AUGUST 2002 VOLUME 15, NUMBER 8 © Copyright 2002 by the American Chemical Society
Articles Substrate Recognition by a Family of Uracil-DNA Glycosylases: UNG, MUG, and TDG Pingfang Liu,†,‡ Artur Burdzy,† and Lawrence C. Sowers*,†,‡ Department of Biochemistry and Microbiology, School of Medicine, Loma Linda University, Loma Linda, California 92350, and Graduate School of Biological Sciences, City of Hope National Medical Center, 1500 East Duarte Road, Duarte, California 91010 Received April 18, 2002
In response to continuous hydrolytic and oxidative DNA damage, cells of all organisms have a complex network of repair systems that recognize, remove, and rebuild the injured sites. Damaged pyrimidines are generally removed by glycosylases that must scan the entire genome to locate lesions with sufficient fidelity to selectively remove the damage without inadvertent removal of normal bases. We report here studies conducted with a series of base analogues designed to test mechanisms of base recognition suggested by structural studies of glycosylase complexes. The oligonucleotide series examined here includes 5-halouracils with increasing substituent size and purine analogues placed opposite the target uracil with hydrogen, amino, and keto substituents in the 2- and 6-positions. The glycosylases studied here include Escherichia coli uracil-DNA glycosylase (UNG), E. coli mismatch uracil-DNA glycosylase (MUG), and the Methanobacterium thermoautotrophicum mismatch thymine-DNA glycosylase (TDG). The results of this study suggest that these glycosylases utilize several strategies for base identification, including (1) steric limitations on the size of the 5-substituent, (2) electronicinductive properties of the 5-substituent, (3) reduced thermal stability of mispairs, and (4) specific functional groups on the purine base in the opposing strand. Contrary to predictions based upon the crystal structure, the preference of MUG for mispaired uracil over thymine is not based upon steric exclusion. Furthermore, the preference for mispaired uracil over uracil paired with adenine is more likely due to reduced thermal stability as opposed to specific recognition of the mispaired guanine. On the other hand, TDG, which exhibits modest discrimination among various pyrimidines, shows strong interactions with functional groups present on the purine opposite the target pyrimidine. These results provide new insights into the mechanisms of base selection by DNA repair glycosylases.
Introduction Endogenous processes including oxidation, hydrolysis, and alkylation constantly damage the DNA of all organ* To whom correspondence should be addressed. Tel: (909) 5584480; FAX: (909) 558-4035; E-mail:
[email protected]. † Loma Linda University. ‡ City of Hope National Medical Center.
isms (1). The hydrolytic deamination of cytosine to uracil is one of the most frequent forms of endogenous DNA damage. If unrepaired, uracil derived from cytosine results in a transition mutation. Uracil residues can also be incorporated into DNA in place of thymine if dUTP levels rise because of folate deficiency or antimetabolite chemotherapy (1, 2). Replacement of thymine with uracil
10.1021/tx020030a CCC: $22.00 © 2002 American Chemical Society Published on Web 07/25/2002
1002
Chem. Res. Toxicol., Vol. 15, No. 8, 2002
in DNA interferes with the interactions of sequencespecific DNA-binding proteins, including transcription factors (3, 4). The hydrolytic deamination of 5-methylcytosine (5mC) to thymine generates a thymine mispaired with guanine (5). Mutations resulting from damage to 5mC comprise one of the most frequent classes of point mutation found in human cancer cells (6). In response to this damage, cells have a complex network of repair systems that recognize and remove the lesions. Damaged pyrimidines are generally removed by glycosylases that cleave the bond between the sugar and the N1 position of the damaged base (2, 7-10). In the human genome, five different glycosylases have been identified that have uracil-DNA glycosylase activity, and many of these share substantial homology with uracilDNA glycosylases found in other organisms (9-36). Some activities will remove uracil residues in single-stranded DNA as well as uracil in duplex DNA paired with either adenine or guanine (11-13, 21). Others, however, are more selective and are reported to remove uracil only when mispaired with guanine (22-25). Some of the glycosylases are highly selective for uracil (11-13, 2125) whereas others will also remove thymine mispaired with guanine (14-20, 26-28). The repair of endogenous lesions presents a formidable challenge for organisms. It is estimated that the total number of endogenous lesions in the human genome is on the order of tens of thousands per cell per day (2). While substantial in number, these lesions are distributed among hundreds of thousands of normal bases. Therefore, the fidelity of the glycosylases must be extremely high, perhaps exceeding the fidelity of DNA polymerases (7). The glycosylases must discriminate damaged from normal bases on the basis of subtle structural or contextual differences (10-36). Otherwise, the glycosylases themselves would promote genetic instability. The basis for the selectivity of these glycosylases is currently an active area of research in several laboratories. The structures of some glycosylases have been studied by X-ray crystallography (29-35). Inferences have been made on the mechanisms by which the glycosylases select lesions based upon critical contact points identified in these structures (29, 34, 35). To test predictions resulting from these structures, we have prepared a series of oligonucleotide substrates containing a series of 5-substituted uracil derivatives paired with a series of modified purines by rearranging the functional groups accounting for these contact points. We have selected a group of bacterial uracil-DNA glycosylases as model systems to investigate the basis for substrate selectivity, including the uracil-DNA glycosylase UNG1 from E. coli that excises uracil from single- or doublestranded DNA (21), the mismatch uracil-DNA glycosylase MUG from E. coli that removes uracil from uracilguanine mispairs (22-25), and the thermostable mismatch thymine-DNA glycosylase TDG from Methanobacterium thermoautotrophicum that removes both uracil 1 Abbreviations: UNG, uracil-DNA glycosylase; MUG, E. coli mismatch uracil-DNA glycosylase; TDG, a thermostable mismatch thymine-DNA glycosylase from Methanobacterium thermoautotrophicum THF; hTDG, human mismatch thymine-DNA glycosylase; A, adenine; G, guanine; T, thymine; U, uracil; 2AP, 2-aminopurine; I, hypoxanthine (deoxyinosine); AA, 2,6-diaminopurine; N, nebularine (purine); FU, 5-fluorouracil; ClU, 5-chlorouracil; BrU, 5-bromouracil; IU, 5-iodouracil; HmU, 5-hydroxymethyluracil; GC/MS, gas chromatography/mass spectrometry.
Liu et al. Table 1. Analytical GC/MS Data of Bases Used in the Studya base
retention time (min)
predominant ion (m/z)
A T C G 2AP I AA N U FU CIU BrU IU
11.322 6.045 7.658 13.059 12.019 10.801 12.032 10.765 5.158 5.051 6.816 7.728 8.863
264 255 254 352 264 265/280 366/351 191 241 259 275 319/321 367
a The oligonucleotides were acid-hydrolyzed, subsequently silylated, and analyzed by GC/MS as previously described (37). Bases are identified by their characteristic retention times and mass spectra.
and thymine mispaired with guanine (26-28). They all have corresponding homologues present in human cells, which are hUNG, hTDG, and MBD4, respectively (911, 14-16, 20, 22-28, 32, 33). The results of this study in many cases contrast with expectations based upon the crystal structures. It is suggested that several mechanisms are important for lesion selection by the glycosylases, including steric fit, thermal stability, glycosidic bond strength, and functional groups located on the purine residue opposite the target pyrimidine. These findings provide new insights into the mechanisms by which damaged pyrimidines may be identified and repaired in DNA.
Materials and Methods Materials. Mismatch uracil-DNA glycosylase (MUG) and thermostable thymine-DNA glycosylase (TDG) were purchased from Trevigen Inc. (Gaithersburg, MD). Uracil-DNA glycosylase (UNG) was from U.S.B. Corp. (Cleveland, OH). Bacteriophage T4 polynucleotide kinase was obtained from New England BioLabs (Beverley, MA). Adenosine 5′-[γ-32P]triphosphate ([γ32P]ATP) was from ICN Life Sciences (Costa Mesa, CA). Sephadex G-50 Quick Spin Columns were purchased from Boehringer Mannheim (Indianapolis, IN). C-18 sep-pak cartridges were from Waters Associates (Milford, MA). Oligonucleotide Synthesis and Characterization. Oligonucleotides were prepared by solid-phase synthesis methods as described previously (36). The phosphoramidites for uracil, fluorouracil, bromouracil, iodouracil, nebularine, hypoxanthine, and 2,6-diaminopurine were from Glen Research (Sterling, VA). The phosphoramidites for 2-aminopurine (37) and 5-chlorouracil (38) were prepared as previously described. Following synthesis and deprotection, oligonucleotides were purified with C-18 seppak cartridges and characterized by GC/MS following acid hydrolysis and conversion to the trimethylsilyl ethers (37). The retention time and m/z of predominant ions from mass spectra for each base are reported in Table 1. Oligonucleotide Labeling and Annealing Procedures. 5′-End radiolabeling was performed using [γ-32P]ATP and T4 polynucleotide kinase under conditions recommended by the enzyme supplier. Labeled mixtures were subsequently centrifuged through a G-50 Sephadex column to remove excess unincorporated nucleotide. Labeled single-stranded oligonucleotides were annealed to a 2-fold molar excess of unlabeled complementary strand in 5 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES)-KOH, pH 7.4, 1 mM NaCl, 0.1 mM EDTA, and 0.1 mM DTT. Annealing mixtures were heated to 95 °C for 5 min and then cooled slowly to room temperature.
Substrate Recognition by Uracil-DNA Glycosylases
Chem. Res. Toxicol., Vol. 15, No. 8, 2002 1003
Figure 1. Oligonucleotide sequences used in this study. (A) 24mer used in the study, where X ) U, T, 5-fluorouracil (FU), 5-chlorouracil (ClU), 5-bromouracil (BrU), or 5-iodouracil (IU), and P ) A, G, 2-aminopurine (2AP), hypoxanthine (I), 2,6diaminopurine (AA), or nebularine (N). (B) 30mer used in MUG experiments under different pH conditions, which is 6 nucleotides longer at the 5′ terminal than the 24mer. Enzymatic Reactions. DNA substrates (4 pmol/reaction) were incubated with 1 µL of UNG (1 unit/µL)2 for 1 min at 37 °C in 5 mM HEPES-KOH, pH 7.4, 1 mM NaCl, 0.1 mM EDTA, and 0.1 mM DTT. Reactions were stopped by adding 5 µL of 0.1 M NaOH and an equal volume of Maxam-Gilbert loading buffer (98% formamide, 0.01 M EDTA, 1 mg/mL xylene cyanol, and 1 mg/mL bromophenol blue). The apyrimidinic sites were cleaved with NaOH by heating at 95 °C for 30 min. MUG reactions were performed with 4 pmol of oligonucleotides, 1 µL of enzyme (1 unit/µL)3 in 20 mM Tris-HCl, pH 8.0, 0.1 mg/mL BSA, 1 mM EDTA, 1 mM EGTA, and 1 mM DTT for 20 min at 37 °C. Similarly, TDG reactions were conducted with 2.5 pmol of oligonucleotides, 5 µL of enzyme (1 unit/µL)4 in 10 mM HEPESKOH, pH 7.4, 100 mM KCl, and 10 mM EDTA for 2 h at 37 °C. Cleavage of abasic sites after glycosylase treatment was performed as described above. Enzymatic activities were also measured as a function of time, and the initial velocities were determined in the linear cleavage region. For experimental ease, the apyrimidinic sites after glycosylase action were cleaved with 1 M piperidine for 15 min at 37 °C for the kinetic studies. MUG activity against BrU:G mispairs was measured at pH 4.9, pH 8.0, and pH 8.9. The reaction volume was 10 µL containing 2 pmol of BrU:G 24mer and 2 pmol of U:G 30mer. Oligonucleotides were incubated with 1 µL of MUG (1 unit/µL) for 10 min at 37 °C. Cleavage of the abasic sites resulting from glycosylase removal of the target base was performed with NaOH by heating at 95 °C for 30 min. The buffers used for this assay were 20 mM Tris-HCl, pH 8.0 (or pH 8.9), 0.1 mg/mL BSA, 1 mM EDTA, 1 mM EGTA, and 1 mM DTT. For pH 4.9 buffer, we used 20 mM MES instead of Tris-HCl, and all other conditions were kept constant. Gel Electrophoresis and Analysis of Cleavage. Reaction samples were electrophoresed on 18% denaturing polyacrylamide gels (8 M urea), and the bands corresponding to substrate and product were quantified using a Molecular Dynamics PhosphorImager (Molecular Dynamics, Sunnyvale, CA).
Results Base Analogues Used in This Study. The oligonucleotide sequences used for these studies are shown in Figure 1. The 24mer and 30mer are homologous, except that the 30mer is 6 nucleotides longer at the 5′ end. Two sets of base analogues were introduced into the assay, corresponding to the X and P bases indicated in Figure 1. The X series corresponds to a series of 5-substituted uracil in which the substituent size increases in the series H < F < Cl < Br < I. The P series corresponds 2 One unit of uracil-DNA glycosylase is defined as the amount of enzyme required to release 1 nmol of [3H]uracil under standard conditions. 3 One unit of mismatch uracil-DNA glycosylase is defined as the amount that cleaves 1 pmol of a 32P-labeled oligonucleotide probe containing 3,N4-ethenocytosine within an oligonucleotide duplex in 1 h at 37 °C. 4 One unit of TDG is defined as the amount that cleaves 1 pmol of an oligonucleotide duplex containing a T/G mismatch in 1 h at 65 °C.
Figure 2. Effect of uracil 5-substituents on glycosylase cleavage. The sequence of the duplex is indicated in Figure 1A where X ) uracil 5-substituents and P ) G. The identity of the base pair at the cleavage site is indicated at the top of the figure. (A) UNG assays. The reactions were performed at 37 °C for 20 min with 1 unit of UNG. (B) MUG assays. The reactions were performed at 37 °C for 20 min with 1 unit of MUG. (C) TDG assays. The reactions were performed at 37 °C for 2 h with 5 units of TDG. The apyrimidinic sites were cleaved with NaOH by heating at 95 °C for 30 min.
to a series of purines substituted in the 2- and 6-positions with H, an amino group, or a keto oxygen, which are the proposed critical contact points for glycosylase selection. The base composition of the oligonucleotides prepared for this study was confirmed by GC/MS. Analytical data are reported in Table 1. Effects of Uracil 5-Substituents on Glycosylase Cleavage. The results of glycosylase activity against the 5-substituted uracil derivatives are shown in Figure 2, and the kinetics of excision are present in Figure 3. Glycosylase removal of the target pyrimidine in the 24base oligonucleotide generates an abasic site that is subsequently cleaved under alkaline conditions, resulting in generation of the 12-base oligonucleotide. UNG has the most limited substrate range (Figure 2A). Only the H- and F-substituted uracils are removed by UNG, and uracil is a better substrate than the F-substituted one (Figure 3A, Table 2). The larger Cl-, CH3-, Br-, and I-substituted uracils are not substrates. The substrate specificity of the mispaired uracil-DNA glycosylase, MUG, is both distinct from UNG and highly surprising (Figure 2B). The H-, F-, and Cl-substituted uracils are removed at comparable rates (Figure 3B, Table 2), compatible with a slightly larger pyrimidine-binding pocket relative to UNG. Thymine (U-methyl) is not a substrate, consistent with previous reports (22-25). However, the larger bromouracil and iodouracil are substrates. The mispaired thymine-DNA glycosylase, TDG, is observed to accept all of the 5-substituted uracils, but with preference for 5-halogenated uracils (Figure 3C, Table 2). Effects of Solution pH on Cleavage of the 5-Substituted Uracils. While similar in size, the inductive properties of the 5-methyl and 5-bromo substituents of thymine and 5-bromouracil are opposite. The difference in the electronic-inductive properties results in a substantial difference in pK values for the N3 proton, with 5-bromouracil being substantially ionized at physiological
1004
Chem. Res. Toxicol., Vol. 15, No. 8, 2002
Liu et al.
Figure 4. MUG activities against BrU:G mispairs under different pH conditions. The sequences of the duplexes are indicated in Figure 1. The 12mer is the cleaved product from the BrU:G 24mer, and the 18mer is the cleaved product from the U:G 30mer. To test the possible effects of base ionization on MUG activity, the glycosylase reaction was conducted over a range of pH 4.9-8.9 that would substantially alter the fraction of ionized 5-bromouracil residues. Since the fraction of the ionized form is independent of the incubation time, this assay was performed at 37 °C for a single time point (10 min). The apyrimidinic sites were cleaved with NaOH by heating at 95 °C for 30 min. For details, see Materials and Methods.
Figure 3. Kinetic study of UNG, MUG, and TDG cleavage of 5-substituted uracil analogues paired with guanine. The reactions were performed with UNG, MUG, and TDG at 37 °C for different periods of time, and at each time point, an equal volume of 98% formamide dye was added to stop the reactions. The apyrimidinic sites after glycosylase action were cleaved with 1 M piperidine for 15 min at 37 °C. For details, see Materials and Methods. Table 2. Initial Velocities of Cloned UNG, MUG, and Thermostable TDG for Uracil 5-Substituents Paired with Guanine [pmol of Oligonucleotide Cleaved (µg of Protein)-1 min-1] U:G FU:G CIU:G BrU:G T:G IU:G
UNG
MUG
TDG
1.4 × 104 1.5 × 103
