The bacterial epigenome

Depósito de investigación de la Universidad de Sevilla https://idus.us.es/ “This version of the article has been accepted for publication, after peer review (when applicable) and is subject to Springer Nature’s AM terms of use, but is not the Version of Record and does not reflect post-acceptance improvements, or any corrections. The Version of Record is available online at: https://doi.org/10.1038/s41579-019-0286-2 ” Accepted Author Manuscript 1" 2" 3" 4" The bacterial epigenome 5" 6" 7" 8" María A. Sánchez-Romero and Josep Casadesús* 9" Departamento de Genética, Facultad de Biología, Universidad de Sevilla, 10" Apartado 1095, 41080 Sevilla, Spain 11" 12" * E-mail [email protected] 13" 14" 15" 16" Corresponding author: Josep Casadesús. E-mail [email protected]. 17" 18" © María Antonia Sánchez-Romero and Josep Casadesús, 2020. The definitive 19" peer reviewed, edited version of this article is published in Nature Reviews 20" Microbiology, Volumen 18, 2020, DOI: 10.1038/s41579-019-0286-2 21" Version of Record: https://doi.org/10.1038/s41579-019-0286-2 22" 23" " 2" Abstract 1" In all domains of life, genomes contain epigenetic information superimposed 2" over the nucleotide sequence. Epigenetic signals control DNA-protein 3" interactions and can cause phenotypic change in the absence of mutation. A 4" nearly universal mechanism of epigenetic signalling is DNA methylation. In 5" bacteria, DNA methylation has roles in genome defense, chromosome 6" replication and segregation, nucleoid organization, cell cycle control, DNA repair 7" and regulation of transcription. In many bacterial species, DNA methylation 8" controls reversible switching (phase variation) of gene expression, a 9" phenomenon that generates phenotypic cell variants. Formation of epigenetic 10" lineages enables the adaptation of bacterial populations to harsh or changing 11" environments and modulates the interaction of pathogens with their eukaryotic 12" hosts. 13" 14" " 3" [H1] Introduction 1" In the second half of the 20th century, eukaryotic genomes were found to 2" contain information superimposed over the nucleotide sequence 1. In 3" multicellular eukaryotes, this additional, epigenetic information controls 4" differentiation and development, and its perturbation can cause pathological 5" conditions 2. The main epigenetic signals in eukaryotic genomes are DNA 6" methylation 3 and post-translational modification of histones 4, and both 7" mechanisms contribute to establish the transcription patterns that govern the 8" formation of specialized cell types. DNA methylation controls binding of proteins 9" to DNA, and provides a primary layer of information to shape the eukaryotic 10" epigenome 3. DNA methylation is also found in bacterial genomes, and controls 11" DNA-protein interactions like in eukaryotes 5-8 (Box 1). 12" The methylated bases found in bacteria are N6-methyl-adenine (6mA) and C5-13" methyl-cytosine (5mC), which exist also in eukaryotes, and N4-methyl-cytosine 14" (4mC), which is only found in bacteria and archaea (Table 1). Base methylation 15" is post-replicative, and the bacterial genome is methylated at both DNA strands 16" in the absence of DNA replication. Upon DNA replication, base methylation is 17" present in one strand only (hemimethylation) 5. Early studies in Escherichia coli 18" and other gammaproteobacteria unveiled roles for DNA adenine methylation in 19" chromosome replication, mismatch repair, control of transposon activity and 20" regulation of transcription 5, 8. Proteins that can discriminate the methylation 21" state (methylated or hemimethylated) of cognate binding sites were also 22" identified 5. In alphaproteobacteria, DNA adenine methylation was found to 23" govern the cell cycle 9. Seminal studies in uropathogenic E. coli unveiled a role 24" for DNA adenine methylation in the formation of phenotypic variants of bacterial 25" cells 10, in a manner reminiscent of the involvement of DNA cytosine 26" methylation in the formation of eukaryotic cell lineages. At the turn of the 27" century, DNA adenine methylation was found to be essential for pathogenesis 28" of Salmonella enterica subsp. enterica serovar Typhimurium 11, 12, and later of 29" other pathogens 13, 14. Together, these studies extended to bacteria the notion 30" that DNA methylation can exert epigenetic control over the phenotype 6, 7, 15-17. 31" The possibility that phosphorothioation of the DNA backbone may provide 32" " 4" additional epigenetic marks, perhaps in a manner complementary to DNA 1" methylation, remains open 18, 19. 2" A hurdle in the investigation of the roles of DNA methylation in bacteria was that 3" most studies focussed on N6-methyl-adenine, which is a base modification that 4" was difficult to detect. This technical limitation has been overcome upon the 5" development of DNA sequencing technologies that enable the high-throughput 6" analysis of DNA methylation across the genome (the methylome [G]). Two such 7" technologies, single-molecule real time sequencing (SMRT) 20, 21 and nanopore 8" sequencing 22, are currently ‘gold standards’ in the study of bacterial 9" methylomes 23. SMRT technology permits to infer the presence of methylated 10" bases on the template DNA strand from kinetic signatures that are specific for 11" each type of methylated base 20. Nanopore sequencing identifies methylated 12" nucleotides by detecting changes in electric current density upon 13" electrophoretic passage of the sample through an orifice 22. Methylome analysis 14" has shown that base methylation is widespread in the genomes of bacteria and 15" archaea 24, including the small genomes of certain (not all) obligate parasites 25. 16" The wealth of information provided by methylome analysis has also updated 17" (and sometimes changed) classic notions on the physiological relevance and 18" adaptive value [G] of DNA methylation in bacteria. For instance, DNA 19" methyltransferases of restriction-modification (R-M) systems [G] were thought 20" to function in genome defense only, and housekeeping functions were ascribed 21" to solitary (’orphan’) DNA methyltransferases not associated with R-M systems. 22" This distinction is not valid anymore as a number of R-M methyltransferases 23" have been shown to have roles in addition to genome defense, including 24" transcriptional regulation and formation of phenotypic cell variants 26-28. The 25" initial notion that solitary DNA methyltransferases were rare did also not hold 26" up, as methylome analysis has shown that they are abundant 24. Finally, the 27" semantics of the bacterial epigenome, which had been traditionally centered on 28" N6-methyl-adenine, has been extended to C5-methyl-cytosine and N4-methyl-29" cytosine. 30" In this Review, we summarize historic and recent developments in epigenetic 31" signalling by DNA methylation, with emphasis on its contribution to phenotypic 32" heterogeneity in bacterial populations. This bias may be justified by the growing 33" " 5" evidence that DNA methylation-mediated control of lineage formation is a 1" widespread phenomenon in bacteria14,15,27, 28, whereas control of physiological 2" processes such as chromosome replication and mismatch repair may be 3" restricted to certain bacterial clades 8. Emphasis on phenotypic heterogeneity 4" may be further warranted by theoretical analysis that outlines its adaptive value 5" 29, 30. We also discuss, often in a speculative manner, evolutionary aspects of 6" bacterial DNA methylation. Complementary information can be found in other 7" reviews 27, 31-34. 8" [H1] DNA methyltransferases 9" Formation of 6mA, 5mC and 4mC is catalyzed by DNA methyltransferases that 10" transfer a methyl group from S-adenosyl-methionine (SAM) to DNA 35, 36. Most 11" DNA methyltransferases are active as monomers 37. For the methylation 12" reaction, the target base is rotated out (‘flipped out’) of the DNA helix and 13" inserted into the catalytic pocket of the enzyme 35, 36. Depending on the position 14" of the target base in the double helix, DNA methyltransferases are classified 15" into two groups: exocyclic (for 6mA and 4mC) and endocyclic (for 5mC) 37. 16" Aside from this difference, all DNA methyltransferases show structural similarity, 17" and the SAM-binding domain (or catalytic domain) is especially well conserved 18" across kingdoms. By contrast, the target-recognition domain (TRD) shows high 19" variability in sequence and structure, which correlates with manifold target 20" specificities 38. Structural similarity and conservation of motifs between different 21" groups of DNA methyltransferases may be indicative of monophyletic origin, 22" and diversification across kingdoms may have occurred upon insertion and 23" permutation of domains 39. 24" DNA methyltransferases associated with restriction-modification (R-M) systems 25" are abundant, with an average of more than two R-M systems per genome 40, 41. 26" Species with higher numbers of R-M systems exist, and extreme examples are 27" found in Helicobacter pylori strains that encode more than 50 R-M systems 42. 28" In addition, many bacterial and archaeal genomes harbour at least one orphan 29" DNA methyltransferase 24. Orphan DNA methyltransferase genes are also 30" found in about one fifth of phage genomes 43. Most orphan DNA 31" methyltransferases are conserved at the genus level, and the most highly 32" represented are the Dam family in gammaproteobacteria and the CcrM family in 33" " 6" alpha-proteobacteria 24. Bioinformatic detection of a gene encoding a DNA 1" methyltransferase does not imply that the enzyme is active or even expressed: 2" for example, transcription of the gene encoding the YhdJ methyltransferase of 3" gammaproteobacteria is low or absent under laboratory conditions 44. 4" Most DNA methyltransferases recognize specific DNA targets: for instance, 5" Dam and CcrM have a single target site (GATC and GANTC, respectively), 6" whereas Dcm has two target sites, CCAGG and CCTGG 5. Orphan DNA 7" methyltransferases that have several TRDs (and thus are multi-specific) have 8" been described in Bacillus subtilis phages 43. Nonspecific DNA 9" methyltransferases also exist but will not be discussed here 45. 10" DNA methyltransferases of R-M systems methylate DNA in a distributive 11" manner; that is, they methylate only one target upon DNA binding 37. 12" Distributive DNA methylation can be seen as a crucial adaptation of these 13" enzymes to their biological function: the need to re-bind DNA after each 14" methylation event can slow down methylation of phage DNA, thus facilitating 15" restriction by the cognate R-M endonuclease 37. Orphan DNA 16" methyltransferases are generally processive, and methylate multiple targets 17" without dissociating from DNA 37, 46. However, the CcrM methyltransferase of 18" alphaproteobacteria seems to be distributive 47. Processivity can vary 19" depending on the genome context; for instance, certain DNA sequences that 20" flank the GATC site can decrease the processivity of the Dam 21" methyltransferase 48. 22" Bacterial DNA methyltransferases can be essential or dispensable 7, 13, 49. In the 23" genus Yersinia, Dam-dependent methylation is essential in Y. enterocolitica and 24" dispensable in two related species, Y. pseudotuberculosis and Y. pestis 50-52."25" Growth conditions can also determine whether a DNA methyltransferase is 26" essential or dispensable: for instance, the CcrM methylase of Caulobacter 27" crescentus is only essential in rich medium 53. Phenotypes of strains that lack or 28" overproduce a DNA methyltransferase can be indicative of physiological roles 29" that DNA methylation might have in the species under study 14, 49, 54, 55. 30" Moreover, altered patterns of gene expression can provide evidence that the 31" methylation state of promoters or regulatory regions controls the activity of the 32" transcriptional machinery 56, 57. However, altered gene expression does not 33" " 7" necessarily imply direct transcriptional control because indirect effects are 1" possible. For instance, increased levels of SAM in DNA methylase mutants may 2" induce changes in gene expression due to alterations in one-carbon metabolism 3" 58. Similarly, transcriptome differences can reflect indirect control. For instance, 4" upregulation of the SOS regulon [G] in dam mutants of E. coli and S. enterica is 5" not caused by Dam-dependent transcriptional control but by DNA strand 6" breakage 59, 60. Enigmatic cases of post-transcriptional regulation by Dam-7" dependent methylation have also been described 61, 62, and a tentative 8" explanation is that Dam methylation may regulate the transcription of genes 9" involved in control of mRNA stability, mRNA translation or other post-10" transcriptional events. 11" [H1] Hemimethylation and demethylation 12" Bacterial DNA methyltransferases are active on both nonmethylated and 13" hemimethylated DNA 46, 63, thus acting as both writers of the methylome and as 14" copyists that transmit the methylation pattern of the genome to daughter cells. 15" Because every round of DNA replication produces hemimethylated DNA from 16" methylated DNA, methylome inheritance requires methylation of the newly 17" synthesized DNA strand (Fig. 1). The duration of the hemimethylated state after 18" DNA replication varies among bacterial clades, and a crucial factor is the 19" availability of the DNA methyltransferase. In gammaproteobacteria, Dam 20" methylase is synthesized at every stage of growth 64, and the hemimethylated 21" state of GATC sites is short-lived because the Dam methylase trails the DNA 22" replication fork at a short distance, estimated to be less than 10 kbp65. By 23" contrast, the genome of alphaproteobacteria remains hemimethylated during 24" most of the cell cycle because the CcrM methylase is only synthesized at the 25" final stage of chromosome replication 9, 32, 34. 26" As a rule, hemimethylation after DNA replication is transient if the cognate DNA 27" methyltransferase is available. However, at specific sites methylation of the 28" daughter strand can be hindered by binding of proteins that prevent DNA 29" methyltransferase activity, causing methylation loss in a passive manner (Fig. 30" 1). This phenomenon is well documented in E. coli and S. enterica ser. 31" Typhimurium: stable hemimethylated GATC sites are formed when a DNA-32" binding protein protects hemimethylated DNA from Dam methylase activity 6, 14, 33" " 8" 66. If Dam methylation hindrance persists during two consecutive rounds of 1" genome replication, a nonmethylated GATC site is formed 14,15. 2" Undermethylation requires that the processivity of Dam methylase is reduced, 3" which occurs at GATC sites that are flanked by specific, A-T rich DNA 4" sequences 48,67. The effect of sequences outside the GATC site is not surprising 5" as DNA-bound Dam methyltransferase spans 10 bp 68. Transcription factors Lrp 6" 69, OxyR 70-72, Fur 73 and HdfR 74, 75 are examples of DNA-binding proteins that 7" are involved in ’passive’ erasure of methyl groups (see below). 8" Undermethylated GANTC sites have also been detected in alphaproteobacterial 9" genomes, and an example of protection from CcrM methylation has been 10" shown to involve a transcription factor, MucR 76. Passive demethylation may be 11" a common phenomenon in the bacterial world: methylatable sites that lack 12" methylation in one or both DNA strands are common in the genomes of 13" gamma- and alphaproteobacteria 24, 77-80, and some of those sites are located 14" within protein-binding sites 24, 80. In E. coli, the methylation state of certain 15" GATC sites depends on growth conditions, which suggests that environmental 16" effects can contribute to shape the methylome 81. 17" ’Active’ DNA demethylation has been described in E. coli, and its physiological 18" relevance remains to be fully understood. AlkB, a DNA repair protein that 19" removes alkyl groups from nucleic acids, targets also 6mA and removes its 20" methyl group 82. Because AlkB expression is strongly induced upon DNA 21" damage, widespread demethylation of m6A may occur under such conditions. It 22" is thus conceivable that demethylation of physiologically relevant, critical GATC 23" sites might contribute to survival, perhaps delaying genome replication or cell 24" division 82. 25" [H1] Readout of methylation marks 26" The methyl group of 6mA, 5mC and 4mC protrudes from the major groove of 27" the double helix, which provides a platform for DNA-binding proteins to bind to 28" cognate nucleotide sequences. As a consequence, the methylation state of 29" critical sites can either favour or hinder the interaction between a DNA-binding 30" domain of the protein and a cognate DNA site 7. In addition, base methylation 31" can induce or enhance DNA curvature 83-85, which can further modulate DNA-32" protein interactions. 33" " 15" by site-specific recombination and expansion and/or contraction of single 1" sequence repeats can also mediate switching of hsd genes 27, 33 (Fig. 4). An 2" example of the contribution of type I R-M systems to phase variation has been 3" described in Streptococcus pneumoniae, in which genetic rearrangements 4" generate six m6A variants of a type I DNA methyltransferase that controls the 5" transparent-opaque transition [G] 123-125. Each DNA methyltransferase variant 6" generates a distinct methylation pattern in the genome, which results in different 7" gene expression profiles and produces pneumococcal lineages with different 8" virulence capacity 123-125. 9" [H1] Transcriptional control by 5mC and 4mC 10" JHP1050 is a 5mC DNA methyltransferase derived from an ancestral R-M 11" system that is highly conserved among H. pylori strains 126. Mutants that lack 12" JHP1050 show alterations in adherence to host cells, natural competence for 13" DNA uptake, bacterial cell shape and susceptibility to copper 126. These 14" phenotypes correlate with altered patterns of gene expression, and site-directed 15" mutagenesis of GCGC targets at promoter regions provides evidence for direct 16" transcriptional control by 5mC 126. Formation of 5mC may further influence gene 17" expression in an indirect manner, perhaps by changes in DNA topology as 18" previously proposed for m6A 91. 19" In Vibrio cholerae, the VhcM orphan 5mC methyltransferase is necessary for 20" optimal growth both in vitro and during infection, and modulates stress 21" responses 127. Transcriptome changes in vhcM deletion mutants may be a 22" consequence of direct transcriptional control but indirect effects are also 23" conceivable 127. A similar situation may apply to upregulation of the general 24" stress response in E. coli mutants lacking the Dcm 5mC methyltransferase 128 25" and to RpoS-dependent control of a drug transporter 129. 26" The literature contains also one example of epigenetic signalling by 4mC: 27" M2.Hpy.AII, a DNA methyltransferase that introduces a methyl group at the N4 28" position in the first C of TCTTC motifs, controls gene expression and virulence-29" related traits in H. pylori 130. As in other studies discussed above, transcriptional 30" control may be direct or indirect 130. 31" " 16" [H1] DNA methylation and bacterial evolution 1" DNA methylation is both a conservative and an innovative force in bacterial 2" evolution. R-M systems function as barriers for acquisition of genetic elements 3" by horizontal transfer 26, 40. Furthermore, R-M systems can limit recombination 4" between bacterial clades that harbour distinct R-M repertoires 131, 132. A likely 5" mechanism is cleavage of incoming DNA molecules with concomitant 6" shortening of recombination tracts. These conservative roles are strengthened 7" by the behaviour of R-M systems as toxin-antitoxin addiction modules: if an R-M 8" system is lost, the decrease in the number of DNA methyltransferase molecules 9" may permit restriction, and a single chromosome break can be sufficient to 10" cause cell death 40, 133. Among orphan methyltransferases, Dam contributes to 11" genome maintenance by repression of transposable elements 90, 134 and 12" prophages 135, 136, and by reduction of the mutation rate via mismatch repair 104. 13" Interestingly, Dam methyltransferase can also be viewed as an addictive 14" component of the genome as its loss causes genome instability upon DNA 15" strand breakage 59, 137. 16" A major role of both Dam and phase-variable DNA methyltransferases is the 17" formation of phenotypic lineages, which enables division of labour in a 18" community or prepares the community for future changes in the environment 19" (bet hedging) 138. An organism can only have a limited set of traits, and 20" improvement of a given phenotype comes at the expense of another phenotype 21" 139. Lineage formation solves this problem by allotting specific phenotypes to 22" each cell type in a manner that can benefit the whole population. Formation of 23" epigenetic cell variants may also provide a strategy for mutation avoidance: 24" mutations are often deleterious, and their accumulation in a population can 25" decrease fitness 140, 141. In fact, bacterial populations with high mutation rates 26" adapt quickly to a given environment but are outcompeted by non-mutators in 27" the long run 142. By contrast, non-mutational variation can adapt bacterial 28" populations to environmental challenge with lower fitness costs. An example is 29" provided by the Dam methylation-dependent OpvAB bet-hedging system of S. 30" enterica which confers bacteriophage resistance by modification of the O-31" antigen 107 (see above). In an opvAB deletion mutant strain, bacteriophage-32" " 17" resistant mutants can be isolated; however, such mutants often harbour 1" mutations that impair fitness 107. 2" Although Dam methylation-dependent switches and phasevarions may have 3" analogous roles in the formation of bacterial lineages, they must be viewed as 4" the products of distinct evolutionary strategies, each with advantages and 5" drawbacks. Dam-dependent switches are the products of long-term co-evolution 6" with the core genome, and formation of Dam-dependent methylation patterns 7" may be an especially robust mechanism for transmission of epigenetic memory 8" to daughter cells. Furthermore, crosstalk with the housekeeping machinery may 9" permit adjustment of the switching frequency in response to environmental cues 10" 72, 99, 100. However, integration into host regulatory networks may heavily 11" constrain further evolution. By contrast, stochastic switching of phasevarions 12" may be prone to perturbation but at the same time be more evolvable, and 13" independence from host genome control may further increase evolvability. 14" Recombination between DNA methyltransferase alleles may be a powerful 15" mechanism for the generation of DNA methyltransferase alleles with novel 16" TRDs, especially in species that harbour multiple DNA methyltransferase genes 17" 113, 118. 18" High rates of horizontal transfer of DNA methylase genes, either R-M or orphan, 19" can further promote bacterial diversification 40, 41. Indeed, DNA-recognition 20" domains seem to be highly mobile 143. In R-M systems, change of DNA 21" methylation specificity can permit the acquisition of novel foreign elements, 22" including prophages, transposons, plasmids and other mobile elements, 23" perhaps accompanied by the loss of resident, old components of the peripheral 24" genome 41. Furthermore, the acquisition of a novel methylation pattern may 25" modify the transcriptome if the transcriptional machinery of the recipient cell 26" contains appropriate readers. The fate of the new-born pattern of epigenetic 27" regulation will be determined by natural selection, and it might instantly produce 28" a novel bacterial clade in case of high adaptive value. Shuffling and exchange 29" of TRD modules may thus be viewed as a trial-and-error game for bacterial 30" diversification. 31" Use of 5mC as an epigenetic mark has an enigmatic side because 5mC is 32" unstable in all cells, including bacteria, and 5mC deamination produces thymine 33" " 18" 144. In the resulting T:G mismatches, discrimination of the incorrect base is not 1" feasible as T and G are normal bases. If left unrepaired, a GC to AT transition 2" will occur upon DNA replication, and one of the daughter DNA molecules will 3" contain a base pair substitution. Gammaproteobacteria possess a repair system 4" (very short patch repair (VSP) which corrects T:G mismatches by replacing T 5" with C 145. However, a system of this kind is absent in many bacterial clades. 6" For instance, 5mC deamination contributes to the high mutation frequency of H. 7" pylori 126. One can speculate thaaat a high mutation rate may have adaptive 8" value in certain species and may represent a burden in others. However, 9" asymmetric mutation towards AT seems to be a universal phenomenon in 10" bacterial genomes 146, and the bias may be explained by high rate of 5mC 11" deamination 147. 12" [H1] Conclusions and future perspectives 13" Conservation of the DNA methyltransferase structure suggests early 14" evolutionary origin 38, and widespread methylation of bacterial and archaeal 15" genomes may be indicative of positive selection during evolution 24. A 16" conceivable scenario is that DNA methylation arose as a genome defense 17" mechanism associated to restriction 40. An alternative possibility is that 18" restriction enzymes did not initially evolve to destroy foreign DNA molecules but 19" to guarantee that DNA methyltransferases remained active 24. High rates of 20" horizontal transfer of DNA methyltransferase genes have contributed to the 21" spread of DNA methylation in the bacterial world 41, and exchange and shuffling 22" of DNA methyltransferase genes have promoted methylome diversification 24, 27, 23" 118, 143. 24" During evolution, perhaps very early, addition of epigenetic information to the 25" nucleotide sequence turned out to confer adaptive traits to bacterial cells. One 26" such trait is the ability to form phenotypic cell variants, often in a reversible 27" manner 15, 27, 111. An asset of reversibility is the ability to restore phenotypic 28" heterogeneity after a bottleneck, a circumstance especially frequent inside 29" animal hosts 27. The abundance of phasevarions in bacterial pathogens may 30" support this view 27, 33, 114. 31" " 19" In certain bacterial clades, long term co-evolution with the host genome has 1" permitted DNA methyltransferases to participate in the control of housekeeping 2" processes, including the cell cycle 32, 34, chromosome replication and 3" segregation 6-8, DNA repair 104 and secretion of virulence determinants 13. 4" Functional analysis of bacterial methylomes may reveal novel roles in years to 5" come. 6" Modification of the host epigenome by bacterial pathogens is a well-established 7" phenomenon, the significance of which remains to be understood 148-150. 8" Epigenetic memory of a bacterial infection might benefit either the host (for 9" example, by boosting immune responses), the pathogen (for example, by 10" weakening immunity) or both organisms (for example, by facilitating commensal 11" or symbiotic co-existence). An intriguing possibility is that DNA 12" methyltransferases of intracellular pathogens may directly methylate the 13" eukaryotic genome 151. 14" Future applications of bacterial DNA methylation in biotechnology may include 15" the use of inhibitors of DNA adenine methylation as antibacterial drugs 152-155 16" and the design of synthetic switches, sensors and memory devices 156, 157 (Box 17" 2). 18" 19" Acknowledgments 20" Work in the author’s laboratory is supported by grant BIO2016-75235-P from 21" the Ministerio de Ciencia, Innovación y Universidades of Spain and the 22" European Regional Fund. They are grateful to M. van der Woude and L. 23" García-Pastor for discussions. 24" 25" Author contributions 26" J. C. and M. A. S.-R. contributed to discussion of content, wrote the article, 27" reviewed and edited the manuscript before submission and researched data for 28" article. 29" " 20" Conflict of Interest 1" The authors declare no conflict of interest. 2" Publisher’s note 3" Springer Nature remains neutral with regard to jurisdictional claims in published maps 4" and institutional affiliations. 5" Reviewer information: 6" Nature Reviews Microbiology thanks M. R. Oggioni and the other anonymous 7" reviewer(s) for their contribution to the peer review of this work. 8" Table of content: ~40 words 9" DNA methylation affects DNA-protein interactions and has important cellular 10" roles, including the control of reversible switching (phase variation) of gene 11" expression. In this Review, Sánchez-Romero and Casadesús review epigenetic 12" signalling by DNA methylation, focussing on its contribution to phenotypic 13" heterogeneity in bacterial populations. 14" Table. 1. Examples of writers and readers of the bacterial epigenome, and physiological processes or cell phenotypes controlled by 1" DNA methylation. 2" Methylated base Writer Target sequence (5’-3’) Readers Physiological process or phenotype under control N6-methyl- adenine Dam GATC DnaA 101, SeqA 86, MutHLS 104, RNA polymerase 90, transposase 90, 134, Lrp 69, 88, 89, 158, OxyR 70-72, 93, Fur 73 and HdfR 74, 75 Chromosome replication and segregation 86, 101, 103, nucleoid organization 87, 159, mismatch repair 104, transposition 90, 134, conjugal transfer 88, motility 56, synthesis and secretion of virulence determinants 11, 12, 56, 62, 160, envelope structure 161, bacteriophage resistance 107 and antibiotic resistance 155 CcrM GANTC GcrA 76, MucR 105, 162 and RNA polymerase 53, 57 Cell cycle control 32, 34 ModA13 AGAAA Unknown Antibiotic resistance, epithelial cell invasion and biofilm formation 113 SpnD39III Allelespecific (six variants) Unknown Opacity (capsule synthesis) and virulence 123 C5-methyl- cytosine Dcm CTAGG and CCTGG Unknown General stress response 128 and drug transport 129 JHP1050 GCGC Unknown Cell morphology, competence, outer membrane composition and copper resistance 126 " 22" N4-methyl- cytosine M2.HpyAII TCTTC Unknown Virulence (adhesion to host cells) 130 Fig. 1. Methylation, hemimethylation and passive demethylation of GATC 1" sites in gammaproteobacteria. a. The bacterial genome is methylated at both 2" DNA strands in the absence of DNA replication. Upon DNA replication, 3" methylation is present in the template strand only (hemimethylation). 4" Hemimethylation after DNA replication is transient as Dam, an orphan DNA 5" methyltransferase that trails the DNA replication fork, targets GATC sites for 6" methylation to restore the methylated state. This cycle occurs once per round of 7" DNA replication, and affects most GATC sites 5. b. Binding of a protein that 8" functions as a methylation blocking factor (MBF) can prevent methylation of a 9" hemimethylated GATC site by the Dam methyltransferase. If hindrance of 10" methylation persists during two consecutive rounds of DNA replication, a 11" nonmethylated GATC site is formed 66. DNA methylation hindrance by blocking 12" factors occurs only at GATC sites flanked by DNA sequences that decrease the 13" processivity of the Dam methylase 48 (not shown). Undermethylated 14" (hemimethylated or nonmethylated) GATC site states can be transmitted to 15" daughter cells in certain cases. 16" Fig. 2. Control of bistable loci by the formation of alternative Dam 17" methylation patterns. a. Shown is an imaginary locus that harbours two 18" protein-binding sites, each containing a GATC (GATC 1 and GATC 2). Two 19" binding patterns are possible, and each binding pattern causes passive 20" demethylation of a specific GATC site because methylation is hindered at the 21" sites bound by the regulator. Nonmethylation can increase the affinity of the 22" protein for its cognate binding site, thus creating a positive feedback loop that 23" stabilizes the DNA methylation pattern and makes it heritable. However, the 24" system can be reset when hemimethylated DNA substrates are formed upon 25" DNA replication. Switching occurs if the protein moves from one cognate site to 26" the other, and a major factor that determines the frequency of switching in each 27" direction is the affinity of the protein towards each binding site 72,94. Ancillary 28" factors (not shown) can skew switching, either in a stochastic manner or in 29" response to environmental signals 99, 100, 163. b. Epigenetic states of bistable loci 30" of Escherichia coli (pap and sci1) and Salmonella enterica (gtr, opvAB and std) 31" under the control of Dam methylation. The diagrams show regulatory regions 32" " 24" upstream of the promoters, with the promoter-proximal side on the right. 33" Transcription is repressed in the OFF state and activated in the ON state. In all 34" cases, the regulatory region contains binding sites for a transcription factor: 6 35" binding sites for Lrp in pap 92, four for OxyR in gtr 93 and opvAB 106, three for Fur 36" in sci 73, and one binding site for HdfR in std 75. The transcription factor binding 37" sites contain GATC sites: two in pap 92, three in sci1 73 and std 75, and four in 38" opvAB 107 and gtr 93. The methylation state of such sites is dictated by the 39" binding pattern of the transcription factor: GATCs located within sites bound by 40" the transcription factor undergo passive demethylation while GATCs within 41" unbound sites are methylated. This mutual exclusion between Dam methylation 42" and transcription factor binding produces distinct patterns of DNA methylation in 43" the regulatory region of OFF and ON cells. In PapOFF cells, the upstream GATC 44" site is methylated and the downstream GATC is nonmethylated, and the 45" opposite GATC methylation pattern is found in the PapON lineage 92. In the 46" regulatory regions of gtr and opvAB, cells in OFF and ON states display 47" opposite Dam methylation patterns involving two methylated and two 48" nonmethylated GATC sites 72, 93. Formation of nonmethylated GATC sites is 49" likewise observed in Sci1OFF cells as the consequence of Fur binding 73, and in 50" StdON cells as the consequence of HdfR binding 75. Under laboratory conditions, 51" formation of ON and OFF subpopulations is reversible (phase variable) in pap 52" 10, gtr 93, opvAB 106 and std 96 but not fin sci1 73. In pap, switching to ON 53" requires an ancillary factor, PapI 94. 54" Fig. 3. Formation of subpopulations controlled by Dam-dependent 55" methylation. a. Phase-variable transcription of the Salmonella enterica opvAB 56" operon produces a lineage of OpvABOFF cells with long O-antigen chains in the 57" lipopolysaccharide (LPS) (OpvABOFF) and a lineage with shorter O-antigen 58" chains (OpvABON) 105. Shortening of the O-antigen renders the OpvABON 59" lineage avirulent but resistant to bacteriophages (many Salmonella-targeting 60" phages use the O-antigen as receptor). In the presence of a phage, the 61" OpvABOFF subpopulation will be eliminated while the OpvABON subpopulation 62" will survive. When phage challenge ceases, OpvABOFF cells produced by phase 63" variation will survive and virulence will be regained. However, a small 64" " 31" References 205" 1. Holliday, R. Epigenetics: a historical overview. Epigenetics 1, 76-80 206" (2006). 207" 2. Henikoff, S. & Greally, J.M. Epigenetics, cellular memory and gene 208" regulation. Curr Biol 26, R644-8 (2016). 209" 3. Jones, P.A. Functions of DNA methylation: islands, start sites, gene 210" bodies and beyond. Nat Rev Genet 13, 484-92 (2012). 211" 4. Lawrence, M., Daujat, S. & Schneider, R. Lateral thinking: How histone 212" modifications regulate gene expression. Trends Genet 32, 42-56 (2016). 213" 5. Marinus, M.G. Methylation of DNA, in Escherichia coli and Salmonella. 214" Cellular and Molecular Biology (ed. Neidhardt, F.C.) 782-791 (ASM 215" Press, Washington, D.C., 1996). 216" 6. Casadesus, J. & Low, D. Epigenetic gene regulation in the bacterial 217" world. Microbiol Mol Biol Rev 70, 830-56 (2006). 218" 7. Wion, D. & Casadesus, J. N6-methyl-adenine: an epigenetic signal for 219" DNA-protein interactions. Nat Rev Microbiol 4, 183-92 (2006). 220" 8. Løbner-Olesen, A., Skovgaard, O. & Marinus, M.G. Dam methylation: 221" coordinating cellular processes. Curr Opin Microbiol 8, 154-160 (2005). 222" 9. Stephens, C., Reisenauer, A., Wright, R. & Shapiro, L. A cell cycle-223" regulated bacterial DNA methyltransferase is essential for viability. Proc 224" Natl Acad Sci U S A 93, 1210-4 (1996). 225" 10. Blyn, L.B., Braaten, B.A., White-Ziegler, C.A., Rolfson, D.H. & Low, D.A. 226" Phase-variation of pyelonephritis-associated pili in Escherichia coli: 227" evidence for transcriptional regulation. EMBO J 8, 613-20 (1989). 228" The first description of bacterial lineage formation under the control 229" of DNA methylation. 230" 11. Heithoff, D.M., Sinsheimer, R.L., Low, D.A. & Mahan, M.J. An essential 231" role for DNA adenine methylation in bacterial virulence. Science 284, 232" 967-70 (1999). 233" 12. Garcia-Del Portillo, F., Pucciarelli, M.G. & Casadesus, J. DNA adenine 234" methylase mutants of Salmonella typhimurium show defects in protein 235" secretion, cell invasion, and M cell cytotoxicity. Proc Natl Acad Sci U S A 236" 96, 11578-83 (1999). 237" 13. Marinus, M.G. & Casadesus, J. Roles of DNA adenine methylation in 238" host-pathogen interactions: mismatch repair, transcriptional regulation, 239" and more. FEMS Microbiol Rev 33, 488-503 (2009). 240" " 32" 14. Sanchez-Romero, M.A., Cota, I. & Casadesus, J. DNA methylation in 241" bacteria: from the methyl group to the methylome. Curr Opin Microbiol 242" 25, 9-16 (2015). 243" 15. Casadesus, J. & Low, D.A. Programmed heterogeneity: epigenetic 244" mechanisms in bacteria. J Biol Chem 288, 13929-35 (2013). 245" 16. Casadesús, J. & Torreblanca, J. Methylation-related epigenetic signals in 246" bacterial DNA, in Epigenetic mechanisms of gene regulation (eds. 247" Russo, V.E.A., Martienssen, R.A. & Riggs, A.D.) 141-153 (Cold Spring 248" Harbor Laboratory, Cold Spring Harbor, New York, 1996). 249" 17. Atack, J.M., Tan, A., Bakaletz, L.O., Jennings, M.P. & Seib, K.L. 250" Phasevarions of bacterial pathogens: Methylomics sheds new Light on 251" old enemies. Trends Microbiol 26, 715-726 (2018). 252" 18. Chen, C. et al. Convergence of DNA methylation and phosphorothioation 253" epigenetics in bacterial genomes. Proc Natl Acad Sci U S A 114, 4501-254" 4506 (2017). 255" 19. Wang, L., Jiang, S., Deng, Z., Dedon, P.C. & Chen, S. DNA 256" phosphorothioate modification-a new multi-functional epigenetic system 257" in bacteria. FEMS Microbiol Rev 43, 109-122 (2019). 258" 20. Flusberg, B.A. et al. Direct detection of DNA methylation during single-259" molecule, real-time sequencing. Nat Methods 7, 461-5 (2010). 260" 21. Fang, G. et al. Genome-wide mapping of methylated adenine residues in 261" pathogenic Escherichia coli using single-molecule real-time sequencing. 262" Nat Biotechnol 30, 1232-9 (2012). 263" 22. Clarke, J. et al. Continuous base identification for single-molecule 264" nanopore DNA sequencing. Nat Nanotechnol 4, 265-70 (2009). 265" 23. Beaulaurier, J., Schadt, E.E. & Fang, G. Deciphering bacterial 266" epigenomes using modern sequencing technologies. Nat Rev Genet 20, 267" 157-172 (2019). 268" This review summarizes the technologies for mapping bacterial 269" methylomes. 270" 24. Blow, M.J. et al. The epigenomic landscape of prokaryotes. PLoS Genet 271" 12, e1005854 (2016). 272" 25. Lluch-Senar, M. et al. Comprehensive methylome characterization of 273" Mycoplasma genitalium and Mycoplasma pneumoniae at single-base 274" resolution. PLoS Genet 9, e1003191 (2013). 275" 26. Vasu, K. & Nagaraja, V. Diverse functions of restriction-modification 276" systems in addition to cellular defense. Microbiol Mol Biol Rev 77, 53-72 277" (2013). 278" " 33" 27. De Ste Croix, M. et al. Phase-variable methylation and epigenetic 279" regulation by type I restriction-modification systems. FEMS Microbiol Rev 280" 41, S3-S15 (2017). 281" 28. Tan, A., Atack, J.M., Jennings, M.P. & Seib, K.L. The capricious nature 282" of bacterial pathogens: Phasevarions and vaccine development. Front 283" Immunol 7, 586 (2016). 284" 29. Thattai, M. & van Oudenaarden, A. Stochastic gene expression in 285" fluctuating environments. Genetics 167, 523-30 (2004). 286" 30. Kussell, E. & Leibler, S. Phenotypic diversity, population growth, and 287" information in fluctuating environments. Science 309, 2075-8 (2005). 288" 31. Adhikari, S. & Curtis, P.D. DNA methyltransferases and epigenetic 289" regulation in bacteria. FEMS Microbiol Rev 40, 575-91 (2016). 290" 32. Mouammine, A. & Collier, J. The impact of DNA methylation in 291" Alphaproteobacteria. Mol Microbiol 110, 1-10 (2018). 292" A recent article reviewing DNA methylation in alphaproteobacteria. 293" 33. Phillips, Z.N., Husna, A.U., Jennings, M.P., Seib, K.L. & Atack, J.M. 294" Phasevarions of bacterial pathogens - phase-variable epigenetic 295" regulators evolving from restriction-modification systems. Microbiology 296" (2019). 297" A recent review of phasevarions. 298" 34. Mohapatra, S.S., Fioravanti, A. & Biondi, E.G. DNA methylation in 299" Caulobacter and other Alphaproteobacteria during cell cycle progression. 300" Trends Microbiol 22, 528-35 (2014). 301" 35. Jurkowska, R.Z. & Jeltsch, A. Mechanisms and biological roles of DNA 302" methyltransferases and DNA methylation: From past achievements to 303" future challenges. Adv Exp Med Biol 945, 1-17 (2016). 304" 36. Cheng, X. Structure and function of DNA methyltransferases. Annu Rev 305" Biophys Biomol Struct 24, 293-318 (1995). 306" 37. Bheemanaik, S., Reddy, Y.V. & Rao, D.N. Structure, function and 307" mechanism of exocyclic DNA methyltransferases. Biochem J 399, 177-308" 90 (2006). 309" 38. Malone, T., Blumenthal, R.M. & Cheng, X. Structure-guided analysis 310" reveals nine sequence motifs conserved among DNA amino-311" methyltransferases, and suggests a catalytic mechanism for these 312" enzymes. J Mol Biol 253, 618-32 (1995). 313" 39. Bujnicki, J.M. Sequence permutations in the molecular evolution of DNA 314" methyltransferases. BMC Evol Biol 2, 3 (2002). 315" " 34" 40. Ershova, A.S., Rusinov, I.S., Spirin, S.A., Karyagina, A.S. & Alexeevski, 316" A.V. Role of restriction-modification systems in prokaryotic evolution and 317" ecology. Biochemistry (Mosc) 80, 1373-86 (2015). 318" 41. Oliveira, P.H., Touchon, M. & Rocha, E.P. The interplay of restriction-319" modification systems with mobile genetic elements and their prokaryotic 320" hosts. Nucleic Acids Res 42, 10618-31 (2014). 321" 42. Nobusato, A., Uchiyama, I. & Kobayashi, I. Diversity of restriction-322" modification gene homologues in Helicobacter pylori. Gene 259, 89-98 323" (2000). 324" 43. Murphy, J., Mahony, J., Ainsworth, S., Nauta, A. & van Sinderen, D. 325" Bacteriophage orphan DNA methyltransferases: insights from their 326" bacterial origin, function, and occurrence. Appl Environ Microbiol 79, 327" 7547-55 (2013). 328" 44. Broadbent, S.E., Balbontin, R., Casadesus, J., Marinus, M.G. & van der 329" Woude, M. YhdJ, a nonessential CcrM-like DNA methyltransferase of 330" Escherichia coli and Salmonella enterica. J Bacteriol 189, 4325-7 (2007). 331" 45. Murray, I.A. et al. The non-specific adenine DNA methyltransferase 332" M.EcoGII. Nucleic Acids Res 46, 840-848 (2018). 333" 46. Urig, S. et al. The Escherichia coli dam DNA methyltransferase modifies 334" DNA in a highly processive reaction. J Mol Biol 319, 1085-96 (2002). 335" 47. Albu, R.F., Jurkowski, T.P. & Jeltsch, A. The Caulobacter crescentus 336" DNA-(adenine-N6)-methyltransferase CcrM methylates DNA in a 337" distributive manner. Nucleic Acids Res 40, 1708-16 (2012). 338" 48. Peterson, S.N. & Reich, N.O. GATC flanking sequences regulate Dam 339" activity: evidence for how Dam specificity may influence pap expression. 340" J Mol Biol 355, 459-72 (2006). 341" This paper describes DNA sequence elements that influence the 342" processivity of the Dam methyltransferase. 343" 49. Payelleville, A. et al. DNA adenine methyltransferase (Dam) 344" overexpression impairs Photorhabdus luminescens motility and 345" virulence. Front Microbiol 8, 1671 (2017). 346" 50. Julio, S.M. et al. DNA adenine methylase is essential for viability and 347" plays a role in the pathogenesis of Yersinia pseudotuberculosis and 348" Vibrio cholerae. Infect Immun 69, 7610-5 (2001). 349" 51. Robinson, V.L., Oyston, P.C. & Titball, R.W. A dam mutant of Yersinia 350" pestis is attenuated and induces protection against plague. FEMS 351" Microbiol Lett 252, 251-6 (2005). 352" 52. Taylor, V.L., Titball, R.W. & Oyston, P.C. Oral immunization with a dam 353" mutant of Yersinia pseudotuberculosis protects against plague. 354" Microbiology 151, 1919-26 (2005). 355" " 35" 53. Gonzalez, D. & Collier, J. DNA methylation by CcrM activates the 356" transcription of two genes required for the division of Caulobacter 357" crescentus. Mol Microbiol 88, 203-18 (2013). 358" 54. Falker, S., Schilling, J., Schmidt, M.A. & Heusipp, G. Overproduction of 359" DNA adenine methyltransferase alters motility, invasion, and the 360" lipopolysaccharide O-antigen composition of Yersinia enterocolitica. 361" Infect Immun 75, 4990-7 (2007). 362" 55. Shell, S.S. et al. DNA methylation impacts gene expression and ensures 363" hypoxic survival of Mycobacterium tuberculosis. PLoS Pathog 9, 364" e1003419 (2013). 365" 56. Balbontin, R. et al. DNA adenine methylation regulates virulence gene 366" expression in Salmonella enterica serovar Typhimurium. J Bacteriol 188, 367" 8160-8 (2006). 368" 57. Gonzalez, D., Kozdon, J.B., McAdams, H.H., Shapiro, L. & Collier, J. The 369" functions of DNA methylation by CcrM in Caulobacter crescentus: a 370" global approach. Nucleic Acids Res 42, 3720-35 (2014). 371" 58. Friso, S., Udali, S., De Santis, D. & Choi, S.W. One-carbon metabolism 372" and epigenetics. Mol Aspects Med 54, 28-36 (2017). 373" 59. Torreblanca, J. & Casadesus, J. DNA adenine methylase mutants of 374" Salmonella typhimurium and a novel Dam-regulated locus. Genetics 144, 375" 15-26 (1996). 376" 60. Marinus, M.G. Recombination is essential for viability of an Escherichia 377" coli dam (DNA adenine methyltransferase) mutant. J Bacteriol 182, 463-378" 8 (2000). 379" 61. Campellone, K.G. et al. Increased adherence and actin pedestal 380" formation by dam-deficient enterohaemorrhagic Escherichia coli 381" O157:H7. Mol Microbiol 63, 1468-81 (2007). 382" 62. Lopez-Garrido, J. & Casadesus, J. Regulation of Salmonella enterica 383" pathogenicity island 1 by DNA adenine methylation. Genetics 184, 637-384" 49 (2010). 385" 63. Herman, G.E. & Modrich, P. Escherichia coli dam methylase. Physical 386" and catalytic properties of the homogeneous enzyme. J Biol Chem 257, 387" 2605-12 (1982). 388" 64. Boye, E., Marinus, M.G. & Lobner-Olesen, A. Quantitation of Dam 389" methyltransferase in Escherichia coli. J Bacteriol 174, 1682-5 (1992). 390" 65. Campbell, J.L. & Kleckner, N. The rate of Dam-mediated DNA adenine 391" methylation in Escherichia coli. Gene 74, 189-90 (1988). 392" 66. Low, D.A. & Casadesus, J. Clocks and switches: bacterial gene 393" regulation by DNA adenine methylation. Curr Opin Microbiol 11, 106-12 394" (2008). 395" " 36" 67. Peterson, S.N. & Reich, N.O. Competitive Lrp and Dam assembly at the 396" pap regulatory region: implications for mechanisms of epigenetic 397" regulation. J Mol Biol 383, 92-105 (2008). 398" 68. Horton, J.R., Liebert, K., Bekes, M., Jeltsch, A. & Cheng, X. Structure 399" and substrate recognition of the Escherichia coli DNA adenine 400" methyltransferase. J Mol Biol 358, 559-70 (2006). 401" 69. van der Woude, M.W., Braaten, B.A. & Low, D.A. Evidence for global 402" regulatory control of pilus expression in Escherichia coli by Lrp and DNA 403" methylation: model building based on analysis of pap. Mol Microbiol 6, 404" 2429-35 (1992). 405" 70. Wallecha, A., Munster, V., Correnti, J., Chan, T. & van der Woude, M. 406" Dam- and OxyR-dependent phase variation of agn43: essential elements 407" and evidence for a new role of DNA methylation. J Bacteriol 184, 3338-408" 47 (2002). 409" 71. Waldron, D.E., Owen, P. & Dorman, C.J. Competitive interaction of the 410" OxyR DNA-binding protein and the Dam methylase at the antigen 43 411" gene regulatory region in Escherichia coli. Mol Microbiol 44, 509-520 412" (2002). 413" 72. Cota, I. et al. OxyR-dependent formation of DNA methylation patterns in 414" OpvABOFF and OpvABON cell lineages of Salmonella enterica. Nucleic 415" Acids Res 44, 3595-609 (2016). 416" 73. Brunet, Y.R., Bernard, C.S., Gavioli, M., Lloubes, R. & Cascales, E. An 417" epigenetic switch involving overlapping Fur and DNA methylation 418" optimizes expression of a type VI secretion gene cluster. PLoS Genet 7, 419" e1002205 (2011). 420" 74. Jakomin, M., Chessa, D., Baumler, A.J. & Casadesus, J. Regulation of 421" the Salmonella enterica std fimbrial operon by DNA adenine methylation, 422" SeqA, and HdfR. J Bacteriol 190, 7406-13 (2008). 423" 75. Garcia-Pastor, L., Sanchez-Romero, M.A., Jakomin, M., Puerta-424" Fernandez, E. & Casadesus, J. Regulation of bistability in the std fimbrial 425" operon of Salmonella enterica by DNA adenine methylation and 426" transcription factors HdfR, StdE and StdF. Nucleic Acids Res 47, 7929-427" 7941 (2019). 428" 76. Fioravanti, A. et al. DNA binding of the cell cycle transcriptional regulator 429" GcrA depends on N6-adenosine methylation in Caulobacter crescentus 430" and other Alphaproteobacteria. PLoS Genet 9, e1003541 (2013). 431" 77. Ringquist, S. & Smith, C.L. The Escherichia coli chromosome contains 432" specific, unmethylated dam and dcm sites. Proc Natl Acad Sci U S A 89, 433" 4539-43 (1992). 434" 78. Wang, M.X. & Church, G.M. A whole genome approach to in vivo DNA-435" protein interactions in E. coli. Nature 360, 606-10 (1992). 436" " 37" 79. Kozdon, J.B. et al. Global methylation state at base-pair resolution of the 437" Caulobacter genome throughout the cell cycle. Proc Natl Acad Sci U S A 438" 110, E4658-67 (2013). 439" 80. Payelleville, A. et al. The complete methylome of an entomopathogenic 440" bacterium reveals the existence of loci with unmethylated Adenines. Sci 441" Rep 8, 12091 (2018). 442" 81. Hale, W.B., van der Woude, M.W. & Low, D.A. Analysis of 443" nonmethylated GATC sites in the Escherichia coli chromosome and 444" identification of sites that are differentially methylated in response to 445" environmental stimuli. J Bacteriol 176, 3438-41 (1994). 446" 82. Li, D. et al. Exocyclic carbons adjacent to the N6 of adenine are targets 447" for oxidation by the Escherichia coli adaptive response protein AlkB. J 448" Am Chem Soc 134, 8896-901 (2012). 449" This paper describes active demethylation of m6A during DNA 450" repair. 451" 83. Polaczek, P., Kwan, K. & Campbell, J.L. GATC motifs may alter the 452" conformation of DNA depending on sequence context and N6-adenine 453" methylation status: possible implications for DNA-protein recognition. Mol 454" Gen Genet 258, 488-93 (1998). 455" 84. Kimura, T., Asai, T., Imai, M. & Takanami, M. Methylation strongly 456" enhances DNA bending in the replication origin region of the Escherichia 457" coli chromosome. Mol Gen Genet 219, 69-74 (1989). 458" 85. Diekmann, S. DNA methylation can enhance or induce DNA curvature. 459" EMBO J 6, 4213-7 (1987). 460" 86. Waldminghaus, T. & Skarstad, K. The Escherichia coli SeqA protein. 461" Plasmid 61, 141-50 (2009). 462" 87. Sanchez-Romero, M.A. et al. Dynamic distribution of SeqA protein 463" across the chromosome of Escherichia coli K-12. MBio 1, e00012-10 464" (2010). 465" 88. Camacho, E.M. & Casadesus, J. Conjugal transfer of the virulence 466" plasmid of Salmonella enterica is regulated by the leucine-responsive 467" regulatory protein and DNA adenine methylation. Mol Microbiol 44, 1589-468" 98 (2002). 469" 89. Camacho, E.M. & Casadesus, J. Regulation of traJ transcription in the 470" Salmonella virulence plasmid by strand-specific DNA adenine 471" hemimethylation. Mol Microbiol 57, 1700-18 (2005). 472" 90. Roberts, D., Hoopes, B.C., McClure, W.R. & Kleckner, N. IS10 473" transposition is regulated by DNA adenine methylation. Cell 43, 117-30 474" (1985). 475" " 38" 91. Camacho, E.M. et al. Regulation of finP transcription by DNA adenine 476" methylation in the virulence plasmid of Salmonella enterica. J Bacteriol 477" 187, 5691-9 (2005). 478" 92. van der Woude, M., Braaten, B. & Low, D. Epigenetic phase variation of 479" the pap operon in Escherichia coli. Trends Microbiol 4, 5-9 (1996). 480" 93. Broadbent, S.E., Davies, M.R. & van der Woude, M.W. Phase variation 481" controls expression of Salmonella lipopolysaccharide modification genes 482" by a DNA methylation-dependent mechanism. Mol Microbiol 77, 337-53 483" (2010). 484" 94. Hernday, A.D., Braaten, B.A. & Low, D.A. The mechanism by which DNA 485" adenine methylase and PapI activate the pap epigenetic switch. Mol Cell 486" 12, 947-57 (2003). 487" 95. Davies, M.R., Broadbent, S.E., Harris, S.R., Thomson, N.R. & van der 488" Woude, M.W. Horizontally acquired glycosyltransferase operons drive 489" salmonellae lipopolysaccharide diversity. PLoS Genet 9, e1003568 490" (2013). 491" 96. Garcia-Pastor, L., Sanchez-Romero, M.A., Gutierrez, G., Puerta-492" Fernandez, E. & Casadesus, J. Formation of phenotypic lineages in 493" Salmonella enterica by a pleiotropic fimbrial switch. PLoS Genet 14, 494" e1007677 (2018). 495" 97. Suwandi, A. et al. Std fimbriae-fucose interaction increases Salmonella-496" induced intestinal inflammation and prolongs colonization. PLoS Pathog 497" 15, e1007915 (2019). 498" 98. Nou, X. et al. Regulation of pyelonephritis-associated pili phase-variation 499" in Escherichia coli: binding of the PapI and the Lrp regulatory proteins is 500" controlled by DNA methylation. Mol Microbiol 7, 545-53 (1993). 501" 99. Hernday, A.D., Braaten, B.A., Broitman-Maduro, G., Engelberts, P. & 502" Low, D.A. Regulation of the pap epigenetic switch by CpxAR: 503" phosphorylated CpxR inhibits transition to the phase ON state by 504" competition with Lrp. Mol Cell 16, 537-47 (2004). 505" 100. White-Ziegler, C.A., Angus Hill, M.L., Braaten, B.A., van der Woude, 506" M.W. & Low, D.A. Thermoregulation of Escherichia coli pap transcription: 507" H-NS is a temperature-dependent DNA methylation blocking factor. Mol 508" Microbiol 28, 1121-37 (1998). 509" 101. Reyes-Lamothe, R. & Sherratt, D.J. The bacterial cell cycle, 510" chromosome inheritance and cell growth. Nat Rev Microbiol 17, 467-478 511" (2019). 512" 102. Campbell, J.L. & Kleckner, N. E. coli oriC and the dnaA gene promoter 513" are sequestered from dam methyltransferase following the passage of 514" the chromosomal replication fork. Cell 62, 967-79 (1990). 515" 103. Cagliero, C., Grand, R.S., Jones, M.B., Jin, D.J. & O'Sullivan, J.M. 516" Genome conformation capture reveals that the Escherichia coli 517" " 39" chromosome is organized by replication and transcription. Nucleic Acids 518" Res 41, 6058-71 (2013). 519" 104. Modrich, P. Methyl-directed DNA mismatch correction. J Biol Chem 264, 520" 6597-600 (1989). 521" 105. Ardissone, S. et al. Cell cycle constraints and environmental control of 522" local DNA hypomethylation in alpha-proteobacteria. PLoS Genet 12, 523" e1006499 (2016). 524" 106. Cota, I., Blanc-Potard, A.B. & Casadesus, J. STM2209-STM2208 525" (opvAB): a phase variation locus of Salmonella enterica involved in 526" control of O-antigen chain length. PLoS One 7, e36863 (2012). 527" 107. Cota, I. et al. Epigenetic control of Salmonella enterica O-antigen chain 528" length: A tradeoff between virulence and bacteriophage resistance. PLoS 529" Genet 11, e1005667 (2015). 530" 108. Chessa, D., Winter, M.G., Jakomin, M. & Baumler, A.J. Salmonella 531" enterica serotype Typhimurium Std fimbriae bind terminal 532" alpha(1,2)fucose residues in the cecal mucosa. Mol Microbiol 71, 864-75 533" (2009). 534" 109. De Bolle, X. et al. The length of a tetranucleotide repeat tract in 535" Haemophilus influenzae determines the phase variation rate of a gene 536" with homology to type III DNA methyltransferases. Mol Microbiol 35, 211-537" 22 (2000). 538" 110. de Vries, N. et al. Transcriptional phase variation of a type III restriction-539" modification system in Helicobacter pylori. J Bacteriol 184, 6615-23 540" (2002). 541" 111. Srikhanta, Y.N., Maguire, T.L., Stacey, K.J., Grimmond, S.M. & Jennings, 542" M.P. The phasevarion: a genetic system controlling coordinated, random 543" switching of expression of multiple genes. Proc Natl Acad Sci U S A 102, 544" 5547-51 (2005). 545" 112. Fox, K.L. et al. Haemophilus influenzae phasevarions have evolved from 546" type III DNA restriction systems into epigenetic regulators of gene 547" expression. Nucleic Acids Res 35, 5242-52 (2007). 548" 113. Srikhanta, Y.N. et al. Phasevarions mediate random switching of gene 549" expression in pathogenic Neisseria. PLoS Pathog 5, e1000400 (2009). 550" 114. Srikhanta, Y.N., Fox, K.L. & Jennings, M.P. The phasevarion: phase 551" variation of type III DNA methyltransferases controls coordinated 552" switching in multiple genes. Nat Rev Microbiol 8, 196-206 (2010). 553" 115. Kwiatek, A., Mrozek, A., Bacal, P., Piekarowicz, A. & Adamczyk-554" Poplawska, M. Type III methyltransferase M.NgoAX from Neisseria 555" gonorrhoeae FA1090 regulates biofilm formation and interactions with 556" human cells. Front Microbiol 6, 1426 (2015). 557" " 40" 116. Blakeway, L.V. et al. Moraxella catarrhalis restriction-modification 558" systems are associated with phylogenetic lineage and disease. Genome 559" Biol Evol 10, 2932-2946 (2018). 560" 117. Atack, J.M., Yang, Y., Seib, K.L., Zhou, Y. & Jennings, M.P. A survey of 561" Type III restriction-modification systems reveals numerous, novel 562" epigenetic regulators controlling phase-variable regulons; phasevarions. 563" Nucleic Acids Res 46, 3532-3542 (2018). 564" A comprehensive bioinformatic search for phase-variable type III R-M 565" DNA methyltransferases. 566" 118. Gawthorne, J.A., Beatson, S.A., Srikhanta, Y.N., Fox, K.L. & Jennings, 567" M.P. Origin of the diversity in DNA recognition domains in phasevarion 568" associated modA genes of pathogenic Neisseria and Haemophilus 569" influenzae. PLoS One 7, e32337 (2012). 570" 119. Srikhanta, Y.N. et al. Methylomic and phenotypic analysis of the ModH5 571" phasevarion of Helicobacter pylori. Sci Rep 7, 16140 (2017). 572" 120. Anjum, A. et al. Phase variation of a Type IIG restriction-modification 573" enzyme alters site-specific methylation patterns and gene expression in 574" Campylobacter jejuni strain NCTC11168. Nucleic Acids Res 44, 4581-94 575" (2016). 576" 121. Sitaraman, R., Denison, A.M. & Dybvig, K. A unique, bifunctional site-577" specific DNA recombinase from Mycoplasma pulmonis. Mol Microbiol 46, 578" 1033-40 (2002). 579" 122. Doberenz, S. et al. Identification of a Pseudomonas aeruginosa PAO1 580" DNA methyltransferase, its targets, and physiological roles. MBio 8, 581" e02312-16 (2017). 582" 123. Manso, A.S. et al. A random six-phase switch regulates pneumococcal 583" virulence via global epigenetic changes. Nat Commun 5, 5055 (2014). 584" This paper, together with references 124 and 125, characterized a 585" phase variation mechanism that controls switching between 586" pneumococcal phenotypic forms. 587" 124. Li, J. et al. Epigenetic switch driven by DNA inversions dictates phase 588" variation in Streptococcus pneumoniae. PLoS Pathog 12, e1005762 589" (2016). 590" 125. Oliver, M.B., Basu Roy, A., Kumar, R., Lefkowitz, E.J. & Swords, W.E. 591" Streptococcus pneumoniae TIGR4 phase-locked opacity variants differ in 592" virulence phenotypes. mSphere 2, e00386-17 (2017). 593" 126. Estibariz, I. et al. The core genome m5C methyltransferase JHP1050 594" (M.Hpy99III) plays an important role in orchestrating gene expression in 595" Helicobacter pylori. Nucleic Acids Res 47, 2336-2348 (2019). 596" This paper demonstrated direct transcriptional control of bacterial 597" genes by 5mC. 598"