Q- =') 1991 Oxford University Press
Nucleic Acids Research, Vol. 19, No. 16 4467-4472
Monoclonal antibody-based, selective isolation of DNA fragments containing an alkylated base to be quantified in defined gene sequences Klaus Hochleitner, Jurgen Thomale, Alexander Yu.Nikitin+ and Manfred F.Rajewsky* Institute of Cell Biology (Cancer Research), West German Cancer Center Essen, University of Essen Medical School, Hufelandstrasse 55, D-4300 Essen 1, FRG Received May 14, 1991; Revised and Accepted July 17, 1991
ABSTRACT We have established a sensitive, monoclonal antibody (Mab)-based procedure permitting the selective enrichment of sequences containing the miscoding alkylation product 06-ethylguanine (06-EtGua) from mammalian DNA. H5 rat hepatoma cells were reacted with the N-nitroso carcinogen N-ethyl-N-nitrosourea in vitro, to give overall levels of >25 06-EtGua residues per diploid genome (corresponding to 06-EtGua/guanine molar ratios of 210-8). For analysis, enzymatically restricted DNA from these cells is incubated with an antibody specific for 06-ethyl-2'-deoxyguanosine, the resulting Mab-DNA complexes are separated from (06-EtGua)-free fragments by filtration through a nitrocellulose (NC) membrane, and the DNA is recovered from the filterbound complexes quantitatively. The efficiency of Mab binding to DNA fragments containing 06-EtGua is constant over a range of 06-EtGua/guanine molar ratios between 10-5 and 10-8. (06-EtGua)-containing restriction fragments encompassing known gene sequences (e.g., the immunoglobulin E heavy chain gene of H5 rat hepatoma cells used as a model in this study) are subsequently amplified by PCR and quantified by slot-blot hybridisation. The content and distribution of a specific carcinogen-DNA adduct in defined sequences of genomic DNA can thus be analyzed as well as the kinetics of intragenomic (toposelective) repair of any DNA lesion for which a suitable Mab is available.
INTRODUCTION The integrity of the cellular genome is a prerequisite for the viability and specialized function of a cell. Both prokaryotic and eukaryotic cells have, therefore, developed sophisticated mechanisms for the enzymatic repair of specific DNA lesions which may occur either 'spontaneously' or as a result of exposure to endogenous and exogenous DNA-reactive agents (1, 2). *
Besides the characterization of specific DNA repair proteins and genes much attention is recently being focused on the differential efficiency and intragenomic 'toposelectivity' of DNA repair processes in mammalian cells (for review, see Ref. 3). Thus, pyrimidine dimers formed in cellular DNA upon exposure to ultraviolet light, and some 'bulky' DNA adducts produced by chemical carcinogens, were found to be repaired non-uniformly throughout the genome (4-10). In some cases faster repair was observed in transcriptionally active versus inactive genes (4, 6-9), with a preference for the coding strand (8, 9). Factors contributing to the non-randomness of repair may include the molecular nature of a modified DNA component, its presentation within a particular nucleotide sequence or chromatin configuration, and the characteristics of the repair-protein(s)
involved. In the case of the well characterised alkylation products formed in cellular DNA after exposure to N-nitroso carcinogens (11, 12) or to therapeutic agents with similar DNA reactivity (13), several studies have indicated a non-random distribution of some of these lesions in DNA as a function of nucleotide sequence and chromatin structure (14-18). However, only limited information is available with respect to intragenomic variations in the rates of enzymatic repair of small alkyl adducts (methyl or ethyl groups; 19-22). With the aid of DNA footprinting analyses, preferential repair of heat-labile DNA methylation products (N7-methylguanine, N3-methyladenine) was observed in an actively transcribed rat insulin gene as compared to its nontranscribed counterpart (22). However, due to the lack of sufficiently sensitive methodology the possible toposelective formation and repair of 0-alkylation products in defined DNA sequences still need to be analyzed. These DNA alkylation products include the miscoding lesions 06-alkylguanine and 04-alkylthymine which are of particular relevance to mutagenesis and carcinogenesis by N-nitroso compounds (11, 12, 17, 18, 23-27). The present report describes a monoclonal antibody (Mab)based immunoaffinity procedure, combined with the polymerase chain reaction (PCR), for the selective enrichment of DNA
To whom correspondence should be addressed
+ On leave of absence from the N.N.Petrov Research Institute of Oncology of the USSR
Ministry of Health, Leningrad, USSR
4468 Nucleic Acids Research, Vol. 19, No. 16 fragments carrying the alkylation product 06-ethylguanine (06-EtGua) from a large excess of DNA devoid of this lesion.
MATERIALS AND METHODS Cell culture and alkylation of cellular DNA in vivo H5 rat hepatoma cells (28) were grown as monolayers at 37°C in Dulbecco's modified Eagle's Medium (DMEM; Flow Laboratories) supplemented with 10% fetal bovine serum (Sebak), penicillin (100 U/ml), and streptomycin (50 Mg/ml), under a humidified atmosphere containing 10% CO2. Log-phase H5 cells were reacted with N-ethyl-N-nitrosourea (EtNU, Roth) as described (29). Briefly, EtNU (recrystallized twice from methanol) was pre-dissolved in H20-free dimethyl sulfoxide (100 mg/ml; stock solution) and stored at -20°C in the dark. Immediately prior to use, the stock solution was diluted 1:50 in serum-free DMEM adjusted with CO2 to a pH value 85 % of all plasmid molecules carrying at least one 06-EtGua (Figure 1). Binding of non-ethylated plasmid SV2gpt DNA was hardly measurable (detection limit: 0.1 %) at Mab concentrations 2 300 jig/ml, but reached measurable values at very high Mab concentrations (Figure 1). Additional experiments were performed in order to optimize the ratio of specific/non-specific Mab binding. Samples of pSV2gpt DNA (20 ,ug) containing a constant amount of 35Slabeled and different amounts of non-labeled ethylated plasmid molecules, were reacted with Mab ER-6 and passed through NC membranes. As evidenced by the fractions of Mab-DNA complexes retained on the membranes as a function of the 06-EtGua/Gua molar ratios in the plasmid DNA (Table 1), the retention of ethylated DNA is constant for 06-EtGua/Gua molar ratios between 10-5 and 10-8.
40-
~0 z 0
20-
0
100
200
300
400
Concentration of Mab ER-6
500
(pg/ml)
06-EtGua/Gua molar ratio in DNA l0o10-6
1o-7 Figure 1. Retention of (06-EtGua)-containing pSV2gpt DNA on NC membranes as a function of antibody concentration. Constant amounts (8 ng) of ethylated plasmid DNA (containing, on average, 1 06-EtGua residue/molecule, A) or nonethylated plasmid DNA (-) were incubated with different concentrations of Mab ER-6 (see Materials and Methods). Mab-DNA complexes retained on the membranes were quantified by liquid scintillation spectrometry. The dotted upper curve was computed from the experimental data for the ethylated plasmid DNA (lower curve), assuming a Poisson distribution of 06-EtGua residues among the plasmid molecules. Vertical bars: standard deviations (>23 independent experiments).
10-8
DNA complexed by Mab ER-6 percent of (06-EtGua)percent of 'input' containing DNA radioactivity fragments 54.2 ± 1.7 54.6 1.4 53.3 2.6 54.5 4.9
84.7 4 2.6 85.3±2.2 83.3 4.0 85.2 7.6
Samples of pSV2gpt DNA with varying 06-EtGua/Gua molar ratios were spiked with constant amounts of 35S-labeled, ethylated pSV2gpt DNA. Twenty itg each of these samples were incubated with 30 i4g of Mab ER-6 and filtrated through NC membranes. The amounts of (06-EtGua)-containing DNA molecules retained on the NC membranes in the form of Mab-DNA complexes were determined by liquid scintillation spectrometry and expressed as% of the 'input' radioactivity, assuming 06-EtGua residues to be distributed according to Poisson. The values are means of 3 independent experiments (± S.D.).
4470 Nucleic Acids Research, Vol. 19, No. 16 was incubated with Mab ER-6, passed through NC membranes, and recovered as described. The eluted plasmid DNA was then lyophilized and re-dissolved in a final volume of 100 iLl of 0.1 M NaCl containing 30 ,g of Mab ER-6. The recovery of ethylated plasmid DNA was determined after each step during three subsequent adsorption/elution cycles (Table 2). While near complete recovery was obtained in all three elution steps, the efficiency of Mab binding to (06-EtGua)-containing molecules in the second and third adsorption step was reduced by a constant factor in each step (Table 2). This was found to be due to a substance co-eluting from the BA85 NC membrane during the first elution step and interfering with the binding of Mab ER-6 to 06-EtGua in the subsequent cycles (data not shown).
encompassing the immunoglobulin E (IgE) heavy chain gene was used as a model DNA sequence. Different amounts (1-200 ng) of EcoRI digested genomic DNA were used as a template for amplification. To each reaction sample 20 fg of linearized pSV2gpt DNA were added as an internal standard. A segment of the plasmid DNA was co-amplified with a segment of the rat IgE gene sequence. DNA samples were not incubated with Mab ER-6 prior to amplification. The PCR products, as determined on agarose gels, were 175 bp (gpt) and 150 bp (IgE), respectively, in length (data not shown), in precise agreement with the values predicted from published sequence data (37, 38). PCR products were blotted onto nylon membranes and hybridized with specific oligonucleotide probes for both sequences (see Materials and Methods). Quantification by slot blot hybridization and autoradiography showed a poor correlation between the amount of template DNA and the amount of PCR product of the IgE gene (Figure 2a). The amount of PCR product from the gpt sequence exhibited considerable variations between samples containing identical amounts of template DNA (Figure 2b). However, after normalization of the data obtained for the IgE sequence to the respective values for the co-amplified gpt sequence, linearity (R=0.99) was found for the whole range of template DNA concentrations examined (Figure 2c). The specific enrichment from H5 cell DNA of (06-EtGua)containing EcoRI restriction fragments encompassing the rat IgE gene is shown in Figure 3. After preselection for (06-EtGua) containing fragments and normalization of the data to the 'internal standard' (ethylated pSV2gpt DNA), a linear relationship was observed between the molar 06-EtGua content of the genomic DNA and the amount of amplified product. The unspecific background from non-ethylated DNA was negligible.
Enrichment and amplification by PCR of a specific sequence of genomic DNA A prerequisite for quantitative DNA analysis involving amplification by PCR is a linear relationship between the amount of genomic template DNA in the PCR reaction and the amount of the resulting PCR product. In order to establish controlled assay conditions, a 15 kb EcoRI fragment of DNA from H5 cells Table 2. Recovery of pSV2gpt DNA fragments containing .1 06-EtGua residue, and of non-ethylated DNA fragments, during three repeated cycles of antibody binding and elution. Fraction (%) of DNA molecules recovered from
after lI" filtration after 1st elution after 2nd filtration after 2nd elution after 3rd filtration after 3rd elution
ethylated pSV2gpt DNA
non-ethylated pSV2gpt DNA
88 85 49 47 27 26
0.5 0.5 n.d. n.d. n.d. n.d.
Determination of the 06-EtGua/Gua molar ratio in a specific DNA restriction fragment isolated from cultured H5 rat hepatoma cells at different times after exposure to EtNU The initial frequency and kinetics of elimination of 06-EtGua from DNA were determined in the EcoRI DNA restriction fragment encompassing the IgE heavy chain gene of H5 rat hepatoma cells. DNA was isolated at different times after a 20 min EtNU-exposure of H5 cells in culture. Samples containing C
Eight ng of ethylated or non_ethylated 355-labeled pSV2gpt DNA were incubated with 30 itg of Mab ER-6. After filtration, DNA molecules complexed by the antibody and retained on the NC membranes were eluted, dried and re-dissolved (see Materials and Methods). This procedure was carried out three times and the recovery was determined after each step. The values are means from > 3 independent experiments. n.d.: Not detectable (detection limit: 0.1%).
b
a
10 1
0
c
Ifi
C
coio
U)
S
en
-6 0
:D
10CL
*I
c-0'
I I
0
t
a
T
c
0 0
10
cn
4_
o
.0_
10'
1'
C
0.
0. -o
I
.0 I
-0 1.0
1-
-
0n
10
10
100 Amount of DNA in PCR Sample
(ng)
3-
1
..
i r
ri
10
100 Amount of DNA in PCR Sample (ng)
Amount of DNA in PCR Sample (ng)
Figure 2. Correlation between the amount of template DNA and the resulting PCR product. Different amounts of genomic DNA restricted with EcoRI were used for amplification of a rat IgE heavy chain gene sequence. As an internal control, each sample contained 20 fg of linearized pSV2gpt DNA, part of which amplified with the genomic DNA. After blotting, oligonucleotide hybridization, and autoradiography, results obtained for both target sequences were plotted separately against the amounts of genomic DNA in the respective samples (a: IgE; b: pSV2gpt). c: Linear correlation between the amount of genomic template DNA and the respective PCR product after normalizing the data for the IgE sequence to those obtained for the pSV2gpt DNA. was co-
Nucleic Acids Research, Vol. 19, No. 16 4471 5 jig of EcoRI restricted DNA and 20 fg of ethylated pSV2gpt DNA were incubated with Mab ER-6 and subsequently analysed as described in Materials and Methods. The initial frequency (after 20 min exposure to EtNU) and the kinetics of elimination of 06-EtGua within 24 hrs in the EcoRI fragment were found to be identical to the corresponding values obtained by ISB for the total genomic DNA (Figure 4).
DISCUSSION A pivotal step common to recent experimental approaches to the detection and quantification of (carcinogen-induced) structural _
' 100
N,
U'
0.
T
.Www
LU 10~ ~ ~ ~ -
o6_tu/u
or Ratio in DNA
Figure 3. Enrichment of an (06-EtGua)-containing DNA restriction fragment encompassing the rat IgE heavy chain gene. Genonmic DNA (10 ytg) from EtNUtreated H5 rat hepatoma cells containing 06-EtGua (06-EtGua/Gua molar ratio, l0-5) and 20 fg of ethylated pSV2gpt DNA (internal standard) were incubated with Mab ER6 (30 ytg), and (06-EtGua)-free DNA fragments were removed from Mab/DNA complexes by filtration through NC membranes. This procedure was repeated once prior to co-amplification by PCR of part of the cellular IgE gene sequence and the plasmid derived E.coli gpt gene. In addition, genomic DNA from EtNU-treated H5 cells was diluted with non-ethylated H5 DNA to give overall 06-EtGua/Gua molar ratios of 106, i0-7, and 10-8, respectively. The resulting PCR products were blotted onto nylon membranes and hybridized with sequence-specific 32P-labeled oligonucleotides. The signals obtained for the IgE gene sequence were normalized to the signals from the co-amplified gpt sequence and plotted against the 06-EtGua/Gua molar ratios of H5 cell DNA in the respective samples. 0
10 C,
a
c) o S
L- 1 0 20 30 10 Time after EtNU-Pulse (hrs)
Figure 4. Initial frequency and kinetics of elimination of 06-EtGua from a specific EcoRI DNA restriction fragment encompassing the IgE heavy chain gene and from total genomic DNA isolated from H5 rat hepatoma cells at different times after a 20 min-exposure to EtNU in culture. The amount of 06-EtGua in total genomic DNA (0) was determined by ISB. The immunoaffinity-PCR technique (see Figs.2 and 3 and Materials and Methods) was used to quantify 06-EtGua in the restriction fragment encompassing the IgE heavy chain gene (A; 5 ztg of Eco RI restricted genomic DNA).
modifications of DNA in defined gene sequences, has been the specific cleavage of DNA (or its inhibition) at the site of the respective lesion (4, 19). A restriction fragment length polymorphism is thus created which permits the analysis of the distribution and toposelective repair of DNA lesions. The applicability of this approach is thus restricted to DNA lesions which occur at rather high frequencies and are associable with site-specific DNA cleavage. In contrast, the present approach is based on the combined application of immunoaffinity and PCR methodology. This procedure was developed to detect low levels of specific lesions in defined gene sequences and, in particular, to quantify specific DNA alkylation products such as the premutational 06-alkylguanine (11, 12, 17, 18, 23-27). The procedure allows the selective isolation of a small fraction of DNA molecules carrying a specific alkylation product (06-EtGua) from a very large excess of unmodified DNA, and to amplify the lesion-containing sequences by PCR for quantitative analysis. The specificity of the method for a given DNA adduct results from the use of a high-affinity Mab exhibiting a very low degree of cross-reactivity with other DNA constituents. As shown by calibration experiments with ethylated SV2gpt plasmid DNA, and assuming a Poisson distribution of 06-EtGua residues among these model DNA molecules, >85% of all DNA fragments containing at least 1 06-EtGua residue are retained on the NC membranes as Mab-DNA complexes, the background retention of non-ethylated DNA is negligible. This result could not necessarily be expected since the affinity of Mab ER-6 towards O6-EtGua contained in ds DNA is significantly reduced in comparison to 06-ethyl-2'-deoxyguanosine as a monodeoxynucleoside (data not shown). Moreover the data seem to rule out pronounced effects of neighbouring bases and/or DNA curvature leading to positional differences in the recognition of 06-EtGua by Mab ER-6, since no differences in relative DNA binding efficiency were found between the second and third adsorption/elution cycle (see Table 2). A 20 min-exposure of cultured H5 rat hepatoma cells to 100 ttg of EtNU/ml is the limiting condition tolerated by these cells without severe cytotoxic effects (data not shown). This exposure level results in an 06-EtGua/Gua molar ratio in DNA of about 10-5. From this value it can be calculated that only a few percent of restriction fragments encompassing a particular gene (e.g., the rat IgE heavy chain gene used as a model in this study) will contain at least 1 06-EtGua residue. The resulting small number of gene copies isolated by the immunoaffinity procedure thus precludes their quantification by conventional blotting and hybridization techniques. This is especially true for alkylation frequencies corresponding to 06-EtGua/Gua molar ratios 2 orders of magnitude was found between the sample concentration of genomic template DNA and the amount of specific PCR product. The quantitative analyses subsequent to the PCR reaction are achieved by slot blotting and hybridization with sequence-specific oligonucleotides. Thus it is not necessary to amplify complete restriction fragments but rather only small segments of 100-200 bp, permitting quantification of DNA fragments with different lengths in the same sample. It should also be emphasized that although each DNA fragment
4472 Nucleic Acids Research, Vol. 19, No. 16 isolated via Mab-DNA complexing contains at least 1 06-EtGua residue, this residue is-for statistical reasons-in most fragments not located within the sequence amplified by PCR. In conclusion, we have described an immunoaffinity-PCR approach to the quantification of a defined DNA lesion in DNA restriction fragments encompassing known gene sequences. This procedure is not limited to the isolation of DNA molecules containing 06-EtGua, but may be applied to any stable DNA lesion for which a suitable Mab is available. The sensitivity of the method in terms of the molar content of lesions detectable in genomic DNA is comparable to that of competitive RIA (35), immuno-slot-blot (30), or 32P-postlabeling procedures (39). However, considering the small amount of genomic DNA required, together with the fact that the DNA lesion is not quantified in bulk DNA but rather in a specific gene sequence, the sensitivity of the method by far exceeds that of the latter assays. It has thus become possible, for the first time, to analyze the formation and repair of 06-alkylguanine in specific gene sequences in vivo as demonstrated here for an EcoRI DNA restriction fragment encompassing the rat IgE heavy chain gene. This method can equally be applied to DNA containing other DNA adducts for which suitable Mabs are available, including 0-alkylation products other than 06-EtGua (27, 30, 40).
ACKNOWLEDGEMENTS We thank K.Sendowski for oligonucleotide syntheses and A.Braun-Saure, B.Smialek and I.Spratte for skilful technical assistance. This work was supported by the Deutsche Forschungsgemeinschaft (SFB 102/A70) and by the Commission of the European Communities (ENV-544-D [B]).
REFERENCES 1. Friedberg, E.C. (1985) DNA Repair, W.H. Freeman, New York. 2. Setlow, R.B. (1989) In Castellani, A. (ed.), DNA Damage and Repair, Plenum press, New York, pp. 1-9. 3. Smith, C.A., Mellon, I. (1990) In Obe, G. (ed.), Advances in Mutagenesis Research. Springer-Verlag, Berlin, Vol. 1, pp. 153-194. 4. Bohr, V.A., Smith, C.A., Okumoto, D.S. and Hanawalt, P.C. (1985) Cell, 40, 359-369. 5. Leadon, S.A. (1986) Nucleic Acids Res., 14, 8979-8995. 6. Madhani, H.D., Bohr, V.A. and Hanawalt, P.C. (1986) Cell, 45, 417-423. 7. Mellon, I., Bohr, V.A., Smith, C.A. and Hanawalt, P.C. (1986) Proc. Natl. Acad. Sci. USA, 83, 8878-8882. 8. Mellon, I., Spivak, G. and Hanawalt, P.C. (1987) Cell, 51, 241-249. 9. Mellon, I. and Hanawalt, P.C. (1989) Nature, 342, 95-98. 10. Poirier, M.C., Fullerton, N.F., Patterson, E.D. and Beland, F.A. (1990) Carcinogenesis, 11, 1343-1347. 11. Preussmann, R. and Stewart, B.W. (1984) N-Nitroso Carcinogens ACS Monograph No. 182. American Chemical Society, Washington, D.C., pp. 643-828. 12. Rajewsky, M.F. (1980) In Montesano, R., Bartsch, H. and Tomatis, L. (eds.), Molecular and Cellular Aspects of Carcinogen Screening Tests. IARC Scientific Publ. No. 27. International Agency for Research on Cancer, Lyon, pp. 41-54. 13. Bohr, V.A. (1990) J. Cancer Res. Clin. Oncol., 116, 384-391. 14. Nehls, P., Rajewsky, M.F., Spiess, E., Werner, D. (1984) EMBO J., 3, 327-332. 15. Ryan, A.J., Billett, M.A. and O'Connor, P.J. (1986) Carcinogenesis, 7, 1497-1503. 16. Topal, M.D., Eadie, J.S. and Conrad, M. (1986) J. Biol. Chem., 261, 9879-9885. 17. Glickman, B.W., Horsfall, M.J., Gordon, A.J.E. and Burns, P.A. (1987) Env. Health Perspect., 56, 29-32. 18. Horsfall, M.J., Gordon, A.J.E., Burns, P.A., Zielenska, M., van der Vliet, G.M.E. and Glicknan, B.W. (1990) Environ. Mol. Mutagen., 15, 107-122. 19. Nose, K. and Nikaido, 0. (1984) Biochim. Biophys. Acta, 781, 273 -278.
20. Milligan, J.R. and Archer, M.C. (1988) Biochem. Biophys. Res. Comm., 155, 14-17. 21. Scicchitano, D.A. and Hanawalt, P.C. (1989) Proc. Natl. Acad. Sci. USA, 86, 3050-3054. 22. LeDoux, S.P., Patton, N.J., Nelson, J.W., Bohr, V.A., Wilson, G.L. (1990) J. Biol. Chem., 265, 14875-14880. 23. Rajewsky, M.F., Augenlicht, L.H., Biessmann, H., Goth, R., Huilser, D.F., Laerum, O.D. and Lomakina, L.Y. (1977) In Hiatt, H.H., Watson, J.D. and Winston, J.J. (eds.), Origins of Human Cancer, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y., pp. 709-726. 24. Pegg, A.E. (1984) Cancer Invest., 2, 223-231. 25. Doniger, J., Day, R.S. and DiPaolo, J.A. (1985) Proc. Natl. Acad. Sci. USA, 82, 421-425. 26. Zarbl, H., Sukumar, S., Arthur, A.V., Martin-Zanca, D. and Barbacid, M. (1985) Nature, 315, 382-385. 27. Thomale, J., Huh, N., Nehls, P., Eberle, G. and Rajewsky, M.F. (1990) Proc. Natl. Acad. Sci. USA, 87. 9883-9887. 28. Deschatrette, J., Moore, E.E., Dubois, M., Cassio, D. and Weiss, M.C. (1979) Somat. Cell Genet., 5, 697-718. 29. Huh, N. and Rajewsky, M.F. (1986) Carcinogenesis, 7, 435-439. 30. Nehls, P., Adamkiewicz, J. and Rajewsky, M.F. (1984) J. Cancer Res. Clin. Oncol., 108, 23-29. 31. Rajewsky, M.F., Muller, R., Adamkiewicz, J. and Drosdziok, W. (1980) In Pullman, B., Ts'o, P.O.P. and Gelboin, H. (eds.), Carcinogenesis: Fundamental Mechanisms and Environmental Effects. Reidel, Dordrecht, pp. 207-218. 32. Samson, L., Thomale, J., Rajewsky, M.F. (1988) EMBO J., 7, 2261-2267. 33. Mulligan, P.C. and Berg, P. (1980) Science, 209, 1422-1427. 34. Nehls, P. and Rajewsky, M.F. (1990) Carcinogenesis, 11, 81-87. 35. Muller, R. and Rajewsky, M.F. (1980) Cancer Res., 40, 887-896. 36. Saiki, R.K., Scharf, S., Faloona, F., Muilis, K.B., Horn, G.T., Erlich, H.A. and Arnheim, N. (1985) Science, 230, 1350-1354. 37. Steen, M.L., Hellman, L. and Petersson, U. (1984) J. Mol. Biol., 177, 19-32. 38. Richardson, K.K., Fostel, J. and Skopek, T.R. (1983) Nucleic Acids Res., 11, 8809-8816. 39. Wilson, V.L., Basu, A.K., Essigmann, J.M., Smith, R.A. and Harris, C.C. (1988) Cancer Res., 48, 2156-2161. 40. Adamkiewicz, J., Eberle, G., Huh, N., Nehls, P. and Rajewsky, M.F. (1985) Environm. Health Perspect., 62, 49-55.
Nucleic Acids Research, Vol. 19, No. 16 4467-4472
Monoclonal antibody-based, selective isolation of DNA fragments containing an alkylated base to be quantified in defined gene sequences Klaus Hochleitner, Jurgen Thomale, Alexander Yu.Nikitin+ and Manfred F.Rajewsky* Institute of Cell Biology (Cancer Research), West German Cancer Center Essen, University of Essen Medical School, Hufelandstrasse 55, D-4300 Essen 1, FRG Received May 14, 1991; Revised and Accepted July 17, 1991
ABSTRACT We have established a sensitive, monoclonal antibody (Mab)-based procedure permitting the selective enrichment of sequences containing the miscoding alkylation product 06-ethylguanine (06-EtGua) from mammalian DNA. H5 rat hepatoma cells were reacted with the N-nitroso carcinogen N-ethyl-N-nitrosourea in vitro, to give overall levels of >25 06-EtGua residues per diploid genome (corresponding to 06-EtGua/guanine molar ratios of 210-8). For analysis, enzymatically restricted DNA from these cells is incubated with an antibody specific for 06-ethyl-2'-deoxyguanosine, the resulting Mab-DNA complexes are separated from (06-EtGua)-free fragments by filtration through a nitrocellulose (NC) membrane, and the DNA is recovered from the filterbound complexes quantitatively. The efficiency of Mab binding to DNA fragments containing 06-EtGua is constant over a range of 06-EtGua/guanine molar ratios between 10-5 and 10-8. (06-EtGua)-containing restriction fragments encompassing known gene sequences (e.g., the immunoglobulin E heavy chain gene of H5 rat hepatoma cells used as a model in this study) are subsequently amplified by PCR and quantified by slot-blot hybridisation. The content and distribution of a specific carcinogen-DNA adduct in defined sequences of genomic DNA can thus be analyzed as well as the kinetics of intragenomic (toposelective) repair of any DNA lesion for which a suitable Mab is available.
INTRODUCTION The integrity of the cellular genome is a prerequisite for the viability and specialized function of a cell. Both prokaryotic and eukaryotic cells have, therefore, developed sophisticated mechanisms for the enzymatic repair of specific DNA lesions which may occur either 'spontaneously' or as a result of exposure to endogenous and exogenous DNA-reactive agents (1, 2). *
Besides the characterization of specific DNA repair proteins and genes much attention is recently being focused on the differential efficiency and intragenomic 'toposelectivity' of DNA repair processes in mammalian cells (for review, see Ref. 3). Thus, pyrimidine dimers formed in cellular DNA upon exposure to ultraviolet light, and some 'bulky' DNA adducts produced by chemical carcinogens, were found to be repaired non-uniformly throughout the genome (4-10). In some cases faster repair was observed in transcriptionally active versus inactive genes (4, 6-9), with a preference for the coding strand (8, 9). Factors contributing to the non-randomness of repair may include the molecular nature of a modified DNA component, its presentation within a particular nucleotide sequence or chromatin configuration, and the characteristics of the repair-protein(s)
involved. In the case of the well characterised alkylation products formed in cellular DNA after exposure to N-nitroso carcinogens (11, 12) or to therapeutic agents with similar DNA reactivity (13), several studies have indicated a non-random distribution of some of these lesions in DNA as a function of nucleotide sequence and chromatin structure (14-18). However, only limited information is available with respect to intragenomic variations in the rates of enzymatic repair of small alkyl adducts (methyl or ethyl groups; 19-22). With the aid of DNA footprinting analyses, preferential repair of heat-labile DNA methylation products (N7-methylguanine, N3-methyladenine) was observed in an actively transcribed rat insulin gene as compared to its nontranscribed counterpart (22). However, due to the lack of sufficiently sensitive methodology the possible toposelective formation and repair of 0-alkylation products in defined DNA sequences still need to be analyzed. These DNA alkylation products include the miscoding lesions 06-alkylguanine and 04-alkylthymine which are of particular relevance to mutagenesis and carcinogenesis by N-nitroso compounds (11, 12, 17, 18, 23-27). The present report describes a monoclonal antibody (Mab)based immunoaffinity procedure, combined with the polymerase chain reaction (PCR), for the selective enrichment of DNA
To whom correspondence should be addressed
+ On leave of absence from the N.N.Petrov Research Institute of Oncology of the USSR
Ministry of Health, Leningrad, USSR
4468 Nucleic Acids Research, Vol. 19, No. 16 fragments carrying the alkylation product 06-ethylguanine (06-EtGua) from a large excess of DNA devoid of this lesion.
MATERIALS AND METHODS Cell culture and alkylation of cellular DNA in vivo H5 rat hepatoma cells (28) were grown as monolayers at 37°C in Dulbecco's modified Eagle's Medium (DMEM; Flow Laboratories) supplemented with 10% fetal bovine serum (Sebak), penicillin (100 U/ml), and streptomycin (50 Mg/ml), under a humidified atmosphere containing 10% CO2. Log-phase H5 cells were reacted with N-ethyl-N-nitrosourea (EtNU, Roth) as described (29). Briefly, EtNU (recrystallized twice from methanol) was pre-dissolved in H20-free dimethyl sulfoxide (100 mg/ml; stock solution) and stored at -20°C in the dark. Immediately prior to use, the stock solution was diluted 1:50 in serum-free DMEM adjusted with CO2 to a pH value 85 % of all plasmid molecules carrying at least one 06-EtGua (Figure 1). Binding of non-ethylated plasmid SV2gpt DNA was hardly measurable (detection limit: 0.1 %) at Mab concentrations 2 300 jig/ml, but reached measurable values at very high Mab concentrations (Figure 1). Additional experiments were performed in order to optimize the ratio of specific/non-specific Mab binding. Samples of pSV2gpt DNA (20 ,ug) containing a constant amount of 35Slabeled and different amounts of non-labeled ethylated plasmid molecules, were reacted with Mab ER-6 and passed through NC membranes. As evidenced by the fractions of Mab-DNA complexes retained on the membranes as a function of the 06-EtGua/Gua molar ratios in the plasmid DNA (Table 1), the retention of ethylated DNA is constant for 06-EtGua/Gua molar ratios between 10-5 and 10-8.
40-
~0 z 0
20-
0
100
200
300
400
Concentration of Mab ER-6
500
(pg/ml)
06-EtGua/Gua molar ratio in DNA l0o10-6
1o-7 Figure 1. Retention of (06-EtGua)-containing pSV2gpt DNA on NC membranes as a function of antibody concentration. Constant amounts (8 ng) of ethylated plasmid DNA (containing, on average, 1 06-EtGua residue/molecule, A) or nonethylated plasmid DNA (-) were incubated with different concentrations of Mab ER-6 (see Materials and Methods). Mab-DNA complexes retained on the membranes were quantified by liquid scintillation spectrometry. The dotted upper curve was computed from the experimental data for the ethylated plasmid DNA (lower curve), assuming a Poisson distribution of 06-EtGua residues among the plasmid molecules. Vertical bars: standard deviations (>23 independent experiments).
10-8
DNA complexed by Mab ER-6 percent of (06-EtGua)percent of 'input' containing DNA radioactivity fragments 54.2 ± 1.7 54.6 1.4 53.3 2.6 54.5 4.9
84.7 4 2.6 85.3±2.2 83.3 4.0 85.2 7.6
Samples of pSV2gpt DNA with varying 06-EtGua/Gua molar ratios were spiked with constant amounts of 35S-labeled, ethylated pSV2gpt DNA. Twenty itg each of these samples were incubated with 30 i4g of Mab ER-6 and filtrated through NC membranes. The amounts of (06-EtGua)-containing DNA molecules retained on the NC membranes in the form of Mab-DNA complexes were determined by liquid scintillation spectrometry and expressed as% of the 'input' radioactivity, assuming 06-EtGua residues to be distributed according to Poisson. The values are means of 3 independent experiments (± S.D.).
4470 Nucleic Acids Research, Vol. 19, No. 16 was incubated with Mab ER-6, passed through NC membranes, and recovered as described. The eluted plasmid DNA was then lyophilized and re-dissolved in a final volume of 100 iLl of 0.1 M NaCl containing 30 ,g of Mab ER-6. The recovery of ethylated plasmid DNA was determined after each step during three subsequent adsorption/elution cycles (Table 2). While near complete recovery was obtained in all three elution steps, the efficiency of Mab binding to (06-EtGua)-containing molecules in the second and third adsorption step was reduced by a constant factor in each step (Table 2). This was found to be due to a substance co-eluting from the BA85 NC membrane during the first elution step and interfering with the binding of Mab ER-6 to 06-EtGua in the subsequent cycles (data not shown).
encompassing the immunoglobulin E (IgE) heavy chain gene was used as a model DNA sequence. Different amounts (1-200 ng) of EcoRI digested genomic DNA were used as a template for amplification. To each reaction sample 20 fg of linearized pSV2gpt DNA were added as an internal standard. A segment of the plasmid DNA was co-amplified with a segment of the rat IgE gene sequence. DNA samples were not incubated with Mab ER-6 prior to amplification. The PCR products, as determined on agarose gels, were 175 bp (gpt) and 150 bp (IgE), respectively, in length (data not shown), in precise agreement with the values predicted from published sequence data (37, 38). PCR products were blotted onto nylon membranes and hybridized with specific oligonucleotide probes for both sequences (see Materials and Methods). Quantification by slot blot hybridization and autoradiography showed a poor correlation between the amount of template DNA and the amount of PCR product of the IgE gene (Figure 2a). The amount of PCR product from the gpt sequence exhibited considerable variations between samples containing identical amounts of template DNA (Figure 2b). However, after normalization of the data obtained for the IgE sequence to the respective values for the co-amplified gpt sequence, linearity (R=0.99) was found for the whole range of template DNA concentrations examined (Figure 2c). The specific enrichment from H5 cell DNA of (06-EtGua)containing EcoRI restriction fragments encompassing the rat IgE gene is shown in Figure 3. After preselection for (06-EtGua) containing fragments and normalization of the data to the 'internal standard' (ethylated pSV2gpt DNA), a linear relationship was observed between the molar 06-EtGua content of the genomic DNA and the amount of amplified product. The unspecific background from non-ethylated DNA was negligible.
Enrichment and amplification by PCR of a specific sequence of genomic DNA A prerequisite for quantitative DNA analysis involving amplification by PCR is a linear relationship between the amount of genomic template DNA in the PCR reaction and the amount of the resulting PCR product. In order to establish controlled assay conditions, a 15 kb EcoRI fragment of DNA from H5 cells Table 2. Recovery of pSV2gpt DNA fragments containing .1 06-EtGua residue, and of non-ethylated DNA fragments, during three repeated cycles of antibody binding and elution. Fraction (%) of DNA molecules recovered from
after lI" filtration after 1st elution after 2nd filtration after 2nd elution after 3rd filtration after 3rd elution
ethylated pSV2gpt DNA
non-ethylated pSV2gpt DNA
88 85 49 47 27 26
0.5 0.5 n.d. n.d. n.d. n.d.
Determination of the 06-EtGua/Gua molar ratio in a specific DNA restriction fragment isolated from cultured H5 rat hepatoma cells at different times after exposure to EtNU The initial frequency and kinetics of elimination of 06-EtGua from DNA were determined in the EcoRI DNA restriction fragment encompassing the IgE heavy chain gene of H5 rat hepatoma cells. DNA was isolated at different times after a 20 min EtNU-exposure of H5 cells in culture. Samples containing C
Eight ng of ethylated or non_ethylated 355-labeled pSV2gpt DNA were incubated with 30 itg of Mab ER-6. After filtration, DNA molecules complexed by the antibody and retained on the NC membranes were eluted, dried and re-dissolved (see Materials and Methods). This procedure was carried out three times and the recovery was determined after each step. The values are means from > 3 independent experiments. n.d.: Not detectable (detection limit: 0.1%).
b
a
10 1
0
c
Ifi
C
coio
U)
S
en
-6 0
:D
10CL
*I
c-0'
I I
0
t
a
T
c
0 0
10
cn
4_
o
.0_
10'
1'
C
0.
0. -o
I
.0 I
-0 1.0
1-
-
0n
10
10
100 Amount of DNA in PCR Sample
(ng)
3-
1
..
i r
ri
10
100 Amount of DNA in PCR Sample (ng)
Amount of DNA in PCR Sample (ng)
Figure 2. Correlation between the amount of template DNA and the resulting PCR product. Different amounts of genomic DNA restricted with EcoRI were used for amplification of a rat IgE heavy chain gene sequence. As an internal control, each sample contained 20 fg of linearized pSV2gpt DNA, part of which amplified with the genomic DNA. After blotting, oligonucleotide hybridization, and autoradiography, results obtained for both target sequences were plotted separately against the amounts of genomic DNA in the respective samples (a: IgE; b: pSV2gpt). c: Linear correlation between the amount of genomic template DNA and the respective PCR product after normalizing the data for the IgE sequence to those obtained for the pSV2gpt DNA. was co-
Nucleic Acids Research, Vol. 19, No. 16 4471 5 jig of EcoRI restricted DNA and 20 fg of ethylated pSV2gpt DNA were incubated with Mab ER-6 and subsequently analysed as described in Materials and Methods. The initial frequency (after 20 min exposure to EtNU) and the kinetics of elimination of 06-EtGua within 24 hrs in the EcoRI fragment were found to be identical to the corresponding values obtained by ISB for the total genomic DNA (Figure 4).
DISCUSSION A pivotal step common to recent experimental approaches to the detection and quantification of (carcinogen-induced) structural _
' 100
N,
U'
0.
T
.Www
LU 10~ ~ ~ ~ -
o6_tu/u
or Ratio in DNA
Figure 3. Enrichment of an (06-EtGua)-containing DNA restriction fragment encompassing the rat IgE heavy chain gene. Genonmic DNA (10 ytg) from EtNUtreated H5 rat hepatoma cells containing 06-EtGua (06-EtGua/Gua molar ratio, l0-5) and 20 fg of ethylated pSV2gpt DNA (internal standard) were incubated with Mab ER6 (30 ytg), and (06-EtGua)-free DNA fragments were removed from Mab/DNA complexes by filtration through NC membranes. This procedure was repeated once prior to co-amplification by PCR of part of the cellular IgE gene sequence and the plasmid derived E.coli gpt gene. In addition, genomic DNA from EtNU-treated H5 cells was diluted with non-ethylated H5 DNA to give overall 06-EtGua/Gua molar ratios of 106, i0-7, and 10-8, respectively. The resulting PCR products were blotted onto nylon membranes and hybridized with sequence-specific 32P-labeled oligonucleotides. The signals obtained for the IgE gene sequence were normalized to the signals from the co-amplified gpt sequence and plotted against the 06-EtGua/Gua molar ratios of H5 cell DNA in the respective samples. 0
10 C,
a
c) o S
L- 1 0 20 30 10 Time after EtNU-Pulse (hrs)
Figure 4. Initial frequency and kinetics of elimination of 06-EtGua from a specific EcoRI DNA restriction fragment encompassing the IgE heavy chain gene and from total genomic DNA isolated from H5 rat hepatoma cells at different times after a 20 min-exposure to EtNU in culture. The amount of 06-EtGua in total genomic DNA (0) was determined by ISB. The immunoaffinity-PCR technique (see Figs.2 and 3 and Materials and Methods) was used to quantify 06-EtGua in the restriction fragment encompassing the IgE heavy chain gene (A; 5 ztg of Eco RI restricted genomic DNA).
modifications of DNA in defined gene sequences, has been the specific cleavage of DNA (or its inhibition) at the site of the respective lesion (4, 19). A restriction fragment length polymorphism is thus created which permits the analysis of the distribution and toposelective repair of DNA lesions. The applicability of this approach is thus restricted to DNA lesions which occur at rather high frequencies and are associable with site-specific DNA cleavage. In contrast, the present approach is based on the combined application of immunoaffinity and PCR methodology. This procedure was developed to detect low levels of specific lesions in defined gene sequences and, in particular, to quantify specific DNA alkylation products such as the premutational 06-alkylguanine (11, 12, 17, 18, 23-27). The procedure allows the selective isolation of a small fraction of DNA molecules carrying a specific alkylation product (06-EtGua) from a very large excess of unmodified DNA, and to amplify the lesion-containing sequences by PCR for quantitative analysis. The specificity of the method for a given DNA adduct results from the use of a high-affinity Mab exhibiting a very low degree of cross-reactivity with other DNA constituents. As shown by calibration experiments with ethylated SV2gpt plasmid DNA, and assuming a Poisson distribution of 06-EtGua residues among these model DNA molecules, >85% of all DNA fragments containing at least 1 06-EtGua residue are retained on the NC membranes as Mab-DNA complexes, the background retention of non-ethylated DNA is negligible. This result could not necessarily be expected since the affinity of Mab ER-6 towards O6-EtGua contained in ds DNA is significantly reduced in comparison to 06-ethyl-2'-deoxyguanosine as a monodeoxynucleoside (data not shown). Moreover the data seem to rule out pronounced effects of neighbouring bases and/or DNA curvature leading to positional differences in the recognition of 06-EtGua by Mab ER-6, since no differences in relative DNA binding efficiency were found between the second and third adsorption/elution cycle (see Table 2). A 20 min-exposure of cultured H5 rat hepatoma cells to 100 ttg of EtNU/ml is the limiting condition tolerated by these cells without severe cytotoxic effects (data not shown). This exposure level results in an 06-EtGua/Gua molar ratio in DNA of about 10-5. From this value it can be calculated that only a few percent of restriction fragments encompassing a particular gene (e.g., the rat IgE heavy chain gene used as a model in this study) will contain at least 1 06-EtGua residue. The resulting small number of gene copies isolated by the immunoaffinity procedure thus precludes their quantification by conventional blotting and hybridization techniques. This is especially true for alkylation frequencies corresponding to 06-EtGua/Gua molar ratios 2 orders of magnitude was found between the sample concentration of genomic template DNA and the amount of specific PCR product. The quantitative analyses subsequent to the PCR reaction are achieved by slot blotting and hybridization with sequence-specific oligonucleotides. Thus it is not necessary to amplify complete restriction fragments but rather only small segments of 100-200 bp, permitting quantification of DNA fragments with different lengths in the same sample. It should also be emphasized that although each DNA fragment
4472 Nucleic Acids Research, Vol. 19, No. 16 isolated via Mab-DNA complexing contains at least 1 06-EtGua residue, this residue is-for statistical reasons-in most fragments not located within the sequence amplified by PCR. In conclusion, we have described an immunoaffinity-PCR approach to the quantification of a defined DNA lesion in DNA restriction fragments encompassing known gene sequences. This procedure is not limited to the isolation of DNA molecules containing 06-EtGua, but may be applied to any stable DNA lesion for which a suitable Mab is available. The sensitivity of the method in terms of the molar content of lesions detectable in genomic DNA is comparable to that of competitive RIA (35), immuno-slot-blot (30), or 32P-postlabeling procedures (39). However, considering the small amount of genomic DNA required, together with the fact that the DNA lesion is not quantified in bulk DNA but rather in a specific gene sequence, the sensitivity of the method by far exceeds that of the latter assays. It has thus become possible, for the first time, to analyze the formation and repair of 06-alkylguanine in specific gene sequences in vivo as demonstrated here for an EcoRI DNA restriction fragment encompassing the rat IgE heavy chain gene. This method can equally be applied to DNA containing other DNA adducts for which suitable Mabs are available, including 0-alkylation products other than 06-EtGua (27, 30, 40).
ACKNOWLEDGEMENTS We thank K.Sendowski for oligonucleotide syntheses and A.Braun-Saure, B.Smialek and I.Spratte for skilful technical assistance. This work was supported by the Deutsche Forschungsgemeinschaft (SFB 102/A70) and by the Commission of the European Communities (ENV-544-D [B]).
REFERENCES 1. Friedberg, E.C. (1985) DNA Repair, W.H. Freeman, New York. 2. Setlow, R.B. (1989) In Castellani, A. (ed.), DNA Damage and Repair, Plenum press, New York, pp. 1-9. 3. Smith, C.A., Mellon, I. (1990) In Obe, G. (ed.), Advances in Mutagenesis Research. Springer-Verlag, Berlin, Vol. 1, pp. 153-194. 4. Bohr, V.A., Smith, C.A., Okumoto, D.S. and Hanawalt, P.C. (1985) Cell, 40, 359-369. 5. Leadon, S.A. (1986) Nucleic Acids Res., 14, 8979-8995. 6. Madhani, H.D., Bohr, V.A. and Hanawalt, P.C. (1986) Cell, 45, 417-423. 7. Mellon, I., Bohr, V.A., Smith, C.A. and Hanawalt, P.C. (1986) Proc. Natl. Acad. Sci. USA, 83, 8878-8882. 8. Mellon, I., Spivak, G. and Hanawalt, P.C. (1987) Cell, 51, 241-249. 9. Mellon, I. and Hanawalt, P.C. (1989) Nature, 342, 95-98. 10. Poirier, M.C., Fullerton, N.F., Patterson, E.D. and Beland, F.A. (1990) Carcinogenesis, 11, 1343-1347. 11. Preussmann, R. and Stewart, B.W. (1984) N-Nitroso Carcinogens ACS Monograph No. 182. American Chemical Society, Washington, D.C., pp. 643-828. 12. Rajewsky, M.F. (1980) In Montesano, R., Bartsch, H. and Tomatis, L. (eds.), Molecular and Cellular Aspects of Carcinogen Screening Tests. IARC Scientific Publ. No. 27. International Agency for Research on Cancer, Lyon, pp. 41-54. 13. Bohr, V.A. (1990) J. Cancer Res. Clin. Oncol., 116, 384-391. 14. Nehls, P., Rajewsky, M.F., Spiess, E., Werner, D. (1984) EMBO J., 3, 327-332. 15. Ryan, A.J., Billett, M.A. and O'Connor, P.J. (1986) Carcinogenesis, 7, 1497-1503. 16. Topal, M.D., Eadie, J.S. and Conrad, M. (1986) J. Biol. Chem., 261, 9879-9885. 17. Glickman, B.W., Horsfall, M.J., Gordon, A.J.E. and Burns, P.A. (1987) Env. Health Perspect., 56, 29-32. 18. Horsfall, M.J., Gordon, A.J.E., Burns, P.A., Zielenska, M., van der Vliet, G.M.E. and Glicknan, B.W. (1990) Environ. Mol. Mutagen., 15, 107-122. 19. Nose, K. and Nikaido, 0. (1984) Biochim. Biophys. Acta, 781, 273 -278.
20. Milligan, J.R. and Archer, M.C. (1988) Biochem. Biophys. Res. Comm., 155, 14-17. 21. Scicchitano, D.A. and Hanawalt, P.C. (1989) Proc. Natl. Acad. Sci. USA, 86, 3050-3054. 22. LeDoux, S.P., Patton, N.J., Nelson, J.W., Bohr, V.A., Wilson, G.L. (1990) J. Biol. Chem., 265, 14875-14880. 23. Rajewsky, M.F., Augenlicht, L.H., Biessmann, H., Goth, R., Huilser, D.F., Laerum, O.D. and Lomakina, L.Y. (1977) In Hiatt, H.H., Watson, J.D. and Winston, J.J. (eds.), Origins of Human Cancer, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y., pp. 709-726. 24. Pegg, A.E. (1984) Cancer Invest., 2, 223-231. 25. Doniger, J., Day, R.S. and DiPaolo, J.A. (1985) Proc. Natl. Acad. Sci. USA, 82, 421-425. 26. Zarbl, H., Sukumar, S., Arthur, A.V., Martin-Zanca, D. and Barbacid, M. (1985) Nature, 315, 382-385. 27. Thomale, J., Huh, N., Nehls, P., Eberle, G. and Rajewsky, M.F. (1990) Proc. Natl. Acad. Sci. USA, 87. 9883-9887. 28. Deschatrette, J., Moore, E.E., Dubois, M., Cassio, D. and Weiss, M.C. (1979) Somat. Cell Genet., 5, 697-718. 29. Huh, N. and Rajewsky, M.F. (1986) Carcinogenesis, 7, 435-439. 30. Nehls, P., Adamkiewicz, J. and Rajewsky, M.F. (1984) J. Cancer Res. Clin. Oncol., 108, 23-29. 31. Rajewsky, M.F., Muller, R., Adamkiewicz, J. and Drosdziok, W. (1980) In Pullman, B., Ts'o, P.O.P. and Gelboin, H. (eds.), Carcinogenesis: Fundamental Mechanisms and Environmental Effects. Reidel, Dordrecht, pp. 207-218. 32. Samson, L., Thomale, J., Rajewsky, M.F. (1988) EMBO J., 7, 2261-2267. 33. Mulligan, P.C. and Berg, P. (1980) Science, 209, 1422-1427. 34. Nehls, P. and Rajewsky, M.F. (1990) Carcinogenesis, 11, 81-87. 35. Muller, R. and Rajewsky, M.F. (1980) Cancer Res., 40, 887-896. 36. Saiki, R.K., Scharf, S., Faloona, F., Muilis, K.B., Horn, G.T., Erlich, H.A. and Arnheim, N. (1985) Science, 230, 1350-1354. 37. Steen, M.L., Hellman, L. and Petersson, U. (1984) J. Mol. Biol., 177, 19-32. 38. Richardson, K.K., Fostel, J. and Skopek, T.R. (1983) Nucleic Acids Res., 11, 8809-8816. 39. Wilson, V.L., Basu, A.K., Essigmann, J.M., Smith, R.A. and Harris, C.C. (1988) Cancer Res., 48, 2156-2161. 40. Adamkiewicz, J., Eberle, G., Huh, N., Nehls, P. and Rajewsky, M.F. (1985) Environm. Health Perspect., 62, 49-55.
