Multiomic map of methylation in Mycoplasma hyorhinis

A multiomic analysis of Mycoplasma hyorhinis strain SK76 has identified three classes of restriction-modification systems, while showing that only a subset of the encoded proteins and methylation patterns were detected under the laboratory conditions examined. By integrating comparative genomics, structural modeling, quantitative proteomics and genome-wide Oxford Nanopore sequencing, researchers linked a prominent N6-methyladenine signature to the Type III methyltransferase Mod1, associated cytosine methylation with the Type II Mhy3-End3 module, and connected a weaker, bipartite adenine methylation pattern to the Type I specificity subunit HsdS2. DNA-cleavage activity was not tested, so HsdR1, End3 and Res1 remain putative cognate restriction components rather than experimentally validated restriction enzymes.

The study focused on M. hyorhinis, a wall-less bacterium primarily known as a porcine pathogen and as a frequent contaminant of mammalian cell cultures. Its small, streamlined genome and limited genetic tractability make it difficult to characterize individual epigenetic systems through conventional gene manipulation. Earlier work has implicated M. hyorhinis restriction-modification proteins and helicases in virulence and host interactions, while Type II DNA methyltransferases have been reported to modify cytosine residues in host DNA following invasion. The present research did not examine host modification, infection outcomes or clinical effects; instead, it characterized methylation in SK76 bacterial DNA and inferred links between individual restriction-modification components and genome-wide modification patterns.

Researchers began by searching the SK76 genome for components of Type I, Type II and Type III restriction-modification systems. The Type I locus contained one HsdM methyltransferase, one HsdR restriction subunit and three HsdS specificity subunits. HsdS1 and HsdS2 occupy adjacent, oppositely oriented genes with a 25-base-pair overlap, whereas HsdS3 is located at a distant chromosomal position. Sequence analysis also corrected the predicted start of hsdS1 in two genome annotations: the actual start lies upstream and includes a putative promoter region. All three HsdS proteins contain two N6 DNA methylase target-recognition domains, but their sequence identities and structural completeness differ. HsdS2 retains a complete coiled-coil central helical region, HsdS1 contains a shortened helix, and HsdS3 lacks the second alpha-helix in that region, suggesting reduced functionality or pseudogenization in the latter variants.

Quantitative proteomics supported a hierarchy among the Type I components. HsdS1 and HsdS2 were detected in both exponential- and stationary-phase cultures, although HsdS1 protein intensity was approximately three log2 units lower than that of HsdS2. No unique peptides could be confidently assigned to HsdS3. The HsdM methyltransferase and HsdR restriction subunit were also detected, and HsdS2, HsdM and HsdR each showed lower abundance in stationary phase than in exponential phase. AlphaFold-Multimer models indicated that the HsdM and HsdS variants could form structurally plausible M2S assemblies, but the models did not establish which specificity subunit functions in vivo. HsdS2 was therefore identified as the most likely principal specificity subunit because of its comparatively high abundance, structural integrity and association with the detected bipartite methylation motif.

The three Type II systems in SK76 contain cytosine methyltransferases Mhy1, Mhy2 and Mhy3, each positioned near a putative restriction-associated protein designated End1, End2 or End3. Phylogenetic analyses supported these local pairings: Mhy1 and End1 clustered with a Mycoplasma agalactiae pair, Mhy2 and End2 with Ureaplasma parvum proteins, and Mhy3 and End3 with Sau3AI-related enzymes from Staphylococcus aureus. Proteomics distinguished the Mhy3-End3 pair from the other candidates. Mhy3 and End3 were consistently detected at substantial abundance, whereas Mhy1 and Mhy2 were not detected and End1 was absent from the measurements. End2 produced a low-confidence identification supported by only two peptides, leaving its biological status uncertain. Failure to detect a protein does not exclude expression below the sensitivity of the mass-spectrometry workflow.

Structural and sequence analysis reinforced the evidence for Mhy3-End3. The predicted Mhy3 structure closely resembled the Staphylococcus aureus Sau3AI methyltransferase, with a root-mean-square deviation of 0.4 angstrom, and contained the PCQ catalytic motif characteristic of C5 cytosine methyltransferases plus an FxGxG motif involved in S-adenosyl-L-methionine binding. The End3 model had a 1.7-angstrom RMSD from the corresponding Sau3AI restriction-enzyme structure and retained a PD-(D/E)XK nuclease motif. Together, these findings support predicted competence and the Mhy3-End3 pairing, but they do not demonstrate that End3 cleaves DNA. The absence of detectable Mhy1 and Mhy2 likewise provides no evidence for their methylation activity in this experiment, without ruling out weak expression or modification below the assays’ detection thresholds.

Oxford Nanopore sequencing supplied a genome-wide view of three modified DNA bases: N6-methyladenine, 5-methylcytosine and 4-methylcytosine. At a moderate motif-discovery threshold of 0.5, the dominant adenine signature was centered on GGAT and occurred at 2,333 genomic sites with 77.2% read coverage. Methylation at this motif was distributed across both strands at similar overall levels, with 47.5% on the positive strand and 50.7% on the negative strand, rather than showing a global strand bias. A second, lower-frequency adenine pattern had the bipartite arrangement TA-N6-ATC and appeared at 890 sites with 47.2% coverage. The predominant cytosine signal was centered on the palindromic GATC sequence, found at 2,320 sites with 90.1% coverage. Only seven 4-methylcytosine sites were detected genome-wide, and the proposed CATCC-centered pattern was supported by just two methylated sites with 0.016 coverage, making a robust biological interpretation unlikely.

Integration of the motifs with protein and sequence data yielded specific but inferential assignments. Mod1 was identified as the principal active Type III methyltransferase because it retained the catalytic DPPY and FxGxG motifs, was detected in the proteome and matched the asymmetric GGAT-centered N6-methyladenine pattern. The compact Type III locus also contained Mod2A, Mod2B, Mod3A and Mod3B adjacent to the Res1 restriction subunit. Mod2A and Mod3A were annotated as truncated proteins lacking the catalytic motif, although short sequence repeats could theoretically restore full-length reading frames through frameshifting. Unique peptides were detected for Mod2B and Mod3B, but not for the truncated variants or putative restored products. The investigators consequently described the additional Mod homologues as possible evolutionary remnants or conditionally regulated elements whose activity under other conditions remains unresolved.

The Type II cytosine pattern was assigned to Mhy3, which targets GATC-containing DNA, with End3 retained as its most plausible restriction partner. The lower-frequency bipartite adenine motif was attributed most plausibly to the Type I system, specifically HsdS2, because bipartite recognition is characteristic of Type I systems and the signal was less abundant than the principal GGAT pattern. Across the three classes, gene presence alone therefore overstated the methylome detected under the tested conditions. The best-supported modules were HsdM with HsdS2, Mhy3 with End3, and Mod1 with Res1, whereas several annotated methyltransferases and specificity subunits lacked corresponding protein or methylation evidence. The study demonstrates how genome-wide modification measurements can narrow predictions generated from sequence alone in a genetically intractable organism, but it does not directly establish restriction activity.

Comparative analysis of 22 M. hyorhinis strains indicated that most restriction-modification genes are broadly conserved, with methyltransferases generally more conserved than restriction enzymes. The truncated organization of the mod2-mod3 region was also conserved across multiple strains, supporting the view that it is a stable but potentially regulatable feature rather than an anomaly confined to SK76. Previously published transposon-mutagenesis data from strain JF5820 classified the methyltransferases as essential, along with End2, while End1, End3 and Res1 were classified as non-essential in that assay. Because the present study used SK76 and did not directly assay restriction activity or essentiality, it cannot establish that these functions are interchangeable across strains. Overall, the multiomic map identifies methylation patterns and probable methyltransferase assignments under one set of culture conditions, while phase variation, low-level expression and DNA cleavage remain to be tested directly.

Subject of Research: Multiomic characterization of DNA methylation and restriction-modification systems in Mycoplasma hyorhinis SK76

Article Title: Integrative multiomic characterization of the methylome and restriction-modification repertoire of Mycoplasma hyorhinis SK76

Article References: Frutos-Grilo, E., D’Angelo, M., Anglada-Girotto, M., & Serrano, L. (2026). Integrative multiomic characterization of the methylome and restriction-modification repertoire of Mycoplasma hyorhinis SK76. Frontiers in Microbiology, 17, Article 1950185. https://doi.org/10.3389/fmicb.2026.1950185

Image Credits: AI Generated

DOI: 10.3389/fmicb.2026.1950185

Keywords: Mycoplasma hyorhinis, DNA methylation, restriction-modification systems, methyltransferases, Nanopore sequencing, proteomics, structural modeling, bacterial epigenetics, Type III methyltransferase, porcine pathogen, Integrative, multiomic