Mycoplasma CG- and GATC-specific DNA methyltransferases selectively and efficiently methylate the host genome and alter the epigenetic landscape in human cells

Abstract

Aberrant DNA methylation is frequently observed in disease, including many cancer types, yet the underlying mechanisms remain unclear. Because germline and somatic mutations in the genes that are responsible for DNA methylation are infrequent in malignancies, additional mechanisms must be considered. Mycoplasmas spp., including Mycoplasma hyorhinis, efficiently colonize human cells and may serve as a vehicle for delivery of enzymatically active microbial proteins into the intracellular milieu. Here, we performed, for the first time, genome-wide and individual gene mapping of methylation marks generated by the M. hyorhinis CG- and GATC-specific DNA cytosine methyltransferases (MTases) in human cells. Our results demonstrated that, upon expression in human cells, MTases readily translocated to the cell nucleus. In the nucleus, MTases selectively and efficiently methylated the host genome at the DNA sequence sites free from pre-existing endogenous methylation, including those in a variety of cancer-associated genes. We also established that mycoplasma is widespread in colorectal cancers, suggesting that either the infection contributed to malignancy onset or, alternatively, that tumors provide a favorable environment for mycoplasma growth. In the human genome, ∼ 11% of GATC sites overlap with CGs (e.g., CGAT(m)CG); therefore, the methylated status of these sites can be perpetuated by human DNMT1. Based on these results, we now suggest that the GATC-specific methylation represents a novel type of infection-specific epigenetic mark that originates in human cells with a previous exposure to infection. Overall, our findings unveil an entirely new panorama of interactions between the human microbiome and epigenome with a potential impact in disease etiology.

Keywords: 5-methylcytosine; CPGI; 5mC; CpG island; MTase; DNA methylation; DNA methyltransferases; Mycoplasma hyorhinis DNA cytosine methyltransferase.; colorectal cancer; epigenetics; host-pathogen interactions; mycoplasma.

Figure 1.

M. hyorhinis MTases. (A) conserved motifs in M. hyorhinis MTases. The amino acid sequence alignment of M. hyorhinis Mhy1, Mhy2 and Mhy3, Mfe1 from M. fermentans, Uur1 from Ureaplasma urealyticus and Mpe1 from M. penetrans. Conserved motifs are indicated by numbers. A consensus sequence is shown at the bottom. (B) MTase loci in M. hyorhinis genomes. Positions of the Mhy1 (40.5–46.5 kb), Mhy2 (92–98 kb), and Mhy3 (472.5–480 kb) genes in M. hyorhinis chromosome. Genes are: dnaJ, chaperone protein; mnmA, tRNA-specific 2-thiouridylase; Mhy1, CG-specific MTase; MYM_0038, hypothetical protein; nox, NADH oxidase; MYM_0074, hypothetical protein; lepA, GTP-binding protein; Mhy2, CG-specific MTase; MYM_0077, hypothetical protein; MYM_0078, ATP-binding ABC transporter; MYM_0384, putative endonuclease or phosphatase; MYM_0385, glutamyl aminopeptidase; Mhy3, GATC-specific MTase; MYM_0387, hypothetical protein; hsdS-3 and hsdS-4, type I DNA MTases. Genes are shown as solid color bars. The direction of transcription is indicated by arrows. Putative MTase and endonuclease genes are red and blue, respectively. Other genes are green.

Figure 2.

Mycoplasma MTases translocate to the human cell nuclei and methylate the human genome. (A) substrate specificity of Mhy1 (^mCG), Mhy2 (^mCG and A^mCATGT), and Mhy3 (GAT^mC). Bisulfite DNA sequencing chromatograms of DNA substrates are shown. Stars indicate 5 mC. (B) confocal images of HT1080/Mhy1, HT1080/Mhy2, HT1080/Mhy3, HTR8/Mhy1, HTR8/Mhy3, and HTR8/mock cells. Upper left panel shows the condensed chromatin in a single HT1080/Mhy1 cell. Red, the V5 immunostaining. Blue, DAPI. (C) DNA Methylation-interference assay. HT1080/Mhy1, HT1080/Mhy2, HT1080/Mhy3 and HT1080/mock genomic DNA was digested using HpaII, MspI (a methylation-insensitive isoschizomer of HpaII), Sau3AI and MboI (a methylation-insensitive isoschizomer of Sau3AI). (D) methylation-specific PCR of a CPGI promoter region of the MMP2 gene in HT1080/Mhy1 and HT1080/mock cells. PCR products specific to methylated (M) and unmethylated (U) DNA are shown.

Figure 3.

Genome-wide methylation induced by mycoplasma MTases. (A) circular histogram representation of DNA methylation induced by Mhy1 (red), Mhy2 (green), and Mhy3 (blue) in HT1080/Mhy1, HT1080/Mhy2, and HT1080/Mhy3 cells, respectively. For simplicity, the MeDIP profile of human chromosome 1 is shown. The data for human chromosomes 2–22, Y and X are in Figure S3A-O. Histogram bar heights are calculated as logarithms of a number of the individual methylation marks within a 100 kb window (see Extended Experimental Procedures). Gray histograms correspond to methylation in HT1080/mock cells. Orange and light blue histograms indicate relative density of CG and GATC sites within a 100 kb window, respectively. (B-D) bars show the methylation level in HT1080/Mhy1 (red), HT1080/Mhy2 (green), HT1080/Mhy3 (blue), and HT1080/mock cells (gray). Methylation was individually calculated for all non-repetitive genomic sequences (B), genic regions (C), and CPGI regions (D). The percentages were calculated relative to HT1080/mock cells (gray bars = 100%).

Figure 4.

Methylation of AKT3, ID1, and MMP2 genes. (A) methylation of the AKT3 gene locus. Top, the AKT3 gene locus with exons indicated as vertical bars. Middle, MeDIP profiles. Solid color bars indicate the genomic position and MeDIP coverage of individual CG- and GATC-methylation marks in HT1080/mock (gray), HT1080/Mhy1 (red) and HT1080/Mhy3 (blue) in a 2.5 kb GATC cluster of the AKT3 gene (positions 243977.5–243980 kb on chromosome 1). Bottom, bisulfite sequencing chromatogram of a 218 bp sub-region with 4 GATC sites methylated in HT1080/Mhy3 cells. Arrows indicate 5 mC. (B) methylation of the ID1 gene. Top, bisulfite sequencing chromatogram of a 213 bp sub-region in the promoter of the ID1 gene locus in HT1080/Mhy1 and HT1080/Mhy3 cells. The converted and original genomic sequences are shown under chromatograms. Middle, MeDIP profiles. Solid blue bars indicate the genomic position and MeDIP coverage of the individual GATC-methylation marks in HT1080/Mhy3 cells in a 2.5 kb ID1 gene promoter (positions 30192–30194.5 kb on chromosome 20). No GATC methylation was recorded in HT1080/Mhy1 and HT1080/mock cells. Bottom, ID1 gene locus with exons (solid black bars) and CPGIs (purple bars). (C-D) methylation of the MMP2 gene. C, MeDIP profiles of a 27.5 kb MMP2 gene locus in HT1080/Mhy3, HT1080/Mhy1 and HT1080/mock cells. Solid color bars indicate the genomic position and MeDIP coverage of the individual CG- and GATC-methylation marks in HT1080/mock (gray), HT1080/Mhy1 (red) and HT1080/Mhy3 (blue) cells. CPGIs, purple rectangles. Genomic coordinates are shown at the panel bottom. (D) bisulfite sequencing diagrams of a 221 bp sub-region (indicated by a red arrow in panel C) in the CPGI promoter (purple bar) of the first exon (solid black bar) of the MMP2 gene in HT1080/Mhy1 and HT1080/mock cells. Stars and arrows indicate 5 mC.

Figure 5.

Methylation of cancer-specific genes. (A) heatmap of the CG- and GATC-methylation in 76 cancer-specific genes in HT1080/Mhy1, HT1080/Mhy3 and HT1080/mock cells. Red and blue correspond to the high and the low methylation levels according to MeDIP, respectively. (B) MeDIP profiles of the selected gene loci. Solid color bars indicate the genomic position and MeDIP coverage of the individual CG- and GATC-methylation marks in HT1080/mock (gray), HT1080/Mhy1 (red) and HT1080/Mhy3 (blue) cells in the APC, SMAD4, MYC, JUN, and TGFB1 gene loci. Solid bars indicate exons (black) and CPGIs (purple). Chromosomal coordinates are shown at the bottom of the diagrams.

Figure 6.

Transcriptome analysis in fibrosarcoma and trophoblasts. (A) top regulated genes in HT1080/Mhy1, HT1080/Mhy3 relative to HT1080/mock cells. Up- and downregulated genes are in red and green, respectively. (B) affected functional (blue) and canonical pathways (orange) in HT1080/Mhy1 and HT1080/Mhy3 cells identified by the functional clustering analysis. The pathways with P ≤ 0.05 are shown. (C) top up- (red) and down- (green) regulated genes in HRT8/Mhy1 cells relative to HRT8/mock cells. (D) affected disease (blue) and canonical pathways (orange) in HRT8/Mhy1 cells. Only pathways with P ≤ 0.05 are shown.

Figure 7.

Activation of the cancer-specific pathways in trophoblasts. (A) top up- and down-regulated genes in HRT8/Mhy3 cells relative to HTR8/mock cells. (B) affected disease (blue) and canonical pathways (orange) in HRT8/Mhy3 cells. The pathways with P ≤ 0.05 are shown. (C) proliferation specific regulatory network in HTR8/Mhy3 cells. Directly interacting molecules are connected by arrows. Up- and downregulated genes are in red and green, respectively.

Figure 8.

Detection of mycoplasma in colorectal cancer specimens. A 270 bp PCR product specific to the mycoplasma rRNA gene was amplified from the total DNA isolated from colorectal cancer tumor (T) and matching normal (N) specimens (negative cases Cs661 and Cs694, and positive cases Cs544 and Cs583). DNA size markers and the mycoplasma-positive (HepG2 cells infected with M. hyorhinis) and negative (uninfected HepG2 cells and H₂O) controls are shown.