The Control of Natural Variation in Cytosine Methylation in Arabidopsis

Corresponding author: Eric J. Richards, Washington University, 1 Brookings Dr., St. Louis, MO 63130., richards{at}biology.wustl.edu (E-mail)

Communicating editor: S. HENIKOFF

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

We explore the extent and sources of epigenetic variation in cytosine methylation in natural accessions of the flowering plant, Arabidopsis thaliana, by focusing on the methylation of the major rRNA gene repeats at the two nucleolus organizer regions (NOR). Our findings indicate that natural variation in NOR methylation results from a combination of genetic and epigenetic mechanisms. Genetic variation in rRNA gene copy number and trans-acting modifier loci account for some of the natural variation in NOR methylation. Our results also suggest that divergence and inheritance of epigenetic information, independent of changes in underlying nucleotide sequence, may play an important role in maintaining natural variation in cytosine methylation.

METHYLATION of cytosine is a common DNA modification widely distributed in both prokaryotic and eukaryotic kingdoms. Cytosine methylation is an important epigenetic mark that modifies the information content of the underlying genetic sequence. Perturbation of cytosine methylation, in mutants or by inhibitor treatment, leads to developmental defects in organisms ranging from plants and fungi to mammals (LI et al. 1992 ; MALAGNAC et al. 1997 ; OKANO et al. 1999 ; MARTIENSSEN and COLOT 2001 ). The phenotypic abnormalities are caused by the accumulation of genetic mutations (CHEN et al. 1998 ; MIURA et al. 2001 ) as well as heritable epigenetic alterations in DNA methylation (KAKUTANI et al. 1996 ; JACOBSEN et al. 2000 ; SOPPE et al. 2000 ) that lead to changes in gene expression. Cytosine methylation also interacts with other epigenetic modifications, such as histone deacetylation and histone methylation (RICHARDS and ELGIN 2002 ), to establish and maintain chromatin structures that influence transcription and recombination.

Both forward and reverse genetic approaches have identified important components of cytosine methylation systems in fungi, plants, and mammals. Common targets for reverse genetics are several different classes of cytosine DNA methyltransferases (LI et al. 1992 ; FINNEGAN et al. 1996 ; RONEMUS et al. 1996 ; OKANO et al. 1999 ; BESTOR 2000 ), which add methyl groups to the five position of cytosine residues after DNA replication. Genetic screens have uncovered other trans-acting proteins necessary for maintenance of cytosine methylation, including a subset of SWI2/SNF2 proteins [Arabidopsis DDM1 (VONGS et al. 1993 ; JEDDELOH et al. 1998 ), mouse LSH (DENNIS et al. 2001 ), and human ATRX (GIBBONS et al. 2000 )], and a histone H3 Lys9 methyltransferase [Neurospora DIM-5 (TAMARU and SELKER 2001 )]. In contrast, forward genetic screens have not identified cis-acting signals for cytosine methylation, although reverse genetic approaches have uncovered DNA sequence elements that may influence cytosine methylation levels (YATES et al. 1999 ; MIAO et al. 2000 ). Despite this progress, the mechanisms that specify and maintain cytosine methylation patterns and levels remain poorly understood.

As an alternative to forward or reverse genetics, we exploited natural variation and quantitative trait locus (QTL) analysis to study the mechanisms by which cytosine methylation is controlled and inherited. The natural variation/QTL approach has the advantage of examining biologically relevant variation, including alleles of small effect, which have been maintained in natural populations. This strategy has the additional advantage of uncovering interactions between different loci involved in control of cytosine methylation. Genetic analysis using natural strains also provides the opportunity to examine the stability of different parental cytosine methylation patterns introduced into the same nucleus in hybrids.

We have chosen the model flowering plant, Arabidopsis thaliana, as a study system for this work, taking advantage of available natural accessions and genetic resources developed for this species (ALONSO-BLANCO and KOORNNEEF 2000 ). We found significant natural variation in cytosine methylation, particularly in the arrays of rRNA genes [nucleolus organizer regions (NOR)], which constitute 6% of the genome (COPENHAVER and PIKAARD 1996B ). Our genetic analyses indicate that NOR cytosine methylation is controlled by a combination of factors, including trans-acting modifiers, variation in rRNA gene copy number, and epigenetic inheritance of parental methylation patterns.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Plant materials:
All plants were grown in a greenhouse or environmental chambers under standard growth conditions with 16 hr/day of light. The recombinant inbred (RI) lines used in this study as well as the natural accessions used in the survey were obtained from the Arabidopsis Biological Resource Center (ABRC) at The Ohio State University. The ABRC stock numbers are given in Table 1. The natural accessions were collected from different locations in Central Eurasia and North America (ALONSO-BLANCO and KOORNNEEF 2000 ). In this study, single-seed descent lines derived from the natural collections were used, with the exception of the accession from Martuba (Mt-0; ABRC stock no. CS1380). Further, we took every natural accession line through at least two additional generations of single-seed descent starting from the seed aliquots from the ABRC. Because A. thaliana is predominantly self-fertilizing, these natural accessions correspond to independent genotypic variants.

Southern blot analysis:
Leaf tissue was collected before flowering, and DNA was extracted from individuals as previously described (JEDDELOH et al. 1998 ). DNA for the initial Southern survey (Fig 1) was prepared from pooled tissue using QIAGEN (Valencia, CA) columns and protocols. Genomic DNA was digested according to the manufacturer's (New England Biolabs, Beverly, MA) recommendations. Southern blots were prepared and hybridized as described previously (JEDDELOH et al. 1998 ), using a radiolabeled fragment encompassing the 5.8S and 25S rRNA gene [pARR17 insert (VONGS et al. 1993 ), probe 1, Fig 1A]. NOR methylation levels were determined from phosphorimager (Bio-Rad, Richmond, CA) files using the Quantity One (Bio-Rad) software. The percentage of the hybridization signal in each lane above 1.0 kb corresponds to the percentage of methylation, based on the reference to the hybridization signal distribution in ddm1 (strain Columbia), a mutant that demethylates the rRNA genes (VONGS et al. 1993 ; see Fig 1B). Methylation of the polymorphic intergenic spacer (IGS) region in Fig 5 was monitored using a hybridization probe corresponding to a 1.7-kb EcoRI fragment in the IGS (probe 2, Fig 1A). We investigated methylation at a single-copy genomic region using a 3.3-kb EcoRI fragment derived from bAt105 (PRUITT and MEYEROWITZ 1986 ) as a hybridization probe. This DNA fragment encompasses a portion of the gene MHC9.7, the entire open reading frame of MHC9.8, as well as intergenic sequences (GenBank accession no. AP001305). The fragment itself includes three HpaII sites; however, due to methylation at some of the neighboring HpaII sites, up to eight different HpaII sites are assayed by this probe.

View larger version (31K): In this window In a new window Download PPT slide   Figure 1. Cytosine methylation of the Arabidopsis NOR. (A) The NOR at the upper tips of chromosomes 2 and 4 are composed of tandem arrays of a 10-kb repeat unit encoding the 18S, 5.8S, and 25S rRNAs. A bent arrow shows the location of the transcription start site, and the positions of the two hybridization probes used in this study are indicated below the monomer repeat unit. (B) Natural variation in NOR methylation. HpaII-digested genomic DNA prepared from pooled tissue was size fractionated by gel electrophoresis and transferred to a nylon membrane. The membrane was hybridized with radiolabeled probe 1 (see A) and the signal was detected by autoradiography. Percentage of NOR methylation was calculated by measuring the percentage of hybridization signal in each lane (based on phosphorimaging) that corresponded to restriction fragments >1 kb. The 1-kb limit was chosen on the basis of the distribution of demethylated NOR hybridization signal seen in the Columbia ddm1 mutant sample. (C) Qualitative differences in NOR hybridization signal (probe 1) between strains Hi and Ler are shown by comparing phosphorimaging intensity plots. The arrow indicates NOR restriction fragment or fragments of intermediate size present in Hi but absent in Ler.

View larger version (22K): In this window In a new window Download PPT slide   Figure 2. QTL analysis of NOR methylation in a Cvi/Ler RI population. QTL associated with NOR methylation are mapped onto the five Arabidopsis chromosomes. Genetic distance on each chromosome is shown in morgans; 0 morgan is at the top of each linkage map. Strength of the association is reflected in the likelihood ratio (LR), which is directly related to the LOD score; LOD = 0.217 LR. The dashed line represents the 5% significance threshold calculated by a permutation test (1000 replicates). The dots indicate the position of the amplified fragment length polymorphism (AFLP) marker used in the ANOVA presented in Table 2.

View larger version (14K): In this window In a new window Download PPT slide   Figure 3. Epistatic interaction between QTL3a and QTL4a. NOR cytosine methylation (means ± standard error) is plotted relative to QTL4a genotype. Solid line denotes this relationship in plants homozygous for the Ler allele at QTL3a; the dashed line plots the relationship in plants homozygous for the QTL3a Cvi allele. The difference in the slopes of the lines indicates that the genotype at one QTL influences the phenotypic effect of the other QTL.

View larger version (15K): In this window In a new window Download PPT slide   Figure 4. Direct correlation between rRNA gene copy number and NOR methylation. The percentages of NOR methylation (Table 1) values were plotted against normalized rRNA gene copy number, which was determined by Southern analysis as described in MATERIALS AND METHODS. Error bars represent standard error; in many cases, the standard error is too small to be visible. The line is a linear regression of the data (linear correlation coefficient, R = 0.83). Removal of the Hi data point changes the linear correlation coefficient to 0.79 (P < 0.02).

View larger version (63K): In this window In a new window Download PPT slide   Figure 5. Inheritance of the NOR hypomethylated state in Cvi/Ler hybrids. Genomic Southern analysis was used to examine the cytosine methylation state of the NOR intergenic region in individuals of the following genotypes: Cvi, Ler, and Ler x Cvi. As a control, a mixture of equal amounts of Cvi and Ler genomic DNA (Mix) was treated in parallel. Genomic DNA samples were digested with EcoRI, in combination with the methylation-sensitive HpaII or its isochizomer, MspI (which will cut CmCGG). After transfer, the membrane was hybridized with a radiolabeled intergenic region fragment, probe 2 (see Fig 1A). The arrow indicates a hypomethylated 2.0-kb Cvi-specific EcoRI fragment from the intergenic region that remains hypomethylated in F1 hybrids. The direction of the cross did not affect the results (data not shown).

rRNA gene copy number:
rRNA gene copy number was determined using Southern blot analysis. Genomic DNA was digested with EcoRI (New England Biolabs), and Southern blots were prepared as described above. The blots were hybridized with a single-copy probe containing MHC9.7 (AT3g21390) and MHC9.8 (AT3g21400) and subsequently rehybridized with the pARR17 probe (probe 1, Fig 1A; VONGS et al. 1993 ). The hybridization signals were quantified using phosphorimager analysis (Bio-Rad), and the ratios between the rRNA gene signal and single-copy signal were calculated. An average of three measurements for each strain was determined, which was then normalized to the ratio for strain Columbia.

QTL analysis:
The QTL Cartographer software suite was used to perform QTL analyses (BASTEN et al. 1994 , BASTEN et al. 2001 ). The genotypic data for all the RI lines is available at http://nasc.nott.ac.uk. Single-marker analysis was followed by interval mapping according to LANDER and BOTSTEIN 1989 using the Kosambi mapping function. In addition, we used a composite interval mapping approach that includes the most significant markers to control for genetic background effects. Our composite interval analysis used Zmap Model 6 in QTL Cartographer with a window size of 10 cM and 10 background markers (determined by stepwise regression in Srmap). Significance levels were set to 5% and estimated by performing 1000 permutations of each analysis (DOERGE and CHURCHILL 1996 ). Epistasis interactions between the five QTL identified were examined by a two-way ANOVA that included the single markers, as well as all two-way interactions possible. A second analysis to determine the amount of variation explained by each factor included only the significant two-way interactions between QTL. Heterozygotes in the RI population were treated as missing data in this analysis, which is presented in Table 2.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Using Southern blot analysis with methylation-sensitive restriction endonucleases, we examined 10 accessions of A. thaliana, collected from wild populations, for differences in cytosine methylation at both repetitive and low-copy genomic loci (Table 1). In all natural accessions examined, highly repetitive centromeric sequences, including the 180-bp HindIII repeats (MARTINEZ-ZAPATER et al. 1986 ), the pericentromeric Athila retrotransposable elements (PELISSIER et al. 1996 ), and 5S rRNA genes (CAMPELL et al. 1992 ), were heavily methylated at HpaII sites and showed no variation in DNA methylation (Table 1 and data not shown). We also found no variation in the methylation status of the single-copy FWA gene, which was previously shown to be a hotspot for epimutation (SOPPE et al. 2000 ). However, we found significant variation in the methylation of two genomic regions in the 10 natural accessions: the single-copy MHC9.7/9.8 locus on chromosome 3 (PRUITT and MEYEROWITZ 1986 ) and the major rRNA gene repeats at the end of chromosomes 2 and 4 (NOR2 and NOR4; COPENHAVER and PIKAARD 1996A ). At the MHC9.7/9.8 region, we found a range of modification from complete methylation of all HpaII sites examined in strain Ws to complete lack of methylation in strain Can (Table 1). There is no clear correlation between methylation at the MHC9.7/9.8 locus and modification of the rRNA gene clusters (NOR; Table 1; Fig 1A and Fig B). NOR methylation ranged from 20% in Can to >90% in Hi (defined by the percentage of total hybridization to HpaII fragments >1 kb; see Fig 1B). However, even in natural accessions with comparable levels of NOR methylation, 5-methylcytosine appears to be differentially distributed along the major rRNA gene repeat. Natural accessions such as Hi show restriction fragments of intermediate size that are absent in other strains, such as Ler, despite similar levels of overall methylation (see Fig 1C). These intermediate-sized HpaII fragments are most likely due to epigenetic variation in methylation patterns rather than to nucleotide sequence polymorphism because the rRNA gene repeat sequences are highly conserved among Arabidopsis strains. To examine the extent of variation of NOR methylation within strains, we quantitated NOR methylation levels from 100 individuals for each natural accession. We found little intrastrain variation in NOR methylation, indicating that methylation of the major rRNA gene repeats is a stable characteristic of each natural accession (Table 1).

To study the basis of the observed variation in DNA methylation among A. thaliana natural accessions, we performed a QTL analysis. We measured NOR methylation in 162 recombinant inbred lines generated from a cross between a high NOR methylation parent (Ler) and a low NOR methylation parent (Cvi; ALONSO-BLANCO et al. 1998 ). Using single-marker analysis as well as interval and composite interval mapping, we identified five QTL that account for the majority of variation in NOR methylation between the parents (Fig 2). The major QTL, which together explain 50% of the variation in NOR methylation (Table 2), map to the tip of chromosomes 2 and 4, where the NOR2 and NOR4 loci reside. The analysis also identified three trans-acting QTL, one each on chromosomes 1, 3, and 5, that explain 20% of the NOR methylation variation between the Cvi and Ler parents (Table 2). We also calculated the effect of pairwise interactions between the five QTL and found three significant (P < 0.1) interactions involving four of the QTL (Table 2). The strongest interaction effect was between QTL3a and QTL4a (Fig 3).

The major QTL assignments to NOR2 and NOR4 in the Cvi/Ler cross indicate that the main determinants of NOR methylation are specified in cis. Cis-acting NOR methylation control could be explained by several mechanisms. One possibility is that the major rRNA gene repeats in Cvi and Ler have diverged at the primary DNA sequence level, which may lead to a difference in cytosine methylation. However, rRNA gene repeats are highly conserved, and sequence polymorphisms are rare in the transcribed region of the rRNA genes among A. thaliana natural accessions (C. PIKAARD, personal communication; our unpublished results). Although the number of short tandem "SalI" repeats in the intergenic region can vary in different A. thaliana strains, leading to the formation of different intergenic spacer length variants (COPENHAVER and PIKAARD 1996B ), there is no clear correlation between the variation in intergenic spacer length and NOR methylation (data not shown). These considerations argue against the hypothesis that sequence variation in the major rRNA gene repeats accounts for differences in cytosine methylation.

Another possibility is that differences in rRNA gene copy number affect NOR methylation levels. In most eukaryotes, rRNA genes are found in multiple copies and only a subset of these genes is expressed in most cells (CONCONI et al. 1989 ; DAMMANN et al. 1993 ; SANTORO and GRUMMT 2001 ). Strains with more rRNA genes may contain on average more transcriptionally quiescent gene copies, which can be subject to cytosine methylation associated with gene inactivity (FLAVELL et al. 1988 ). As shown in Fig 4, the relative copy number of 25S rRNA genes is positively correlated with NOR methylation in the 10 natural accessions surveyed. However, other factors must play a role as well. Strains with similar rRNA gene copy number, such as Cvi and Ler, can have markedly different levels of NOR methylation.

Epigenetic inheritance of parental methylation patterns is a third possible explanation for the strong QTL association with NOR2 and NOR4. To test this hypothesis, we measured NOR methylation in F₁ hybrids between Cvi and Ler. 25S rRNA gene methylation in F₁ hybrids created by reciprocal crosses was intermediate between the two parental values (Table 1). This result could be due to strict inheritance of parental methylation patterns or a homogenization of methylation at intermediate values on all NOR in the hybrids. The latter hypothesis would be predicted if NOR methylation was set by an rRNA gene-counting mechanism acting on the whole nucleus level. We distinguished between these two mechanisms by examining the methylation of the two parental NOR types in the polymorphic intergenic region. Fig 5 shows that a polymorphic EcoRI fragment specific for the low-methylation Cvi parent remains hypomethylated in the hybrids, as evidenced by complete digestion with HpaII. These data indicate that parental NOR methylation patterns are maintained in Cvi/Ler hybrids.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Our survey of A. thaliana strains has uncovered significant natural variation in cytosine methylation restricted to certain genomic sequences. There is little variation among A. thaliana strains in terms of total cytosine methylation and most genomic regions examined show equivalent levels of methylation. In contrast, the different A. thaliana strains are highly variable in their cytosine methylation of the single-copy region encompassing predicted genes MHC9.7 and -9.8. We found the most striking variation in cytosine methylation in the long arrays of rRNA genes found at the tips of chromosomes 2 and 4 (NOR2 and NOR4). At the moment, we were unable to make any inference regarding the phenotypic effect of different NOR methylation levels due to the numerous genetic differences among the different A. thaliana strains.

After completing our survey of natural variation in cytosine methylation, we turned our attention to discerning the sources of this variation. Our investigation indicates that both genetic and epigenetic mechanisms lead to natural variation in cytosine methylation in Arabidopsis. One type of genetic variation that appears to specify NOR methylation content is rRNA gene copy number. The correlation between rRNA repeat number and cytosine methylation may relate to the general observation that highly reiterated sequences in eukaryotic genomes are often highly methylated. The mechanistic basis of the association between rRNA repeat number and cytosine methylation is not understood, but is consistent with the hypothesis that excess rRNA genes may be archived and transcriptionally silenced in association with increased cytosine methylation (FLAVELL et al. 1988 ). A simple gene-counting mechanism, however, cannot explain all of the natural variation in NOR methylation, because strains with similar rRNA gene copy numbers can possess widely divergent NOR methylation levels (e.g., Cvi and Ler strains).

Some of the natural variation in cytosine methylation is likely to result from the genetic variation in genes encoding factors that direct and/or regulate DNA methylation. Our analysis suggests that genetic variation in such trans-acting modifier loci exists between the Cvi and Ler strains. Three trans-acting QTL affecting NOR methylation were identified in this study, mapping to chromosomes 1, 3, and 5. The identity of the genetic loci corresponding to these QTL is not known. We note that previously identified cytosine methylation modifiers, such as DDM1 (JEDDELOH et al. 1998 ) and the cytosine methyltransferase gene MET1 (FINNEGAN et al. 1996 ; RONEMUS et al. 1996 ), fall outside of the windows encompassing the three trans-acting QTL. Both the Cvi and Ler parents must carry functional alleles for these established cytosine methylation modifiers because no relevant variation at these loci was segregating in the RI mapping populations examined. Of the 10 remaining predicted cytosine methyltransferase genes in Arabidopsis (http://www.chromdb.org), 3 of the predicted de novo cytosine methyltransferase genes in the DOMAINS REARRANGED METHYLTRANSFERASE class (CAO et al. 2000 ) fall within windows defined by QTL on chromosomes 3 and 5. DRM3 maps close to the upper boundary of the QTL3a window, and DRM1 and DRM2 map in the vicinity of QTL5a. Additionally, the chromosome 5 QTL window corresponds to the genetic neighborhood containing the KYP histone H3 lysine 9 methyltransferase gene (At5g13960), which was recently shown to be important for specification of cytosine methylation at CpNpG sites (JACKSON et al. 2002 ). Given the low resolution of QTL localization, these connections may be coincidental, but identification of such candidate genes may facilitate the large amount of work necessary to define the gene(s) corresponding to the trans-acting QTL.

The major QTL in our study map directly on the target loci, NOR2 and NOR4. While several genetic mechanisms are consistent with this result, the finding that parental NOR methylation patterns are retained in Cvi/Ler F₁ hybrids (Table 1; Fig 5) points to epigenetic inheritance of diverged parental NOR methylation patterns as the simplest explanation for the major QTL. Transmission of differential NOR methylation patterns through meiosis contradicts the expectation in mammals, where cytosine methylation patterns are reset early in development (MONK et al. 1987 ; REIK et al. 2001 ). Our results, however, are corroborated by the inheritance of DNA methylation-based epialleles in plants (JACOBSEN and MEYEROWITZ 1997 ; CUBAS et al. 1999 ; SOPPE et al. 2000 ; KNOX and ELLIS 2001 ). Further support for inheritance of parental cytosine methylation comes from the observation that ddm1-induced hypomethylation of genomic sequences is inherited in a Mendelian manner (VONGS et al. 1993 ; KAKUTANI et al. 1999 ).

It is not clear how parental cytosine methylation patterns are specified, but two general mechanisms can be envisioned. First, cytosine methylation patterns may serve as a template for maintaining identical or similar methylation patterns after DNA replication, as first proposed by Holliday and Riggs in 1975 (HOLLIDAY and PUGH 1975 ; RIGGS 1975 ). This model requires that genomic sequences are not hypomethylated during early plant development; however, a report of a dramatic loss of cytosine methylation in tobacco pollen generative nuclei suggests that a DNA methylation erasure phase may exist in plants (OAKELEY et al. 1997 ). An alternative model for inheritance of parental cytosine methylation patterns calls for specification by epigenetic marks independent of 5-methyl-cytosine. Such a model is consistent with the recent demonstration that methylation of histone H3 on lysine 9 is required for cytosine methylation in Neurospora (TAMARU and SELKER 2001 ) and important for ^mCpNpG methylation in Arabidopsis (JACKSON et al. 2002 ). Further clues to the mechanisms responsible may develop from study of situations in which the inheritance of parental methylation breaks down. The genetic interaction between QTL3a and NOR4 (QTL4a) described here may represent such a situation. The genotype at QTL3a strongly influences the strength of QTL4a (Fig 3), suggesting that the inheritance of parental NOR4 methylation patterns can be modified in trans.

While the inheritance of diverged DNA methylation patterns can contribute to the maintenance of natural epigenetic variation, a mechanism is still required to generate a divergent epigenetic state initially. Genetic variation at trans-acting modifier loci may play an important role. Indeed, the significance of trans-acting control of NOR methylation may have been underestimated in the present study because inheritance of widely divergent parental cytosine methylation patterns may overshadow the influence of factors in trans. Epigenetic mechanisms may also contribute directly to the generation of divergent cytosine methylation states. Environmental conditions have been documented to alter cytosine methylation patterns in plants (BURN et al. 1993 ; SCHMITT et al. 1997 ), and it is reasonable to propose that environmental stresses could lead to significantly divergent DNA methylation NOR haplotypes. Alternatively, small changes in cytosine methylation may accumulate over time, in a manner analogous to genetic drift of neutral genetic variation. We are currently testing both of these possibilities in an attempt to evaluate the role of epigenetic mechanisms in generating divergent cytosine methylation patterns.

	ACKNOWLEDGMENTS

We thank our colleagues J. Cheverud and P. Kover for technical advice throughout the project. We are grateful to J. Cheverud, C. S. Pikaard, and T. L. Stokes for critical reading of the manuscript. This work was supported by a grant from the Ministry of Agriculture, Forestry and Fisheries (MAFF), Japan, and a grant from the National Science Foundation to E.J.R. (MCB 9985348).

Manuscript received March 27, 2002; Accepted for publication May 31, 2002.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ALONSO-BLANCO, C. and M. KOORNNEEF, 2000  Naturally occurring variation in Arabidopsis: an underexploited resource for plant genetics. Trends Plant Sci. 5:22-29.[Medline]

ALONSO-BLANCO, C., A. J. PEETERS, M. KOORNNEEF, C. LISTER, and C. DEAN et al., 1998  Development of an AFLP based linkage map of Ler, Col and Cvi Arabidopsis thaliana ecotypes and construction of a Ler/Cvi recombinant inbred line population. Plant J. 14:259-271.[Medline]

BASTEN, C. J., B. S. WEIR and Z-B. ZENG, 1994 Zmap: a QTL cartographer. Fifth World Congress on Genetics Applied to Livestock Production: Computing Strategies and Software, edited by C. SMITH, J. S. GAVORA, B. BENKEL, J. CHESNAIS, W. FAIRFULL et al., pp. 65–66. Organizing Committee, Fifth World Congress on Genetics Applied to Livestock Production, Guelph, ON, Canada.

BASTEN, C. J., B. S. WEIR and Z-B. ZENG, 2001 QTL Cartographer. Department of Statistics, North Carolina State University, Raleigh, NC.

BESTOR, T. H., 2000  The DNA methyltransferases of mammals. Hum. Mol. Genet. 9:2395-2402.[Abstract/Free Full Text]

BURN, J. E., D. J. BAGNALL, J. D. METZGER, E. S. DENNIS, and W. J. PEACOCK, 1993  DNA methylation, vernalization, and the initiation of flowering. Proc. Natl. Acad. Sci. USA 90:287-291.[Abstract/Free Full Text]

CAMPELL, B. R., Y. SONG, T. E. POSCH, C. A. CULLIS, and C. D. TOWN, 1992  Sequence and organization of 5S ribosomal RNA-encoding genes of Arabidopsis thaliana. Gene 112:225-228.[Medline]

CAO, X., N. M. SPRINGER, M. G. MUSZYNSKI, R. L. PHILLIPS, and S. KAEPPLER et al., 2000  Conserved plant genes with similarity to mammalian de novo DNA methyltransferases. Proc. Natl. Acad. Sci. USA 97:4979-4984.[Abstract/Free Full Text]

CHEN, R. Z., U. PETTERSSON, C. BEARD, L. JACKSON-GRUSBY, and R. JAENISCH, 1998  DNA hypomethylation leads to elevated mutation rates. Nature 395:89-93.[Medline]

CONCONI, A., R. M. WIDMER, T. KOLLER, and J. M. SOGO, 1989  Two different chromatin structures coexist in ribosomal RNA genes throughout the cell cycle. Cell 57:753-761.[Medline]

COPENHAVER, G. P. and C. S. PIKAARD, 1996a  RFLP and physical mapping with an rDNA-specific endonuclease reveals that nucleolus organizer regions of Arabidopsis thaliana adjoin the telomeres on chromosomes 2 and 4. Plant J. 9:259-272.[Medline]

COPENHAVER, G. P. and C. S. PIKAARD, 1996b  Two-dimensional RFLP analyses reveal megabase-sized clusters of rRNA gene variants in Arabidopsis thaliana, suggesting local spreading of variants as the mode for gene homogenization during concerted evolution. Plant J. 9:273-282.[Medline]

CUBAS, P., C. VINCENT, and E. COEN, 1999  An epigenetic mutation responsible for natural variation in floral symmetry. Nature 401:157-161.[Medline]

DAMMANN, R., R. LUCCHINI, T. KOLLER, and J. M. SOGO, 1993  Chromatin structures and transcription of rDNA in yeast Saccharomyces cerevisiae. Nucleic Acids Res. 21:2331-2338.[Abstract/Free Full Text]

DENNIS, K., T. FAN, T. GEIMAN, Q. YAN, and K. MUEGGE, 2001  Lsh, a member of the SNF2 family, is required for genome-wide methylation. Genes Dev. 15:2940-2944.[Abstract/Free Full Text]

DOERGE, R. W. and G. A. CHURCHILL, 1996  Permutation tests for multiple loci affecting a quantitative character. Genetics 142:285-294.[Abstract]

FINNEGAN, E. J., W. J. PEACOCK, and E. S. DENNIS, 1996  Reduced DNA methylation in Arabidopsis thaliana results in abnormal plant development. Proc. Natl. Acad. Sci. USA 93:8449-8454.[Abstract/Free Full Text]

FLAVELL, R. B., M. O'DELL, and W. F. THOMPSON, 1988  Regulation of cytosine methylation in ribosomal DNA and nucleolus organizer expression in wheat. J. Mol. Biol. 204:523-534.[Medline]

GIBBONS, R. J., T. L. MCDOWELL, S. RAMAN, D. M. O'ROURKE, and D. GARRICK et al., 2000  Mutations in ATRX, encoding a SWI/SNF-like protein, cause diverse changes in the pattern of DNA methylation. Nat. Genet. 24:368-371.[Medline]

HOLLIDAY, R. and J. E. PUGH, 1975  DNA modification mechanisms and gene activity during development. Science 187:226-232.[Free Full Text]

JACKSON, J. P., A. M. LINDROTH, X. CAO, and S. E. JACOBSEN, 2002  Control of CpNpG DNA methylation by the KRYPTONITE histone H3 methyltransferase. Nature 416:556-560.[Medline]

JACOBSEN, S. E. and E. M. MEYEROWITZ, 1997  Hypermethylated SUPERMAN epigenetic alleles in Arabidopsis. Science 277:1100-1103.[Abstract/Free Full Text]

JACOBSEN, S. E., H. SAKAI, E. J. FINNEGAN, X. CAO, and E. M. MEYEROWITZ, 2000  Ectopic hypermethylation of flower-specific genes in Arabidopsis. Curr. Biol. 10:179-186.[Medline]

JEDDELOH, J. A., J. BENDER, and E. J. RICHARDS, 1998  The DNA methylation locus DDM1 is required for maintenance of gene silencing in Arabidopsis. Genes Dev. 12:1714-1725.[Abstract/Free Full Text]

KAKUTANI, T., J. A. JEDDELOH, S. K. FLOWERS, K. MUNAKATA, and E. J. RICHARDS, 1996  Developmental abnormalities and epimutations associated with DNA hypomethylation mutations. Proc. Natl. Acad. Sci. USA 93:12406-12411.[Abstract/Free Full Text]

KAKUTANI, T., K. MUNAKATA, E. J. RICHARDS, and H. HIROCHIKA, 1999  Meiotically and mitotically stable inheritance of DNA hypomethylation induced by ddm1 mutation of Arabidopsis thaliana. Genetics 151:831-838.[Abstract/Free Full Text]

KNOX, M. R. and T. H. ELLIS, 2001  Stability and inheritance of methylation states at PstI sites in Pisum. Mol. Genet. Genomics 265:497-507.[Medline]

LANDER, E. S. and D. BOTSTEIN, 1989  Mapping Mendelian factors underlying quantitative traits using RFLP linkage maps. Genetics 121:185-199.[Abstract/Free Full Text]

LI, E., T. H. BESTOR, and R. JAENISCH, 1992  Targeted mutation of the DNA methyltransferase gene results in embryonic lethality. Cell 69:915-926.[Medline]

MALAGNAC, F., B. WENDEL, C. GOYON, G. FAUGERON, and D. ZICKLER et al., 1997  A gene essential for de novo methylation and development in Ascobolus reveals a novel type of eukaryotic DNA methyltransferase structure. Cell 91:281-290.[Medline]

MARTIENSSEN, R. A. and V. COLOT, 2001  DNA methylation and epigenetic inheritance in plants and filamentous fungi. Science 293:1070-1074.[Abstract/Free Full Text]

MARTINEZ-ZAPATER, J. M., M. A. ESTELLE, and C. R. SOMERVILLE, 1986  A highly repeated DNA sequence in Arabidopsis thaliana.. Mol. Gen. Genet. 204:417-423.

MIAO, V. P., M. FREITAG, and E. U. SELKER, 2000  Short TpA-rich segments of the zeta-eta region induce DNA methylation in Neurospora crassa. J. Mol. Biol. 300:249-273.[Medline]

MIURA, A., S. YONEBAYASHI, K. WATANABE, T. TOYAMA, and H. SHIMADA et al., 2001  Mobilization of transposons by a mutation abolishing full DNA methylation in Arabidopsis. Nature 411:212-214.[Medline]

MONK, M., M. BOUBELIK, and S. LEHNERT, 1987  Temporal and regional changes in DNA methylation in the embryonic, extraembryonic and germ cell lineages during mouse embryo development. Development 99:371-382.[Abstract]

OAKELEY, E. J., A. PODESTA, and J. P. JOST, 1997  Developmental changes in DNA methylation of the two tobacco pollen nuclei during maturation. Proc. Natl. Acad. Sci. USA 94:11721-11725.[Abstract/Free Full Text]

OKANO, M., D. W. BELL, D. A. HABER, and E. LI, 1999  DNA methyltransferases Dnmt3a and Dnmt3b are essential for de novo methylation and mammalian development. Cell 99:247-257.[Medline]

PELISSIER, T., S. TUTOIS, S. TOURMENTE, J. M. DERAGON, and G. PICARD, 1996  DNA regions flanking the major Arabidopsis thaliana satellite are principally enriched in Athila retroelement sequences. Genetica 97:141-151.[Medline]

PRUITT, R. E. and E. M. MEYEROWITZ, 1986  Characterization of the genome of Arabidopsis thaliana. J. Mol. Biol. 187:169-183.[Medline]

REIK, W., W. DEAN, and J. WALTER, 2001  Epigenetic reprogramming in mammalian development. Science 293:1089-1093.[Abstract/Free Full Text]

RICHARDS, E. J. and S. C. R. ELGIN, 2002  Epigenetic codes for heterochromatin formation and silencing: rounding up the usual suspects. Cell 108:489-500.[Medline]

RIGGS, A. D., 1975  X inactivation, differentiation, and DNA methylation. Cytogenet. Cell Genet. 14:9-25.[Medline]

RONEMUS, M. J., M. GALBIATI, C. TICKNOR, J. CHEN, and S. L. DELLAPORTA, 1996  Demethylation-induced developmental pleiotropy in Arabidopsis. Science 273:654-657.[Abstract]

SANTORO, R. and I. GRUMMT, 2001  Molecular mechanisms mediating methylation-dependent silencing of ribosomal gene transcription. Mol. Cell 8:719-725.[Medline]

SCHMITT, F., E. J. OAKELEY, and J. P. JOST, 1997  Antibiotics induce genome-wide hypermethylation in cultured Nicotiana tabacum plants. J. Biol. Chem. 272:1534-1540.[Abstract/Free Full Text]

SOPPE, W. J., S. E. JACOBSEN, C. ALONSO-BLANCO, J. P. JACKSON, and T. KAKUTANI et al., 2000  The late flowering phenotype of fwa mutants is caused by gain-of-function epigenetic alleles of a homeodomain gene. Mol. Cell 6:791-802.[Medline]

TAMARU, H. and E. U. SELKER, 2001  A histone H3 methyltransferase controls DNA methylation in Neurospora crassa. Nature 414:277-283.[Medline]

VONGS, A., T. KAKUTANI, R. A. MARTIENSSEN, and E. J. RICHARDS, 1993  Arabidopsis thaliana DNA methylation mutants. Science 260:1926-1928.[Abstract/Free Full Text]

YATES, P. A., R. W. BURMAN, P. MUMMANENI, S. KRUSSEL, and M. S. TURKER, 1999  Tandem B1 elements located in a mouse methylation center provide a target for de novo DNA methylation. J. Biol. Chem. 274:36357-36361.[Abstract/Free Full Text]

This article has been cited by other articles:

				R. Matyasek, J. A. Tate, Y. K. Lim, H. Srubarova, J. Koh, A. R. Leitch, D. E. Soltis, P. S. Soltis, and A. Kovarik Concerted Evolution of rDNA in Recently Formed Tragopogon Allotetraploids Is Typically Associated With an Inverse Correlation Between Gene Copy Number and Expression Genetics, August 1, 2007; 176(4): 2509 - 2519. [Abstract] [Full Text] [PDF]

				C. Shindo, G. Bernasconi, and C. S. Hardtke Natural Genetic Variation in Arabidopsis: Tools, Traits and Prospects for Evolutionary Ecology Ann. Bot., June 1, 2007; 99(6): 1043 - 1054. [Abstract] [Full Text] [PDF]

				A. L. Keyte, R. Percifield, B. Liu, and J. F. Wendel Infraspecific DNA Methylation Polymorphism in Cotton (Gossypium hirsutum L.) J. Hered., September 1, 2006; 97(5): 444 - 450. [Abstract] [Full Text] [PDF]

				R. T. GRANT-DOWNTON and H. G. DICKINSON Epigenetics and its Implications for Plant Biology 2. The 'Epigenetic Epiphany': Epigenetics, Evolution and Beyond Ann. Bot., January 1, 2006; 97(1): 11 - 27. [Abstract] [Full Text] [PDF]

				H.-C. Jing, J. H. M. Schippers, J. Hille, and P. P. Dijkwel Ethylene-induced leaf senescence depends on age-related changes and OLD genes in Arabidopsis J. Exp. Bot., November 1, 2005; 56(421): 2915 - 2923. [Abstract] [Full Text] [PDF]

				M. S. Lewis, J. M. Cheverud, and C. S. Pikaard Evidence for Nucleolus Organizer Regions as the Units of Regulation in Nucleolar Dominance in Arabidopsis thaliana Interecotype Hybrids Genetics, June 1, 2004; 167(2): 931 - 939. [Abstract] [Full Text] [PDF]