Carcinogenesis vol.17 no.5 pp.895-9O2, 1996
COMMENTARY
The role of adduct site-specific mutagenesis in understanding how
carcinogen—DNA adducts cause mutations: perspective, prospects
and problems
Edward L.Loechler
Department of Biology, Boston University, Boston, MA 02215, USA
Usually, a particular mutagen/carcinogen forms adducts at
many sites in DNA, making it impossible to determine
which type of adduct causes which mutation and why.
Adduct site-specific mutagenesis studies, in which a single
adduct is built into a vector, can be used to overcome this
problem. The adduct can be situated in double-stranded
DNA, single-stranded DNA or in a single-stranded gap,
and the benefit and concerns associated with each are
addressed. An adduct site-specific study is most useful
when it is compared to a mutagenesis study with its
corresponding mutagen/carcinogen. Mutations induced
by a particular mutagen/carcinogen can be influenced
by DNA sequence context, mutagen/carcinogen dose (and
other changes in conditions), level of SOS induction, cell
type and other factors. Thus, it is important to match
the conditions of the adduct study versus the mutagen/
carcinogen study as closely as possible. DNA sequence
context can profoundly affect the quantitative and qualitative pattern of adduct mutagenesis, which is addressed.
In vitro studies with DNA polymerases, frameshift mutagenesis and semi-targeted mutagenesis, whereby a mutation
is induced near but not at the site of the adduct, are
each discussed. Finally, the relationship between structural
studies on adducts and mutagenesis is considered.
Introduction
Carcinogens are usually mutagens and are thought to initiate
tumorigenesis by reacting with DNA to form DNA adducts,
which can be misreplicated when encountered by a DNA
polymerase. Thus, the ability of DNA adducts to induce
mutations is an integral part of the mechanism of carcinogenesis. A particular mutagen/carcinogen generally reacts at a
variety of atoms in DNA to generate a variety of adducts,
making it difficult to determine definitively which type of
adduct caused which mutation and why (for examples, see ref.
1). One potential means to assess the biological consequences
of individual DNA adducts involves their study in defined
DNA sequences using adduct site-specific techniques.
In brief, adduct site-specific mutagenesis involves the
synthesis of an oligonucleotide that contains an adduct of
known structure at a defined position, and the use of recombinant DNA techniques to incorporate this oligonucleotide into
an autonomously replicating plasmid- or viral-based vector.
This vector can be studied in vitro or placed into cells, where
biological processing occurs, the nature of which can be
deduced from an analysis of the progeny. I shall use the term
•Abbreviations: MF, mutation frequency, Me-O*-Gua, C^-methylguanine;
AFB|, aflatoxin B,; ss, single-stranded; ds, double-stranded; 2-AAF, 2acerylaminofluorene; BPDE, benzo[a]pyrene diol epoxide; MNU, methylnitrosourea; Bz-O6-Gua, O*-benzylguanine; B[a]P, benzo[a]pyrene.
© Oxford University Press
'adduct site-specific' mutagenesis to describe this process
instead of 'site-specific' or 'site-directed' mutagenesis, which
are more generally associated with mutagenesis for protein
technology. In addition, I shall use the term 'random mutagenesis' when I refer to a mutational study with a mutagen/
carcinogen itself. Of course the latter is far from random,
except in the sense that it is stochastic. Finally, I shall use the
term 'quantitative effects' to refer to issues related to differences in mutation frequency (MF*), which is frequently
discussed in terms of 'hotspots/coldspots', and 'qualitative
effects' to refer to differences in kinds of mutations (e.g. GC
—» TA versus GC —> AT), which is commonly called 'mutagenic
specificity'.
Bea Singer and John Essigmann outlined the strategies
and techniques involved in doing adduct site-specific mutagenesis, as well as many of the fundamental issues, in a fine
review of this topic in Carcinogenesis in 1991 (1). The
issues they raised will not be reiterated here, and whereas
they principally used the adducts of alkylating agents
[e.g. O^methylguanine (Me-O6-Gua)] for purposes of illustration, I shall focus on the adducts of bulky mutagens/carcinogens, such as polycyclic aromatic hydrocarbons, aromatic
amines and mycotoxins [e.g. aflatoxin B\ (AFB,)]. I will only
incidentally touch on adduct site-specific studies of exocyclic
adducts, UV lesions and lesions formed by oxidative DNA
damage, each of which deserves its own review.
In this Commentary I give examples of how mutagenesis
and its study can be complex. Accordingly, the tone at times
may seem somewhat negative. This simply reflects my opinion
that mutagenesis studies in general, and adduct site-specific
studies in particular, must be done carefully, and important
assumptions and caveats must be both recognized and stated,
hi spite of this, I believe that adduct site-specific mutagenesis
studies are valuable—indeed critical—to answering certain
kinds of mechanistic questions.
Vector-related considerations
hi an adduct site-specific study, the choice of vector is
significant. Both single-stranded (ss) and double-stranded (ds)
vectors, as well as vectors where the lesion is in a ss gap,
have been used. Some issues related to these choices are
described below.
A variety of plasmid- and phage-based systems have been
used to perform adduct site-specific mutagenesis in dsDNA
(1). The advantage of studying a lesion in a ds vector is that
the system comes as close as possible to resembling genomic
replication, including with representative DNA repair. One of
the major problems associated with a system where one DNA
strand contains a lesion(s) from a bulky mutagen/carcinogen,
while the complementary strand does not, is that progeny
plasmids are preferentially derived from the strand containing
no lesion(s). This phenomenon is referred to as 'strand bias'
(2), and may also occur with simple adducts such as Me-O6Gua (3). In order to minimize the generation of progeny
895
E.L.Loechler
plasmids from the strand not containing the adduct site-specific
modification, either uracil or UV damage has been introduced
into the unmodified strand (e.g. see 4—7). Both seem to help,
although evidence suggests that progeny plasmids from the
non-adduct-containing strand are not completely eliminated
even when that strand contains a saturating dose of UV damage
(4), or a large fraction of uracils (7). The argument for using
UV damage is that it mimics having bulky DNA damage in
both strands. One unexplored issue in this regard is that MF/
adduct determined in adduct site-specific studies can be more
than an order of magnitude greater than MF/adduct estimated in
random mutagenesis studies with the corresponding mutagen.
MF/adduct of ~0.05% was estimated for 2-AF assuming
mutations from its major adduct 2-AF-C8-Gua. In an adduct
site-specific study with 2-AF-C8-Gua, MF/adduct > ~ 1 % was
obtained. MF/adduct of -0.001-0.01% for (+)-anti-BPDE can
be estimated from Rodriguez and Loechler (27), assuming
mutations from its major adduct (+)-o/iri-B[a]P-N2-Gua, which
when studied site specifically gave MF/adduct in the 0.1-1%
range (4). This may be attributable to the fact that the strategies
to eliminate strand bias may markedly enhance mutagenesis
by the adduct
A variety of means of performing adduct site-specific studies
in ss vectors have been developed, notably a very simple and
flexible system developed by Lawrence and LeClerc for
Escherichia coli (8), whose strategy was adapted for use in
some mammalian (e.g. COS) cells (9). The study of lesions
in ssDNA has proven attractive, because it eliminates the
problem of strand bias, and the observed MF is likely not to
be influenced by DNA repair, because DNA repair will either
be inefficient or not generate progeny. One potential problem
with the use of ss vectors (for others, see below) is that there
is no guarantee that the conformation of a lesion is the same
in ssDNA as it is in dsDNA, which may have an influence on
the pathway of mutagenesis when DNA polymerase forms the
ss/ds replication junction (see below). There is no evidence
for or against this possibility.
Concerns have been raised that replication of an adduct in
a ss gap may involve DNA polymerase I in E.coli (10,11),
which may not give a representative mutational response, since
DNA polymerase En is generally believed to be associated
with most adduct-directed mutagenesis. It is uncertain how
problematic this can be, but if gapped vectors are used,
then the validity of the approach should be established. For
example, it was shown that replication of Me-06-Gua in a
short gap is likely to involve DNA polymerase HI and not be
problematic (3).
Fuchs and co-workers showed that MF was ~20 times higher
when an 2-AAF-C8-Gua adduct was in the lagging rather than
in the leading strand (12). This result emphasizes that there is
nothing innate about the biological response to an adduct, and
even factors such as leading versus lagging strand DNA
synthesis, which both involve DNA polymerase Ed, may affect
the mutational response. (This is discussed at greater length
below.) In this regard, although ss <J>X174 is known to be
replicated by a process that mimics lagging strand DNA
synthesis, Ml3 (an Ff phage), which supplies the origin for
most E.coli ss vectors, is replicated by a less well-defined
mechanism (13). To the best of my knowledge nothing is
known about the replication of ss vectors in mammalian cells,
so we do not know as yet how the results from these systems
relate to replication of ds, genomic DNA. Strandedness may
also affect mutagenesis in a second way, since an adduct may
896
be repaired to a different extent depending upon whether it
is in the non-transcribed versus transcribed strand due to
'transcription coupled repair' (14), and this may affect the MF
obtained when ds vectors are used.
When doing adduct site-specific mutagenesis studies, the
adduct-containing vector must be transformed or transfected
into bacterial or mammalian cells respectively, and it is possible
that this procedure may affect the mutational patterns. For
example, the so-called Hanahan procedure of transformation
(15) employs certain metal ions (e.g. manganese), which are
known to associate with DNA polymerases and alter replication
fidelity (see 16 and refs therein). There is no experimental
evidence that the use of the Hanahan procedure is problematic,
but it is a potential concern.
When using a phage-based vector, one has to be cognizant
of the fact that during the recovery period, which is commonly
employed following transformation, phage particles may be
produced and can reinfect surrounding cells, and thereby
preferentially amplify any progeny genomes that are replicated
relatively rapidly. Thus, an argument can be made that a
recovery period following transformation should be avoided
if at all possible when using phage-based vectors (9).
Finally, it is worth mentioning that although an adduct is
replicated autonomously in an extrachromosomal genome in
most adduct site-specific studies, there are viable strategies (at
least for mammalian cells) whereby an adduct can be located
in a vector that ultimately integrates into the host cell genome,
where it is replicated (17).
The elusive mutational spectra
As discussed by Singer and Essigmann (1), the ultimate goal
of an adduct site-specific study should be to investigate whether
a particular adduct is responsible for a particular mutation
obtained in a random mutagenesis (or carcinogenesis) study.
A positive result would be a correlation, i.e. the adduct would
give a qualitative and quantitative pattern of mutation that
was consistent with what was observed with the mutagen/
carcinogen itself at the site of interest.
In the previous section, an example was cited where mutagenesis was affected quantitatively by whether the adduct was
in the leading versus lagging strand. In this section seven
additional examples are provided, illustrating that the mutational spectra of a particular mutagen/carcinogen in a random
mutagenesis study can vary depending on a variety of experimental parameters. These results (and others presented in
subsequent sections) imply both that the absence or presence
of a correlation between a random and an adduct site-specific
study may not be definitive, and that the conditions in an
adduct site-specific and a random mutagenesis study should
be matched as closely as possible to avoid gratuitous differences
(or similarities). Finally, these (and other) results show that
the relationship between adducts and mutations is complex,
and that one cannot definitively infer from a single adduct
site-specific study what the biological consequences of a
particular adduct will be in all circumstances.
An early random mutagenesis study of AFB| in E.coli
showed that GC -» TA mutations were induced almost exclusively (>90%) (18). However, GC -> AT and GC -» TA
mutations were shown to contribute approximately equally to
base substitution mutations in a subsequent study using a
different mutational system (19,20). While there are many
possible explanations for this difference, the most likely is
The role of adduct dte-spedfic mutagenests
that GC -» AT mutations are only detectable in certain
sequence contexts, which were not present in the mutational
target in the first study (discussed in greater detail in ref. 21).
This is one of many examples that illustrates that a particular
mutational spectra is not definitive, but is dependent on the
system employed, which critically depends on factors, such as
DNA sequence contexts, where mutations can be detected.
A higher fraction of mutations at A:T base pairs was
obtained at low than at high concentrations of (+)-antibenzo[a]pyrene diolepoxide (BPDE) when studied in the HPRT
gene of V-79 cells (22,23). This shows that a mutational
spectra can vary when changing as simple a parameter as dose
(i.e. adducts/genome).
The result described in the previous paragraph may relate
to a variety of factors, such as dose-dependent effects on DNA
repair, but may also be due to the induction of a bacterial-like
SOS response in mammalian cells at higher doses of (+)-antiBPDE. The SOS response in bacteria involves the induction
of a multitude of genes by DNA damage, and is generally
believed to enhance survival, although with increased mutagenesis (24). There are examples where the induction of the
SOS response is known to change the qualitative pattern of
mutagenesis, including with (+)-anti-BPDE (25), as well as
even with simple methylating agents (26). Thus, it is important
to consider that the level of induction of an SOS-like response
in bacteria, or possibly mammalian cells, might affect the
quantitative and/or qualitative pattern of mutagenesis.
The pattern of mutagenesis was shown to change both
quantitatively and qualitatively when a (-t-)-anfi-BPDEadducted plasmid was either stored at — 80°C and freeze thawed or heated at 80°C for 10 min, then cooled prior to
transformation when studied in E.coli (27). This observation
suggests that in addition to sequence context, dose, level of
SOS induction, etc., a mutational spectra—even with the same
number and type of adducts/genome—may be sensitive to other
factors, such as how the DNA is handled.
Moriya and Grollman have performed an adduct site-specific
study with several exocyclic adducts (11). When the same
adduct-containing shuttle vector was studied in monkey (COS)
versus E.coli cells, the mutations differed both quantitatively
and qualitatively. This shows that mutations from a single
adduct in the same context and in the same vector may change
when processed in different cell types. However, this is not a
universal rule, and there are other examples where disparate
cell types show remarkably similar patterns, e.g. the pattern
of mutations in the p53 gene associated with some human skin
cancers (28) is very reminiscent of UV-induced mutagenesis in
E.coli (29; discussed in ref. 30).
AT —> GC mutations predominated in the mutational spectrum of nitric oxide, which differs significantly from the
spectrum obtained with several nitric oxide-generating chemicals, where GC —» AT mutations predominated (see 31 and
refs therein). This example shows that conclusions must be
drawn cautiously when using model compounds. (A similar
concern can be expressed regarding the use of model DNA
adducts).
I wish to give one example to show that even when an adduct
studied site specifically gives a pattern of mutagenesis that
correlates with that predicted for its corresponding mutagen/
carcinogen, it does not necessarily prove a mechanistic link.
Barbacid and co-workers showed that a specific GC -» AT
mutation in the 12th codon of H-ras was reproducibly associated with methylnitrosourea (MNU)-induced rat mammary
tumorigenesis (32). The specificity of the mutation was exactly
that predicted for an MNU-based pathway involving Me-OGua as the premutagenic lesion (reviewed in 1). An adduct
site-specific study with Me-O6-Gua in a sequence context
corresponding to the 12th codon of H-ras gave G —> A
mutations, which confirmed the plausibility of this mechanism
(33), and ostensibly concluded this story. However, it has
recently been shown that this GC -» AT mutation is likely to
be a spontaneous event that in fact pre-existed in ras prior
to MNU treatment (34). Although MNU does not appear to
induce this mutation, it plays some kind of role, such as in
the selective outgrowth of the pre-existing mutants. This
demonstrates the need to undertake additional experiments
beyond correlative studies to confirm a conclusion about the
relationship between an adduct and a mutation.
The effects of DNA sequence context on adduct mutagenesis
In the previous section an example with AFB] was given of
how DNA sequence context can influence mutational spectra
in a random mutagenesis study. This theme is expanded in
this section based on several adduct site-specific studies, and
not only raises questions about how sequence context can
influence mechanism, but also emphasizes that the DNA
sequence context in an adduct site-specific study should match
that of the random mutagenesis study as closely as possible.
Quantitative effects
It is well known that DNA sequence context can affect the
quantitative pattern of mutagenesis, as this is the basis of the
familiar concept of a 'hotspot' in a mutational spectra (35).
Mutational hotspots can be due to: (i) a hotspot for adduction;
(ii) a coldspot for DNA repair; and/or (iii) a hotspot during
mutagenic processing, such as DNA replication. Fuchs and
co-workers showed that hotspot, —2 frameshift mutations in
Atari sequences (5'-G|G2CG3CC-3') induced by 2-AAF were
the result of efficient mutagenesis by 2-AAF-C8-Gua situated
at G3 (but not G| or G2) (6), and had nothing to do with G3
being a hotspot for adduction (36) or a coldspot for DNA
repair (37). This work is probably the most definitive on this
subject. However, hotspots for AFB| adduction are likely to
play a major role in its mutational hotspots given that both
occur preferentially in GC-rich regions (19,20,38). There has
been considerable interest recently in the possibility that
coldspots for DNA repair may contribute to hotspots for
mutagenesis, notably in p53 (39). An ideal study to explain
mutational hotspots should address all three factors.
Two other adduct site-specific studies showing the effect
of sequence context on the biological consequences of
DNA adducts are of particular interest. N6-Ade adducts
from styrene oxide have been placed at A2 and A3 corresponding to the 61st codon of H-ras (5'-CA2A3-3') (40). Using a
ssM13-based vector in E.coli cells, the N6-Ade adduct from
the fl-stereoisomer had a 300-fold higher plaque yield at
position A3 than A2. Interestingly, the S-stereoisomeric adduct
at A2 gave A -» G mutations (MF ~1.6%), while 5- at A3,
and R- at both A2 and A3 gave no mutations (MF <~0.1%).
DNA sequence context has also been shown to affect the
quantitative pattern of Me-O6-Gua mutagenesis using purified
DNA polymerases (discussed in ref. 1).
Seidman and co-workers have shown that a change in a
single base pair in the supF gene can depress a UV-induced
mutational hotspot up to 48 bp away when studied in xeroderma
pigmentosum cells (41). Although not an adduct site-specific
897
E.L.Loechler
study, this shows that the quantitative effects of changing
sequence context can be unexpectedly distant.
Qualitative effects
The first example that DNA sequence context can affect the
qualitative pattern of mutagenesis came from the work of
Moschel and Barbacid, who showed that C^-benzylguanine
(Bz-O6-Gua) induced G -> A mutations at G,, but G -» A, T
and C at G2 in the 12th codon of ras (5'-GiG2A-3') (33).
Since then, 2-AAF-C8-Gua has been shown to induce different
kinds of frameshift mutations depending upon DNA sequence
context (see below), and (+)-a/zf/-B[a]P-N2-Gua has been
shown to induce principally G —»T mutations in one sequence
context (5'-TGC-3') (4), but G -> T, A and C in another (5'CGG-3') (42) and principally G -> A in yet another (5'-COTS'; R.Shukla and E.L.Loechler, unpublished results). From this
it can be concluded that in many cases it is impossible to
perform a single adduct site-specific study and learn the
qualitative mutagenic potential of an adduct, since it can vary
with, for example, sequence context.
How do DNA adducts induce different mutations?
In the case of simple alkylating agents it appears that each
adduct principally induces a single kind of mutation, and that
different mutations are induced by different adducts, e.g. GC
-» AT, AT -> GC and AT -» TA mutations by Alk-O6Gua, Alk-C^-Thy and Alk-O2-Thy (probably) respectively (1).
However, in the case of bulky mutagens/carcinogens it seems
that it is more likely that a single adduct can induce multiple
mutations, e.g. as cited above for Bz-O6-Gua, 2-AAF-C8Gua and (+)-trans-anti-B[a]P-N2-Gua. Loechler (21) has a
discussion of possible mechanisms of mutation, including both
mis-informational mechanisms, such as pathways whereby the
carcinogen moiety of an adduct might induce the base moiety
of an adduct to tautomerize, ionize, rotate or wobble, as well
as non-informational mechanisms. Better evidence for the
existence of a non-informational pathway of mutagenesis with
the lesions of bulky mutagens/carcinogens is emerging (43).
One question about adducts from bulky mutagens is: by what
mechanism can a single adduct induce multiple mutations? This
is an open question, but we have proposed that a single DNA
adduct may adopt different conformations in DNA, where
each may induce a different type of mutation (27,42,44-47).
(This implies that bulky adducts must be able to form multiple
conformations, which is discussed below.) Perhaps the best
example that different adduct conformations can lead to
different biological endpoints comes from the study of oxoC8-Gua, which is a lesion formed via activated oxygen damage
in DNA (see 48, and 49 and refs therein). A cogent case has
been made that this lesion pairs with dCTP when it is in its
an/j-conformation with respect to the glycosylic bond, but
with dATP when it is in its .ryn-conformation, which would
lead to G -» T mutations.
The A-rule
It has been observed that dAMP is preferentially incorporated
opposite some lesions in certain cases, and this preference has
been called the 'A-rule' (see 50 and refs therein). The A-rule
is simply a categorization and not an explanation for a
mutational pattern. While it is important to note such tendencies, it is also important to recognize that such a rule is not
898
an explanation, and its invocation should not be considered as
being the equivalent of a proposed mechanism.
Frameshift mutagenesis
Three adduct site-specific studies have proven most revealing
regarding mechanisms of frameshift mutagenesis. In the study
mentioned above involving —2 frameshifts induced by 2AAF-C8-Gua in Narl sequences, the authors proposed that the
mutagenic mechanism is likely to be dependent upon the
action of topoisomerase and/or nuclease acting on the adduct
in this sequence, which appears capable of adopting a Z-DNAlike conformation (see 6 and refs therein). This is one of
several examples where a mutational process probably involves
more than simply DNA replication of a lesion (e.g. see 51 and
refs therein).
The qualitative pattern of mutagenesis changes to - 1
frameshifts when 2-AAF-C8-Gua is in runs of three or four
consecutive G:C base pairs (52,53). MF increased dramatically
as the adduct was moved from the 5'- to the 3'-most Gua in
the run. However, it is not a universal rule that the highest
MF will be obtained when an adduct is placed at the 3'-end
of a run. Wang and Taylor (54) studied a cis-syn thymine
dimer at all five positions in a run of six consecutive thymine
residues in vitro. Frameshift mutations only occurred at a
significant frequency (~30% and ~ 5 % for —1 and —2
frameshifts respectively) in the case of one construct: 5'TDTTT-3', where D represents the thymine dimer and is near
the 5'-end of the run. In spite of this difference, some
similar conclusions emerged about the mechanism of frameshift
mutagenesis when comparing 2-AAF-C8-Gua and D. In both
cases, a Streisinger slippage-type event in the lesion-containing
strand was favored. In both cases, the authors concluded that
the correct base must have been incorporated opposite the lesion
(i.e. dCMP opposite 2-AAF-C8-Gua and dAMPs opposite
D) prior to the slippage event in the mutagenic pathway.
Furthermore, both concluded that the slippage event must have
occurred near the 3'-end of the run. It may be that 2-AAFC8-Gua is sufficiently flexible to form a slipped intermediate
at the site of the lesion, while D is not, such that slippage
must occur at a distance.
If we accept for the moment that adduct conformational
complexity is at the root of mutational complexity, then this
raises the question: what is the relationship between the
conformations that lead to base substitution versus frameshift
mutations? We found evidence for a coupling between
frameshift and base substitution mutagenesis for (+)-antiB[a]PDE (46). The simplest rationale for this was that a single
(+)-/ra/K-a/ift-B[a]P-N2-Gua adduct in a single conformation
can potentially give rise to either a frameshift or a base
substitution mutation. In fact a more persuasive example
comes from the work of Wang and Taylor cited above (54);
remarkably, they found that the only construct that gave rise
to base substitution mutations at a significant frequency (~30%)
was also 5'-TDTTT-3'. Thus, only one of five constructs gave
misincorporations, and it gave both base substitutions and
frameshifts. This strongly suggests coupling and something
unique about the thymine dimer in the 5'-TDTTT-3' context
(perhaps conformation) that promotes both base substitution
and frameshift mutagenesis. Although no experiments have
yet proven that base substitution and frameshift mutations can
be induced from the same conformation of an adduct, it is the
most sensible possibility. Clearly, the pathway for frameshift
The role of adduct site-specific mutagenesls
versus base substitution mutagenesis must diverge, and this
has been addressed in a model (47), which suggests that the
same adduct conformation in dsDNA can give rise to either
type of mutation, and that divergence only occurs after the
formation of the replicative ss/ds junction by DNA polymerase.
Semi-targeted mutagenesis
In the absence of evidence to the contrary, it has generally
been assumed when evaluating mutational spectra that a
particular mutation was targeted to the site of an adduct.
Examples are now emerging to show that this is not always
the case. Fuchs provided the first convincing evidence for
what he calls 'semi-targeted mutagenesis', by which an adduct
induces a mutation near but not at the site of the lesion (52).
2-AAF-C8-Gua at position G, or G2 (but not G3) in the
sequence context 5'-CCCGiG2G3-3' preferentially induced a
- 1 deletion in the run of three Cyts. This result has no simple,
obvious mechanistic explanation. Approximately 25% of the
mutations induced by (+)-fra/w-anri-B[a]P-N2-Gua situated at
G, in one particular sequence context (5'-CG|G 2 C-3') were
semi-targeted G2 —» A mutations (42). Essigmann and coworkers have preliminary evidence for significant semitargeted mutagenesis occurring with an AFB,-N7-Gua adduct
(J.Essigmann, personal communication).
There are examples in the literature where the mutational
spectra of bulky mutagens/carcinogens show tandem mutations,
and it has been argued that these events must result from
coupled targeted and semi-targeted events during the replication
of a single adduct. AFBi-oxide was shown to induce a high
fraction (~13%) of tandem double mutations in xeroderma
pigmentosum cells (55). With (+)-cnfi-benzo[aj]anthracene
diol epoxide, the major mutations were G —> A (~36%), but
tandem GG —» AA mutations comprised ~15% of the spectrum
in E.coli (56). Finally, both B[a]P and 4-aminobiphenylinduced frameshift mutations were studied in the hisD3052
allele of the Salmonella typhimurium TA98 Ames tester strain
(57,58). Under certain conditions as much as 23% of the
mutants were complex and involved both base substitution
and frameshift mutations.
The role of in vitro studies using adduct site-specifically
modified templates
Although this is by no means a consensus opinion, I believe
that the study of the adducts of bulky mutagens/carcinogens
using purified DNA polymerases in vitro has generally been
less revealing about mechanisms of mutagenesis. The main
reason for concern is that in most cases the DNA polymerase
employed either is not representative (e.g. DNA polymerase I
from E.coli, which is a repair polymerase), or is without
important host factors (e.g. the SOS mutagenesis factors UmuD
and C, which are currently unavailable for study, should
probably be used with DNA polymerase HI from E.coli
in most cases). In many cases, results obtained in vitro
significantly deviate from the results obtained in cells, and one
can have no confidence that a finding in vitro will necessarily
be reflective of anything with a true replicative complex in
cells. P hasten to add that the study of non-bulky lesions (e.g.
Me-O6-Gua) using purified DNA polymerases appears to me
to be less problematic] A second problem is that the adducts of
bulky mutagens/carcinogens principally block DNA replication
in vitro, and bypass is usually an infrequent event. Thus, one
has to be concerned that the small fraction of bypass observed
is really due to the lesion of interest and not a minor
contaminant. There are other problems as well.
In spite of these concerns, there have been several revealing
in vitro studies, such as the thymine dimer study of Wang and
Taylor (54; see above). Furthermore, some significant in vitro
studies have been done with both T7 DNA polymerase and
E.coli DNA polymerase m, which are both true replicative
enzymes. For example, Dipple and co-workers have shown
that a modified T7 DNA polymerase incorporates dAMP
preferentially opposite both N2-Gua and N^Ade adducts of
7-bromomethylbenz[a]anthracene, which would lead to the
formation of the most frequently observed mutations with
bulky mutagens/carcinogens, namely GC —> TA and AT -»
TA transversions (59). Fuchs and co-workers did a singleturnover kinetic analysis of the bypass of both 2-AF- and 2AAF-C8-Gua adducts using T7 DNA polymerase (60). dCTP
was preferentially incorporated opposite the adduct at a reduced
rate (-lO 3 - and 107-fold slower respectively than with no
adduct); dATP could also be incorporated. These workers have
also been studying replication of 2-AF- and 2-AAF-C8-Gua
by DNA polymerase HI (61).
Relating structural studies using adduct site-specifically
modified oligonucleotides to mutational studies
Just as adduct site-specifically synthesized oligonucleotides
have revolutionized the study of mutations from adducts, they
have also revolutionized the study of the structure of these
same adducts. I shall not review this literature, but I do wish
to reflect on several issues about how structural and mutational
studies interrelate.
Physical studies are beginning to show that the adducts of
bulky mutagens/carcinogens can adopt multiple conformations
in DNA. Evidence for multiple conformations for a single
(+)-fra/w-anf/-B[a]P-N2-Gua adduct was obtained in several
fluorescence studies (62-64), although the actual conformations
themselves cannot be determined using this approach. Based
on NMR studies, evidence for multiple conformations in the
same oligonucleotide was obtained for (+)-trans-anti-B[a]PN2-Gua (65), (+)-c«-anri-B[a]P-N2-Gua (66) and 2-AAF-C8Gua (67). Unfortunately, only the spectrum in the latter study
was sufficiently resolved to permit the assignment of both
conformations; one had the 2-AF moiety in the major groove,
while the second was a base-displacement type structure.
If adduct conformational complexity proves to be important
to mutagenesis, then it may be confounding, because it will
be difficult to assign a particular biological result to a particular
conformation. In this regard, it is worth reflecting on the
biological consequences of having bulky lesions in DNA. This
varies depending on the adduct, sequence context, cell type,
etc.; however, the following is my view of a composite for a
typical bulky DNA adduct based upon data from studies where
DNA repair is unlikely to be significant. (More details in this
regard can be found in ref. 21.) The most significant consequence of having a bulky adduct in a vector genome is the
blocking of DNA replication as inferred from an adductdependent decrease in the yield of progeny vectors. Typically
progeny yield is reduced ~2- to- 4-fold when compared to the
progeny yield of the same vector with no adduct. The secondmost significant event is non-mutagenic bypass of the adduct.
When progeny vectors that do arise are analyzed, >90% of
the time no mutations are observed at or near the original
genome location of the bulky adduct. While it might be argued
899
EX.Loechler
that the high yield of non-mutant progeny vectors is due to
DNA repair of lesions when located in dsDNA vectors, this
seems unlikely to be the case when lesions are in ssDNA
vectors (see below). It is more likely that faithful replication
is the consequence of the fact that the adducts of most bulky
mutagens/carcinogens do not disrupt the hydrogen bonding
face of the base moiety of the adduct, such that proper
Watson—Crick base pairing is possible—at least when the
adduct is in some conformations. [This is obviously not the
only possibility given that certain exocyclic adducts, such as
ethenoadenine and ethenocytosine, which each have their
hydrogen bonding face obliterated, can be replicated correctly
a majority of the time in many (but not all) circumstances
(10,11).] Quantitatively, the least most significant biological
event is mutagenesis, which occurs typically in <10% of
progeny vectors. For example, MF <0.1% was obtained for
the 5'-G-G-3' intrastrand cross-links of cis-DDP when studied
in a ssDNA vector in E.coli; in fact, detectable mutagenesis
was only observed at the base on the 5'-side following SOS
induction (MF -1-3%) (68). Furthermore, both N6-Ade and
N2-Gua adducts of polycyclic aromatic hydrocarbons have
given MF <~2% when studied in ss vectors in E.coli
(40,43,69), or when studied in a ds vector in repair-deficient
E.coli (4).
Given that mutagenesis is an infrequent event, it is entirely
possible that it may result from the replication of an adduct
in a minor conformation. If true, this will significantly complicate the effort to determine what conformation is responsible
for what mutation.
Finally, it is worth mentioning that the relationship between
structure and mutation is likely to be complex, since it involves
a ss/ds replicative junction and a DNA polymerase, which
must play a role in defining the mutations beyond simply
being a vehicle for forming hydrogen bonds for many reasons,
including the results cited above concerning the faithful replication of exocyclic adducts. Furthermore, if indeed the catalytic
step involving misincorporation is the rate-determining step
during mutagenic bypass, which is sensible and has been
shown to be likely in several cases (60,61), then the critical
structure for mutagenesis is actually that of a transition state.
Transition states have historically been difficult to study;
however, the use of structure—activity relationships represent
one classic method that has proven valuable in the past. There
is at least one group that is taking this challenging approach
in the hopes of understanding the mechanism of Me-O6-Gua
mutagenesis (70).
Conclusion
The simplest hypothesis for adduct mutagenesis, which had
been assumed implicitly for many years principally based upon
the prototypical lesion Me-O6-Gua, was that an adduct was
responsible for a single kind of mutation and had a single
relevant structure and conformation. Given this premise, a
single adduct site-specific study would be sufficient to define
the role that a particular adduct played in the mutagenic/
carcinogenic process. As elaborated above, the situation now
appears to be considerably more complex, and the quantitative
and qualitative pattern of mutagenesis for an adduct is potentially quite variable and changeable, and can be influenced by
factors as diverse as cell type, DNA sequence context and
even solute conditions. If this adduct mutational complexity
is due to adduct conformational complexity, then perhaps
900
methods for 'conformational, adduct site-specific' mutational
studies may have to be developed. In lieu of this, it may be
that certain mutations will dominate in certain sequence
contexts, which can be used to unravel the mechanism of a
particular mutation. Finally, another big challenge for elaborating mechanism will be understanding the role of DNA polymerase in defining adduct mutagenesis.
In the long term, I am confident that there will come a time
when we will be able to provide more than a simple catalog
of the mutations induced by a particular agent or adduct in a
particular set of conditions. I believe that once an understanding
of the mechanisms of the process has emerged, then we will
be able to understand the patterns of mutations induced by
carcinogens based upon more fundamental principles.
Acknowledgements
I thank Larry Marnett and John Essigmann for communicating unpublished
results. I am grateful to Anthony Dipple, John Essigmann, Scott Jelinsky,
Rajiv Shukla and the reviewer for numerous helpful comments. E.L.L. was
supported by grants from the NIH (ES03775, CA50432) while working on
this document.
Note added in proof
In retrospect, I have not sufficiently emphasized the need to purify, as well
as characterize, adducted oligonucleotides prior to their incorporation into a
plasmid. For example, we purify our unadducted oligonucleotide by HPLC
prior to adduction, and then purify the adducted oligonucleotide by HPLC, as
well as both by denaturing and native polyacrylamide gel elctrophoresis after
adduction (42). At least in some cases, we have found that all of these steps
must be taken in order to insure punty. Our method to convince ourselves
that our adducted oligonucleotide is >99% pure is described in Figure 2 of
reference 42.
References
l.Singer.B. and EssigmannJ.M. (1991) Site-specific mutagenesis:
retrospective and prospective. Carcinogenesis, 12, 949-955.
2. Koffel-Schwartz.N., Maenhaut-Michel.G. and Fuchs.R.P.P. (1987) Specific
strand loss in JV-2-acetylaminofluorene-modified DNA. J. Mol. Biol*, 193,
651-659.
3. Pauly,G.T., Hughes.S.H. and Moschel.R.C. (1995) Mutagenesis in
Escherichia coli by three C^-substituted guanine residues in doublestranded or gapped plasmids. Biochemistry, 34, 8924-8930.
4. Mackay,W., Benasutti,M., Drouin.E. and Loechler,E.L (1992) Mutagenesis
by the major adduct of activated benzo[a]pyrene, (+)-anri-BP-N2-Gua,
when studied in an Escherichia coli plasmid using site-directed methods.
Carcinogenesis, 13, 1415-1425.
5.Reid,T.M., Lee,M.-S. and King,C.M. (1990) Mutagenesis by site-specific
arylamine adducts in plasmid DNA: enhancing replication of the adducted
strand alters mutation frequency. Biochemistry, 29, 6153—6161.
6. Burnouf,D., KoehlJ". and Fuchs,R.P.P. (1989) Single adduct mutagenesis:
strong effect of the position of a single acetylaminofluorene adduct within
a mutation hot spot. Proc. Natl Acad. Sci. USA, 86, 4147-4151.
7.Fmk,S.P., Reddy.R. and Mamett,LJ. (1996) Relative contribution of
cytosine deamination and error-prone replication to the propanodeoxyguanosine —> deoxyadenosine mutation in Escherichia coli. Chem.
Res. Toxicol., 9, 277-283.
8. Banerjee.S.K., Christensen.R.B., Lawrence.C.W. and LeClercJ.E. (1988)
Frequency and spectrum of mutations produced by a single cis-syn
thymine-thymine cyclobutane dimer in a single-stranded vector. Proc.
Natl Acad. Sci. USA, 85, 8141-8145.
9. Moriya,M. (1993) Single-stranded shuttle phagemid for mutagenesis studies
in mammalian cells: 8-oxoguanine in DNA induces targeted GC —> TA
transversion in simian kidney cells. Proc. Natl Acad. Sci. USA, 90,
1122-1126.
10. Basu^A.K., Wbod.M.L., Niedemhofer.LJ., Ramos.L.A. and Essigmann,
J.M. (1993) Mutagenic and genotoxic effects of three vinyl chlorideinduced DNA lesions: 1 JV'-ethenoadenine. 3-A^-ethenocytosine and 4amino-5-{imidazole-2-yl)imidazole. Biochemistry, 32, 12793-12801.
ll.Moriya.M., Zhang.W., Johnson,F. and Grollman.A.P. (1994) Mutagenic
potency of exocyclic DNA adducts: marked differences between
The role of adduct site-spedfic mutagenesis
Escherichia coli and simian kidney cells. Proc. Natl Acad. Sci. USA, 91,
11899-11903.
12.Veaute.X. and Fuchs,R.P.P. (1993) Greater susceptibiltiy to mutations in
lagging strand of DNA replication in Escherichia coli than in leading
strand. Science, 261, 598-600.
13.Komberg,A. (1992) DNA replication. W.H.Freeman, New York, pp. 557586.
14. Hanawalt,P.C. (1994) Transcription-coupled repair and human disease.
Science, 266, 1957-1958.
15.Hanahan,D. (1983) Studies of transformation of Escherichia coli with
plasmids. J. Mol. Biol., 166, 557-580.
16. Downey.K.M. and So,AG. (1989) Metal ion-induced mutagenesis in vitro:
molecular mechanisms. In Sigel.H. (ed.). Metal Ions in Biological Systems.
Marcel Dekker, New York, Vol. 25, pp. 1-30.
17.Ellison,K.S., Dogliotti.E., Connors.T.D., Basu^A.K. and EssigmannJ.M.
(1989) Sue-specific mutagenesis by O6-alkylguanines located in the
chromosomes of mammalian cells: influence of the mammalian O6alkylguanine alkyltransferase. Proc. Natl Acad. Sci. USA, 86, 8620-8624.
18.Foster,P.L., EisenstadtJS. and MillerJ.H. (1983) Base substitution mutations
induced by metabolically activated AFB,. Proc. Natl Acad. Sci. USA, 80,
2695-2698.
19.Sahasrabudhe,S.R., Sambamurti.K. and Humayun.M.Z. (1989) Basesubstitution mechanisms and the origin of strand bias. Mol. Gen. Genet.,
217, 20-25.
20. Sambamurti.K., CallahanJ., Perkins.C.P., JacobsonJ.S and Humayun.M.Z.
(1988) Mechanisms of mutagenesis by bulky DNA lesions at the guanine
N7 position. Genetics, 120, 863-873.
21.Loechler,E.L. (1994) Mechanism by which aflatoxins and other bulky
carcinogens induce mutations. In Eaton and Groopman (eds). The
Toxicology of Aflatoxins: Human Health, Veterinary, and Agricultural
Significance. Academic Press, Orlando, Chap. 8, pp. 149—178.
22.Wei,S.-J.C, Chang.R.L., Wong,C.-Q., Bhachech,N., Cui.X.X., Hennig.E.,
Yagi.H., SayerJ.M., Jerina,D.M., Prcston,B.D. and Conney.A.H. (1991)
Dose-dependent differences in the profile of mutations induced by an
ultimate carcinogen from benzo[a]pyrene Proc. Natl Acad. Sci. USA, 88,
11227-11230.
23.Wei,S.-J.C, Chang.R.L., Bhachech.N., Cui,X.X., Merkler.K.A., Wong,
C.Q., Hennig.E., Yagi.H., Jerina,D M. and Conney,A.H. (1993) Dosedependent differences in the profile of mutations induced by (+)-7/?,85dihydroxy-95,10/f-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene in the coding
region of the hypoxanthine (guanine) phosphoribosyltransferase gene in
Chinese hamster V79 cells. Cancer Res., 53, 3294-3301.
24. Friedberg.E.C, Walker,G.W. and Siede.W. (1995) DNA Repair and
Mutagenesis. American Society for Microbiology, Washington, DC,
pp. 407-464.
25.Rodriguez,H. and Loechler,E.L. (1993) Mutational specificity of the (+)anf/-diol epoxide of benzo[a]pyrene in a supF gene of an Escherichia coli
plasmid: DNA sequence context influences hotspots, mutagenic specificity
and the extent of SOS enhancement of mutagenesis. Carcinogenesis, 14,
373-383.
26. Chaudhuri.l. and EssigmannJ.M. (1991) 3-Methyladenine mutagenesis
under conditions of SOS induction in Escherichia coli. Carcinogenesis,
12, 2283-2289.
27. Rodriguez.H. and Loechler,E.L. (1993) Mutagenesis by the (+)-antidiol epoxide of benzo[a]pyrene: what controls mutagenic specificity?
Biochemistry, 32, 1759-1769.
28. Brash.D.E. el at. (1991) A role for sunlight in skin cancer UV-induced
p53 mutations in squamous cell carcinoma. Proc. Natl Acad. Sci. USA,
88, 10124.
29.MillerJ.H. (1985) Mutagenic specificity of ultraviolet light. J. Mol. Biol.,
182, 45-68.
30.Greenblatt,M.S., Bennett.W.P., Hollstein,M. and Harris.C.C. (1994)
Mutations in the p53 tumor suppressor gene: clues to cancer etiology.
Cancer Res., 54, 4855-4878.
3I.Routledge,M.N., Wink.D.A., Keefer.L.K. and DippleA (1994) DNA
sequence changes induced by two nitric oxide donor drugs in the supF
assay. Chem. Res. Toxicol., 7, 628-632.
32.Sukumar,S., Notariom.V., Martin-Zanca,D. and Barbacidjvl. (1983)
Induction of mammary carcinomas in rats by nitroso-methylurea involves
malignant activation of the H-ras-\2 locus by single point mutations.
Nature, 306, 658-661.
33.Mitra,G., Pauly.G.T, Kumar.R., Pei.G.K., Hughes.S.H., Moschel.R.C. and
Barbacid.M. (1989) Molecular analysis of O6-substituted guanine-induced
mutagenesis of ras oncogenes. Proc. Natl Acad. Sci. USA, 86, 8650-8654.
34.Cha,R.S., Thilly.W.G. and Zarbl.H. (1994) N-Nitroso-jV-methylurcainduced rat mammary tumors arise from cells with preexisting oncogenic
Hrasl gene mutations. Proc. Natl Acad. Sci. USA, 91, 3749.
35.MillerJ.H. (1978) The lac\ gene: its role in lac operon control and its use
as a genetic system. In MillerJ.H. and Reznikoff.W.S. (eds), The Operon.
Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, pp. 31-88.
36. Fuchs.R.P.P. (1985) DNA binding spectrum of the carcinogen W-acetoxyN-2-acetylaminofluorene significantly differs from the mutation spectrum.
J. Mol. Biol., 177, 173-180.
37. Seeberg,E. and Fuchs.R.P.P. (1990) Acetylaminofluorene bound to different
guanines of the sequence -GGCGCC- is excised with different efficiencies
by the UvrABC excision nuclease in a pattern not correlated to the potency
of mutation induction. Proc. Natl Acad. Sci. USA, 87, 191-194.
38.Benasutti,M., Ejadi.S., Whitlow.M.D. and LoechlerJE.L. (1988) Mapping
the binding site of aflatoxin B| in DNA: systematic analysis of the
reactivity of aflatoxin Bi with guanines in different DNA sequences.
Biochemistry, 27, 472-481.
39.TomaIetti,S. and Pfeifer.G.P. (1994) Slow repair of pyrimidine dimers at
the p53 mutation hotspots in skin cancer. Science, 263, 1436-1438.
4O.Latham,GJ., Zhou.L., Hams.C.M., Harris.T.M. and Lloyd,R.S. (1993) The
replication fate of R- and S-styrene oxide adducts on adenine N6 is
dependent on both chirality of the lesion and the local sequence context
J. Biol. Chem., 268, 23427-23434.
41.Parris,C.N., Levy.D.D., JesseeJ. and SeidmanJvl.M. (1994) Proximal and
distal effects of sequence context on ultraviolet mutational hotspots in a
shuttle vector replicated in xeroderma cells. J. Mol. Biol., 236, 491-502.
42.Jelinsky,S.A., Mao.B., Geacintov.N.E. and LoechlerJS.L. (1995) The major,
N2-Gua adduct of the (+)-ann-benzo[a]pyrene diol epoxide is capable of
inducing G —> A and G —» C, in addition to G —> T, mutations.
Biochemistry, 34, 13545-13553.
43. Min.Z., Gill.R.G., Cortez.C, Harvey,R.G., Loechler.E.L. and DiGiovanni J.
(1996) Targeted A —> T, and G -» T mutations induced by site-specific
deoxyadenosine and deoxyguanosine adducts, respectively, from the (+)onri'-diol-epoxide of dibenztajjanthracene in M13mp7L2. Biochemistry,
in press.
44. Drouin.E. and Loechler.E.L. (1993) Mutagenesis by the (+)-anti diol
epoxide benzo[a]pyrene does not significantly involve AP-sites: evidence
that the complexity of the mutational spectra is due to adduct
conformational polymorphism. Biochemistry, 32, 6555-6562.
45. Drouin.E. and Loechler.E.L. (1995) The major adduct of the (+)-anti diol
epoxide of benzo[a]pyrene can be unstable in double stranded DNA.
Biochemistry, 34, 2251-2259.
46. Rodriguez.H. and Loechler.E.L. (1995) Are base substitution and frameshift
mutagenesis pathways interrelated? An analysis based upon studies of
mutagenesis by the (+}-anti diol epoxide of benzo[a]pyrene. Mutat. Res.,
326, 29-37.
47. Loechler.E.L. (1995) How are potent bulky carcinogens able to induce
such a diverse array of mutations? Mol. Carcinogenesis, 13, 213-219.
48.Wood,M.L., Dizdaroglu.M., Gajewski.E. and EssigmannJ.M. (1990)
Mechanistic studies of ionizing radiation and oxidative mutagenesis:
genetic effects of a single 8-hydroxyguanine (7-hydro-8-oxoguanine)
residue inserted at a unique site in a viral genome. Biochemistry, 29,
7024-7032.
49.Shibutani,S., Bodepudi.V., Johnson.F. and Grollman.A.P. (1993)
Translesion synthesis on DNA templates containing 8-oxo-7,8dihydrodeoxyadenosine. Biochemistry, 32, 4615—4621.
50.Larson,K.L. and Strauss.B.S. (1987) Influence of template strandedness
on in vitro replication of mutagen-damaged DNA. Biochemistry, lib,
2A1X-2419.
51. Kaiser,V.L. and Ripley.L.S. (1995) DNA nick processing by exonuclease
and polymerase activities of bacteriophage T4 DNA polymerase accounts
for acridine-induced mutation specificities in T4. Proc. Natl Acad. Sci.
USA, 92, 2234-2238.
52.Lambert,I.B., Napolitano.R.L. and Fuchs.R.P.P. (1992) Carcinogen-induced
frameshift mutagenesis in repetitive sequences. Proc. Natl Acad. Set USA,
89, 1310-1314.
53. Napolitano.R.L., Lambert,I.B. and Fuchs.R.P.P. (1994) DNA sequence
determinants of carcinogen-induced frameshift mutagenesis. Biochemistry,
33, 1311-1315.
54.Wang,C.-I. and TaylorJ.S. (1992) In vitro evidence that UV-induced
frameshift and substitution mutations at T tracts are the result of
misalignment-mediated replication past a specific thymine dimer.
Biochemistry, 31, 3671-3681.
55. Levy,D.D., GroopmanJ.D., Lim,S.E., SeidmanJvl.M. and Kraemer.K.H.
(1992) Sequence specificity of aflatoxin B1 -induced mutations in a plasmid
replicated in xeroderma pigmentosum and DNA repair proficient human
cells. Cancer Res., 52, 5668-5673.
56.Gill,R.D., Rodriguez.H., Cortez,C, Harvey.R.G., Loechler^.L. and
DiGiovanniJ. (1993) Mutagenic specificity of the (+)-an/i-diol epoxide
901
EX.Loechler
of dibenz[a,/]anthracene in the supF gene of an Escherichia coli plasmid
Mol. Carcinogenesis, 8, 145—154.
57.DiMarini,D.M., Shelton.M.L. and Bell.D.A. (1994) Mutation spectra in
Salmonella of complex mixtures: comparisons of urban air to
benzo[a]pyrene. Environ. Mol. Mutagenesis, 24, 262-275.
58. LevineJ.G., Schaaper,R.M. and DiMarini.D.M. (1994) Complex frameshift
mutations mediated by plasmid pKMlOl: mutational mechanisms deduced
from 4-aminobiphenyl-induced mutation spectra in Salmonella. Genetics,
136, 731-746.
59.Reardon,D.B., Bigger.C.A. and Dipple.A. (1990) DNA polymerase action
on bulky deoxyguanosine and deoxyadenosine adducts. Carcinogenesis,
11, 165-168.
60. LindsleyJ.E. and Fuchs.R.P.P. (1994) Use of single-turnover kinetics to
study bulky adducts bypass by T7 DNA polymerase. Biochemistry, 33,
764-772.
61.Belguise-Valladier,R, Maki.H., Sekiguchi.M. and Fuchs.R.P.P. (1994)
Effect of single DNA lesions in in vitro replication with DNA polymerase
III holoenzyme. J. Mol. Biol., 236, 151-164.
62.Suh,M., Jankowiak,R., Ariese.R, Mao.B., GeacintovJvl.E. and Small.G.J.
(1994) Flanking base effects on the structural conformation of the (+)iranj-am/-benzo[a]pyrene diol epoxide adduct to N-2-dG in sequencedefined oligonucleotides Carcinogenesis, 15, 2891-2898.
63.Marsch,G.A., Jankowiak.R., Suh.M. and Small.GJ. (1994) Sequence
dependence of benzo[a]pyrene diol epoxide-DNA adduct conformation
distribution: a study by laser-induced fluorescence/polyacrylamide gel
electrophoresis. Chem. Res. Toxicol, 7, 98-109.
64. LeBreton.P.E., Huang,C.-R., Femando,H., Zajc.B., Lakshman.M.K.,
SayerJ.M. and Jerina,D.M. (1995) Multiple fluorescence lifetimes for
oligonucleotides containing single, site-specific modifications of guanines
and adenines corresponding to trans addition of the exocyclic amino
groups to (+H7R,8S,9S,\0R)- and (-)-(7S,8R,9R, 105>7,8-dihydroxy9,10-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene. Chem. Res. Toxicol., 8,
338-348.
65. Fountain^.A. and Krugh.T.R. (1995) Structural characterization of a (+)/ra/Ls-a/ir<-benzo[a]pyrene —DNA adduct using NMR, restrained energy
minimization and molecular dynamics. Biochemistry, 34, 3152-3161.
66.Cosman,M., de los Santos,C, Fiala,R., Hingerty.B.E., Ibanez.V, Luna.E.,
Harvey.R., Geacintov.N.E, Broyde,S. and Patel.D.J. (1993) Solution
conformation of the (+)-cis-anli-[BP]dG adduct in a DNA duplex:
intercalation of the covalently attached benzo[a]pyrenyl nng into the helix
and displacement of the modified deoxyguanosine. Biochemistry, 32,
4145-4155.
67.Eckel,L.M. and Krugh.T.R. (1994) Structural characterization of two
interchangeable conformations of a 2-aminofluorene-modified DNA
oligomer by NMR and energy minimization. Biochemistry, 33, 1361113624.
68.Naser,L.J.N., Yarema,K.J., Lippard.SJ. and EssigmannJ.M. (1993)
Mutagenicity and genotoxicity of the major DNA adduct of the antitumor
drug cii-diamminedichloroplatinum(II). Biochemistry, 32, 982-988.
69.Chary,P., Latham.GJ., Robberson.D.L., Kim.SJ., Harns.C.M., Harris.T.M.
and Lloyd.R.S. (1995) In vivo and in vitro replication consequences of
stereoisomeric benzo[a]pyrene-7,8-dihydrodiol 9,10-epoxide adducts on
adenine N* at the second position of N-raj codon 61.7. Bio/. Chem., 270,
4990-5000.
70. Spratt.T.E. and Campbell.C.R. (1995) Hydrogen bonding interaction
between O6-methylgiianine and cytosine or thymine during DNA
replication. Proc. Am. Assoc. Cancer Res., 36, 161.
Received on August 25, 1995; revised on January 26, 1996; accepted on
January 30, 1996
902