Sequence Context Modulation of Translesion Synthesis at a Single N

When polymerase I Klenow fragment exo− is used, translesion synthesis (TLS) is observed with efficiencies varying according to the position of the a...
0 downloads 0 Views 96KB Size
144

Chem. Res. Toxicol. 1999, 12, 144-150

Sequence Context Modulation of Translesion Synthesis at a Single N-2-Acetylaminofluorene Adduct Located within a Mutation Hot Spot Dominique Y. Burnouf,* Roman Miturski,† and Robert P. P. Fuchs Groupe d’Epide´ miologie Mole´ culaire du Cancer, UPR 9003, Centre National de la Recherche Scientifique, Institut de Recherche sur les Cancers de l’Appareil Digestif, 1 place de l’hopital, 67097 Strasbourg Cedex, France Received August 17, 1998

Oligonucleotides containing a single N-(deoxyguanosin-8-yl)acetylaminofluorene lesion (dGuoC8-AAF) at each guanine residue of the sequence (5′-G1G2G3) have been used as templates for in vitro primer extension reactions by several DNA polymerases [Escherichia coli DNA polymerase III holoenzyme, its R subunit, DNA polymerase I Klenow fragment proficient (exo+) or deficient (exo-) in its 3′ f 5′ exonuclease activity, and Sequenase]. The dGuo-C8-AAF lesion appears to be a strong block for all DNA polymerases: exo+ DNA polymerases stop one nucleotide before encountering the lesion, while partial incorporation opposite the lesion is observed only with enzymes devoid of the exonuclease activity. The efficiency of incorporation across from the adduct depends on both the DNA polymerase and the position of the lesion. When polymerase I Klenow fragment exo- is used, translesion synthesis (TLS) is observed with efficiencies varying according to the position of the adduct (G2 > G1 > G3). Sequencing of the TLS products shows that error-free TLS is observed only when the AAF lesion is bound to G1, while all TLS events occurring at G2- or G3-AAF adducts are mutagenic. The major mutational event is a G deletion (27, 76, and 55% of the events for G1, G2, and G3, respectively), while two-G deletions occur to a lesser extent (17-30%). These results are discussed in view of the slippage model developed for frameshift mutagenesis occurring during translesion synthesis at replication blocking lesions.

Introduction Occurrence of mutations in the genome of organisms is considered the initial event in the processes that lead to cancer or genetic diseases. However, the molecular mechanisms that trigger the mutational event and specify the nature of the mutations are not clearly understood. Reactions of mutagens or carcinogens with DNA alter its chemical structure and result in the formation of lesions, the large majority of which are processed by highly efficient repair mechanisms that restore the original DNA sequence (1). If these repair processes are saturated or defective, persistent lesions will interfere with DNA replication and eventually give rise to induced mutations. In Escherichia coli, a set of coregulated inducible genes, referred to as the SOS system, is induced in response to DNA damage (for a review, see ref 2). Several SOS genes, including recA, umuD, and umuC, are required for UVand chemical-induced mutagenesis (2, 3), but their biochemical functions in the mutational processes are still unknown. A two-step model has been proposed for UV-induced mutagenesis, in which the DNA polymerase incorporates a nucleotide opposite the DNA lesion, but is no longer able to elongate the distorted primertemplate junction (4). A ternary complex umuD′2C [umuD′ * To whom correspondence should be adressed. E-mail: Dominique. [email protected]. Fax: (33) 03 88 11 90 99. † Present address: 2nd Department of Gynecological Surgery, Medical Academy, Jaczewski Street, 8, 20-290 Lublin, Poland.

is the active form of the umuD protein in mutagenesis that results from a RecA*-mediated self-cleavage of umuD (5-7)], along with the recA protein, interacts with the stalled polymerase and helps it to resume elongation past the lesion (4, 8-10). We have studied mutagenesis induced by N-2-acetylaminofluorene, a rat hepatocarcinogen which mainly binds in vivo to the C8 position of guanines. Two types of adducts are found in rats: N-(deoxyguanosin-8-yl)aminofluorene (dGuo-AF)1 (70%) and the acetylated form N-(deoxyguanosin-8-yl)-acetylaminofluorene (dGuo-AAF) (25%). In a forward mutation assay carried out in E. coli, AF adducts induce mainly base substitutions (11), while AAF adducts trigger almost exclusively frameshift mutations at repetitive sequences, which are thus considered mutational hot spots (12, 13). Such sequences were extensively studied to elucidate the molecular processes of AAF-induced frameshift mutagenesis (14-19). Recently, the use of single-stranded AAF-monomodified vectors led to the identification, in vivo, of two translesion synthesis (TLS) pathways: the nonslipped and the slipped TLS pathways (18). A common step shared by both pathways is the incorporation of a dCMP residue opposite the dGuo-C8-AAF lesion. Elongation then pro1 Abbreviations: AAF, N-2-acetylaminofluorene; AF, N-2-aminofluorene; PolIKF(exo-), E. coli DNA polymerase I Klenow fragment, 3′ f 5′ exonuclease deficient; X-Gal, 5-bromo-4-chloro-3-indolyl β-Dgalactoside; IPTG, isopropyl β-D-thiogalactopyranoside; TLS, translesion synthesis; LT, lesion terminus; SMI, slipped mutagenic intermediate.

10.1021/tx9801920 CCC: $18.00 © 1999 American Chemical Society Published on Web 01/20/1999

Sequence Context Modulation of Translesion Synthesis

Chem. Res. Toxicol., Vol. 12, No. 2, 1999 145

Figure 1. (A) Sequences of the four different oligonucleotides used to construct the 63-mer templates used in the elongation experiments. The 14-SmaI oligonucleotide contains a SmaI site (5′-CCCGGG) which is the target sequence for in vitro modification by N-acetoxy-N-2-acetylaminofluorene at G1, G2, or G3. The purified monomodified or unmodified oligonucleotides were hybridized onto a 54-mer scaffold C, along with oligonucleotides A (14-mer) and B (35-mer). The 63-mer long ligation products are purified on polyacrylamide gel and then used as templates for in vitro primer elongation experiments. (B) Scheme of in vitro elongation experiments by DNA polymerases of 5′-32P-labeled P20 primer hybridized onto a 63-mer template bearing a single G1-AAF adduct. The oval represents the AAF moiety covalently bound to either of the three guanines of the target site in the different constructs. (C) Nomenclature used for intermediate products of elongation and full-length TLS products.

ceeds, in the nonslipped TLS pathway, from the so-called lesion terminus (LT), i.e., the replication intermediate where the incorporated dCMP is located opposite the lesion, to yield the wild-type sequence. Alternatively, in the slipped TLS pathway, the LT isomerizes into a socalled slipped mutagenic intermediate (SMI), where the incorporated dCMP pairs with the guanine residue located 5′ to the lesion, thus forming a post-LT (15). The SMI, stabilized by the presence of the AAF lesion (20), is further elongated to yield a frameshift mutation. These two TLS pathways exhibit distinct genetic requirements; both require the induction of the SOS system, but while the nonslipped TLS pathway is dependent upon the umuDC function, slipped TLS is not, although it requires another, as yet unidentified, SOS-controlled function called npf (nar processing factor) (18, 21). In this work, we analyzed the sequence context modulation of translesion synthesis at a single AAF adduct located within a mutation hot spot. Oligonucleotides bearing a single AAF lesion at each guanine of a three-G sequence are used as templates for in vitro primer elongation reactions performed by several DNA polymerases. The AAF lesion appears to strongly block all DNA polymerases, with the exception of DNA polymerase I KF exo- which is found to perform TLS. Slipped and nonslipped TLS events are both observed, but their occurrence and efficiencies depend on the position of the lesion within the three-G target sequence.

Materials and Methods Enzymes and Reagents. E. coli DNA polymerase I Klenow fragment (PolIKF) and the corresponding exonuclease deficient enzyme [PolIKF(exo-)], T4 DNA polynucleotide kinase, T4 DNA

ligase, calf intestinal alkaline phosphatase (CIAP), and EcoRI and BamHI restriction enzymes were purchased from New England Biolabs (Beverly, MA). Sequenase 2.0 and T7 DNA polymerase were from Amersham Pharmacia Biotech (Saclay, France), and pfu polymerase was from Stratagene (La Jolla, CA). The enzymes were used as suggested by the suppliers. E. coli polymerase III holoenzyme (PolIII HE) was prepared as described previously (21, 22). The R subunit of E. coli polymerase III was a gift from H. Maki (Nara, Japan). Ultrapure grade dNTP, [γ-32P]ATP, and [R-35S]dCTP were purchased from Amersham Pharmacia Biotech. Oligonucleotides were synthesized on an Applied Biosystems model 380B synthesizer and purified by reverse-phase HPLC on a C18 column (Zymark, Hopkinton, MA). Some of them were purchased purified from Eurogentec (Seraing, Belgium) or from Bioprobe Systems (Montreuil sous Bois, France). Construction of Site Specifically Modified Template. The construction of site specific AAF-monomodified oligonucleotides has been described previously (24-26). The 14-SmaI mer (Figure 1) was treated in vitro with N-acetoxy-N-2-acetylaminofluorene as previously described (27). The purification and characterization of the various AAF-monomodified oligonucleotides have already been published (24, 27). The 63-mer template was obtained by ligation of three oligonucleotides: A (14-mer), AAF-modified or unmodified 14-SmaI, and B (35-mer), hybridized on a complementary scaffold C (54-mer) (Figure 1). Oligonucleotide A was radiolabeled using [γ-32P]ATP at a low specific activity (20 Ci/mmol) to follow the ligation reaction. Hybridization was carried out in 50 mM Tris-HCl (pH 7.8), 10 mM MgCl2, and 10 mM DTT buffer, using equimolar amounts of each oligonucleotide. The hybridization mix was heated for 15 min to 70 °C and cooled slowly to room temperature. ATP (1 mM) and 25 units of T4 DNA ligase were then added for 1 h at 16 °C. DNA was then ethanol precipitated, resuspended in 80% formamide/20% blue mix, and loaded onto a 20% denaturing (8 M urea) polyacrylamide gel. The radioactive band corresponding

146 Chem. Res. Toxicol., Vol. 12, No. 2, 1999

Burnouf et al.

to the 63-mer ligation product was sliced out from the gel, eluted overnight in a 0.16 M ACONa, 1 mM EDTA solution, ethanol precipitated, and counted to determine the amount of material recovered. The 5′-32P was then removed with CIAP to avoid any interference of the 32P label from the template with the 32P label introduced during the primer elongation experiments. The products were kept at -20 °C in TE. In Vitro Replication Assays. The purified oligonucleotides (15 fmol), AAF-monomodified or not, were used as templates for the elongation of a 5′-32P-labeled primer P (20-mer) (3000 Ci/mmol) (Figure 1), which was previously purified on a 20% denaturing polyacrylamide gel and processed as described for the 63-mers. Hybridization was carried out in the corresponding polymerase buffer, using a 2:1 template:primer ratio. After being heated at 65 °C for 15 min, the mix was slowly cooled to room temperature. The efficiency of the hybridization was checked by running an aliquot of the hybridization mix on a 12% nondenaturing polyacrylamide gel. Primer extension was started by adding the DNA polymerase to the hybridization mix. PolIKF reaction buffer was 50 mM Tris-HCl (pH 7.6), 10 mM MgCl2, 0.1 mM DTT, 2% glycerol, and each dNTP at 40 µM; 0.1 unit of PolIKF(exo+) and 0.4 unit of PolIKF(exo-) were added for 10 min at 30 °C in 10 µL. Sequenase reaction was carried out at 37 °C for 10 min with 0.7 unit of enzyme in 10 µL containing 40 mM Tris-HCl (pH 7.6), 20 mM MgCl2, 12 mM DTT, 50 mM NaCl, and each dNTP at 40 µM. DNA polymerase III reactions were carried out in 10 µL of 20 mM Tris-HCl (pH 7.5), 8 mM MgCl2, 8 mM DTT, 4% glycerol, each dNTP at 300 µM, 1 mM ATP, and 80 mg/mL BSA (Boehringer). Enzyme (80 units) was added for 5 min at 30 °C. Conditions for the R subunit were the same except that 120 units was used for 20 min at 30 °C and no ATP was added in the reaction mix. Reactions were stopped by placing the reaction vials in a cold ethanol bath, followed by proteinase K digestion at 55 °C. The elongation products were then ethanol precipitated and loaded onto a 12% denaturing (8 M urea) polyacrylamide gel. Analysis was carried out on a Phosphorimager 445SI (Molecular Dynamics), and the bands were quantified using the ImageQuaNT software (Molecular Dynamics). Analysis of TLS Products. The 62- and 63-mer replication products, resulting from TLS performed by PolIKF(exo-), were purified on 12% PAGE as described. They were further PCR amplified with a NJR PTC 150 minicycler, using P1 and P2 primers (5′-CGCGGATCCCCGGTGATGTCGGCG and 5′-CCGGAATTCCATCGTGGCCGGC, respectively). We first performed asymmetric PCR using primers P1 (1 pmol) and P2 (50 pmol) and pfu polymerase, to selectively amplify the TLS products synthesized by PolIKF(exo-). Contamination of the full-length TLS products by the modified template is unlikely because, as compared to an unmodified oligonucleotide, the AAF adduct is responsible of a slower migration of the monoadducted template on PAGE (27, 39). Moreover, the pfu polymerase, which has a proofreading activity, would not be able to achieve TLS on any copurified AAF-modified template. The linearly amplified singlestranded PCR products, resulting from elongation of P2 primer, were purified on denaturing PAGE and further PCR amplified using P1 and P2 primers (50 pmol). P1 and P2 primers were designed so that the PCR products derived from the 62- and 63-mer products will disrupt the reading frame of the LacZ gene when cloned into the polylinker of vector pUC118. Using the LacZ′ R complementation assay, such recombinant plasmids yield white clones on indicator plates containing X-Gal and IPTG. PCRs were carried out in a 25 µL volume supplemented with each dNTP at 40 µM and 1 unit of pfu DNA polymerase. After 5 min at 94 °C, cycling conditions were 30 s at 94 °C, 30 s at 50 °C, and 30 s at 72 °C for 30 cycles, with a final extension at 72 °C for 5 min. The final PCR products were gel purified, digested by EcoRI and BamHI restriction enzymes, and cloned into the polylinker of vector pUC118. The ligation products were transformed by electroporation (Bio-Rad gene pulser) into XL1Blue E. coli cells. Bacteria were plated on LB agar plates containing ampicillin (100 µg/mL), X-Gal (50 µg/mL), and IPTG

Figure 2. Primer extension experiments performed with E. coli DNA polymerase I proficient (exo+) or deficient (exo-) in its 3′ f 5′ exonuclease (proofreading) activity, Sequenase, E. coli DNA polymerase III holoenzyme, and its R subunit. The lane noted L corresponds to the loading of a mix containing the four sequencing reaction mixtures of the 63-mer. Lanes Go correspond to the elongation of 5′-32P-labeled P20 hybridized to the unmodified 63-mer template. Lanes G1-G3 correspond to the elongation of primer hybridized to the templates modified on the corresponding guanine. The numbering of the template sequence and position of the SmaI site are indicated on the left side of the figure. Table 1. Percentages of Incorporation Opposite (Position L0) or One Nucleotide Before (Position L-1) the Different AAF Adducts (G1, G2, or G3) by Different DNA Polymerasesa Klenow fragment Pol III Holo (exo+)

Klenow fragment (exo-) Sequenase R subunit

G1 G2 G3 G1 G2 G3 G1 G2 G3 G1 G2 G3 G1 G2 G3 L0 6 9 9 9 15 9 82 70 45 67 73 79 48 30 9 L-1 94 91 91 91 85 91 18 30 55 33 27 21 52 70 91 a

Data of three independent experiments.

(60 µg/mL). White clones were further sequenced with T7 DNA polymerase and analyzed on a 8% denaturing (8 M urea) polyacrylamide gel.

Results Primer Elongation on AAF-Monomodified Templates by Different DNA Polymerases. Replication of the unmodified 63-mer by DNA polymerase I Klenow fragment (Figure 2), or by any of the other DNA polymerases (data not shown), gives full-length elongation products. With any of the modified templates, shorter products, which correspond to the arrest of the enzyme in the vicinity of the lesion, are observed (Figure 2). DNA polymerases proficient in the 3′ f 5′ exonuclease activity, such as Pol III holo or PolIKF, display a single major block, which represents about 90% of the elongated material (Table 1) and corresponds to the arrest of the polymerase one nucleotide before the lesion (position

Sequence Context Modulation of Translesion Synthesis

Chem. Res. Toxicol., Vol. 12, No. 2, 1999 147

Figure 3. Steady state kinetic elongation of 5′-32P-labeled P20 by E. coli DNA polymerase I deficient in its 3′ f 5′ exonuclease (proofreading) activity. Elongation was performed under the same conditions as described in the legend of Figure 2. At each indicated time (from 0 to 75 min), an aliquot was taken and processed as described. Pausing sites at positions L-1, L0, L+1, and L+2 (one nucleotide before, opposite, and one and two nucleotides past the lesion, respectively) and the 62-, 63-, and 64-mers are indicated. G0, G1-, G2-, and G3-AAF refer to the templates used for the elongation experiments. P20 denotes the nonelongated primer. 62- and 63-mer products were eluted from the gel and PCR amplified for sequence analysis.

L-1). Alternatively, when DNA polymerases deficient in the proofreading activity are used [PolIKF(exo-), Sequenase 2.0, and the PolIII R subunit], doublets are observed one nucleotide before (L-1) and opposite (L0) the dGuoC8-AAF adduct. The incorporation efficiency opposite the lesion (L0) varies with the DNA polymerase. Sequenase and PolIKF(exo-) efficiently incorporate at this position, while the R subunit is at least 2-fold less efficient (Table 1). The position of the lesion within the three-G sequence also modulates the incorporation efficiency at L0; while Sequenase incorporates roughly to the same extent (∼70%), irrespective of the adduct position (Table 1), decreases in the incorporation efficiencies of PolIKF(exo-) and the R subunit are observed when the adduct is located at G1 > G2 > G3, respectively (Figure 2 and Table 1). PolIKF(exo-) Elongation Products Analysis. Primer elongation of G2-AAF templates by PolIKF(exo-) revealed two weak bands, 62 and 63 nucleotides long, suggesting that this enzyme is able to achieve complete translesion synthesis (TLS) of AAF adducts, in vitro. To further investigate the nature of the complete TLS events mediated by PolIKF(exo-), we analyzed the appearance of full-length elongation products for each monomodified template (G1-, G2-, and G3-AAF) as a function of time of incubation (Figure 3). Kinetics of Appearance of the TLS Products. Unmodified template replication yields 63-mer products which are rapidly converted into 64-mers due to the addition of a non-template-encoded extra nucleotide by the exo- enzyme. For all AAF-monomodified templates, increasing the reaction times results in the formation and differential accumulation of 62- and 63-mer products (Figure 3). 64-mer products are only observed with the

G1-AAF templates. The accumulation of full-length products reflects the ability of the polymerase to incorporate across from each adduct and to elongate past the lesion. Elongation proceeds up to the L-1 position for G2- and G3-AAF templates, but for G1-AAF, the polymerase is slowed several nucleotides before the lesion, as stop bands appear at positions L-3 and L-2. However, the L-1 band does not accumulate, while intermediate products at positions L0 and L+1 and full-length TLS products do. Consequently, the incorporation of a nucleotide at position L0 can be visualized by the kinetics of disappearance of the L-1 band (Figure 3). As judged by the intensity of this band, incorporation across from G-AAF adducts occurs with various efficiencies depending on the position of the lesion (G1 > G2 > G3) (Figure 3). Moreover, the polymerase differentially elongates the replication intermediate from position L0; while a faint band is detected at position L+1 for G3-AAF templates, one or two bands are observed immediately after the adduct position for G1- or G2-AAF constructs, respectively (Figure 3). The material at position L+1 for G1-AAF accumulates as a function of time, while such an accumulation is not observed for G2-AAF at position L+1 or L+2. The efficiencies of these incorporation steps strongly influence the kinetics of appearance of the fulllength TLS products; for G2-AAF templates, they appear after incubation for only 2 min and are readily seen after 10 min, while TLS products derived from G1- and G3AAF templates appear after reaction for 10 min and slowly accumulate. These TLS products represent 12, 45, and 5% of the total elongated primer for G1-, G2-, and G3-AAF constructs, respectively, after incubation for 1 h. Thus, the yield of full-length TLS products depends not only on the kinetics of the incorporation step across

148 Chem. Res. Toxicol., Vol. 12, No. 2, 1999

Burnouf et al.

Table 2. Types and Percentages of AAF-Induced Mutations Detected within the SmaI Sequence of PolIKF(exo-) TLS Productsa G1-AAF total wt -G -2G -GC -C -3G G>T G>A G>C

G2-AAF

G3-AAF

#

%

#

%

#

%

29 12 8 5 1 1 2 -

100 41.3 27.5 17.2 3.4 3.4 6.8 -

21 16 5 -

100 76.2 23.8 -

26 15 8 1 1 1 1

100 57 30 3.8 3.8 3.8 3.8

a # represents the total number of clones analyzed. The relative contribution of dGMP misincorporation to the overall mutagenesis is calculated as the ratio of -1G, -2G, and -3G for G1-, G2-, and G3-AAF, respectively, to the total number of mutants, expressed as a percentage (see the text).

from the lesion but also on the subsequent elongation steps (L+1 and L+2), all these reactions being sequencedependent (G2 > G1 > G3). Sequence Analysis of Complete TLS Products. Fulllength TLS products 62- and 63-mers were purified on polyacrylamide gels and PCR amplified for further analysis by sequencing. No mutation was observed within the three-G site among the elongation products derived from the unmodified template G0. Sequence analysis of the elongation products of the AAF-modified templates shows that error-free TLS is observed only for the G1AAF adduct and accounts for 41% of all G1 sequences (Table 2). The intensity of the 64-mer band, which contains 63-mer products further elongated with an extra nucleotide, is half of the intensity of the 63-mer band. Consequently, error-free TLS ranges between 60 and 70% and represents the major TLS process occurring when AAF is bound to G1. In contrast, all elongation products derived from G2- or G3-AAF templates present a mutation within the three-3G sequence. The major mutational event observed is a G deletion and accounts for 76, 57, and 27% of mutated plasmids for G2, G3, and G1, respectively (Table 2). Two-G deletions occurred at a lower frequency (17, 24, and 30% for G1, G2, and G3, respectively). It should be noted that these -2 frameshift events, which were not detected previously in forward mutation assays (13), were selected here as 62mer products due to the incorporation by the exopolymerase of a nontemplated nucleotide. Some other minor mutationnal events, frameshifts and substitutions accounting for less than 10% of all sequences, are also detected for G1 and G3 constructs (Table 2).

Discussion In vivo, replication of double-stranded DNA molecules bearing blocking lesions, such as UV-induced pyrimidine dimers or AAF adducts, occurs via two major strategies: damage avoidance, which represents the major pathway and occurs through post-replication recombinationnal repair or polymerase strand switching, and translesion synthesis (TLS) where the DNA polymerase reads through the lesion (17, 28). TLS results from successful completion of kinetically unfavorable replication steps, i.e., incorporation of a nucleotide opposite and subsequent elongation past the lesion (29). We will refer to the intermediate products derived from replication one nucleotide before,

Figure 4. Error-free and mutagenic TLS pathways of an AAF lesion. In the pre-LT step, the DNA polymerase is blocked one nucleotide before the dGuo-C8-AAF lesion which can adopt a syn or anti configuration. In the latter case, the DNA polymerase incorporates a dCMP opposite the adduct, forming a lesion terminus (LT) which can be further elongated, provided that the adducted base retains the anti configuration. If it adopts a syn configuration, elongation is hindered and a slippage can occur. The resulting SMI can be further elongated.

opposite, or one nucleotide past the lesion as L-1, L0, and L+1, respectively (Figure 1C). Moreover, we will distinguish two kinds of full-length TLS products: nonslipped and slipped. Nonslipped TLS results from the elongation of a dC::dG-AAF lesion terminus (LT) (Figure 4), thus yielding an error-free TLS product. Alternatively, slipped TLS results from the elongation of a slipped mutagenic intermediate (SMI), where the nucleotide incorporated opposite the adduct pairs with the template guanine residue located 5′ to the lesion, thus forming a postlesion terminus (post-LT) (Figure 4), the elongation of which leads to a -1 frameshift mutation (15, 18, 19). Parameters Which Modulate TLS at Single AAF Adducts. Results from elongation experiments presented here show that several parameters, i.e., the position of the lesion within the three-G sequence (Figure 2 and Table 1), the local sequence context, and the nature of the DNA polymerase, modulate TLS of AAF adducts in vitro. Position of the Lesion and Local Sequence Context. Normal bases assume mainly the anti configuration in DNA, but AAF modification of a guanine shifts the equilibrium in favor of a syn configuration (30-33), strongly precluding the pairing of the modified guanine with any incoming nucleotide. Experimentally, however, dCMP was found to be incorporated opposite the AAFmodified guanine. This may occur only when the adducted base transiently adopts the anti configuration (Figure 4). This conformationnal constraint is thought to account for the delay in nucleotide incorporation opposite the lesion (29). The effect of the position of the lesion within the three-G sequence on the incorporation efficiency at position L0, as observed for PolIKF(exo-) and the R subunit, could thus reflect a different synanti equilibrium for each adducted base, this equilibrium being modulated by the local sequence context. Indeed, a combined NMR/molecular mechanics approach has recently revealed that the interconversion between syn and anti configurations of the AF-modified guanines within a NarI site (5′-GGCGCC), another hot spot for AAF-induced frameshift mutagenesis, was dependent on

Sequence Context Modulation of Translesion Synthesis

the nearest neighbor sequences (34). In our case, the kinetics of disappearance of the L-1 band (Figure 3), which reflects the facility of incorporation at position L0, would suggest that the syn-anti equilibrium is differentially displaced in favor of the anti configuration in the following order: G1 > G2 > G3. This equilibrium can also affect the subsequent elongation step (L0 f L+1), performed by PolIKF(exo-). Once dCMP is incorporated opposite the lesion, the adducted base either may retain its anti configuration (Figure 4), thus allowing elongation from the LT to occur, or may adopt the syn configuration, which would preclude further elongation of the LT. In this latter case, given the proper local sequence context, a SMI may form, which upon elongation will yield a frameshift product (Figure 4). Sequencing data are in good agreement with this hypothesis as nonslipped TLS is observed only for the G1-AAF adduct, while G2- and G3-AAF lesions induce only slipped mutagenic TLS. Slipped TLS at the G1-AAF adduct can only occur if a dGMP residue is misincorporated opposite the lesion. This process is reminiscent of the misincorporation/ misalignment mechanism proposed for frameshift mutagenesis occurring at solitary guanines (35). As judged by the weak intensity of the 62-mer band, compared to that obtained for G2 and G3 templates (Figure 3), this is a rare event, in agreement with other data (36). This misincorporation event may also rarely occur opposite G2and G3-AAF adducts and may result in two-G and three-G deletions, respectively (Table 2). However, the relative contribution of this event to the overall mutagenesis decreases as the lesion is located on G1 (47%), G2 (23%), or G3 (3,8%), respectively (Table 2). Nature of the Polymerases. In vitro, TLS at AAF adducts depends on the 3′ f 5′ exonuclease proofreading activity of the DNA polymerase. In accordance with earlier reports (24, 29), we found that when proofreading proficient enzymes are used, a major stop band appears at position L-1, while reactions performed by exonuclease deficient enzymes yield bands at positions L-1 and L0. This reveals the ability of the polymerase to incorporate a nucleotide across from the adducted base, this nucleotide being excised by the proofreading exonuclease if present. Further elongation past the adduct is hindered [except with PolIKF(exo-)], suggesting that a limiting step in pursuing elongation is the incorporation of a nucleotide immediately after the lesion (L+1). Indeed, single-turnover kinetic experiments showed that incorporation by T7 DNA polymerase of a correct nucleotide opposite or one nucleotide after an AAF lesion is reduced about (4 × 106)- and more than (1 × 107)-fold, respectively, as compared to incorporation at the same positions on the unmodified template (29). Due to this delay in the elongation process, and because the rate of the 3′ f 5′ exonuclease reaction is much faster than the rate of incorporation (37), a proofreading proficient polymerase idles opposite the lesion, thus showing the single-band arrest (L-1) observed for PolIKF(exo+) and PolIII holoenzyme (Figure 1). Among the three exonuclease deficient polymerases we used, only PolIKF was able to perform TLS at AAF adducts. Previous studies in the absence of lesion have shown that PolIKF(exo-) preferentially extends mispair termini, leading to base substitutions, while the R subunit rather elongates misaligned termini, thus giving rise to frameshift mutations (35, 38). In our experiments, these enzymes behave differently; PolIKF(exo-) readily

Chem. Res. Toxicol., Vol. 12, No. 2, 1999 149

elongates a SMI, presumably because of the stabilizing effect of the AAF adduct (20), while the R subunit does not elongate the SMI. This subunit, which works usually as part of a greater assembly, is highly sensitive to secondary structures (35), and the presence of a bulky lesion close to a 3′ terminus may alter its stability or capability to elongate further. In vivo, single-adduct mutagenesis experiments have shown that the G3-AAF lesion induces 10- and 100-fold more -1 frameshift mutations than G2- and G1-AAF lesions, respectively (15). This contradicts our present results. However, a direct correlation between in vivo and in vitro data is limited because of the difference in the replicative systems. Specifically, the involvement in vivo under SOS-induced conditions of additionnal factors, such as the UmuD′2C complex, helps the polymerase to elongate from the lesion terminus (error-free TLS) (18). Indeed, TLS analysis at the G3-AAF lesion in vivo has shown that elongation from the LT, which is not observed in vitro for such an adduct, is the major route in wildtype E. coli cells, while it is totally abolished in strains defective for the UmuDC functions (17, 18). Similarly, in vivo, the Npf factor, which is known to strongly stimulate AAF-induced frame-shift mutagenesis under SOS-induced conditions (13, 18), increases the efficiency of slipped TLS (18, 19). In conclusion, efficient in vitro TLS at AAF-modified repetitive sequences results from the successful combination of at least two steps: (1) an efficient incorporation opposite the lesion, which is modulated both by the local sequence context and by the nature of the polymerase, and (2) the elongation from the resulting LT or from a SMI that can form in the proper local sequence context.

Acknowledgment. We are grateful to professor H. Maki for the gift of the R subunit of E. coli DNA polymerase III, to Dr. Bertrand-Burgraff for providing the E. coli DNA polymerase III holoenzyme, to Dr. M. Bichara for careful reading of the manuscript and helpful comments, and to Marc Nothisen for technical assitance.

References (1) Friedberg, E. C., Walker, G. C., and Siede, W. (1995) DNA repair and Mutagenesis, pp 191-232, American Society for Microbiology, Washington, DC. (2) Murli, S., and Walker, G. C. (1993) SOS mutagenesis. Curr. Opin. Genet. Dev. 3, 719-725. (3) Kato, T., and Shinoura, Y. (1977) Isolation and characterization of mutants of Escherichia coli deficient in induction of mutations by ultraviolet light. Mol. Gen. Genet. 156, 121-131. (4) Bridges, B. A., and Woodgate, R. (1985) The two step model of bacterial UV mutagenesis. Mutat. Res. 150, 133-139. (5) Nohmi, T., Battista, J. R., Dodson, L. A., and Walker, G. C. (1988) RecA-mediated cleavage activates UmuD for mutagenesis: mechanistic relationship between transcriptional derepression and posttranslational activation. Proc. Natl. Acad. Sci. U.S.A. 85, 1816-1820. (6) Shinagawa, H., Iwasaki, H., Kato, T., and Nakata, A. (1988) RecA protein-dependent cleavage of UmuD protein and SOS mutagenesis. Proc. Natl. Acad. Sci. U.S.A. 85, 1806-1810. (7) Burckhardt, S. E., Woodgate, R., Scheuermann, R. H., and Echols, H. (1988) UmuD mutagenesis protein of Escherichia coli: overproduction, purification, and cleavage by RecA. Proc. Natl. Acad. Sci. U.S.A. 85, 1811-1815. (8) Echols, H., and Goodman, M. F. (1991) Fidelity mechanisms in DNA replication. Annu. Rev. Biochem. 60, 477-511. (9) Hevroni, D., and Livneh, Z. (1988) Bypass and termination at apurinic sites during replication of single-stranded DNA in vitro: a model for apurinic site mutagenesis. Proc. Natl. Acad. Sci. U.S.A. 85, 5046-5050. (10) Rajagopalan, M., Lu, C., Woodgate, R., O’Donnell, M., Goodman, M. F., and Echols, H. (1992) Activity of the purified mutagenesis

150 Chem. Res. Toxicol., Vol. 12, No. 2, 1999

(11)

(12) (13)

(14)

(15) (16) (17)

(18) (19) (20)

(21) (22) (23)

(24)

(25)

proteins UmuC, UmuD′, and RecA in replicative bypass of an abasic DNA lesion by DNA polymerase III. Proc. Natl. Acad. Sci. U.S.A. 89, 10777-10781. Bichara, M., and Fuchs, R. P. (1985) DNA binding and mutation spectra of the carcinogen N-2-aminofluorene in Escherichia coli. A correlation between the conformation of the premutagenic lesion and the mutation specificity. J. Mol. Biol. 183, 341-351. Fuchs, R. P., Schwartz, N., and Daune, M. P. (1981) Hot spots of frameshift mutations induced by the ultimate carcinogen N-acetoxy-N-2-acetylaminofluorene. Nature 294, 657-659. Koffel-Schwartz, N., Verdier, J. M., Bichara, M., Freund, A. M., Daune, M. P., and Fuchs, R. P. P. (1984) Carcinogen-induced mutation spectrum in wild-type, uvrA and umuC strains of Escherichia coli. Strain specificity and mutation-prone sequences. J. Mol. Biol. 177, 33-51. Burnouf, D., Koehl, P., and Fuchs, R. P. (1989) Single adduct mutagenesis: strong effect of the position of a single acetylaminofluorene adduct within a mutation hot spot. Proc. Natl. Acad. Sci. U.S.A. 86, 4147-4151. Lambert, I. B., Napolitano, R. L., and Fuchs, R. P. (1992) Carcinogen-induced frameshift mutagenesis in repetitive sequences. Proc. Natl. Acad. Sci. U.S.A. 89, 1310-1314. Napolitano, R. L. (1994) DNA sequence determinants of carcinogeninduced frameshift mutagenesis. Biochemistry 33, 1311-1315. Koffel-Schwartz, N., Coin, F., Veaute, X., and Fuchs, R. P. P. (1996) Cellular strategies for accommodating replication-hindering adducts in DNA: Control by the SOS response in Escherichia coli. Proc. Natl. Acad. Sci. U.S.A. 93, 7805-7810. Napolitano, R. L., Lambert, I. B., and Fuchs, R. P. P. (1997) SOS factors involved in translesion synthesis. Proc. Natl. Acad. Sci. U.S.A. 94, 5733-5738. Fuchs, R. P. P., and Napolitano, R. L. (1998) Inactivation of DNA proofreading obviates the need for SOS induction in frameshift mutagenesis. Proc. Natl. Acad. Sci. U.S.A. 95, 13114-13119. Garcia, A., Lambert, I. B., and Fuchs, R. P. P. (1993) DNA adducts induced stabilization of slipped frameshift intermediates within repetitive sequences: implications for mutagenesis. Proc. Natl. Acad. Sci. U.S.A. 90, 5989-5993. Maenhaut-Michel, G., Janel-Bintz, R., and Fuchs, R. P. P. (1992) A umuDC-independent SOS pathway for frameshift mutagenesis. Mol. Gen. Genet. 235, 373-380. Maki, H., and Kornberg, A. (1985) The polymerase subunit of DNA polymerase III of Escherichia coli. J. Biol. Chem. 260, 6570-6578. Maki, H., Maki, S., and Kornberg, A. (1988) DNA polymerase III holoenzyme of Escherichia coli. IV. The holoenzyme is an asymmetric dimer with twin active sites. J. Biol. Chem. 263, 65706578. Belguise-Valladier, P., Maki, H., Sekiguchi, M., and Fuchs, R. P. (1994) Effect of single DNA lesions on in vitro replication with DNA polymerase III holoenzyme. Comparison with other polymerases. J. Mol. Biol. 236, 151-164. Belguise, V. P., and Fuchs, R. P. (1991) Strong sequencedependent polymorphism in adduct-induced DNA structure: analysis of single N-2-acetylaminofluorene residues bound within the NarI mutation hot spot. Biochemistry 30, 10091-10100.

Burnouf et al. (26) Seeberg, E., and Fuchs, R. P. (1990) Acetylaminofluorene bound to different guanines of the sequence -GGCGCC- is excised with different efficiencies by the UvrABC excision nuclease in a pattern not correlated to the potency of mutation induction. Proc. Natl. Acad. Sci. U.S.A. 87, 191-194. (27) Koehl, P., Burnouf, D., and Fuchs, R. P. (1989) Construction of plasmids containing a unique acetylaminofluorene adduct located within a mutation hot spot. A new probe for frameshift mutagenesis. J. Mol. Biol. 207, 355-364. (28) Livneh, Z., Cohen-Fix, O., Skaliter, R., and Elizur, T. (1993) Crit. Rev. Biochem. Mol. Biol. 28, 465-513. (29) Lindsley, J. E., and Fuchs, R. P. (1994) Use of single-turnover kinetics to study bulky adduct bypass by T7 DNA polymerase. Biochemistry 33, 764-772. (30) Leng, M., Ptak, M., and Rio, P. (1980) Conformation of acetylaminofluorene and aminofluorene modified guanosine and guanosine derivatives. Biochem. Biophys. Res. Commun. 96, 10951102. (31) Lipkowitz, K. B., Chevalier, T., Widdifield, M., and Beland, F. A. (1982) Force field conformational analysis of aminofluorene and acetylaminofluorene substituted deoxyguanosine. Chem- Biol. Interact. 40, 57-76. (32) O’Handley, S. F., Sanford, D. G., Xu, R., Lester, C. C., Hingerty, B. E., Broyde, S., and Krugh, T. R. (1993) Structural characterization of an N-acetyl-2-aminofluorene (AAF) modified DNA oligomer by NMR, energy minimization, and molecular dynamics. Biochemistry 32, 2481-2497. (33) Patel, D., Mao, B., Gu, Z., Hingerty, B., Gorin, A., Basu, A., and Broyde, S. (1998) Nuclear Magnetic Resnance solution structures of covalent aromatic amine-DNA adducts and their mutagenic relevance. Chem. Res. Toxicol. 11, 391-407. (34) Mao, B., Hingerty, B. E., Broyde, S., and Patel, D. J. (1998) Solution structure of the Aminofluorene (AF)-external conformer of the anti-(AF)-C8-dG adduct opposite dC in a DNA duplex. Biochemistry 37, 95-106. (35) Mo, J. Y., and Schaaper, R. M. (1996) Fidelity and error specificity of the R catalytic subunit of Escherichia coli DNA polymerase III. J. Biol. Chem. 271, 18947-18953. (36) Shibutani, S., and Grollman, A. P. (1993) On the mechanism of frameshift (deletion) mutagenesis in vitro. J. Biol. Chem. 268, 11703-11710. (37) Wong, I., Patel, S. S., and Johnson, K. A. (1991) An induced-fit kinetic mechanism for DNA replication fidelity: direct measurement by single-turnover kinetics. Biochemistry 30, 526-537. (38) Bebenek, K., Joyce, C. M., Fitzgerald, M. P., and Kunkel, T. A. (1990) The fidelity of DNA synthesis catalyzed by derivatives of Escherichia coli DNA polymerase I. J. Biol. Chem. 265, 1387813887. (39) Fuchs, R. P. P. (1984) DNA binding spectrum of the carcinogen N-acetoxy-N-2-acetylaminofluorene significantly differs from the mutation spectrum. J. Mol. Biol. 177, 173-180.

TX9801920