Cytosine methylase and hydroxymethylase activity in mammalian mitochondria

Abstract

Introduction: Mitochondria are integral components of eukaryotic cells, functioning as energy powerhouses and key mediators of diverse metabolic and signaling cascades. As endosymbiotic remnants, these unique organelles retain and express their own DNA. Mitochondrial DNA (mtDNA) is packaged into DNA-protein complexes called nucleoids, and is also subject to epigenetic modification. We identified a mitochondrial isoform of DNA methyltransferase 1 (mtDNMT1) that binds to mtDNA in critical control regions; however, its enzymatic activity remained unexplored.

Results: Here, we show that endogenously-tagged mtDNMT1 purified from mitochondria exhibits time- and concentration-dependent CpG-specific DNA methyltransferase activity, but it is not working alone: DNMT3b cooperates with mtDNMT1 to methylate mtDNA and regulate mitochondrial transcription. In addition, we detect ten-eleven translocase (TET)-like hydroxymethylase activity in mitochondria, demonstrating that mechanisms for both writing and erasing 5-methylcytosine marks are functional in this organelle. CRISPR/Cas9-mediated inactivation of mtDNMT1 and/or DNMT3b activity resulted in a stepwise decrease in mitochondrial methylation across the heavy and light strand promoters of mtDNA, with a significant reduction in transcription of several mtDNA-encoded OXPHOS genes. Interestingly, the effects of mtDNA methylation on mitochondrial transcription are diametrically opposed to the role of promoter methylation in the nucleus, suggesting a novel mode of gene regulation in mitochondria. Cells lacking mtDNMT1 and/or DNMT3b also exhibited a modest reduction in mtDNA content, suggesting that methylation impacts both mtDNA transcription and replication.

Discussion: These observations implicate mtDNA methylation in the fine-tuning of mitochondrial function and suggest a role for aberrant mitochondrial methylase activity in disease.

Keywords: DNA demethylation; DNA methylation; DNA methyltransferase; DNA replication; epigenetics; mitochondrial DNA (mtDNA); transcription.

FIGURE 1

mtDNMT1 and DNMT3b are catalytically active in mitochondria. (A) The fusion PCR strategy used to generate the HCT116 TAP-tagged DNMT1 cell line. LHA, left homology arm; RHA, right homology arm; TAP, tandem affinity purification tag; Neo^R, neomycin resistance cassette. (B) Immunoblots showing DNMT1-TAP localizes properly to the mitochondrial compartment and is expressed at near-endogenous levels. VDAC, voltage-dependent anion channel (mitochondrial loading control); H3K4me³, histone 3 lysine 4 tri-methylation (nuclear contamination control). (C) Schematic of the CpG-specific methyltransferase assay used in (D–F). (D) DNMT1-TAP purified from mitochondria exhibits CpG-specific methyltransferase activity that is linear with respect to time. (E) Slope and R ² linear correlation values for CpG methyltransferase activity of mtDNMT1-TAP, plotted in (D) (F) mtDNMT1-TAP exhibits classic Michaelis-Menten kinetics when methyltransferase activity is plotted over enzyme concentration. (G) DNMT1 and DNMT3b, but not DNMT3a, contribute to total methylation activity in mitochondria isolated from murine ES cells. (+) ctrl, positive M. SssI CpG methylase control; (-) ctrl, Boiled WT negative control; DNMT1+/c, heterozygous catalytic knockout of DNMT1 (1 allele); DNMT1c/c, homozygous catalytic knockout of DNMT1 (both alleles); DNMT3a−/−, DNMT3a knockout; DNMT3b−/−, DNMT3b knockout; TKO, triple knockout, lacking all three DNMTs. Experiments consisted of two biological replicates, each with at least three technical replicates. Error bars represent the standard deviation about the mean. Statistical significance was determined by one-way ANOVA with Tukey’s post-hoc test, *p < 0. 05.

FIGURE 2

Tet hydroxymethylase activity and TET2 in mitochondria. (A) In vitro Tet enzyme assay detects hydroxymethylase activity in mitochondria. (B) Densitometry of slot blots in (A) (C) Immunoblot analysis of whole cell and purified mitochondria from HCT116 cells show that TET1 localizes to the outer mitochondrial membrane, and is not resistant to enzymatic digestion of the intact mitochondrial pellet. M(-), untreated mitochondria; M(+), trypsin-digested mitochondria; VDAC, mitochondrial loading control; H3K4me³, nuclear contamination control. (D) Immunoblot analysis of whole cell and purified mitochondria demonstrate that TET2 is protected from enzymatic digestion of the mitochondrial pellet, suggesting its localization within mitochondria. Error bars represent the standard deviation about the mean for two independent experiments, each with two technical replicates. Statistical significance was determined by two-tailed, two-sample Student’s t-test, *p < 0.05, **p < 0.01.

FIGURE 3

Upregulation of mtDNMT1/loss of p53 increases mitochondrial 5mC and 5hmC levels and exerts gene-specific effects on transcription. (A) MeDIP analysis shows increased 5mC enrichment in the HSP and LSP regions of mtDNA upon mtDNMT1 upregulation/loss of p53. (B) Hydroxy-MeDIP analysis shows similar increases in 5hmC levels upon mtDNMT1 upregulation/loss of p53. (C,D) Upregulation of mtDNMT1/loss of p53 induces gene-specific effects on mitochondrial transcription. mTERF, mitochondrial termination factor site; HSP, heavy-strand promoter; LSP, light strand promoter; ND4c, light strand complementary to ND4; ND5c, light strand complementary to ND5. Experiments consisted of at least two biological replicates, each with at least three technical replicates. Statistical significance was determined by two-tailed, two-sample Student’s t-test, *p < 0.05, **p < 0.01, ***p < 0.005, ****p <0 .0005, n. s., not significant.

FIGURE 4

CRISPR/Cas9 targeting of the DNMT1 MTS in HCT116 WT and 3bKO cells. (A) Schematic of the CRISPR/Cas9-targeted mtDNMT1 locus. Bottom panel displays genetic deletion of Exons 2–21 of DNMT3b in the 3bKO parental cell line. Figure adapted from Rhee, et al. (Rhee et al., 2002). (B) Representative sequence trace for the mt1KO2 clone showing the 19bp deletion within the MTS for mtDNMT1 creates a new stop codon (TGA, red box) between the mitochondrial start (ATG, green box, line 1) and nuclear start (ATG, green box, line 2) codons. Lower case font on line 1 represents the 5′ UTR. (C) Table lists specific frameshift mutations and outcomes for all HCT-GE clones, including the probability of mitochondrial localization of genome-edited MTSs as predicted by MitoProtII algorithms. (D) Immunoblots of purified mitochondrial lysates display a substantial reduction in mitochondrial localization of DNMT1 in HCT-GE cells and complete genetic deletion of DNMT3b in the 3bKO-derived cells. WCL WT, Whole cell lysate from WT HCT116s (positive control for all antibodies); TFAM, transcription factor A of mitochondria (mitochondrial loading control); H3K4me³, histone 3 lysine 4 tri-methylation (nuclear contamination control). (E) Densitometric quantitation of DNMT1 expression in mitochondrial lysates relative to TFAM demonstrates substantial reduction in mitochondrial localization of mtDNMT1 in the HCT-GE cells. (F) Immunoblots of whole cell demonstrate that total DNMT protein levels remain unchanged upon CRISPR/Cas9-mediated deletion of the mtDNMT1 MTS. A DNMT3b antibody confirms the genotype of 3bKO-derived cells. TFAM, transcription factor A of mitochondria (mitochondrial loading control); H3K4me³, histone 3 lysine 4 tri-methylation (nuclear loading control). Error bars represent the standard deviation about the mean. Statistical significance was determined by one-way ANOVA with Tukey’s post-hoc test, ***p < 0.005.

FIGURE 5

Functional consequences of genetic deletion of mtDNMT1 and/or DNMT3b in the HCT-GE cells. (A) MeDIP analysis shows significant alterations in 5mC enrichment across various regions of mtDNA. The most significant changes occurred within the critical control regions HSP and LSP, which displayed stepwise decreases in 5mC levels upon sequential disruption of mitochondrial DNMT activity. (B) ddPCR analysis detects a modest but significant reduction in mtDNA copy number upon genetic loss of both mtDNMT1 and DNMT3b. (C) Mitochondrial transcription originating from both HSP and LSP is dramatically reduced upon loss of mitochondrial methylation. OriL, origin of light strand replication; mTERF, termination factor site; HSP, heavy strand promoter; LSP, light strand promoter; ND1, mitochondrial NADH dehydrogenase 1; Cox1, mitochondrial cytochrome oxidase 1; ND6, mitochondrial NADH dehydrogenase 6. Experiments consisted of at least two biological replicates, each with at least three technical replicates. Error bars represent the standard deviation about the mean. Statistical significance was determined by Welch’s two-sample, two-tailed t-test, *p < 0.05, **p < 0.005, ***p < 0.0005, n. s., not significant.