Chem. Res. Toxicol. 2002, 15, 1619-1626
1619
Site-Specific Mutagenesis in Escherichia coli by Adducts Derived from the Highly Carcinogenic Fjord-Region Benzo[c]phenanthrene 3,4-Diol 1,2-Epoxides
N2-Deoxyguanosine
Leilani A. Ramos,† Ingrid Ponte´n,†,‡ Anthony Dipple,†,§ Subodh Kumar,| Haruhiko Yagi,⊥ Jane M. Sayer,⊥ Heiko Kroth,⊥ Govind Kalena,⊥ and Donald M. Jerina*,⊥ Laboratory of Comparative Carcinogenesis (Formerly Chemistry of Carcinogenesis Laboratory), National Cancer InstitutesFrederick, Frederick, Maryland 21702, Environmental Toxicology and Chemistry Laboratory, Great Lakes Center, State University of New York College at Buffalo, 1300 Elmwood Avenue, Buffalo, New York 14222, and Laboratory of Bioorganic Chemistry, National Institute of Diabetes and Digestive and Kidney Diseases, The National Institutes of Health, DHHS, Bethesda, Maryland 20892 Received July 22, 2002
Although there have been numerous studies of site-specific mutagenesis by dGuo adducts of benzo[a]pyrene diol epoxides (B[a]P DEs), the present study represents the first example of site-specific mutagenesis by dGuo adducts of the highly carcinogenic benzo[c]phenanthrene 3,4-diol 1,2-epoxides (B[c]Ph DEs). The eight adducts that would result from cis- and transopening at C-1 of four optically active isomers of B[c]Ph DEs by the N2-amino group of dGuo were incorporated into 5′-TTCGAATCCTTCCCCC (context III) and 5′-GGGGTTCCCGAGCGGC (context IV) at the underlined site. These modified oligonucleotides along with unmodified controls were ligated into single-stranded M13mp7L2, which were then used to transfect SOSinduced Escherichia coli. Upon replication of the lesions in each of the two sequence contexts, mutational analysis of the progeny was performed by differential hybridization. For the 16 adducts, the mutation frequencies varied over 2 orders of magnitude with a reasonably even distribution (0.4-1% for three adducts, 1-2% for six adducts, 3-7.4% for five adducts, and one adduct each at 11 and 39%). For all but this last adduct, the mutation frequency for a given B[c]Ph DE adduct was less than for its B[a]P analogue with the same stereochemistry in the same sequence. For the vectors containing adducts with S configuration at the site of attachment of the hydrocarbon to the dGuo base, the main base substitution was G f T followed by G f A. In contrast, for the vectors containing adducts with R configuration, the main base substitution was G f A. The most notable observation in the present study is the low frequency of mutations induced by the B[c]Ph DE-dGuo adducts relative to their B[a]P counterparts. A possible structural basis for this difference is proposed.
Introduction Mutations in oncogenes and tumor suppressor genes are believed to be the keystones in initiation of the carcinogenic process (1). The mutations can result from replication via translesion synthesis of DNA damaged by physical agents such as UV irradiation as well as by chemicals. The environmentally ubiquitous polycyclic aromatic hydrocarbons (PAHs)1 (2) are metabolized in mammals to reactive diol epoxides (DE) (3) which damage * To whom correspondence should be addressed. † NCIsFrederick (L.A.R. and I.P. have contributed equally to the experimental work). ‡ Current address: AstraZeneca R&D, Safety Assessment, SE-151 85 So¨derta¨lje, Sweden. § Deceased. | SUNY College at Buffalo. ⊥ National Institutes of Health. 1 Abbreviations: DE, diol epoxide; DE-1, diol epoxide in which the epoxide oxygen and the benzylic hydroxyl groups are cis (also called syn); DE-2, diol epoxide in which the epoxide oxygen and the benzylic hydroxyl groups are trans (also called anti); B[c]Ph DE, benzo[c]phenanthrene 3,4-diol 1,2-epoxide; B[a]P DE, benzo[a]pyrene 7,8-diol 9,10-epoxide; PAH, polycyclic aromatic hydrocarbon.
10.1021/tx020073r
DNA by forming stable, covalent adducts predominantly at the exocyclic N2 and N6 amino groups of deoxyguanosine (dGuo) and deoxyadenosine (dAdo), respectively (reviewed in refs 4 and 5). Such damage is most likely responsible for the mutagenic and carcinogenic properties (6-8) of the DEs. To study the mutagenic effects of individual carcinogen-DNA adducts, site-specific mutation studies using individual adducts have been developed (reviewed in refs 9 and 10). Utilizing the M13mp7L2 vector in Escherichia coli, we have previously investigated the effects of adduct stereochemistry and sequence context of the bay-region benzo[a]pyrene 7,8-diol 9,10-epoxide (B[a]P DE) adducts in two DNA 16-mer sequences (see Figure 1) containing dGuo adducts (∼CGA∼ and ∼GGT∼, contexts III and IV, respectively) and in two 16-mers containing dAdo adducts (∼TAG∼ and ∼GAT∼, contexts I and II, respectively) (11-13). We also examined the fjord-region benzo[c]phenanthrene 3,4-diol 1,2-epoxide (B[c]Ph DE) adducts in the same two dAdo sequence contexts (14). It has been of particular interest to compare the mutagenic responses
This article not subject to U.S. Copyright. Published 2002 by the American Chemical Society Published on Web 12/16/2002
1620
Chem. Res. Toxicol., Vol. 15, No. 12, 2002
Ramos et al. Table 1. HPLC Retention Times and Absolute Configurations of Oligonucleotides Containing N2-Deoxyguanosine Adducts Derived from cis- and trans-Opening of B[c]Ph DE-2a N2-dGuo adduct
retention time (min)
configuration
5′-TTCGAATCCTTCCCCC-3′ [III] cis cis trans trans
17.2,b 19.1c 18.8,b 20.8c 19.7,b 15.1,d 18.4c 20.3,b 15.9,d 19.4c
1R 1S 1R 1S
5′-GGGGTTCCCGAGCGGC-3′ [IV] cis cis trans trans
Figure 1. Structures of the adducts formed from cis- and transopening at C-1 of benzo[c]phenanthrene 3,4-diol 1,2-epoxide (B[c]Ph DE) isomers by the exocyclic N2 amino group of dGuo. The ring of the parent hydrocarbon shown in boldface is metabolized to the DEs whose adducts are shown in the partial structures. DE-1 and DE-2 refer to the DE isomers with the benzylic 4-hydroxyl group and the epoxide oxygen cis and trans, respectively. Adducts were inserted at the underlined positions in the sequence contexts shown. Contexts I and II used in previous studies (11, 12, 14) of dAdo adducts are also shown.
of the B[c]Ph DE adducts with the corresponding B[a]P DE adducts, since the fjord-region B[c]Ph DEs are among the most tumorigenic PAH DEs known (7), and in contrast to the bay-region B[a]P DEs, their high tumorigenic response is not limited to a single stereoisomer (7, 8). Thus, in the present work, we have determined the mutational consequences of B[c]Ph DE-dGuo adducts in the same two sequence contexts III and IV as for the previously reported B[a]P DE adducts. Interestingly, we observe that the present B[c]Ph DE-dGuo adducts are in general appreciably less mutagenic in E. coli than their B[a]P counterparts.
Experimental Procedures Materials. The enzymes T4 polynucleotide kinase, T4 DNA ligase and uracil DNA glycosylase, and chemicals X-gal and IPTG were purchased from USB Corp. (Cleveland, OH). EcoRI restriction enzyme and [γ-32P]ATP were obtained from Amersham Corp. (Piscataway, NJ). For preparation and purification of the M13 DNA, QIAprep 8 M13 kits from Qiagen (Valencia, CA) were used. ABI-PRISM Dye terminator Cycle Sequencing Ready Reaction kits, obtained from Perkin-Elmer (Foster City, CA), were used for DNA sequencing. Bacteriophage M13mp7L2 and E. coli strain SMH77 were generous gifts from Dr. C. W. Lawrence (University of Rochester, NY). Synthesis of Oligonucleotides Containing cis- and trans-Opened B[c]Ph DE-2 dGuo Adducts. Context III containing cis- and trans-opened adducts as well as context IV containing cis-opened adducts corresponding to DE-2 (benzylic 4-hydroxyl group and epoxide oxygen trans, cf. Figure 1) were made on a 2-µmol scale by a semiautomated procedure as described (15, 16). Support bound 12-mers were prepared on a DNA synthesizer using commercial dC-controlled pore glass (cpg). The manual coupling step (16) utilized the mixed dia-
11.7d 13.1d 13.1d,e 13.1d,e
1R 1S 1R 1S
a The modified base is underlined. For all chromatographic procedures, a linear gradient of acetonitrile in 0.1 M (NH4)2CO3 buffer, pH 7.5, that increased the percentage of acetonitrile from zero to 17.5% over 25 min was used. Where more than one retention time is given, the compounds were purified by successive chromatographic steps utilizing the columns and conditions as indicated in the footnotes. b On a Hamilton PRP-1 column (7 µm, 9.5 × 250 mm) at 50 °C eluted at 4.5 mL/min. c On a Phenomenex Luna phenyl-hexyl column (4.6 × 250 mm) at 50 °C eluted at 1.5 mL/min. d On a Hamilton PRP-1 column, (10 µm, 4.1 × 250 mm) at 75 °C eluted at 1.5 mL/min. e The pure diastereomers were prepared from separated phosphoramidites that were single diastereomers at C-1; see text.
stereomers (1R/1S) of the appropriate adducted O-allyl dGuo phosphoramidites (10-12 mg) prepared as described (16). After completion of the sequence and removal of the 5′-DMT group on the synthesizer, the O-allyl protecting group was cleaved using a palladium catalyst as described (16). After cleavage (NH4OH, 58 °C, 16 h) from the cpg support, the oligonucleotides were purified by HPLC (Table 1). Final yields of each purified diastereomer ranged from ∼1 to 4 A260. For CD spectra of selected oligonucleotides, see the Supporting Information. Absolute configurations (Table 1) of the adducts in the oligonucleotides were assigned by enzymatic hydrolysis (17) to the nucleoside level with snake venom phosphodiesterase/alkaline phosphatase, followed by HPLC to isolate the nucleoside adducts, whose CD spectra in MeOH were compared with the known B[c]Ph DE-2 dGuo adducts (18). Context IV oligonucleotides containing trans-opened DE-2 dGuo adducts as their 1R/1S diastereomers were co-chromatographic in several HPLC systems. Thus, the four phosphoramidite isomers (diastereomers at C-1 and at phosphorus) were separated by HPLC on a Vertex LiChrosorb silica-60 column (16 × 250 mm) (Sonntek, Inc.) eluted at 9 mL/min with a gradient of EtOAc in n-hexane that increased the proportion of EtOAc from 40 to 70% over 60 min: tR 30.5 (1), 33.8 (2), 36.2 (3), and 43.3 (4) min. CD spectra of these O-allyl dGuo phosphoramidites were not directly diagnostic of their absolute configuration. However, CD spectra of fractions 1 and 3 were identical. Similarly, CD spectra of fractions 2 and 4 were identical but different from 1 and 3. Thus, fractions 1 and 3 have the same chirality at C-1 and differ only in chirality at phosphorus, whereas fractions 2 and 4 have the same chirality at C-1 that is opposite to that of 1 and 3. The separated phosphoramidite fraction 3 was used for oligonucleotide synthesis following the procedure described above, and the absolute configuration of the adduct in the fully deprotected oligonucleotide prepared from this optically pure phosphoramidite was shown to be 1R on the basis of the CD spectrum (cf. ref 18) of the nucleoside adduct obtained on enzymatic hydrolysis of the oligonucleotide (17). Thus fractions 1 and 3 as well as their derived oligonucleotides have the same 1R chirality at C-1. Conversely, phosphoramidite fractions 2 and 4 as well as their derived oligonucleotides must have 1S chirality at C-1. Synthesis of Oligonucleotides Containing cis- and trans-Opened B[c]Ph DE-1 dGuo Adducts. All oligonucle-
Mutagenesis by Benzo[c]phenanthrene dGuo Adducts Table 2. HPLC Retention Times and Absolute Configurations of Oligonucleotides Containing N2-Deoxyguanosine Adducts Derived from cis- and trans-Opening of B[c]Ph DE-1a N2-dGuo adduct
retention time (min)
configuration
5′-TTCGAATCCTTCCCCC-3′ [III] cis (E) cis (L) trans (L) trans (E)
17.7,b 18.5c 18.6b 17.4d 22.1d
1R 1S 1R 1S
5′-GGGGTTCCCGAGCGGC-3′ [IV] cis (E) cis (L) trans (L) trans (E)
15.4b 14.5b 14.4d 18.1d
1R 1S 1R 1S
a The modified base is underlined. E and L refer to the elution order of the precursor O6-allyl 3′,5′-di-TBDMS 2,3,4-triacetates (19) on silica. For all chromatographic purifications of oligonucleotides, a linear gradient of acetonitrile in 0.1 M (NH4)2CO3 buffer, pH 7.5, that increased the percentage of acetonitrile from zero to 17.5% over 25 min was used. b On a Hamilton PRP-1 column (10 µm, 4.1 × 250 mm) at 75 °C eluted at 1.5 mL/min. c On a Phenomenex Luna phenyl-hexyl column (5 µm, 4.6 × 250 mm) at 60 °C eluted at 1.2 mL/min. d On a Waters Associates Xterra MS C18 column (5 µm, 4.6 × 250 mm) at 75 °C eluted at 1.5 mL/min.
otides containing adducts corresponding to DE-1 (benzylic 4-hydroxyl group and epoxide oxygen cis, cf. Figure 1) were synthesized from phosphoramidites that were single diastereomers at C-1 and were derived from a pair of diastereomerically pure disilyl triacetates (19) of known absolute configuration. Oligonucleotides were made as above (15, 16) on a 2-µmol scale using 10 mg of the appropriate phosphoramidites. For chromatographic purification see Table 2. Final yields of purified oligonucleotides ranged from ∼1 to 26 A260 depending on manual coupling yield and difficulty of purification. CD spectra of selected oligonucleotides are given in the Supporting Information. Unmodified oligonucleotides were obtained from Life Technologies. All oligonucleotides were further purified by denaturing 20% polyacrylamide gel electrophoresis (PAGE) as previously described (13, 14). Concentrations of oligonucleotides were determined by UV measurement, and their purity was confirmed by further electrophoresis after end labeling with [γ-32P]ATP. Construction of M13 Genomes. M13mp7L2 vectors were constructed following a basic protocol developed by Lawrence and co-workers (20, 21) with modifications as described (cf. ref. 13 and references therein). The apparent ligation efficiency, defined as the ratio of circular DNA to total DNA, was estimated by loading an aliquot of the genome construct (∼100 ng) on an agarose gel, separating the components by electrophoresis followed by Southern blotting and radioactive probing (11-14). The amounts of closed circular and linear DNA were quantified by PhosphorImager analysis (Storm 860, Molecular Dynamics, Inc.). Transfection into E. coli and Mutation Analysis of Progeny Phage. E. coli (SMH77) cells were SOS-induced by irradiation with UV light (254 nm) at 40 J/m2 for 40 s (22) and then made competent by CaCl2 treatment. Transfection and replication protocols were as described (cf. ref 13 and references therein). Progeny phage containing the 16-mer insert were detected as blue plaques (11). The survival is defined as the number of blue plaques obtained upon transfection with equal amounts of adducted relative to control DNA and is a function of both ligation efficiency and subsequent replication past the adduct. The progeny phage was analyzed for base substitution mutations by differential hybridization with probes complementary to each of the four possible base substitutions at the adduct site, followed by autoradiography as previously described (11). The DNA from any blue plaques that did not align to any signal in the autoradiograph was extracted and sequenced.
Chem. Res. Toxicol., Vol. 15, No. 12, 2002 1621
Results Oligonucleotides. The present set of sixteen oligonucleotides (eight diastereomeric dGuo adducts derived from cis- and trans-opening of both enantiomers of the diastereomeric B[c]Ph DE-1 and DE-2 (cf. Figure 1) in sequence contexts III and IV) complete the “matrix” of all possible dGuo and dAdo adducts in these two sequences, which we had designed to probe the effects of adduct stereochemistry and neighboring sequence on mutagenic response. Use of O-allyl protected dGuo derivatives (16, 19) as intermediates in the synthesis of the required phosphoramidites made possible the facile preparation of these new, B[c]Ph DE-dGuo adducted oligonucleotides. The allyl group was removed prior to cleavage of the oligonucleotides from the cpg support as described (16). We had previously observed (16) that the pairs of oligonucleotides formed from the mixed diastereomers (R and S at the point of attachment of the hydrocarbon to N2 of dGuo) of both B[a]P and B[c]Ph DEadducted dGuo phosphoramidites separated quite readily on reversed-phase HPLC and that the R-adducted oligonucleotides derived from both hydrocarbons eluted earlier than their S-diastereomers, independent of sequence. These empirical observations are borne out in the present study for both the cis- and trans-opened B[c]Ph DE-2 dGuo adducts in context III and for the cis-opened B[c]Ph DE-2 adducts in context IV (Table 1). However, the two context IV (∼GGT∼) oligonucleotides containing 1R and 1S trans-opened B[c]Ph DE-2 adducts, derived from the mixed diastereomers of the phosphoramidite, unexpectedly failed to separate in several reversed-phase HPLC systems and were thus synthesized from the separated 1R and 1S phosphoramidite diastereomers. To preclude this potential difficulty, phosphoramidites that were single diastereomers at C-1 were used in subsequent syntheses of both oligonucleotide sequences containing the B[c]Ph DE-1 adducts. Construction of M13 Genomes. The circular vectors were constructed by annealing and ligating a 56-mer scaffold/16-mer oligonucleotide duplex into linearized M13 DNA. The ligation efficiencies, defined as the amount of circular DNA as a fraction of the sum of linear and circular DNA, for the control oligonucleotides (unadducted) were 38 and 41% in context III and context IV, respectively (for sequence contexts see Figure 1). The adducted oligonucleotides gave somewhat lower ligation efficiencies, which in most cases did not differ significantly for different adducts (Table 3). Upon transfection into SOS-induced E. coli cells, the adducted M13 constructs gave fewer blue plaques compared to the control constructs. The survival values, defined as the ratio of the number of blue plaques obtained from equal amounts of total DNA from adducted relative to unadducted samples, ranged from 7.9 to 50.7% (context III) and 2.1 to 59.4% (context IV). Survival values were not corrected for ligation efficiency. For details of survival values in each context and each experiment see the Supporting Information. Mutational Analysis. The mutational frequencies for the adducted deoxyguanosine inserted into context III and context IV (Figure 1) are given in Table 3. Except for a few of the adducted constructs, the overall mutation frequencies were relatively low compared to our previously observed B[a]P DE-dGuo adducts. In context III, the mutation frequencies ranged from 0.6 to 38.6%, while
1622
Chem. Res. Toxicol., Vol. 15, No. 12, 2002
Ramos et al.
Table 3. Ligation Efficiency (LE), Survivala Relative to Control and Number and Frequency (MFtot) of Targeted Base Substitutions for B[c]Ph DE-dGuo Adducts sequence context context III (5′-TTCGAATCCTTCCCCC) number of plaques controlb trans DE-1/R trans DE-2/R cis DE-1/R cis DE-2/R trans DE-1/S trans DE-2/S cis DE-1/S cis DE-2/S a
GfG
GfT
GfA
GfC
MFtot (%)
11 394 5110 3521 2221 4603 3871 3438 2346 865
0 21 14 207 13 22 9 55 82
45 33 38 1187 47 6 8 25 15
1 2 0 4 0 47 2 17 7
0.4 1.1 1.5 38.6 1.3 1.9 0.6 4.0 10.7
context IV (5′-GGGGTTCCCGAGCGGC) number of plaques
LE (%)
survival (%)
GfG
GfT
GfA
GfC
MFtot (%)
LE (%)
survival (%)
38 30 29 35 31 31 18 35 31
100 50.7 38.1 17.1 36.4 28.6 37.3 11.3 7.9
5124 2901 748 983 918 1461 1007 1361 87
0 9 7 18 15 9 4 14 5
0 24 14 27 38 2 0 11 1
0 0 2 1 4 0 0 1 1
