Interplay between Two Epigenetic Marks

Lianna M. Johnson, Xiaofeng Cao, Steven E. Jacobsen
2002 Current Biology  
CpG sites, but it is also found at asymmetric sites and , in plants, at CpNpG sites. Once established, DNA methylation is inherited through mitosis, and often through chromatin and provides a binding site for the heterochromatin protein 1 (HP1) [6-9]. H3-K9 methylation is Los Angeles, California 90095-1606 carried out by the SU(VAR)3-9 class of proteins [8, 10] and is correlated with silenced genes in mammals [11], chickens [12], Schizosaccharomyces pombe [8], and Summary Neurospora crassa [13]
more » ... . Several lines of evidence suggest that H3-K9 methylation is an early event in the Background: The heterochromatin of many eukaryotes formation of heterochromatin: in humans, H3-K9 methylis marked by both DNA methylation and histone H3 ation has been shown to mark the X chromosome shortly lysine 9 (H3-K9) methylation, though the exact relationafter coating with Xist RNA and prior to gene inactivation ship between these epigenetic modifications is un- [14] . This is followed much later by DNA methylation known. In Neurospora, H3-K9 methylation is required [15] . In Neurospora, it has been shown that all DNA for the maintenance of all known DNA methylation. In methylation is dependent on H3-K9 methylation, sug-Arabidopsis, H3-K9 methylation directs some of the gesting that methylation of H3-K9 occurs prior to DNA CpNpG and asymmetric methylation. However, it is not methylation [13] . Similarly, in Arabidopsis, CpNpG methknown in any organism whether DNA methylation may ylation is partially dependent on H3-K9 methylation [16]. also direct histone H3 methylation. It has been proposed that these two epigenetic marks Results: Using chromatin immunoprecipitation (ChIP) may signal to one another to ensure propagation of the assays, we show that Arabidopsis heterochromatin is silenced state [1]. However, it is still not known whether associated with H3-K9 methylation. This histone methyl-DNA methylation can actually direct H3-K9 methylation ation is dependent on the KRYPTONITE and DDM1 or whether DNA methylation can act independently to genes (SU[VAR]3-9 and SWI2/SNF2 homologs, respecsilence gene expression in the absence of H3-K9 methyltively). We also find that a decrease in DNA methylation ation. does not directly cause a loss of H3-K9 methylation. Arabidopsis serves as an ideal system in which to Instead, a decrease in H3-K9 methylation is only seen examine these questions, as many viable mutants that at loci where transcription is derepressed. affect DNA methylation have been isolated. One of the Conclusions: We conclude that DNA methylation does first mutations isolated was in the SWI2/SNF2 homolog, not control the methylation of histone H3-K9. We pro-DDM1 (decrease in DNA methylation), encoding a putapose that loss of H3-K9 methylation is due to transcriptive chromatin remodeling protein [17, 18] . Mutations in tional reactivation, coupled with deposition of unmethylthis gene result in the loss of approximately 70% of ated nucleosomes. These findings are consistent with total DNA methylation. More recently, loss-of-function recent observations of DNA replication-independent demutations in a SU(VAR)3-9 homolog, KRYPTONITE position of histone H3.3 in Drosophila. Our results also (KYP), were isolated in a suppressor screen for reactivasuggest that, in Arabidopsis, DNA methylation is suffition of hypermethylated and silenced superman (sup) cient for gene silencing, but H3-K9 methylation is not. alleles [16]. The KYP protein was shown to methylate H3-K9 in vitro, and kyp mutants were found to reduce Introduction the overall levels of CpNpG methylation in vivo. Two mutations in DNA methyltransferases are also available: Heterochromatic regions of the genome, located pri- met1, a CpG methyltransferase that is homologous to marily in centromeres and telomeres, are generally charmammalian Dnmt1 [19], and cmt3, a chromomethylacterized by increased chromatin condensation and transferase that is important for CpNpG and asymmetric decreased access to regulatory proteins [1]. Many repetmethylation [20, 21]. itive genes, transposable elements, imprinted genes, In this paper, we use chromatin immunoprecipitation and transgenes are silenced in a sequence-independent (ChIP) assays to show that methylation of histone H3manner and have some or all of the characteristics of K9 is preferentially localized to heterochromatin in heterochromatin [2]. Heterochromatin is stably inherited plants. We also show that a mutation in the histone and thus must contain one or more epigenetic marks to methyltransferase gene, KYP, reduces methylation of direct its maintenance during cell division [1]. Two such H3-K9 at all loci tested: the centromeric 180-bp repeats, marks have been under intense scrutiny: DNA methylathe Ta3 and Ta2 retrotransposons, and the hypermethyltion [3] and histone H3-K9 methylation [4]. ated SUP gene. Histone H3-K9 methylation is also re-DNA methylation is found most often at symmetrical duced in lines carrying a mutation in the SWI2/SNF2 homolog, DDM1. Finally, we find that the reduction of DNA methylation itself does not lead to a reduction in 4 Correspondence bation overnight with rotation, 60 l SS DNA/Protein A agarose was added and incubation continued for 2 hr. The agarose beads were We thank Yoo Lee, Michael Huang, and Channy Hyun for technical then washed with 1 ml of each of the following: 2ϫ lysis buffer, 1ϫ assistance, Jim Jackson, Anders Lindroth, and Daniel Zilberman LNDET (0.25 M LiCl, 1% NP40, 1% deoxycholate, 1 mM EDTA, 10 for many stimulating discussions and for critically reviewing this mM Tris [pH 8]; [39]), and 3ϫ TE (10 mM Tris-HCl [pH 8], 1 mM manuscript, and Eric Richards for seeds of the met1 mutant. This EDTA). The immunocomplexes were eluted from the beads with 300 work was supported by National Institutes of Health grant GM60398 l 1% SDS, 0.1 M NaHCO3. A total of 12 l 5 M NaCl was then to S.E.J. added to each tube, and crosslinks were reversed by incubation at 65ЊC for 5-6 hr. Residual protein was degraded by the addition of Received: May 31, 2002 20 g Prot K (in 10 mM EDTA and 40 mM Tris [pH 8]) at 45ЊC for 1 Revised: June 27, 2002 hr, followed by phenol/chloroform/isoamyl alcohol extraction and Accepted: June 28, 2002 ethanol precipitation. Pellets were washed with 70% EtOH and resuspended in 75 l TE. Approximately 1-2 l was used for PCR. Each of the immunoprecipitations was performed at least three independent times, and control precipitations with other antibodies References were done at the same time (e.g., anti-dimethylated-lys4-histone H3, anti-diacetylated-lys9,14-histone H3, anti-tetra-acetylated histone 1. Richards, E.J., and Elgin, S.C. (2002). Epigenetic codes for het-H4, data not shown). erochromatin formation and silencing: rounding up the usual suspects. Cell 108, 489-500. 2. Bird, A.P., and Wolffe, A.P. (1999). Methylation-induced repres-ChIP PCR Quantitative PCR was used to determine the amounts of genomic sion-belts, braces, and chromatin. Cell 99, 451-454. 3. Bird, A. (2002). DNA methylation patterns and epigenetic mem-DNA immunoprecipitated in the ChIP experiments. The primer pairs used were as follows: 180-bp repeats (JP1623: 5Ј-ACCATCAAAGC ory. Genes Dev. 16, 6-21. 4. Zhang, Y., and Reinberg, D. (2001). Transcription regulation by CTTGAGAAGCA-3Ј and JP1624: 5Ј-CCGTATGAGTCTTTGTCTTTG
doi:10.1016/s0960-9822(02)00976-4 pmid:12194816 fatcat:lribrw6q2ffg7acrf5jaeidgdu