Article No. jmbi.1999.3304 available online at http://www.idealibrary.com on
J. Mol. Biol. (1999) 294, 1223±1230
Crystallographic Studies on Damaged DNAs.
II. N6-Methoxyadenine can Present Two Alternate
Faces for Watson-Crick Base-pairing, Leading to
Pyrimidine Transition Mutagenesis
Toshiyuki Chatake1, Takaaki Hikima1, Akira Ono2, Yoshihito Ueno3
Akira Matsuda3 and Akio TakeÂnaka1*
1
Graduate School of Bioscience
and Biotechnology, Tokyo
Institute of Technology
Nagatsuda, Midori-ku,
Yokohama, 226-8501, Japan
2
Graduate School of Science
Tokyo Metropolitan University
Hachioji, Minamiosawa, Tokyo
192-0364, Japan
3
Graduate School of
Pharmaceutical Sciences
Hokkaido University, Nishi 6
Kita 12, Kita-ku, Sapporo
060-0812, Japan
In a previous paper, 20 -deoxy-N6-methoxyadenosine (mo6A) was shown
to form a mismatch base-pair with 20 -deoxycytidine with a Watson-Cricktype geometry. To fully understand the structural basis of genetic
mutations with damaged DNA, it is necessary to examine whether the
methoxylated adenine residue still has the ability to form the regular
Watson-Crick pairing with a thymine residue. Therefore, a DNA dodecamer with the sequence d(CGCGmo6AATTCGCG) has been synthesized
and its crystal structure determined. The methoxylation has no signi®cant
effect on the overall DNA conformation, which is that of a standard Bform duplex. The methoxylated adenine moieties adopt the amino tautomer with an anti conformation around the C6-N6 bond to the N1 atom,
and they form a Watson-Crick base-pair with thymine residues on the
opposite strand, similar to an unmodi®ed adenine residue. It is concluded that methoxylated adenine can present two alternate faces for
base-pairing, thanks to the amino $ imino tautomerism allowed by
methoxylation. Based on this property, two gene transition routes are
proposed.
# 1999 Academic Press
6
*Corresponding author
Keywords: N -methoxyadenine; mutagenesis; X-ray structure; mispairing;
damaged DNA
Introduction
Genetic information is transmitted from generation to generation through DNA replication, in
which a new complementary strand is synthesized
starting from either strand of an original DNA
duplex. In this process, the highest accuracy is
achieved by forming Watson-Crick base-pairs
between adenine and thymine bases and between
guanine and cytosine bases, as an absolute rule in
every organism. When DNA is attacked with certain chemicals, however, this rule is disturbed and
Abbreviations used: dATP, 20 -deoxyadenosine
5 -triphosphate; TTP, thymidine 50 -triphosphate; dGTP,
20 -deoxyguanosine 50 -triphosphate; dCTP, 20 deoxycytidine 50 -triphosphate; mo6A, 20 -deoxy-N6methoxyadenosine; mo6dATP, 20 -deoxy-N6methoxyadenosine 50 -triphosphate; MPD, 2-methyl-2,4pentanediol.
E-mail address of the corresponding author:
[email protected]
0
0022-2836/99/501223±8 $30.00/0
errors are introduced into the synthesized DNA.
As a result, genetic mutations will occur. These
chemicals are called mutagens (Singer &
Kusmierek, 1982). As mutations can lead to several
diseases, including cancer, extensive studies on
mutagens have been performed.
Oxyamines such as hydroxylamine and methoxylamine are known to be mutagens (Singer &
Kusmierek, 1982) which predominantly attack and
modify the exocyclic amino groups of nucleic acid
bases (Kochetkov & Budowsky, 1969). Substitution
at the N6 amino groups of adenine bases has been
established (Budowsky et al., 1975), and is thought
to be the origin of the observed mutations. It is
expected that the N6-methoxyadenine moiety has
different chemical and physicochemical properties
from the unmodi®ed base. When a DNA fragment
containing 20 -deoxy-N6-methoxyadenosine (hereafter designated as mo6A) is used as a template,
the non-complementary 20 -deoxycytidine 50 -triphosphate (dCTP) as well as the complementary thymidine 50 -triphosphate (TTP) are incorporated by
# 1999 Academic Press
1224
N6-Methoxyadenine for Watson-Crick Base-pairings
the Klenow fragment of DNA polymerase I into
the corresponding sites of the newly synthesized
DNA strand (Nishio et al., 1992). In addition, free
20 -deoxyadenosine 50 -triphosphate (dATP), which
is a reactant to be incorporated in the synthesized
DNA, is also modi®ed by oxyamines (Singer,
1975). Methoxylated dATP (hereafter designated
as mo6dATP) is incorporated at both templates, 20 deoxycytidine and thymidine (Singer & Spengler,
1982; Abdul-Maish & Bessman, 1986; Hill et al.,
1998). The DNA polymerase accepts only WatsonCrick base-pairs, therefore it is plausible that methoxylation allows the adenine moiety to form
Watson-Crick-type pairs with both thymine and
cytosine. Indeed, we recently determined the crystal structure of a DNA dodecamer with the nucleotide sequence d(CGCGmo6AATCCGCG) (hereafter
designated as Dmo6A C), and found that the
mo6A residue forms a Watson-Crick-type pair with
cytosine residues in B-form duplex DNA (Chatake
et al., 1999). The adenine moiety has the imino
tautomeric form, with the methoxy group in
the anti conformation around the C6-N6 bond to
the N1 atom. X-ray analyses of the related derivatives N9-benzyl-N6-methoxyadenine (Fujii et al.,
1990) and N6-methoxy-20 ,30 ,50 -tri-O-methyladeno-
Figure 1 (legend opposite)
N6-Methoxyadenine for Watson-Crick Base-pairings
sine (Birnbaum et al., 1984) also indicate that the
N6-methoxylated base takes the imino form,
although with the methoxy group in the syn conformation. In solution, however, these derivatives
are in equilibrium between the imino form and the
amino form (Stolarski et al., 1984), the latter being
preferred by the unmodi®ed adenine base
(Wolfenden, 1969). To understand the structural
basis of the mutation mechanism, it is necessary to
con®rm that the methoxylated adenine residue can
adopt the amino tautomer, and that it still has the
ability to form a canonical Watson-Crick pair with
a thymine residue. For this purpose, we have synthesized a DNA dodecamer with the sequence
d(CGCGmo6AATTCGCG) (Dmo6A T) and determined its crystal structure. Electron density maps
clearly show that the mo6A residue does indeed
form Watson-Crick-type pairs with thymine residues in the B-form duplex. This demonstrates the
existence of two alternate faces of the methoxylated
20 -deoxyadenosine through the amino $ imino
tautomerism, allowing its base-pairing with either
20 -deoxycytidine or thymidine, and indicates several possible gene transition mechanisms.
Results and Discussion
Crystal structures
The two Dmo6A T strands are associated to
form a right-handed double helix as shown in
Figure 1(a). All torsion angles and local helical parameters (calculated by the program NUPARM;
Bansal et al., 1995) are given in the Supplementary
1225
Material (Tables 1 and 2, respectively). The crystal
structure is isomorphous to the original dodecamer
with
the
sequence
d(CGCGAATTCGCG)
(Dickerson et al., 1981; Shui et al., 1998), and to the
modi®ed (``damaged'') dodecamer reported in the
previous paper (Chatake et al., 1999). Similar to
those structures, two duplexes related by 21 symmetry along the c-axis form columns in a head-totail fashion. To connect them, two inter-duplex
base-pairs are formed through N2-H N3 hydrogen bonds between Ga12 and Ga2, and between Gb2
and Gb12 (see Figure 2 by Chatake et al., 1999). An
octahedrally hydrated magnesium cation is located
in the major groove of each duplex. Three of the
bound water molecules form hydrogen bonds with
the O6 and N7 atoms of Ga2 and Gb10 and the other
three are hydrogen bonded to the phosphate oxygen atoms of Aa6, and Ta7, in another duplex,
related by 21 symmetry along the b-axis. This situation is the same as that found in the Dmo6A C
crystal (see Figure 4 by Chatake et al., 1999).
Effect of methoxylation on DNA conformation
The structures of both Dmo6A T (Figure 1(a))
and Dmo6A C (Figure 1(b)) are similar to that of
the original dodecamer (Figure 1(c)). The overall
root-mean-square displacement of non-hydrogen
Ê for
Ê for Dmo6A T and 0.31 A
atoms is 0.31 A
Dmo6A C, when they are superimposed on the
original dodecamer. Figure 2 shows plots of some
representative helical parameters along the nucleotide sequence. These parameters ¯uctuate around
average values close to those of typical B-form
Figure 1. Stereo views of DNA dodecamers (a) Dmo6A T, (b) Dmo6A C (Chatake et al., 1999) and (c) the original
dodecamer (Dickerson, et al., 1981; Shui et al., 1998). The diagrams were drawn with the program MOLSCRIPT
(Kraulis, 1991). The mo6A residues are shown in red and the 20 -deoxycytidine residues paired with mo6A are in
green.
1226
N6-Methoxyadenine for Watson-Crick Base-pairings
Hydrogen-bonding scheme of mo6A
Figure 3 shows 2jFoj ÿ jFcj maps at the two sites
of the mo6A residues. Each N6-methoxyadenine
moiety forms a pair with the thymine base on the
opposite strand. It is noteworthy that the geometry
of the pairing is essentially the same as the usual
Watson-Crick adenine thymine pair. The methoxy
groups of the mo6A bases take an anti conformation with respect to the N1 atom around the C6N6 bond, like those found in the Dmo6A C crystal
(Chatake et al., 1999). The hydrogen bond distances
and angles in the base-pairs are given in Table 1:
those of the other hydrogen bonds are in the Supplementary Material (Table 3). The distances
between the N6 (mo6Aa5) and O4 (Tb8) atoms and
between the N1 (mo6Aa5) and N3 (Tb8) atoms, and
the corresponding distances in the mo6Ab5 Ta8
pair, indicate that mo6A can form a base-pair
through hydrogen bonds with thymine just like the
unmodi®ed adenine residue. In the Dmo6A C
crystal, a Watson-Crick-like mo6A C pairing was
Figure 2. A comparison of some representative helical
parameters in Dmo6A T (triangles), Dmo6A C (Chatake
et al., 1999) (circles) and the original dodecamer (Dickerson et al., l981; Shui et al., 1998) (boxes). The top, middle
and bottom diagrams represent the helical rise, the helical twist angles and the displacements, respectively, of
the three duplexes. They are plotted along the nucleotide sequence to show their ¯uctuations. These parameters were calculated by the program NUPARM
(Bansal et al., 1995). In the right shaded column, their
average values are indicated, together with the typical
ones for standard A and B-form DNA (Arnot & Hukins,
1972).
DNA, and their variations along the sequence are
almost the same, even at the mo6A residues, indicating that methoxylation of adenine residues does
not signi®cantly affect the overall DNA conformation. The X-ray structure (Kiefer et al., 1998) of a
large fragment of DNA polymerase I from Bacillus
stearothermophilus, complexed with DNA primer
and template strands, shows that the two DNA
strands form a duplex with a B-form conformation
in the polymerase active site. In addition, there is
an open, solvent-accessible space in the major
groove of the bound DNA. Therefore, when adenine bases in DNA are methoxylated, the damaged
DNA can still adopt an acceptable conformation in
the polymerase, and the methoxy group will not
interfere with the binding.
Figure 3. Final 2jFoj ÿ jFcj electron density on
the
20 -deoxy-N6-methoxyadenosine:thymidine
pairs
mo6Aa5 Tb8 and mo6Ab5 Ta8 in Dmo6A T. Broken lines
show possible hydrogen bonds. Maps were contoured at
the 1.4s level by the program O (Jones et al., 1991). The
nucleotides are numbered from the 50 end independently in the two strands, a and b.
1227
N6-Methoxyadenine for Watson-Crick Base-pairings
Ê ) and angles (deg.) involved in base pairing with 20 -deoxy-N6-methoxyadenoTable 1. Hydrogen bond distances (A
sine
mo6Aa5 Tb8
Ta8 mo6Ab5
N1 N3
C2-N1 N3
N1 N3-C2
N3 N1
C2-N3 N1
N3 N1-C2
2.85
121
111
2.80
114
118
N6 O4
C6-N1 N3
N1 N3-C4
O4 N6
C4-N3 N1
N3 N1-C6
found (Chatake et al., 1999). When both the
mo6A T pair and the mo6A C pair are superimposed on the canonical A T pair found in the original dodecamer (Dickerson et al., 1981; Shul et al.,
1998), the base positions are almost the same. In
other words, the methoxylated adenine base can
form a pair with either thymine or cytosine bases
in the same geometry as a Watson-Crick pair, with
no signi®cant changes in atomic positions. Therefore, these pairings can be accepted as WatsonCrick pairs in the polymerase (Kiefer et al., 1998).
Tautomerism of mo6A and base-pairing
In the crystalline state, only the imino form has
been found in N6-methoxyadenine derivatives
(Fujii et al., 1990; Birnbaum et al., 1984). The imino
form is also preferred in the Dmo6A C crystal,
forming a Watson-Crick-type base-pair with a
cytosine residue (Chatake et al., 1999). In the present Dmo6A T crystal, however, to form the
observed base-pairs with thymine residues, the
chemical structure of the methoxylated adenine
base must be the amino form, as shown in Figure 4.
An unmodi®ed adenine base almost always exists
in the amino form, its tautomerization occurring
very rarely (Wolfenden, 1969). The existence of
either tautomer of mo6A in the crystalline state,
depending on the base in the opposite strand, is
consistent with the fact that in solution, an equilibrium is observed between the amino and the
imino forms (Stolarski et al., 1984), the ratio
between the two tautomers being a function of the
polarity of the solution. Furthermore, when a cytosine derivative is added to a solution containing an
mo6A derivative, the imino form increases in population, and when a uridine derivative is added, the
amino form increases (Stolarski et al., 1987). Therefore, it is concluded that it is the formation of a
Watson-Crick-type base-pair with either 20 -deoxycytidine or thymidine that stabilizes one or the
other tautomer.
The difference in total energy between the two
tautomers and the stabilization energy on basepairing were calculated by the molecular orbital
method (see Figure 5). When 9-methyl-N6-methoxyadenine is alone in vacuum, the amino form is
more stable than the imino form, as in a hydrophobic solution. But when mo6A makes a base-pair
with thymine or with cytosine in the Watson-Crick
type, the resulting base-pairs are stabilized at
almost the same level, consistent with the results of
2.96
120
121
2.91
119
123
C6-N6 O4
N6 O4-C4
118
120
C4-O4 N6
O4 N6-C6
123
116
O6-N6 O4
128
O4 N6-O6
128
the present X-ray analyses that 20 -deoxy-N6-methoxyadenosine can form a base-pair with either thymidine or 20 -deoxycytidine.
In summary, when a thymine residue is located
on the opposite site in a DNA duplex, mo6A takes
the amino tautomer and behaves like an unmodi®ed adenine, forming a base-pair with the thymidine. On the other hand, when a cytosine residue
is located on the opposite strand, as discussed in
the previous paper for Dmo6A C, the mo6A residue takes the imino form and behaves like a 20 deoxyguanosine, forming a base-pair with the two
hydrogen bonds between N1 (mo6A) and N3 (C)
and between N6 (mo6A) and N4 (C). The existence
of such ``alternate faces'' of 20 -deoxy-N6-methoxyadenosine in base-pairing points clearly to a possible mutation mechanism, discussed below.
Biological implication
The possibility of two alternate faces of N6-methoxyadenine is applicable to adenine residues in
template DNA strands and to dATP as a reactant
being incorporated during DNA replication. Considering with successive replication steps, possible
routes of gene transition are shown in Figure 6.
Figure 4. Chemical structures for mo6A T and
mo6A C pairings.
1228
Figure 5. Change in total energy between the two tautomers (upper half) and their stabilization energies on
base-pairing (lower half) with 1-methylthymine (left)
and with 1-methylcytosine (right), from molecular orbital calculations. To construct initial model structures,
atomic coordinates of adenosine, b-cytidine, 1-(2-deoxyb-D-ribofuranosyl)-5-methyluracil, N6-methoxy-20 ,30 ,50 -triO-methyladenosine and 11-methoxygelsemamide were
obtained from the CSD ®le (Cambridge Structural Database System). The geometry of the Watson-Crick type
pairing was obtained from those of the 9-(2-carboxyethyl)-guanine:1-methylcytosine complex (Fujita et al.,
1984) and the 9-ethyladenine:1-methyluracil complex
(Cambridge Structural Database System). The ribose of
each nucleotide was replaced with a methyl group to
save computer time. The atomic coordinates were optimized using a 3-21G basic set by the ab initio method
with the program Gaussian94 (Frisch et al., 1995).
When an adenine base in DNA is methoxylated,
the original A T pair can be replaced by a G C
pair after two steps. In the ®rst step, if dCTP is
incorporated instead of TTP at the methoxylated
adenine residue, then dGTP can be incorporated at
the 20 -deoxycytidine site in the second step. In the
case of mo6dATP incorporation, A T to G C and
G C to A T mutations are possible with three
replication steps. In the A T ! G C case, if
mo6dATP is incorporated at the template thymine
residue by forming a mo6A T pair, then the incorporated mo6A residue will next accept dCTP by
mimicking G; ®nally, dGTP will be incorporated at
N6-Methoxyadenine for Watson-Crick Base-pairings
Figure 6. A gene transition mechanism. There are two
possible routes, which occur when (a) the template adenosine residue on DNA is methoxylated and when (b)
dATP, which is a reactant for DNA replication, is methoxylated.
the cytosine residue of the second template. In the
G C ! A T case, if mo6dATP is incorporated at
the template cytosine residue by mimicking G,
then the incorporated mo6A residue will accept
TTP; ®nally dATP will be incorporated at the template thymine residue.
Experimental Procedures
Synthesis and crystallization
DNA dodecamer of Dmo6A T was synthesized and
crystallized in the same way as that reported in the previous paper (Chatake et al., 1999). A reservoir solution
containing 25 % (v/v) 2-methyl-2,4-pentanediol (MPD),
18 mM magnesium acetate, 6 mM spermine tetrahydrochloride, 80 mM sodium chloride and 10 mM sodium
cacodylate at pH 7.0 was used.
1229
N6-Methoxyadenine for Watson-Crick Base-pairings
Data collection
To prevent nucleation and growth of hexagonal ice,
the crystals were bathed in a reservoir solution containing 35 % (v/v) MPD for one minute before ¯ash freezing.
X-ray data were collected at 110 K on the Sakabe-Weissenberg camera (Sakabe, 1991) with synchrotron radiation at the Photon Factory (BL-6B) in Tsukuba. To
compensate the blind region, another crystal with a
different orientation was used at 100 K. The two sets of
diffraction patterns were processed by the program
DENZO (Otwinowski & Minor, 1997), and intensity data
of a total of 28,875 observed re¯ections were merged by
the programs SCALA and AGROVATA in CCP4
(Collaborative Computational Project Number 4, 1994)
into 7518 independent re¯ections with Rmerge of 3.6 %.
The completeness of the data was 79.7 % in the
Ê resolution range and 59.5 % for the outer
100 1.6 A
Ê resolution shell. The unit cell dimensions
1.67 1.60 A
Ê , b ˆ 39.8 A
Ê and c ˆ 66.5 A
Ê , the space
are a ˆ 25.5 A
group being P212121. Statistics of data collection and
crystal data are given in Table 2.
Structure determination
The initial structure was determined by the molecular
replacement method, using the structure of the original
DNA dodecamer d(CGCGAATTCGCG) (Shui et al.,
1998) with the program AMoRe (Navaza, 1994). To
Table 2. Crystal data, data collection and structure
determination
Space group
Ê)
Unit cell (A
Asymmetric unit (duplexes)
Ê)
Resolution (A
Measured reflections
Unique reflections
Completeness (%)
in the outer shell (%)
Rmerge a(%)
Structure refinement
Resolution range
Used reflections
Final model
Number of DNA atoms
Number of water molecules
Number of spermine atoms
Number of magnesium
atoms
Ê 2)
Average B-factors (A
DNA
Water molecules
Spermine atoms
Magnesium atom
R-factorb (%)
Rfree c (%)
r.m.s. deviation
Ê)
Bond lengths (A
Bond angles (deg.)
Improper angles (deg.)
Ê)
Average coordinates errord (A
a
P212121
a ˆ 25.5, b ˆ 39.8, c ˆ 66.5
4
100 1.6
28,875
7518
79.7
Ê)
59.5(1.67 1.60 A
3.6
5.0 1.6
6836
490
126
14
1
16.4
33.1
56.6
13.8
21.1
27.0
0.005
1.1
1.6
0.22
Rmerge ˆ 100 hkljjIhklj ÿ hIhklij/hklhIhkli.
R factor ˆ 100 jjFoj ÿ jFcjj/jFoj, where jFoj and jFcj are
observed and calculated structure factor amplitudes, respectively.
c
Calculated using a random set containing 10 % of observations, that were omitted during re®nement (BruÈnger, 1992b).
d
Estimated from a Luzzati plot (Luzzati, 1952).
b
inspect the molecular structure and to add water molecules, the programs O (Jones et al., 1991) and QUANTA
(distributed by Molecular Simulations, Inc.) were used.
Taking the hydrogen-bonding scheme into consideration,
the mo6A bases were assumed to have an amino form.
For X-PLOR (BruÈnger, 1992a) re®nement, the stereochemical parameters of the amino form of mo6A were
derived by combination of the structures of an unmodi®ed 20 -deoxy-adenosine and the methoxy group of 11methoxygelseamide (Lin et al., 1989). A combination of
simulated annealing and positional re®nements was
applied, followed by interpretation of omit maps at
every nucleotide residue. One magnesium cation
hydrated octahedrally and one spermine molecule was
found, and 126 water molecules were assigned. During
structural re®nement, no constraints were applied
between the paired mo6A and thymidine residues. The
Ê resolution data
®nal R-factor was 21.1 % for 5.0 1.6 A
(Rfree ˆ 27.0 % for 10 % of the observed data). Statistics of
the structure determination are summarized in Table 2.
Data Bank accession code
The atomic coordinates have been deposited in the
Nucleic Acid Database (NDB) (entry code BD0010).
Acknowledgments
We thank N. Sakabe and N. Watanabe for facilities
and help during data collection at the Photon Factory
(Tsukuba), and T. Simonson for proofreading of the
manuscript. This work was supported in part by a grant
for the RFTF(97L00503) (Research For The Future) from
the Japanese Society for the Promotion of Science, by
Grants-in-Aid for Scienti®c Research on Priority Area
(Nos. 08124203 and 07250205) from the Ministry of Education, Science, Sport and Culture of Japan, and by the
Sakabe project of TARA (Tsukuba Advanced Research
Alliance), University of Tsukuba.
References
Abdul-Masih, M. T. & Bessman, M. J. (1986). Biochemical studies, 6-N-hydroxyaminopurine. J. Biol. Chem.
261, 2020-2026.
Arnott, S. & Hukins, D. W. L. (1972). Optimized parameters for A-DNA and B-DNA. Biochem. Biophys.
Res. Commun. 47, 1504-1510.
Bansal, M., Battacharyya, D. & Ravi, B. (1995).
NUPARM and NUCGEN: software for analysis and
generation of sequence dependent nucleic acid
structures. Comput. Appl. Biosci. 13, 281-287.
Birnbaum, G. I., Kierdaszuk, B. & Shugar, D. (1984).
Tautomerism and conformation of the promutagenic analogue N6-methoxy-20 ,30 ,50 -tri-O-methyladenosine. Nucl. Acids Res. 12, 2447-2460.
BruÈnger, A. T. (1992a). X-PLOR Version 3.1. A System for
X-ray Crystallography and NMR, Yale University
Press, New Haven, CT.
BruÈnger, A. T. (1992b). Free-R value: a novel statistical
quantity for assessing the accuracy of crystal structures. Nature, 355, 472-475.
Budowsky, E. I., Sverdlov, E. D., Spasokukotskaya, T. N.
& Koudelka, J. (1975). Mechanism of the mutagenic
action of hydroxyamine VIII. Functional properties
1230
N6-Methoxyadenine for Watson-Crick Base-pairings
of the modi®ed adenosine residues. Biochim. Biophys. Acta, 390, 1-13.
Chatake, T., Ono, A., Ueno, Y., Matsuda, A. &
TakeÂnaka, A. (1999). Crystallographic studies on
damaged DNAs, I. An N6-methoxyadenines forms a
Watson-Crick pair with a cytosine residue in a
B-DNA duplex. J. Mol. Biol. 294, 1215-1222.
Collaborative Computational Project Number 4 (1994).
The CCP4 suite: programs for protein crystallography. Acta Crystallog. sect. D, 50, 760-763.
Dickerson, R. E. & Drew, H. R. (1981). Structure of a
B-DNA dodecamer. II. In¯uence of base sequence
on helix structure. J. Mol. Biol. 149, 761-786.
Frisch, M. J., Frisch, A. E. & Foresman, J. B. (1995). Gaussian94 User's Reference, Gaussian Inc., Pittsburgh.
Fujii, T., Saito, T., Date, T. & Nishibata, Y. (1990). Purines. XXXIX. The crystal structure of 9-benzyl-N6methoxyadenine. Chem. Pharm. Bull. 38, 912-916.
Fujita, S., Takenaka, A. & Sasada, Y. (1984). A model for
interactions of amino acid side chains with WatsonCrick base pair of guanine and cytosine; crystal
structure of 9-(2-carboxyethyl)guanine and its crystalline complex with 1-methylcytosine. Bull. Chem.
Soc. Jpn, 57, 1707-1712.
Hill, F., Williams, D. M., Loakes, D. & Brown, D. M.
(1998). Comparative mutagenicities of N6-methoxy2,6-diaminopurine and N6-methoxyaminopurine
20 -deoxyribonucleosides and their 50 -triphosphates.
Nucl. Acids Res. 26, 1144-1149.
Jones, T. A., Zou, J. Y., Cowman, S. W. & Kjeldgaard,
M. (1991). Improved methods for building protein
models in electron density maps and the location of
errors in these models. Acta Crystallog. sect. A, 47,
110-119.
Kiefer, J. R., Mao, C., Braman, J. C. & Beese, L. S. (1998).
Visualizing DNA replication in a catalytically active
Bacillus DNA polymerase crystal. Nature, 391, 304307.
Kochetkov, N. K. & Budowsky, E. I. (1969). The chemical modi®cation of nucleic acids. Prog. Nucl. Acid
Res. Mol. Biol. 9, 403-438.
Kraulis, P. J. (1991). MOLSCRIPT: a program to produce
both detailed and schematic plots of protein structures. J. Appl. Crystallog. 24, 946-950.
Lin, L. Z., Cordell, G. A., Ni, C. Z. & Clardy, J. (1989).
Gelseamide and 11-methoxygelseamide, two novel
secoindole alkaloids from gelsemium elegans. Tetrahedron Letters, 30, 1177.
Luzzati, P. V. (1952). Traitement statistique des erreurs
dans la determination des structures cristalines.
Acta Crystallog. 5, 802-810.
Navaza, J. (1994). AMoRe: an automated package for
molecular replacement. Acta Crystallog. sect. A, 50,
157-163.
Nishio, H., Ono, A., Matsuda, A. & Ueda, T. (1992). The
synthesis and properties of oligodeoxyribonucleotides containing N6-methoxyadenine. Nucl. Acids
Res. 20, 777-782.
Otwinowski, Z. & Minor, W. (1997). Processing of X-ray
diffraction data collected in oscillation mode.
Macromolecular crystallography Part A. Methods
Enzymol. 276, 307-326.
Sakabe, N. (1991). X-ray diffraction data collection system for modern protein crystallography with a
Weissenberg camera and an imaging plate using
synchrotron radiation. Nucl. Instrum. Methods Phys.
Res., A303, 448-463.
Shui, X., McFail-Isom, L., Hu, G. G. & Williams, L. D.
(1998). The B-DNA dodecamer at high resolution
reveals a spine of water on sodium. Biochemistry,
37, 8341-8355.
Singer, B. (1975). The chemical effects of nucleic acid
alkylation and their relation to mutagenesis and
carcinogenesis. Prog. Nucl. Acid Res. Mol. Biol. 15,
219-284.
Singer, B. & Kusmierek, J. T. (1982). Chemical mutagenesis. Annu. Rev. Biochem. 51, 655-693.
Singer, B. & Spengler, S. (1982). Reaction of O-methylhydroxylamine with adenosine shifts tautomeric
equilibrium to cause transitions. FEBS Letters, 139,
69-71.
Stolarski, R., Kierdaszuk, B., Hagberg, C. E. & Shugar,
D. (1984). Hydroxylamine and methoxyamine mutagenesis: Displacement of the tautomeric equilibrium
of the promutagen N6-methoxyadenosine by complementary base pairing. Biochemistry, 23, 2906-2913.
Stolarski, R., Kierdaszuk, B., Hagberg, C. E. & Shugar,
D. (1987). Mechanism of hydroxylamine mutagenesis: Tautomeric shifts and proton exchange
between the promutagen N6-methoxyadenosine and
cytidine. Biochemistry, 26, 4332-4337.
Wolfenden, R. V. (1969). Tautomeric equilibria in inosine
and adenosine. J. Mol. Biol. 40, 307-310.
Edited by I. Tinoco
(Received 29 June 1999; received in revised form 6 October 1999; accepted 12 October 1999)