ABSTRACT
5-Formyluracil (5-foU) is a major lesion of thymine produced in DNA by ionizing radiation and various chemical oxidants. To assess its biochemical effects on DNA replication, 22mer oligonucleotide templates containing an internal 5-foU at defined sites were synthesized by the phosphoramidite method and examined for ability to serve as a template for various DNA polymerases in vitro. Klenow fragments with and without 3' -> 5' exonuclease of DNA polymerase I, Thermus thermophilus DNA polymerase (exonuclease-deficient) and Pyrococcus furiosus DNA polymerase (exonuclease-proficient) read through the site of 5-foU in the template. Primer extension assays revealed that the 5-foU directed not only incorporation of dAMP but also dCMP opposite the lesion during DNA synthesis. Misincorporation opposite 5-foU was unaffected by 3' -> 5' exonuclease activity. DNA polymerases had different dissociation rates from a dCMP/T mispair and from a dCMP/5-foU mispair. The incorporation of an `incorrect' nucleotide was dependent on the sequence context and DNA polymerase used. These results suggest that 5-foU produced in DNA has mutagenic potential leading to T -> G transversions during DNA synthesis.
Active oxygen species are generated in living cells by normal metabolism and by exogenous sources such as ionizing radiation and various chemical oxidants (1 -3 ). They modify the base and sugar moieties in DNA (4 -7 ). Thymine glycols, 8-hydroxyguanine and 5-hydroxypyrimidines are formed in DNA by exposure to such oxidizing agents (5 -8 ). 5-Formyluracil (5-foU) is a novel type of oxidatively modified thymine in DNA (9 ,10 ). The methyl group of thymine is vulnerable to hydroxyl radical attack and it produces 5-hydroperoxymethyluracil, which is spontaneously decomposed to form 5-hydroxymethyluracil and 5-foU (4 ,11 ). 5-FoU is formed in a yield comparable with that of thymine glycols and 8-hydroxyguanine by ionizing radiation (4 ,9 ,12 ) and quinone-sensitized UVA photooxidation (13 ,14 ). Bjelland et al. (15 ,16 ) and Zhang et al. (12 ) have shown that Escherichia coli and mammalian cells have DNA glycosylase activity that removes 5-foU from DNA exposed to ionizing radiation. 5-Formyl-2'-deoxyuridine is mutagenic to Salmonella TA102 when added to the culture medium (9 ). However, there are no direct indications regarding the biological consequences of 5-foU formed in DNA.
Recent developments in the chemical synthesis of oligonucleotides have allowed modified bases to be introduced into the oligonucleotide templates at defined sites (17 -21 ). This is a useful means of predicting their lethal and mutagenic consequences based on the effect on DNA synthesis in vitro. In this study we synthesized oligonucleotide templates with one 5-foU at various sites using phosphoramidite chemistry (22 ) and examined their interaction with various DNA polymerases in vitro. The present experiments demonstrate that 5-foU directed not only incorporation of dAMP but also dCMP during DNA synthesis, suggesting that it is a potent mutagenic lesion leading to T -> G transversions.
T4 polynucleotide kinase was purchased from TOYOBO Co. Klenow fragments of DNA polymerase I with and without 3' -> 5' exonuclease (KF+ and KF-, respectively) (23 ) were from TOYOBO Co. and Ambion Inc., respectively. Thermus thermophilus (Tth) DNA polymerase came from Epicentre Technologies Corp. and Pyrococcus furiosus (Pfu)DNA polymerase from Stratagene. Four normal HPLC-grade 2'-deoxyribonucleotide 5'-triphosphates (dNTPs) were purchased from Takara Shuzo. [[gamma]-32P]ATP (>259 TBq/mmol) was the product of ICN Biomedicals Inc. Dithiothreitol (DTT) and nuclease-free BSA were from Wako Pure Chemicals.
Oligonucleotides containing an internal 5-foU at desired sites were synthesized as described by Sugiyama et al. (22 ). In brief, oligonucleotides containing 5-(1,2-dihydroxyethyl)uracil were constructed by the phosphoramidite chemistry using an ABI 381A DNA synthesizer. Saturated NaIO4 was added to a solution of oligonucleotides containing the precursor of 5-foU and the reaction mixture was vortex-mixed for 1 min at room temperature. The oligonucleotides containing 5-foU were purified by HPLC (22 ). The synthesized oligonucleotides are shown in Figure 1 .
Oligonucleotide primers (100 pmol) (Fig. 1 ) were labeled at the 5'-end by T4 polynucleotide kinase (4 U) in the presence of [[gamma]-32P]ATP in 70 mM Tris-HCl pH 7.6, 10 mM MgCl2, 5 mM DTT in a total volume of 20 µl. After an incubation at 37oC for 60 min, the reaction was stopped by adding 3 µl of 0.25 M EDTA and non-incorporated [[gamma]-32P]ATP was removed using Spin-columns. The mixtures of 5'-labeled primers and their complementary oligonucleotide templates (molar ratio of template:primer was 8:1) in annealing buffer containing 50 mM Tris-HCl pH 7.5, 10 mM MgCl2 and 0.1 mM DTT were heated at 90oC for 5 min, then cooled to room temperature over a period of 1.5 h.
To assess the overall effect of 5-foU on DNA synthesis, templates 1 or 3/primer 2 (Fig. 1 ) (0.1 pmol as the primer) in a reaction buffer (10 mM Tris-HCl pH 7.5, 5 mM MgCl2, 7.5 mM DTT and 200 µg/ml BSA) were incubated with 0.1 U of KF+ or KF- in the presence of four dNTPs (100 µM) at 25oC. One unit of KF+ and KF- is the amount of enzyme activity that incorporates 10 nmol of total nucleotides into acid-insoluble materials in 30 min at 25oC. The reaction with Tth DNA polymerase and Pfu DNA polymerase was carried out at 74oC. One unit of the thermostable DNA polymerases converts 10 nmol of dNTP into acid-insoluble materials in 30 min at 74oC.
To determine the nucleotides incorporated opposite 5-foU during DNA synthesis, the complementary pairs template 1, 2 or 3/primer 1 or 3 (Fig. 1 ) (50 fmol) in 10 µl reaction mixture were incubated with various DNA polymerases. The reaction proceeded in a buffer containing oligonucleotide templates annealed with primers (50 fmol), 10 mM Tris-HCl, 5 mM MgCl2, 7.5 mM DTT, 0.2 mg/ml BSA, 100 µM dNTP and 0.02 U enzyme for 5 min. The reaction was stopped by adding termination solution (95% formamide, 0.1% bromophenol blue, 0.1% xylene cyanol and 20 mM EDTA).
Kinetic studies of incorporation of nucleotides opposite 5-foU during DNA synthesis were done under the conditions described above using 0.01-500 µM dNTP and 0.2 U DNA polymerase. The complementary pairs template 1 or 3/primer (50 fmol) in 10 µl reaction mixture were incubated at 25oC (KF+ or KF-) or at 74oC (Tth and Pfu DNA polymerases) for 5 min. The Michaelis constant (Km) and the maximal velocity of the reaction (Vmax) were obtained from Lineweaver-Burk plots of the kinetic experimental results. The kcat/Km value was calculated according to Dong et al. (24 ).
The reaction mixtures were heated at 95oC for 5 min, cooled and loaded onto 20% polyacrylamide gels in the presence of 7 M urea, followed by resolution at a constant voltage of 1800 V. Approximately 2 fmol oligonucleotides were loaded in each lane. After electrophoresis the gels were dried and autoradiographed using Fuji RX films at -80oC. Band intensity was quantified by densitometry of the autoradiographs.
To examine the effects of 5-foU on DNA synthesis in vitro, a 22mer template containing one 5-foU at position 17 from the 3'-end (template 3) was primed with a 13mer primer (primer 2) and replicated with various DNA polymerases in the presence of four dNTPs. Reaction products were analyzed by denaturing polyacrylamide gel electrophoresis. The results with KF+ are shown in Figure 2 . After incubation for >5 min full-length DNA (22mer) was synthesized with the templates containing thymine and 5-foU. Termination bands due to pausing of DNA synthesis were not observed 1 nt prior to and opposite the modified base in the template. Similar results were obtained with KF-, Tth DNA polymerase and Pfu DNA polymerase (data not shown). These results indicate that DNA polymerases read through the site of 5-foU in the template.
We identified the nucleotides incorporated opposite 5-foU in the template during DNA synthesis. Primer 3 (16mer) annealed to template 3 was extended by various DNA polymerases in the presence of a single dNTP. The extension of primer 3 annealed with template 3 by KF- was analyzed by 7 M urea-20% polyacrylamide gel electrophoresis. The results are shown in Figure 3 . KF- incorporated dCMP in addition to dAMP opposite 5-foU. dGMP was slightly incorporated. Some incorporation of dCMP was seen with the normal template (Figs 3 and 4 ). This probably reflects an intrinsic low fidelity of KF-. As shown in Figure 4 , this misincorporation was significantly lowered when 3' -> 5' exonuclease-proficient KF+ was used. In contrast, incorporation opposite 5-foU was unaffected by exonuclease activity (Fig. 4 ). KF+ also inserted dCMP opposite the lesion as well as dAMP. The ratios of dCMP/dAMP incorporated opposite the 5-foU during reactions catalyzed by KF+ and KF- were 0.3 and 0.32 respectively. These results indicated that the 5-foU-dCMP mispair could be proofread.
When cellular DNA is exposed to oxidizing agents, several thymine hydroperoxides are formed. The major products are cis-6-hydroperoxy-5-hydroxy-5,6-dihydrothymine and cis-5-hydroperoxy-6-hydroxy-5,6-dihydrothymine, which gradually undergo secondary reactions to form thymine glycols (4 ,11 ). On the other hand, 5-hydroperoxymethyluracil spontaneously decomposes to produce 5-hydroxymethyluracil and 5-foU (4 ,11 ). Because 5-foU is readily formed in DNA by exposure to ionizing radiation and active oxygen species (4 ,9 ,12 -14 ), it is important to assess its mutagenic potential. Kasai et al. (9 ) have reported that 5-formyl-2'-deoxyuridine is mutagenic to Salmonella TA102 when added to the culture medium. However, the biological consequences of 5-foU formed in DNA remain unknown.
Many types of DNA damage have been identified in cells exposed to ionizing radiation and oxidizing agents (4 -8 ). If left unrepaired these could have lethal and mutagenic consequences. Recent developments in the chemical synthesis of oligonucleotides have made it possible to introduce a modified base into oligonucleotide templates at defined sites (17 -21 ). This is a useful means of predicting its lethal and mutagenic consequences based on the effects on DNA synthesis in vitro. In the present study oligonucleotide templates containing one 5-foU were synthesized by phosphoramidite chemistry (22 ). It is noteworthy that these oligonucleotides are stable to heat and alkali (data not shown).
Replication of the templates containing 5-foU with KF+ or KF- was not arrested 1 nt prior to or opposite the lesion (Fig. 2 ). Either a correct or an incorrect nucleotide other than dAMP could be incorporated opposite 5-foU. If the latter occurs in vivo it would result in base substitution mutations. In the presence of a single dNTP, KF+ incorporated dCMP in addition to dAMP opposite the site of 5-foU. KF- also inserted dCMP and dAMP opposite 5-foU. The ratios of dCMP to dAMP incorporated opposite the 5-foU during reactions catalyzed by KF+ and KF- were 0.3 and 0.32 respectively. Therefore, the 5-foU-dCMP mispair is stable to proofreading activity. This argument was supported by the results obtained with the Tth (exonuclease-deficient) and Pfu (proofreading exonuclease-proficient) DNA polymerases. These results of this study predict that T -> G transversions are induced by 5-foU.
The mutation spectrum will potentially be sequence context dependent, since the dCMP incorporation frequency opposite 5-foU was affected by the nearest neighbour base pair. If this is true in vivo, the mutagenic effect of 5-foU needs to be examined in association with nearest neighbour influence. The precise mechanisms for the sequence context-dependent manner of dCMP misincorporation opposite 5-foU remain unsolved. Nearest neighbour base stacking interactions appear to have different effects on the Km and Vmax of DNA polymerase reactions.
Powerful blocking lesions, such as urea residues and [beta]-ureidoisobutyric acid, are also mutagenic (27 ). Although these lesions direct specific misincorporations that are influenced by sequence context (27 ), their mutagenic activities might also be SOS dependent. Although 5-foU is not a blocker of DNA synthesis in vitro, it is of interest to assess the mutagenic properties of 5-foU introduced into reporter genes such as lacZ and supF on plasmid vectors in E.coli cells exposed to UV or not.
The results of this study suggest that 5-foU produced in DNA constitutes a lethal and mutagenic lesion by directing misincorporation of incorrect nucleotides. Lethal and mutagenic DNA lesions are usually removed by cellular DNA repair enzymes (23 ,28 -30 ). 5-foU in DNA has recently been identified as a substrate for the AlkA enzyme (3-methyladenine DNA glycosylase II) of E.coli (15 ). We have shown that mammalian liver extracts contain a DNA glycosylase activity for removing 5-foU from X-irradiated DNA (12 ). Furthermore, Bjelland et al. (16 ) have also reported that human blood cells have a similar DNA repair activity. The chemical synthesis of oligonucleotides containing 5-foU will allow the DNA repair pathways for 5-foU and the interaction of the lesion with DNA polymerases to be clarified (31 -34 ). Purification of 5-foU DNA glycosylase from mammalian cells is under investigation in our laboratory using the heteroduplex oligonucleotides containing one 5-foU molecule.
This study was partly supported by grants from the Ministry of Education, Science, Sports and Culture of Japan. This study was also supported by grants from the Inoue Foundation (to Q.-M.Z.) and from the Core Research for Evolutional Science and Technology of Japan Science and the Technology Corporation (to H.S.).
*To whom correspondence should be addressed. Tel: +81 75 753 4097; Fax: +81 75 753 4087; Email: yonei@kingyo.zool.kyoto-u.ac.jp
REFERENCES
+Present address: Institute for Medical and Dental Engineering, Tokyo Medical and Dental University, Surugadai, Chiyoda-ku, Tokyo 101, Japan