Mobility of DNA sequence recognition domains in DNA methyltransferases suggests epigenetics-driven adaptive evolution

Abstract

DNA methylation is one of the best studied epigenetic modifications observed in prokaryotes as well as eukaryotes. It affects nearby gene expression. Most DNA methylation reactions in prokaryotes are catalyzed by a DNA methyltransferase, the modification enzyme of a restriction-modification (RM) system. Its target recognition domain (TRD) recognizes a specific DNA sequence for methylation. In this commentary, we review recent evidence for movement of TRDs between non-orthologous genes and movement within a gene. These movements are likely mediated by DNA recombination machinery, and are expected to alter the methylation status of a genome. Such alterations potentially lead to changes in global gene expression pattern and various phenotypes. The targets of natural selection in adaptive evolution might be these diverse methylomes rather than diverse genome sequences, the target according to the current paradigm in biology. This “epigenetics-driven adaptive evolution” hypothesis can explain several observations in the evolution of prokaryotes and eukaryotes.

Keywords: :

• DNA methylation
• Helicobacter pylori
• domain movement
• epigenetics
• evolution
• genome evolution
• methylome
• protein evolution
• restriction-modification

This article refers to:

Roles of Epigenetic DNA Methylation and Its Diversity

Among the possible epigenetic modifications of genes, DNA methylation has been well studied both in eukaryotes and prokaryotes. Recent innovations in genome sequencing technology have led to detection of methylated bases even in large genomes and have revealed relationships between DNA methylation and gene expression regulation among others.¹ The various roles of DNA methylation have been well studied in model prokaryotes. For example, methylation status of target sites is used to switch between on/off states for nearby gene expression. Differential recognition between fully-methylated sites and hemi-methylated sites are utilized for strand discrimination during mismatch repair in Escherichia coli.² Methylation status around the genome replication origin leads to regulation of replication in Caulobacter crescentus.²

DNA methylation in prokaryotes is often performed by a DNA methyltransferase, which forms a restriction-modification (RM) system.³ RM systems show methyltransferase (modification) activity and restriction enzyme activity, where the restriction enzyme cuts unmethylated DNA. Both activities show high specificity with respect to DNA sequence recognition. Most of them recognize specific four- to eight-bp long target sequence, unlike eukaryotic DNA methyltransferases that recognize two to three nucleotides.¹

Because DNA methylation affects gene expression, its changes may lead to changes in cell physiology and contribute to adaptive evolution.

Such changes can involve destruction and reconstruction of constituent genes by various mutations, especially frameshift mutations. Length variation of single nucleotide repeats within an open reading frame may lead to phase variation of modification genes,⁴ bringing about global changes in the transcriptome.

Changes in expression of RM genes may lead to diversity in the level of DNA methylation at target sequences for a particular genome. A modification enzyme or special regulatory protein may work as a transcriptional regulator of a RM system. For example, reverse promoters and antisense RNA within coding regions affect gene expression.⁵ The promoter region of RM systems can include its own target sequence of the systems, and the methylation of promoter region leads to repression of expression.^6,7

Changes in the repertoire of RM systems within a genome may result from horizontal gene transfer. By genome context comparison analysis, many RM systems are found to be linked with genes of mobile genetic elements such as those for transposases and integrases or are found to be in a transposon-like structure without transposases.⁸ Diversity in RM system repertoire has been observed through intraspecific genome comparisons, for example, in Helicobacter pylori and Neisseria meningitidis.^9,10 These observations suggest the gain and loss of RM genes by horizontal transfer.

The target sequence of an RM system is primarily determined by the region within a gene called the target recognition domain (TRD).¹¹ TRD sequence variation among the same ortholog group within a species is observed in the Type III mod gene (methyltransferase gene) and likely corresponds to variation in the target sequence.¹² Diversity of TRD sequences of Type I S (specificity) genes, composed of two TRD regions, is also well known.^13-15 A genome inversion between two Type I S genes in a head-to-head manner causes swapping of TRD sequences between the genes.^16,17 The previously discovered allelic diversity of TRD regions were, however, only observed between the same orthologous group or between the same domain sites.

In this commentary, we discuss two novel TRD movement mechanisms that may lead to changes in the DNA methylation target sequence.^18,19 These mechanisms may increase variety of target DNA sequences and may also be responsible for diversity in global gene expression patterns.

Change in Sequence Specificity of DNA Methylation by Movement of Target Recognition Domains

By comparison of RM systems in various strains of Helicobacter pylori, we found two novel mechanisms for movement of the associated TRDs: movement between non-orthologous Type III mod genes and movement between different domain sites of Type I S genes. Both mechanisms utilize DNA recombination at sequences flanking TRD regions.

TRD movement between non-orthologous Type III mod genes

H. pylori has five orthologs of Type III mod genes at five different loci, four of which have large allelic diversity in the TRD region.^12,19,20 We found that TRDs with the same or nearly same amino acid sequence are shared by different orthologs (Fig. 1C). For example, TRD homology group A is observed in loci 1 and 3; TRD C in locus 1 and 3 and TRD D in loci 1, 2 and 4. Sequences outside of the TRD are not well conserved between the orthologs, so how can this TRD sharing occur?

Figure 1. Movement of target DNA recognition domains between non-orthologous genes of Type III mod genes. (A) Gene organization in mod genes. TRD, target recognition domain. Roman numerals, amino-acid sequence motifs conserved among m6A DNA methyltransferases. (B) A likely process of the movement of target recognition domains: DNA recombination at conserved DNA sequences flanking the target recognition domain that encode the conserved amino-acid motifs. (C) Repertoire of orthologs of mod genes in global strains of Helicobacter pylori. Members of the same homology group of target recognition domains are in the same color. Small vertical bars in green and small vertical bars in orange: start codon and stop codon generated by frameshift mutations. Modified from Furuta et al.¹⁹

Sequence comparison revealed that the TRD region in all of the orthologs is flanked by amino-acid sequence motifs conserved among DNA methyltransferases²¹ (Fig. 1A). The DNA sequences encoding these motifs have weak sequence similarity and seem to be subjects for DNA recombination during the movement of TRD region between non-orthologous genes (Fig. 1B).

We also found similar TRD sequences in different ortholog genes for other Helicobacter species. This TRD mobility mechanism may occur not only within a species but also between species.¹⁹ Notably, some of these TRD homology groups are found in various classes such as Bacilli, Clostridia, Fusobacteria in addition to most of Proteobacteria.¹⁹

TRD movement between two domain sites within a Type I S gene and within a Type IIG S gene

A Type I RM system consists of three genes: restriction, modification and specificity. The specificity gene is known to determine the recognition sequence of the whole Type I RM system and is essential for both restriction and modification enzyme activities.³ The specificity gene has two TRD sites, TRD1 and TRD2, each of which recognizes one half of a bipartite recognition sequence (Fig. 2A). H. pylori has four orthologs (six loci in total) of Type I S genes. In three of them (five loci in total), TRD1 and TRD2 are flanked by the same pair of short sequences (Fig. 2A). For these three orthologs, we found that a TRD sequence can move between TRD1 and TRD2 sites within a gene (Fig. 2D). For example, TRD homology groups labeled a, b, c, d and f are observed at both the TRD1 site and the TRD2 site of the gene on locus 1. The amino acid sequences of the same TRD homology groups are identical or almost identical to each other. This movement likely occurred by DNA recombination at the flanking sequences (Fig. 2B and C). This novel process was designated as Domain Movement (DoMo).¹⁸

Figure 2. Domain movement (DoMo) between two domain sites within a gene for the specificity subunit of Type I RM systems. (A) Organization of the specificity (S) gene. TRD1 and TRD2 recognize a 5′ half site and 3′ half site, respectively. Copy number of tandem repeats in the middle striped region defines the distance between the two half sites. (B) A likely process of replacement of TRD sequences by DNA recombination between the flanking sequences, x and y. (C) A likely process of domain movement by recombination between the flanking sequences, x and y. (D) Repertoire of S genes in two loci of global strains of H. pylori. The number in the central white box indicates copy number of the tandem repeat sequences. A white circle indicates a start codon, whereas a black circle indicates a stop codon. Modified from Furuta et al.¹⁸

TRD movement also took place between genes at different loci presumably taking advantage of the similarity of the flanking sequences. For example, TRD homology groups labeled a, b, c, d, e, f and h are shared by locus 1 and locus 2 (Fig. 2D).

Domain Movement was also found for a S gene of Type IIG restriction-modification system in H. pylori.¹⁸

Differences from Other Gene Diversification Mechanisms

Various mechanisms for changing gene sequences are known. For example, exon shuffling and alternative splicing observed in eukaryotes result in changes to domain combinations.^22,23 Recruitment of domain sequences from pseudogenes at a different locus by gene conversion is a well-known mechanism for antigenic variation of cell surface component genes in several bacteria.^24-26 The uniqueness of the mechanisms we have discovered is in movement between different intragenic sites, either inter-locus or intra-locus. Our mechanisms occur without exon-intron structure or presumably other mobile genetic elements. Domain Movement found for TRDs in Type I S genes occurs between two domains within a gene in contrast to gene conversion involving two genes. Therefore Domain Movement represents a novel mechanism for gene/protein alteration.

Adaptive Evolution through Alteration in DNA Methylation Specificity?

What is the biological significance of these gene mechanisms for TRD movement? They very likely lead to alteration and diversification of DNA sequences targeted for methylation. Mechanisms we found may lead to drastic variations of target recognition sequences per lineage. Especially in the case of Domain Movement for Type I S genes, each locus has about 10 types of TRD sequences that can be present at two sites. If we suppose that repeat length between the two domain sites has 10 variations, each locus has 10 × 10 × 10 = 10³ diversity in recognition sequence. H. pylori carries Type I S genes in up to five different loci. Therefore the overall methylation sequence diversity totals to 10¹⁵ (Fig. 3). One genome sequence can take any one of these 10¹⁵ methylome states.

Figure 3. Epigenetics-driven adaptive evolution, a hypothesis. Movements of target DNA recognition domains generate a wide diversity in sequence specificity in a DNA methyltransferase at one locus. Combination of DNA methyltransferases of multiple loci results in huge overall diversity in DNA sequences to be methylated. If one locus can show 1000 DNA sequence specificities, five such loci would generate 10¹⁵ specificities in DNA methylation. One genome sequence may take one of a huge number of epigenome states differing in DNA methylation pattern. Each of these epigenomes (methylomes) may define a specific pattern of global gene expression and a specific set of phenotypic traits. The diverse epigenomes may be the target of natural selection in adaptive evolution. See text for evidence and further detail.

Recent innovation in genome sequencing has made it possible to reveal genome methylation at the single nucleotide resolution.²⁷ This technology may be used for determination of methylome status in vivo and for comparison of diverse methylomes in various strains. The level of expression of genes involved in DNA methylation is another potentially important factor affecting methylome diversity. Although transcription of all S loci in an H. pylori strain was confirmed by transcriptome analysis,²⁸ methylation activity of them must be confirmed by the methylome analysis.

DNA methylation affects nearby transcription. Therefore diversification of methylation specificity should lead to diversification of a cell’s transcriptome. Each epigenome (methylome) may correspond to a specific transcriptome and a specific set of phenotypes. These diverse epigenomes may provide targets of natural selection in adaptation. The paradigm for adaptive evolution in current biology is “selection from preformed genome diversity.” We propose that the epigenome diversity is at least as important as the genome diversity and can provide targets of selection.

Furthermore, alteration of expression of restriction-modification systems discussed above suggests that epigenome variation might be inducible by environmental and internal factors. Adaptive evolution is now often explained by selection from genome sequence variants, but several phenomena are difficult to explain by this paradigm.²⁹ For example, it is difficult to explain by current paradigm about the crossing of a valley on the fitness landscape during acquisition of complex adaptive traits. This “epigenetics-driven adaptive evolution” hypothesis may overcome these difficulties in the current paradigm in evolution and explain recent findings in the epigenetics of various forms of life.^1,2

Acknowledgments

We are grateful to David Dryden for comments on the manuscript and Jacob Albritton for editing grammar, style and contents. This work was supported by the Grants-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (21370001 to I.K., 24790412 to Y.F.) and from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) (24113506 to I.K., 24119503 to Y.F.); the Global COE Project of Genome Information Big Bang from MEXT to I.K.; Grant in Promotion of Basic Research Activities for Innovative Biosciences from Bio-oriented Technology Research Advance Institution to I.K. and the Takeda Science Foundation to Y.F.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.