Drosophila evolved alternate regulatory mechanisms to compensate for this loss? The search for the active DNA methyltransferase for both the adenine and cytosine methylation in Drosophila continues and further analysis is likely to reveal novel features of the fine tuning of the epigenetic machinery.

Summary

DNA methylation thus has a pivotal role in epigenetic processes. DNA methylation seems to have evolved in multicellular organisms to further enrich the messages encoded within the DNA sequence to add newer connotations and meaning. Nutrition and environment orchestrate phenotypes by interplaying with the basic genetic information and allowing subtle changes. It is hence important to unravel the signals which decide the exact locations of methylation marks and imprints. The powerful model organism fruit fly and the human have major differences in the DNA methylation machinery. A basic understanding of the molecular genetic mechanisms in adding the epigenetic marks and interpreting their meaning will throw further light onto the networks governed by readers and writers of epigenetic processes and help design better strategies for treatment of cancer and life style diseases.