Promega Notes Magazine Number 42, June 1993, p.

22

A practical guide to DNA methylation

Tom Schoenfeld
Promega Corporation

The pattern and extent of DNA methylation can significantly affect the success of restriction digestions
or bacterial transformations. Prokaryotic DNA may be methylated by host restriction/modification
systems, while eukaryotic DNA often is methylated in CG or CNG di- or trinucleotides. This article
presents background information on the effects of methylation on common cloning applications as well
as troubleshooting tips for situations where DNA methylation interferes with the handling of DNA.

Introduction
Most problems attributable to DNA methylation fall into one of two categories:

inability to cut DNA with restriction enzymes

inability to obtain transformants

This article contains background information that explains why these problems occur in cloning,
followed by troubleshooting tips that will help you determine if methylation is the problem and, if so,
how to solve it. A final section describes common eukaryotic methylation patterns and how to use this
information to your advantage.

DNA methylation refers to the covalent modification of DNA by transfer of a methyl group from S-
adenosylmethionine to one of a few possible sites on cytosine or adenine. The molecular biologist will
primarily encounter m5C methylation of cytosine and m6N methylation of adenine (Figure 1). The
guidelines in this article, however, also apply to other types of methylation such as m4C methylation of
cytosine and m4C hydroxymethylation of cytosine (1).

Figure 1. Structure of m5C-methylated cytidine and m6N-methylated adenosine.

Prokaryotic methylation
Both adenine and cytosine methylation occur in prokaryotes at 0.5-5% of the available sites. In
prokaryotes, DNA methylation is involved in restriction/modification systems and in DNA mismatch
repair systems.

Restriction/modification systems

Restriction enzymes were first identified based on their ability to limit or "restrict" the range of infection
of bacterial hosts by a given bacteriophage. Restriction enzymes also may play a role in preventing
genetic exchange between groups of bacteria. All restriction enzyme genes studied thus far are
associated with closely-linked methylase genes (2). The methylases limit restriction digestion to foreign
DNA, whether it is of phage or other origin. The methylase modifies one of the bases in the recognition
sequence of its cognate restriction enzyme, thereby preventing digestion at this site and protecting the
host DNA (3).

Besides the methylation specific to each restriction/modification system, the presence of other
methylated bases within a restriction site can block or significantly reduce cleavage in about half of the
possible cases (1). This inhibition is probably due to steric hindrance and can vary for different
restriction enzymes. For example, the isoschizomer pair of Msp I and Hpa II both recognize the
sequence CCGG. Msp I is sensitive to methylation of the external cytosine only (m5CCGG), but Hpa II
is sensitive to methylation of both the internal and external cytosine (Cm5CGG or m5CCGG).

Isoschizomer pairs such as Msp I and Hpa II that differ in their methylation sensitivity are very useful
tools. By comparing the digestion patterns using such an isoschizomer pair, one can determine the extent
to which a DNA sample is methylated. This can be used to measure percent methylation of a genomic
DNA fragment or, following a transfection, to reveal the ratio of transfected to host DNA. A table of
isoschizomer pairs that differ in methylation sensitivity can be found in reference 5.

The sensitivity of a restriction enzyme to a specific methylation pattern may differ if the site is
hemimethylated (methylated on only one strand). This information is commonly listed in methylation
sensitivity tables such as those in reference 1 and the Promega Catalog.

Restriction and methylation systems in common lab strains of E. coli

The methylation genetics of the host used to propagate a cloned insert will determine that DNA's
methylation pattern. Since most of the DNA handled in a typical lab is prepared in E. coli, its
methylation genetics are described here in detail (Table 1). Many common lab strains of E. coli are
deficient in one or more restriction/modification system. The genetics of specific strains are described in
the Promega Catalog.

dam and dcm These are the most commonly encountered E. coli methylations. The dam gene product
methylates the N6 position (m6N) of adenine in the sequence GATC while the dcm gene product
methylates the C5 position (m5C) of the internal cytosine in the sequences CCAGG or CCTGG (5). Most
common laboratory strains of E. coli contain both of these activities. In laboratory strains that are recA-
to prevent recombination of transforming DNA, the dam gene will always be present. This is because
the combination of recA- and dam- results in a lethal phenotype. The same is not true for the dcm gene,
which is compatible with a recA- genotype (6).

EcoK and EcoB The genes encoding the EcoK and EcoB restriction/modification systems occupy the
same locus. Therefore a strain has either the EcoK or EcoB system, but not both. Each of these
restriction systems cleaves unmethylated DNA. Most lab strains of E. coli are derived from an EcoK
parent. The EcoK restriction enzyme gene generally is mutant in these strains, resulting in a restriction
minus genotype (rK-). Some strains also are mutant in the EcoK methylase gene (mK-). In those strains
in which the methylase is active, however (mK+), m6N adenine methylation occurs at Am6ACGTGC and
GCm6ACGTT. Since these sites are rare, this methylation generally does not interfere with restriction
digests.

McrA, McrBC and Mrr These restriction systems require methylation of the target DNA for cleavage.
McrA cleaves m5CG-methylated DNA, McrBC cleaves (A/G)m5C-methylated DNA and Mrr cleaves
m6N adenine-methylated DNA. These three systems are of little concern when subcloning DNA from E.
coli, since none of them cleaves Dcm or Dam-modified DNA. Because McrA and McrBC may cleave
some sites containing m5C in the CG dinucleotide, however, avoid these strains when cloning eukaryotic
genomic DNA.

Table 1. Guide to the Restriction/Modification Genetics of E. coli.

Gene Wild Type Phenotype Comments

dam Adenine is m6N-methylated in the DNA prepared from dam+ cells will
sequence Gm6ATC. contain Gm6ATC and will not be
restricted in vitro by enzymes blocked
by this methylation pattern.
dcm Cytosine is m5C-methylated in the DNA prepared from dcm+ cells will
sequence Cm5C(A/T)GG . contain Cm5C(A/T)GG and will not
be restricted in vitro by enzymes
blocked by this methylation pattern.
ecoK/ecoB Produces the EcoK or EcoB DNA that is not methylated at the
restriction enzyme and corresponding appropriate site before introduction
methylase. Adenine is methylated in into ecoK rK+ or ecoB rB+ cells will
the sequences Am6ACGTGC and be cleaved. Restriction minus (rB/K-)
m6
GC ACGTT. strains may or may not produce the
methylase (mB/K+ or mB/K-). If not,
DNA propagated in these cells will
not be protected when transformed
into ecoK rK+ cells.
mcrA Produces the McrA restriction mcrA+ cells will digest m5CG-
enzyme. methylated DNA. Avoid these strains
when cloning mammalian or plant
genomic DNA.
mcrBC Produces the McrBC restriction mcrBC+ cells will digest (A/G)m5C-
enzyme. methylated DNA. Since these sites
may overlap CG or CNG methylation
sites, avoid these stains when cloning
DNA from cells with this methylation
pattern.
mrrA Produces the MrrA restriction mrrA+ cells will digest m6A-
enzyme. methylated DNA.
Troubleshooting methylation problems
Restriction digestion of prokaryotic DNA

Our Technical Services representatives sometimes are asked "Which restriction enzymes are sensitive to
methylation?" Strictly speaking, every restriction enzyme is sensitive to methylation on at least one base
in its recognition site (by its cognate methylase). With this in mind, however, a better question would be
"Is the restriction enzyme I would like to use sensitive to the methylation pattern in my DNA?"

Figure 2 shows how to determine if methylation is preventing cleavage of your target DNA and, if so,
how to overcome this problem. Table 2 lists the methylation sensitivities of selected commercially
available restriction enzymes.

Methylation typically interferes with DNA cleavage in one of two situations: cutting DNA cloned and
propagated in dam+ and dcm+ E. coli strains (discussed below) and cutting eukaryotic genomic DNA
from certain organisms (discussed later in this article). Most commercially available cloning vectors
contain multiple cloning regions designed to avoid common types of methylation.

Figure 2. Troubleshooting methylation-related problems with digestion of cloned

DNA.

Table 2. Sensitivity of Selected Restriction Enzymes to Common Types of Methylation. Use this
table to determine if any of these common types of methylation could interfere with cutting by a given
restriction enzyme. For example, an "s" (sensitive) in the dcm column means that dcm methylation can
inhibit cleavage by the restriction enzyme in question.

Enzyme Site dam dcm CG CNG

Aat II GACGTC i i s i
Acc III TCCGGA s(ol) i i i
Apa I GGGCCC i s(ol) s(ol) i
Ava I C(T/C)CG(A/G)G i i s i
Ava II GG(A/T)CC i s(ol) s(ol) s(ol)
Bal I TGGCCA i s(ol) i s(ol)
BamH I GGATCC i i i s(ol)
Bbu I GCATGC i i i i
Bgl II AGATCT i i i s(ol)
BsaO I CG(A/G)(T/C)CG i i n/a n/a
BstO I CC(A/T)GG i i i n/a
BstZ I CGGCCG i i s(ol) s(ol)
Cla I ATCGAT s(ol) i s i
Csp I CGG(A/T)CCG i i s(ol) s
Csp45 I TTCGAA i i s i
Eco47 III AGCGCT i i s i
Eco52 I CGGCCG i i n/a n/a
EcoR I GAATTC i i i i
Hae III GGCC i i i s(ol)
Hinc II GT(T/C)(A/G)AC i i i i
Hind III AAGCTT i i i i
Hpa II CCGG i i s s
Kpn I GGTACC i i i i
Mbo I GATC s i i i
Mbo II GAAGA(8/7) s(ol) i i i
Mlu I ACGCGT i i s i
Msp I CCGG i i i s
Nar I GGCGCC i i s i
Not I GCGGCCGC i i s s
Pst I CTGCAG i i i s
Pvu I CGATCG i i s s(ol)
Pvu II CAGCTG i i i s
Sac I GAGCTC i i i n/a
Sac II CCGCGG i i s s
Sal I GTCGAC i i s n/a
Sau3A I GATC i i s(ol) s(ol)
Sau96 I GGNCC i s(ol) s(ol) s(ol)
Sfi I GGCCNNNNNGGCC i s(ol) s(ol) s(ol)
Sin I GG(A/T)CC i i i s(ol)
Sma I CCCGGG i i s s
SnaB I TACGTA i i s i
Sph I GCATGC i i i i
Spo I TCGCGA i i s i
Stu I AGGCCT i s(ol) i s(ol)
Taq I TCGA s(ol) i i i
Xba I TCTAGA s(ol) i i i
Xho I CTCGAG i i s i
Xho II (A/G)GATC(T/C) i i i s(ol)
Xma I CCCGGG i i i n/a

Information provided in this table is based on reference 1.

Key i = insensitive to this methylation

s = sensitive to this methylation
s(ol) = potentially sensitive where the methylation site partially overlaps the
restriction site
n/a = information not available

dam and dcm methylation and restriction digests

If the recognition site of a restriction enzyme does not contain all or part of the dam recognition
sequence (GATC) or the dcm sequence (CC(A/T)GG), it will be insensitive to dam or dcm methylation.
Generally, dam or dcm methylation will be a problem only if the site contains one of these two
sequences or if it ends in GA, GAT, CC or CC(A/T), and therefore potentially overlaps with a dam or
dcm site.

Enzymes containing one of these sequences vary in their sensitivity to methylation. For instance, both
Mbo I (GATC) and BamH I (GGATCC) restriction sites contain complete dam sites, but Mbo I is
inhibited and BamH I is not inhibited by dam methylation. In rare cases, methylation of a base outside,
but adjacent to, a restriction site may inhibit cleavage. For example, Nar I (GGCGCC) is inhibited by
the presence of m4C methylation adjacent to its recognition site (GGCGCCm4C) (1).

When cleavage by a given restriction enzyme is blocked by dam or dcm methylation, check the available
isoschizomers of that enzyme for one that is insensitive to methylation.

Transformation of bacterial cells

In general, methylation does not lead to transformation problems when using DNA prepared in E. coli or
other common hosts. If you have difficulty obtaining transformants, however, check the genetics of the
source and recipient host strains (see Figure 3). One common problem occurs when CG-methylated
eukaryotic DNA is cloned into an E. coli host with an Mcr restriction system, which cleaves methylated
DNA (5). Remember that a clone takes on the methylation pattern of the host in which it was most
recently propagated. For example, human genomic DNA cloned in bacterial strain JM109 will not be
CG-methylated.
 Figure 3. Troubleshooting methylation-related problems with obtaining
transformants.

Also take care when cloning unmodified DNA into a newly developed bacterial host, since examples
exist of unsuspected restriction/modification systems causing instability of the cloned DNA (7).

Digesting eukaryotic genomic DNA

Eukaryotic methylation patterns

Methylation in eukaryotes is variable. The degree of methylation depends not only on the species, but
also on the cell type and the developmental stage of the cell. In some mammals, methylation normally is
limited to the m5C position of cytosine in the dinucleotide CG, while in plants methylation also occurs at
CNG sequences where N is any base (1). The role of eukaryotic CG methylation is not fully understood.
Up to 30% of the cytosine may be methylated in plants, while normally only 2-5% is methylated in
mammals. In yeast and Drosophila, no methylation is detected. Other lower eukaryotes also have
distinct DNA methylation patterns.

Although most CG dinecleotides in mammalian genomes are highly methylated, clusters of stably
unmethylated CGs exist throughout the genome. Such clusters of unmethylated CGs are referred to as
"CG islands." CG islands have a higher than average G+C content, approximately ten times higher than
the rest of the genome, and almost always occur in the 5´ region of transcribed DNA (8).

Restriction enzymes that tend to cut inside unmethylated CG islands obviously are less likely to be
inhibited by CG methylation (8). Such enzymes typically have restriction sites containing more than six
bases, a high G+C content and more than one CG in the recognition sequence. Not I (GGCGCGCC) is
an example of such an enzyme.

Strategies for generating large fragments of genomic DNA

The CG dinucleotide occurs about five times less often than other dinucleotides in eukaryotic DNA.
Thus, enzymes such as Pst I that contain CG in the recognition site can be used to generate large
fragments due to the infrequency of their cut sites.

The differing sensitivities of restriction enzymes to methylation also can be used to advantage when
digesting eukaryotic genomic DNA. For example, the Asu II/Csp45 I isoschizomers both recognize the
sequence TTCGAA. This site will most commonly occur outside of CG islands and therefore will tend
to be methylated. When the m5C position of cytosine is methylated, Asu II will cleave this site but Csp45
I will not. If very large restriction fragments are needed, Csp45 I is desirable because it will cleave CG-
methylated DNA very infrequently. Alternatively, Asu II should be used if a researcher wishes to cut at
specific sites and wants some assurance that all sites are cut.

Methylation can complicate interpretation of mapping data

Eukaryotic cells are not known to possess restriction/modification systems. However, evidence is
mounting that eukaryotic methylation is involved in gene regulation. CG methylation of eukaryotic
DNA varies with cell type and developmental stage, which can lead to difficulty in interpreting mapping
data. For instance, variations in restriction patterns have been used as evidence of translocations and
deletions. If these variations are, in fact, caused by differences in methylation patterns, the researcher
may be misled into believing that a translocation or a deletion has occurred when one has not (8).

Troubleshooting methylation-related problems with digestion of genomic DNA

The troubleshooting approach for genomic DNA is similar to that for cloned DNA (Figure 3), except
that methylation affects the C in the CG dinucleotide and, in the case of plant DNA, the CNG
trinucleotides. In these cases, the only option for addressing a methylation problem is to use a different
restriction enzyme. For some applications, such as RFLP mapping, use enzymes that have sites lacking
CG dinucleotides or that are insensitive to methylation, such as Hae III, Taq I and Pst I. Where the the
choice of a restriction enzyme site is more constrained, search for methylation-insensitive
isoschizomers. The awareness that a problem may exist can be sufficient to prevent misinterpretation of
data. As mentioned above, it is important to avoid mcrA+ and mcrBC+ strains when cloning eukaryotic
genomic DNA because these may cleave some sites containing m5C in the CG dinucleotide.

Conclusions
An understanding of DNA methylation is useful for troubleshooting cloning and transformation
problems as well as for correct interpretation of experimental results, especially those obtained from
digestions of eukaryotic genomic DNA. In addition, DNA methylation may be used as a way to limit the
extent of cleavage of a target DNA. This may be useful for generating large fragments for megabase
mapping, protecting specific sites or directing digestion to a specific site. DNA methylation can be a
useful tool or a source of frustration, depending on the care taken to consider methylation during the
experimental design.

References
1. Nelson, M. and McClelland, M. (1991) Nucl. Acids Res. 19(sup), 2045.

2. Yuan, R. and Smith, D.L. (1984) In: DNA Methylation: Biochemistry and Biological Significance,
Razin, A., Cedar, H. and Riggs, A.D., eds., Springer-Verlag, New York.

3. Modrich, P. and Roberts, R. (1985) In: Nucleases, Linn, S.M. and Roberts, R.J., eds., Cold Spring
Harbor Laboratory, Cold Spring Harbor, New York, 109.

4. Protocols and Applications Guide (1991) Promega Corporation.

5. Sambrook, J., Fritsch, E.F. and Maniatis, T. (1991) Molecular Cloning: A Laboratory Manual 2nd
 ed., Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York, 5.15-5.26.

6. Bhagwat, A. (1992) Meth. Enzymol. 216, 199.

7. Mermelstein, L.D. et al. (1992) Bio/Technology 10, 190.

8. Bickmore, W. A. and Bird, A. P. (1992) Meth. Enzymol. 216, 224.

© 1993 Promega Corporation. All Rights Reserved.