! ~',

Fundamental and Molecular Mechanisms of Mutagenesis

ELSEVIER Mutation Research 309 (1994) 11-15

Induction of tandem-base change mutations

Frankl in H u t c h i n s o n *

Radiobiology Laboratories, Department of Therapeutic Radiology, Yale University School of Medicine, 333 Cedar Street, New Haven, CT 06510-8039, USA

Received 13 December 1993; accepted 26 January 1994

Abstract

Four databases with sequence changes for 12000 mutations in mammalian and bacterial cells were searched for genotoxic agents inducing tandem mutations, in which two adjacent base pairs are changed. Ultraviolet light induces about one CC > "IT per 10-20 mutations, and other tandem-base changes at about half that frequency. There is strong evidence that c/s-diammine dichloroplatinum (II) induces tandems. These results suggest that tandem-base changes are induced by agents that damage two adjacent base pairs in DNA. Tandems, particularly CC > q-T, can be used as indicators of exposure to ultraviolet light, as in genes in skin cells exposed to sunlight. Oxidizing agents such as superoxide and ionizing radiation do not induce a significant level of tandem-base changes (such as CC > 11") in double-strand DNA, so such mutations are not a useful indicator of exposure to mutagens of this type. All conclusions are equally valid for bacterial and mammalian cells.

Key words: Base change mutations; Tandem mutations; Mutation databases; Oxidizing agents; Ultraviolet light

1. Introduction

A distinctive type of mutation induced by some mutagens is one in which two adjacent base pairs have changed: for example, adjacent G : C pairs might be altered to A : T pairs, a mutation re- ferred to here as CC > ' I T (GG > A A in the complementary strand). Since the probability that two independently formed lesions cause the dou-

* Corresponding author, Tel. 203 785 5614; Fax 203 785 6309.

ble change is extremely small *, the mutation is ascribed to a single event, and is commonly called a tandem-base change.

It has long been recognized that ultraviolet light induces such mutations with fairly high fre- quency. Recently, a significant level of CC > TT

* A high mutat ion frequency of one base change per 104 genomes in a gene with 103 base pairs leads to an average of one base change in 107 bases. Even if a particular site is 100-fold more mutable than average, a base on either side of a given base change will be muta ted at a frequency of only 2 × 10 -5, so tandem-base changes will be formed at a rate of less than one per 50000 base changes.

0027-5107/94/$07.00 © 1994 Elsevier Science B.V. All rights reserved SSD1 0 0 2 7 - 5 1 0 7 ( 9 4 ) 0 0 0 4 6 - 8

12 F. Hutchinson/Mutation Research 309 (1994) 11-15

mutations has been reported in single-strand DNA (ssDNA) treated with oxidizing agents (Reid and Loeb, 1992, 1993; Tkeshelashvili et al., 1991), and Loeb and his associates have suggested this mutation might be an indication of mutagenesis by oxygen-based radicals. A way to identify such mutagenesis would be of great importance, be- cause of the role such radicals generated in nor- mal cell metabolism are thought to play in "spontaneous" mutagenesis and carcinogenesis (Ames and Gold, 1991).

The availability of several databases with base- sequence changes in many thousands of muta- tions has made feasible a search for the kinds of mutagenic agents that induce tandem mutations. This paper reports the results of this search.

2. Methods

The following databases in electronic form were scanned, with the commercial database pro- gram Paradox, for tandem-base changes and the mutagens that made them. To confirm conclu- sions reached in the search, additional results were retrieved from the literature, as described.

A mutational database assembled at Yale in- cludes (Aug. 1993 version) more than 16000 se- quenced mutants in 10 genes: the E. coli genes

crp, gpt, lacL lacZ(a) complementing factor, supF and tetA, lambda phage cI repressor, yeast sup4-o, hamster aprt and human hprt. Of these, about 8600 are sequenced mutations that were searched in this work; the remainder are mutations creat- ing stop codons in lacl, from which sequence changes were deduced by methods developed by Miller and his associates (Coulondre and Miller, 1977), and were not searched because detection of tandem-base changes in this way is inefficient. Data sets were included in the database because of their relevance to the research interests of the organizer (F. Hutchinson), with emphasis on the mutagens UV and ionizing radiation. The Yale database is available on Internet: anon(ccnathl. med.yale.edu; when asked for a password, answer "mutagen."

A database of 800 mutations (Spring 1993 ver- sion) in the human hprt gene has been described (Cariello et al., 1992), including both mutations induced in cultured human cells by exogenous agents and those found in DNA from humans. It is also available on Internet: skopek@uncvxl.oit. unc.edu; for password, answer "database."

A database of mutations in the human p53 gene collected from human tissue (Caron de Fro- mentel and Soussi, 1992) has more than 1900 entries (Aug. 1993 version). It is available on a computer diskette from Prof. T. Soussi, Institut

Table 1 Tandem mutat ions induced by various mutagens

Mutagen Database Total number of mutat ions

Number of tandem mutat ions at

CC > TT PyPy 5'-PyPu 5'-PuPy

UV Yale 2216 hprt 103 p53 75

Oxidative agents Yale-dsDNA 782 Yale-ssDNA 363

Spontaneous Yale 2 628 p53 1 830 HaemB 806 hprt 167

cis-Platin Yale 225

Other Yale 2 374 hprt 511

98 51 4 8 11 3 6 1

1 5 3 12

4 4 t 1 11 2 3

1 1 I

2 1

3 I 1 1

F. Hutchinson/Mutation Research 309 (1994) 11-15 13

de G6n&ique Mol6culaire, Unit6 301, INSERM, 27 rue J. Dodu, 75010 Paris, France.

A database of 806 mutations (July 1993 ver- sion, Giannelli et al., 1993) at the haemophilia B locus, coding for the human factor IX protein, is available on Internet: anonymous@ftp.embl- heidelberg.de; enter e-mail address for password.

3. Results

The results of the search for tandem base changes are summarized in Table 1. Instead of the 4 x 4 = 16 possible sites in a single strand, the symmetry of double-strand DNA (dsDNA) re- duces the number of different sites to 10: for example, a site with CC in the transcribed strand is, for a mutagenic agent, the same as a site with GG in that strand and CC in the complementary strand. Each of the ten sites can mutate in 3 x 3 = 9 ways, for a total of 90 different tandem-base changes. It quickly became apparent that the predominant type of tandem-base change is CC > TF (GG > AA), which was made a separate entry in the table. Presentation of data on other tandem base changes is complicated by the large number of possibilities. Table 1 gives the number of tandems (except CC > TF) at PyPy sequences, at 5'-PyPu, and at 5'-PuPy.

UV is the mutagen inducing by far the largest number of tandem base changes in the Yale and human hprt databases. The data show that one in 10 or 20 UV-induced mutations are CC > T-F, and that other tandem-base changes were in- duced at roughly half that rate, mostly at PyPy sites of possible pyrimidine cyclobutane dimers or (6-4) dipyrimidine products. The conclusion is valid for both bacterial and mammalian cells. In the 66 UV-induced tandem base changes (other than CC > TT) in the two databases, there were 55 transitions and 77 transversions. For the hu- man p53 gene, all CC > TT except one were from skin cells; mutations in skin are listed in Table 1 as UV-induced because such cells are more likely to be exposed to sunlight than other human cells, but there may be other factors, such as a high level of oxygenation.

The anticancer drug c/s-diamminedichloro- platinum(II), or cis-platin, induced a significant frequency of tandems in data sets for bacterial and yeast cells in the Yale database. A search of the literature for sequence changes induced by cis-platin turned up an additional data set for mammalian cells, with two tandem-base changes of 28 sequenced mutations (de Boer and Glick- man, 1989). Of the total of 5 tandem-base changes (these two plus the three in the Yale database), four are at PuPu sequences, thought to be the major site of intrastrand crosslinking (Eastman, 1986).

A search was made for tandem-base changes induced by oxidizing agents such as transition metal ions plus oxygen, superoxide, hydrogen peroxide, singlet oxygen and ionizing radiation. For DNA mutagenized in single-strand form, two data sets in the Yale database with 129 mutants showed an average of one CC > TT per 10 muta- tions (Reid and Loeb, 1992; Tkeshelashvili et al., 1991), a conclusion supported by reversion assays using a specially engineered plasmid that found about one CC > TT per 30 mutations (Reid and Loeb 1993). There were no tandem-base changes in the 6 other data sets (234 mutants) in the Yale database for ssDNA exposed to oxidizing agents, including two (McBride et al. 1991, 1992) from same (L.A. Loeb's) laboratory that reported the CC > TT mutations.

Of the 9 tandems in mammalian and bacterial cells (including one CC > TT) in dsDNA treated with oxidizing agents, 7 were among 337 muta- tions in genes on plasmids mutagenized in solu- tion; because induced mutagenesis was only mod- erately increased over background, at least some of these could be from spontaneous processes (see below). Clearly, neither CC > TT nor other tandem-base changes are good indicators of ex- posure of dsDNA to oxidizing mutagens.

The class of spontaneous mutations was taken to include all data sets so listed in the Yale and human hprt databases, all mutations in the p53 database from cells other than skin, and all muta- tions in the haemophilia B database. Of the 9 spontaneous tandems in the Yale database,~seven were among 145 mutants in genes on untreated dsDNA transfected into mammalian cells, and

14 F. Hutchinson/Mutation Research 309 (1994) 11-15

might have been introduced by the process of transfection (Seidman et al., 1987). The other two tandems among spontaneous mutations in the Yale database, together with the data in the Haemophil ia B database, suggest a very low inci- dence of tandem-base changes induced by spon- taneous intracellular processes, the order of one in a thousand. Data for p53 show about one tandem per 100 mutations of spontaneous origin, as do the relatively few data in the hprt database, and the reason for the difference is not known.

For completeness, mutagens associated with tandem mutations in the "Othe r " category (Table 1) are listed here. The three CC > TT mutations in the Yale database are all from a collection of 39 mutants induced in aprt in cultured hamster ceils by 1 mM thymidine (Phear and Meuth 1989). The only relevant paper found in the literature showed no CC > TT induced by 3 mM thymidine among 25 revertants at two CC sites * in mouse A912 cells (Kresnak and Davidson, 1992); the two data sets are consistent with one CC > TT per tens of C > T mutations, presumably from a G : T pair increasing the probability of an adjacent mispair, but more data are needed. 2 of the 3 tandem base changes from the hprt database are from a collection of 35 mutants induced in cul- tured human lymphoblast ceils by N-ethyl-N- nitrosourea (Bronstein et al., 1991), but this re- sult is contrary to data in the Yale database: no tandem base changes among 130 mutants (5 data sets) induced by the same agent. The remaining tandem (CC > TT) in the human hprt database was in D N A from a smoker (Vrieling and Thijssen, 1990).

A nonsystematic search through papers not included in these databases turned up two with significant fractions of tandem-base changes. Ds- D N A incubated for 30 days at 37°C with echino- mycin, a bis-intercalating anti tumor drug thought to partially denature DNA, showed 33 CC > TT mutations in 184 mutants, mostly C > T base

* Although CC > TT mutations were not explicitly assayed in the experiments, Dr. Richard Davidson pointed out that such a mutation at either site would insert leucine, producing an amino acid change known to be detectable at both sites.

changes from deamination of cytosine (Moyer et al., 1993). SsDNA incubated at 70°C at pH 4.8, to encourage depurination, showed a CC > TT, a G G > CC, and an A G > TC mutation, as well as two G G G > AAA, in 93 sequenced mutants (Cabral Neto et al., 1992).

4. Discussion

A search of about 12 000 sequenced mutations listed in 4 databases showed an incidence of tandem-base change mutations that differs greatly from one mutagen to another.

The highest incidence of tandems is from mu- tagens that form products involving two adjacent bases in the same D N A strand: UV with pyrimi- dine dimers and (6-4) products, and cis-platin with intrastrand crosslinks. This suggests that the rule, that base change mutations occur at the inducing lesion (Miller 1982), applies also to tan- dem-base changes. The relatively large fraction of CC > TT and other tandems induced by UV vali- dates the use of these mutations in skin cells as a indicator of mutagenesis by UV-B from the sun (Brash et al., 1991), a matter of particular impor- tance because of evidence for increasing levels of such radiation at the earth 's surface (Kerr and McElroy, 1993).

One in 200 UV-induced mutations in Table 1 (12 of 2394) were tandems at PuPy sites, but most of these cannot be of spontaneous origin. Sponta- neous mutations form only a small fraction of the mutations in irradiated ceils, since UV is an effi- cient mutagen, and among those of spontaneous origin only one in a hundred or less will be tandems (see Results). Tandems at PuPy sites might arise from purine-pyr imidine photoprod- ucts (Bose and Davies 1984), or be rare cases of base changes at undamaged bases.

The data show that tandem mutations are not an effective indicator of exposure of dsDNA in bacterial or mammalian cells to oxygen-based free radicals or ionizing radiation. This conclusion may be reconciled with reports of a CC > TT muta- tion per 10-30 mutations induced by such agents (Reid and Loeb, 1992, 1993; Tkeshelashvili et al., 1991) as follows: the latter data are for ssDNA,

F. Hutchinson/Mutation Research 309 (1994) 11-15 15

for which both products and processing can differ from dsDNA; also, the absence of tandem muta- tions in 6 other data sets with 234 mutants in- duced in ssDNA by oxidizing agents (see Results) suggests that some as yet unknown factor needs to be considered. It is worth noting that, if oxidiz- ing agents do induce tandems in dsDNA, the low level of tandems in the spontaneous spectrum indicates that oxidizing agents are not important in spontaneous mutagenesis, a conclusion that differs from current opinion.

Acknowledgements

I thank Dr. J. Edward Donnellan Jr., for much of the computer programming and for discussion. This research was supported by grant 5 RO1 CA58952 from the U.S. National Institutes of Health.

References

Ames, B.N., and L.S. Gold (1991) Endogenous mutagens and the causes of aging and cancer, Mutation Res., 250, 3-16.

Bose, S.N., and R.J.H. Davies (1984) The photoreactivity of T - A sequences in oligodeoxyribonucleotides and DNA, Nucl. Acids Res., 12, 7903-7914.

Brash, D.E., J.A. Rudolph, J.A. Simon, A. Lin, G.J. McKenna, H.P. Baden, A.J. Halperin and J. Ponten (1991) A role for sunlight in skin cancer: UV-induced p53 mutations in squamous cell carcinoma, Proc. Natl. Acad. Sci. (U.S.A.), 88, 10124-10128.

Bronstein, S.M., J.E. Cochrane, T.R. Craft, J.A. Swenberg and T.R. Skopek (1991) Toxicity, mutagenesis, and muta- tional spectra of N-ethyl-N-nitrosOurea in human cell lines with different DNA repair phenotypes, Cancer Res., 51, 5188-5197.

Cabral Neto, J.B., A. Gentil, R.E. Caseira Cabral and A. Sarasin (1992) Mutation spectrum of heat-induced abasic sites on a single-stranded shuttle vector replicated in mammalian cells, J. Biol. Chem., 267, 19718-19723.

Cariello, N., T.R. Craft, H. Vrieling, A.A. van Zeeland, T. Adams and T.R. Skopek (1992) Human HPRT mutant database: software for data entry and retrieval, Environ. Mol. Mutagen., 20, 81-83.

Caron de Fromentel, A.R. and T. Soussi (1992) TP53 tumor

suppressor gene: a model for investigating human mutage- nesis, Genes, Chromosomes, Cancer, 4, 1-15.

Coulondre, C., and J.H. Miller (1977) Genetic studies of the lac repressor IV, J. Mol. Biol., 117, 577-606.

de Boer, J.G., and B.W. Glickman (1989) Sequence specificity of mutation induced by the antitumor drug cisplatin in the CHO aprt gene, Carcinogenesis, 10, 1363-1367.

Eastman, A. (1986) Reevaluation of interaction of cis-di- chloro(ethylene-diamine)platinum(II) with DNA, Bio- chemistry, 25, 3912-3915.

Giannelli, F., P.M. Green, K.A. High, S. Sommer, M.-C. Pooh et al. (1993) Haemophilia B: database of point mutations and short additions and deletions-fourth edition, 1993, Nucl. Acids Res., 13, 3075-3087.

Kerr, J.B., and C.T. McElroy (1993) Evidence for large up- ward trends of ultraviolet-B radiation linked to ozone depletion, Science, 262, 1032-1034.

Kresnak, M.T., and R.L. Davidson (1992) Thymidine-induced mutations in mammalian cells: Sequence specificity and implications for mutagenesis in vivo, Proc. Natl. Acad. Sci. (U.S.A.), 89, 2829-2833.

McBride, T.J., B.D. Preston and L.A. Loeb (1991) Mutagenic spectrum resulting from DNA damage by oxygen radicals, Biochemistry, 30, 207-213.

McBride, T.J., J.E. Schneider, R.A. Floyd and L.A. Loeb (1992) Mutations induced by methylene blue plus light in single-stranded M13mp2, Proc. Natl. Acad. Sci. (U.S.A.), 89, 6866-6870.

Miller, J.H (1982) Carcinogens induce targeted mutations in E. coli, Cell, 31, 5-7.

Moyer, R., D. Briley, A. Johnsen, U. Stewart and B.R. Shaw (1993) Echinomycin, a bis-intercalating agent, induces C > T mutations via cytosine deamination, Mutation Res., 288, 291-300.

Phear, G., and M. Meuth (1989) The genetic consequences of DNA precursor pool imbalance: sequence analysis of mu- tations induced by excess thymidine at the hamster aprt locus, Mutation Res., 214, 201-206.

Reid, T.M., and L.A. Loeb (1992) Mutagenic specificity of oxygen radicals produced by human leukemia cells, Cancer Res., 52, 1082-1086.

Reid, T.M., and L.A. Loeb (1993) Tandem double CC > TT mutations are produced by reactive oxygen species, Proc. Natl. Acad. Sci. (U.S.A.), 90, 3904-3907.

Seidman, M.M., A. Bredberg, S. Seetharam and K.H. Krae- met (1987) Multiple point mutations in a shuttle vector propagated in human cells: evidence for an error-prone DNA polymerase activity, Proc. Natl. Acad. Sci. (U.S.A.), 84, 4944-4948.

Theshelashvili, L.K., T. McBride, K. Spence and L.A. Loeb (1991) Mutation spectrum of copper-induced DNA dam- age, J. Biol. Chem., 266, 6401-6406.

Vrieling, H., and J.C.P. Thijssen (1990) Unpublished data, Submitted to Human HPRT Database.