¹Centre de recherche en cancérologie de l'Université Laval, Pavillon L'Hôtel-Dieu de Québec, Centre hospitalier universitaire de Québec, Québec G1R 2J6 Canada

²Département de biologie, Université Laval, Québec G1K 7P4 Canada

Correspondence to: Alan Anderson, Centre de recherche, L'Hotel-Dieu de Québec, 11 côte du Palais, Québec, Canada GIR 2J6

p53 has been postulated to be the guardian of the genome. However, results supporting the prediction that point mutation frequencies are elevated in p53-deficient cells either have not been forthcoming or have been equivocal. To analyse the effect of p53 on point mutation frequency, we used the supF gene of the pYZ289 shuttle vector as a mutagenic target. pYZ289 was treated in vitro by ultraviolet irradiation, aflatoxin B₁, (±)7,8-dihydroxy-9,10-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene and meta-chloroperoxybenzoic acid and then transfected into p53-deficient cells with or without a p53 expression vector. p53 reduced the mutant frequency up to fivefold when pYZ289 was treated with aflatoxin B₁, (±)7,8-dihydroxy-9,10-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene or meta-chloroperoxybenzoic acid but not when it was ultraviolet-irradiated. The p53-dependent mutation frequency reduction was higher at a higher level of premutational lesions for aflatoxin B₁ and (±)7,8-dihydroxy-9,10-epoxy-7,8,9,10-tetrahydrobenzo[a]pyrene and at a lower level of lesions for meta-chloroperoxybenzoic acid. This suggests that the chemical mutagens produce, in a dose-dependent fashion, two kinds of DNA damage, one subject to p53-dependent mutation frequency reduction and the other not. These results indicate that p53 can reduce the point mutation frequency in a shuttle vector treated by chemical mutagens and suggest that p53 can act as guardian of the genome for at least some kinds of point mutations.

p53; mutation frequency; shuttle vector

Introduction

p53 mutations are the most common specific genetic change in human cancer cells (Selivanova and Wiman, 1995). Furthermore, both humans (Frebourg and Friend, 1992) and mice (Harvey et al., 1993a) heterozygous in the germ line for a p53 null allele have a high incidence of spontaneous tumors in a wide spectrum of tissues, and mice homozygous for a null allele are even more tumor prone (Donehower et al., 1992). p53 deficiency permits gene amplification and leads to an enhanced frequency of chromosome abnormalities and aneuploidy (Livingstone et al., 1992; Yin et al., 1992; Harvey et al., 1993b; Bouffler et al., 1995), which may explain in part the genomic instability of cancer cells (Hartwell, 1992). DNA damage, and particularly DNA strand breaks, induces a transient increase in the level of p53 in wild type mammalian cells (Maltzman and Czyzyk, 1984; Kastan et al., 1991; Hall et al., 1993; Lu and Lane, 1993; Nelson and Kastan, 1994) due largely to an increase in p53 stability (Oren et al., 1981; Maltzman and Czyzyk, 1984; Giaccia and Kastan, 1998; Lane, 1998). In cells containing normal p53, but not in those lacking it, DNA damage can lead to a transitory block in the cell cycle (G1 arrest) (Kastan et al., 1991; Kuerbitz et al., 1992). Since cells lacking functional p53 continue to divide after treatment with DNA damaging agents, it has been sugested that they would accumulate genetic damage because of their inability to repair DNA damage before entry into S phase. Lane (1992) proposed that p53 monitors the integrity of the genome such that when DNA damage occurs, p53 accumulates, and a p53-mediated G1 cell cycle checkpoint switches off replication to permit repair. Accordingly, tumor cells in which p53 is inactivated by mutation or by binding to viral proteins would be less stable genetically and would accumulate rearrangements and mutations. However, present data do not support the view that genomic instability due to p53 deficiency is a consequence of the cell's failure to repair DNA damage before undertaking DNA replication and suggest rather that genomic instability arises from a defect in one of the other p53-dependent cellular pathways (Murnane and Schwartz, 1993; Smith and Fornace, 1996).

Although the guardian of the genome hypothesis receives support from the observation that p53-deficient cells are more subject than their p53-positive counterparts to gene amplification, chromosome aberrations and aneuploidy (Livingstone et al., 1992; Yin et al., 1992; Harvey et al., 1993b; Bouffler et al., 1995), results supporting the prediction that point mutation frequencies are elevated in p53-deficient cells either have not been forthcoming or have been equivocal. It has been reported that p53 has no effect on spontaneous lacI mutation frequencies in wild type and p53-negative transgenic mice (Nishino et al., 1995), on spontaneous and 4-nitroquinoline-1-oxide-induced lacI mutations in wild type and p53-deficient mouse embryonic fibroblasts (Sands et al., 1995), on spontaneous hprt mutations in p53-inducible human osteosarcoma cells (Yamagishi et al., 1997b), or on ultraviolet (UV)-induced supF mutations in a shuttle vector in wild type and p53-deficient mouse embryonic fibroblasts (Ishizaki et al., 1996). Furthermore, there is no X-ray-induced hypermutability of the hprt gene in B cell precursors from p53^-/p53^- mice (Griffiths et al., 1997). On the other hand, p53 was reported to reduce UV-induced hprt mutations in human osteosarcoma cells (Yamagishi et al., 1997b; Yagi et al., 1998) and in a supF target gene in an integrated genome (Yuan et al., 1995) and in a shuttle vector (Yamagishi et al., 1997b).

To study the effect of p53 on point mutation frequency, we used the pYZ289 shuttle vector system (Moriwaki et al., 1991) in which the plasmid DNA was treated in vitro with four mutagens, (±)7,8-dihydroxy-9,10-epoxy-7,8,9,10-tetrahydrobenzo[a] pyrene(BPDE), aflatoxin B₁ (AFB1), meta-chloroperoxy-benzoic acid (CPBA), and UV irradiation at 254 nm (UVC). BPDE, AFB1 and UVC are well-known point mutagens. BPDE and AFB1 generate bulky DNA adducts, largely on guanine residues, that lead mainly to guanine substitutions (Muench et al., 1983; Yang et al., 1987). CPBA is a known DNA-damaging agent (Jacobsen and Humayun, 1986; Rosenthal et al., 1990), but its mutagenic effects have apparently not been observed previously. We found that p53 reduces the frequency of supF point mutations induced by BPDE, AFB1 and CPBA, but not by UVC treatment. The p53-dependent mutation frequency reduction (MFR) was dependent on the mutagen dose. These results suggest that the guardian of the genome hypothesis is applicable to point mutations under some conditions.

Results

Transient expression of p53 in a p53-deficient cell line does not lead to increased apoptosis

The p53-deficient cell line 10(3) does not synthesize an active p53 protein (Harvey and Levine, 1991). p53 was detected by Western blot analysis after transient transfection of 10(3) cells with the pCMV53k expression vector. p53 was present at a low level 6 h after transfection, reached a peak at 24 h, and declined thereafter to reach undectable levels at 72 h (Figure 1a). Immunostaining of p53 after transient transfection with pECM53k revealed that p53 was present mostly in the nucleus and that a small quantity was also present in the cytoplasm of positive cells (Figure 1b). The transfection efficiency was 8 - 10%. Similar immuno-staining results were obtained after transient transfection with pCMV53k, except that the p53 levels were higher (data not shown).

Expression vector-driven synthesis of p53 in p53-deficient cells can in some cases lead to increased apoptosis, depending on p53 levels and the presence of DNA damage (Chen et al., 1996). To determine whether expression vector-derived p53 increased apoptosis in our system, 10(3) cells were transfected with pECM53k (or pECMmutk) or cotransfected with pECM53k and BPDE-treated pYZ289 and evaluated for apoptosis after 24 and 48 h. In all cases less than 0.2% of cells bound Annexin V and exhibited morphological changes, such as plasma membrane blebbing and nuclear fragmentation, typical of apoptosis (data not shown). The 10(3) cells could indeed undergo apoptosis, because essentially all cells bound Annexin V and showed characteristic apoptotic changes after exposure to actinomycin D for 24 h and were dead at 48 h (data not shown). Furthermore, transfected cells immunostained for p53 did not exhibit any morphological changes characteristic of apoptotic cells (Figure 1b). Hence, under the conditions of our experiments, expression vector-driven synthesis of p53 did not lead to a detectable increase in the frequency of apoptosis.

p53 reduces supF mutation frequency after treatment with some mutagens

The pYZ289 shuttle vector was treated with AFB1, BPDE, CPBA or UVC, purified and cotransfected into 10(3) cells with the pECM53k p53 expression vector or with the pECMmutk control vector. After three days, replicated shuttle vectors were purified and electroporated into E. coli MBM7070, the bacteria were plated on medium containing X-Gal, and supF mutants were identified as white colonies. The apparent p53-dependent MFR was calculated as the frequency of white colonies observed for pYZ289 from pECMmutk-transfected cells over that for pYZ289 from pECM53k-transfected cells.

Of the four agents tested, three exhibited p53-dependent MFR that varied with the dose of mutagen (Figure 2). AFB1 and BPDE gave apparent p53-dependent MFR values that varied from undetectable at lower white colony frequencies to 4.3 (for AFB1) or 4.7 (for BPDE) at higher white colony frequencies. For CPBA on the other hand, the apparent p53-dependent MFR was highest at the lowest white colony frequency and decreased as the white colony frequency increased. With UVC, no apparent p53-dependent MFR was detected at any white colony frequency. This latter result is similar to that reported previously for the pYZ289 shuttle vector (Ishizaki et al., 1996).

The point mutation frequency differs from the white colony frequency

Mutant supF genes were sequenced to permit calculation of the point mutation frequency and the generation of point mutation spectra. Mutants with deletions of >3 bp and with multiple mutations separated by >3 bp were excluded. Putative sibling mutations, i.e., identical mutations found in plasmids isolated from the same transfection dish, were also excluded; to reduce the frequency of such putative siblings, each experiment was done in 6 - 12 different transfection dishes and the mutants were analysed separately. The sequence analysis made it possible to calculate the point mutation frequency, and this had the effect of altering some of the apparent p53-dependent MFR values calculated previously from the initial white colony frequency (Table 1).

For the UVC-treated shuttle vector there was no appreciable difference in the p53-dependent MFR as calculated for point mutations in comparison to apparent MFR (Table 1). This was not unexpected since there was no apparent p53-dependent MFR at any UVC dose tested (Figure 2). AFB1-treated pYZ289 carrying 63 adducts per plasmid had an apparent p53-dependent MFR of 3.7 that increased to 6.6 when only point mutations were considered (Table 1). In the presence of p53 there were relatively fewer point mutations so the other kinds of mutations were proportionally more represented. For BPDE-treated pYZ289, the lower dose was such that no apparent p53 response was evident, whereas the higher dose gave an apparent p53-dependent MFR of 3.4 (Figure 2). However, when only point mutations were considered, BPDE gave p53-dependent MFR values of 2.5 and 4.9 at the lower and higher doses respectively (Table 1) again indicating that in the presence of p53 there were relatively fewer point mutations.

The presence of p53 does not affect the kinds of point mutations

The kinds of supF point mutations generated by UVC, AFB1 and BPDE were not affected appreciably by the presence of p53 (Table 2). The kinds of point mutations found after UVC treatment of the shuttle vector were typical of UVC-induced mutations in this system (Ishizaki et al., 1996). There was a preponderance of G:C to A:T transitions (73%) and a large proportion of tandem substitutions (45%). For AFB1, the kinds of point mutations were similar to those observed in our previous work (Courtemanche and Anderson, 1994). There was a total of 89% of guanine substitutions, the most frequent of which was the G:C to T:A transversion (56%). Most guanine substitutions (84%) affected only a single base pair. BPDE, a mutagen known to exhibit dose-dependent differences in the profile of mutations (Wei et al., 1993), was tested at two doses. The distribution of mutations followed those obtained previously (Wei et al., 1993, 1994) for low and high doses of (+)-BPDE and (-)-BPDE in that there was a tendency for G:C to T:A transversions to be more frequent at the higher dose (51%) than at the lower dose (43%). Frameshifts (deletions of 1 - 2 bp plus insertions) tended to be less frequent at the higher dose (10%) than at the lower dose (18%) and there was a significant increase in the proportion of tandem base mutations at the higher dose (45%) compared to the lower dose (28%). For AFB1 and BPDE, the relatively high proportions of tandem mutations as compared to previous reports (Courtemanche and Anderson, 1994; Yang et al., 1987; Wei et al., 1994) may be related to the use of the pYZ289 shuttle vector or the 10(3) cells or both.

p53 has subtle effects on point mutational spectra

Mutational spectra were determined for supF point mutations generated by UVC, AFB1 and BPDE (Figure 3). The distribution of the substitutions was not random and 6 - 9 hotspots, representing 5% of all mutations, were observed per spectrum. The only mutational spectra showing statistically significant differences in the presence and absence of p53 were those for BPDE at the higher dose (Figure 3d; ² test, P=0.025). Nevertheless, several hot spots changed in the presence of p53. All UVC hotspots were at pyrimidine pairs, and two, at positions 160 and 175 (composed of tandem and 3 bp mutations respectively), disappeared in the presence of p53. Three AFB1 hotspots (at positions 102, 159 and 168) disappeared or were diminished in the presence of p53, while one (at position 141) appeared. At the lower BPDE dose, one hotspot disappeared (at position 111) and two new ones appeared (at positions 127 and 164) in the presence of p53. At the higher BPDE dose, five hotspots (at positions 113, 123, 144, 156 and 168) disappeared and one new one (at position 127) appeared in the presence of p53.

Discussion

In this study we analysed the effect of the p53 tumor suppressor protein on supF mutagenesis in a pYZ289 shuttle vector treated with four different mutagenic agents, AFB1, BPDE, CPBA and UVC, and observed that cotransfection with the p53 expression vector pECM53k can reduce point mutation frequency for AFB1, BPDE and CPBA, but not for UVC.

We used the 10(3) line of p53-deficient mouse fibroblasts (Harvey and Levine, 1991) and the polyoma-based pYZ289 shuttle vector for our experiments. pYZ289 will replicate in rodent cells and polyoma large T antigen does not form a complex with murine p53 under conditions where SV40 large T does (Wang et al., 1989). Western blot and immunofluorescence analyses (Figure 1) demonstrated that p53 accumulated in 10(3) cells transfected with pECM53k or pCMV53k and that the accumulated p53 was primarily nuclear, as is the case for the endogenous protein. The kinetics of p53 accumulation and degradation (Figure 1a) indicated that p53 is relatively unstable and suggested that the normal p53 degradation system (Oren et al., 1981) is active in 10(3) cells. The biological activity of the newly synthesized p53 protein was confirmed by the arrest of the replication of an SV40-based pSP189 shuttle vector in Ad293 cells, as well as by a decrease in the replication of the polyoma-based pYZ289 in 10(3) cells (C Courtemanche and A Anderson, unpublished observations). However, the presence of p53 was not associated with a detectable increase in apoptosis. Hence, the observed effects of p53 on supF mutation frequency are not confounded by preferential death of cells cotransfected with the p53 expression vector and the mutagen-treated shuttle vector. While apoptosis was not observed in cells transfected with the p53 expression vector, it is quite likely that these cells arrest in G1 as a consequence of the presence of high levels of p53, whereas the cells transfected with the control vector continue to replicate. If this does occur it seems to be without effect on the measured supF mutation frequency, given that cotransfection with the p53 expression vector did not affect point mutation frequency for UVC-treated pYZ289.

The shuttle vector mutagenesis system we have used has advantages and disadvantages. In addition to the relative ease with which mutations can be characterized in the shuttle vector context, an important advantage is that since p53 is added to the p53-deficient 10(3) cells by transient expression there can be no extraneous confounding factors due to clonal differences between p53-positive and p53-negative cells. However, one has to recognize that p53 expression vector-driven synthesis in p53-deficient cells will lead to high non-physiological levels of p53. Furthermore, the mutation frequency at the supF locus is measured in a replicating plasmid, as opposed to a native gene within a chromatin context.

Our observations that mutagenesis of supF in pYZ289 by UVC is indifferent to the presence or absence of p53 are in agreement with those of Ishizaki et al. (1996) made with a system very similar to ours. They transfected UV-irradiated pYZ289 into p53-proficient and p53-deficient mouse embryonic fibroblasts and reported neither any difference in the supF mutation frequencies nor any marked changes in the spectrum of base change mutations. Nor indeed are there any marked differences between their UV supF mutation spectra for p53-proficient and p53-deficient embryonic mouse fibroblasts and ours for 10(3) cells transfected with pECM53K or pECMmutK. In the case of UV mutagenesis of pYZ289 therefore, physiological levels or high expression vector-driven levels of p53 gave the same results. This also implies that the p53-dependent MFR observed for AFB1, BPDE and CPBA was not solely a result of high non-physiological p53 levels.

There have been several recent reports of an effect of p53 in reducing mutation frequencies. Havre et al. (1995) reported increased spontaneous mutagenesis at the hprt locus in human cells in which p53 was inactivated by HPV16 E6. However, most of the mutants analysed carried hprt deletions or rearrangements, such that the observed increases in spontaneous mutagenesis may reflect the known increases in genomic instability (Livingstone et al., 1992; Yin et al., 1992; Harvey et al., 1993b) in p53-negative cells. Similarly, Clarke et al. (1997) reported that at high doses of -irradiation there was a 5 - 6-fold increase in the mutation frequency at the endogenous Dlb-1 locus in p53-deficient mice, Havre et al. (1995) reported increased low-dose UV mutagenesis at the hprt locus in human cells in which p53 was inactivated by HPV16 E6, and Yamagishi and coworkers reported p53-dependent suppression of UV-induced (Yamagishi et al., 1997b) or X-ray-induced (Yamagishi et al., 1997a) mutations at the endogenous hprt locus in human osteosarcoma cells, but in none of these cases were the mutations characterized at the molecular level. Results supporting a p53-dependent supression in point mutation frequencies have come from Yuan et al. (1995), who reported that p53 induction in mouse cells caused a fourfold decrease in UV-induced mutagenesis in a supF target gene in an integrated genome, and from Yamagishi et al. (1997b), who reported a modest p53-dependent decrease in UV-induced supF mutations in a shuttle vector passed through human osteosarcoma cells.

In our experiments, mutagen-treated shuttle vectors passed through 10(3) cells exhibited a p53-dependent decrease in white colony frequency and point mutation frequency for AFB1, BPDE and CPBA. AFB1 and BPDE are well-studied genotoxic carcinogens that are known to cause point mutations (Foster et al., 1983; Trottier et al., 1992b; Levy et al., 1992; Yang et al., 1987; Wei et al., 1991, 1993). Although CPBA is a known DNA-damaging agent (Jacobsen and Humayun, 1986; Rosenthal et al., 1990), its mutagenic properties seem not to have been reported previously. None of these agents has been used in studies of possible effects of p53 on point mutation frequency.

The striking observation that p53-dependent MFR was more evident at high doses of AFB1 and BPDE and at low doses of CPBA requires comment. We have considered two kinds of explanation for these results. One possibility is that a certain threshold of DNA damage is required before p53-mediated MFR can be detected. This model implicitly assumes that for a given agent all DNA damage is equivalent with respect to p53-mediated MFR and that a certain level of damage must be attained before the presence of p53 will reduce mutagenesis. A second possibility is that agents such as BPDE, AFB1 and CPBA produce, in a dose-dependent fashion, two kinds of DNA damage, one of which is subject to p53-mediated MFR and the other is not. The second possibility is favored because p53-dependent MFR was observed at low doses but not at high doses of CPBA, a result hard to reconcile with a requirement for a threshold of DNA damage to obtain p53-dependent MFR. In addition, there are dose-dependent differences in the profile of BPDE-induced mutations in the hprt gene of V79 cells (Wei et al., 1991, 1993), suggesting dose-dependent differences in the type of BPDE-induced DNA damage. Furthermore, it is certainly not unreasonable to suspect the existence of (dose-dependent) differences in the type of damage caused by diverse DNA-damaging agents. If indeed there are two kinds of DNA damage, one subject to p53-dependent MFR and the other not, this could help explain why p53-dependent MFR has been observed in some studies and not in others. Also, even though there are differences between the mutational spectra for AFB1 and BPDE in the presence and absence of p53, they are nevertheless relatively subtle which sugests that the major effect of p53 is to reduce the overall mutation frequency rather than to modify the nature of the mutations generated.

In conclusion, we have shown p53 can reduce the supF point mutation frequency in a shuttle vector treated by the chemical mutagens AFB1, BPDE and CPBA. Furthermore, preliminary results from our laboratory indicate that supF mutations induced by hydrogen peroxide and by UV irradiation at 366 nm in the presence of riboflavin are also subject to p53-dependent MFR (C Courtemanche and A Anderson, unpublished observations). Hence, our results indicate that p53 can play the role of guardian of the genome for point mutations, at least for certain kinds of DNA damage. The efficacy of p53-dependent MFR may also depend on the level of DNA damage as well as on the cell type (Griffiths et al., 1997). The kinds of DNA damage subjected to p53-dependent MFR remain to be determined. Another key experiment for the future will be to determine whether AFB1, BPDE and CPBA are also subject to p53-dependent MFR when the mutational target is a cellular gene.

Materials and methods

Cells, plasmids and transfections

The p53-deficient mouse fibroblast cell line 10(3) (Harvey and Levine, 1991) was obtained from AJ Levine. Cells were grown in Dulbecco's modified Eagle's medium containing 4.5 g/l glucose, 10% fetal bovine serum, 100 U/ml penicillin and 100 g/ml streptomycin. Escherichia coli MBM7070, which carries a lacZ amber mutation, was obtained from MM Seidman. pYZ289, which is a polyoma-based shuttle vector suitable for use in mouse cells (Moriwaki et al., 1991), was obtained from T Yagi. The SV40 early promoter-driven pECM53 expression vector for wild type mouse p53 (Johnson et al., 1991), was from S Benchimol. pECM53 was modified by replacing its ampicillin-resistance gene with a kanamycin-resistance gene that does not interfere with the subsequent shuttle vector selection in E. coli, thus obtaining pECM53k. The p53 coding sequence of pECM53k was eliminated to obtain pECMmutk for use as a control p53-negative expression vector. To obtain an expression vector giving a higher level of synthesis, a cytomegalovirus promoter-driven pCMV53k expression vector for wild type mouse p53 was constructed using the p53-coding sequence of pECM53k. Cells were transfected in suspension by a calcium phosphate method (Trottier et al., 1992a).

Western blot analysis

For Western blot analysis, 3´10⁶ cells were transfected with 40 g of pCMV53k and plated in 150-mm culture dishes. Six hours later the calcium phosphate-DNA co-precipitates were removed and the cells were washed and incubated in fresh medium. At various times after transfection, the cells were harvested and lysed on ice in 250 l of Laemmli sample buffer (62 mM Tris HCl pH 6.8, 10% glycerol, 2% sodium dodecyl sulfate (SDS), 5% -mercaptoethanol, 0.003% bromophenol blue, 0.4 mM phenylmethylsulfonyl fluoride) and the DNA was sheared by passage through a 27 gauge needle (Bristow et al., 1994). After boiling for 10 min and centrifugation, the supernatant proteins were assayed (Bradford, 1976) and samples were loaded onto a 10% polyacrylamide gel containing SDS and subjected to electrophoresis (Laemmli, 1970). Proteins were transferred to nitrocellulose paper by electroblotting. The membranes were blocked by treatment with 5% dried milk in 50 mM Tris HCL pH 7.4/150 mM NaCl/0.1% Tween 20 and incubated with anti-p53 antibody PAb421 (Oncogene Science) (1 : 1000 dilution). Immunodetection was carried out using an enhanced chemiluminescence detection kit (Amersham) with horseradish peroxidase-conjugated goat anti-mouse immunoglobulin G (1 : 10 000 dilution) as a secondary antibody.

Immunostaining of p53 and assay for apoptosis

Cells (3´10⁵) were transfected with 3.3 g of pECM53k or pECMmutk or with 2.5 g of BPDE-treated pYZ289 together with 3.3 g of pECM53k, plated on glass coverslips in 35-mm culture dishes and incubated for 24 or 48 h. For immunostaining of p53, the cells were washed with phosphate-buffered saline, fixed in methanol at -20°C for 10 min and incubated for one h with anti-p53 antibody PAb421 (1 : 1000 dilution) and then with Texas Red-conjugated goat anti-mouse immunoglobulin G (Amersham) (1 : 100 dilution). Fluorescence was viewed using a confocal laser microscope (Bio-Rad). For detection of apoptosis, an Annexin-V-FLUOS Staining Kit (Boehringer Mannheim) was used.

In vitro treatment of pYZ289 with mutagens

Thirty g of plasmid was irradiated using a germicidal lamp (254 nm). For AFB1, adducts were generated as previously described (Courtemanche and Anderson, 1994). The reaction mixture, in a final volume of 150 l of 50 mM sodium phosphate buffer pH 7.4, contained pYZ289 (40 g), liver microsomes from an Aroclor 1254-treated male rat (Ames et al., 1975) (75 g protein), NADP (1.6 mM), glucose-6-phosphate (16 mM), glucose-6-phosphate dehydrogenase (0.4 U) and [³H]AFB1 (214 M, 0.16 Ci/mmol, obtained by diluting [³H]AFB1 at 16 Ci/mmol from Moravek Chemicals with unlabeled AFB1 from the US National Cancer Institue Chemical Carcinogen Repository). Incubation was at 37°C for up to 60 min. Plasmid DNA was purified by proteinase K digestion (3 mg/ml), phenol/chloroform extraction and ethanol precipitation. The DNA concentration was determined by fluorimetry (Kapp et al., 1974) and the number of residues bound per plasmid molecule was calculated from the specific activity. Treatment with BPDE (Yang et al., 1987) was performed in 200 l of 10 mM Tris · HCl pH 7.5, 1 mM EDTA containing 45 g of pYZ289. The reaction was initiated by the addition of 1 - 6 l of BPDE (0.3 mg/ml in tetrahydrofuran) and incubation was for 90 min at room temperature with protection from visible light. Unbound BPDE was removed by ethanol precipitation and the DNA concentration determined by fluorimetry. For BPDE, the estimated numbers of adducts per vector were 13 and 108 for the lower and higher doses respectively, as derived by comparison with the supF mutant frequency for a [³H]BPDE-treated plasmid preparation carrying 47 adducts per vector (data not shown). The method for CPBA treatment of pYZ289 was adapted from a procedure used to activate AFB1 chemically (Mariën et al., 1987). The reaction mixture contained 45 g of plasmid DNA in 200 l of 20 mM sodium phosphate buffer, pH 7.4, at 37°C. For each modification, 10 l of a freshly prepared solution of CPBA in methylene chloride (at 2, 20, 60 or 70 mg/ml) was added every 7.5 min for 22.5 min. At 30 min the reaction was stopped by extraction twice with 200 l of chloroform and the modified DNA was precipitated with ethanol and the DNA content determined by fluorimetry.

Characterization of supF mutants

Cells (3´10⁶) were transfected with 30 g of mutagen-treated pYZ289 and 40 g of pECM53k or pECMmutk and incubated in 150-mm culture dishes for 72 h. Plasmid DNA was then isolated (Hirt, 1967), purified by digestion with proteinase K (2 mg/ml) and phenol/chloroform extraction, treated with ribonuclease A (0.2 mg/ml) and digested with DpnI (Peden et al., 1980). Plasmids harvested from individual transfection dishes were purified and analysed separately. The progeny plasmids were electroporated (Dower et al., 1988) into E. coli MBM7070 in cuvettes with a 0.1 cm electrode gap using a Bio-Rad gene pulser (15 kV/cm, 25 Fd, 400 ohm). The bacteria were plated on LB agar plates supplemented with ampicillin (100 g/ml) and covered with LB agar (10 ml per 150 mm diameter dish) containing 5-bromo-4-chloro-3-indolyl--D-galactosidase (X-Gal) (0.8 mg/ml), isopropyl--D-thiogalactoside (190 g/ml), and ampicillin (100 g/ml). White or light blue Lac^- mutant colonies were picked from the background of blue wild type colonies and purified by restreaking. Plasmids from individual mutant colonies were purified by a mini-scale method (Del Sal et al., 1989) and their electrophoretic mobility was analysed on 0.8% agarose gels. The supF DNA sequence of plasmids that co-migrated with the wild type plasmid and of those with minor alterations of their gel electrophoretic pattern was determined (Sanger et al., 1977) using double-stranded plasmid DNA as template and T7 DNA polymerase.

Acknowledgements

We thank T Yagi for the pYZ289 shuttle vector, S Benchimol for the pECM53 p53 expression vector, AJ Levine for the 10(3) cells, and MM Seidman for E. coli MBM7070 as well as for advice and encouragement concerning the use of shuttle vector mutagenesis systems. We also thank M Desrochers for providing the microsomes, A Loranger and N Marceau for advice on immunofluorescence techniques, as well as P Françon and E Pellerin for technical assistance. C Courtemanche held a graduate fellowship from the FCAR of Québec. This work was supported by the Medical Research Council.

Ames BN, McCann J and Yamasaki E. (1975). Mutation Res. 31, 347-364. MEDLINE

Bouffler SD, Kemp CJ, Balmain A and Cox R. (1995). Cancer Res. 55, 3883-3889. MEDLINE

Bradford MM. (1976). Anal. Biochem. 72, 248-254. Article MEDLINE

Bristow RG, Jang A, Peacock J, Chung S, Benchimol S and Hill RP. (1994). Oncogene 9, 1527-1536. MEDLINE

Chen XB, Ko LJ, Jayaraman L and Prives C. (1996). Genes Dev. 10, 2438-2451. MEDLINE

Clarke AR, Howard LA, Harrison DJ and Winton DJ. (1997). Oncogene 14, 2015-2018. MEDLINE

Courtemanche C and Anderson A. (1994). Mutation Res. 306, 143-151. MEDLINE

Del Sal G, Manfioletti G and Schneider C. (1989). Biotechniques 7, 514-519. MEDLINE

Donehower LA, Harvey M, Slagle BL, McArthur MJ, Montgomery CA, Butel JS and Bradley A. (1992). Nature 356, 215-221. MEDLINE

Dower WJ, Miller JF and Ragsdale CW. (1988). Nucleic Acids Res. 16, 6127-6145. MEDLINE

Foster PL, Eisenstadt E and Miller JH. (1983). Proc. Natl. Acad. Sci. USA 80, 2695-2698. MEDLINE

Frebourg T and Friend SH. (1992). J. Clin. Invest. 90, 1637-1641. MEDLINE

Giaccia AJ and Kastan MB. (1998). Genes Dev. 12, 2973-2983. MEDLINE

Griffiths SD, Clarke AR, Healy LE, Ross G, Ford AM, Hooper ML, Wyllie AH and Greaves M. (1997). Oncogene 14, 523-531. MEDLINE

Hall PA, McKee PH, Menage HD, Dover R and Lane DP. (1993). Oncogene 8, 203-207. MEDLINE

Hartwell L. (1992). Cell 71, 543-546. MEDLINE

Harvey DM and Levine AJ. (1991). Genes Dev. 5, 2375-2385. MEDLINE

Harvey M, McArthur MJ, Montgomery CA, Butel JS, Bradley A and Donehower LA. (1993a). Nat. Genet. 5, 225-229. MEDLINE

Harvey M, Sands AT, Weiss RS, Hegi ME, Wiseman RW, Pantazis P, Giovanella BC, Tainsky MA, Bradley A and Donehower LA. (1993b). Oncogene 8, 2457-2467. MEDLINE

Havre PA, Yuan J, Hedrick L, Cho K and Glazer PM. (1995). Cancer Res. 55, 4420-4424. MEDLINE

Hirt B. (1967). J. Mol. Biol. 26, 365-369. MEDLINE

Ishizaki K, Nishizawa K, Mimaki S and Aizawa S. (1996). Mutation Res. 364, 43-49. MEDLINE

Jacobsen JS and Humayun MZ. (1986). Carcinogenesis 7, 491-493. MEDLINE

Johnson P, Gray D, Mowat M and Benchimol S. (1991). Mol. Cell. Biol. 11, 1-11. MEDLINE

Kapp LN, Brown SL and Klevecz RR. (1974). Biochim. Biophys. Acta 361, 140-143. MEDLINE

Kastan MB, Onyekwere O, Sidransky D, Vogelstein B and Craig RW. (1991). Cancer Res. 51, 6304-6311. MEDLINE

Kuerbitz SJ, Plunkett BS, Walsh WV and Kastan MB. (1992). Proc. Natl. Acad. Sci. USA 89, 7491-7495. MEDLINE

Laemmli UK. (1970). Nature 227, 680-685. MEDLINE

Lane DP. (1998). Nature 394, 616-617. Article MEDLINE

Lane DP. (1992). Nature 358, 15-16. MEDLINE

Levy DD, Groopman JD, Lim SE, Seidman MM and Kraemer KH. (1992). Cancer Res. 52, 5668-5673. MEDLINE

Livingstone LR, White A, Sprouse J, Livanos E, Jacks T and Tisty TD. (1992). Cell 70, 923-935. MEDLINE

Lu X and Lane DP. (1993). Cell 75, 765-778. MEDLINE

Maltzman W and Czyzyk L. (1984). Mol. Cell. Biol. 4, 1689-1694. MEDLINE

Mariën K, Moyer R, Loveland P, van Holde K and Bailey G. (1987). J. Biol. Chem. 262, 7455-7462. MEDLINE

Moriwaki S, Yagi T, Nishigori C, Imamura S and Takebe H. (1991). Cancer Res. 51, 6219-6223. MEDLINE

Muench KF, Misrar RP and Humayun MZ. (1983). Proc. Natl. Acad. Sci. USA 80, 6-10. MEDLINE

Murnane JP and Schwartz JL. (1993). Nature 365, 22. MEDLINE

Nelson WG and Kastan MB. (1994). Mol. Cell. Biol. 14, 1815-1823. MEDLINE

Nishino H, Knoll A, Buettner VL, Frisk CS, Maruta Y, Haavik J and Sommer SS. (1995). Oncogene 11, 263-270. MEDLINE

Oren M, Maltzman W and Levine AJ. (1981). Mol. Cell. Biol. 1, 101-110. MEDLINE

Peden KWC, Pipas JM, Pearson-White S and Nathans D. (1980). Science 209, 1392-1396. MEDLINE

Rosenthal A, Sproat BS and Brown DM. (1990). Biochem. Biophys. Res. Commun. 173, 272-275. MEDLINE

Sands AT, Suraokar MB, Sanchez A, Marth JE, Donehower LA and Bradley A. (1995). Proc. Natl. Acad. Sci. USA 92, 8517-8521. MEDLINE

Sanger F, Nicklen S and Coulson AR. (1977). Proc. Natl. Acad. Sci. USA 74, 5463-5467. MEDLINE

Selivanova G and Wiman KG. (1995). Adv. Cancer Res. 66, 143-180. MEDLINE

Smith ML and Fornace Jr AJ. (1996). Am. J. Pathol. 148, 1019-1022. MEDLINE

Trottier Y, Waithe WI and Anderson A. (1992a). Mutation Res. 281, 39-45.

Trottier Y, Waithe WI and Anderson A. (1992b). Mol. Carcinogen. 6, 140-147.

Wang EH, Friedman PN and Prives C. (1989). Cell 57, 379-392. MEDLINE

Wei SJC, Chang RL, Wong CQ, Bhachech N, Cui XX, Hennig E, Yagi H, Sayer JM, Jerina DM, Preston BD and Conney AH. (1991). Proc. Natl. Acad. Sci. USA 88, 11227-11230. MEDLINE

Wei SJC, Chang RL, Bhachech N, Cui XX, Merkler KA, Wong CQ, Hennig E, Yagi H, Jerina DM and Conney AH. (1993). Cancer Res. 53, 3294-3301. MEDLINE

Wei SJC, Chang RL, Hennig E, Cui XX, Merkler KA, Wong CQ, Yagi H, Jerina DM and Conney AH. (1994). Carcinogenesis 15, 1729-1735. MEDLINE

Yagi T, Mohri-Nakanishi K, Natsuda T, Yamagishi N, Miyakoshi J and Takebe H. (1998). Cancer Lett. 123, 71-76. MEDLINE

Yamagishi N, Miyakoshi J and Takebe H. (1997a). Carcinogenesis 18, 695-700.

Yamagishi N, Miyakoshi J, Yagi T and Takebe H. (1997b). Mutagenesis 12, 191-194.

Yang J-L, Maher VM and McCormick JJ. (1987). Proc. Natl. Acad. Sci. USA 84, 3787-3791. MEDLINE

Yin YX, Tainsky MA, Bischoff FZ, Strong LC and Wahl GM. (1992). Cell 70, 937-948. MEDLINE

Yuan JL, Yeasky TM, Havre PA and Glazer PM. (1995). Carcinogenesis 16, 2295-2300. MEDLINE

Figure 1 p53 is present in (10)3 cells transfected with p53 expression vectors. (a) Western immunoblotting of extracts of pCMV53k-transfected (10)3 cells. Proteins were extracted at indicated times after transfection. The antibody was PAb421. (b) Immunofluorescence detection showing nuclear localization of p53 in (10)3 cells transfected with pECM53k (right) and the absence of p53 in cells transfected with the pECMmutk control vector (left). The cells were fixed 24 h after transfection and immunostained with PAb421. Positive cells represented 8 to 10% of the total cells

Figure 2 The apparent p53-dependent MFR as a function of the white colony frequency. pYZ289 was treated by different doses of four mutagens and cotransfected into (10)3 cells with pECMmutk or pECM53k. The apparent p53-dependent MFR is the white colony frequency for pECMmutk-transfected cells divided by that for pECM53k-transfected cells. The white colony frequency shown is that for pECMmutk-transfected cells

Figure 3 Distribution of mutagen-induced mutations in the supF gene of pYZ289 in the presence or absence of p53. The suppressor tRNA sequence starts at position 99 and ends at position 183. Only substitutions are shown, tandems are underlined, one bp deletions in a tandem are represented by `-' and a dotted underline is used to distinguish adjacent tandems. (a) Spectrum obtained with UV-irradiated pYZ289 giving a white colony frequency of 29.4´10^-3. Other mutations are (above) 164 - 166 GAA AAAT and (below) 166 - 168 ATC AATT; 172 - 176 CCCCC TCCTTT. (b) Spectrum obtained with AFB1-treated pYZ289 giving a white colony frequency of 23.8´10^-3. Other mutations are (above): 102 - 103 GG TGT; 164 delete G; 172 insert A and (below): 102, 103, 104 or 105 delete G; 105 insert GT; 134 - 135 delete TA; 160 insert A. (c) Spectrum obtained with BPDE-treated pYZ289 giving a white colony frequency of 8.4´10^-3. Other mutations are (above): 102, 103, 104 or 105 delete G; 111 - 112 delete GA; 134 insert A; 142 or 143 delete C; 169 insert TTCCCCC; 172, 173, 174, 175 or 176 delete C (five times); 173 - 178 CCCCACCACCAT ACCTAACCA and (below): 99 insert C; 122, 123 or 124 delete C; 142 or 143 delete C (three times); 172, 173, 174, 175 or 176 delete C. (d) Spectrum obtained with BPDE-treated pYZ289 giving a white colony frequency of 71.4´10^-3. Other mutations are (above): 142 or 143 delete C; 172, 173, 174, 175 or 176 delete C (two times) and (below): 111 - 114 TGGG ATTGT; 142 or 143 delete C; 172, 173, 174, 175 or 176 delete G

 Analysis of supF mutants obtained after replication of mutagen-treated pYZ289 in (10)3 cells cotransfected with pECMmutk or pECM53k

 Kinds of supF point mutations obtained after replication of mutagen-treated pYZ289 in (10)3 cells cotransfected with pECMmutk or pECM53k