CpG methylation is targeted to transcription units in an invertebrate genome

Abstract

DNA is methylated at the dinucleotide CpG in genomes of a wide range of plants and animals. Among animals, variable patterns of genomic CpG methylation have been described, ranging from undetectable levels (e.g., in Caenorhabditis elegans) to high levels of global methylation in the vertebrates. The most frequent pattern in invertebrate animals, however, is mosaic methylation, comprising domains of methylated DNA interspersed with unmethylated domains. To understand the origin of mosaic DNA methylation patterns, we examined the distribution of DNA methylation in the Ciona intestinalis genome. Bisulfite sequencing and computational analysis revealed methylated domains with sharp boundaries that strongly colocalize with approximately 60% of transcription units. By contrast, promoters, intergenic DNA, and transposons are not preferentially targeted by DNA methylation. Methylated transcription units include evolutionarily conserved genes, whereas the most highly expressed genes preferentially belong to the unmethylated fraction. The results lend support to the hypothesis that CpG methylation functions to suppress spurious transcriptional initiation within infrequently transcribed genes.

Figure 1.

A mosaic pattern of methlyated domains (MDs) and unmethylated domains (UMDs) in the C. intestinalis genome. Bisulfite sequencing showed the DNA methylation status at every CpG site in ribosomal protein S6 (scaffold 13: 227234–231211) (A), a randomly selected region of chromosome 9q (2421536–2429959) (B), a region of chromosome 7q corresponding to 29,300 bp of cosmid insert cos41 (C), and a region of chromosome 4q corresponding to 38,500 bp of cosmid insert cos2 (D). The top panels show gene annotation and plots of CpG methylation are shown below. Each dot represents a specific CpG corresponding to the CpG map at the bottom of each figure. MD/UMD boundaries are shown by broken lines. Repetitive sequences Cifo-1 (blue), Cimi-1 (green), and Cics-1 (orange) are shown.

Figure 2.

Methylated domains (MDs) within two 1-Mb genomic regions from chromosome 7q (A) and chromosome 4q (B) colocalize with transcription units. MDs (green bars) were inferred from CpG[o/e] ratios as described in the text. The inferred pattern was experimentally verified by methylation-sensitive PCR at sites marked by closed triangles (methylated) and open triangles (unmethylated). Yellow arrows show the size and orientation of predicted transcription units. The positions of bisulfite sequenced cosmid inserts cos41 and cos2 (see Fig. 1) are shown by black boxes. Green boxes indicate regions that were bisulfite sequenced in order to test the accuracy of predicted unmethylated domain (UMD)/MD boundaries (Supplemental Fig. 2). (C) Frequency of predicted genes that overlap to varying degrees with MDs. (D) Frequency of MDs that overlap to varying degrees with predicted genes. (E) Frequency of UMDs with gene overlap.

Figure 3.

Features that distinguish C. intestinalis methylated and unmethylated genes. (A) Histogram showing the frequency of all genes with CpG[o/e] frequencies between 0 and 2. CpG[o/e] of genes whose methylation status was analyzed by bisulfite sequencing are shown in closed (methylated) and open (unmethylated) triangles below. (B) Frequencies of genes that are evolutionarily conserved between C. intestinalis, Homo sapiens, and D. melanogaster plotted against CpG[o/e]. (C) Frequencies of C. intestinalis orthologs of human CpG island genes plotted against CpG[o/e]. (D) Frequencies of C. intestinalis orthologs of human non-CpG island genes plotted against CpG[o/e]. (E) Frequencies of 188 genes that recorded the highest number of EST hits (>500) in the UniGene database plotted against CpG[o/e].

Figure 4.

Variable DNA methylation of repetitive elements in C. intestinalis matches flanking DNA. Elements as named here or by Simmen and Bird (2000) are diagrammed as solid bars above each bisulfite profile showing primer pair location. A horizontal row within each box corresponds to one sequenced clone in which specific CpGs were methylated (solid) or unmethylated (open). Mutated CpG sites are shown as blanks within a row. (A) Results using primer sets within the element showed a mixture of methylated and unmethylated clones. (B) Locus-specific primers showed that the methylation patterns of two Cics-1 elements and two Cimi-1 elements at specific genomic locations are homogeneous and match the methylation status of flanking genomic DNA (open bars above). The locations of the elements analyzed in B are shown in Fig. 1.