Vol. 169, No. 3
N4-Methylcytosine as a Minor Base in Bacterial DNA
MELANIE EHRLICH,'* GEOFFREY G. WILSON,2 KENNETH C. KUO,3 AND CHARLES W. GEHRKE3
Department ofBiochemistry, Tulane Medical School, New Orleans, Louisiana 701121; New England BioLabs, Beverly,
Massachusetts 019152; and Department ofBiochemistry, Experiment Station Chemical Laboratories, University of
Missouri, Columbia, Missouri 652013
Received 16 September 1986/Accepted 2 December 1986
The DNA base composition, including the minor base content, of 26 strains of bacteria was determined. The
studied bacteria are sources of widely used restriction endonucleases. Approximately 35% of the bacterial
DNAs contained N4-methylcytosine, about 60% contained 5-methylcytosine, and about 90% had
Both N6-methyladenine (m6A) and 5-methylcytosine
(m5C) have been long known to be minor bases in bacterial
DNA (7, 9, 16). One or both of these methylated bases are
present in most bacterial DNAs examined (4, 11, 33). Re-
cently, a third minor base, N4-methylcytosine (m4C), has
been found in DNA from eight types ofthermophilic bacteria
(11) and in that from one type of mesophilic bacteria (19, 20).
Previously, m4C residues had been detected only in the RNA
ofthe small ribosomal subunit of bacteria and of mammalian
and insect mitochondria (8, 13, 40). This modified base is not
present in a variety of eucaryotic DNAs (13; Gehrke et al.,
The minor base composition of bacterial DNA is partially
determined by restriction-modification systems (30). The
level of minor bases should be consistent with the known
specificities of the host restriction and modification en-
methyltransferases. For example, if the known restriction
enzymes in a bacterium have recognition sites containing
only G. C base pairs and if m6A as well as m5C is found in
this cellular DNA, then one or more other unrelated modi-
fication pathways must be present. These could be involved
in previously unidentified restriction-modification systems in
the bacterium or in the control of various DNA functions (3,
15, 29, 35).
The present study demonstrates that genomic m4C is not
mostly limited to the DNA of thermophilic bacteria. Rather,
it is present as a minor base in the DNA of many bacterial
mesophiles. Of the 26 bacterial species examined in this
study, 9 contained m4C in their genomes. Most of the
examined bacteria are mesophiles, and all of them are
sources ofcommercially available restriction endonucleases.
The prevalence of m4C as a minor base in these bacterial
DNAs indicates that many restriction endonucleases may be
inhibited by N4-methylation of cytosine residues in their
recognition sites in vivo and that bacterial DNA cytosine
methyltransferases must be carefully checked to determine
whether they catalyze the formation of m4C.
it can reveal the presence of new DNA
MATERIALS AND METHODS
Strains. All bacteria came from the New England BioLabs
strain collection except Xanthomonas campestris
oryzae, which was from the collection of M. Ehrlich. The
strains are listed in Table 1 with the sources from which they
were originally obtained and the temperature at which the
cultures were grown.
Purification of bacterial DNA. DNA was prepared from 10
g of wet packed cells suspended in 20 ml of 25% sucrose-50
mM Tris hydrochloride (pH 8.0), and then 10 ml of 0.25 M
disodium EDTA (pH 8.0) and 6 ml of 10-mg/ml lysozyme in
0.25 M Tris (pH 8.0) were added. After 2 h at 0°C, 24 ml of
1% Triton X-100 in 50 mM Tris hydrochloride-67 mM
disodium EDTA (pH 8.0) and 5 ml of 10% sodium dodecyl
sulfate were added. The solutions were shaken gently for
several minutes to achieve complete mixing and cell lysis.
They then were extracted with phenol and chloroform,
dialyzed, and treated with RNase I, and the DNA was
precipitated by standard techniques (26).
Digestion ofDNA and chromatography ofdeoxynucleosides.
DNA was digested to nucleosides with nuclease P1 and
Escherichia coli alkaline phosphatase (14). The resulting
deoxynucleosides from 20 to 90 ,ug of DNA were separated
by high-performance liquid chromatography. A previously
described elution system (11, 14) was improved to increase
the resolution, sensitivity, and speed ofthe chromatography.
We used a specially developed reversed-phase column
(Supelcosil LC-18S; 250 by 4.6 mm; Supelco); a two-buffer,
single-ramp elution gradient; a 1.0-ml/min flow rate; and a
26°C isothermal column temperature. Buffer A was 2.5%
methanol in 0.05 M potassium phosphate (pH 4.5), and
buffer B was 20% methanol in 0.05 M potassium phosphate
(pH 4.0). To assure consistency, all of the columns were
pretested by using a mixture of nucleoside standards. The
gradient consisted of100% bufferA isocratic from 0 to 5 min,
a linear ramp from 100% buffer A to 100% buffer B from 5 to
20 min, and 100% buffer B isocratic from 20 to 30 min. The
column was washed with 70% methanol in water for 5 min
and equilibrated with bufferA for 15 min before the next run.
The ramp rate, pH, and column temperature must be rigor-
ously maintained to ensure separation of m5dCyd and
m4dCyd. A Hewlett-Packard model 1090A liquid chromatog-
raphy instrument equipped with DR-5 solvent delivery sys-
tem, automatic injector, automatic sampler, diode-array
detector, HP-80B controller, HP-7470A plotter, HP-9133
hard disk drive, HP Think Jet printer, and data process unit
multichannel integrator was used. Absorption of light was
measured at 254
2 nm and 280 + 2 nm. Given a sample size
of >30 ,ug, a methylated deoxynucleoside representing 0.01
mol% of the total deoxynucleosides could be detected.
JOURNAL OF BACTERIOLOGY, Mar. 1987, p. 939-943
Copyright © 1987, American Society for Microbiology
EHRLICH ET AL.
TABLE 1. Sources of bacteria
Bacillus amyloliquefaciens H
C. Duncan and
C. Hutchinson III
Fusobacterium nucleatum D
Haemophilus influenzae Rd
Haemophilus influenzae Rf
Klebsiella pneumoniae OK8
Thermus aquaticus YT1
Xanthomonas campestris pv.
Xanthomonas campestris pv.
Xanthomonas campestris pv.
Xanthomonas campestris pv.
RESULTS AND DISCUSSION
In a previous study, m4C was found in the genomes of
approximately one-half of the 17 types of examined
thermophilic bacteria (11). In that study, 15 types of meso-
philic bacteria were also examined; most were anaerobes,
like the thermophiles. None of these contained m4C in their
DNA. In the present study, we looked at a greater variety of
bacteria, almost all of which are mesophiles (Table
Because of their similar chemical nature, specially designed
chromatography systems (5, 11) have to be used to resolve
m4dCyd and m5dCyd (or the corresponding bases), which
are difficult to separate. With a high-performance liquid
chromatography system that resolves m4dCyd and m5dCyd,
we have quantitated these deoxynucleosides in DNA di-
gests. Besides using the different retention times of these
deoxynucleosides to identify the m4dCyd and m5dCyd
peaks, we also rely on their different UV-light absorbance
ratios. In the eluting buffer, A280 relative to A254 is 3.0 for
m5dCyd and 1.5 for m4dCyd.
The base composition of the different bacterial DNAs is
given in Tables 2 and 3. There were no irregularities in their
major base content, and these data are in agreement with
previously reported values available for some of these spe-
cies (33). The similarity of the major base compositions of
the DNA of the five examined Haemophilus species (Tables
2 and 3) reflects their genetic relatedness. In contrast, in
duplicate determinations, the major base content of the
DNA from Bacillus caldolyticus differed considerably from
that ofthe other Bacillus species (Tables 2 and 3). This result
suggests that B. caldolyticus is genetically quite divergent
from the rest of the examined species of its genus.
All of the bacteria studied contain at least one modified
base (m6A, m5C, or m4C) as a minor component. Of the 26
types of bacteria examined in this study, 9 had m4C in their
DNA (Tables 2 and 3). m5C was present in the genomes of 16
of the DNAs studied (Tables 2 and 3). This is in good
agreement with the previous findings of m5C in the DNA of
about 60% of 44 examined strains of mesophilic bacteria (10,
11, 21, 32, 33).
As was observed in previous studies (11, 33), most (88%)
of the types ofDNA studied had m6A as a minor base. Often
it was present at rather high levels (>0.3 mol%) (Tables 2
and 3). These bacteria may have a dam-type (5'-Gm6ATC-3')
methylation system directing mismatch repair (15, 23, 29).
The dam methylase of E.
restriction enzyme with corresponding specificity for
unmethylated 5'-GATC-3' sites. Consistent with their mod-
erately high m6A content (Tables 2 and 3), Haemophilus
gallinarum, Haemophilus parainfluenzae, Proteus vulgaris,
Klebsiella pneumoniae, and Serratia marcescens have E.
coli-like dam methylation as determined by restriction and
DNA hybridization analysis (3).
The 11 bacterial strains containing <0.2 mol% m6A in
their DNA (Tables 2 and 3) may lack the dam-type modifi-
cation pathway, which has been implicated in the regulation
of transcription, of transposition, and of initiation of DNA
replication as well as in directing mismatch repair in E. coli
(15, 29, 35). Indeed, DNAs from Bacillus globigii and X.
campestris pathovars holcicola, malvacearum, and oryzae,
which have <0.2 mol% m6A (Table 2), were previously
shown not to possess dam-type methylation (3). The m6A
content ofMoraxella bovis DNA can be accounted for by its
methylation at the A residues of 5'-GATC-3' sites as part of
a restriction-modification system (4). B. caldolyticus has
only one known restriction endonuclease, BclI, which rec-
ognizes a 6-base-pair sequence (30). The relatively high level
of m6A in its DNA and the previously reported absence in
this bacterium of sequences that hybridize with those of the
E. coli dam gene (3) suggest that B. caldolyticus harbors
either another restriction-modification system probably rec-
ognizing a 4-base-pair sequence or a DNA (adenine-
N6)methyltransferase other than a dam methylase or restric-
More than one modified base was present in most of the
bacteria examined (Tables 2 and 3). In some cases, as for H.
parainfluenzae, this may reflect the presence in one bacte-
rium of dam methylation (3) as well as several restriction-
modification systems involving different modified bases (25,
41). In contrast, Haemophilus influenzae Rd, whose only
modified base in its DNA is m6A (Table 3), has multiple
DNA methyltransferases, all of which methylate adenine
residues (31). Also, B. caldolyticus, B. globigii, Streptococ-
cus pneumoniae, H.
influenzae Rf, Streptomyces
achromogenes, and X. campestris pv. oryzae contained only
one detectable modified base in their DNA (Tables 2 and 3).
From their major base compositions and the sequences of
their restriction recognition sites, the genomic frequency of
modified bases resulting from methylation at a given class of
restriction sites can be estimated by assuming a random
sequence of bases in the genome and one modified base on
each strand per recognition site. For example, in the cases of
Streptomyces achromogenes (Sacd, GAGCTC; SacIl,
CCGCGG ) and X. campestris pv. oryzae (XorI,
coli is unaccompanied by a
N4-METHYLCYTOSINE IN BACTERIAL DNA
TABLE 2. Minor and major base composition of DNAs from bacteria containing restriction endonucleases that had been tested for
sensitivity to cytosine methylation
Inhibition by cytosine
m6AA + T
H. influenzae Rf
X. campestris pv. holcicola
X. campestris pv. malvacearum
X. campestris pv. oryzae
aOnly the commonly used restriction endonucleases found in the given bacterial strain (Table 1) are listed. See reference 30 for a listing of all known restriction
enzymes and their recognition sequences in these bacteria.
bThe bacteria contain restriction endonucleases at least one of which was tesed for inhibition by site-specific cytosine 5-methylation. AluI, BamHI, BglII,
HaeIII, HhaI, HpaII, PvuI, XhoII, XmaI, and XorII were inhibited by methylation of their DNA substrates catalyzed by at least one bacterial DNA
(cytosine-5)methyltransferase, and BcII was not inhibited by such methylation (4, 17, 25, 27, 34). In the cases of AIuI, BamHI, HaeII, HhaI, and HpaIj, this
inhibition was catalyzed by a DNA (cytosine-5)methyltransferase isolated from the host bacterium and, hence, implicated in that restriction-modification system.
HpaII, HhaI, Bgll, HaeIII, SmaI, and XhoI were inhibited by the vertebrate-type DNA methylation, namely, cytosine 5-methylation at CpG sites within or
overlapping their recognition sites (2, 36, 37, 42). MboI and Hinfl are able to cleave DNA in which virtually all of the cytosine residues are methylated, although
the rate ofcatalysis is unusually slow (18); Hinfl was also partially inhibited bympthylationofaDNA substrate at its CpG sites by a human DNA methyltranferase
(R. Y.-H. Wang and M. Ehrlich, unpublished results). Catalysis by TaqI is unusual in being unaffected by methylation of all the cytosine residues of a DNA
substrate (18). Although almost all restriction endonucleases are inhibited by complete substitution of cytosine residues in their DNA by m5C residues (18), that
is not considered in this table because the extensive nature of such substitution might have indirect effects on enzyme-DNA interactions.
CTGCAG; XorII, CGATCG ), the predicted frequency
of methylated base resulting from any one restriction-
modification system (-0.03 mol% each for Sacl, XorI, and
XorII and -0.1 mol% for SacIl) is much more than the
detection limits for m4C and m6A (<0.01 mol% m4C or m6A;
Tables 2 and 3). Therefore, Streptomyces achromogenes
with 0.23 mol% m5C in its DNA and X. campestris pv.
oryzae with 0.09 mol% m5C probably possess only DNA
Evidence for unexpectedly low frequencies of restriction
TABLE 3. Minor and major base composition of some bacterial
H. influenzae Rd
N. aerocolonigenes NaeI
X. campestris pv.
aThese bacteria, unlike those listed in Table 2, have restriction endo-
nucleases that have not been tested for inhibition by site-specific cytosine
bOnly the commonly used restriction endonucleases found in the given
bacterial strain (Table 1) are listed. See reference 30 for alistingof all known
restriction enzymes and their recognition sequences in these bacteria.
haemolyticus and Haemophilus aegyptius. They both have
DNA (cytosine-5)methyltransferases that recognize a 4-
base-pair site containing only C G base pairs (30). From the
major base compositions of H. haemolyticus and H.
aegyptius DNAs (Table 2), the methylation of the HhaI or
HaeIII recognition sequences would be expected to yield
-0.15 mol% m5C in the genome if a random distribution of
bases is assumed. The actual level of m5C in these genomes
was 0.07 mol%, implying at least a twofold underrepresen-
tation ofthe recognition sites. A similar underrepresentation
of about fourfold in 5'-GATC-3' sites was previously ob-
served for Thermobacteroides acetoethylicus (11). An even
greater difference between the predicted and the observed
minor base compositions was seen for Fusobacterium
nucleatum D. This bacterium harbors three restriction
endonucleases, all ofwhich recognize 4-base-pair sequences
containing only C
CGCG-3', and 5'-GCGC-3' (24). F. nucleatum has only 28
mol% G+C in
tetranucleotide sequences occurred at the frequency ex-
pected for a random distribution of bases in this DNA, then
there should have been -0.12 mol% methylated cytosine to
confer resistance at the recognition sequences to the host
restriction endonucleases. However, induplicate determina-
tions, no m4C was detected in this DNA, and only 0.023
mol% m5C was found (Table 3). As expected, DNA from this
organism was resistant to digestion by FnuDII and the
FnuDI and FnuDIII isochizomers HaeIII and HhaI. This
underrepresentation ofthe modified base involves thousands
of sites per bacterial genome. It might be due to selective
pressure against too many potential restriction sites or too
many modification sites which could be subjectto interac-
tions with sequence-specific DNA-binding proteins (35, 39).
sites was observed for Haemophilus
-G base pairs, namely, 5'-GGCC-3', 5'-
its genome (Table
If the above
VOL. 169, 1987
EHRLICH ET AL.
Alternatively, it might result from selective pressures gener-
ating nonrandom dinucleotide and trinucleotide frequencies
such as selective codon usage (1, 28). On the other hand,
amyloliquefaciens (Table 2) (17), are present at much higher
concentrations than expected based on their identified re-
striction endonuclease recognition sites.
The strain of Streptococcus pneumoniae used for this
study has a rare, although not unique (I. Schildkraut, unpub-
lished results), methylation-dependent restriction system.
The single known restriction enzyme of this strain, DpnI,
cleaves DNA only if the genome is methylated at 5'-GATC-
3' sites to yield bifilarly modified 5'-Gm6ATC-3' sequences
(22). This appears to be a s'trategy to restrict foreign dam-
methylated DNA, especially from bacteriophages. The DNA
of this strain is unmethylated at its 5'-GATC-3' sites and is
thereby protected from such restriction (22). Because this
strain contains 0.14 mol% m6A in its DNA (Table 3), we
predict that it will be found to possess another restriction
system of the more conventional type that requires adenine
methylation for inhibiting the restriction enzyme from hy-
drolyzing the host DNA. Alternatively, it might have a DNA
adenine methyltransferase with functions other than protec-
tion against restriction (15, 23, 29, 35).
Methylated cytosine (m5C or m4C) residues were found in
the DNA of Serratia marcescens, Nocardia aerocoloni-
genes, Bacillus aneurinolyticus, and X. campestris
malvacearum (Tables 2 and 3), which possess at least one
restriction endonuclease (30) able to cleave recognition sites
containing only C G base pairs. The first two genomes
contain m5C and no detectable m4C. Because the last two
species contain genomic m4C as well as m5C, either DNA
might methylate the corresponding XmaI (CCCGGG),
XmaIII (CGGCCG), BanI (GGYRCC), or BanII (GRGCYC)
recognition sites in these bacteria.
In this study, we have found that m4C is frequently present
in mesophilic bacteria as a minor genomic base, just as
previously found in thermophilic bacteria (11) and' in the
mesophile Bacillus centrosporus (19). m5C was not found in
the DNA of any of 15 examined thermophiles which grow
thermophiles examined in this study, B. caldolyticus and
Thermus aquaticus, contained no detectable genomic m5C
(Table 2). Also, Thermoplasma acidophilum 122-1B2, which
is grown optimally at 59°C (and contained 54 mol% A+T in
its genome), had m4C (0.21 mol%), m6A (2.05 mol%), and no
m5C (<0.01 mol%) in its DNA (D. Swinton, S. Hattman, D.
Searcy, and C. Gehrke, unpublished results). These results
are consistent with the hypothesis that bacteria which grow
at high temperatures avoid m5C in their genomes because of
the propensity of this base to heat-induced deamination and
because of the inefficient excision of T from the resulting
T. G mismatched base pairs (12, 38; S. Shenoy, K. Erlich,
and M. Erlich, submitted for publication).
DNA (cytosine-N4)methyltransferases, like the analogous
modification systems. Janulaitis and co-workers (6, 19, 20)
have demonstrated that site-specific DNA methylases from
centrosporus, Micrococcus varians, and Citrobacter
freundii catalyze the formation of m4C residues within
sequences recognized by the restriction endonucleases of
might control DNA repair, expression, replication, or trans-
position like the adenine-specific dam methylase in E. coli
(15, 23, 29, 35).
some bacteria, such
(11). Similarly, the two extreme
some cytosine methyltransferases
We thank Ira Schildkraut of New England BioLabs for suggesting
this study and for critical reading of the manuscript. We are also
very grateful to Dat Phan for his excellent technical assistance in the
This research was supported in part by National Science Foun-
dation grant PCM-8219000.
phylogenetic relationships from DNA restriction patterns and
selection of endonuclease cleavage sites. Proc. Natl. Acad. Sci.
2. Bird, A. P., and E. M. Southern. 1978. Use of restriction
enzymes to study eukaryotic DNA methylation. J. Mol. Biol.
3. 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.
4. Brooks, J. E., and R. J. Roberts. 1982. Modification profiles of
bacterial genomes. Nucleic Acids Res. 10:913-924.
5. Butkus, V. V., S. J. Klimasauskas, and A. A. Janulaitis. 1985.
Analysis of products of DNA modification by methylases: a
procedure for the determination of 5- and N4-methylcytosines in
DNA. Anal. Biochem. 148:194-198.
6. Butkus, V. V., S. J. Klimasauskas, D. Kersulyte, D. Vaitkevicius,
A. Lebionka, and A. Janulaitis. 1985. Investigation of restric-
tion-modification enzymes from M. varians RFL19 with a new
type of specificity toward modification of substrate. Nucleic
Acids Res. 13:5727-5746.
7. Doskocil, J., and Z. Sormova. 1965. The occurrence of 5-
methylcytosine in bacterial deoxyribonucleic acids. Biochim.
Biophys. Acta 95:513-515.
8. Dubin, D. T., and C. C. HsuChen. 1983. The 3'-terminal region
of mosquito mitochondrial small ribosomal subunit RNA: se-
quence and localization of methylated residues. Plasmid
9. Dunn, D. B., and J. D. Smith. 1958. The occurrence of 6-
10. Dybvig, K., D. Swinton, J. Maniloff, and S. Hattman. 1982.
Cytosine methylation of the sequence GATC in a mycoplasma.
J. Bacteriol. 151:1420-1424.
11. Ehrlich, M., M. Gama-Sosa, L. Carreira, L. Ljungdahl, K. Kuo,
and C. Gehrke. 1985. DNA methylation in thermophilic bacte-
methyladenine. Nucleic Acids Res. 13:1399-1412.
12. Ehrlich, M., K. F. Norris, R. Y.-H. Wang, K. C. Kuo, and C. W.
Gehrke. 1986. DNA cytosine methylation and heat-induced
dea,mination. Biosci. Rep. 6:387-393.
13. Fellner, P. 1969. Nucleotide sequences from specific areas ofthe
16S and 23S ribosomal RNAs of E. coli. Eur. J. Biochem.
14. Gehrke, C. W., R. A. McCune, M. A. Gama-Sosa, M. Ehrlich,
and K. C. Kuo. 1984. Quantitative reversed-phase high-
performance liquid chromatography of major and modified
nucleosides in DNA. J. Chromatogr. 301:199-219.
15. Glickman, B. W., and M. Radman. 1980. E. coli mutator
mutants deficient in methylation-instructed DNA mismatch cor-
rection. Proc. Natl. Acad. Sci. USA 77:1063-1067.
16. Hattman, S. 1981. DNA methylation, p. 517-548. In P. D. Boyer
(ed.), The Enzymes, vol. 14. Academic Press, Inc., New York.
17. Hattman, S., T. Keister, and A. Gottehrer. 1978. Sequence
specificity of DNA methylases from Bacillus amyloliquefaciens
and Bacillus brevis. J. Mol. Biol. 124:701-711.
18. Huang, L.-H., C. MI. Farnet, K. C. Ehrlich, and M. Ehrjich.
1982. Digestion of highly modified bacteriophage DNA by
restriction endonucleases. Nucleic Acids Res. 10:1579-1591.
19. Janulaitis, A., S. Klimasauskas, M. Petrusyte, and V. Butkus.
1983. Cytosine modification in DNA by BcnI methylase yields
N4-methylcytosine. FEBS Lett. 161:131-134.
20. Janulaitis, A. A., P. S. Stakenas, M. P. Pyatrushite, Y. B.
Bitinaite, S. I. Klimashauskas, and V. V. Butkus. 1984. Speci-
J., and E. D. Rothman.
1982. Estimation of
in deoxyribonucleic acids.' Biochem.
N4-METHYLCYTOSINE IN BACTERIAL DNA
ficity of new restrictases and methylases. Unusual modification
of cytosine in position 4. Mol. Biol. (Engl. Transl. Mol. Biol.
21. Korch, C., P. Hagbiom, and S. Normark. 1983. Sequence-
specific DNA modification in Neisseria gonorrhoeae. J. Bacte-
22. Lacks, S., and B. Greenberg. 1977. Complementary specificity
of restriction endonucleases of Diplococcus pneumoniae with
respect to DNA methylation. J. Mol. Biol. 114:153-168.
23. Lu, A.-L., S. Clark, and P. Modrich. 1983. Methyl-directed
repair ofDNA base-pair mismatches in vitro. Proc. Natl. Acad.
Sci. USA 80:4639-4643.
24. Lui, A. C. P., B. C. McBride, G. F. Vovis, and M. Smith. 1979.
Site specific endonuclease from Fusobacterium nucleatum. Nu-
cleic Acids Res. 6:1-15.
25. Mann, M. B., and H. 0. Smith. 1977. Specificity ofHpa II DNA
methylation. Nucleic Acids Res. 4:4211-4221.
26. Marmur, J. 1961. A procedure for the isolation of deoxyribo-
nucleic acid from micro-organisms. J. Mol. Biol. 3:208-218.
27. McClelland, M. 1983. The effect of site specific methylation on
restriction endonuclease cleavage (update). Nucleic Acids Res.
28. Nussinov, R. 1984. Doublet frequencies in evolutionary distinct
groups. Nucleic Acids Res. 12:1749-1763.
29. Pukkila, P. J., J. Peterson, G. Herman, P. Modrich, and M.
Meselson. 1983. Effects of high levels of DNA adenine methyla-
tion on methyl-directed mismatch repair in E. coli. Genetics
30. Roberts, R. J. 1985. Restriction and modification enzymes and
their recognition sequences. Nucleic Acids Res. 13:rl65-r200.
31. Roy, P. H., and H. 0. Smith. 1973. DNA methylases of
Haemophilus influenzae Rd. II. Partial recognition site base
sequences. J. Mol. Biol. 81:445-459.
32. Schein, A, B. J. Berdahl, M. Low, and E. Borek. 1972. DNA
methylases of Haemophilus influenzae Rd. Biochim. Biophys.
33. Shapiro, H. S. 1976. Distribution of purines and pyrimidines in
deoxyribonucleic acids, p. 242-283. In G. D. Fasman (ed.),
Handbook of biochemistry and molecular biology, nucleic acids
section, vol. 2. CRC Press, Inc., Cleveland.
34. Smith, H. O., and S. V. Kelly. 1984. The methylases of type II
restriction and modification systems, p. 39-71. In H. Cedar,
A. P. Riggs, and A. Razin (ed.), DNA methylation. Springer-
Verlag, New York.
35. Sternberg, N. 1985. Evidence that adenine methylation influ-
ences DNA-protein interactions in Escherichia coli. J. Bacte-
36. Ulrich, A., T. J. Dull, A. Gray, J. A. Phillips, and S. Peter. 1982.
Variation in the sequence and modification state of the human
insulin gene flanking regions. Nucleic Acids Res. 10:2225-2240.
37. van der Ploeg, L. H. T., and R. A. Flavell. 1980. DNA methyla-
nonerythroid tissues. Cell 19:947-958.
38. Wang, R. Y.-H., K. C. Kuo, C. W. Gehrke, L.-H. Huang, and
M. Ehrlich. 1982. Heat and alkali-induced deamination of 5-
methylcytosine and cytosine residues in DNA. Biochem. Bio-
phys. Acta 697:371-377.
39. Wang, R. Y.-H., X.-Y. Zhang, and M. Ehrlich. 1986. A human
specific. Nucleic Acids Res. 14:1599-1614.
40. Woese, C. R., G. E. Fox, L. Zablen, T. Uchida, L. Bonen, K.
Pechman, B. J. Lewis, and D. Stahl. 1975. Conservation of
primary structure in 16S ribosomal RNA. Nature (London)
41. Yoo, 0. J., P. Dwyer-Hallquist, and K. L. Agarwal. 1982.
Purification and properties of the Hpa I methylase. Nucleic
Acids Res. 20:6511-6519.
42. Youssofian, H., and C. Mulder. 1981. Detection of methylated
DNA sequences in eukaryotic DNA with the restriction
endonucleases SmaI and XmaI. J. Mol. Biol. 150:133-136.
in the human -y4B3-globin locus
in erythroid and
is methylation-specific and sequence-
VOL. 169, 1987