Methylation of DNA
Chapter
53
M. G. MARINUS
DNA methylation in bacteria is most often thought of in its role to protect DNA from restriction endonucleases. In addition to this role, however, studies in Escherichia coli have shown that methylated bases have other biological functions. As described below, DNA adenine methylation is frequently used to control the rate at which these functions exert their effects. Thus it is primarily used for regulatory purposes.
Since the last edition of this book, several reviews of DNA methylation have appeared that are either short (6, 51, 61) or long (43, 54) and are considered complementary to this article.
The DNA of E. coli K-12 contains two modified bases; 6-methyladenine (6-meAde) and 5-methylcytosine (5-meCyt). About 1.5% of all adenines and 0.75% of all cytosines in the chromosome are methylated, and the modifications occur in specific sequences resulting from the action of three DNA methyltransferases (Table 1). The EcoK methyltransferase, which is encoded by the hsd (host specificity) genes, is part of the classical EcoK restriction/modification system described in detail in chapter 52 of this volume.
Table 1DNA methyltransferases in E.coli K-12 |
The Dam (DNA adenine methyltransferase) enzyme, which modifies GATC sequences, forms over 99% of the 6-meAde in E. coli DNA, since strains lacking this enzyme contain only the contribution expected from the EcoK enzyme (46, 69). The Dcm (DNA cytosine methyltransferase) protein, methylating CC(A/T)GG sites, is responsible for all the 5-meCyt in DNA, since none of this modified base can be detected in cells deleted for the dcm gene (2). A dam dcm hsdS mutant contains no detectable modified bases in DNA, indicating that such bases are not essential for E. coli viability.
Analyses of E. coli DNA sequences (5, 23) have indicated the following details about the GATC tetranucleotide. (i) It is represented, on average, once every 214 nucleotides, which is close to the 1/256 expected in a random base sequence. (ii) It is present at a higher than expected frequency in numerous chromosomal locations (e.g., dnaA, rpsP, metL, malP, rplS, xylB, gltX, and guaBA) in addition to oriC. The significance, if any, of this clustering is unknown for all these genes except dnaA and oriC (see below). (iii) It is found more frequently in translated regions than in noncoding or nontranslated regions, which is consistent with more frequent mismatch repair surveillance. In particular, rRNA- and tRNA-encoding genes exhibited the lowest GATC content of all genes examined. This deficiency may be correlated with unwanted palindromic secondary structures. (iv) Finally, the GATC tetranucleotide is never separated from another GATC sequence by more than 2 kb. This allows for dam-directed mismatch repair to occur over the whole genome, since it is not efficient at distances greater than 2 kb (53).
The statistical data described above give the frequency of GATC sites in chromosomal DNA. These sites, however, can be present in unmethylated, hemimethylated, or fully methylated configurations. All GATC sites appear to be methylated in chromosomal and plasmid DNA isolated from E. coli by standard methods and using restriction enzymes such as DpnI and DpnII to monitor methylation status. DpnI cleaves only at methylated sites, DpnII cleaves only at unmethylated sites, and Sau3AI will cut regardless of methylation status. Neither DpnI or DpnII digests hemimethylated sequences (Table 1). Newer techniques such as pulse-field gel electrophoresis of digested DNA and specific end-labeling procedures (67, 76), however, indicate that the E. coli chromosome contains about 20 specific, unmethylated dam sites. The number and intensity of unmethylated sites in the chromosome vary depending on growth phase and growth rate, suggesting that the proteins which bind to them could be involved in gene expression or maintaining chromosome structure. The unmethylated dam sites appear to be mostly (67) or completely (56) modified in strains overproducing Dam, suggesting that the enzyme competes with other DNA binding proteins at these specific sites. Alternatively, the increased Dam concentration may allow for modification of DNA structures (e.g., non-B-form DNA) relatively resistant to methylation at the normal cellular level of the enzyme.
Evidence for competition between Dam and other DNA binding proteins at seven unmethylated sites has been obtained. These findings indicate that these seven sites are involved in regulation of gene expression, and they are discussed in more detail in that section below.
In addition to the unmethylated GATC sites discussed above, persistent hemimethylated sequences have been detected in the chromosome (16, 55). These are distinct from the hemimethylated GATC sites which occur transiently immediately behind the replication fork due to the time lag in modifying new Dam methylation sites. The persistent hemimethylated sites are discussed in more detail in the section on chromosome replication below.
Analyses of the Dcm recognition sequences, CCTGG and its complement CCAGG, indicated that these occur at a higher than expected frequency: every 351 bp instead of every 512 bp as predicted from random sequence (23). As discussed below, the Dcm sites are constantly subjected to cycles of deamination of the 5-meCyt followed by repair of the resulting T-G mismatch and subsequent remethylation by Dcm. This cycle may have affected the abundance of nucleotide sequences in the genome (9, 49).
The state of methylation at dcm sites can be monitored by digestion with EcoRII, which cuts only if the sequence is unmethylated, and BstNI, which cleaves regardless of methylation status (Table 1). As is the case with dam sites, a small but undetermined number of unmethylated dcm sites have been detected in E. coli chromosomal DNA (67). One assumes that there must also be hemimethylated dcm sites in chromosomal DNA, but their existence has not yet been demonstrated.
The investigation of unmethylated and hemimethylated dam and dcm sites in chromosomes has demonstrated the utility of this approach to identify regions of the chromosome with interesting biological features such as sequences that are bound by proteins which have regulatory functions. These kinds of studies should allow a functional dissection of the E. coli chromosome to be integrated with DNA sequence information.
Methylated dam and dcm sites are found in most other enterobacteria (4, 15), and the E. coli dam gene DNA hybridizes under stringent conditions to the DNA of other enterobacteria (15). The biological functions associated with dam and dcm sequences in these related bacteria are presumed to be the same as those in E. coli. Methylated dam sites have also been detected in various gram-positive and gram-negative bacteria as well as in some archaebacteria (56). Methylated dcm sites are found only in members of the Enterobacteriaceae (M. Lieb, personal communication). Strains with a dcm gene also contain the vsr gene (see below).
There are many bacterial species (and all eukaryotes) which do not contain methylated dam and/or dcm sites and presumably have other mechanisms to substitute for Dam and Dcm functions.
The hsd genes which specify the EcoK methyltransferase map at 99 min on the genetic map and 4,615 kb on the physical map. The corresponding map locations for the dam gene are 74 min and 3,536 kb, and those for the dcm gene are 43 min and 2,042 kb. The DNA methylation genes are thus unlinked to one another. Both the dam and dcm genes have been cloned and sequenced (10, 15). All known dcm mutations map to a single gene and are recessive. Similarly, all dam mutations, except one, are in a single complementation group and are recessive. The exceptional mutation is very leaky, making it difficult to assign conclusively within the same complementation group. These genetic data suggest that no other functional methylation genes exist in E. coli.
The 834-bp dam gene (GenBank accession no. J01600) is part of a transcriptional unit containing at least four genes (34, 40) and perhaps six or seven (A. Lyngstadaas and E. Boye, personal communication). The locations of promoters and a transcriptional terminator are shown in Fig. 1. Each promoter has been cloned individually, and the order of promoter strength is P2>P1>P3>P4>P5 (L. J. Rasmussen, personal communication). Insertion of the cat (chloramphenicol acetyltransferase) coding sequence into the aroK gene reduces transcription across the dam gene by 70%, and a mini-Tn10 insertion in urf74.3 does so by 90% (40). These data indicate that promoters P1 and P2 (situated about 3.5 kb upstream of dam) and P3 (located 2 kb upstream) are the most important for dam gene transcription. Promoters P1 through P4 all show the typical RNA polymerase sigma-70 recognition sequences.
Only promoter P2 has thus far been shown to be regulated (65). This promoter is growth rate regulated by a mechanism distinct from that used for rRNA and tRNA gene promoters. Transcription initiation from P2 is not affected by the stringent response, ribosomal feedback, or the level of Fis protein, all of which affect growth rate-dependent rRNA and tRNA promoters (see chapter 90). Conversely, mutations in the cde (constitutive dam gene expression) gene, located at 15 min on the genetic map and 670 kb on the physical map, abolish growth rate regulation of the dam but not rRNA growth rate-dependent promoters (65). It is not yet known what product the cde gene specifies.
The rationale for growth rate regulation of the dam gene may be to correlate Dam levels with the amount of hemimethylated DNA close to the replication fork and at oriC. Cells growing with different doubling times synthesize DNA at different rates. To maintain the optimal level of hemimethylated DNA, the amount of Dam must be adjusted accordingly (65). If dam gene expression were not regulated, too much or too little Dam would result in increased mutagenesis, asynchronous initiation of chromosomes, and alteration of the frequency of transposition.
The 1,419-bp dcm gene (GenBank accession no. X13330; M32307) is overlapped at its 3' end by the first six codons of the vsr gene, which is in a +1 register relative to dcm (19). Such an overlap is uncommon in E. coli and in this case may serve to link the expression of these genes. Both genes appear to be transcribed into a single mRNA, and translation of vsr appears to be dependent upon translation of the upstream dcm coding sequence (19). The mechanism by which this is achieved is not known. The location of the promoter(s) and its mode of regulation are also unknown. The possibility of growth rate regulation of dcm gene expression has not been tested. There is no obvious phenotype associated with the under- or overproduction of Dcm.
DNA methyltransferases transfer the methyl group from S-adenosyl-l-methionine (SAM) to specific residues in double-stranded DNA. For the HhaI methyltransferase, the base to be modified is flipped out of the DNA helix into the active site of the enzyme (36). It is probable that the same basic mechanism is used for the E. coli methyltransferases. In E. coli the normal substrate is hemimethylated DNA. That is, the parental strand is methylated and methyl transfer occurs only onto the newly synthesized unmethylated strand.
The methylation of specific GATC sites in DNA of exponentially growing cells is rapid, occurring within the minimum time (about 1 min) allowed by the sensitivity of the method (16). An already mentioned exception to this involves GATC sites in oriC and the dnaA promoter, which remain hemimethylated longer than other sites; this is possibly due to binding by a membrane fraction discussed in Initiation of Chromosome Replication, below.
Dam has been purified 3,000-fold and is a single polypeptide chain of 278 amino acids with an apparent molecular size of 32 kDa. It has an s 20,w of 2.8S and a Stokes radius of 2.4 nm and exists in solution as a monomer (31). The enzyme has a turnover number of 19 methyl transfers per min and an apparent Km of 3.6 nM for DNA. Double-stranded DNA is a better methyl acceptor than denatured DNA, and there is little difference in the rate of methylation between unmethylated and hemimethylated DNA. Dam transfers one methyl group per DNA binding event even when binding a fully unmethylated site.
Dam has been suggested to have two SAM-binding sites: a catalytic site, and one which increases specific binding to DNA perhaps as a result of an allosteric change in the protein (7). DNA binding and/or methyl transfer is influenced by flanking sequence; the optimal sequence is 3 bp of GC 5' to the GATC and 2 bp of AT and 1 bp of GC 3' to the GATC. Dam is thought to bind the template and to slide processively along the DNA searching for substrate sequences (8).
In fast-growing bacteria there are 130 Dam molecules per cell in K-12 strains and 100 molecules per cell in B strains (14). Each of these molecules would need to transfer 39 methyl groups per min to methylate all available GATC sites in a cell with a doubling time of 30 min. Given the experimentally determined value of 19 turnovers per min per molecule, either the level of Dam in the cell must be limiting or the turnover number is at least twofold higher under in vivo conditions.
The Dam enzyme of E. coli is related at the amino acid level to the corresponding Dam enzymes of phages P1, T1, T2, and T4 and to the DpnII, EcoRV, and Porphyromonas methyltransferases (3, 28, 38). A comparison of the amino acid sequences indicates several conserved regions, and one of these, the DPPY motif, appears to be involved in SAM binding (37). No functional roles have yet been unambiguously described for the other conserved regions. In T4 Dam, mutations in Pro-126 and Phe-127 alter the capability of enzyme to methylate noncognate sites (52), suggesting that the conserved region in E. coli Dam containing these amino acids may serve the same function.
The properties of the purified Dcm protein have not been published, but from the DNA sequence, a 472-amino-acid protein of 53,404 kDa should be produced. Protein sequence comparisons indicate that, like other 5-meCyt methyltransferases, Dcm contains 10 conserved motifs including a Pro-Cys motif (62). The cysteine residue is essential for catalysis but not for DNA binding, suggesting a mechanism of methyl transfer (Fig. 2) similar to that for thymidylate synthase (25), i.e., attack of the C-6 of cytosine by cysteine 177 of Dcm to activate the C-5 position for methylation (24, 78).
To identify the biological role for cytosine methylation, mutant strains lacking this modified base in DNA were isolated (27, 45). Unfortunately, no obvious phenotype has yet been found associated with the dcm mutations. In discussing possible functions for 5-meCyt, it is worth remembering that in contrast to E. coli K-12, E. coli B is naturally dcm –.
The most widely used dcm allele, dcm-6 (45), is defective in both methylation and VSP repair (see below) and shows mutational changes in codons 26 and 45 compared to wild type (19). The polar effect of the nonsense codon (TGA) at position 45 in dcm would most easily explain the effect on vsr.
In addition to dcm-6, the mec mutant allele of dcm (27) has been frequently employed, although the location of the mutation in the gene is not known. Two large deletions which remove dcm and additional genes have been shown to lack Dcm methylation (2, 10).
Spontaneous mutational hotspots for amber nonsense mutations occur in lacI at the 5-meCyt residue in the Dcm recognition site CCAGG, altering it to CTAGG (17). A similar result was obtained in the cI gene of phage lambda, and anomalous recombination frequencies were obtained when these amber mutations were used in genetic crosses (39). This led to the discovery of a very short patch (VSP) repair system correcting T-G mismatches in Dcm recognition sequences.
Such T-G mismatches can occur in nonreplicating DNA by the deamination of 5-meCyt. This reaction is analogous to the deamination of cytosine to form a uracil-guanine mismatch, which is a substrate for uracil-N-glycosylase. In a similar manner, the T-G mismatch is a substrate for the strand- and sequence-specific Vsr endonuclease, followed by conventional DNA polymerase I-dependent excision repair (29). VSP repair can thus be viewed as counteracting the potential mutagenic effects of 5-meCyt deamination (Fig. 3).
Unmethylated dcm sites are substrates for the EcoRII restriction endonuclease (27), suggesting that one function may be protection of DNA from group N plasmids which produce these restriction endonucleases (23). Such plasmids, however, appear to turn on the cognate modification enzyme before turning on the restriction protein, after transfer into a naive cell. It is of interest that the EcoRII methyltransferase shows about 70% amino acid sequence similarity with Dcm. Both enzymes methylate the same DNA sequence, and the function of M.EcoRII is known: it protects DNA from cleavage by EcoRII.
Mutant strains lacking DNA adenine methylation were isolated in order to identify the role of this methylated base in cell metabolism. Unlike the dcm mutants, there are several phenotypic traits associated with dam mutants which have helped to define the multiple roles of 6-meAde in DNA metabolism.
The most commonly used dam mutant alleles are dam-3, dam-4, dam-13::Tn9 (chloramphenicol resistance), and dam-16::Kanr. The latter allele has a substitution of most of the normal dam sequence by a kanamycin resistance determinant (57). The locations of the mutations in the other mutant alleles are not yet known.
Two dam mutations in Salmonella typhimurium have also been described (66). The dam-1 allele confers properties similar to those of dam mutant alleles in E. coli. In contrast, bacteria bearing dam-2 appear to have fully methylated DNA and lack most dam phenotypes (66). It is possible either that the allele is "leaky," resulting in slower methylation at the replication fork, or that the mutant enzyme is more sensitive to its substrate (SAM) or to inhibition by one of the products (S-adenosylhomocysteine).
The dam mutants exhibit a variety of phenotypic traits and other properties (Table 2). The bewildering array of differences compared to the wild type suggests that dam methylation and Dam itself have multiple functions in the cell. For several of the functions described in detail below, the amount of hemimethylated DNA trailing the replication fork is critical. Decreasing or increasing the level of hemimethylated DNA by using a Dam- overproducing plasmid or a dam mutant, respectively, profoundly alters the function involved.
Table 2Altered physiological properties in a dam mutant |
The most direct and convincing evidence for the involvement of dam methylation in mismatch repair comes from the use of in vitro-constructed heteroduplexes of phage DNA (63). Heteroduplexes containing a mismatched base pair were constructed with one strand methylated, both strands methylated, or neither strand methylated. The unmethylated strand was preferentially repaired in heteroduplexes containing one methylated and one unmethylated strand. If neither strand was methylated, repair occurred equally on both strands. No repair was observed when both strands were fully methylated (63). These results indicate that the function of Dam methylation is to impart strand selectivity. That is, repair of base mismatches or deletion/insertions of up to four nucleotides (58) is confined to the unmethylated strand.
Further evidence for the role of Dam in strand discrimination is that dam mutants and cells overproducing Dam show a mutator phenotype (30, 47). These results suggested that the role of the repair system in the wild type is to remove replication errors in the newly synthesized undermethylated DNA strand trailing the replication fork (Fig. 4). In dam mutants, strand discrimination for repair is lost and mutations are introduced into the parental strand using the newly synthesized mutant strand as template. In Dam-overproducing cells, the high concentration of Dam greatly reduces the transient lifetime of hemimethylated GATCs in newly replicated DNA, thereby preventing mismatch repair. For further details about this repair system see chapter 121 in this volume and the review by Modrich (53).
The misdirected mismatch repair in a dam strain can explain many of the dam-associated phenotypes listed above: the mutator phenotype; hyper-recombination and increased transposition; increased basal level of SOS gene transcription; increased breaks in DNA and induction of lysogenic phages; the double mutant inviabilities and suppression by mut alleles (see Table 2 for references).
The double-strand DNA breaks, which are primarily responsible for many of the phenotypes, are undoubtedly produced by MutH endonuclease action on complementary strands at the same GATC or very closely spaced GATCs. These double-strand DNA breaks must be repaired in order for the dam cell to remain viable (77). This would explain the necessity for RecA and RecBCD enzymes, since these are required for double-strand break repair.
The SOS system is not only induced by the DNA strand breaks but is also required for cell survival in a dam mutant (60). This is reflected in the higher basal level of SOS gene expression (subinduction) in dam cells. The products of the recA and ruv SOS genes are essential for dam mutant survival. Surprisingly, elimination of MutH by mutation does not prevent SOS subinduction, indicating that an additional SOS-inducing signal must be generated (59).
The roles of RecA and Ruv, both of which are part of the SOS regulon and essential components of homologous recombination, in maintaining dam mutant viability are unknown. The observations above suggest that SOS-inducible recombinational repair is essential for E. coli survival in the absence of adenine methylation at GATC sites.
Initiation of chromosome replication at the origin, oriC, requires the participation of the DnaA initiation protein. Dam methylation helps to regulate this process but is not an absolute requirement for successful initiation. Rapidly growing E. coli cells contain multiple origins which initiate simultaneously, and Dam methylation is part of the synchronization mechanism. The effect of Dam on each of these processes is described in more detail below.
The oriC region contains 10 times more GATC sites than expected in a random sequence, and initiation of chromosome replication occurs most efficiently on fully methylated oriC DNA. This was demonstrated by the low transformation efficiency of a dam mutant with fully methylated oriC plasmid DNA compared to the efficiency obtained with wild-type cells (70). Unmethylated oriC plasmid DNA, however, transformed both the wild-type and dam strains at high frequency. The low transformation efficiency of dam bacteria with methylated oriC DNA can thus be explained by the inability of hemimethylated DNA, formed after the first round of semiconservative replication, to initiate chromosome replication (70). Efficient initiation of hemimethylated origins can occur in vitro, however, suggesting that the inhibition observed in vivo could be explained by the binding of specific proteins to the hemimethylated oriC (12, 33).
A protein fraction from the outer membrane was described with the desired properties; it bound specifically to oriC only when it was hemimethylated but not when fully methylated or unmethylated (55). Additional evidence for a trans-acting factor was obtained by the isolation of a mutant P1 oriR plasmid which, unlike the parent plasmid, could be established in a dam mutant host strain (1) (the requirements for oriC and the phage P1 oriR initiation are very similar). The mutant plasmid has a much higher copy number than the parent, suggesting that it titrates out an inhibitor of initiation at hemimethylated origins. Although the mutant plasmid can be established in dam mutants, it replicates poorly. A similar reduction in replication of unmethylated compared to methylated or hemimethylated P1 oriR DNA in an in vitro system was noted (1). These results suggest that initiation at the ori region is controlled by two distinct steps: a negative control at the level of hemimethylation and a positive control requiring methylated DNA for efficient function (1).
The dnaA gene is located close to oriC, and GATC sites within both of these regions remain hemimethylated for 30 to 40% of the cell cycle (16). It is reasonable to assume that the outer membrane protein fraction which binds to hemimethylated oriC also binds to the dnaA region to prevent its immediate remethylation. In a hemimethylated form, the dnaA promoter is less active and dnaA gene transcription is less than 10% of the methylated state (54). Since initiation at oriC requires a critical concentration of DnaA, the effect of hemimethylation of the promoter region would delay the process.
Thus Dam methylation can modulate the initiation process at two distinct points to ensure that initiation at oriC occurs only once per generation. Immediately after initiation, both oriC and the dnaA gene remain hemimethylated, keeping the origin inert and reducing expression of the dnaA gene. Later, both the origin and the dnaA promoter become fully methylated and initiation occurs in the next cell cycle after the DnaA protein reaches the required concentration. It follows that overproduction of Dam in wild-type cells should interfere with the initiation process; indeed, the time that oriC and dnaA are hemimethylated is reduced (16).
Dam methylation is part of a mechanism which ensures that all chromosomal origins initiate at the same time in the cell cycle. Consider a cell with two chromosomal origins; synchronous initiation at each of these will produce four origins, and repeating the process will produce eight origins. Thus the number of origins in the cell is 2n if initiation is synchronous. On the other hand, asynchronous initiation in a cell with two chromosomal origins will produce three origins and then four after the next initiation, and so on. Thus the number of origins in this case is a simple arithmetic progression.
Flow cytometry measurements indicated that the number of chromosomal origins in the cell is dependent on the level of Dam (13). At either low or high levels of Dam the distribution of origins was an integral number between 2 and 10, i.e., asynchronous. At intermediate levels of Dam, the distribution of origins was 2n. These data show that the amount of Dam, and hence hemimethylated oriC DNA, is critical for synchronous initiation. The data are consistent with the model that newly synthesized hemimethylated origins are inert for reinitiation.
Further evidence for the above conclusion is shown in Fig. 5. Bacterial cells with different levels of dam gene transcription were analyzed by flow cytometry (40). Cells with normal levels of dam gene transcription show synchronous initiation, but as the level decreases the cells become asynchronous.
Since chromosome initiation at oriC appears to occurs normally in dam mutants, the initiation mechanism can utilize fully methylated or unmethylated DNA. One can ask what happens when dam mutants arise; how can the initiation machinery pass the block imposed by the hemimethylated state? Presumably, hemimethylated DNA can also be used for initiation but at a much reduced efficiency than unmethylated or fully methylated DNA.
The results of Ogden et al. (55) were originally interpreted as a possible mechanism for chromosome segregation. The finding that chromosome segregation in dam cells is normal (75) suggests that the Ogden et al. (55) results are better interpreted on the basis of chromosome initiation.
Further details about the initiation of chromosome replication can be found in chapter 99.
Since the state of GATC methylation can affect protein-DNA binding, then the presence of this tetranucleotide in promoter or regulatory sequences could affect gene expression. The P2 promoter region of the dnaA gene discussed above is active only in the fully methylated state, consistent with its biological role. In contrast, there is evidence that specific protein binding results at about 20 unmethylated GATCs in the E. coli chromosome (76). Four of these are in the cyclic AMP binding protein (CAP)-binding sites preceding the mtl, cdd, flh, and srl operons (76), suggesting modulation of gene expression by Dam methylation through differential CAP binding.
In addition to the four operons listed above, three others were found to have an unmethylated GATC: car, psp, and fep (76). The level of methylation of the carbamoyl phosphate synthase (car) gene GATC was dependent upon cultural conditions; more methylation was detected when arginine and pyrimidines were present than in their absence, suggesting a possible regulatory effect. For all seven operons containing unmethylated GATCs it would be interesting to determine whether the level of expression in the operons listed above changes in cells overproducing Dam. If so, it would suggest that methylation of the sites is important for regulating operon expression.
In contrast to the seven operons listed above, there is good evidence that the unmethylated GATCs in the pap operon are involved in controlling gene expression (74). Pyelonephritis-associated pilus (Pap) expression is regulated by a phase variation mechanism in which individual cells either express pili (phase-on) or not (phase-off). When Pap pilus gene expression is in the phase-off state, GATC1028 is fully methylated and GATC1130 is unmethylated. Conversely, in the phase-on state, the methylation state at these two sites is reversed. In a strain overproducing Dam, the transition of phase-off to phase-on is prevented, whereas in a dam mutant the opposite transition did not occur. The mechanism of phase variation is that Dam competes with the transcriptional activators Lrp and PapI such that Lrp is required for protection of GATC1130 and both Lrp and PapI are required for methylation protection of GATC1028 (74). Other pilus systems also appear to be Dam controlled, although the evidence is not as complete as for pap (74).
There is also evidence that hemimethylated GATC sites are important to control gene expression. The transposition frequency of Tn10 is directly related to the cellular concentration of Dam acting at two specific GATC sites in IS10 right (68). Overproduction of Dam decreases transposition, whereas it is increased in a dam mutant. One of the GATC sites overlaps the –10 region of the transposase promoter, while the other is near the inner end of IS10 in the target area for transposase action. In DNA that is not being replicated these sites are methylated and inert for transposition. Upon replication, these sites become hemimethylated and are activated for transposition. The transposase (tsp) promoter, in a wild-type strain, is active only in the configuration of methylated transposase coding strand and unmethylated noncoding strand.
The coupling of transposase activation and action to hemimethylation means that transposition is repressed for most of the cell cycle but induced when the element is replicated. The asymmetry imposed at the replication fork means that only one of the two copies of the element can transpose. Hence one copy can remain in place while the other finds an alternative location. The coupling to replication helps to prevent the potentially deleterious effects of excessive transposition (68). Other transposons such as Tn5 and Tn903 and the insertion element IS3 also use Dam methylation to control transposition.
The expression of the mom gene of bacteriophage Mu was found to be influenced by Dam methylation (26). The mom gene encodes an unusual modification function whose biological role is unknown and which is expressed late in the phage life cycle. Dam methylation prevents binding of the E. coli OxyR regulatory protein to a 43-bp region upstream of the phage Mu mom gene which contains three GATCs (11). Although it was demonstrated that OxyR binding in vitro occurred on unmethylated (but not methylated) DNA substrates, it is probable that binding occurs at hemimethylated sites in vivo. Once bound, the OxyR repressor prevents Dam from methylating the three critical sites and prevents transcription initiation, perhaps by interfering with the action of the trans-activating C protein on RNA polymerase.
Several E. coli promoters have GATC sites in either the –10 or –35 regions. These include promoter regions for the sulA, trpS, trpR, tyrR, and glnS genes, and expression of these genes is increased in dam mutants compared to wild type (reviewed in references 6, 43, 54, and 61). It is not known whether expression is increased in a hemimethylated configuration, but even if it were, the physiological role for a coupling to replication is not obvious.
In Table 2, the properties of a dam mutant which can be explained by methylation-dependent gene expression include increased transposition by transposons and altered expression of chromosomal and plasmid genes.
Phage P1 encodes a dam gene which produces a Dam protein that is related to that of the E. coli host. P1 dam mutant phage growth is normal on a wild-type host but severely restricted on a dam mutant host (73). The requirement for Dam by phage P1 is related to the mechanism of DNA encapsidation. Packaging begins at a fixed site (pac) and proceeds by a headful mechanism such that greater-than-genome-length units are packaged. This results in terminal redundancy (same DNA sequence at both ends), which allows the linear genome to circularize by recombination upon infection of the next host.
The pac cleavage site is flanked by seven GATCs in a region of 162 bp. Both in vitro and in vivo experiments indicate that cleavage of both strands occurs only on fully Dam methylated pac regions (73). Since the phage requires concatemeric DNA for packaging and because the phage proteins necessary for pac cleavage are synthesized early in the life cycle, there must be a regulatory mechanism to prevent methylation prior to the initiation of packaging. Furthermore, it is necessary that only one or two pac sites in the concatemer are cleaved and that the remainder are protected from cleavage to ensure proper encapsidation. Presumably this occurs by the competitive binding to GATCs of some phage or host protein. Since the phage enzyme(s) responsible for cleavage ("pacase") can bind to hemimethylated pac sites but cannot cleave, it may be that pacase itself prevents methylation (73).
The most frequent use of methylation-deficient strains is for the propagation of DNA molecules lacking either or both dam and dcm methylated sites (56). This allows for the digestion by various endonucleases which have recognition sites overlapping methylation sequences. A list of such endonucleases can be found in most catalogs of restriction enzyme suppliers. A list of dam and dcm strains containing various genetic markers has been published (56). Such strains are available from various commercial sources or from the author.
An alternative use for Dam is the generation of specific cleavage sites. For example, the action of Dam on a specific DNA molecule will prevent the action of ClaI at overlapping dam sites (ATCGATC) but not at nonoverlapping sites (ATCGATG/A/T).
DNA prepared in a dcm mutant avoids loss of bands corresponding to 5-meCyt by the Maxam-Gilbert chemical sequencing procedure.
Shuttle-vector plasmid DNAs isolated from wild-type E. coli transform Streptomyces lividans at a very low frequency. Transformation efficiencies are increased 400- to 10,000-fold when the vector DNAs are prepared from a dam dcm mutant (42). Similar results have been obtained with various Bacillus and Paracoccus species.
dam-directed mismatch repair can reduce the yield of mutants in site-directed mutagenesis by removal of the desired mutated base. This can be avoided by preparing the template strand in a dam mutant, followed by annealing of the mutagenic primer (80) and transformation of the appropriate dam + strain. Alternatively, the DNA can be methylated by Dam in vitro before transforming cells. DNA fragments produced by PCR should be unmethylated, and after transformation into a wild-type strain the mutant yield should not be affected (63).
Many mutagenic and carcinogenic agents induce the SOS system of E. coli. Some mutagenic agents (e.g., 2-aminopurine) are too weak to elicit an SOS response in normal E. coli cells but do so in dam bacteria (18, 64).
Expression of the cloned E. coli dam gene in organisms that do not have Dam methylation can be used as a probe for chromosome structure and function provided methylation is not lethal. The regions which are methylated can be identified by their susceptibility to restriction endonucleases which cleave methylated GATCs. For example, this technique has been used in yeasts, where expressed genes tend to be Dam methylated to a greater extent than repressed genes in cells containing the dam plasmid (71).
DNA methylation sites can exist in three possible states: unmethylated, hemimethylated, and methylated. This allows the cell a mechanism of control through the use of proteins which act specifically in one of these states. The use of two or more DNA methylation sites to control a single biological parameter could multiply the possible combinations of switches. Although some of the functional dam sites have already been identified, the remaining DNA dam and dcm sites need to be found and characterized. At the protein level, we need to define the protein-DNA contacts for binding to the various methylation states. As this information becomes available, it should serve as a model for the analogous methyltransferases in eukaryotic cells.
Dam has evolved in status from an obscure academic curiosity to a commercially available reagent. Furthermore, its use in probing chromosome structure and function in E. coli and other organisms offers a new way to approach this area. Little is known about the domain structure of Dam; the substrate binding and catalytic domains remain to be identified. In this regard it would be useful to have a solved crystal structure of the protein bound to DNA. We are also ignorant about the mechanism of methyl group transfer from SAM to the adenine base at the chemical level.
The various roles of Dam in cellular physiology will be explored further. The hypothetical protein(s) binding to the hemimethylated oriC region will have to be isolated and characterized. The relationship between phage P1 head morphogenesis, dam methylation, and protein binding should prove to be fascinating. The mode of dam gene regulation and how the level of Dam in the cell is controlled are still largely unknown. More genes regulated by Dam methylation will certainly be found, and a dam regulon may exist. In this context the finding (32) that the Ec67 retron encodes both a reverse transcriptase and a dam gene which is 44% identical at the amino acid sequence level to E. coli Dam is tantalizing. That Dam may play some regulatory role is suggested by the finding that there are three GATCs in the promoter region for the reverse transcriptase gene.
Similarly, a crystal structure for Dcm and its relative M.EcoRII would be useful to investigate further the functional domains of the protein. The mechanism of dcm gene expression and its relationship to vsr will be explored. For both Dam and Dcm, variants with altered sequence recognition patterns would be useful additional tools.
I thank A. Bhagwat, R. M. Blumenthal, E. Boye, U. von Freiesleben, S. Hattman, A. Løbner-Olesen, B. R. Palmer, K. R. Peterson, L. J. Rasmussen, and T. Wu for their comments and suggestions to improve this manuscript. I also thank Lene Juel Rasmussen, who provided Fig. 3.
References
1. Abeles, A., T. Brendler, and S. Austen. 1993. Evidence for two levels of control of P1 oriR and host oriC replication origins by DNA adenine methylation. J. Bacteriol. 175:7801–7807.
2. Bale, A., M. d’Alarcao, and M. G. Marinus. 1979. Characterization of DNA adenine methylation mutants of Escherichia coli K-12. Mutat. Res. 59:157–165.
3. Banas, J. A., J. J. Ferretti, and A. Progulske-Fox. 1991. Identification and sequence analysis of a methylase gene in Porphyromonas gingivalis. Nucleic Acids Res. 19:4189–4192.
4. Barbeyron, T., K. Kean, and P. Forterre. 1984. DNA adenine methylation of GATC sequences appeared recently in the Escherichia coli lineage. J. Bacteriol. 160:586–590.
5. Barras, F., and M. G. Marinus. 1988. Arrangement of Dam methylation sites (GATC) in the Escherichia coli chromosome. Nucleic Acids Res. 16:9821–9838.
6. Barras, F., and M.G. Marinus. 1989. The great GATC. Trends Genet. 5:139–143.
7. Bergerat, A., W. Guschlbauer, and G. V. Fazakerley. 1991. Allosteric and catalytic binding of S-adenosylmethionine to Escherichia coli DNA adenine methyltransferase monitored by 3H NMR. Proc. Natl. Acad. Sci. USA 88:6394–6397.
8. Bergerat, A., A. Kriebardis, and W. Guschlbauer. 1989. Preferential site-specific hemimethylation of GATC sites in pBR322 DNA by Dam methyltransferase from Escherichia coli. J. Biol. Chem. 264:4064–4070.
9. Bhagwat, A. S., and M. McClelland. 1992. DNA mismatch correction by very short patch repair may have altered the abundance of oligonucleotides in the E. coli genome. Nucleic Acids Res. 20:1663–1668.
10. Bhagwat, A. S., A. Sohail, and R. J. Roberts. 1986. Cloning and characterization of the dcm locus of Escherichia coli K-12. J. Bacteriol. 166:751–755.
11. Bölker, M., and R. Kahmann. 1989. The Escherichia coli regulatory protein OxyR discriminates between methylated and unmethylated states of the phage Mu mom promoter. EMBO J. 8:2403–2410.
12. Boye, E. 1991. The hemimethylated replication origin of Escherichia coli can be initiated in vitro. J. Bacteriol. 173:4537–4539.
13. Boye, E., and A. Løbner-Olesen. 1990. The role of dam methyltransferase in the control of DNA replication in E. coli. Cell 62:981–989.
14. Boye, E., M. G. Marinus, and A. Løner-Olesen. 1992. Quantitation of Dam methyltransferase in Escherichia coli. J. Bacteriol. 174:1682–1685.
15. Brooks, J. E., R. M. Blumenthal, and T. R. Gingeras. 1983. The isolation and characterization of the Escherichia coli DNA adenine methylase dam gene. Nucleic Acids Res. 11:837–851.
16. Campbell, J. L., and N. Kleckner. 1990. E. coli oriC and the dnaA gene promoter are sequestered from dam methyltransferase following passage of the chromosomal replication fork. Cell 62:967–979.
17. Coulondre, C., J. H. Miller, P. J. Farabaugh, and W. Gilbert. 1978. Molecular basis of substitution hotspots in Escherichia coli. Nature (London) 274:775–780.
18. Craig, R. J., J. A. Arraj, and M. G. Marinus. 1984. Induction of damage-inducible (SOS) repair in dam mutants of Escherichia coli exposed to 2-aminopurine. Mol. Gen. Genet. 194:539–540.
19. Dar, M. E., and A. S. Bhagwat. 1993. Mechanism of expression of DNA repair gene vsr, an Escherichia coli gene that overlaps the DNA cytosine methylase gene, dcm. Mol. Microbiol. 9:823–833.
20. Efimova, E. P., E. P. Delver, and A. A. Belogurov. 1988. Alleviation of type I restriction in adenine methylase (dam) mutants of Escherichia coli. Mol. Gen. Genet. 214:313–316.
21. Gammie, A. E., and J. H. Crosa. 1991. Roles of DNA adenine methylation in controlling replication of the REPI replicon of plasmid pColV-K30. Mol. Microbiol. 5:495–503.
22. Glickman, B. W., and M. Radman. 1980. Escherichia coli mutator mutants deficient in methylation-instructed DNA mismatch correction. Proc. Natl. Acad. Sci. USA 77:1063–1067.
23. Gomez-Eichelmann, M. C., and J. Ramirez-Santos. 1993. Methylated cytosine at Dcm (CCA/TGG) sites in Escherichia coli: possible function and evolutionary implications. Mol. Evol. 37:11–24.
24. Hanck, T., S. Schmidt, and H. J. Fritz. 1993. Sequence-specific and mechanism-based crosslinking of Dcm DNA cytosine-C5 methyltransferase of E. coli K-12 to synthetic oligonucleotides containing 5-fluoro-2'-deoxycytidine. Nucleic Acids Res. 21:303–309.
25. Hardy, L. W., J. S. Finer-Moore, W. R. Montfort, M. O. Jones, D. V. Santi, and R. M. Stroud. 1987. Atomic structure of thymidylate synthase: target for rational drug design. Science 235:448–455.
26. Hattman, S. 1982. DNA methyltransferase-dependent transcription of the phage Mu mom gene. Proc. Natl. Acad. Sci. USA 79:5518–5521.
27. Hattman, S., S. Schlagman, and L. Cousens. 1973. Isolation of a mutant of Escherichia coli defective in cytosine-specific deoxyribonucleic acid methylase activity and partial protection of bacteriophage lambda against restriction by cells containing the N-3 drug-resistance factor. J. Bacteriol. 115:1103–1107.
28. Hattman, S., J. Wilkinson, D. Swinton, S. Schlagman, P. M. MacDonald, and G. Mosig. 1985. Common evolutionary origin of the phage T4 dam and host Escherichia coli dam DNA-adenine methyltransferase genes. J. Bacteriol. 164:932–937.
29. Hennecke, F., H. Kolmar, K. Brundl, and H. J. Fritz. 1991. The vsr gene product of E. coli K-12 is a strand and sequence-specific DNA mismatch endonuclease. Nature (London) 353:776–778.
30. Herman, G. E., and P. Modrich. 1981. Escherichia coli clones that overproduce dam methylase are hypermutable. J. Bacteriol. 145:644–646.
31. Herman, G. E., and P. Modrich. 1982. Escherichia coli dam methylase. Physical and catalytic properties of the homogeneous enzyme. J. Biol. Chem. 257:2605–2612.
32. Hsu, M.-Y., M. Inouye, and S. Inouye. 1990. Retron for the 67-base multicopy single-stranded DNA from Escherichia coli: a potential transposable element encoding both reverse transcriptase and Dam methylase function. Proc. Natl. Acad. Sci. USA 87:9454–9458.
33. Hughes, P., F. Squali-Houssani, P. Forterre, and M. Kohiyama. 1988. In vitro replication of a dam methylated and non-methylated oriC plasmid. J. Mol. Biol. 176:155–159.
34. Jonczyk, P., R. Hines, and D. W. Smith. 1989. The Escherichia coli dam gene is expressed as a distal gene of a new operon. Mol. Gen. Genet. 217:85–96.
35. Karran, P., and M. G. Marinus. 1992. Mismatch correction at O6-methylguanine residues in E. coli DNA. Nature (London) 296:868–869.
36. Klimasauskas, S., S. Kumar, R. J. Roberts, and X. Cheng. 1994. HhaI methyltransferase flips its target base out of the DNA helix. Cell 76:357–369.
37. Kossykh, V. G., S. L. Schlagman, and S. Hattman. 1993. Conserved sequence motif DPPY in region IV of the phage T4 Dam DNA-[N6-adenine]-methyltransferase is important for S-adenosyl-l-methionine binding. Nucleic Acids Res. 21:4659–4662.
38. Lauster, R. 1989. Evolution of type II DNA methyltransferases. J. Mol. Biol. 206:313–321.
39. Lieb, M. 1983. Specific mismatch correction in bacteriophage lambda crosses by very short patch repair. Mol. Gen. Genet. 191:118–125.
40. Løbner-Olesen, A., E. Boye, and M. G. Marinus. 1992. Expression of the Escherichia coli dam gene. Mol. Microbiol. 6:1841–1851.
41. Lundblad, V., and N. Kleckner. 1984. Mismatch repair mutations of Escherichia coli K-12 enhance transposition excision. Genetics 109:3–19.
42. MacNeil, D. J. 1988. Characterization of a unique methyl-specific restriction system in Streptomyces avermitilis. J. Bacteriol. 170:5607–5612.
43. Marinus, M. G. 1987. DNA methylation in Escherichia coli. Annu. Rev. Genet. 21:113–131.
44. Marinus, M. G., and E. B. Konrad. 1976. Hyper-recombination in dam mutants of Escherichia coli K-12. Mol. Gen. Genet. 149:273–277.
45. Marinus, M. G., and N. R. Morris. 1973. Isolation of deoxyribonucleic acid methylase mutants of Escherichia coli K-12. J. Bacteriol. 114:1143–1150.
46. Marinus, M. G., and N. R. Morris. 1974. Biological function for 6-methyladenine residues in the DNA of Escherichia coli K-12. J. Mol. Biol. 85:309–322.
47. Marinus, M. G., A. Poteete, and J. A. Arraj. 1984. Correlation of DNA adenine methylase activity with spontaneous mutability in Escherichia coli K-12. Gene 28:123–125.
48. McGraw, B. R., and M. G. Marinus. 1980. Isolation and characterization of Dam+ revertants and suppressor mutations that modify secondary phenotypes of dam-3 strains of Escherichia coli K-12. Mol. Gen. Genet. 178:309–315.
49. Merkl, R., M. Kroger, P. Rice, and H. J. Fritz. 1992. Statistical evaluation and biological interpretation of non-random abundance in the E. coli K-12 genome of tetra- and pentanucleotide sequences related to VSP mismatch repair. Nucleic Acids Res. 20:1657–1662.
50. Messer, W., U. Bellekes, and H. Lother. 1985. Effect of dam methylation on the activity of the E. coli replication origin, oriC. EMBO J. 4:1327–1332.
51. Messer, W., and M. Noyer-Weidner. 1988. Timing and targeting: the biological functions of dam methylation in E. coli. Cell 54:735–737.
52. Miner, Z., S. L. Schlagman, and S. Hattman. 1989. Single amino acid changes that alter the sequence specificity of the DNA-[N6-adenine] methyltransferase (dam) of bacteriophage T4. Nucleic Acids Res. 17:8149–8157.
53. Modrich, P. 1991. Mechanisms and biological effects of mismatch repair. Annu. Rev. Genet. 25:229–253.
54. Noyer-Weidner, M., and T. A. Trautner. 1993. Methylation of DNA in prokaryotes, p. 39–108. In J. P. Jost and H. P. Saluz (ed.), DNA Methylation: Molecular Biology and Biological Significance. Birkhauser Verlag, Basel, Switzerland.
55. Ogden, G. B., M. J. Pratt, and M. Schaechter. 1988. The replicative origin of the E. coli chromosome binds to cell membranes only when hemimethylated. Cell 54:127–135.
56. Palmer, B. R., and M. G. Marinus. 1995. The dam and dcm strains of Escherichia coli: a review. Gene 143:1–12.
57. Parker, B., and M. G. Marinus. 1988. A simple and rapid method to obtain substitution mutations in Escherichia coli: isolation of a dam deletion/insertion mutation. Gene 73:531–535.
58. Parker, B., and M. G. Marinus. 1992. Repair of DNA heteroduplexes containing small heterologous sequences in Escherichia coli. Proc. Natl. Acad. Sci. USA 89:1730–1734.
59. Peterson, K. R., and D. W. Mount. 1993. Analysis of the genetic requirements for viability of Escherichia coli K-12 DNA adenine methylase (dam) mutants. J. Bacteriol. 175:7505–7508.
60. Peterson, K. R., K. F. Wertman, D. W. Mount, and M. G. Marinus. 1985. Viability of Escherichia coli K-12 DNA adenine methylase (dam) mutants requires increased expression of specific genes in the SOS regulon. Mol. Gen. Genet. 201:14–19.
61. Plumbridge, J. 1987. The role of dam methylation in controlling gene expression. Biochimie 69:439–443.
62. Posfai, J., A. S. Bhagwat, G. Posfai, and R. J. Roberts. 1989. Predictive motifs derived from cytosine methyltransferases. Nucleic Acids Res. 17:2421–2435.
63. Pukkila, P., J. Peterson, G. Herman, P. Modrich, and M. Meselson. 1983. Effects of high levels of DNA adenine methylation on methyl directed mismatch repair in E. coli. Genetics 104:571–582.
64. Quillardet, P., and M. Hofnung. 1987. Induction of the SOS system in a dam-3 mutant: a diagnostic strain for chemicals causing DNA mismatches. Mutat. Res. 177:17–26.
65. Rasmussen, L. J., M. G. Marinus, and A. Løbner-Olesen. 1995. Novel growth rate control of dam gene expression in E. coli. Mol. Microbiol. 12:631–638.
66. Richie, L., D. M. Podger, and R. M. Hall. 1988. A mutation in the DNA adenine methylase gene (dam) of Salmonella typhimurium decreases susceptibility to 9-aminoacridine-induced frameshift mutagenesis. Mutat. Res. 194:131–141.
67. Ringquist, S., and C. Smith. 1992. The Escherichia coli chromosome contains specific unmethylated dam and dcm sites. Proc. Natl. Acad. Sci. USA 89:4539–4543.
68. Roberts, D., B. C. Hoopes, W. McClure, and N. Kleckner. 1985. IS10 transposition is regulated by DNA adenine methylation. Cell 43:117–130.
69. Russel, D. W., and R. K. Hirata. 1989. The detection of extremely rare DNA modifications. J. Biol. Chem. 264:10787–10794.
70. Russel, D. W., and N. D. Zinder. 1987. Hemimethylation prevents DNA replication in E. coli. Cell 50:1071–1079.
71. Singh, J., and A. J. S. Klar. 1992. Active genes in budding yeast diplay enhanced in vivo accessibility to foreign DNA methylases: a novel in vivo probe for chromatin structure of yeast. Genes Dev. 6:186–196.
72. Smith, D. W., A. M. Garland, G. Herman, R. E. Enns, T. A. Baker, and J. W. Zyskind. 1985. Importance of state of oriC GATC sites in initiation of DNA replication in Escherichia coli. EMBO J. 4:1319–1327.
73. Sternberg, N., and J. Coulby. 1990. Cleavage of the bacteriophage P1 packaging site (pac) is regulated by adenine methylation. Proc. Natl. Acad. Sci. USA 87:8070–8074.
74. Van der Woude, M. W., B. A. Braaten, and D. A. Low. 1993. Evidence for global regulatory control of pilus expression in Escherichia coli by Lrp and DNA methylation: model building based on analysis of pap. Mol. Microbiol. 6:2429–2435.
75. Vinella, D., A. Jaffe, R. D’Ari, M. Kohiyama, and P. Hughes. 1992. Chromosome partitioning in Escherichia coli in the absence of Dam-directed methylation. J. Bacteriol. 174:2388–2390.
76. Wang, M. X., and G. M. Church. 1992. A whole genome approach to in vivo DNA-protein interactions in E. coli. Nature (London) 360:606–610.
77. Wang, T.-C., and K. C. Smith. 1986. Inviability of dam recA and dam recB cells of Escherichia coli is correlated with their inability to repair DNA double-strand breaks produced by mismatch repair. J. Bacteriol. 165:1023–1025.
78. Wyszynski, M. W., S. Gabbara, E. A. Kubareva, E. A. Romanova, T. S. Oretskaya, E. S. Gromova, Z. A. Shabarova, and A. S. Bhagwat. 1993. The cysteine conserved among DNA cytosine methylases is required for methyl transfer but not specific binding. Nucleic Acids Res. 21:295–301.
79. Yallaly, P., and A. Eisenstark. 1990. Influence of DNA adenine methylase on the sensitivity of Escherichia coli to near-ultra-violet radiation and hydrogen peroxide. Biochem. Biophys. Res. Commun. 169:64–69.
80. Zoller, M., and M. Smith. 1982. Oligonucleotide-directed mutagenesis using M13-derived vectors: an efficient and general procedure for the production of point mutations in any fragment of DNA. Nucleic Acids Res. 10:6487–6500.
Chapter53
