Insights & Perspectives Think again
Now you see it: Genome methylation makes a comeback in Drosophila Dario Boffelli*, Sachiko Takayama and David I. K. Martin* Drosophila melanogaster is often considered to lack genomic 5-methylcytosine (m5C), an opinion reinforced by two whole genome bisulfite-sequencing studies that failed to find m5C. New evidence, however, indicates that genomic methylation is indeed present in the fly, albeit in small quantities and in unusual patterns. At embryonic stage 5, m5C occurs in short strand-specific regions that cover 1% of the genome, at tissue levels suggesting a distribution restricted to a subset of nuclei. Its function is not obvious, but methylation in subsets of nuclei would obscure functional associations since transcript levels and epigenetic modifications are assayed in whole embryos. Surprisingly, Mt2, the fly’s only candidate DNA methyltransferase, is not necessary for the observed methylation. Full evaluation of the functions of genome methylation in Drosophila must await discovery and experimental inactivation of the DNA methyltransferase, as well as a better understanding of the pattern and developmental regulation of genomic m5C.
.
Keywords: development; Drosophila; methylation
The functions of DNA methylation are incompletely understood 5-Methylcytosine (m5C) is very widely distributed in eukaryotic organisms [1], where it serves as a tag that can be recognized and read by proteins that act at or near the methylated site in a genome [2, 3]. It has been most inten-
sively studied in vertebrates, where it typically occurs as a symmetrical modification of CpG dinucleotides; when it occurs in regions of high CpG density (“CpG islands”) m5C is a central component in systems of epigenetic regulation that suppress transcription initiation in both animals and plants [4–6]. While the prominence of CpG island methylation in plant and vertebrate epigenetics might give the impression that the function of
m5C is well understood, the broad scope of methylation within genomes and among species has left many open questions [7]. m5C occurs in multiple distinct patterns: in vertebrates, it is found over the whole genome except for CpG islands; in other species it is targeted to transposable elements, or to gene bodies; in yet other species it is excluded from gene bodies [1, 8, 9]. No general theme is evident, and in most cases the functional significance of these patterns has not been established [1, 10]. Even in the vertebrates, where m5C has been most intensively studied, there are many unknowns. CpGs that occur outside of CpG islands are usually methylated, but except for silencing of retroelements [11, 12] the function (or more probably, functions) of this methylation is unclear [1, 9]. Also, it has been known for decades that human cancer genomes display widespread loss of methylation [13], but it is not understood why this happens or what its consequences for the malignant phenotype are. Non-CpG methylation is common in plants [1, 14], and has been observed in mammals [15], but its functions are still the subject of investigation in plants [16, 17] and are unknown in mammals. Given the variety of m5C patterns that have been observed, it seems likely that it serves multiple functions.
DOI 10.1002/bies.201400097 Children’s Hospital Oakland Research Institute, Oakland, CA, USA *Corresponding authors: Dario Boffelli E-mail: [email protected] David I. K. Martin E-mail: [email protected]
Abbreviations: bp, base pairs; DNMT, DNA methyltransferase; hm5C, 5-hydroxymethylcytosine; m5C, 5-methylcytosine; MBD, methyl binding domain; MeDIP, methylcytosine immunoprecipitation; TET, teneleven translocation.
Bioessays 36: 0000–0000, ß 2014 WILEY Periodicals, Inc.
DNA methyltransferases establish and maintain CpG methylation A class of related DNA methyltransferases (DNMTs) establish and maintain
www.bioessays-journal.com
1
Think again
D. Boffelli et al.
genomic m5C in animals [18]. The genetic and biochemical characterization of the DNMTs has been carried out mostly in human and mouse, and function in other species has been inferred by the presence of a conserved DNA sequence in the catalytic domains [18]. DNMT1 is a “maintenance” methyltransferase, acting to complete methylation of hemimethylated CpGs [18]. Two members of the DNMT3 family function as de novo methyltransferases [18]. DNMT2, while related to the other methyltransferases, is not clearly a DNA methyltransferase: it can methylate RNA, and this may be its principal function [19, 20]. The importance of m5C for animal and plant development has been established by experimental inactivation of DNMT1 [21], DNMT3a [22], and DNMT3b [22], which results in hypomethylation, with broad effects on development, gene expression, and the suppression of retrotransposons. DNMT1 deficiency gives rise to several neurodevelopmental defects [23, 24], while DNMT3b deficiency has been associated with ICF1 (immunodeficiency-centromeric instability-facial anomalies syndrome-1; OMIM 242860) [22].
Drosophila has a DNA methylation system Drosophila melanogaster once served as an instructive example of a species that carries out complex gene regulation and cell differentiation without cytosine methylation [25, 26]. The fly does not have any apparent homologue of DNMT1 or DNMT3 [18], and m5C had not been detected in the fly genome [25, 26]. The discovery of a fly DNMT2 homologue (Mt2) [27, 28] prompted searches for m5C using more sensitive methods: these studies produced evidence of extremely small quantities of m5C, most abundant in the early embryo, and in the context of CA and CT dinucleotides rather than CpG [29– 31]. The methods could detect the presence of m5C in bulk, but could not establish where in the genome it was located. Immunostaining for m5C in fly embryos revealed a nuclear staining pattern that was weaker when Mt2 was knocked down, and bisulfite sequencing of random genomic fragments 2
Insights & Perspectives
showed apparent methylation in CA and CT contexts [32]. Overexpression of mouse DNMT1 and DNMT3 in Drosophila induced m5C and resulted in developmental defects and lethality [33], prompting speculation that the fly expresses proteins that can interact with genomic m5C and produce phenotypic effects (which were abnormal presumably because the induced methylation was ectopic). One of these proteins may be the methyl binding domain protein MBD2/3 [28] (FlyBase ID CG8208). In contrast to vertebrate MBDs, Drosophila MBD2/3 binds DNA with a preference for CA/ CT contexts [34], consistent with evidence from nearest neighbor analysis of fly genomic methylcytosine [29], and with our recent findings discussed below [35]. MBD2/3 also co-fractionates with components of the NuRD complex [36], suggesting a model in which it binds DNA and recruits chromatinremodeling complexes. Although this MBD may have other functions, its ability to bind methylcytosine is most easily explained by the presence and functional significance of genomic methylcytosine. It is expressed ubiquitously throughout fly development [37], suggesting a similarly broad function for m5C. Another MBD, MBD-R2, is part of the non-specific lethal (NSL) complex and contains a CpG binding domain [38], but its ability to bind methylated DNA has not been demonstrated. Further circumstantial evidence is provided by the presence in Drosophila of dTet (FlyBase ID CG43444). dTet is a homologue of the ten-eleven translocation (TET) protein [39], a methylcytosine oxidase which in mammals is responsible for metabolizing m5C to 5-hydroxymethylcytosine (hm5C), the initial step in the pathway that eventually leads to demethylation [40, 41]. dTet has an intact CXXC CpG-binding domain and HXD catalytic site, strongly implying that it is functional in the same way as mammalian TET proteins [39]. Dunwell et al. [39] argue that the domain of dTet that binds methylated DNA is not restricted to CpG, and that the conservation of TET function in the fly strongly implies that methylcytosine is present on the genome in some tissue or tissues. dTet is expressed from embryonic stage 7/8, predominantly in neural tissues [37], and so could act to
.....
demethylate or create hm5C from m5C laid down at earlier stages.
The presence of genomic methylation in Drosophila has been controversial The extremely low levels of m5C detected, even in the early embryo, seemed incompatible with cytosine methylation having a role in gene regulation and epigenetic phenomena similar to that seen in mammals. The issue of function was further clouded because the methods originally used to detect m5C cannot establish where in the fly genome it is located. Later developments cast further doubts on the biological significance of m5C in the fly. Depletion or mutation of Mt2 has only minor phenotypic effects [20, 32], similarly to the mouse DNMT2 knockout, which is also viable and normal [19, 42]. Inactivation of Mt2 suppresses variegated silencing of white in Drosophila [43], but a claimed link to retrotransposon methylation was refuted [44, 45]. Evidence that Mt2 methylates tRNA suggested an alternative function for the protein, albeit without showing that the function is important [19, 20, 46]. This alternative function for Mt2 cast doubt on the evidence for genomic m5C: if the only DNMT in the fly were in fact an RNA methyltransferase, then genomic methylation might be, as some have described it, “spurious” [44, 47]. Most recently, two whole genome bisulfite sequencing studies concluded that no genomic m5C is present in the fly; one study sequenced the fly embryo genome at an average depth of 6 [8], and the other at 32 [47]. Although unconverted cytosines (indicative of methylcytosine) were detected in the deeper study, they were also present in DNA derived from Mt2 mutants, leading the authors to conclude that the unconverted cytosine was an artifact of incomplete bisulfite conversion. These observations conflict with the earlier evidence for genomic methylation in the fly, and have convinced even one of its discoverers that the evidence was artifactual. Nevertheless the evidence from whole genome bisulfite sequencing does not demonstrate that the earlier results obtained by bulk analysis were wrong.
Bioessays 36: 0000–0000, ß 2014 WILEY Periodicals, Inc.
.....
Insights & Perspectives
New evidence for genome methylation in the fly
bisulfite sequencing. We chose 66 methylated regions for resequencing after bisulfite conversion; sequencing was carried out at extremely high depth to obtain accurate measures of the level at each cytosine. This method validated the presence of methylation at all of the 66 regions, with peaks in roughly the same sites as the peaks detected by MeDIPBseq [35]. However, methylation at these sites is never complete: even the most highly methylated positions are methylated on only 50% of genomes in the stage 5 embryo, and most methylated positions are methylated in 5,000 might be required to detect these methylated regions; enrichment, such as with MeDIP, would reduce the amount of sequencing required to detect them by an unknown factor. Such a low frequency of methylation would be difficult to distinguish from an artifact of bisulfite non-conversion. Our study indicated an 1% nonconversion rate in control DNA using a commercial conversion kit, so that methylation would need to be present in a much higher proportion to be detectable without enrichment. However, if bisulfite non-conversion is randomly distributed (still a murky subject [50]), then the presence of consistent allelic patterns of methylation could be used to detect methylated alleles at much lower frequency, particularly if they occurred in a single characteristic sequence context such as CpG (our data suggest that CG methylation might be present at extremely low frequency – see Fig. 1).
.....
The function of genome methylation in the fly remains enigmatic While the evidence indicates that the fly genome does possess methylation, the function of this methylation is not at all clear. The methylation pattern we have observed is completely unlike that known in any other species; thus no inferences can be made by comparison to other systems. Indeed the strongly focal pattern, with methylation concentrated in small areas characterized by an extended sequence motif, has been observed only in the fly, and it suggests that methylation could have functions that are unique to the fly (although it should be noted that 5mC in the quantity and pattern that we observe might be present in other species that have more abundant 5mC, but its relative rarity would obscure its presence). The distribution of methylation in the fly genome does not produce any obvious hints as to its function: we found no association with several histone modifications that mediate gene regulation, and no relationship with gene regulatory elements or gene bodies. Because Mt2 has been eliminated as the factor responsible for genome methylation, and there is no alternative candidate, the option of assessing a mutant lacking DNA methylation is currently unavailable. We found a small but significant association of gene body methylation with lower expression levels of developmental genes at stage 5 [35]. This suggests the possibility that genome methylation in the fly is – as in many systems – involved in gene regulation; however, much methylation is located at a distance from genes, suggesting that it could have other functions. The weak association of methylation with gene expression, taken together with the presence of methylation in only a small subset of genomes at stage 5, is consistent with a scenario in which methylation is heterocellular (or more properly heteronuclear) at stage 5. This could imply that its effects on gene expression in specific nuclei are larger than those observed over the entire embryo. Most of the methylated regions we have identified are methylated on less than 10% of sequenced DNA strands. We have noted a tendency for
Bioessays 36: 0000–0000, ß 2014 WILEY Periodicals, Inc.
.....
Insights & Perspectives
methylation at specific sites. One recent study detected methylcytosine in adult flies at 0.03% of all cytosine, consistent with the presence of methylated regions in a small but not insignificant subset of cells [48]. Dunwell et al. [39] have pointed out that dTet is expressed in fly neural tissues, and also functions to generate hm5C in the mammalian brain. On this basis they speculate that methylcytosine, or hydroxymethylcytosine, has a functional role in those tissues. The small amounts of methylcytosine detected in adult flies [48] might reflect this. The precise location of such methylation, as experience to date demonstrates, could be very difficult to establish: even at stage 5, where methylcytosine is most abundant, localization required enrichment and deep sequencing; if it is present in lower proportions in other tissues, even our method might not suffice to detect it. One of the most baffling aspects of the picture produced by our analysis of the fly methylome is the non-dependence of methylation on Mt2, the fly DNMT2 homologue. This was an unexpected result: we analyzed an Mt2
strain as a means of buttressing the existence of genome methylation, and when we found methylation retained in one mutant strain we analyzed a second mutant. In both mutant strains, we found methylation at stage 5 at sites and in amounts similar to wild type strains. The result clearly indicates that Drosophila must possess some other factor capable of methylating cytosine. Since no other DNA methyltransferase homologue is apparent in the sequenced portions of the fly genome, the factor must either reside in some unsequenced region, or lack sufficient similarity to other DNMTs to make it recognizable by protein domain similarity searches. We considered the possibility that the genome methylation we observe is derived from a DNA methyltransferase expressed by one of the several Drosophila endoparasites, such as Wolbachia. However, we did not find any reduction in methylation after three generations of treatment with tetracycline (unpublished result); while this result does not eliminate the possibility of an endoparasite origin, it strongly argues against it. Thus Drosophila apparently carries a cryptic DNA methyltransferase of a novel type.
Conclusions and outlook
Figure 2. Allelic pattern of m5C in a single methylated region revealed by deep bisulfite resequencing. The coordinates of the resequenced region are shown at the bottom. Stage 5 genomic DNA was bisulfite converted, PCR amplified, and subjected to paired-end Illumina sequencing to extreme depth. To simplify visualization of the pattern, only the cytosine positions in selected reads are shown (black: methylated; gray: unmethylated). The thin gray lines visible in some reads denote the unsequenced portion of a fragment (between paired ends). Methylation is concentrated in a subset of the sequenced alleles, consistent with a subset of stage 5 nuclei being methylated in that region.
Bioessays 36: 0000–0000, ß 2014 WILEY Periodicals, Inc.
Multiple lines of evidence support the presence of DNA methylation in Drosophila (Box 1). Drosophila possesses proteins that are capable of interacting with genomic m5C [34, 39]. Bulk assays have consistently found small amounts of m5C, most abundant at stage 5 but also present at later stages of development [29, 30, 48], and our recent study has located foci of genome methylation at stage 5 [35]. This evidence indicates that genomic methylation is present, but the pattern is unique, and provides no obvious clues to its function. The role of DNA methylation in Drosophila remains subject to speculation, as is the case in many other species. Nevertheless, the presence of m5C in this important model organism, along with conserved proteins that can mediate its function, resurrects genome methylation as a subject for study and presents the challenge of explaining how it is established and what it does. 5
Think again
methylation to be concentrated in a subset of sequence reads, indicating an allelic pattern to methylation (Fig. 2). This may indicate that all methylation is concentrated in a subset of embryonic nuclei, or that all nuclei have some methylation, but at different subsets of the regions we identified: our study is unable to distinguish these possibilities, because of the short length of the sequence reads (Fig. 3). A technology capable of direct detection of methylcytosine on long-read sequences could resolve this question [51, 52]. A heterocellular pattern is entirely consistent with the differentiation of epigenomes that takes place during development: if m5C functions in gene regulation, its precise positioning would be expected to vary according to cell type or fate. Such variation occurs on a grand scale in vertebrates; in the fly we find tens of thousands of methylated regions, and each of these potentially represents a source of functional differentiation. Heterocellular methylation patterns might extend well beyond the embryonic period, but the increased number and complexity of cells would make it much more difficult to detect
D. Boffelli et al.
Insights & Perspectives
.....
Think again
D. Boffelli et al.
Figure 3. Schematic illustrating hypothetical distribution patterns of m5C in the Drosophila embryo. Two methylated regions are shown at the top in blue and red. In model 1, a subset of nuclei is methylated at Region A but not region B, while a different subset is methylated at Region B but not region A. In model 2, both regions are methylated in a subset of nuclei. In model 3, methylation of the two regions is diffusely distributed throughout the embryo, so that many or all nuclei have a little methylation in both of the regions. The three models describe distinctly different patterns. However, when methylation is assayed over the entire embryo (as in all studies to date) the patterns are indistinguishable.
Box 1 Experimental evidence for DNA methylation in Drosophila Methylated cytosine detected in the fly genome with a method capable of distinguishing DNA and RNA methylation.
The fly genome encodes proteins capable of interacting with methylated cytosines.
Immunoprecipitation of methylated cytosines coupled with bisulfite
sequencing shows that 1% of the fly genome at stage 5 is methylated at low levels. Targeted bisulfite-PCR resequencing provides quantitative information on methylation levels and shows that methylation is dynamic in early development.
The biological significance of fly genomic methylation could be clarified by identifying the responsible DNA methyltransferase, and observing the phenotypic effects of its inactivation. However, no DNMT homologue other than Mt2 is detectable in the fly genome sequence by computational methods. The possibility that such a homologue resides in an unsequenced portion of the fly genome becomes increasingly unlikely, as more sequence is available; we are left with the possibility that the responsible methyltransferase does not resemble the other DNMTs. In that case, identification might require purely functional screens based on detection of methylation at some of the regions we have identified. 6
As discussed above, there remains a distinct possibility that other patterns of genomic m5C are present, either in rare embryonic cells or at later stages of fly development. Discovery of these patterns will require more sequencing, albeit possibly at intimidating depths or in carefully selected subsets of cells or nuclei. The outcomes of such endeavors are at present unknowable, and await the enterprising and adventurous investigator.
References 1. Suzuki MM, Bird A. 2008. DNA methylation landscapes: provocative insights from epigenomics. Nat Rev Genet 9: 465–76. 2. Klose RJ, Bird AP. 2006. Genomic DNA methylation: the mark and its mediators. Trends Biochem Sci 31: 89–97.
3. Bogdanovic´ O, Veenstra GJC. 2009. DNA methylation and methyl-CpG binding proteins: developmental requirements and function. Chromosoma 118: 549–65. 4. Deaton AM, Bird A. 2011. CpG islands and the regulation of transcription. Genes Dev 25: 1010–22. 5. Bird AP, Wolffe AP. 1999. Methylationinduced repression – belts, braces, and chromatin. Cell 99: 451–4. 6. Smith ZD, Meissner A. 2013. DNA methylation: roles in mammalian development. Nat Rev Genet 14: 204–20. 7. Reddington JP, Pennings S, Meehan RR. 2013. Non-canonical functions of the DNA methylome in gene regulation. Biochem J 451: 13–23. 8. Zemach A, McDaniel IE, Silva P, Zilberman D. 2010. Genome-wide evolutionary analysis of eukaryotic DNA methylation. Science 328: 916–9. 9. Jones PA. 2012. Functions of DNA methylation: islands, start sites, gene bodies and beyond. Nat Rev Genet 13: 484–92. 10. Zilberman D. 2008. The evolving functions of DNA methylation. Curr Opin Plant Biol 11: 554–9. 11. Yoder JA, Walsh CP, Bestor TH. 1997. Cytosine methylation and the ecology of intragenomic parasites. Trends Genet 13: 335–40. 12. Crichton JH, Dunican DS, Maclennan M, Meehan RR, et al. 2014. Defending the genome from the enemy within: mechanisms of retrotransposon suppression in the mouse germline. Cell Mol Life Sci 71: 1581– 605. 13. Feinberg AP, Vogelstein B. 1983. Hypomethylation distinguishes genes of some human cancers from their normal counterparts. Nature 301: 89–92. 14. Feng S, Jacobsen SE. 2011. Epigenetic modifications in plants: an evolutionary perspective. Curr Opin Plant Biol 14: 179–86. 15. Lister R, Pelizzola M, Dowen RH, Hawkins RD, et al. 2009. Human DNA methylomes at base resolution show widespread epigenomic differences. Nature 462: 315–22. 16. Feng S, Cokus SJ, Zhang X, Chen P-Y, et al. 2010. Conservation and divergence of methylation patterning in plants and animals. Proc Natl Acad Sci USA 107: 8689–94. 17. Stroud H, Do T, Du J, Zhong X, et al. 2014. Non-CG methylation patterns shape the epigenetic landscape in Arabidopsis. Nat Struct Mol Biol 21: 64–72. 18. Goll MG, Bestor TH. 2005. Eukaryotic cytosine methyltransferases. Annu Rev Biochem 74: 481–514. 19. Goll MG, Kirpekar F, Maggert KA, Yoder JA, et al. 2006. Methylation of tRNAAsp by the DNA methyltransferase homolog Dnmt2. Science 311: 395–8.
Bioessays 36: 0000–0000, ß 2014 WILEY Periodicals, Inc.
.....
Insights & Perspectives
33. Lyko F, Ramsahoye BH, Kashevsky H, Tudor M, et al. 1999. Mammalian (cytosine5) methyltransferases cause genomic DNA methylation and lethality in Drosophila. Nat Genet 23: 363–6. 34. Marhold J, Kramer K, Kremmer E, Lyko F. 2004. The Drosophila MBD2/3 protein mediates interactions between the MI-2 chromatin complex and CpT/A-methylated DNA. Development 131: 6033–9. 35. Takayama S, Dhahbi J, Roberts A, Mao G, et al. 2014. Genome methylation in D. melanogaster is found at specific short motifs and is independent of DNMT2 activity. Genome Res 24: 821–30. 36. Marhold J, Brehm A, Kramer K. 2004. The Drosophila methyl-DNA binding protein MBD2/3 interacts with the NuRD complex via p55 and MI-2. BMC Mol Biol 5: 20. 37. Graveley BR, Brooks AN, Carlson JW, Duff MO, et al. 2011. The developmental transcriptome of Drosophila melanogaster. Nature 471: 473–9. 38. Raja SJ, Charapitsa I, Conrad T, Vaquerizas JM, et al. 2010. The nonspecific lethal complex is a transcriptional regulator in Drosophila. Mol Cell 38: 827–41. 39. Dunwell TL, McGuffin LJ, Dunwell JM, Pfeifer GP. 2013. The mysterious presence of a 5-methylcytosine oxidase in the Drosophila genome: possible explanations. Cell Cycle 12: 3357–65. 40. Tahiliani M, Koh KP, Shen Y, Pastor WA, et al. 2009. Conversion of 5-methylcytosine to 5-hydroxymethylcytosine in mammalian DNA by MLL partner TET1. Science 324: 930–5. 41. Ito S, Shen L, Dai Q, Wu SC, et al. 2011. Tet proteins can convert 5-methylcytosine to 5formylcytosine and 5-carboxylcytosine. Science 333: 1300–3. 42. Okano M, Xie S, Li E. 1998. Dnmt2 is not required for de novo and maintenance methylation of viral DNA in embryonic stem cells. Nucleic Acids Res 26: 2536–40. 43. Phalke S, Nickel O, Walluscheck D, Hortig F, et al. 2009. Retrotransposon silencing and telomere integrity in somatic cells of
Bioessays 36: 0000–0000, ß 2014 WILEY Periodicals, Inc.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
Drosophila depends on the cytosine-5 methyltransferase DNMT2. Nat Genet 41: 696–702. Schaefer M, Lyko F. 2010. Lack of evidence for DNA methylation of Invader4 retroelements in Drosophila and implications for Dnmt2-mediated epigenetic regulation. Nat Genet 42: 920–1 author reply 1. Phalke S, Nickel O, Reuter G. 2010. Reply to “Lack of evidence for DNA methylation of Invader4 retroelements in Drosophila and implications for Dnmt2-mediated epigenetic regulation”. Nat Genet 42: 921. Durdevic Z, Hanna K, Gold B, Pollex T, et al. 2013. Efficient RNA virus control in Drosophila requires the RNA methyltransferase Dnmt2. EMBO Rep 14: 269–75. Raddatz G, Guzzardo PM, Olova N, Fantappie MR, et al. 2013. Dnmt2-dependent methylomes lack defined DNA methylation patterns. Proc Natl Acad Sci USA 110: 8627–31. Capuano F, Mulleder M, Kok R, Blom HJ, et al. 2014. Cytosine DNA methylation is found in Drosophila melanogaster but absent in Saccharomyces cerevisiae, Schizosaccharomyces pombe, and other yeast species. Analyt Chem 86: 3697–702. Jurkowski TP, Meusburger M, Phalke S, Helm M, et al. 2008. Human DNMT2 methylates tRNAAsp molecules using a DNA methyltransferase-like catalytic mechanism. RNA 14: 1663–70. Warnecke PM, Stirzaker C, Song J, Grunau C, et al. 2002. Identification and resolution of artifacts in bisulfite sequencing. Methods 27: 101–7. Flusberg BA, Webster DR, Lee JH, Travers KJ, et al. 2010. Direct detection of DNA methylation during single-molecule, real-time sequencing. Nat Meth 7: 461–5. Wallace EVB, Stoddart D, Heron AJ, Mikhailova E, et al. 2010. Identification of epigenetic DNA modifications with a protein nanopore. Chem Commun 46: 8195–7. Bailey TL. 2011. DREME: motif discovery in transcription factor ChIP-seq data. Bioinformatics 27: 1653–9.
7
Think again
20. Schaefer M, Pollex T, Hanna K, Tuorto F, et al. 2010. RNA methylation by Dnmt2 protects transfer RNAs against stress-induced cleavage. Genes Dev 24: 1590–5. 21. Li E, Bestor TH, Jaenisch R. 1992. Targeted mutation of the DNA methyltransferase gene results in embryonic lethality. Cell 69: 915–26. 22. Okano M, Bell DW, Haber DA, Li E. 1999. DNA methyltransferases Dnmt3a and Dnmt3b are essential for de novo methylation and mammalian development. Cell 99: 247–57. 23. Klein CJ, Botuyan M-V, Wu Y, Ward CJ, et al. 2011. Mutations in DNMT1 cause hereditary sensory neuropathy with dementia and hearing loss. Nat Genet 43: 595–600. 24. Winkelmann J, Lin L, Schormair B, Kornum BR, et al. 2012. Mutations in DNMT1 cause autosomal dominant cerebellar ataxia, deafness and narcolepsy. Hum Mol Genet 21: 2205–10. 25. Tweedie S, Charlton J, Clark V, Bird A. 1997. Methylation of genomes and genes at the invertebrate-vertebrate boundary. Mol Cell Biol 17: 1469–75. 26. Urieli-Shoval S, Gruenbaum Y, Sedat J, Razin A. 1982. The absence of detectable methylated bases in Drosophila melanogaster DNA. FEBS Lett 146: 148–52. 27. Hung M-S, Karthikeyan N, Huang B, Koo HC, et al. 1999. Drosophila proteins related to vertebrate DNA (5-cytosine) methyltransferases. Proc Natl Acad Sci USA 96: 11940–5. 28. Tweedie S, Ng H-H, Barlow AL, Turner BM, et al. 1999. Vestiges of a DNA methylation system in Drosophila melanogaster? Nat Genet 23: 389–90. 29. Lyko F, Ramsahoye BH, Jaenisch R. 2000. DNA methylation in Drosophila melanogaster. Nature 408: 538–40. 30. Gowher H, Leismann O, Jeltsch A. 2000. DNA of Drosophila melanogaster contains 5-methylcytosine. EMBO J 19: 6918–23. 31. Lyko F. 2001. DNA methylation learns to fly. Trends Genet 17: 169–72. 32. Kunert N, Marhold J, Stanke J, Stach D, et al. 2003. A Dnmt2-like protein mediates DNA methylation in Drosophila. Development 130: 5083–90.
D. Boffelli et al.
Now you see it: Genome methylation makes a comeback in Drosophila Dario Boffelli*, Sachiko Takayama and David I. K. Martin* Drosophila melanogaster is often considered to lack genomic 5-methylcytosine (m5C), an opinion reinforced by two whole genome bisulfite-sequencing studies that failed to find m5C. New evidence, however, indicates that genomic methylation is indeed present in the fly, albeit in small quantities and in unusual patterns. At embryonic stage 5, m5C occurs in short strand-specific regions that cover 1% of the genome, at tissue levels suggesting a distribution restricted to a subset of nuclei. Its function is not obvious, but methylation in subsets of nuclei would obscure functional associations since transcript levels and epigenetic modifications are assayed in whole embryos. Surprisingly, Mt2, the fly’s only candidate DNA methyltransferase, is not necessary for the observed methylation. Full evaluation of the functions of genome methylation in Drosophila must await discovery and experimental inactivation of the DNA methyltransferase, as well as a better understanding of the pattern and developmental regulation of genomic m5C.
.
Keywords: development; Drosophila; methylation
The functions of DNA methylation are incompletely understood 5-Methylcytosine (m5C) is very widely distributed in eukaryotic organisms [1], where it serves as a tag that can be recognized and read by proteins that act at or near the methylated site in a genome [2, 3]. It has been most inten-
sively studied in vertebrates, where it typically occurs as a symmetrical modification of CpG dinucleotides; when it occurs in regions of high CpG density (“CpG islands”) m5C is a central component in systems of epigenetic regulation that suppress transcription initiation in both animals and plants [4–6]. While the prominence of CpG island methylation in plant and vertebrate epigenetics might give the impression that the function of
m5C is well understood, the broad scope of methylation within genomes and among species has left many open questions [7]. m5C occurs in multiple distinct patterns: in vertebrates, it is found over the whole genome except for CpG islands; in other species it is targeted to transposable elements, or to gene bodies; in yet other species it is excluded from gene bodies [1, 8, 9]. No general theme is evident, and in most cases the functional significance of these patterns has not been established [1, 10]. Even in the vertebrates, where m5C has been most intensively studied, there are many unknowns. CpGs that occur outside of CpG islands are usually methylated, but except for silencing of retroelements [11, 12] the function (or more probably, functions) of this methylation is unclear [1, 9]. Also, it has been known for decades that human cancer genomes display widespread loss of methylation [13], but it is not understood why this happens or what its consequences for the malignant phenotype are. Non-CpG methylation is common in plants [1, 14], and has been observed in mammals [15], but its functions are still the subject of investigation in plants [16, 17] and are unknown in mammals. Given the variety of m5C patterns that have been observed, it seems likely that it serves multiple functions.
DOI 10.1002/bies.201400097 Children’s Hospital Oakland Research Institute, Oakland, CA, USA *Corresponding authors: Dario Boffelli E-mail: [email protected] David I. K. Martin E-mail: [email protected]
Abbreviations: bp, base pairs; DNMT, DNA methyltransferase; hm5C, 5-hydroxymethylcytosine; m5C, 5-methylcytosine; MBD, methyl binding domain; MeDIP, methylcytosine immunoprecipitation; TET, teneleven translocation.
Bioessays 36: 0000–0000, ß 2014 WILEY Periodicals, Inc.
DNA methyltransferases establish and maintain CpG methylation A class of related DNA methyltransferases (DNMTs) establish and maintain
www.bioessays-journal.com
1
Think again
D. Boffelli et al.
genomic m5C in animals [18]. The genetic and biochemical characterization of the DNMTs has been carried out mostly in human and mouse, and function in other species has been inferred by the presence of a conserved DNA sequence in the catalytic domains [18]. DNMT1 is a “maintenance” methyltransferase, acting to complete methylation of hemimethylated CpGs [18]. Two members of the DNMT3 family function as de novo methyltransferases [18]. DNMT2, while related to the other methyltransferases, is not clearly a DNA methyltransferase: it can methylate RNA, and this may be its principal function [19, 20]. The importance of m5C for animal and plant development has been established by experimental inactivation of DNMT1 [21], DNMT3a [22], and DNMT3b [22], which results in hypomethylation, with broad effects on development, gene expression, and the suppression of retrotransposons. DNMT1 deficiency gives rise to several neurodevelopmental defects [23, 24], while DNMT3b deficiency has been associated with ICF1 (immunodeficiency-centromeric instability-facial anomalies syndrome-1; OMIM 242860) [22].
Drosophila has a DNA methylation system Drosophila melanogaster once served as an instructive example of a species that carries out complex gene regulation and cell differentiation without cytosine methylation [25, 26]. The fly does not have any apparent homologue of DNMT1 or DNMT3 [18], and m5C had not been detected in the fly genome [25, 26]. The discovery of a fly DNMT2 homologue (Mt2) [27, 28] prompted searches for m5C using more sensitive methods: these studies produced evidence of extremely small quantities of m5C, most abundant in the early embryo, and in the context of CA and CT dinucleotides rather than CpG [29– 31]. The methods could detect the presence of m5C in bulk, but could not establish where in the genome it was located. Immunostaining for m5C in fly embryos revealed a nuclear staining pattern that was weaker when Mt2 was knocked down, and bisulfite sequencing of random genomic fragments 2
Insights & Perspectives
showed apparent methylation in CA and CT contexts [32]. Overexpression of mouse DNMT1 and DNMT3 in Drosophila induced m5C and resulted in developmental defects and lethality [33], prompting speculation that the fly expresses proteins that can interact with genomic m5C and produce phenotypic effects (which were abnormal presumably because the induced methylation was ectopic). One of these proteins may be the methyl binding domain protein MBD2/3 [28] (FlyBase ID CG8208). In contrast to vertebrate MBDs, Drosophila MBD2/3 binds DNA with a preference for CA/ CT contexts [34], consistent with evidence from nearest neighbor analysis of fly genomic methylcytosine [29], and with our recent findings discussed below [35]. MBD2/3 also co-fractionates with components of the NuRD complex [36], suggesting a model in which it binds DNA and recruits chromatinremodeling complexes. Although this MBD may have other functions, its ability to bind methylcytosine is most easily explained by the presence and functional significance of genomic methylcytosine. It is expressed ubiquitously throughout fly development [37], suggesting a similarly broad function for m5C. Another MBD, MBD-R2, is part of the non-specific lethal (NSL) complex and contains a CpG binding domain [38], but its ability to bind methylated DNA has not been demonstrated. Further circumstantial evidence is provided by the presence in Drosophila of dTet (FlyBase ID CG43444). dTet is a homologue of the ten-eleven translocation (TET) protein [39], a methylcytosine oxidase which in mammals is responsible for metabolizing m5C to 5-hydroxymethylcytosine (hm5C), the initial step in the pathway that eventually leads to demethylation [40, 41]. dTet has an intact CXXC CpG-binding domain and HXD catalytic site, strongly implying that it is functional in the same way as mammalian TET proteins [39]. Dunwell et al. [39] argue that the domain of dTet that binds methylated DNA is not restricted to CpG, and that the conservation of TET function in the fly strongly implies that methylcytosine is present on the genome in some tissue or tissues. dTet is expressed from embryonic stage 7/8, predominantly in neural tissues [37], and so could act to
.....
demethylate or create hm5C from m5C laid down at earlier stages.
The presence of genomic methylation in Drosophila has been controversial The extremely low levels of m5C detected, even in the early embryo, seemed incompatible with cytosine methylation having a role in gene regulation and epigenetic phenomena similar to that seen in mammals. The issue of function was further clouded because the methods originally used to detect m5C cannot establish where in the fly genome it is located. Later developments cast further doubts on the biological significance of m5C in the fly. Depletion or mutation of Mt2 has only minor phenotypic effects [20, 32], similarly to the mouse DNMT2 knockout, which is also viable and normal [19, 42]. Inactivation of Mt2 suppresses variegated silencing of white in Drosophila [43], but a claimed link to retrotransposon methylation was refuted [44, 45]. Evidence that Mt2 methylates tRNA suggested an alternative function for the protein, albeit without showing that the function is important [19, 20, 46]. This alternative function for Mt2 cast doubt on the evidence for genomic m5C: if the only DNMT in the fly were in fact an RNA methyltransferase, then genomic methylation might be, as some have described it, “spurious” [44, 47]. Most recently, two whole genome bisulfite sequencing studies concluded that no genomic m5C is present in the fly; one study sequenced the fly embryo genome at an average depth of 6 [8], and the other at 32 [47]. Although unconverted cytosines (indicative of methylcytosine) were detected in the deeper study, they were also present in DNA derived from Mt2 mutants, leading the authors to conclude that the unconverted cytosine was an artifact of incomplete bisulfite conversion. These observations conflict with the earlier evidence for genomic methylation in the fly, and have convinced even one of its discoverers that the evidence was artifactual. Nevertheless the evidence from whole genome bisulfite sequencing does not demonstrate that the earlier results obtained by bulk analysis were wrong.
Bioessays 36: 0000–0000, ß 2014 WILEY Periodicals, Inc.
.....
Insights & Perspectives
New evidence for genome methylation in the fly
bisulfite sequencing. We chose 66 methylated regions for resequencing after bisulfite conversion; sequencing was carried out at extremely high depth to obtain accurate measures of the level at each cytosine. This method validated the presence of methylation at all of the 66 regions, with peaks in roughly the same sites as the peaks detected by MeDIPBseq [35]. However, methylation at these sites is never complete: even the most highly methylated positions are methylated on only 50% of genomes in the stage 5 embryo, and most methylated positions are methylated in 5,000 might be required to detect these methylated regions; enrichment, such as with MeDIP, would reduce the amount of sequencing required to detect them by an unknown factor. Such a low frequency of methylation would be difficult to distinguish from an artifact of bisulfite non-conversion. Our study indicated an 1% nonconversion rate in control DNA using a commercial conversion kit, so that methylation would need to be present in a much higher proportion to be detectable without enrichment. However, if bisulfite non-conversion is randomly distributed (still a murky subject [50]), then the presence of consistent allelic patterns of methylation could be used to detect methylated alleles at much lower frequency, particularly if they occurred in a single characteristic sequence context such as CpG (our data suggest that CG methylation might be present at extremely low frequency – see Fig. 1).
.....
The function of genome methylation in the fly remains enigmatic While the evidence indicates that the fly genome does possess methylation, the function of this methylation is not at all clear. The methylation pattern we have observed is completely unlike that known in any other species; thus no inferences can be made by comparison to other systems. Indeed the strongly focal pattern, with methylation concentrated in small areas characterized by an extended sequence motif, has been observed only in the fly, and it suggests that methylation could have functions that are unique to the fly (although it should be noted that 5mC in the quantity and pattern that we observe might be present in other species that have more abundant 5mC, but its relative rarity would obscure its presence). The distribution of methylation in the fly genome does not produce any obvious hints as to its function: we found no association with several histone modifications that mediate gene regulation, and no relationship with gene regulatory elements or gene bodies. Because Mt2 has been eliminated as the factor responsible for genome methylation, and there is no alternative candidate, the option of assessing a mutant lacking DNA methylation is currently unavailable. We found a small but significant association of gene body methylation with lower expression levels of developmental genes at stage 5 [35]. This suggests the possibility that genome methylation in the fly is – as in many systems – involved in gene regulation; however, much methylation is located at a distance from genes, suggesting that it could have other functions. The weak association of methylation with gene expression, taken together with the presence of methylation in only a small subset of genomes at stage 5, is consistent with a scenario in which methylation is heterocellular (or more properly heteronuclear) at stage 5. This could imply that its effects on gene expression in specific nuclei are larger than those observed over the entire embryo. Most of the methylated regions we have identified are methylated on less than 10% of sequenced DNA strands. We have noted a tendency for
Bioessays 36: 0000–0000, ß 2014 WILEY Periodicals, Inc.
.....
Insights & Perspectives
methylation at specific sites. One recent study detected methylcytosine in adult flies at 0.03% of all cytosine, consistent with the presence of methylated regions in a small but not insignificant subset of cells [48]. Dunwell et al. [39] have pointed out that dTet is expressed in fly neural tissues, and also functions to generate hm5C in the mammalian brain. On this basis they speculate that methylcytosine, or hydroxymethylcytosine, has a functional role in those tissues. The small amounts of methylcytosine detected in adult flies [48] might reflect this. The precise location of such methylation, as experience to date demonstrates, could be very difficult to establish: even at stage 5, where methylcytosine is most abundant, localization required enrichment and deep sequencing; if it is present in lower proportions in other tissues, even our method might not suffice to detect it. One of the most baffling aspects of the picture produced by our analysis of the fly methylome is the non-dependence of methylation on Mt2, the fly DNMT2 homologue. This was an unexpected result: we analyzed an Mt2
strain as a means of buttressing the existence of genome methylation, and when we found methylation retained in one mutant strain we analyzed a second mutant. In both mutant strains, we found methylation at stage 5 at sites and in amounts similar to wild type strains. The result clearly indicates that Drosophila must possess some other factor capable of methylating cytosine. Since no other DNA methyltransferase homologue is apparent in the sequenced portions of the fly genome, the factor must either reside in some unsequenced region, or lack sufficient similarity to other DNMTs to make it recognizable by protein domain similarity searches. We considered the possibility that the genome methylation we observe is derived from a DNA methyltransferase expressed by one of the several Drosophila endoparasites, such as Wolbachia. However, we did not find any reduction in methylation after three generations of treatment with tetracycline (unpublished result); while this result does not eliminate the possibility of an endoparasite origin, it strongly argues against it. Thus Drosophila apparently carries a cryptic DNA methyltransferase of a novel type.
Conclusions and outlook
Figure 2. Allelic pattern of m5C in a single methylated region revealed by deep bisulfite resequencing. The coordinates of the resequenced region are shown at the bottom. Stage 5 genomic DNA was bisulfite converted, PCR amplified, and subjected to paired-end Illumina sequencing to extreme depth. To simplify visualization of the pattern, only the cytosine positions in selected reads are shown (black: methylated; gray: unmethylated). The thin gray lines visible in some reads denote the unsequenced portion of a fragment (between paired ends). Methylation is concentrated in a subset of the sequenced alleles, consistent with a subset of stage 5 nuclei being methylated in that region.
Bioessays 36: 0000–0000, ß 2014 WILEY Periodicals, Inc.
Multiple lines of evidence support the presence of DNA methylation in Drosophila (Box 1). Drosophila possesses proteins that are capable of interacting with genomic m5C [34, 39]. Bulk assays have consistently found small amounts of m5C, most abundant at stage 5 but also present at later stages of development [29, 30, 48], and our recent study has located foci of genome methylation at stage 5 [35]. This evidence indicates that genomic methylation is present, but the pattern is unique, and provides no obvious clues to its function. The role of DNA methylation in Drosophila remains subject to speculation, as is the case in many other species. Nevertheless, the presence of m5C in this important model organism, along with conserved proteins that can mediate its function, resurrects genome methylation as a subject for study and presents the challenge of explaining how it is established and what it does. 5
Think again
methylation to be concentrated in a subset of sequence reads, indicating an allelic pattern to methylation (Fig. 2). This may indicate that all methylation is concentrated in a subset of embryonic nuclei, or that all nuclei have some methylation, but at different subsets of the regions we identified: our study is unable to distinguish these possibilities, because of the short length of the sequence reads (Fig. 3). A technology capable of direct detection of methylcytosine on long-read sequences could resolve this question [51, 52]. A heterocellular pattern is entirely consistent with the differentiation of epigenomes that takes place during development: if m5C functions in gene regulation, its precise positioning would be expected to vary according to cell type or fate. Such variation occurs on a grand scale in vertebrates; in the fly we find tens of thousands of methylated regions, and each of these potentially represents a source of functional differentiation. Heterocellular methylation patterns might extend well beyond the embryonic period, but the increased number and complexity of cells would make it much more difficult to detect
D. Boffelli et al.
Insights & Perspectives
.....
Think again
D. Boffelli et al.
Figure 3. Schematic illustrating hypothetical distribution patterns of m5C in the Drosophila embryo. Two methylated regions are shown at the top in blue and red. In model 1, a subset of nuclei is methylated at Region A but not region B, while a different subset is methylated at Region B but not region A. In model 2, both regions are methylated in a subset of nuclei. In model 3, methylation of the two regions is diffusely distributed throughout the embryo, so that many or all nuclei have a little methylation in both of the regions. The three models describe distinctly different patterns. However, when methylation is assayed over the entire embryo (as in all studies to date) the patterns are indistinguishable.
Box 1 Experimental evidence for DNA methylation in Drosophila Methylated cytosine detected in the fly genome with a method capable of distinguishing DNA and RNA methylation.
The fly genome encodes proteins capable of interacting with methylated cytosines.
Immunoprecipitation of methylated cytosines coupled with bisulfite
sequencing shows that 1% of the fly genome at stage 5 is methylated at low levels. Targeted bisulfite-PCR resequencing provides quantitative information on methylation levels and shows that methylation is dynamic in early development.
The biological significance of fly genomic methylation could be clarified by identifying the responsible DNA methyltransferase, and observing the phenotypic effects of its inactivation. However, no DNMT homologue other than Mt2 is detectable in the fly genome sequence by computational methods. The possibility that such a homologue resides in an unsequenced portion of the fly genome becomes increasingly unlikely, as more sequence is available; we are left with the possibility that the responsible methyltransferase does not resemble the other DNMTs. In that case, identification might require purely functional screens based on detection of methylation at some of the regions we have identified. 6
As discussed above, there remains a distinct possibility that other patterns of genomic m5C are present, either in rare embryonic cells or at later stages of fly development. Discovery of these patterns will require more sequencing, albeit possibly at intimidating depths or in carefully selected subsets of cells or nuclei. The outcomes of such endeavors are at present unknowable, and await the enterprising and adventurous investigator.
References 1. Suzuki MM, Bird A. 2008. DNA methylation landscapes: provocative insights from epigenomics. Nat Rev Genet 9: 465–76. 2. Klose RJ, Bird AP. 2006. Genomic DNA methylation: the mark and its mediators. Trends Biochem Sci 31: 89–97.
3. Bogdanovic´ O, Veenstra GJC. 2009. DNA methylation and methyl-CpG binding proteins: developmental requirements and function. Chromosoma 118: 549–65. 4. Deaton AM, Bird A. 2011. CpG islands and the regulation of transcription. Genes Dev 25: 1010–22. 5. Bird AP, Wolffe AP. 1999. Methylationinduced repression – belts, braces, and chromatin. Cell 99: 451–4. 6. Smith ZD, Meissner A. 2013. DNA methylation: roles in mammalian development. Nat Rev Genet 14: 204–20. 7. Reddington JP, Pennings S, Meehan RR. 2013. Non-canonical functions of the DNA methylome in gene regulation. Biochem J 451: 13–23. 8. Zemach A, McDaniel IE, Silva P, Zilberman D. 2010. Genome-wide evolutionary analysis of eukaryotic DNA methylation. Science 328: 916–9. 9. Jones PA. 2012. Functions of DNA methylation: islands, start sites, gene bodies and beyond. Nat Rev Genet 13: 484–92. 10. Zilberman D. 2008. The evolving functions of DNA methylation. Curr Opin Plant Biol 11: 554–9. 11. Yoder JA, Walsh CP, Bestor TH. 1997. Cytosine methylation and the ecology of intragenomic parasites. Trends Genet 13: 335–40. 12. Crichton JH, Dunican DS, Maclennan M, Meehan RR, et al. 2014. Defending the genome from the enemy within: mechanisms of retrotransposon suppression in the mouse germline. Cell Mol Life Sci 71: 1581– 605. 13. Feinberg AP, Vogelstein B. 1983. Hypomethylation distinguishes genes of some human cancers from their normal counterparts. Nature 301: 89–92. 14. Feng S, Jacobsen SE. 2011. Epigenetic modifications in plants: an evolutionary perspective. Curr Opin Plant Biol 14: 179–86. 15. Lister R, Pelizzola M, Dowen RH, Hawkins RD, et al. 2009. Human DNA methylomes at base resolution show widespread epigenomic differences. Nature 462: 315–22. 16. Feng S, Cokus SJ, Zhang X, Chen P-Y, et al. 2010. Conservation and divergence of methylation patterning in plants and animals. Proc Natl Acad Sci USA 107: 8689–94. 17. Stroud H, Do T, Du J, Zhong X, et al. 2014. Non-CG methylation patterns shape the epigenetic landscape in Arabidopsis. Nat Struct Mol Biol 21: 64–72. 18. Goll MG, Bestor TH. 2005. Eukaryotic cytosine methyltransferases. Annu Rev Biochem 74: 481–514. 19. Goll MG, Kirpekar F, Maggert KA, Yoder JA, et al. 2006. Methylation of tRNAAsp by the DNA methyltransferase homolog Dnmt2. Science 311: 395–8.
Bioessays 36: 0000–0000, ß 2014 WILEY Periodicals, Inc.
.....
Insights & Perspectives
33. Lyko F, Ramsahoye BH, Kashevsky H, Tudor M, et al. 1999. Mammalian (cytosine5) methyltransferases cause genomic DNA methylation and lethality in Drosophila. Nat Genet 23: 363–6. 34. Marhold J, Kramer K, Kremmer E, Lyko F. 2004. The Drosophila MBD2/3 protein mediates interactions between the MI-2 chromatin complex and CpT/A-methylated DNA. Development 131: 6033–9. 35. Takayama S, Dhahbi J, Roberts A, Mao G, et al. 2014. Genome methylation in D. melanogaster is found at specific short motifs and is independent of DNMT2 activity. Genome Res 24: 821–30. 36. Marhold J, Brehm A, Kramer K. 2004. The Drosophila methyl-DNA binding protein MBD2/3 interacts with the NuRD complex via p55 and MI-2. BMC Mol Biol 5: 20. 37. Graveley BR, Brooks AN, Carlson JW, Duff MO, et al. 2011. The developmental transcriptome of Drosophila melanogaster. Nature 471: 473–9. 38. Raja SJ, Charapitsa I, Conrad T, Vaquerizas JM, et al. 2010. The nonspecific lethal complex is a transcriptional regulator in Drosophila. Mol Cell 38: 827–41. 39. Dunwell TL, McGuffin LJ, Dunwell JM, Pfeifer GP. 2013. The mysterious presence of a 5-methylcytosine oxidase in the Drosophila genome: possible explanations. Cell Cycle 12: 3357–65. 40. Tahiliani M, Koh KP, Shen Y, Pastor WA, et al. 2009. Conversion of 5-methylcytosine to 5-hydroxymethylcytosine in mammalian DNA by MLL partner TET1. Science 324: 930–5. 41. Ito S, Shen L, Dai Q, Wu SC, et al. 2011. Tet proteins can convert 5-methylcytosine to 5formylcytosine and 5-carboxylcytosine. Science 333: 1300–3. 42. Okano M, Xie S, Li E. 1998. Dnmt2 is not required for de novo and maintenance methylation of viral DNA in embryonic stem cells. Nucleic Acids Res 26: 2536–40. 43. Phalke S, Nickel O, Walluscheck D, Hortig F, et al. 2009. Retrotransposon silencing and telomere integrity in somatic cells of
Bioessays 36: 0000–0000, ß 2014 WILEY Periodicals, Inc.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
Drosophila depends on the cytosine-5 methyltransferase DNMT2. Nat Genet 41: 696–702. Schaefer M, Lyko F. 2010. Lack of evidence for DNA methylation of Invader4 retroelements in Drosophila and implications for Dnmt2-mediated epigenetic regulation. Nat Genet 42: 920–1 author reply 1. Phalke S, Nickel O, Reuter G. 2010. Reply to “Lack of evidence for DNA methylation of Invader4 retroelements in Drosophila and implications for Dnmt2-mediated epigenetic regulation”. Nat Genet 42: 921. Durdevic Z, Hanna K, Gold B, Pollex T, et al. 2013. Efficient RNA virus control in Drosophila requires the RNA methyltransferase Dnmt2. EMBO Rep 14: 269–75. Raddatz G, Guzzardo PM, Olova N, Fantappie MR, et al. 2013. Dnmt2-dependent methylomes lack defined DNA methylation patterns. Proc Natl Acad Sci USA 110: 8627–31. Capuano F, Mulleder M, Kok R, Blom HJ, et al. 2014. Cytosine DNA methylation is found in Drosophila melanogaster but absent in Saccharomyces cerevisiae, Schizosaccharomyces pombe, and other yeast species. Analyt Chem 86: 3697–702. Jurkowski TP, Meusburger M, Phalke S, Helm M, et al. 2008. Human DNMT2 methylates tRNAAsp molecules using a DNA methyltransferase-like catalytic mechanism. RNA 14: 1663–70. Warnecke PM, Stirzaker C, Song J, Grunau C, et al. 2002. Identification and resolution of artifacts in bisulfite sequencing. Methods 27: 101–7. Flusberg BA, Webster DR, Lee JH, Travers KJ, et al. 2010. Direct detection of DNA methylation during single-molecule, real-time sequencing. Nat Meth 7: 461–5. Wallace EVB, Stoddart D, Heron AJ, Mikhailova E, et al. 2010. Identification of epigenetic DNA modifications with a protein nanopore. Chem Commun 46: 8195–7. Bailey TL. 2011. DREME: motif discovery in transcription factor ChIP-seq data. Bioinformatics 27: 1653–9.
7
Think again
20. Schaefer M, Pollex T, Hanna K, Tuorto F, et al. 2010. RNA methylation by Dnmt2 protects transfer RNAs against stress-induced cleavage. Genes Dev 24: 1590–5. 21. Li E, Bestor TH, Jaenisch R. 1992. Targeted mutation of the DNA methyltransferase gene results in embryonic lethality. Cell 69: 915–26. 22. Okano M, Bell DW, Haber DA, Li E. 1999. DNA methyltransferases Dnmt3a and Dnmt3b are essential for de novo methylation and mammalian development. Cell 99: 247–57. 23. Klein CJ, Botuyan M-V, Wu Y, Ward CJ, et al. 2011. Mutations in DNMT1 cause hereditary sensory neuropathy with dementia and hearing loss. Nat Genet 43: 595–600. 24. Winkelmann J, Lin L, Schormair B, Kornum BR, et al. 2012. Mutations in DNMT1 cause autosomal dominant cerebellar ataxia, deafness and narcolepsy. Hum Mol Genet 21: 2205–10. 25. Tweedie S, Charlton J, Clark V, Bird A. 1997. Methylation of genomes and genes at the invertebrate-vertebrate boundary. Mol Cell Biol 17: 1469–75. 26. Urieli-Shoval S, Gruenbaum Y, Sedat J, Razin A. 1982. The absence of detectable methylated bases in Drosophila melanogaster DNA. FEBS Lett 146: 148–52. 27. Hung M-S, Karthikeyan N, Huang B, Koo HC, et al. 1999. Drosophila proteins related to vertebrate DNA (5-cytosine) methyltransferases. Proc Natl Acad Sci USA 96: 11940–5. 28. Tweedie S, Ng H-H, Barlow AL, Turner BM, et al. 1999. Vestiges of a DNA methylation system in Drosophila melanogaster? Nat Genet 23: 389–90. 29. Lyko F, Ramsahoye BH, Jaenisch R. 2000. DNA methylation in Drosophila melanogaster. Nature 408: 538–40. 30. Gowher H, Leismann O, Jeltsch A. 2000. DNA of Drosophila melanogaster contains 5-methylcytosine. EMBO J 19: 6918–23. 31. Lyko F. 2001. DNA methylation learns to fly. Trends Genet 17: 169–72. 32. Kunert N, Marhold J, Stanke J, Stach D, et al. 2003. A Dnmt2-like protein mediates DNA methylation in Drosophila. Development 130: 5083–90.
D. Boffelli et al.
