N(6)-Methyladenine in eukaryotes

Abstract

DNA modifications are a major form of epigenetic regulation that eukaryotic cells utilize in concert with histone modifications. While much work has been done elucidating the role of 5-methylcytosine over the past several decades, only recently has it been recognized that N(6)-methyladenine (N⁶-mA) is present in quantifiable and biologically active levels in the DNA of eukaryotic cells. Unlike prokaryotes which utilize N⁶-mA to recognize “self” from “foreign” DNA, eukaryotes have been found to use N⁶-mA in varying ways, from regulating transposable elements to gene regulation in response to hypoxia and stress. In this review, we examine the current state of the N⁶-mA in research field, and the current understanding of the biochemical mechanisms which deposit and remove N⁶-mA from the eukaryotic genome.

Introduction

One of the mechanisms of epigenetic regulation is that of DNA modification, or the addition of covalent adjuncts to the extant base pairs that make up DNA. The most well-characterized DNA modification in living systems is 5-methylcytosine and its derivatives 5-hydroxymethylcytosine, 5-formylcytosine, and 5-carboxylcytosine (for a review on cytosine methylation see Cedar et al. [1]). Over the past several years, investigation of the modification of adenine in the form of N(6)-methyladenine (N⁶-mA) has increased due to its discovery in multicellular eukaryotes, specifically mammals, bringing with it a whole possibility of new functions [2–5]. The function of N⁶-mA in prokaryotes has been widely investigated in the decades since its discovery in bacteria in 1958 [6], where it has been most famously found to play numerous roles in foreign DNA recognition and restriction endonuclease protection [7, 8]. While the function of N⁶-mA remains to be fully elucidated, recent research into the presence and function of N⁶-mA, its writers and erasers, has provided several lines of evidence as to its important role in eukaryotic biology.

Discovery and function of N⁶-mA in prokaryotes

While searching for nucleotide structural analogues in E. coli in 1957, Dunn and Smith measured an abundant level of a novel modified base, N(6)-methyladenine (then named 6-methylaminopurine) [6]. Their subsequent research identified N⁶-mA in Aerobacter aerogenes, Mycobacterium tuberculosis as well as the bacteriophages T-even and Salmonella bacteriophage in the range of 0.5–2.7% [6]. The levels of N⁶-mA in the bacteriophages were lower than those of their corresponding bacterial hosts [6]. Further experimentation into the function of N⁶-mA in prokaryotes found that loss-of-function mutations to the various Dam prokaryotic DNA methyltransferases were not lethal to E. coli; however, the loss of N⁶-mA methylation did lead to a defective response to λ phage DNA, including a loss of restriction [7]. Investigation into the N⁶-mA Dam methyltransferase found that it acts upon a specific recognition sequence: 5′-GATC-3′ [9]. The palindromic nature of this sequence allows for methylation of adenine on both strands of the DNA in an equal nature allowing for identification of that DNA as “self” or “foreign” depending on the bacterial species in question and the class of restriction-modification systems utilized by the organism [10]. This sequence is then recognized by methylation-sensitive restriction enzymes which either digest (restrict) the DNA in question (i.e. DpnI) or non-methylated DNA (i.e. DpnII) [11, 12]. Additional functions of N⁶-mA in prokaryotes include a role in DNA mismatch repair by identification of the original strand [7, 13], regulation of gene expression in hemi/demethylated GATC promotor sites [14, 15], and preventing secondary initiation of genome replication by requiring dual-strand methylation at GATC sites at replication origins [16].

N⁶-mA in single-celled eukaryotes

While N⁶-mA was only recently discovered in metazoans, it has been recognized as the primary DNA modification in single-cell protists such as Tetrahymena thermophila and Tetrahymena pyriformis since 1973 [17, 18]. While methylation systems have been studied in great depth within prokaryotes, eukaryotes and metazoans do not contain obvious homologs to Dam methyltransferase or Dpn-like restriction enzymes, indicating that while the modification has the same resultant chemical structure of N(6)-methyladenine, the process by which that methylation occurs and is read is of a different evolutionary origin and serves a discrete function in the context of nucleated cells. As protozoan ciliates, T. thermophila and T. pyriformis exhibit nuclear dimorphism, that is, they contain both a reproductive micronucleus and a somatic macronucleus. The somatic macronucleus is generated by processing a copy of the micronucleus via DNA excision, polyploid replication, and substantial epigenetic modifications [19]. The differences between these nuclei are partially regulated by epigenetic factors including histone modification and DNA methylation, which makes it an important model organism in the investigation of novel epigenetic functions. In the initial papers outlining the presence of N⁶-mA in T. pyriformis, the authors noted that while N⁶-mA is present at roughly 0.75% of all adenine sites in the macronucleus, it is all but undetectable in the micronucleus [17]. They also identified that N⁶-mA in single-celled eukaryotes exists in a motif distinct from that of prokaryotes, with the only identifiable motif being 5′-AT-3′, indicating a Dam-like independent methyltransferase pathway [18]. Subsequent work in Tetrahymena indicated that N⁶-mA in T. pyriformis and T. thermophila is neither involved in the prokaryotic-like restriction-modification/DNA degradation pathways, nor is it utilized as a mismatch repair process due to its total absence in the micronucleus which is responsible for the reproduction of the organism [17, 20]. More recent research found that N⁶-mA is enriched at AT sites in the T. thermophila genome, indicating the presence of a methyltransferase due to the significantly increased levels of N⁶-mA at these sites versus non-AT adenines [20]. This site-specific sequencing was produced via SMRT-seq, which enables nucleotide resolution of DNA modifications (see below) [21]. Luo et al. identified a methyltransferase in T. thermophila they named TAMT-1, a gene containing the MTA70 methyltransferase domain and ZZ-type zinc finger DNA-binding domain [22]. In addition to nucleotide motifs, N⁶-mA was found to be enriched near nucleosomes containing the histone variant H2A.Z, a transcription-associated histone variant in T. thermophila.

In addition to T. thermophila,N⁶-mA has been discovered at quantifiable levels (4000 ppm) in Chlamydomonas reinhardtii, a green algae [23] and in numerous species of fungi [24]. Within green algae, N⁶-mA was present at 85% of all genes and was enriched at transcription start sites of genes undergoing transcription indicating a potential activating effect [23]. The deposition of N⁶-mA at active genes in single-celled eukaryotes is in contrast to the repressive role of N⁶-mA seen in most mammalian tissues (see below); however, it still unclear whether N⁶-mA can activate transcription based on recent studies [22]. Thus, further genetic studies are needed for investigate the role of N⁶-mA in transcription in simple eukaryotes [2, 25, 26].

The search for N⁶-mA in metazoa

While N⁶-mA had been identified in prokaryotes and single-celled eukaryotes for several decades, its identification in multicellular eukaryotes and metazoans proved to be a far more difficult task due to the limits of detection in the 1970’s to 80’s which could only detect N⁶-mA in concentrations > 1:10,000 [27]. The resurgence of metazoan N⁶-mA research in recent years has primarily been brought about by technological advances in base modification quantification that allows for independent lines of quantitative evidence such as antibody-based quantification (immunoblot), UPLC–MS/MS, and single-molecule real-time sequencing (SMRT-seq). With the maturation of these technologies, N⁶-mA was found to be present in the germ cells and embryos of D. melanogaster [3] and was discovered in C. elegans [28]. The discovery of N⁶-mA in D. melanogaster was accompanied by data indicating that the alteration of N⁶-mA through perturbation of the Drosophila-specific demethylase (DMAD) yielded altered germ cell differentiation [3]. Subsequent publications from other groups have found N⁶-mA in Mus musculus embryonic stem cells [2], Xenopus laevis testes, M. musculus kidney [29], Danio rerio (zebrafish) embryos [5], Sus domesticus (pig) embryos [5], M. musculus neuronal tissue [30], and H. sapiens glioblastoma [25]. These findings have all been confirmed using immunoblot/immunofluorescence imaging in tandem with UPLC–MS/MS, which reliably detected low-abundance N⁶-mA (several ppm of total adenines) to relatively high levels (1000 ppm) [25].

The 5-methylcytosine research community has benefited greatly from the comparatively high levels of 5mC found in the mammalian genome and the bisulfite conversion process which allows for base resolution sequencing of 5mC/5hmC deposition using Sanger and next-generation sequencing tools, although it is important to note that bisulfite sequencing cannot distinguish 5hmC from 5mC. No such chemical conversion process exists for N⁶-mA as of writing, thus studies analyzing the function of N⁶-mA through site-specific sequencing rely on DNA-immunoprecipitation followed by next-generation sequencing (DIP-seq) or SMRT-seq. The recent development of DIP-seq has significantly increased the resolution and sensitivity of N⁶-mA mapping [2, 3, 5, 26]. While SMRT-seq, which relies on single-molecule real-time sequencing to identify various DNA modifications via characteristic parameters of the DNA polymerase reaction, is a promising platform to recognize N⁶-mA at single-nucleotide resolution, it needs to be improved and optimized for recognizing N⁶-mA in eukaryotic genomes [21]. Current limitations to SMRT-seq include the need for high sequence coverage (25-fold for each strand), and the biologic finding that N⁶-mA deposition in eukaryotes is often heterogeneous within a cell population (which is in contrast to that of prokaryotes) [21]. Given high N⁶-mA levels in microorganisms, researchers should consider and rule out contamination as a possible source of N⁶-mA in their reactions. Sequencing approaches can unbiasedly evaluate the contribution from different genomes, which provides a screening tool to confirm the source of N⁶-mA peaks. Of note, re-inspection of the non-mammalian contribution in published mammalian datasets is trivial (0.0001–0.5%) [2, 25, 31], in contrast to a recent report [32]. Another concern of possible bias of for simple repeats was raised against DIP-seq including N⁶-mA and 5mC, which was attributed to non-specificity of IgG for these sequences. However, re-analyses of our published data from murine ESCs and human glioblastoma do not support this claim either [32]. Further improvement in sequencing approaches and bioinformatic analysis are needed for these intriguing yet challenging genomic elements.

Methylation and demethylation machinery in eukaryotes

Identification of N⁶-mA demethylases in multicellular eukaryotes

Several recent reports have demonstrated through genetic and in vitro biochemical experimentation that ALKBH1 is a DNA N⁶-mA demethylase in mammalian cells [2, 25, 31]. ALKBH1 is a homolog of the prokaryotic AlkB α-ketogluterate-dependent dioxygenase which is utilized by prokaryotes to remove DNA adducts that arise by DNA damage, but not N⁶-mA [33]. AlkB homologs are conserved through the tree of life; metazoans share seven AlkB homologues (ALKBH1, 2, 3, 4, 6, 7, 8) and vertebrates have evolved an additional two (ALKBH5 and FTO). AlkB dioxygenase enzymes function by hydroxylating the methyl group of N⁶-mA, generating an unstable hydroxymethyl group which is spontaneously released as a formaldehyde, resulting in direct demethylation of the base [33–36] (Fig. 1). The N⁶-mA demethylation process is rapid and complete, requiring only one oxygenation, with no downstream oxygenation products in contrast to the 5mC demethylation pathway in mammals (Fig. 1) [1]. Although N(6)-hydroxymethyladenine (N⁶-hmA) is chemically unstable, a recent study found its presence in mammalian genomes [37], which indicates that yet unknown protein factors may recognize N⁶-hmA and thereby potentially prevent it from spontaneous hydrolyzation by storage in a hydrophobic binding pocket [38]. Interestingly, N⁶-hmA has been reported in RNA as well as an oxidation product of FTO [38]. Thus, although no proteins that bind N⁶-hmA in DNA or RNA have yet been discovered, unknown pathways may be involved in the regulation of the completion of N⁶-mA demethylation.

Fig. 1.

N⁶-mA is demethylated by ALKBH1 by Fe(II)-mediated oxidation using α-ketoglutarate and O₂ to oxidize the methyl group. The intermediary N(6)-hydroxymethyladenine is inherently unstable, and rapidly decays to adenine and formaldehyde. 5mC is demethylated after several rounds of TET mediated oxidation followed by base excision repair

The discovery of multiple highly conserved AlkB homologs in multicellular eukaryotes provided in silico evidence that each of the homologs may have a unique catalytic reaction, either in DNA/RNA modification, DNA repair, or other nucleic acid modifications; however, reactions for all nine homologs in vertebrates have not been found [35]. The function of each of these individual enzymes has been studied, but in some cases remains unelucidated (see Table 1 and Fedeles et al. for an in depth review of AlkB homologs [34]). In M. musculus, the perturbation of ALKBH1 by truncation leads to altered placental/fetal development and skews sex ratios in the pups [2, 39, 40]. The authors noted several developmental phenotypes such as reduced birth weight/size, impaired trophoblast lineage differentiation, and eye/craniofacial defects [39]. Along with the discovery of N⁶-mA in embryonic stem cells in 2016, ALKBH1 was demonstrated to be a strong demethylase of DNA N⁶-mA in vitro [2] and further experimentation has confirmed that ALKBH1 is a strong demethylase of N⁶-mA in genomic DNA in both mice and humans [25, 31, 41]. It should be noted that ALKBH1 has been reported to function as a tRNA m1A demethylase, but only in the context of m1A58 in the stem-loop structure; abolishment of the stem-loop structure decreases the affinity of ALKBH1 for m1A [42] or 5mC in mitochondrial tRNA [43]. A recent study also identified ALKBH1 function in regulating N⁶-mA in mitochondrial DNA, which may explain previous observations of ALKBH1 functions in metabolism [26]. However, it is well known that ALKBH1 is primarily localized to the nucleus in human and mouse embryonic stem cells and neural systems [2, 25, 44]; thus, the physiological significance of ALKBH1 mitochondrial function which was mainly observed in cell culture remains unclear. How DNA N⁶-mA demethylation activity and tRNA activity contribute to ALKBH1 mitochondrial function has yet to be ascertained. It is also worth noting that ALKBH1 seems to function mainly as a genomic DNA demethylase in cells where ATP production mainly relies on glycolysis and not on mitochondria-based oxidative phosphorylation, such as in embryonic stem cells and cancer cells where ALKBH1 locates to the nucleus, but not mitochondria [44]. Thus, the function of ALKBH1 may be tissue/cell type dependent and further experimentation is needed to determine the pathways that delineate different aspects of ALKBH1 function.

Table 1.

AlkB homologs found in metazoans, their known substrates and known phenotypes on overexpression or depletion

AlkB homolog	Found in	Substrate	Known phenotypes	References
Alkbh1	Metazoans	DNA N6-mA demethylase, mitochondrial nucleic acid demethylase, putative histone lysine demethylase	Necessary for proper development of M. musculus	[2, 25, 31, 44, 45]
Alkbh2	Metazoans	Nuclear DNA repair enzyme (alkylation damage)	Knockdown impairs rDNA transcription and leads to increased ss/dsDNA breaks	[45–52]
Alkbh3	Metazoans	Nuclear DNA repair enzyme	Necessary for proper recovery from inflammatory response/DNA alkylation	[52–56]
Alkbh4	Metazoans	Actin lysine demethylase	Embryonic lethal, dysregulation of actin/myosin	[57, 58]
Alkbh5	Vertebrates	RNA m6A demethylase	Knockouts have decreased fertility (spermatogenesis-related)	[59–62]
Alkbh6	Metazoans	No available information	No available information
Alkbh7	Metazoans	Unknown, potentially proteins involved in necrosis	Knockouts have increased obesity	[63–65]
Alkbh8	Metazoans	tRNA maturation (methyltransferase and hydroxylation)	No observable phenotype. Complete loss of uridine modifications upon knockout	[66, 67]
FTO	Vertebrates	RNA m6A demethylase, 3-methylthymine/3-methyluracil DNA demethylase. Small nuclear RNA demethylase	Overexpression linked with obesity and diabetes	[60, 68, 69]

In D. melanogaster, DMAD (originally CG2083), the only  αKG-dependent dioxygenase homolog in the Tet/ALKB subfamily, was originally identified as a demethylase of N⁶-mA in early development, germline cells, and subsequently was confirmed in neurogenesis [3, 70]. Although one study also implicated this protein in demethylating 5mC in RNA, a recent biochemical and genetic study that investigated DMAD function in neurogenesis demonstrated that it serves as a genomic DNA N⁶-mA demethylase, with minimal/no impact on RNA 5mC levels [70]. Interestingly, the same elegant study of DMAD in neurogenesis also revealed that N⁶-mA directly recruits the polycomb complex which contributes to PcG/H3K27me-mediated epigenetic silencing in Drosophila, which is in contrast to the proposed activation function in the Drosophila germ line [70]. Consistently, a recent biochemical study demonstrated that N⁶-mA causes site-specific RNA polymerase pausing [71]. In light of these discoveries, it is of great importance to determine the functional significance of N⁶-mA depositions at actively transcribed genes in plants, simple eukaryotes and within the Drosophila germline. It will be of great interest to determine the yet unknown mechanisms by which N⁶-mA has tissue-dependent functions within the same species, such as Drosophila.

Putative N⁶-mA methyltransferases in eukaryotes

The identification of the predominant DNA N⁶-mA methyltransferase in eukaryotes is a vital step in both understanding the function of N⁶-mA in metazoans and obtaining important experimental tools to modulate N⁶-mA. While the methyltransferase in prokaryotes (Dam) has been known for decades, research into eukaryotic DNA N⁶-mA methyltransferases has yielded several putative enzymes with no definitive consensus on the primary methyltransferase. Experimental data such as SMRT-seq and N⁶-mA-IP-seq (DIP-Seq) in eukaryotes have demonstrated a lack of 5′-GATC-3′ motif enrichment, and thus it follows that the primary methyltransferase enzyme in metazoans may be quite different than Dam. The targeted search for active N⁶-mA methyltransferases in metazoans and mammals has relied on searching for highly conserved motifs and secondary structures found within all methyltransferases which utilize SAM (AdoMet) as a methyl donor in their methyltransferase reactions [72]. As of now, the effort of searching for eukaryotic N⁶-mA methyltransferases has only generated limited success. These types of screens have yielded the putative N⁶-mA DNA methyltransferase, TAMT-1 in T. thermophila [31]. Utilizing TAMT-1 knockouts in T. thermophila, Luo et al. found that TAMT-1 is responsible for around 1/3 the N⁶-mA deposition in the macronucleus [22]. These results coupled with the modest reduction in N⁶-mA in T. thermophila indicates that while TAMT-1 and its homolog METTL4 may have DNA methyltransferase function, it is likely not the sole DNA N⁶-mA methyltransferase in single-celled eukaryotes. It is also of interest to determine the role of N⁶-mA in transcription with TAMT-1 knockouts in future studies. TAMT-1 is a member of the MTA70 family of methyltransferases, and its homologs in C. elegans are DAMT-1 and in humans METTL4 [22]. DAMT-1 is a putative methyltransferase evolved from the MTA-70 family of methyltransferases within C. elegans, and the mutation of specific amino acids in DAMT-1 in catalytic pockets and substrate recognition pockets is able to reduce N⁶-mA deposition on the C. elegans genome in vivo, whereas the expression of DAMT-1 in SF9 insect cells led to an increase in N⁶-mA [28]. Researchers have yet been unable to isolate DAMT-1 in cell culture in quantities necessary in vitro to test its N⁶-mA methylation capacity [28].

A second possible candidate N⁶-mA DNA methyltransferase is N6AMT-1. A recent study demonstrated in vitro via immunoblot and LC/MS–MS that N6AMT-1 has catalytic activity methylating adenine containing oligos to N⁶-mA [31]. Additionally, the authors demonstrated via siRNA knockdown that reducing N6AMT-1 reduces N⁶-mA levels in vitro by roughly 50% in human cancer cell lines or more recently in neurons under stress [30, 31]. However, a previous report using similar biochemical reaction conditions did not observe such activities [73]. The latest attempt to replicate N6AMT-1 activity in vitro or in human glioblastoma cells has also been unsuccessful [25], and additional experimentation and investigation via protein purification and substrate specific assay conditions is called for.

The function of N⁶-mA in single-celled eukaryotes and invertebrates

In prokaryotes, it has been well established that the role of N⁶-mA DNA modification is to serve as a signal to identify “self” DNA from “foreign” DNA. N⁶-mA plays a critical role in the restriction-modification system, and in simple single-celled organisms provides an elegant way to defend against the constant threat of phages and other forms of DNA transfer that would harm the organism. In eukaryotes, especially higher metazoans which contain their own immune system, and thus do not utilize the restriction-modification system of organismal defense, this frees up N⁶-mA as an additional modification to be used for alternative purposes. Evidence from studies performed in T. thermophila indicates that N⁶-mA has been adopted as an epigenetic regulator of the macronucleus in several different ways. Recent work has elucidated that N⁶-mA is anti-correlated with nucleosomes in T. thermophila, indicating an epigenetic effect inverse to that of 5mC in eukaryotes [20, 22]. Indeed, in an entirely in vitro reconstituted experiment using isolated histones and synthesized strands of DNA with a single N⁶-mA residue, methylation or demethylation of the single residue was able to alter nucleosome positioning indicating a potential function of N⁶-mA in fine-tuning nucleosome placement in a highly sensitive manner. Experimentation in T. thermophila using high-resolution SMRT-seq identified a significant enrichment of N⁶-mA in the genic region of the genome, with specific enrichment near the 5′ end of the gene body in actively transcribed genes [20]. It will be of interest to see whether this function is also conversed in metazoans.

Function of N⁶-mA in mammals

Over the past 3 years, several labs have started exploring the function of N⁶-mA in eutherians, utilizing mouse and human samples in their studies. Given the novelty of the discovery of N⁶-mA in eutherians, our knowledge of the function of N⁶-mA in higher organisms is limited, and we expect greater elucidation into the normal function of N⁶-mA and its role in disease states over the next decade.

N⁶-mA in embryonic stem cells

The initial paper discovering the presence of N⁶-mA in eutherians identified not only the demethylase responsible for demethylation in murine cells, but also a significant repressive effect of N⁶-mA on its deposition sites within embryonic stem cells, especially on young LINE-1 (long interspersed nuclear element) transposons [2]. Via DIP-seq and SMRT-seq, Wu et al. identified significant enrichment of N⁶-mA at young and middle-aged LINE-1 transposons [2]. Transposable elements are a class of genetic elements which have (or had) the ability to copy and insert themselves to a new region of the genome utilizing their own sequence [74]. Transposable elements comprise a large portion of the mammalian genome, roughly 50% in mouse and human, and display lower diversity within eutherians than other animals with similarly sized genomes and families [74, 75]. Retrotransposable elements, such as LINE, have retained their ability to insert themselves back into the genome, and in eutherians are thought to be suppressed by epigenetic mechanisms. In embryonic stem cells, however, N⁶-mA is significantly correlated with epigenetic silencing of LINE-1, but specifically only those with intact reading frames and promotors, in other words elements that pose an active threat to the organism if they reactivate [2]. N⁶-mA is enriched towards the 5′UTR and ORF1 of LINE1 elements in murine tissues, and studies performed in human tissues have recapitulated this result, reinforcing the hypothesis that N⁶-mA is a strong transcriptional repression mark in metazons [2, 26].

The examination of N⁶-mA in single-celled eukaryotes and mammals has identified two functions thus far of N⁶-mA, as being enriched at transcription start sites and actively transcribed gene bodies mostly in single-celled eukaryotes and the Drosophila germline, and as a repressive mark in eutherians and probably other metazoans [2, 20, 23, 30, 70, 76]. While these roles of N⁶-mA appear diametrically opposed, the function of epigenetic marks often is often dependent on numerous factors including stage of development, organism, and disease state, thus the role of N⁶-mA in eukaryotic biology is far from fully elucidated. It is possible that N⁶-mA has adopted a new, repressive function in metazoans, but several alternative explanations may also exist. Given the difference in genome size between mammals and single-cell eukaryotes, and the sensitivity of the current sequencing techniques which have been used to examine site-specific depositions, it is possible that an enrichment at transcription start sites within actively transcribed genes has escaped observation in multicellular eukaryotes due to low abundance at single-cell level and/or cell-to-cell heterogeneity [71]. Biochemical experiments have identified N⁶-mA deposition in transcription sites as sufficient to cause RNA PolII pausing/stalling [71], which provides an alternative explanation for the enrichment of N⁶-mA at actively transcribed genes. Further genetic studies in single-cellular eukaryotes and human/murine cell culture models examining the impact of increased N⁶-mA methylation on RNA PolII activity will be critical in the synthesis of these discreet findings.

N⁶-mA has also been found to dynamically alter cell fate decisions during ESC differentiation, one example where the genomic and phenotypic data line up is in placental development. Murine studies using ALKBH1-deficient mice found placental and developmental defects in ALKBH1-knockout mice, and embryonic stem cell ALKBH1-knockout cell lines, which have significantly increased N⁶-mA deposition, have altered differentiation factor expression including but not limited to CDX2, Lefty1, Foxa2, Gata4 and Gata6 [2, 39, 40]. These data strongly implicate N⁶-mA as playing a role in regulating embryonic and extraembryonic lineage development and cell potency.

N⁶-mA as an epigenetic mark in response to stress signals

In addition to its role in downregulating retrotransposon expression in developing stem cells, N⁶-mA has been found to be enriched in other genes that are downregulated in the mouse, specifically in neuronal tissue of animals undergoing a stress response [30, 76]. While there was no significant overlap between N⁶-mA deposition and genic regions in murine embryonic stem cells, N⁶-mA has been found to deposit on genes in the developing brain. Yao et al. examined global dynamic changes to N⁶-mA deposition in 8-week-old male mice exposed to two hours per day of stress for 14 days, which elicited a chronic stress response in the experimental cohort [30]. Using this technique, they then examined N⁶-mA levels in various parts of the brain using UHPLC–MS/MS, and found an enrichment of N⁶-mA over adenine in the pre-frontal cortex of the experimental animals versus the control animals, which was confirmed by dot blot and explored in a site-specific manner via N⁶-mA-IP-seq [30]. Their initial results confirm the finding that N⁶-mA is recruited to retrotransposons and intergenic regions. They found that during the stress-response-mediated aberrant methylation LINE sites are significantly downregulated due to N⁶-mA hypermethylation. Intriguingly they found that the remaining N⁶-mA enrichment was heavily deposited in intragenic sites (33%) versus exons (0.3%) [30]. Yao et al. found a large number of differentially methylated sites but determined that hypomethylated sites are located at the TSS and TES, whereas hypermethylated sites are located in the gene body between those sites. They additionally determined that N⁶-mA deposition is enriched at genes associated with psychiatric disease-correlated loci (depression, schizophrenia and autism spectrum disorder) which indicates that N⁶-mA plays a critical role in epigenetic silencing in the pre-frontal cortex’s stress response pathway [30]. These were some of the first data in eutherians to indicate that N⁶-mA plays a directed epigenetic role in stimuli response, and since then this pattern has been seen in neoplasms as well as healthy tissues.

N⁶-mA in human tumors

While Yao et al. examined stressed but otherwise healthy neuronal tissue in a murine model, recent studies from the Xiao and Rich labs examined the role of both N⁶-mA and ALKBH1 in the cellular processes and prognosis for adult human patients with glioblastoma [25]. Glioblastomas are highly invasive neoplasms that drive from glial tissue in the brain, where alterations to global and gene-specific levels of DNA modifications such as 5mC and 5hmC have been previously observed. Xie et al. found that N⁶-mA is dramatically (500-fold) upregulated in glioma stem cell lines (cells that can give rise to remote neoplasms in a mouse model) in vitro and in primary glioblastoma cells directly from patients. These findings were relatively uniform across samples from 67 patients, indicating a substantial aberrant regulation of N⁶-mA during the carcinogenesis and proliferative phase of the cancer. Xie et al. also interrogated pathways regulated by ALKBH1 in human glioblastoma cells by knocking down ALKBH1 and discovered a significant enrichment of N⁶-mA near key oncogenic pathways characterized in glioblastoma carcinogenesis, including hypoxia response pathways. Glioblastoma is a fast-growing tumor that struggles to maintain adequate vascularization to the entire tumor, leading to hypoxic necrosis of the core of the tumor. Upregulation of genes that help a cell adjust to these hypoxic environments are a hallmark of glioma stem cells [25]. Previous research had investigated the role of chromatin modification in the form of histone modifications and had linked histone modification H3K9me3 with the repression of hypoxia-induced genes; in their analysis of N⁶-mA deposition sites, Xie et al. found that 83% of genes associated with H3K9me3 hypoxia-induced sites had an increase in N⁶-mA after ALKBH1 knockout [25]. These data indicate that in the case of glioblastoma, the cancer cells have may have co-opted N⁶-mA as a regulatory marker and use ALKBH1 to upregulate the expression of hypoxia-induced genes, allowing them to proliferate in the severely hypoxic conditions within the tumor. These findings coupled with the relative lack of N⁶-mA signaling in a majority of sampled tissues [2, 27, 29] indicate that the N⁶-mA regulatory network may be an ideal therapeutic target for glioblastoma and other cancers with aberrant N⁶-mA/ALKBH1/N⁶-mA methyltransferase activity [25]. While this work provides novel insights into the potential function of N⁶-mA in cancers, further research is necessary to confirm these results in other tumor types and patient cohorts.

While the discovery of N(6)-methyladenine dates back to 1957, our understanding of this necessary DNA modification in eukaryotes is in its infancy in comparison to other DNA modifications such as 5mC and its derivatives. In general, N⁶-mA in mammalian genomes seems to be sensitive to various environmental signals, such as fear conditioning and hypoxia, which indicate that this modification may quickly respond to environmental cues and plays vital regulatory functions. For example, a study reported that when ES cells are cultured under chemical inhibitors against MAPK and GSK3B, N⁶-mA cannot be easily detected, in contrast to the standard serum/LIF medium [77, 78]. This condition, which is often referred as “2i”, greatly compromises ESC pluripotency when used in long-term culture and triggers global DNA and histone demethylation [78, 79]. This example underscores the importance of understanding the regulatory pathways of N⁶-mA and complexity of mammalian epigenetic regulation.

Future perspectives

N⁶-mA has emerged as a potent regulator of gene expression, chromatin configuration, stem cell potency, stress response, tumorigenesis, and retrotransposon suppression. With renewed interest in this field, our understanding of the role N⁶-mA plays in metazoan and eutherian biology grows weekly providing us new answers in the field of epigenetics and opening new avenues of scientific inquiry with the promise of potential therapeutic interventions. The research performed into the existence and function of N⁶-mA in eukaryotes is in its early stages, and additional research into the machineries involved as writers (methyltransferases), readers (binding proteins), and erasers (demethylases) of N⁶-mA is necessary to further our knowledge of the function N⁶-mA. One major challenge of N⁶-mA research is its low abundance in a population of cells. Immunofluorescence and SMRT-seq studies implicated cell-to-cell heterogeneity as a contributing factor and, therefore, developing novel approach for enriching the N⁶-mA presenting subpopulation would greatly facilitate this line of investigation. Another key hurdle to investigating N⁶-mA is the difficulty in quantifying and mapping N⁶-mA deposition in low numbers of cells. We encourage research into enhancing the specificity of existing tools such as SMRT-seq and DIP-seq, as these technical limitations currently limit research to samples with abundant levels of N⁶-mA or large numbers of cells.