SEARCH

SEARCH BY CITATION

Keywords:

  • DNA methylation;
  • epigenetics;
  • histone modification;
  • plant

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. DNA methylation in plants
  5. Histone modifications in plants
  6. Other epigenetic patterns
  7. Determination of DNA methylation and histone modifications’ loci
  8. Perspectives
  9. Acknowledgments
  10. References

Epigenetic research is at the forefront of plant biology and molecular genetics. Studies on higher plants underscore the significant role played by epigenetics in both plant development and stress response. Relatively recent advances in analytical methodology have allowed for a significant expansion of what is known about genome-wide mapping of DNA methylation and histone modifications. In this review, we explore the different modification patterns in plant epigenetics, and the key factors involved in the epigenetic process, in order to illustrate various putative mechanisms. Experimental technology to exploit these modifications, and proposed focus areas for future plant epigenetic research, are also presented.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. DNA methylation in plants
  5. Histone modifications in plants
  6. Other epigenetic patterns
  7. Determination of DNA methylation and histone modifications’ loci
  8. Perspectives
  9. Acknowledgments
  10. References

Epigenetics, as originally coined by Conrad Waddington in 1942, referred to the intermediate factor between genotype and phenotype (Waddington & Kacser 1957). Today epigenetics refers mainly to the heritable changes that do not relate to the DNA sequence but can be reliably inherited from one generation to the next. Several mechanisms for epigenetic inheritance have been found, such as DNA methylation, histone modifications, siRNAs, paramutation, nucleosome arrangement, and others. However, the definition of epigenetics is still being debated by biologists and epigeneticists (Bird 2007). Over the last decade, epigenetic studies have focused mainly on mammals; plants have received much less attention. Some of those mammalian studies have identified human diseases accompanied by epigenetic abnormalities (Jiang et al. 2004; Jirtle 2009).

Because of the rapid advancement of parallel sequencing techniques, some DNA epigenetic information formerly hidden in the genome has been revealed; the first human epigenomic map has recently been presented (Lister et al. 2009). Although there have been fewer epigenetic studies on plants, there is a fair amount of epigenetic information on certain plant models such as Arabidopsis (Zhang et al. 2006; Kasschau et al. 2007; Zhang et al. 2007a,b; Zilberman et al. 2007; Bernatavichute et al. 2008; Cokus et al. 2008; Lister et al. 2008; Zhang et al. 2009), Oryza sativa (Kapoor et al. 2008; Li et al. 2008; Lu et al. 2008) and maize (Wang et al. 2009). Several labs in multiple countries have identified enzymes and homologous factors that participate in the regulation of plant epigenetics. Depending on analytical resolution and data precision, the epigenetic code of the DNA sequence may be determined in the near future. In this paper we have highlighted the key factors that play a significant role in plant epigenetics (mainly DNA methylation and histone modifications), in order to elucidate interactions and putative mechanisms (Table 1).

Table 1.   Performers (enzymes – in bold type) involved in DNA methylation and histone modification in plants
 SitesModification enzymeDemodification enzymeFunction
DNA methylation
GeneCG sitesMaintain: MET1ROS1, DMERegulate gene expression
Establishment: DRM2
TransposonCG, CHG, and CHH sitesMaintain: MET1, CMT3, DRM2ROS1, DMERepress transposons activation
Establishment: DRM2
Histone modifications
AcetylationH2A K5acUnknownHDACs, RPD3-like superfamily: HDA2, HDA5, HDA6, HDA7, HDA8, HDA9, HDA10, HDA14, HDA15, HDA17, HDA18, HDA19, HD-tuin family: HDT1, HDT2, HDT3, HDT4, sirtuin family: SRT1, SRT2Hollender & Liu 2008)Active gene expression, control plant development, and defend stress
H2B K6acUnknown
H2B K11acUnknown
H2B K27acUnknown
H2B K32acUnknown
H3 K9acUnknown
H3 K14acHAG1 (Earley et al. 2007)
H3 K23acUnknown
H3 K27acUnknown
H4 K5acHAM1, HAM2 (Earley et al. 2007)
H4 K8acUnknown
H4 K12acHAG2 (Earley et al. 2007)
H4 K16acUnknown
H4 K20acUnknown
MethylationH3 K4me1/2/3TrxG class of histone methyltransferasesAtJmj4, ELF6 (Jeong et al. 2009)Control specific genes repression
H3 K4me1UnknownLDL1 (Spedaletti et al. 2008)
H3 K4me2ATX1 (Alvarez-Venegas & Avramova 2005), CLF (Doyle & Amasino 2009)LDL1 (Spedaletti et al. 2008)
H3 K4me3ATX1 (Alvarez-Venegas & Avramova 2005), EFS/SDG8 (Kim et al. 2005), CLF (Saleh et al. 2007) 
H3 K9me1/2/3Su(var) class of HMTs, SDG714 (Ding et al. 2007)JmjC-domain and LSD1-type HDMs
H3 K9me1SUVH2 (Naumann et al. 2005), SUVH4 (KYP), SUVH5, SUVH6 (Ebbs & Bender 2006) 
H3 K9me2SUVH2 (Naumann et al. 2005), SUVH4 (KYP), SUVH6 (Ebbs & Bender 2006), SUVR4 (Thorstensen et al. 2006)JMJ706 (Sun & Zhou 2008)
H3 K9me3SDG8 (Dong et al. 2008)JMJ706 (Sun & Zhou 2008)
H3 K27me1/2/3Unknown 
H3 K27me1ATXR5, ATXR6 (Jacob et al. 2009) 
H3 K27me2SUVH2 (Ay et al. 2009) 
H3 K27me3CLF, SWN, MEA (Makarevich et al. 2006; Schubert et al. 2006), SUVH2 (Ay et al. 2009) 
H3 K36me1/2/3Set domain HMTsJmjC-domain HDMs
H3 K36me1Unknown 
H3 K36me2EFS/SDG8 (Zhao et al. 2005) 
H3 K36me3SDG8 (Dong et al. 2008; Grini et al. 2009) 
H4 K20me1/2/3UnknownJmjC-domain HDMs
H4 K20me1SUVH2 (Naumann et al. 2005) 
H4 K20me2Unknown 
H4 K20me3Unknown 
PhosphorylationH2A S129ph, H2A S141ph, H2A S145ph, H2B S15ph, H3 T3ph, H3 T11ph, H3 S10ph,H3 S28phKinases (Houben et al. 2007)Phosphatases (Houben et al. 2007)Cell mitosis and apoptosis
UbiquitinationH2B K143ub1Ring-type E3 ligases HUB1, HUB2, UBC1, UBC2 (Gu et al. 2009)SUP32/UBP26 (Sridhar et al. 2007)Control plant development and defend stress

DNA methylation in plants

  1. Top of page
  2. Abstract
  3. Introduction
  4. DNA methylation in plants
  5. Histone modifications in plants
  6. Other epigenetic patterns
  7. Determination of DNA methylation and histone modifications’ loci
  8. Perspectives
  9. Acknowledgments
  10. References

DNA methylation is one of the most important processes in epigenetics as it results in a direct modification of the DNA sequence through mitosis and amitosis. The same modification pattern found in parental cells can be inherited in subsequent generations. For the most part, DNA methylation occurs at the fifth carbon position of a cytosine ring. DNA methylation in mammals is restricted to symmetrical CG sequences (Bird 2002), while plant DNA methylation is found at CG sites, CHG sites (H indicating for A, C or T) and CHH sites (an asymmetrical site). Genome-wide analysis of DNA methylation in plants by McrBC digestion, methylcytosine immunoprecipitation (MeDIP), or sequencing of bisulfate-treated DNA indicate that transposons are heavily methylated at both CG and non-CG sites, whereas the density of methylcytosines in genes is much lower and limited to CG sites (Zhang et al. 2006; Zilberman et al. 2007; Cokus et al. 2008; Lister et al. 2008). METHYLTRANSFERASE1 (MET1), the counterpart of mammalian DNMT1 (Mammalian DNA methyltransferase1), maintains DNA methylation at CG sites (Finnegan & Dennis 1993; Kankel et al. 2003; Saze et al. 2003). Plant-specific DNA methyltransferase, CHROMOMETYLASE3 (CMT3) can maintain DNA methylation at CHG sites (Jackson et al. 2002). DOMAINS REARRANGED METHYLASE 1 and 2 (DRM1/2), together with siRNAs, maintain DNA methylation at CHH sites (Zhang 2008).

There are two primary consequences of DNA methylation: (i) methylcytosines in the gene body of the genome play a role in regulating the gene expression and (ii) methylcytosines in repeat sequences, such as transposable elements, are believed to prevent repetitive sequences from compromising normal genome function (Miura et al. 2001; Simon et al. 2005). In this paper we have focused on DNA methylation in gene bodies and DNA methylation of transposable elements (TEs) to illustrate the interactions between different enzymes and their regulators.

DNA methylation of genes

DNA methylation of genes is a highly flexible method to control gene expression or repression in an organism containing the same DNA sequence information in every cell. Busslinger et al. (1983) showed that transfecting methylated DNA into mouse cells can inhibit gene transcription. Subsequent research suggested that, when all cytosine is methylated, genes were silenced either in the 5′ portion or in the 3′ portion (Keshet et al. 1985). Nevertheless, methylcytosines in the promoter regions were more efficient in repressing gene transcription than in the coding regions (Hsieh 1997). Genome-wide massive sequencing of Arabidopsis DNA methylation loci revealed that body-methylated genes demonstrate substantially less methylation than transposons, and are limited to the CG sites (Zhang et al. 2006; Zilberman et al. 2007; Cokus et al. 2008; Lister et al. 2008). However, a recent study with Petunia showed that the expression of ect-pMADS3, a class-C floral homeotic gene, was active due to RNA-directed DNA methylation (RdDM) at CGs (Shibuya et al. 2009), which suggests that DNA methylation may play varied roles in different regions of the genome.

Methylation of promoter DNA is extremely important in mammals because the genes are generally related to ontogenesis and differentiation (Zemach & Grafi 2003). In plants, many tissue-specific gene promoters are regulated through DNA methylation (Zhang et al. 2006; Johnson et al. 2007). A recent study noted that DNA methylation in the promoter regions mostly occurs at CG sites and depends on MET1 and DRM2 (Berdasco et al. 2008). With regard to DNA methylation at CG sites maintained by MET1, it is likely that there are additional demands for RdDM at certain loci to modify the DNA sequence, with the assistance of DRM2. Analysis of rice gene methylation also showed that the rice genome contains more methylated promoters than does the Arabidopsis genome (Li et al. 2008). This raises an interesting question about the comparability of dicotyledon and monocotyledon DNA methylation patterns.

High-throughput DNA methylation sequencing in Arabidopsis thaliana, Oryza sativa, and maize (Zea mays) genomes also revealed that gene coding regions may be candidates of moderate methylation (Zhang et al. 2006; Zilberman et al. 2007; Cokus et al. 2008; Li et al. 2008; Lister et al. 2008; Wang et al. 2009). Although DNA methylation is usually considered a process for silencing genes, Zhang et al. (2006) found that many house-keeping genes that were methylated in coding regions actually showed a higher level of expression. It is significant that protein-coding gene methylation is polymorphic in different Arabidopsis ecotypes (Vaughn et al. 2007). These data indicate that while methylation in coding regions is not directly related to gene expression, there are some methylation patterns, such as DNA methylation in promoters, which can influence gene expression levels and are pivotal for plant development.

Intragenic DNA methylation mechanisms are essentially those that repress gene expression; how those mechanisms operate remains an important question. It is now well established that DNA methylation can induce chromatin remodeling, and methylation at some gene promoters can recruit methylcytosine-binding proteins such as KRYPTONITE (KYP), histone H3 lysine 9 (H3K9), methyltransferase and VARIANT IN METHYLATION 1 (VIM1) (Woo et al. 2008) to bind to DNA. Transcription initiation is blocked because Pol II cannot gain access to the initial site. Recent research revealed that the ibm1 mutation results in hyper-methylation in thousands of genes (Miura et al. 2009). Gene methylation is restricted to CG sites in wild-type Arabidopsis, whereas many CHG sites are methylated in the ibm1 mutation. These data suggest that one or more critical factors are playing a pivotal role in directing DNA methylation to correct gene sites and preventing other methyltransferases from modifying gene regions such as the CHG or CHH sites. In conclusion, there is strong evidence that disparate factors work in an orderly manner to control gene expression by adding methyl groups to the correct targets.

DNA methylation of transposons

Genome-wide sequencing has revealed that TEs and other repetitive elements constitute a large proportion of most eukaryotic genomes. Active TEs can insert into protein coding regions and influence the gene expression of nearby regions, which threatens to disrupt normal genome function. Eukaryotic organisms have evolved mechanisms to silence and immobilize TEs, such as RNA interference and epigenetic DNA methylation. In fact, TEs and other repeat sequences are the main targets of cytosine methylation and various studies have demonstrated that DNA methylation primarily functions as a TE silencer (Suzuki & Bird 2008; Sekhon & Chopra 2009). For example, TEs with inverted repeats at termini can be targeted for DNA methylation when they are transformed in Arabidopsis (Zilberman et al. 2004).

DNA methylation at CG, CHG and CHH sites in TEs is conducted by MET1 and CMT3 when it is maintained between cell generations (Kato et al. 2003). Meanwhile, DRM2 drives DNA methylation of TEs at CHH sites (Teixeira & Colot 2009). In contrast, Miura et al. (2009) identified a jmjC domain-containing protein IBM1 (increase in BONSAI methylation 1) that can limit the degree to which DNA methylation targets CG sites in genes, but not in transposons. This suggests that the methylases do not necessarily select a special site to add a methyl group, but certain special regulators act to protect functional sites against methylation, thus maintaining their activity.

Although significant questions remain, valuable research has clarified the mechanisms that drive DNA methylation at TEs, revealing that RdDM serves as one of the most important control mechanisms for various de novo DNA methylation patterns, in order to regulate gene expression or repress the activation of TEs.

RNA-directed DNA methylation

Until very recently, researchers were puzzled by how DNA methylation is guided to specific sequences that are in need of genome modification. It is now clear that RdDM accomplished this feat through an RNA complementary pathway. RdDM was first observed in tobacco infected with viroids. When these circular pathogenic RNA molecules were infected, homologous DNA sequences became methylated (Wassenegger et al. 1994). Several studies reported that promoter methylation may be introduced by this smRNA combined mechanism (Dorweiler et al. 2000; Okano et al. 2008), indicating that genes are also regulated via this de novo DNA methylation system. However, other evidence suggests that TEs and other repeat sequences are more likely to be targeted by this pathway (Perez-Hormaeche et al. 2008). Considering that gene expression can be regulated by nearby repetitive elements, RdDM appears to be the most likely method for indirectly controlling gene activation or inhibition. In addition, through this dynamic regulation pathway, TEs and other deleterious repetitive elements are effectively silenced.

More details of the RdDM pathway have clarified the process, particularly the roles played by Pol IV (formerly Pol IVa) and Pol V (formerly Pol IVb). Pol II and Pol III are present in eukaryotes; Pol IV and Pol V are specific to plants and are involved in the siRNA silencing mechanism (Herr et al. 2005; Kanno et al. 2005; Onodera et al. 2005; Pontier et al. 2005). In RdDM, Pol IV is necessary for siRNA production (Wierzbicki et al. 2008) and Pol V can guide arogonaute protein4 (AGO4) to chromatin (Wierzbicki et al. 2009). According to at least one model, RdDM processes are quite distinct, in that through RNA-directed RNA polymerase (RDR2) and RNA polymerase IV (Pol lV), specific single-stranded RNAs form dsRNAs (Huettel et al. 2007). The dsRNAs are then cleaved by Dicer-like 3 (DCL3) into ∼20 to 24-nt siRNAs, one of which is uploaded to AGO4 containing RNA-induced silencing complex (RISC) (Henderson & Jacobsen 2007). In addition, Hua Enhancer1 (HEN1) participates in this procedure, methylating siRNAs at their 3′ ends (Matzke et al. 2009). Furthermore, the AGO4-siRNA-KTF1 (KOW-domain transcription factor1)-Pol V (RNA polymerase V)-DRM2 complex methylates the DNA sequence complementary to the siRNA (Matzke et al. 2009; Wierzbicki et al. 2009). Investigations have also revealed that DEFECTIVE IN MERISTEM SILENCING 4 (DMS4), also known as RDM4, is indispensable in RdDM, indicating that it assists in the regulation of several RNA polymerases (He et al. 2009; Kanno et al. 2010).

DNA demethylation

Cytosine demethylation depends on a maternal gene encoding a DNA glycosylase-lyse, DEMETER (DME), which can excise the base five methylcytosine directly (Choi et al. 2002; Gehring et al. 2006; Morales-Ruiz et al. 2006). DME belongs to an Arabidopsis gene family that includes the somatically-expressed gene REPRESSOR OF SILENCING 1 (ROS1). ros1 mutants showed specific locus DNA hypermethylation and enhanced TGS. Through genome analysis of demethylases in Arabidopsis, many transposons are the targets of ROS1 (Penterman et al. 2007; Zhu et al. 2007). Also, genome-wide analysis of Arabidopsis endosperm has revealed that DME regulates the whole endosperm genome CG demethylation, which means imprinting in plants is accompanied by DNA demethylation (Hsieh et al. 2009). Recently, ROS3, containing an RNA-recognition motif (RRM), was found to bind on ROS1 which, it can be speculated, reflects putative DNA demethylation directed by smRNAs or perhaps the RdDM pathway (Zheng et al. 2008).

Histone modifications in plants

  1. Top of page
  2. Abstract
  3. Introduction
  4. DNA methylation in plants
  5. Histone modifications in plants
  6. Other epigenetic patterns
  7. Determination of DNA methylation and histone modifications’ loci
  8. Perspectives
  9. Acknowledgments
  10. References

In addition to DNA methylation, histone N-terminal tail modifications such as acetylation, methylation, phosphorylation, ubiquitination, ribosylation and biotinylation constitute an attractive area in epigenetics. The nucleosome is considered a barrier to regulatory factors approaching the DNA sequence (Berger 2007) and histones are fundamental components of chromatin, and thus the nucleosome. Eight histone proteins form a nucleosome, which contains two copies each of H2A, H2B, H3 and H4 proteins (Fig. 1). Amino acids on N-peptide tails protruding from the nucleosome of H3 and H4 are easier to modify than other histone amino acids. By various modifications, recruited enzymes, such as chromatin remodeling ATPases, use adenosine tri-phosphate (ATP)-derived energy to catalyze nucleosome arrangement and chromatin remodeling; activators or inhibitors can then bind to DNA sequences. Gene expression is upregulated by acetylation, phosphorylation and ubiquitination, while it is repressed by dimethylation of H3K9 (histone H3 Lysine 9) and H3K27 (histone H3 Lysine 27), biotinylation and sumoylation. Clearly, histone modifications function as gene regulators (Vaillant & Paszkowski 2007). In addition, studies on different plant genera have shown that histone modifications play a critical role in plant development (Wagner 2003) and in plant defense mechanisms (Sokol et al. 2007; Kim et al. 2008). Investigations of diverse modifications and genome-wide histone modifications in plants have revealed that one histone modification can interact with another histone modification or DNA methylation (Zhang et al. 2007a,b; Bernatavichute et al. 2008; Li et al. 2008; Wang et al. 2009; Wu et al. 2009; Zhang et al. 2009).

image

Figure 1.  Brief illustration of DNA methylation and histone modifications in the plant genome. The nucleosome is composed of two copies each of H2A, H2B, H3 and H4 (one copy shown, each histone is in a different color). Amino acid tails protruding from the histone can be modified by diverse marks. The modification enzymes on the histone are listed in Table 1. By MET1, CMT3 and DRM2, DNA methylation patterns can be maintained through several generations both on genes and transposons. Associated with siRNA, de novo methylation of transposons is established by a RdDM mechanism. A methyl group on methylcytosine is removed by glycosylase (ROS1/DME) at any site it locates.

Download figure to PowerPoint

H3 Lysine acetylation

Lysine acetylation is thought to co-occur with transcription activation (Benhamed et al. 2008). Recent studies showed nucleosomes with lysine-acetylated residues usually rearrange during plant development (Li et al. 2009), and acetylation modification is related to root elongation (Krichevsky et al. 2009), flowering (He et al. 2003; Bond et al. 2009) and cold tolerance (Zhu et al. 2008). About 15 histone acetyltransferases (HAT) have been found in Arabidopsis. They are classified into several families, such as the GNAT family, CBP/p300 family and TAFII family. One of the lysine acetyltransferases, GCN5, was found to interfere with the miRNA pathway in Arabidopsis (Kim et al. 2009). This suggests that miRNA silences gene expression in diverse ways because GCN5 is required for the expression of a large number of genes. Histone deacetylases (HDAC) in plants belong to three families: RDP-like, HD-tuin and sirtuin (Hollender & Liu 2008). Jiao et al. (2007) conducted a review of existing data and found that a variety of light-induced developmental processes in plants require acetylation and deacetylation of specific lysine residues. Acetylation of lysine residues is very flexible and may play a versatile role in plant life cycles.

H3 Lysine methylation

The majority of covalent modifications on histones are H3 Lysine methylation. Histone lysine residues can be mono-, di- or trimethylated, where each methyl state confers different meaning from a biological standpoint (Cloos et al. 2008). Typically, H3K9me2 (histone H3 Lysine 9 dimethylation), H3K9me3 and H3K27me3 downregulate gene expression, while H3K4me1, H3K4me2, H3K4me3, H3K36me2 and H3K36me3 upregulate target gene expression (Zhou 2009). In plants, seven classes of histone methyl transferases (HMTs), which are called SET (Su(var)3-9, E(Z) and Trithorax)-domain proteins, catalyze methylation of diverse targets (Ng et al. 2007). For example, H3K9 is mono-, di- and trimethylated by the Su (var) class of HMTs; H3K27 is mono- and dimethylated by the E (Z) class of HMTs (Polycomb Repressive Complex 2) (Pfluger & Wagner 2007). Thirty-nine SET genes have been found in the Arabidopsis genome. While histone methylation was believed to be an irreversible process, recent experimental results have shown that enzymes such as FLOWERING LOCUS D (FLD), LSD1-LIKE 1 (LDL1) and LSD1-LIKE 2 (LDL2), that are involved in the regulation of plant flowering, are histone demethylases. Sun & Zhou (2008) reported that Jumonji C (jmjC) domain-containing proteins, associated with flower development, could cause H3K9 demethylation in rice. In atxr5 and atxr6 mutations, H3K27 monomethylation is lost and activity is detected in the heterochromatin regions (Jacob et al. 2009). This observation supports the experimental result that H3K27me1 is enriched at the heterochromatin and is a silence marker (Mathieu et al. 2005; Fuchs et al. 2006).

Other histone modifications

In addition to H3 Lysine acetylation and methylation, various modification patterns have been found in Arabidopsis such as acetylation on H2A Lysine5, H2B Lysine6, 11, 27, 32, H4 Lysine5, 8, 12, 16, 20; methylation on H4 Lysine20; phosphorylation on H2A Serine129, 141, 145, H2B Serine15, H3 Threonine3, 11 and Serine10, 28; ubiquitination on H2B Lysine143 (Zhang et al. 2007a; Kim et al. 2008; Cerutti & Casas-Mollano 2009). Phosphorylation on histone occurs mainly during cell mitosis (Cerutti & Casas-Mollano 2009). Zhang et al. (2007a) reported that alteration of H2B Serine phosphorylation is connected to apoptosis. Ubiquitination of H2B Lysine is more complex and is catalyzed by two RING E3 ligases, and removed by deubiquitinases. Through this modification plants can defend against necrotrophic fungal pathogens as well as regulate seed development (Luo et al. 2008; Dhawan et al. 2009). Cao et al. (2008) found that hub1 and hub2 mutants lost H2Bub1 and flowered early, indicating the essential nature of ubiquitination in plants. Histone modification enzymes are listed in Table 1.

Interplay between DNA methylation and histone modifications

Studies of interactions between DNA methylation and histone modification in plants have confirmed their significant role in jointly silencing gene expression (Fuks 2005). The two processes may share the same feedback loop model to regulate gene activation or inactivation, as evidenced by factors such as DDM1 (decrease in DNA methylation) (Jeddeloh et al. 1999), jmjC-domain-containing proteins and KYP engaged in both DNA methylation and histone modifications (Johnson et al. 2007; Iwase et al. 2007; Saze et al. 2008; Miura et al. 2009). High-resolution DNA methylation and histone H3K4 di- and trimethylation sequencing of rice reveal two entire chromosomes and two centromeres; enriched DNA methylation is evident while H3K4 methylation does not occur. However, protein coding genes embody both DNA methylation and H3K4me2/H3K4me3 modifications. The outcomes unambiguously mirror a distinct regulatory model driven by DNA methylation and H3K4 methylation in intragenic sequences and transposable sequences. Since DNA methylation maintenance is an immediate process following DNA duplication, certain methylated cytosines may recruit enzymes and other factors involved in histone modification to establish and maintain the modifications. Published reports support this possibility (Lorincz et al. 2004; Okitsu & Hsieh 2007). In addition, histone modifications are more easily altered than epigenetic mark-DNA methylation that is more stable and resistant to change. With the smRNA targeting and silencing mechanism, through DNA methylation and histone modification, plants can protect genomes from activation of malignant transposable and repetitive elements, resist virus or viroid invasion, and regulate the expression of special genes under the influence of various environmental factors.

Other epigenetic patterns

  1. Top of page
  2. Abstract
  3. Introduction
  4. DNA methylation in plants
  5. Histone modifications in plants
  6. Other epigenetic patterns
  7. Determination of DNA methylation and histone modifications’ loci
  8. Perspectives
  9. Acknowledgments
  10. References

Other epigenetic patterns such as smRNAs, imprinting and paramutation are also essential for plant inheritance and development. The roles played by smRNAs in plants have been reviewed in several comprehensive articles (Carthew & Sontheimer 2009; Voinnet 2009). Here we introduce and briefly discuss imprinting and paramutation in plants. Although they are more or less related to DNA methylation and tandem repeats, imprinting and paramutation follow more specific rules in different stages of the plant life cycle. Imprinting is the influence of the maternal or paternal genomes in gene expression by the offspring. Epigenetic marks (DNA methylation and histone modification) and correlative regulators (Polycomb group proteins, DNA methyltransferases, DNA demethylating, DNA glycosylases) establish and maintain plant gene imprinting (Huh et al. 2008). We consider endosperm development in Arabidopsis a typical example that demonstrates imprinting. Both maternal and paternal alleles of the FLOWERING WAGENINGEN (FWA) and the C2H2 zinc-finger protein FERTILIZATION INDEPENDENT SEED2 (FLS2) are silenced by methylation on CG sites in the central cell and sperm. Through demethylation by DME DNA glycosylase, FWA and FLS2 alleles are activated in the central cell. In contrast, the methylation state of the endosperm can be changed only when one of the two maternal alleles is unmethylated in the central cell and endosperm. Imprinting is critical to seed development (Huh et al. 2007; Nowack et al. 2007; Berger & Chaudhury 2009).

Paramutation is an interaction between alleles that can result in heritable expression change of one allele. Ground-breaking research by R.A. Brink about 50 years ago revealed that, in Zea mays, the maize locus red1 (r1) interacted with its allele and controlled the pigmentation of corn kernels. Subsequent studies, particularly on the maize b1 locus, identified the molecular basis for the control, in that b1 encodes a transcription factor that is needed for accumulation of the pigment anthocyanin (Chandler & Stam 2004). The B’ allele induces a light purple plant while B-I induces a dark purple plant. The B-I transcription level is 10–20 × higher than that of the B’ allele. However, the methylation pattern of the promoters and coding sequences are no different between the B’ and B-I alleles. When combined, the paramutagenic B’ allele can interact with paramutagenic B-I and change it to the B’ allele, resulting in a light purple plant. This paramutation can be inherited in the progeny. Studies have shown that RdDM plays an important role in the paramutation process; however, smRNAs alone are not sufficient to establish paramutation (Chandler 2007); further research is required to identify the significant factors involved.

Determination of DNA methylation and histone modifications’ loci

  1. Top of page
  2. Abstract
  3. Introduction
  4. DNA methylation in plants
  5. Histone modifications in plants
  6. Other epigenetic patterns
  7. Determination of DNA methylation and histone modifications’ loci
  8. Perspectives
  9. Acknowledgments
  10. References

There are several methods for detecting methylcytosine, some of them focusing on single genes and some on the wide-scale genome. Depending on distinct reaction mechanisms and processes, the methods are divided into two classes. The first is the bisulfate treatment method. Unmethylated cytosine can be converted to uracil while methylcytosine remains unaltered. In a polymerase chain reaction (PCR) reaction, uracil is a template-like thymidine and all original cytosine residues are converted to thymidine. Ultimately, the PCR products are sequenced and mapped onto genome data in order to locate unchanged cytosine, which is the methylcytosine. The advantage of this method is that DNA integrity is sustained; however, the rate of conversion must be controlled in order to minimize bias. The second class is a biological process that needs some methylation-sensitive enzymes such as HpaII and NotI. The DNA digestion process is often performed with isoschizomers as well. Therefore, one of the two enzymes is methylation loci sensitive and the other recognizes the same DNA sequence but is independent of methylcytosine. DNA methylation loci can be determined through analysis of the different fragment patterns. This method is easy to perform and robust. However, failure to properly control the activity of the isoschizomers on their targets could lead to incomplete DNA digestion. The development of new sequencing technologies has renewed interest in these methods.

Antibody techniques were used for several years to detect histone modifications. Unfortunately, the diversity of modification patterns severely restricted the design of specific antibodies that could be used for genome-wide analysis (Villar-Garea & Imhof 2006). The evolution of analytical techniques led to the use of mass spectrometry to locate the exact modification sites, allowing for a more widely applicable technique for evaluating histone modifications. Villar-Garea et al. (2008), for example, used a mass spectrometric method to investigate histone modifications. Most recently, histone variant analyses were also conducted using mass spectrometry (Zhang et al. 2007a; Wu et al. 2009).

Perspectives

  1. Top of page
  2. Abstract
  3. Introduction
  4. DNA methylation in plants
  5. Histone modifications in plants
  6. Other epigenetic patterns
  7. Determination of DNA methylation and histone modifications’ loci
  8. Perspectives
  9. Acknowledgments
  10. References

Although the epigenetic code has not been fully defined, there is an increasing number of opportunities for its application, particularly with the rapid development of genome-wide high-resolution sequencing. In mammals, epigenetic modification can control ontogenesis and can be changed by some diseases such as cancer and some mental diseases. In the case of double fertilization in plants, epigenetic modification is quite distinct between the zygote and primary endosperm cell. Some epigenetic research of endosperm reveals that it is a typical model for imprinting studies. Moreover, enzymes of de novo DNA methylation in plants are more diverse than in mammals since plants accomplish methylation via a siRNA combined mechanism. Analysis of histone modifications in Arabidopsis show that some nonconserved sites on different histones are modified (Zhang et al. 2007a). Plants are excellent models for studying epigenetics, and epigenetic performers in plants are diverse and intricate as they participate in both plant development and regulation of the stress response. The main modifications and pathways that have been clearly detected and discussed in this article are shown in Figure 1. While there are surprisingly few differences in DNA sequences among advanced organisms, gene expression can be radically different. Studies on epigenetics can deepen our understanding of how epigenetic-regulated DNA expression controls plant development and stress tolerance. DNA methylation plays a major role in either regulating gene expression or guiding other epigenetic mechanisms to function at the right time and place. However, many questions remain, such as how DNA methylation influences other modification patterns and what mechanisms are engaged to realize those modifications. Further, it has been shown that the same DNA methylation pattern may function differently under different circumstances, and that DNA methylation is very dynamic at different stages of plant growth and development (Law & Jacobsen 2009). With regard to histone modifications like lysine acetylation and methylation, and considering that there have been several studies on N-tail modifications, attention should be focused on how these diverse modifications function jointly in gene regulation or cell differentiation, and on the triggers that induce these modifications. The review completed by Cedar & Bergman (2009) on the correlations between DNA methylation and histone modification in mammals provided significant examples that could be applied to plant research. Although some epigenetic analyses in plants have uncovered fascinating models of the interplay between DNA methylation and histone modifications, comparative research on plant epigenetics will be much more complex because plants are easily affected by the external environment and there are multiple epigenetic responses in plants.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. DNA methylation in plants
  5. Histone modifications in plants
  6. Other epigenetic patterns
  7. Determination of DNA methylation and histone modifications’ loci
  8. Perspectives
  9. Acknowledgments
  10. References

This work was supported by the National Natural Sciences Foundation of China (30771326, 30971743), the National High Technology Research and Development Program of China (“863” Program) (2008AA10Z125), and the Program for New Century Excellent Talents in University of China ([NCET-07-0740). We are grateful to our German colleagues for the kind support and communication under the MoST and BMBF collaboration program (2009DFA32030, CHN 08/001).

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. DNA methylation in plants
  5. Histone modifications in plants
  6. Other epigenetic patterns
  7. Determination of DNA methylation and histone modifications’ loci
  8. Perspectives
  9. Acknowledgments
  10. References