Proc. NatL Acad. Sci.

USA
Vol. 80, pp. 2695-2698, May 1983
Genetics

Base substitution mutations induced by metabolically activated

aflatoxin B1
(mutagenesis/carcinogenesis/lac operon/Escherichia coli/SOS response)

P. L. FOSTER*, E. EISENSTADTt,

AND

J. H. MILLERt

*Interdiscjplinary Programs in Health and tDepartment of Cancer Biology and Laboratory of Toxicology, Harvard School of Public Health, 665 Huntington Avenue,
Boston, Massachusetts 02115; and *Departement de Biologie Moleculaire, UniversitM de Geneve, Geneva, Switzerland

Communicated by Eugene P. Kennedy, January 31, 1983

We have determined the base substitutions generated by metabolically activated aflatoxin B1 in the lad gene of
a uvrB- strain of Eacherichia coli. By monitoring over 70 different nonsense mutation sites, we show that activated aflatoxin B1

ABSTRACT

TA transversions. One possible pathspecifically induced G-C

way leading to this base change involves depurination at guanine
-*

residues. We consider this mechanism of mutagenesis in the light

of our other findings that the carcinogens benzo[a]pyrene diol
epoxide and N-acetoxyacetylaminofluorene also specifically induce G-C
TA transversions.
-+

Knowledge of the mutational specificity of a mutagen, combined with an understanding of the biochemistry of its DNA
reactions, is crucial to the identification of its important premutational lesions. For this reason, we have examined the
spectrum of base substitution mutations generated in the lacl
gene of Escherichia coli by the carcinogen aflatoxin B1 (AFB1).
AFB1 is a potent animal hepatocarcinogen (1), and epidemiological evidence correlates aflatoxin consumption with human
liver cancer in Asia and Africa (2). AFB1 is activated by liver
microsomal enzymes to the highly reactive 2,3-oxide (3, 4), which
forms covalent adducts in vivo and in vitro to the N-7 position
of guanine in DNA (5-8). Can this guanine N-7 adduct, which
represents more than 90% of the AFB1 bound to DNA (9), account for the mutagenic specificity of AFB1?
The lacI system consists of >70 characterized amber and ochre
sites that can be used to detect all possible base changes except
show here
AT -* GC transitions (10). Using this system,
that AFB1-induced base substitution mutations arise almost exTA transversions. We consider mechaclusively by G-C
nisms that can account for this specificity and for its similarity
to the mutagenic specificities of the carcinogens benzo[a]pyrene (BaP) diol epoxide (11) and N-acetoxyacetylaminofluorene
(N-acetoxy-AAF) (unpublished data), both of which also show
preference for inducing base substitutions by G-C -* TA
we

-*

transversions.

MATERIALS AND METHODS

Bacterial Strains. PF101, the E. coli strain used here for AFB1
mutagenesis, was constructed by first transducing the uvrB5
allele (12) linked to a TnlO insertion in chlA into the F' lac+
pro+ strain GM1 (13). This strain was then transformed with
DNA from plasmid pGW270, which retains the mutagenesisenhancing muc loci but not the conjugal functions nor the growthretarding slo gene of plasmid pKM101 (14, 15). Thus, PF101
carries the lac operon on a F' episome, is excision repair defective (uvrB-), and has a slightly increased (2-fold) spontaneous mutation rate and greatly increased (10- to 20-fold) in-

duced mutation rates (muc'). Bacteriophage P1 transduction

(16), transformation (17), and preparation of plasmid DNA (18)
were as described. The set of E. coli strains for identifying lad
nonsense mutants has been described in detail (13).
Media. Selective and indicator plates (13, 16), the minimal
medium used to culture bacteria for mutagenesis (11), and the
metabolic activation buffer (19) were prepared as described.
Chemicals. AFB1 was obtained from Sigma (lot. 89C4015),
dissolved in methylene chloride, distributed in conveniently
sized aliquots, evaporated under N2 gas, and stored at -20'C.
For mutagenesis, aliquots were dissolved in dimethyl sulfoxide.
Aroclor-induced rat liver microsomes were obtained from Litton Bionetics (lot. REK085) and stored at -70C.
Mutagenesis. Bacteria were grown in minimal medium to a
density of 2.5 x 108 cells per ml, washed once with E salts, and
concentrated 10-fold by centrifugation. The cell suspension was
then diluted 1:6 into the metabolic activation buffer containing
rat liver microsomes (60 ,ul/ml) and exposed to 180 ,uM AFB1
for 60 min at 37C, which resulted in 19% cell survival. Cells
were then washed once by centrifugation, the cell suspension
was diluted 1:8 into E salts, and 0.1 ml (containing 5 x 106 cells)
was added to 110 tubes containing 10 ml each of minimal medium. After incubation at 37C overnight (to a density of 1.5
109 cells per ml), Lac constitutive mutations were selected
from these cultures as described (13). Eventually, 193 independent nonsense mutations in the lad gene (95 ambers and
98 ochres) were identified.
RESULTS
Activated AFB1 is a potent base substitution mutagen for Salmonella typhimurium strains that are excision repair defective
and that bear the mutagenesis-enhancing plasmid pKM101 (20).
To recover AFB1-induced mutations above background in E.
coli, it was also necessary to introduce not only a defect in the
excision-repair pathway but also plasmid pKM101 or its derivative plasmid pGW270, both of which carry the muc+ loci (14,
15). After strain PF101 was treated with AFB1 and rat liver microsomes, Lac constitutive mutations occurred at a frequency
of 10-4, which was 10-fold more than the mutation frequency
of the untreated controls exposed to microsomes but not to AFB1.
Of the induced mutants, 16% carried amber or ochre mutations
in the lad gene, whereas only 3% of the spontaneous mutations
resulted from nonsense mutations in this gene. Thus, AFB1
generated a 50-fold increase in lad nonsense mutations compared with the level observed in the untreated cells.
Fig. 1 shows the striking distribution of nonsense mutations
induced by AFB1 in the lad gene-close to 90% arose through
X

Abbreviations: AFB1, aflatoxin B1; BaP, benzo[a]pyrene; N-acetoxy-AAF,

N-acetoxyacetylaminofluorene.
Current address: Dept. of Biology, Univ. of California, Los Angeles,

The publication costs of this article were defrayed in part by page charge
payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. 1734 solely to indicate this fact.

CA 90024.
2695

Genetics: Foster et aL

2696

10

Proc. Natl. Acad. Sci. USA 80 (1983)

Table 1. Summary of base substitution mutations in the lad
gene induced by metabolically activated AFB1
% of
Available
Sites
Total
analyzed
Substitution
sites
found
occurrences
mutations
G-C A-T
26
10
12
6
G-C TEA
23
22
169
89
A-T T-A
15
3
5
3
A-T C-G
5
1
1
0.5
3
0
0
GC-C-G
0

G'C- A*T

-rI

II

G C-. T A

20

Total sites
73
37
189
100
Ambers
37
22
92
49
Ochres
36
15
97
51
The total includes two mutations at 06, which may be either an A-T
-* T-A site or a G-C --*COG site.

15
10

C-)

cr
C-)

H
0

tlIl

III[........II.
I
I
I

. . .

other than those for G&C -+ T A transversions and (ii) with the
two best-induced sites (03 and A20) excluded, the distribution
among only the sites for G-C -3 TEA transversions fit random
distributions (P = 0.75-0.90 for each).
At three sites we can monitor both G&C -- TEA and G{-G
C-G transversions at the same base pair; the former yields TAA
(ochre) and the latter TAG (amber) mutations at the third position of TAC tyrosine codons. G:C -+ T:A transversions occurred strikingly more frequently at these three sites than did
G-C -G C-G transversions (P < 0.005; Table 3).

.-

0
LL

10

G'C- C G

cUco
5
m

10

AT-wT'A

DISCUSSION
5

0
10

,I

11.,

II

~iu-

Sixteen percent of the Lac constitutive mutants -generated by

AFB1 resulted from amber and ochre mutations in the lad gene.
Thus, base substitutions probably represent a sizable fraction
of the mutations induced by this carcinogen in E. coli bearing
the muc' loci. AFB1 appeared to be able to induce nonsense
mutations only through G'C TEA transversions; for example,
even at the same sites, G-C
T A transversions were greatly
preferred over G'C -) CG transversions (Table 3). However,
with the exception of two "hotspots, " 03 and A20, AFBL seemed
to show little preference among the remaining 21 G'C -- TEA
transversion sites at the level of resolution of a collection of this

III

A.T-0-C.G

0
0

I.

100

200

300

size.

AMINO ACID RESIDUE

FIG. 1. The frequencies of mutations across the lad gene. Since

amber and ochre mutations occurred with nearly equal frequencies after
AFB1 treatment and approximately equal numbers of each were analyzed, their occurrences have been combined without normalization.
Mutations at 06, X5, and X6 have not been included (see legends to
Tables 1 and 2). Bars below the base line indicate sites not mutated by
AFB1. *, Spontaneous hotspots (13).
G'C -) T A transversions, and almost all of the available

-)

T-A transversion sites

one

G-C

--

were

GAC

induced (Table 1). In fact, the

T'A transversion site that

was

not

detected after

mutagenesis by AFB1-025--has a strong selection bias against

it because it is poorly suppressed by nonsense suppressors. There
is little indication of a specific induction of any other base change,
as the remaining mutations were scattered among many different sites (Table 2). Overall, the frequency distribution of
mutations is distinctly nonrandom (P << 0.001); for example,
mutations were recovered at only 37 of the 72 available sites,

whereas, with a random distribution, we would expect only 5

or 6 sites not to be included in a collection. of this size (21).
However, both (i) the distribution of mutations among all sites

We can conclude for the following reasons that AFB1-induced base substitutions, although SOS dependent, occur predominantly at the sites of premutational lesions. First, the nonsense mutations induced by AFB1 arose almost exclusively as
a result of G-C -* TEA transversions; were the mutations untargeted, we would expect other base substitutions to appear
in the collection. Second, mutational spectra, including both
the sites well mutated and the sites not mutated, are unique to
each SOS-dependent mutagen that has been. studied with the
lad system [e.g., ultraviolet light and 4-nitroquinoline-1-oxide
(13), BaPdiol epoxide (11), and neocarzinostatin (22)]. Moreover, although both the AFB1 and BaP diol epoxide spectra are
dominated by GC -- T-A transversions, the sites that are best
induced by these. two mutagens are not the same (Table 2 and
ref. 11). As we have discussed in more detail elsewhere (21, 23),
these findings strongly suggest that, in our studies, any untargeted mutagenesis occurring as a result of the induction of the
SOS response is minor.
The principal AFB1 adduct to DNA is at the N-7 position of
guanine (6-9); 7-alkyldeoxyguanosine residues can yield a formamidopyrimidine derivative by opening of the imidazole ring
(24, 25). A persistent, ring-opened AFB1-guanine residue has
been identified in vivo (26) and could be a premutational lesion.

Proc. Natd Acad. Sci. USA 80 (1983)

Genetics: Foster et aL
Table 2. Amber and Ochre mutations induced in the lad gene by
metabolically activated AFB1
Ichres
Ambers
Occurrences
Site
Occurrences
Site
Ssubstitution
1
0
09
A5
GC A*T
0
1
010
A6*
011
1
0
A9
0
2
013
A15*
0
017
0
A16
0
0
021
A19
0
1
A21
024
0
0
027
A23
0
0
028
A24
0
1
029
A26
0
2
034
A31
0
1
035
A33
1
A34*
1
A35
GLC-

TA

A2

A7
A10
A12
A13
A17
A20
A25
A27
A28

A-T

-*

T*A

All

A18
A32
X9
A36

A*T --+ COG

G-C--C-G

7
2
14
9
3
10
18
2
4
7

03
07
08
014
015
019
020
025
026
030
031
032
036

3
0
1
0
1

01
02
04
05
012
016
018
022
023
033

A3
A4
A14
A22

A30

Al
A8
A29

23
6
5
3
6
7
4
0

11

8
12
4
4
0

0
0

0
0
0
0
0

0
0

0
0

Designations are according to Coulondre and Miller (13). In addition

to the 187 mutants listed above and the two mutations at 06 (see the
legend to Table 1), three amber mutations were generated by tandem
double base changes at CTG leucine residues (i.e., CTG TAG). One
of these was at site X5 (amino acid residue 128) and two were at site
X6 (amino acid residue 184). *, Sites containing 5-methylcytosines, which
are spontaneous hotspots (13).

However, N-7 purine adducts also can induce depurination by

destabilizing the N-glycosylic bond (25). The release of free N7AFB1-guanine from AFB1-treated DNA is spontaneous in vitro
at slightly alkaline pH (27) and may be enzymatically accelerated in vivo (28). Thus, the production of aguaninic sites may
be a major result of DNA damage by AFB1.
Apurinic sites appear to be mutagenic for bacteriophage (29-

2697

Table 3. AFB1-induced mutations at the three TAC tyrosine

residues in the lad gene
Independent occurrences
TAC to TAG
TAC to TAA
Coding
(G-C -C C-G)
(G-C -. TEA)
position
Site
0
23
Tyr-7
AZ, 03
0
5
Tyr-47
A8,08
0
8
Tyr-273
A29,030
0
36
Total
Both amber and ochre mutations can be generated at these sites, allowing both G0 -C T-A and G-C -* C-G transversions to be monitored.

32). For example, point mutations can be recovered in ()X174

phage by transfecting spheroplasts from SOS-induced bacteria
with hot-acid-treated (i.e., depurinated) single-stranded 4'X174
DNA (30). In addition, after exposure to the carcinogen f3-propiolactone, the mutation rate of e$X174 DNA was increased by
treatment presumed to convert /-propiolactone-guanine residues to aguaninic sites (31). These observations have led Schaaper et al to suggest that apurinic sites may be important intermediates in the mutagenicity of carcinogens, such as /&
propiolactone and AFB1, that form adducts predominantly to
the N-7 position of purines (31).
Recently, Schaaper et aL have demonstrated that the base
substitutions that are recovered after transfection with hot-acidtreated single-stranded 4X174 DNA are dominated by A -+ T
and G -- T transversions, suggesting that there is a strong preference for adenine insertion opposite apurinic sites (32). In vitro evidence also has indicated a preference for insertion of adenine residues during bypass of apurinic sites in template DNA
(33). Thus, depurination at guanine residues could be expected
to lead specifically to G-C -- T A transversions during replication of the damaged DNA.
Therefore, the principal AFB1-induced mutations in the lad
gene can be accounted for by postulating that AFB1 generates
apurinic sites at guanine residues and that, during replication
after DNA damage in bacteria, there is a preference for inserting adenine opposite apurinic sites. The few base substitutions other than G-C -- T A transversions that were induced
by AFB1 in the lad gene could have occurred as a result of (i)
rare insertions of other bases across from a depurinated residue, (ii) infrequent depurinations of adenines, (iii) intrusions of
the spontaneous background or low levels of untargeted mutations, or (iv) other rarer premutational lesions.
Two other chemical carcinogens-BaP diol epoxide and Nacetoxy-AAF-also preferentially induce G-C -* TEA transversions in the lad gene (ref. 11; unpublished data). Because apurinic sites are not the major lesions produced by these agents,
and because other DNA lesions-such as guanine N-2 adducts-may be mutagenic, different pathways leading to the

G-C -- T A transversion can be imagined (11, 34). It seems reasonable, however, to consider a mechanism common to all three
carcinogens that accounts for their induction of G-C -- TEA
transversions. Because BaP diol epoxide and N-acetoxy-AAF
have been shown to induce apurinic sites in DNA in vitro (35),
the similar mutagenic specificities of AFB1, BaP diol epoxide,
and N-acetoxy-AAF may simply derive from their common ability
to create apurinic sites in DNA and from the preferential insertion of adenine opposite such lesions. Alternatively, the bulky
DNA lesions that these compounds produce may themselves be
noninformational sites opposite which adenines are preferentially inserted during replication after DNA damage. Finally,
there is the possibility that under the conditions of our experiments, lesions or base mispairs leading to base changes other

2698

Genetics: Foster et al.

than GC -- T-A transversions are preferentially corrected.

A role for mutations in the carcinogenic process has long been
suspected (36, 37) and has received support from the recent
observation that a single base pair substitution apparently is
sufficient to confer transforming properties on a normal mammalian gene (38-40). Our finding that three different chemical
carcinogens are similar in the base pair substitutions they generate in bacteria leads us to speculate that the carcinogenicity
of such agents may derive, in part, from their ability to induce
specific mutations in specific mammalian genes.
We thank Graham Walker and Steven Elledge for helpful discussions, Steven Winans for aid in preparing plasmid DNA, and Allan
Campbell for a bacterial strain. P.L.F. was a Fellow of the Interdisciplinary Programs in Health and was supported by a grant from the
Andrew Mellon Foundation and Grant CR807809-01-1 of the U.S. Environmental Protection Agency. This work was supported in part by
U.S. National Institutes of Health Grants CA26135 and ES02021 to E.E.;
J. H. M. was sponsored by grants from the Swiss League Against Cancer
and from the Swiss National Fund.
1. Wogan, G. N. (1973) in Methods in Cancer Research, ed. Busch,
H. (Academic, New York), Vol. 7, pp. 309-344.
2. Linsell, C. A. & Peers, F. G. (1977) in Origins of Human Cancer,
eds. Hiatt, H. H., Watson, J. D. & Winsten, J. A. (Cold Spring
Harbor Laboratory, Cold Spring Harbor, NY), pp. 549-556.
3. Garner, R. C. (1973) Chem.-BioL Interact. 6, 125-129.
4. Swenson, D. H., Miller, E. C. & Miller, J. A. (1974) Biochem.
Biophys. Res. Commun. 60, 1036-1043.
5. Essigman, J. M., Croy, R. G., Nadzan, A. M., Busby, W. F., Jr.,
Reinhold, V. N., Buchi, G. & Wogan, G. N. (1977) Proc. Nati Acad.
Sci. USA 74, 1870-1874.
6. Lin, J. K., Miller, J. A. & Miller, E. C. (1977) Cancer Res. 37,
4430-4438.
7. Martin, C. N. & Garner, R. C. (1977) Nature (London) 267, 863865.
8. Croy, R. G., Essigman, J. M., Reinhold, V. N. & Wogan, G. N.
(1978) Proc. Nati Acad. Sci. USA 75, 1745-1749.
9. Essigman, J. M., Green, C. L., Croy, R. G., Fowler, K. W, Buchi,
G. H. & Wogan, G. N. (1983) Cold Spring Harbor Symp. Quant.
BioL 47, in press.
10. Miller, J. H. (1978) in The Operon, eds. Miller, J. H. & Reznikoff, W. S. (Cold Spring Harbor Laboratory, Cold Spring Harbor,
NY), pp. 31-88.
11. Eisenstadt, E., Warren, A. J., Porter, J., Atkins, D. & Miller, J.
H. (1982) Proc. Nati Acad. Sci. USA 79, 1945-1949.
12. Howard-Flanders, P., Boyce, R. P. & Theriot, L. (1966) Genetics
53, 1119-1136.

Proc. NatL Acad. Sci. USA 80 (1983)

13. Coulondre, C. & Miller, J. H. (1977)J. Mol Biol 117, 525-575.
14. Langer, P. J., Shanabruch, W. G. & Walker, G. C. (1981)J. Bacteriol. 145, 1310-1316.
15. Perry, K. L. & Walker, G. C. (1982) Nature (London) 300, 278281.
16. Miller, J. H. (1972) Experiments in Molecular Genetics (Cold Spring
Harbor Laboratory, Cold Spring Harbor, NY).
17. Morrison, D. A. (1977)J. Bacteriol 132, 349-351.
18. Shanabruch, W. G. & Walker, G. C. (1980) Mol Gen. Genet. 179,
289-297.
19. Ames, B. N., McCann, J. & Yamasaki, E. (1975) Mutat. Res. 31,

347-364.
20. McCann, J., Spingarn, N. E., Kobori, J. & Ames, B. N. (1975)
Proc. Natl Acad. Sri. USA 72, 979-983.
21. Foster, P. L., Eisenstadt, E. & Cairns, J. (1982) Nature (London)
299, 365-367.
22. Foster, P. L. & Eisenstadt, E. (1983)J. Bacteriol 153, 379-383.
23. Miller, J. H. (1982) Cell 31, 5-7.
24. Lawley, P. D. & Brookes, P. (1963) Biochem. J. 89, 127-138.
25. Shapiro, R. (1968) Prog. Nucleic Acid Res. Mol Biol 8, 73-112.
26. Croy, R. G. & Wogan, G. N. (1981) Cancer Res. 41, 197-203.
27. Groopman, J. D., Croy, R. G. & Wogan, G. N. (1981) Proc. Natl
Acad. Sci. USA 78, 5445-5449.
28. Cerutti, P. A., Wang, V. T. & Amstad, P. (1980) in Carcinogenesis: Fundamental Mechanisms and Environmental Effects, eds.
Pullman, B., Ts'o, P. 0. P. & Gelboin, H. (Reidel, Holl and Kluwer, Boston), pp. 465-477.
29. Freese, E. B. (1961) Proc. Natl Acad. Sci. USA 47, 540-545.
30. Schaaper, R. M. & Loeb, L. A. (1981) Proc. Natl Acad. Sci. USA
78, 1773-1777.
31. Schaaper, R. M., Glickman, B. W. & Loeb, L. A. (1982) Cancer
Res. 42, 3480-3485.
32. Schaaper, R. M., Kunkel, T. A. & Loeb, L. A. (1983) Proc. Natl
Acad. Sci. USA 80, 487-491.
33. Strauss, B., Rabkin, S., Sagher, D. & Moore, P. (1983) Biochimie
64, 829-838.
34. Kadlubar, F. F. (1980) Chem.-Biol Interact. 31, 255-263.
35. Drinkwater, N. R., Miller, E. C. & Miller, J. A. (1980) Biochemistry 19, 5087-5092.
36. Boveri, T. (1914) The Origin of Malignant Cells (Williams & Wil-

kins, Baltimore).

37. McCann, J., Choi, E., Yamasaki, E. & Ames, B. N. (1975) Proc.
Natl. Acad. Sci. USA 72, 5135-5139.
38. Tabin, C. J., Bradley, S. M., Bargmann, C. I., Weinberg, R. A.,
Papageorge, A. G., Scolnick, E. M., Dhar, R., Lowy, D. R. &
Chang, E. H. (1982) Nature (London) 300, 143-149.
39. Reddy, E. P., Reynolds, R. K., Santos, E. & Barbacid, M. (1982)
Nature (London) 300, 149-152.
40. Taparowsky, E., Suard, Y., Fasano, O., Shimizu, K., Goldfarb,
M. & Wigler, M. (1982) Nature (London) 300, 762-765.