Limits...
Dynamic DNA cytosine methylation in the Populus trichocarpa genome: tissue-level variation and relationship to gene expression.

Vining KJ, Pomraning KR, Wilhelm LJ, Priest HD, Pellegrini M, Mockler TC, Freitag M, Strauss SH - BMC Genomics (2012)

Bottom Line: We validated MeDIP-seq results by bisulfite sequencing, and compared methylation and gene expression using published microarray data.We found striking differences among tissues in methylation, which were apparent at the chromosomal scale and when genes and transposable elements were examined.In contrast to other studies in plants, gene body methylation had a more repressive effect on transcription than promoter methylation.

View Article: PubMed Central - HTML - PubMed

Affiliation: Department of Forest Ecosystems and Society, Oregon State University, Corvallis, 97331, USA.

ABSTRACT

Background: DNA cytosine methylation is an epigenetic modification that has been implicated in many biological processes. However, large-scale epigenomic studies have been applied to very few plant species, and variability in methylation among specialized tissues and its relationship to gene expression is poorly understood.

Results: We surveyed DNA methylation from seven distinct tissue types (vegetative bud, male inflorescence [catkin], female catkin, leaf, root, xylem, phloem) in the reference tree species black cottonwood (Populus trichocarpa). Using 5-methyl-cytosine DNA immunoprecipitation followed by Illumina sequencing (MeDIP-seq), we mapped a total of 129,360,151 36- or 32-mer reads to the P. trichocarpa reference genome. We validated MeDIP-seq results by bisulfite sequencing, and compared methylation and gene expression using published microarray data. Qualitative DNA methylation differences among tissues were obvious on a chromosome scale. Methylated genes had lower expression than unmethylated genes, but genes with methylation in transcribed regions ("gene body methylation") had even lower expression than genes with promoter methylation. Promoter methylation was more frequent than gene body methylation in all tissues except male catkins. Male catkins differed in demethylation of particular transposable element categories, in level of gene body methylation, and in expression range of genes with methylated transcribed regions. Tissue-specific gene expression patterns were correlated with both gene body and promoter methylation.

Conclusions: We found striking differences among tissues in methylation, which were apparent at the chromosomal scale and when genes and transposable elements were examined. In contrast to other studies in plants, gene body methylation had a more repressive effect on transcription than promoter methylation.

Show MeSH
Association of tissue predominant expression with methylation. RPKM data from lists of genes with strong bias in gene expression compared to those in the entire genome for those tissues. Asterisks indicate significantly different means (P < 0.05) based on comparisons of "all genes" to "tissue bias" genes within a tissue type group, calculated for promoters and gene bodies separately.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC3298464&req=5

Figure 8: Association of tissue predominant expression with methylation. RPKM data from lists of genes with strong bias in gene expression compared to those in the entire genome for those tissues. Asterisks indicate significantly different means (P < 0.05) based on comparisons of "all genes" to "tissue bias" genes within a tissue type group, calculated for promoters and gene bodies separately.

Mentions: for which tissues were divided into subsets s and o, with ns assigned to the number of tissues in subset s, and no was assigned to the number of tissues not in subset s (all other tissues); ws denotes the weight applied to tissues in subset s, i.e. max(ns, no)/ns, wo denotes the weight applied to tissues in subset o, i.e. max(ns, no)/no, Es denotes the sum of expression values of tissues in subset s, and Eo denotes the sum of expression values of tissues in subset. A gene was considered biased at or above a calculated bias of 0.9. The number of tissue biased genes varied from 320 for leaves to 6,729 for male and female catkins (pooled for the calculation of expression bias). From this set, we found that 2.5 to 5.8% and 0.5 to 12.5% of genes called as biased based on expression were called methylated by our criteria at promoters or gene bodies, respectively. When all of the genes showing bias for a tissue type were compared to all genes in our dataset for their RPKM levels, the differences were small; however, 9 of 12 comparisons were statistically significant and consistent in direction (Figure 8). In all cases where a difference was significant, the tissue predominant genes had lower methylation, both for promoters and/or gene bodies. In every case, whether significant or not, the genes with upwardly biased expression in that tissue never showed a higher RPKM value for promoters or for gene bodies. Excluding male catkins as outliers due to their unusual GO patterns (discussed above), for gene bodies all six tissues were consistent in having lower RPKM for the expression-biased tissue set (P < 0.05).


Dynamic DNA cytosine methylation in the Populus trichocarpa genome: tissue-level variation and relationship to gene expression.

Vining KJ, Pomraning KR, Wilhelm LJ, Priest HD, Pellegrini M, Mockler TC, Freitag M, Strauss SH - BMC Genomics (2012)

Association of tissue predominant expression with methylation. RPKM data from lists of genes with strong bias in gene expression compared to those in the entire genome for those tissues. Asterisks indicate significantly different means (P < 0.05) based on comparisons of "all genes" to "tissue bias" genes within a tissue type group, calculated for promoters and gene bodies separately.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC3298464&req=5

Figure 8: Association of tissue predominant expression with methylation. RPKM data from lists of genes with strong bias in gene expression compared to those in the entire genome for those tissues. Asterisks indicate significantly different means (P < 0.05) based on comparisons of "all genes" to "tissue bias" genes within a tissue type group, calculated for promoters and gene bodies separately.
Mentions: for which tissues were divided into subsets s and o, with ns assigned to the number of tissues in subset s, and no was assigned to the number of tissues not in subset s (all other tissues); ws denotes the weight applied to tissues in subset s, i.e. max(ns, no)/ns, wo denotes the weight applied to tissues in subset o, i.e. max(ns, no)/no, Es denotes the sum of expression values of tissues in subset s, and Eo denotes the sum of expression values of tissues in subset. A gene was considered biased at or above a calculated bias of 0.9. The number of tissue biased genes varied from 320 for leaves to 6,729 for male and female catkins (pooled for the calculation of expression bias). From this set, we found that 2.5 to 5.8% and 0.5 to 12.5% of genes called as biased based on expression were called methylated by our criteria at promoters or gene bodies, respectively. When all of the genes showing bias for a tissue type were compared to all genes in our dataset for their RPKM levels, the differences were small; however, 9 of 12 comparisons were statistically significant and consistent in direction (Figure 8). In all cases where a difference was significant, the tissue predominant genes had lower methylation, both for promoters and/or gene bodies. In every case, whether significant or not, the genes with upwardly biased expression in that tissue never showed a higher RPKM value for promoters or for gene bodies. Excluding male catkins as outliers due to their unusual GO patterns (discussed above), for gene bodies all six tissues were consistent in having lower RPKM for the expression-biased tissue set (P < 0.05).

Bottom Line: We validated MeDIP-seq results by bisulfite sequencing, and compared methylation and gene expression using published microarray data.We found striking differences among tissues in methylation, which were apparent at the chromosomal scale and when genes and transposable elements were examined.In contrast to other studies in plants, gene body methylation had a more repressive effect on transcription than promoter methylation.

View Article: PubMed Central - HTML - PubMed

Affiliation: Department of Forest Ecosystems and Society, Oregon State University, Corvallis, 97331, USA.

ABSTRACT

Background: DNA cytosine methylation is an epigenetic modification that has been implicated in many biological processes. However, large-scale epigenomic studies have been applied to very few plant species, and variability in methylation among specialized tissues and its relationship to gene expression is poorly understood.

Results: We surveyed DNA methylation from seven distinct tissue types (vegetative bud, male inflorescence [catkin], female catkin, leaf, root, xylem, phloem) in the reference tree species black cottonwood (Populus trichocarpa). Using 5-methyl-cytosine DNA immunoprecipitation followed by Illumina sequencing (MeDIP-seq), we mapped a total of 129,360,151 36- or 32-mer reads to the P. trichocarpa reference genome. We validated MeDIP-seq results by bisulfite sequencing, and compared methylation and gene expression using published microarray data. Qualitative DNA methylation differences among tissues were obvious on a chromosome scale. Methylated genes had lower expression than unmethylated genes, but genes with methylation in transcribed regions ("gene body methylation") had even lower expression than genes with promoter methylation. Promoter methylation was more frequent than gene body methylation in all tissues except male catkins. Male catkins differed in demethylation of particular transposable element categories, in level of gene body methylation, and in expression range of genes with methylated transcribed regions. Tissue-specific gene expression patterns were correlated with both gene body and promoter methylation.

Conclusions: We found striking differences among tissues in methylation, which were apparent at the chromosomal scale and when genes and transposable elements were examined. In contrast to other studies in plants, gene body methylation had a more repressive effect on transcription than promoter methylation.

Show MeSH