Introduction of genome-wide CG methylation into E. coli

A – Plasmids encoding active 5mC DNA methyltransferases (M.SssI and M.MpeI) and inactive M.SssI were introduced into Esherichia coli strain B C2523 without active 5mC nuclease McrA to protect methylated DNA before degradation. B Methyl-sensitive restriction enzyme digest. HpaII is blocked by 5mC in the recognized C5mCGG palindrome and MspI can digest methylated C5mCGG and unmethylated CCGG sequences. C Genomic DNA was isolated from strains expressing active methyltransferases (M.SssI and M.MpeI) and inactive M.SssI, and global 5mC level was calculated as a percentage of total cytosine.

MMS-induced sensitivity in strains carrying active 5mC methyltransferases.

WT strains and alkB mutants with M.SssI and M.MpeI methyltransferases were treated with 20 mM MMS. The y-axis shows the percentage survival relative to the survival under the same conditions without MMS. Each set of 3 points shows three separate clones derived from a single transformation with the indicated plasmid. Three separate transformations were performed for each plasmid.

H2O2-induced sensitivity in strains carrying an active 5mC methyltransferase.

WT strains and alkB and fpg mutants with M.SssI methyltransferase were treated with 10 mM H2O2. The percentage of survivals was calculated for each strain and normalized to the number of colonies growing under the same conditions without H2O2 treatment. Each experiment shows independent clones from the same transformation. Three separate transformations were performed for each plasmid.

Reactive Oxygen Species production in the presence of M.SssI and H2O2. M.SssI was either induced with arabinose or kept repressed at the beginning of the experiment. Fluorescence per cell is plotted over 1200 minutes for each of the indicated conditions. The lines represent the loess fit from pooled data across two replicate experiments (two separate clones from a single transformation for each condition). The dashed lines show the standard error of the loess fit. Note that the error of the fit is too small to be visible as separate lines for conditions with very low reactive oxygen species.

Production of 5hmC and 5fC from upon H2O2 treatment in the presence of M.SssI. A and B show representative mass spectrometry traces for peaks identified as 5hmdC and 5fdC respectively. Quantitation is in figure C and D, where data points are the mean of two technical replicates and each data point is a separate clone from the transformation. Y-axis shows level of 5hmC or 5fC measured by mass spectrometry as a percentage of unmodified cytosine. P-value from a one way anova for effect of the experimental condition on the measurement is indicated on each plot.

H2O2-induced sensitivity in strains carrying an active 5mC methyltransferase and expressing TET responsible for oxidative derivatives of 5mC.

The WT strain and a strain carrying IPTG-inducible TETs (Negleria TET – NgTET or phage TET – TET43, more details in Materials section), which modifies 5mC to its oxidative derivatives that, in addition, also overexpressed inactive or active M.SssI methyltransferase, were treated with 10 mM H2O2. A – Drop plate test for E. coli cells without and after 10 mM H2O2 treatment (more technical details on Figure 3). B – The percentage of survivals from WT and NgTET cells was calculated for each strain and each condition and normalized to the number of colonies growing under the same conditions without H2O2 treatment. Each set of 3 points represents 3 independent clones from a single transformation and three separate transformations were performed.

Model: sources of DNA damage due to DNMT expression. Repair pathways are shown in green, enzymatic pathways in blue and DNA damage mechanisms in red.

Drop test screening sensitivity of M.SssI expressing cells to various DNA damaging agents. WT, alkB mutants and Fpg mutants are shown.

Induction of M.SssI expression under control of pBAD promoter by arabinose. The plasmid was digested with methyl-sensitive restriction enzyme HpaII. The presence of the upper band indicates undigested plasmid. Arabinose maintains methyltransferase expression even in the presence of 0.15% glucose, whereas removal of arabinose leads to repression of the methyltransferase. Three independent clones from a single transformation are shown.

Removal of glucose affects ROS production in E. coli. Fluorescence per cell was measured over 1200 minutes for TOP10 cells in the absence of the M.SssI plasmid. 0.15% glucose is sufficient to repress ROS levels, demonstrating that it is removal of glucose rather than addition of arabinose that causes ROS accumulation.

Transcriptomics analysis. A shows a volcano plot of genome wide transcript level changes measured by RNAseq comparing M.SssI expressing cells (arabinose) to M.SssI repressed cells (glucose). The x axis shows log2 fold change and the y axis shows the Benjamani Hochberg adjusted p-value calculated through DESeq2. alkB and ada are highlighted. B shows the log2 fold change in transcript levels upon H2O2 treatment of cells in which M.SssI is repressed (x axis) compared to cells where M.SssI is expressed (y axis). Catalase genes katG and katE are highlighted.