Characterisation of Type II DNA Methyltransferases of Metamycoplasma hominis

Abstract

Bacterial virulence, persistence and defence are affected by epigenetic modifications, including DNA methylation. Solitary DNA methyltransferases modulate a variety of cellular processes and influence bacterial virulence; as part of a restriction-modification (RM) system, they act as a primitive immune system in methylating the own DNA, while unmethylated foreign DNA is restricted. We identified a large family of type II DNA methyltransferases in Metamycoplasma hominis, comprising six solitary methyltransferases and four RM systems. Motif-specific 5mC and 6mA methylations were identified with a tailored Tombo analysis on Nanopore reads. Selected motifs with methylation scores >0.5 fit with the gene presence of DAM1 and DAM2, DCM2, DCM3, and DCM6, but not for DCM1, whose activity was strain-dependent. The activity of DCM1 for C^mCWGG and of both DAM1 and DAM2 for G^mATC was proven in methylation-sensitive restriction and finally for recombinant rDCM1 and rDAM2 against a dam-, dcm-negative background. A hitherto unknown dcm8/dam3 gene fusion containing a (TA) repeat region of varying length was characterized within a single strain, suggesting the expression of DCM8/DAM3 phase variants. The combination of genetic, bioinformatics, and enzymatic approaches enabled the detection of a huge family of type II DNA MTases in M. hominis, whose involvement in virulence and defence can now be characterized in future work.

Keywords: DNA MTase; Metamycoplasma hominis; Nanopore; RM system; Tombo; dam; dcm; orphan.

Figure 1

Phylogenetic tree of type II DNA MTases presents two main branches of DAM and DCM homologs in M. hominis. MTase sequences of the M. hominis strains tested, whose accession numbers are listed in Supplementary Table S1, were used in ClustalW-based Multiple Sequence Alignment using software program MegAlign 5.08 with default setting for phylogenetic tree construction.

Figure 2

Type II DNA MTase gene loci and cassettes. (A) Position of the DNA MTase gene loci with respect to the genome sequence of ATCC23114 (type strain PG21); dam loci (1–3) in blue, dcm loci (1–8) in green. (B) Composition of the MTase gene loci. Genes are represented by arrows and coloured as follows: chromosomal core genes (accord. to PG21) in grey, conserved genes in the MTase cassettes (red), dam homologs in blue, dcm homologs in green, restriction endonuclease (REase) genes in yellow with blue frame; genes, in some isolated affected by truncation, are shaded.

Figure 3

Prevalence and co-occurrence of type II DNA MTases in M. hominis. (A) Bargraphs of the different dam (blue) and dcm (green) gene prevalence (%) calculated on presence in the 115 M. hominis strains tested (Kruskal-Wallis p = 2.33 × 10⁻⁵⁴). (B) Bargraphs of the number of M. hominis strains with zero to four different solitary MTases (orange) or MTases of RM systems (red). The combination of dam and dcm genes in view on solitary- and RM-MTases is listed in Table S3 for all strains (Mann–Whitney U p = 0.556).

Figure 4

DNA MTase and REase transcripts in selected M. hominis strains. Relative transcript levels of each DNA MTase and, if associated, REase gene (rease) were calculated to the mean of three housekeeping genes (gap, mho_0150, and lgt). Presence of the respective genes is marked by (+); absence of next neighboured RM-genes is marked by (−). Bar graphs show the mean of the transcripts with the standard deviation of two independent cultures, each tested in duplicate.

Figure 5

Alignment of conserved motifs of C5-cytosine MTases of M. hominis. (A) Schematic representation of motif order (I to X) in C5-cytosine MTases with SAM-binding motifs in red and catalytic motifs in green. (B) Consensus sequences of motifs I, IV, VI, VIII, IX, and X in DCM1-7 of M. hominis strains (Mhom) in multiple alignment with each highest homolog non-mollicutes species. Acronyms and homologies of the bacterial species are listed in Table 3.

Figure 6

Alignment of conserved motifs of N-6-adenine and N-4-cytosine MTases of M. hominis. (A) Schematic representation of motif order (I to IX) in N-6-adenine and N-4-cytosine MTases. (B) Consensus sequences of motifs I to X in N-6-adenine MTases (DAM1, DAM2, and DCM8α/DAM3α) and N-4-cytosine MTases (DCM3β and DCM8β) of M. hominis strains (Mhom) in multiple alignment with each highest homologous non-mollicutes. Colour coding of conserved amino acids corresponds to that of Malone et al. [11]. Acronyms and homologies of the bacterial species are listed in Table 3.

Figure 7

Proof of DCM1 activity in methylation sensitive restriction (MSR) analysis. (A) Number of dcm1 genes per genome. (B) Relative transcript levels of dcm1, calculated to the mean of three housekeeping genes (gap, mho_0150 and lgt) with the standard deviation of two independent cultures, each tested in duplicates. (C) Plots of methylation scores calculated with tailored Tombo script (CC^mAGG Kruskal–Wallis chi-squared = 649.26, df = 6, p-value < 2.2 × 10⁻¹⁶, C^mCTGG Kruskal–Wallis chi-squared = 843.05, df = 6, p-value < 2.2 × 10⁻¹⁶). (D) MSR analysis. Restricted DNA was separated on 1% (w/v) agarose gels.

Figure 8

Proof of DAM1 and DAM2 activity in MSR analysis. (A) Presence of full-length genes was assigned a value of 1, truncated genes a value of 0.5, and absent genes a value of 0. (B) Relative transcript levels of dam1, dam2, and dcm2, calculated to the mean of three housekeeping genes (gap, mho_0150, and lgt) with the standard deviation of two independent cultures, each tested in duplicates. (C) Plots of methylation scores calculated with tailored Tombo (G^mATC Kruskal–Wallis chi-squared = 4365.2, df = 6, p-value < 2.2 × 10⁻¹⁶, GAT^mC Kruskal–Wallis chi-squared = 10,077, df = 6, p-value < 2.2 × 10⁻¹⁶). (D) MSR analysis. Restricted DNA was separated on 1% (w/v) agarose gels.

Figure 9

rDCM1 and rDAM2 expression led to respective methylation activity of transformed E. coli. Coommassie blue staining of total protein (A) and immunostaining of histidine-tagged proteins (B) in cell lysate of dam-/dcm-deficient E. coli (−), transformed by pcDNA3.1-(codon-optimized) dcm1 of PG21 (rDCM1) or –(codon-optimized) dam2 of 8958 (rDAM2). (C) MSR analysis of the respective genomic E. coli DNAs; unrestricted (DCM(−) and DAM(−)) or restricted with MvaI (1), SgeI (2), and EcoRII (3) for rDCM1 clone and restricted with Sau3AI (4), DpnI (5), and MboI (6) for rDAM2 clone.

Figure 10

Boxplots of MTases and MGEs. Presence of selected MTases (+ present, − not present) correlated with number of selected MGEs, according to the significance in Pearson’s R correlation analysis (see Figure S6). Median bar is shown in red. (Mann–Whitney U MGE_all vs. dcm1 p = 0.0001; MhoV1 vs. dcm5A p = 0.4677; ICEHo-I vs. dcm8 p = 0.2818; ICEHo-II vs. dcm8 p = 0.0930).