Effects of O6-Alkylguanine-DNA Alkyltransferase and Mismatch Repair

Defective mismatch repair is known to be associated with tolerance to the cytotoxic effects of agents that produce m6G in DNA in both E. coli (16) and...
0 downloads 0 Views 102KB Size
Download PDF
894

Chem. Res. Toxicol. 2001, 14, 894-900

Mutagenesis by O6-Methyl-, O6-Ethyl-, and O6-Benzylguanine and O4-Methylthymine in Human Cells: Effects of O6-Alkylguanine-DNA Alkyltransferase and Mismatch Repair Gary T. Pauly and Robert C. Moschel* Chemistry of Carcinogenesis Laboratory, National Cancer Institute at Frederick, PO Box B, Frederick, Maryland 21702 Received February 14, 2001

Double-stranded and gapped shuttle vectors were used to study mutagenesis in human cells by O6-methyl (m6G)-, O6-ethyl (e6G)-, and O6-benzylguanine (b6G), and O4-methylthymine (m4T) when these bases were incorporated site-specifically in the ATG initiation codon of a lacZ′ gene. Vectors were transfected into either human kidney cells (293) or colon tumor cells (SO) or into mismatch repair defective human colon tumor cells (H6 and LoVo). Cellular O6alkylguanine-DNA alkyltransferase (alkyltransferase) was optionally inactivated by treating cells with O6-benzylguanine prior to transfection. In alkyltransferase competent cells, the mutagenicity of all the modified bases was substantially higher in gapped plasmids than in double-stranded plasmids. Alkyltransferase inactivation increased mutagenesis by the three O6-substituted guanines in both double-stranded and gapped plasmids but did not affect m4T mutagenesis. In the absence of alkyltransferase, mutagenesis by m6G and to a lesser extent e6G in double-stranded vectors was higher in the mismatch repair defective H6 and LoVo cells than in SO or 293 cells indicating that e6G as well as m6G were subject to mismatch repair processing in these cells. The level of mutagenesis by m4T and b6G was not affected by mismatch repair status. When incorporated in gapped plasmids and in the absence of alkyltransferase, the order of mutagenicity for the modified bases was m4T > e6G = m6G > b6G. The O6-substituted guanines primarily produced GfA transitions while m4T primarily produced TfC transitions. However, m4T also produced a significant number of TfA transversion mutations in addition to TfC transitions in mismatch repair deficient LoVo cells.

Introduction The production of O6-substituted guanines and O4substituted thymines in DNA has long been associated with the mutagenicity and carcinogenicity of alkylating carcinogens (e.g., N-alkyl-N-nitroso compounds). Aspects of the mutagenicity and repair of these types of modified bases have been studied previously in mammalian cells using the site-specific mutagenesis approach (1-9). We have used this approach here to examine mutagenesis and repair of O6-methyl (m6G)-, O6-ethyl (e6G)-, and O6benzylguanine (b6G), as well O4-methylthymine (m4T) in human cells. For these studies, we constructed a shuttle vector that could harbor the modified bases site-specifically within the ATG initiation codon of a lacZ′ gene and be replicated in both human cells and Escherichia coli. The vector can harbor a modified base in either a singleor double-stranded DNA context. This newly developed system was adapted from the bacterial plasmid system we used previously to study mutagenesis by these compounds in E. coli (10-13). It is well known that the principal cellular defense against O6-substituted guanine damage is O6-alkylguanine-DNA alkyltransferase (alkyltransferase) (14). This protein transfers O6-substituents from the modified guanines to a reactive cysteine residue in the active site * To whom correspondence should be addressed. Phone: (301) 8465852. Fax: (301) 846-6146. E-mail: [email protected].

10.1021/tx010032f

thereby restoring a normal guanine in DNA. This stoichiometric reaction results in the inactivation of the protein. Although the protein has been shown to demethylate m4T in vitro, this reaction is far less efficient than demethylation of m6G residues (15). Altshuler et al. (7) compared the mutagenicity of m6G and m4T in Chinese hamster ovary cells that were either proficient or deficient in alkyltransferase. They showed that m4T was more mutagenic than m6G in both cell types, but m4T unlike m6G, was not significantly repaired by alkyltransferase. In addition to alkyltransferase repair, mismatch repair is known to play a role in modulating the mutagenicity and genotoxicity of m6G residues. Defective mismatch repair is known to be associated with tolerance to the cytotoxic effects of agents that produce m6G in DNA in both E. coli (16) and in human cells (17-19) and we have previously demonstrated a role for mismatch repair in modulating m6G mutagenesis in E. coli (10-12). Our objectives in this study were to compare mutagenesis and repair of m6G, e6G, and b6G as well as m4T in human cells when the modified bases were incorporated in either double-stranded or gapped vectors and to assess the ability of our shuttle vector system to confirm and extend previous observations made for some of these bases by other laboratories (1, 7-9, 17-19). In addition, we sought to compare our results in human cells with those obtained previously in E. coli (10-13).

This article not subject to U.S. Copyright. Published 2001 by the American Chemical Society Published on Web 06/26/2001

m6G, e6G, b6G, and m4T Mutagenesis in Human Cells

Figure 1. Characteristics of the shuttle vector pGP50 and a schematic for the construction of vectors containing a modified base in the ATG initiation codon of a lacZ′ gene. P is the promoter and lacZ′ is the coding sequence of the gene. The asterisks mark the positions for incorporation of a modified base. The complementary oligonucleotide contains either thymine or uracil residues to permit experiments with either doublestranded or gapped plasmids (see text).

Experimental Procedures Materials and Methods. The synthesis of the modified oligodeoxyribonucleotides and complementary strands were described (10-13). These oligonucleotides were repurified by reverse phase HPLC using the conditions described for detritylated oligonucleotides (20) prior to use. All restriction enzymes and T4 DNA ligase used in this study were from New England Biolabs (Beverly, MA). The restriction enzymes were used in the buffers provided by the supplier. The human embryonic kidney cell line 293 (ATCC CRL 1573) was obtained from the American Type Culture Collection (Rockville, MD). The 293 cells were maintained in minimal essential medium with Earle’s salts and L-glutamine plus 10% heat inactivated horse serum and antibiotic-antimycotic (Life Technologies, Grand Island, NY). Human colon tumor cell lines SO, H6 (21, 22), and LoVo (23) were a kind gift from Dr. Paul Modrich, Duke University Medical Center. These colon cell lines were maintained in McCoy’s 5A medium with 10% fetal bovine serum and antibioticantimycotic (Life Technologies, Grand Island, NY). All cells were grown in a 5% CO2 environment. Preparation of Adduct Containing Plasmids. The procedure for the incorporation of oligonucleotides into the shuttle vector pGP50 (Figure 1A) was similar to that described previously (13). The vector was isolated from E. coli strain DH5 (24) using a Wizard Megaprep System (Promega Corporation, Madison, WI), and was purified by CsCl density gradient centrifugation. In a 20 mL reaction, 2 mg of pGP50 were digested with 500 units of BspMI restriction enzyme for 4 h in order to remove

Chem. Res. Toxicol., Vol. 14, No. 7, 2001 895 the insert sequence and create a pair of noncomplementary and nonpalindromic four base overhanging ends (Figure 1B). Preparative electrophoresis of the digested plasmid through 0.8% low melting point agarose gave three bands of 0.6, 0.7, and 5.2 kb. The 5.2 kb fragment was recovered from the gel. Pairs of oligonucleotides (Figure 1C) were annealed by mixing approximately 80 pmol of each oligonucleotide in 30 µL of 100 mM NaCl, 10 mM Tris-HCl, and 1 mM EDTA (pH 8.0) and by heating to 70 °C followed by slow cooling to room temperature. A total of 225 µg (65 pmol) of 5.2 kb plasmid fragment was ligated with 80 pmol of annealed oligonucleotides in 8 mL of a buffer containing 66 mM Tris-HCl (pH 7.6), 6.6 mM MgCl2, 10 mM dithiothreitol, 1 mM ATP, and 100 Weiss units of T4 DNA ligase at 16 °C overnight. The reaction was made 2 M in ammonium acetate and the DNA was precipitated with 2.5 vol of absolute ethanol. The DNA was resuspended and digested with 300 units of XhoI restriction enzyme for 3 h at 37 °C. Proteins were extracted with phenol/chloroform/isoamyl alcohol (25:25:1) followed by extraction with chloroform and the DNA was then precipitated with 2.5 vol of ethanol. Preparative gel electrophoresis of the XhoI digested DNA on 0.8% low melting point agarose gave three bands of 5.2, 3.9, and 1.3 kb. DNA in the 5.2 kb band was isolated. The 5.2 kb fragment was dissolved in 7.5 mL of ligation buffer (see above) and was treated with 125 Weiss units of T4 DNA ligase overnight at 16 °C. At the end of this ligation, 2.5 mL of 8 M ammonium acetate was added and the DNA was precipitated by addition of 2.5 vol of ethanol. DNA from this second ligation reaction was electrophoresed through 0.8% low melting point agarose, containing 1 µg/mL ethidium bromide. This electrophoresis gave bands corresponding to single, covalently closed plasmids, with an apparent size of 2.6 kb, as well as bands corresponding to linear, nicked, and multimeric plasmids. The covalently closed plasmids were recovered from the gel and quantified. From 10 to 15 µg of each adduct-containing plasmid was prepared in this way. Two types of plasmid were constructed. Double-stranded plasmids contained thymine residues in the complementary strand (Figure 1C). Gapped plasmids contained four uracil residues in the complement in place of the thymine residues (Figure 1C). In host cells, these plasmids would be acted on by uracil-DNA glycosylase and AP endonuclease to create a gap in the complementary strand (11, 12). Transfection of Cells and Detection of Mutations in Replicated Vectors. The procedures used to pass shuttle vectors through human cells and score adduct induced mutations in E. coli are diagramed in Figure 2. On the first day of a typical transfection (2.5-3.0) × 105 cells were plated onto a 35 mm dish containing 2 mL of growth medium. In experiments where alkyltransferase was depleted, the growth medium was supplemented with 100 mM O6-benzylguanine in DMSO to a final concentration of 50 µM (25). On the second day the growth medium was replaced with fresh medium with or without O6benzylguanine depending on whether alkyltransferase depletion was intended or not. Two hours later, the medium was removed, the cells were washed with Opti-MEM (Life Technologies), and the cells were overlaid with 1 mL of lipid-DNA complex. The lipid-DNA complexes were formed by mixing 250 ng of plasmid DNA in 100 µL of Opti-MEM with 7 µL of lipofectamine (Life Technologies), also in 100 µL of Opti-MEM. The complex was allowed to form for 30 min at room temperature and then 800 µL of Opti-MEM at 37 °C with or without 62.5 µM O6benzylguanine was added immediately before transfer to the cells. After 5-h exposure to the lipid-DNA complex, 1 mL of growth medium containing twice the normal amount of serum and no antibiotic-antimycotic was added with or without O6benzylguanine. After 15-17 h, the medium was replaced by fresh growth medium with or without O6-benzylguanine. On the fourth day of the experiment, the medium was poured off and the cells were twice scraped from the dish into 500 µL of

896

Chem. Res. Toxicol., Vol. 14, No. 7, 2001

Pauly and Moschel ing X-gal1 and IPTG1 and incubated (13). The remainder of the electroporation was diluted with an equal volume of 30% glycerol in LB broth and quick frozen at -80 °C. A dilution of the electroporation reaction was determined that would give approximately 100 colonies per plate. The frozen stock was then used to innoculate 10 such plates which were incubated and scored for blue and white colonies. To demonstrate that DpnI treatment was effective in eliminating plasmids that had not been replicated in human cells, 250 ng of an e6G-containing double-stranded plasmid was digested with DpnI as described above and the resulting DNA was used to transform E. coli DH10B cells. No transformants were produced. In a parallel experiment with no DpnI enzyme, 35 000 transformants were produced by the same amount of undigested plasmid. Identification of Mutations. Fifty white E. coli colonies were picked from experiments in which double-stranded and gapped vectors carrying m6G, e6G, b6G, or m4T were replicated in alkyltransferase depleted 293 or LoVo cells. These colonies were spotted onto nitrocellulose filters. Colonies derived from plasmids containing O6-substituted guanines were probed with 32P-end labeled 16-base oligonucleotides directed against G or A at the site of the O6-substituted guanine as described previously (13). Colonies derived from plasmids containing O4methylthymine were probed with 32P-end labeled 15-base oligonucleotides directed against T or C at the site of the modified base as described in ref 12. Mutants that were not identified in this way, together with 20 mutants from the control plasmid that contained uracil residues in the complementary strand and had been replicated in 293 cells, were sequenced using an ABIPRISM dye terminator cycle sequencing kit and an ABI 373A DNA sequencer (Perkin-Elmer, Foster City, CA).

Results

Figure 2. A schematic illustrating the transfection of shuttle vectors into human cells, the recovery of replicated vectors and the subsequent transformation of E. coli where adduct induced mutations are scored by the growth of white (mutant) rather than blue (wild-type) bacterial colonies. phosphate-buffered saline (PBS).1 The cells were pelleted by centrifugation and the PBS was removed. Plasmid DNA was isolated from the cells with a Wizard mini-prep purification system (Promega Corporation) using the same protocol described by the manufacturer for the isolation of plasmids from bacterial cells. Typically, 10-100 ng of plasmid DNA was recovered. Upon recovery, the entire sample of recovered DNA was digested with 10 units of DpnI restriction enzyme for 1 h in 50 µL of buffer to degrade unreplicated plasmid DNA (26) which contains 19 potential DpnI restriction sites. The DNA was precipitated by the addition of 2.5 vol of ethanol, was briefly air-dried, and was resuspended in 25 µL of 10 mM Tris-HCl (pH 8), 1 mM EDTA. Electroporation competent E. coli DH10B cells (24) were prepared as previously described for polA strains (11) and 1 µL of recovered shuttle-vector DNA was mixed with 100 µL of DH10B cells on ice, transferred to a 0.1 cm cuvette and pulsed at 1.8 kV with an E. coli pulser apparatus (Bio-Rad Laboratories, Melville, NY). After pulsing, 900 µL of LB broth was added and the cells were incubated at 37 °C for 30 min. Several dilutions of the electroporation reaction were plated onto media contain1 Abbreviations: m6G, O6-methylguanine; e6G, O6-ethylguanine; b6G, O6-benzylguanine; m4T, O4-methylthymine; alkyltransferase, O6-alkylguanine-DNA alkyltransferase; PBS, phosphate-buffered saline; IPTG, isopropyl-β-D-thiogalactopyranoside; X-gal, 5-bromo-4-chloro-3-indolylβ-D-galactoside.

The design of the nonintegrating shuttle vector described here was adapted from the bacterial plasmid system we used to study mutagenesis by these modified bases in E. coli (10-13). The shuttle vector, which is illustrated in Figure 1A, contains an E. coli plasmid origin of replication (Col E1 ori) in addition to a viral origin of replication (SV40 ori) which allows replication in human and simian cells. The product of the SV40 large T antigen gene is necessary for utilization of the SV40 origin of replication. A beta lactamase gene (Amp) provides for selection in E. coli. The plasmid contains a nonfunctional lacZ R complementation gene in which the lacZ promoter (P in Figure 1A) is separated from the coding sequence of the lacZ R complementation peptide (lacZ′) by a 1.3 kilobase insert (insert). The insert sequence is flanked by outward facing sites for the restriction enzyme BspMI (Figure 1B). Pairs of oligonucleotides were incorporated into the plasmid by first digesting with BspMI and isolating the major fragment and then utilizing the noncomplementary four base overhanging ends to direct ligation of a pair of oligonucleotides (Figure 1C) into the plasmid. This resulted in a covalently closed circular vector that contained a functional lacZ′ gene. The initiation codon (ATG) of the lacZ′ gene contained normal bases or a carcinogenmodified base as indicated in Figure 1C. Double-stranded vectors contained thymine residues in the complementary strand and gapped vectors contained uracil residues in the complement as illustrated in Figure 1C. The procedure followed for mutagenesis experiments is diagramed in Figure 2. Adduct containing vectors were transfected into both repair competent or mismatch repair defective human cell lines which were optionally treated with O6-benzylguanine to inactivate O6-alkylgua-

m6G, e6G, b6G, and m4T Mutagenesis in Human Cells

Figure 3. Percentage of white (mutant) colonies produced following passage of double-stranded control and adduct containing shuttle vectors through 293 cells (solid bars), mismatch repair competent SO colon tumor cells (open bars), mismatch repair defective H6 colon tumor cells (hatched bars) or mismatch repair defective LoVo colon tumor cells (crosshatched bars). Panel A shows the results of experiments in cells not pretreated with O6-benzylguanine, and panel B shows results of experiments with cells pretreated with 50 µM O6-benzylguanine to deplete alkyltransferase. All data are the average of three transfections. Error bars represent the standard deviation.

nine-DNA alkyltransferase. Upon recovery of replicated vectors from the human cells, plasmids were used to transform lacZ R complementing E. coli where mutations in the lacZ′ initiation codon were detected by screening for lacZ′ activity. Modified base induced mutations gave rise to growth of white bacterial colonies. In the absence of mutation the colonies were blue. Figure 3A shows the percentage of mutant (white) colonies produced when double-stranded vectors were replicated in the four different human cell lines 293, SO, H6, and LoVo. For control plasmids, (i.e., adduct free, double-stranded plasmids) the fraction of lacZ′ inactivating mutations was very low, as expected. The presence of a single O6-substituted guanine only slightly increased the frequency of mutations over that for the control. However, the percentage of mutants produced by m4T was significantly higher than that for any of the O6substituted guanines in these vectors. Figure 3B shows the percentage of mutant colonies produced when double-stranded vectors were replicated in cells that had been pretreated with 50 µM O6benzylguanine to inactivate alkyltransferase. As ex-

Chem. Res. Toxicol., Vol. 14, No. 7, 2001 897

pected, alkyltransferase inactivation had no effect on the percentage of mutants produced with the control plasmids. However, the percentage of mutants was substantially increased for the three O6-substituted guanines indicating that all these modified guanines were readily repaired by alkyltransferase. In contrast, the fraction of mutants produced by the m4T-containing vectors did not increase after alkyltransferase inactivation indicating that m4T is not subject to significant alkyltransferase repair. Figure 3B also indicates that after alkyltransferase inactivation in the two mismatch repair defective cell lines H6 and LoVo, there was a greater increase in the percentage of mutant colonies produced by m6G- and e6G-containing vectors than there was in the mismatch repair competent 293 and SO cells. The percentage was increased approximately 3-fold for m6G and 2-fold for e6G in the H6 and LoVo cells compared to the 293 and SO cells. No similar effect was observed with b6G or m4T indicating that both m6G and to a lesser extent e6G residues were subject to mismatch repair processing in 293 and SO cells but b6G and m4T residues were not. Gapped plasmids were constructed by inserting four uracil residues in the complementary strand in the sequence surrounding the modified base (Figure 1C). Cellular uracil-DNA glycosylase would be expected to create several abasic sites in these plasmids at the sites of the uracil residues which would subsequently lead to formation of gaps in the complementary strand as a result of action by AP endonuclease. This would promote translesional synthesis past the modified bases and eliminate possible preferential DNA synthesis over the normal complementary strand containing thymine resdiues (11). The percentages of mutant colonies produced by these gapped vectors in alkyltransferase competent cells are shown in Figure 4A. Although we observed a slight increase in the percentage of mutants produced by gapped control plasmids compared to double-stranded control plasmids (Figure 3, panels A and B), the increase in percentage of mutants produced by the four modified bases was much greater with gapped vectors. For the case of gapped vectors containing m4T, the percentage of mutants increased to near 100%. This indicates that m4T essentially always directs incorporation of a base other than adenine during translesional DNA synthesis and indicates that m4T is not repaired in these gapped plasmids. The very high mutation frequency also indicates that cellular processing of these vectors provides an effective means of forcing translesional DNA synthesis over the modified bases. The lower mutation frequencies with the O6-modified guanines reflects their reduced miscoding potency compared to m4T. The results with gapped vectors in alkyltransferase inactivated cells are shown in Figure 4B. Under these conditions, the O6-substituted guanines produced higher percentages of mutants indicating again that these bases are readily subject to alkyltransferase repair as was observed with double-stranded vectors (Figure 3). Unfortunately, it is not possible from the data of Figure 4 to determine whether alkyltransferase repair occurs prior to, during or after gap processing in these vectors. Additionally, our data suggest that nucleotide excision repair is not operating on these gapped plasmids. Our system would detect possible excision repair with these vectors only if excision of the modified base-containing strand occurred prior to gap filling. This would reduce mutagenesis since normal wild-type coding information

898

Chem. Res. Toxicol., Vol. 14, No. 7, 2001

Pauly and Moschel

gapped control plasmids replicated in 293 cells. With the latter plasmids, mutations were primarily deletions of the region surrounding the ATG initiation codon together with point mutations at sites outside the initiation codon. For the O6-substituted guanines, a GfA transition at the site of the modified base was the most commonly observed mutation in all cases. However, mutagenesis by b6G residues in double stranded vectors in LoVo cells was very low (Figure 3). Therefore, a large fraction of mutations resulting from deletions similar to those seen with the control plasmids were also observed with the b6Gcontaining plasmids. We also observed an infrequent 5′TfC transition at the site of the T residue 5′ to the b6G. This same 5′-TfC transition was also observed with an e6G-containing vector. Although mutations at sites adjacent to b6G and e6G residues have been identified in our previous site-specific mutagenesis experiments in rat cells (2, 6), we have not observed these mutations in E. coli (10-13). With m4T containing vectors, TfC transitions were most common. However, in LoVo cells, m4T residues also induced TfA transversions that were especially prevalent with m4T in gapped vectors. Irrespective of the modified base, we occasionally observed a targeted transition that was accompanied by a one or two nucleotide deletion within five nucleotides of the modified base. Point mutations were also observed at sites distant (10 nucleotides or more) from the modified base. Since these were also seen with control plasmids (see above) they were presumed to be unrelated to the modified base. We also noted infrequent insertions of unidentified sequences, which were also regarded as unrelated to the presence of a modified base. Figure 4. Percentage of white (mutant) colonies produced following passage of gapped control and adduct containing shuttle vectors through 293 cells (solid bars), mismatch repair competent SO colon tumor cells (open bars), mismatch repair defective H6 colon tumor cells (hatched bars) or mismatch repair defective LoVo colon tumor cells (crosshatched bars). Panel A shows the results of experiments in cells not pretreated with O6-benzylguanine, and panel B shows results of experiments with cells pretreated with 50 µM O6-benzylguanine to deplete alkyltransferase. All data are the average of three transfections. Error bars represent the standard deviation.

would be provided by the uracil-containing strand. The fact that m4T mutagenesis is nearly 100% indicates that resynthesis of the complement across from the m4T residue resulting in formation of an m4T/G or perhaps an m4T/T basepair must occur much faster than nucleotide excision repair. Additionally, excision repair would not be expected to occur on these gapped substrates while the segment surrounding the modified base is single stranded because nucleotide excision repair requires a double-stranded substrate. Finally, possible excision repair of the modified strand after gap filling would not be observable in our system since it would not alter the sequence of the gap-filled complement and would not alter the observed mutation frequency. Therefore, if we assume that in the absence of alkyltransferase there is no other repair mechanism operative for these four modified bases in gapped plasmids, then the data of Figure 4B suggest that the miscoding potency of the four bases follows the order m4T > e6G = m6G > b6G. The types of mutations induced by these modified bases in double-stranded and gapped vectors replicated in either 293 or LoVo cells (after alkyltransferase inactivation) are summarized in Table 1 together with data for

Discussion It is well-known that during translesional DNA synthesis, O6-substituted guanine residues code for the incorporation of thymine in the complementary strand while O4-methylthymine residues code for incorporation of guanine (1-9). This results in the expected GfA and TfC transitions induced by each modified base, respectively. Our data for human cells (Table 1) and E. coli (1013) are clearly consistent with these expectations. We recognize that the DNA polymerase system in human cells that encounters a modified base in a gapped vector may not be the same polymerase responsible for translesional synthesis in double-stranded vectors and that base insertion across from a modified base may be highly polymerase dependent. Shibutani et al. (27) have demonstrated this for insertion of cytosine vs adenine opposite 8-oxoguanine residues in vitro. Although our data for O6-substituted guanines and m4T in human cells indicate that the identity of the base inserted across from these modified bases is not polymerase sensitive, the frequency of this insertion may be. Such sensitivity may explain why mutagenesis by m4T in double-stranded vectors is less than 50%. This is the theoretical maximum mutation frequency for a modified base in a doublestranded context that is not repaired and always codes for a base that differs from that coded for by the unmodified parent. Alternatively, it may be that m4T is subject to nucleotide excision repair in double-stranded vectors in human cells as has been suggested by Klein et al. (9) or that it induces preferential DNA synthesis using the unmodified strand as a template (a strand bias

m6G, e6G, b6G, and m4T Mutagenesis in Human Cells

Chem. Res. Toxicol., Vol. 14, No. 7, 2001 899

Table 1. Mutations Identified in Shuttle Vectors Replicated in Either 293 or LoVo Cells modified base

vector type

GfA

GfT

5′-TfC

TfC

TfA

other point mutations

compound mutations

deletions

insertions

293 cells m 6G e6G b6G m4 T control m6 G e6G b6G m4 T

dba db db db gappede gapped gapped gapped gapped

50 50 47

m6G e6G b6G m4 T m6 G e6G b6G m4 T

db db db db gapped gapped gapped gapped

50 49 29

1b

1 47

7f 47 47 47

2d 1b 1h 1h 2d

1j 2c

1k

19g 1c

2 46

1c 1c 12g 1c

1i 2i

LoVo cells

50 50 47

1 1b

1 46

2

42

8

2g

1

a Double-stranded vector containing thymine residues in the complementary strand (Figure 1C). b Point mutation(s) more than 10 nucleotides away from the initiation codon. c Large deletions (20-250 bp). d TfC at the site of an m4T plus a one or two base deletion within five basepairs. e A vector containing uracil residues in the complementary strand (Figure 1C). f Six remote point mutations (see footnote b, above) and a single AfG in the initiation codon. g A mix of both small (1-5 bp) and large (20-250 bp) deletions. h GfA at the site of a modified guanine plus a single or tandem CfT within five basepairs. i Insertions of 1-4 kb of unidentified sequence. j Deletion of the ATG initiation codon. k Dual AfT transversion in and near the initiation codon, i.e., ATATGfTTTTG.

effect). Strand bias effects are commonly observed with double-stranded vectors in E. coli (10-13, 28) although not for m4T residues (12). Significantly, we observed that m4T in human cells can produce TfA transversion mutations in addition to TfC transitions especially in the mismatch repair defective LoVo cells. Klein et al. (8) noted similar TfA transversions with O4-ethylthymine in HeLa cells. The fact that the m4T-induced transversions were much more prevalent in a mismatch repair defective cell line suggests that the human mismatch repair system effectively recognizes and removes initially formed m4T/T mismatches. In the absence of repair and when incorporated into gapped vectors, the miscoding potency of the four modified bases in human cells was very similar to that which we observed in E. coli (10-13). Thus, m4T was the most miscoding base followed by e6G and m6G, which were essentially equipotent, and all three were more potent miscoding bases than b6G. Although the former modified guanines coded for T incorporation a bit less effectively in human cells (75-90%) (Figure 4B) than in E. coli (9096%) (10-13), they still exhibited a marked preference for T vs C incorporation in the complementary strand. This was not the case for b6G residues in E. coli or human cells. In E. coli, the preference for coding for T vs C was near 70% (10-13) compared to 45-60% in human cells (Figure 4B). These observations indicate that this aspect of mutagenesis by these modified bases in human cells differs from that observed in E. coli albeit subtly. Our data for the repairability of the O6-substituted guanines and the lack of repair of O4-methylthymine by alkyltransferase in human cells (Figures 3 and 4) is consistent with the known sensitivity of repair of these bases to the human protein (14, 15). The data of Altshuler et al. (7) also indicated that m6G was subject to alkyltransferase repair in CHO cells but m4T was not. We found in addition, however, that both e6G and b6G are much more sensitive to alkyltransferase repair in human cells than in E. coli (10-13). In E. coli, only m6G mutagenesis was modulated by alkyltransferase repair

mechanisms (10-13) while in human cells, mutagenesis by all three O6-substituted guanines was enhanced by alkyltransferase inactivation. Another difference in the mechanism of mutagenesis by these bases in human cells vs E. coli was revealed by our experiments in the mismatch repair defective lines. Interestingly, m6G and to a lesser extent e6G were both apparently sensitive to mismatch repair processing in the 293 and SO human cells although only m6G was subject to mismatch repair processing in E. coli (10-13, 16). The sensitivity of m6G to mismatch processing in human cells is well-known (17-19, 29, 30) as is the fact that cells that are mismatch repair deficient are tolerant to the toxic effects of methylating agents that produce m6G lesions in DNA (29, 30). Models to explain genotoxic or “futile” mismatch repair suggest that the repair system attempts to correct an apparently mismatched basepair (e.g., an m6G/T basepair) by first excising a segment of the newly synthesized T-containing strand and resynthesizing a new complement across from the m6G lesion. This leads to creation of a new m6G/T “mispair” which stimulates additional “futile” rounds of excision and resynthesis that is regarded as genotoxic and ultimately lethal. In mismatch repair deficient cells, there is no attempt made to correct m6G/T mispairs which results in increased survival at the expense of increased mutation (10, 29, 30). This mechanism explains the increased mutagenesis we observed for both m6G and e6G residues when incorporated in double-stranded vectors in H6 and LoVo cells (Figure 3). Interestingly, e6G damage has not been thought to be processed by mismatch repair systems in mammalian cells since mismatch repair deficient cells treated with ethylating agents that produce e6G in DNA have generally not exhibited a significantly enhanced tolerance to ethylation damage (31-35). Therefore, it was surprising that our shuttle vector system indicated that e6G residues were also processed by mismatch repair. On this basis, it seems important to reexamine tolerance to e6G damage in a greater variety of mismatch repair deficient cells to support our conclusion that both e6G and

900

Chem. Res. Toxicol., Vol. 14, No. 7, 2001

m6G are subject to mismatch repair processing in human cells.

References (1) Ellison, K. S., Dogliotti, E., Connors, T. D., Basu, A. K., and Essigmann, J. M. (1989) Site-specific mutagenesis by O6-alkylguanines located in the chromosomes of mammalian cells: Influence of the mammalian O6-alkylguanine-DNA alkyltransferase. Proc. Natl. Acad. Sci. U.S.A. 86, 8620-8624. (2) Mitra, G., Pauly, G. T., Kumar, R., Pei, G. K., Hughes, S. H., Moschel, R. C., and Barbacid, M. (1989) Molecular analysis of O6substituted guanine-induced mutagenesis of ras oncogenes. Proc. Natl. Acad. Sci. U.S.A. 86, 8650-8654. (3) Kamiya, H., Miura, K., Ohtomo, N., Nishimura, S., and Ohtsuka, E. (1991) Transforming activity of a synthetic c-Ha-ras gene containing O6-methylguanine in codon 12. Jpn. J. Cancer Res. 82, 997-1002. (4) Pletsa, V., Gentil, A., Margot, A., Armier, J., Kyrtopoulos, S. A., and Sarasin, A. (1992) Mutagenesis by O6-meG residues within codon 12 of the human Ha-ras proto-oncogene in monkey cells. Nucleic Acids Res. 20, 4897-4901. (5) Bishop, R. E., Dunn, L. L., Pauly, G. T., Dolan, M. E., and Moschel, R. C. (1993) The role of O6-alkylguanine-DNA alkyltransferase in protecting Rat4 cells against the mutagenic effects of O6substituted guanine residues incorporated in codon 12 of the H-ras gene. Carcinogenesis 14, 593-598. (6) Bishop, R. E., Pauly, G. T., and Moschel, R. C. (1996) O6Ethylguanine and O6-benzylguanine incorporated site-specifically in codon 12 of the rat H-ras gene induce semi-targeted as well as targeted mutations in Rat4 cells. Carcinogenesis 17, 849-856. (7) Altshuler, K. B., Hodes, C. S., and Essigmann, J. M. (1996) Intrachromosomal probes for mutagenesis by alkylated DNA bases replicated in mammalian cells: a comparison of the mutagenicities of O4-methylthymine and O6-methylguanine in cells with different DNA repair backgrounds. Chem. Res. Toxicol. 9, 980-987. (8) Klein, J. C., Bleeker, M. J., Lutgerink, J. T., van Dijk, W. J., Brugghe, H. F., van den Elst, H., van der Marel, G. A., van Boom, J. H., Westra, J. G., Berns, A. J. M., and Kriek, E. (1990) Use of shuttle vectors to study the molecular processing of defined carcinogen-induced DNA damage: mutagenicity of single O4ethylthymine adducts in HeLa cells. Nucleic Acids Res. 18, 41314137. (9) Klein, J. C., Bleeker, M. J., Roelen, H. C. P. F., Rafferty, J. A., Margison, G. P., Brugghe, H. F., van den Elst, H., van der Marel, G. A., van Boom, J. H., Kriek, E., and Berns, A. J. M. (1994) Role of nucleotide excision repair in processing of O4-alkylthymines in human cells. J. Biol. Chem. 269, 25521-25528. (10) Pauly, G. T., Hughes, S. H., and Moschel, R. C. (1994) Response of repair-competent and repair-deficient Escherichia coli to three O6-substituted guanines and involvement of methyl-directed mismatch repair in the processing of O6-methylguanine residues. Biochemistry 33, 9169-9177. (11) Pauly, G. T., Hughes, S. H., and Moschel, R. C. (1995) Mutagenesis in Escherichia coli by three O6-substituted guanines in double-stranded or gapped plasmids. Biochemistry 34, 89248930. (12) Pauly, G. T., Hughes, S. H., and Moschel, R. C. (1998) Comparison of mutagenesis by O6-methy- and O6-ethyl-guanine and O4methylthymine in E. coli using double-stranded and gapped plasmids. Carcinogenesis 19, 457-461. (13) Pauly, G. T., Hughes, S. H., and Moschel, R. C. (1991) A sectored colony assay for monitoring mutagenesis by specific carcinogenDNA adducts in Escherichia coli. Biochemistry 30, 11700-11706. (14) Pegg, A. E., Dolan, M. E., and Moschel, R. C. (1995) Structure, function, and inhibition of O6-alkylguanine-DNA alkyltransferase. In Progress in Nucleic Acid Research and Molecular Biology (Cohn, W. E., and Moldave, K., Eds.) Vol. 51, pp 167-223, Academic Press, New York. (15) Zak, P., Kleibl, K., and Laval, F. (1994) Repair of O6-methylguanine and O4-methylthymine by the human and rat O6-methyguanine-DNA methyltransferases. J. Biol Chem. 269, 730-733. (16) Karran, P., and Marinus, M. G. (1982) Mismatch correction at O6-methylguanine residues in E. coli DNA. Nature 296, 868-869.

Pauly and Moschel (17) Branch, P., Aquilina, G., Bignami, M., and Karran, P. (1993) Defective mismatch binding and a mutator phenotype in cells tolerant to DNA damage. Nature 362, 652-654. (18) Kat, A., Thilly, W. G., Fang, W.-H., Longley, M. J., Li, G.-M., and Modrich, P. (1993) An alkylation-tolerant, mutator human cell line is deficient in strand-specific mismatch repair. Proc. Natl. Acad. Sci. U.S.A. 90, 6424-6428. (19) Branch, P., Hampson, R., and Karran, P. (1995) DNA mismatch binding defects, DNA damage tolerance, and mutator phenotypes in human colorectal carcinoma cell lines. Cancer Res. 55, 23042309. (20) Pauly, G. T., Powers, M., Pei, G. K., and Moschel, R. C. (1988) Synthesis and properties of H-ras DNA sequences containing O6substituted 2′-deoxyguanosine residues at the first, second or both positions of codon 12. Chem. Res. Toxicol. 1, 391-397. (21) Parsons, R., Li, G.-M., Longley, M. J., Fang, W.-H., Papadopoulos, N., Jen, J., de la Chappelle, A., Kinzler, K. W., Vogelstein, B., and Modrich, P. (1993) Hypermutability and mismatch repair deficiency in RER+ tumor cells. Cell 75, 1227-1236. (22) Papadopoulos, N., Nicolaides, N. C., Wei, Y.-F., Ruben, S. M., Carter, K. C., Rosen, C. A., Haseltine, W. A., Fleischmann, R. D., Fraser, C. M., Adams, M. D., Venter, J. C., Hamilton, S. R., Petersen, G. M., Watson, P., Lynch, H. T., Peltomaki, P., Mecklin, J.-P., de la Chappelle, A., Kinzler, K. W., and Vogelstein, B. (1994) Mutation of a mutL homolog in hereditary colon cancer. Science 263, 1625-1629. (23) Umar, A., Boyer, J. C., Thomas, D. C., Nguyen, D. C., Risinger, J. I., Boyds, J., Ionov, Y., Perucho, M., and Kunkel, T. A. (1994) Defective mismatch repair in extracts of colorectal and endometrial cancer cell lines exhibiting microsatellite instability. J. Biol. Chem. 269, 14367-14370. (24) Grant, S. G. N., Jessee, J., Bloom, F. R., and Hanahan, D. (1990) Differential plasmid rescue from transgenic mouse DNAs into Escherichia coli methylation-restriction mutants. Proc. Natl. Acad. Sci. U.S.A. 87, 4645-4649. (25) Dolan, M. E., Moschel, R. C., and Pegg, A. E. (1990) Depletion of mammalian O6-alkylguanine-DNA alkyltransferase activity by O6-benzylguanine provides a means to evaluate the role of this protein in protection against carcinogeneic and therapeutic alkylating agents. Proc. Natl. Acad. Sci. U.S.A. 87, 5368-5372. (26) Peden, K. W. C., Pipas, J. M., Pearson-White, S., and Nathans, D. (1980) Isolation of mutants of an animal virus in bacteria. Science 209, 1392-1396. (27) Shibutani, S., Takeshita, M., and Grollman, A. P. (1991) Insertion of specific bases during synthesis past the oxidation-damaged base 8-oxodG. Nature 349, 431-434. (28) Koffel-Schwartz, N., Maenhaut-Michel, G., and Fuchs, R. P. P. (1987) Specific strand loss in N-2-Acelylaminofluorene-modified DNA. J. Mol. Biol. 193, 651-659. (29) Aquilina, G., Biondo, R., Dogliotti, E., and Bignami, M. (1993) Genetic consequences of tolerance to methylation DNA damage in mammalian cells. Carcinogenesis 14, 2097-2103. (30) Karran, P., and Hampson, R. (1996) Genomic instability and tolerance to alkylating agents. Cancer Surveys 28, 69-85. (31) Ishida, R., and Takahashi, T. (1987) N-Methyl-N′-nitro-N-nitrosoguanidine-resistant HeLa S3 cells still have little O6methylguanine-DNA methyltransferase activity and are hypermutable by alkylating agents. Carcinogenesis 8, 1109-1113. (32) Aquilina, G., Zijno, A., Moscufo, N., Dogliotti, E., and Bignami, M. (1989) Tolerance to methylnitrosourea-induced DNA damage is associated with 6-thioguanine resistance in CHO cells. Carcinogenesis 10, 1219-1223. (33) Karran, P., and Bignami, M. (1992) Self-destruction and tolerance in resistance of mammalian cells to alkylation damage. Nucleic Acids Res. 20, 2933-2940. (34) Galloway, S. M., Greenwood, S. K., Hill, R. B., Bradt, C. I., and Bean, C. L. (1995) A role for mismatch repair and production of chromosome aberrations by methylating agents in human cells. Mutat. Res. 346, 231-245. (35) Armstrong, M. J., and Galloway, S. M. (1997) Mismatch repair provokes chromosome aberrations in hamster cells treated with methylating agents or 6-thioguanine, but not with ethylating agents. Mutat. Res. 373, 167-178.

TX010032F