Chem. Res. Toxicol. 1996, 9, 1167-1175
1167
Comparison of the Efficiency of Synthesis Past Single Bulky DNA Adducts in Vivo and in Vitro by the Polymerase III Holoenzyme Gary J. Latham,†,‡,§ Andrew G. McNees,‡ Bart De Corte,† Constance M. Harris,† Thomas M. Harris,† Mike O’Donnell,| and R. Stephen Lloyd*,‡ Center in Molecular Toxicology and Departments of Biochemistry and Chemistry, Vanderbilt University, Nashville, Tennessee 37232, Microbiology Department, Hearst Research Foundation, and Howard Hughes Medical Institute, Cornell University Medical College, New York, New York 10021, and Sealy Center for Molecular Science, The University of Texas Medical Branch, Galveston, Texas 77550 Received March 26, 1996X
Previous studies from our laboratory revealed that site-specific and stereospecific styrene oxide (SO) lesions in M13 DNA were readily bypassed when transfected into Escherichia coli cells, but these same lesions blocked the progress of several purified polymerases in vitro when situated in oligodeoxynucleotide templates (Latham, G. J., et al. (1993) J. Biol. Chem. 268, 23427-23434; Latham, G. J., et al. (1995) Chem. Res. Toxicol. 8, 422-430). To resolve this apparent discrepancy, we constructed single-stranded M13 genomes containing single SO adducts and compared their replication efficiencies in E. coli cells to the extent of bypass synthesis in vitro using three different complexes of the purified E. coli polymerase III (Pol III) holoenzyme. The transformation efficiencies of the SO-adducted M13 templates were comparable to those of the nonadducted controls, indicating facile bypass in E. coli. When the identical adducted M13 vectors were replicated in vitro with the reconstituted complexes of the Pol III holoenzyme, the results were consistent with the in vivo data: Synthesis past two of the three SO adducts in M13 was unhindered relative to synthesis on the unadducted M13 control template. Since our previous in vitro assays indicated that SO adducts in 33-mer templates largely blocked polymerases other than Pol III, we repeated these studies using reconstituted Pol III. Significantly, Pol III replication was poorly processive and strongly terminated by SO lesions in 33-mer templates. This result was in stark contrast to the efficient bypass in vitro of the same adducts in M13 DNA. In fact, Pol III-mediated bypass was enhanced to >75-fold on adducted circular M13 templates as compared to adducted linear oligodeoxynucleotides. The implications of the effects of polymerase processivity and template‚primer structure upon lesion bypass are discussed.
Introduction III)1
The polymerase III (Pol holoenzyme (HE) is a multisubunit protein complex responsible for chromosomal replication in Escherichia coli (1). The interaction of the Pol III accessory proteins with DNA confers both a high rate of synthesis (750 nucleotides/s) and a remarkable processivity (>100 kb per binding event) to the core polymerase, resulting in an efficient protein machine (2). Although recent studies probing the assembly of the Pol III HE indicate that more than one highly processive complex can be reconstituted from the available protein components (3), the best studied HE contains Pol III* (4). Pol III* is a moderately processive polymerase comprised of nine separate subunits (R22θ2τ2γ2δ1δ′1χ1ψ1) including * To whom correspondence should be addressed. † Vanderbilt University. ‡ The University of Texas Medical Branch. § Present address: Institute of Molecular Biology, University of Oregon, Eugene, OR 97403-1229. | Cornell University Medical College. X Abstract published in Advance ACS Abstracts, September 15, 1996. 1 Abbreviations: Pol III, Escherichia coli polymerase III; HE, holoenzyme; Pol III*, R22θ2τ2γ2δ1δ′1χ1ψ1; Pol III-γθ, Pol III* lacking subunits γ and θ, subunit stoichiometry R22τ4δ1δ′1χ1ψ1; Pol III-γθ, Pol III* lacking subunits γ, θ, and , subunit stoichiometry R2τ4δ1δ′1χ1ψ1; SO, styrene oxide; ssb, single-stranded binding protein; ssDNA, singlestranded DNA; HPLC, high performance liquid chromatography; NMR, nuclear magnetic resonance; EDTA, ethylenediaminetetraacetate; TAE, Tris-acetate, ethylenediaminetetraacetate.
S0893-228x(96)00055-0 CCC: $12.00
the core polymerase (Rθ) and the ATPase-containing γ complex (γ2δ1δ′1χ1ψ1) (4, 5). Assembly of the processive Pol III HE2 depends on the “clamp loading” activity of the γ complex, which uses the energy from ATP hydrolysis to mount the doughnut-shaped β protein onto circular DNA. The τ subunit, which can link two HEs together for asymmetric DNA synthesis (6), also has ATPase activity. This is not surprising when one considers that τ and γ are translational variants from the same gene, dnaX (7-9). Significantly, a τ complex formed with δδ′χψ can load β dimers onto DNA to form a processive HE in the absence of γ (3). Thus, Pol III HE molecules can be reconstituted using the clamp loading activity of γ or τ complexes, depending on the protein levels available for assembly. Indeed, it is currently unclear whether Pol III HE containing an integrated clamp loader activity in an E. coli. cell uses τ or γ for this function (10, 11). The possibility that other “forms” of Pol III HE participate in DNA repair, recombination, or mutagenesis has been considered (3). However, the consequences of DNA damage on the replicative characteristics of the Pol III HE subassemblies have not been investigated. Thus, one of the goals of this study was to characterize the ability of different complexes of Pol III HE to bypass site-specific 2 Here, the term “HE” refers to polymerases that use a DNA sliding clamp.
© 1996 American Chemical Society
1168 Chem. Res. Toxicol., Vol. 9, No. 7, 1996
bulky DNA adducts and lend insight into the cellular roles of these complexes in E. coli. The advent of strategies for the chemical synthesis of oligodeoxynucleotides containing single DNA adducts (12, 13) has revolutionized the study of how DNA lesions are processed by cellular enzymes. We have modified these methodologies to allow both the site-specific and stereospecific placement of bulky DNA adducts within essentially any sequence (14). To understand the influence of sequence context, adduct chirality, and polymerase specificity upon adduct processing, we previously described the replication fate of DNA modified by styrene oxide (SO) both in vivo and in vitro (15-18). As one of the most heavily produced industrial organic chemicals, styrene is a possible carcinogen (19) and a serious health concern to exposed workers in certain occupational settings (20). Importantly, the major DNA-reactive metabolite of styrene, SO, is mutagenic in several test systems (21). In addition, SO adducts are an excellent model for general bulky modifications on DNA since styryl modifications can be readily manipulated in an array of DNA sequences by site-specific techniques (14). Thus, we chose to study the processing of SO lesions in pursuit of the molecular details of how DNA damage blocks DNA polymerization, is faithfully replicated, or is converted into mutations. During the course of this work, we observed an apparent paradox: Although all of the SO lesions examined (when inserted into M13 vectors) were bypassed efficiently in E. coli cells (15, 18), many of these adducts (when localized to 33-mer or 63mer templates) severely inhibited the bypass of several different purified polymerases (e.g., Klenow fragment, Sequenase 2.0, polymerase R, polymerase β, T4 polymerase holoenzyme, and others) in vitro (15-18). To seek a resolution to this discrepancy between in vivo and in vitro lesion bypass, we have compared the ability of the reconstituted Pol III HE, the polymerase responsible for M13 replication in E. coli, to bypass several SO adducts in M13 DNA in vitro to the efficiency of lesion bypass effected by the cellular Pol III HE (as measured by the plaque-forming abilities of the identical adducted M13 constructs in repair-deficient E. coli cells). In contrast to the polymerases tested previously (15, 17, 18), in vitro Pol III HE-mediated synthesis past two of the three SO adducts in M13 was highly efficient, consistent with the in vivo data suggesting facile bypass. Furthermore, the extent of translesion synthesis was 3- to >75fold greater for the processive Pol III HE than for the nonprocessive Pol III complex when comparing polymerase activities on adducted circular M13 templates and adducted linear oligodeoxynucleotides. These results underscore the importance of reconstituting the in vivo cellular machinery with the relevant DNA substrate when investigating adduct processing in vitro.
Materials and Methods Synthesis, Purification, and Characterization of SOAdducted 11-mers. The chemical synthesis and purification of the SO-containing oligodeoxynucleotides have been described previously (14, 15). The adducted 11-mers were characterized by capillary gel electrophoresis, enzymatic digestion (snake venom phosphodiesterase and alkaline phosphatase) of the oligomers and HPLC analysis of resulting nucleotides, PAGE analysis of 32P end-labeled oligomers, and/or 1H NMR. In each case, the oligomers were essentially homogeneous and contained a single adduct consistent with modification by SO. Construction of Unmodified and SO-Modified N-rasM13mp7L2. Single-stranded M13mp7L2 DNAs containing
Latham et al. site-specific and stereospecific SO adducts were prepared as described (15). Briefly, M13mp7L2 DNA was digested with EcoRI, purified, and ligated with SO-modified 11-mers in the presence of a 51-mer scaffold. To eliminate the possibility that a trace contaminant of the adducted oligodeoxynucleotide synthesis might preferentially ligate into M13 when a large excess of the 11-mer was used, the molar ratio of 11-mer:51mer:M13 was limited to 5:2:1. Since the ligation of 11-mer into M13 to form a covalently-closed circular molecule was 99% blocked in the immediate vicinity of the damaged base (18), suggesting that >99% of the DNA templates were modified. (2) PAGE, capillary gel electrophoresis, 1H 1NMR, and HPLC analyses of adducted oligomers indicated that a minimum of 95% of the 11-mer substrates contained an SO-DNA lesion (see above). (3) SO adducts were stable under the conditions used to assay replication competence both in vivo and in vitro, as measured by several criteria (18). (4) Differential hybridization studies demonstrated that ∼95% of the ssM13 molecules contained an N-ras insert and therefore likewise contained an SO lesion after ligation with the substrate 11-mer (see below). Taken together, these findings strongly argue that the construction of adducted ssM13 produced templates that were nearly homogeneously modified with single SO lesions. Plaque-Forming Ability and Mutagenesis of N-rasM13mp7L2 Constructs Transfected into AB2480 Cells. The transformation procedures and methods for screening mutant plaques formed by SO-adducted M13mp7L2 constructs are described elsewhere (15). Three different preparations of the ligated M13 products were used to transform repair-deficient AB2480 (recA- uvrA-) cells. (1) The unpurified ligation mixture, which contains primarily covalently-closed circular M13 with ligated N-ras inserts and linear (unligated) M13mp7L2, along with slight amounts of linear, singly-ligated N-ras-M13mp7L2
Bypass of Bulky DNA Adducts in Vivo and in Vitro
Chem. Res. Toxicol., Vol. 9, No. 7, 1996 1169
Table 1. SO-Adducted M13 Is Nonblocking and Poorly Mutagenic in Transformed Repair-Deficient AB2480 Cellsa transformation with: N-ras 61 R(61,2)-SO
PFA/input DNAb
ligation efficiency (%)
(1) Crude Ligation 470 100 350 75
adjust PFA/ input DNAb 470 470
N-ras 12 @(12,2) @(11,3)
1300 1100
100 100
1300 1100
R(12,2)-SO R(11,3)-SO
1200 1600
65 ND
1800 >1600
MF
