The base excision repair pathway.

The base excision repair pathway has evolved to protect cells from the deleterious effects of endogenous DNA damage induced by hydrolysis, reactive oxygen species and other intracellular metabolites that modify DNA base structure. However, base excision repair is also important to resist lesions produced by ionizing radiation and strong alkylating agents, which are similar to those induced by endogenous factors.

Repair and mutagenesis at oxidized DNA lesions in the developing brain of wild-type and Ogg1-/- mice.

OGG1 (8-oxoguanine DNA glycosylase-1) is one of the main DNA glycosylases present in mammalian cells. The enzyme removes 7,8-dihydro-8-oxoguanine (8-oxoG) lesions, believed to be the most important oxidized lesions due to their relatively high incidence and their miscoding properties. This study shows that in prenatal mice brains the repair capacity for 8-oxoG is 5-10-fold higher than in adult mice brains. Western blot analysis and repair activity in extracts from Ogg1(-/-) mice revealed that OGG1 was responsible for the efficient 8-oxoG removal from prenatal mice. To investigate how OGG1 protects against oxidative stress-induced mutagenesis, pregnant Big Blue/wild-type and Big Blue/Ogg1(-/-) mice were exposed to nontoxic doses of gamma radiation. A 2.5-fold increase in the mutation frequency in Ogg1(-/-) mouse brains was obtained by exposure to 3.5 Gy at day 19 postfertilization. This was largely due to GC to TA transversions, believed to originate from 8-oxoG mispairing with A during replication. Furthermore, rapid cell divisions seemed to be required for fixation of mutations, as a similar dose of radiation did not increase the mutation frequency, or the frequency of GC to TA transversion, in the adult brain.

The Saccharomyces cerevisiae homologues of endonuclease III from Escherichia coli, Ntg1 and Ntg2, are both required for efficient repair of spontaneous and induced oxidative DNA damage in yeast.

Endonuclease III from Escherichia coli is the prototype of a ubiquitous DNA repair enzyme essential for the removal of oxidized pyrimidine base damage. The yeast genome project has revealed the presence of two genes in Saccharomyces cerevisiae, NTG1 and NTG2, encoding proteins with similarity to endonuclease III. Both contain the highly conserved helix-hairpin-helix motif, whereas only one (Ntg2) harbors the characteristic iron-sulfur cluster of the endonuclease III family. We have characterized these gene functions by mutant and enzyme analysis as well as by gene expression and intracellular localization studies. Targeted gene disruption of NTG1 and NTG2 produced mutants with greatly increased spontaneous and hydrogen peroxide-induced mutation frequency relative to the wild type, and the mutation response was further increased in the double mutant. Both enzymes were found to remove thymine glycol and 2, 6-diamino-4-hydroxy-5-N-methylformamidopyrimidine (faPy) residues from DNA with high efficiency. However, on UV-irradiated DNA, saturating concentrations of Ntg2 removed only half of the cytosine photoproducts released by Ntg1. Conversely, 5-hydroxycytosine was removed efficiently only by Ntg2. The enzymes appear to have different reaction modes, as judged from much higher affinity of Ntg2 for damaged DNA and more efficient borhydride trapping of Ntg1 to abasic sites in DNA despite limited DNA binding. Northern blot and promoter fusion analysis showed that NTG1 is inducible by cell exposure to DNA-damaging agents, whereas NTG2 is constitutively expressed. Ntg2 appears to be a nuclear enzyme, whereas Ntg1 was sorted both to the nucleus and to the mitochondria. We conclude that functions of both NTG1 and NTG2 are important for removal of oxidative DNA damage in yeast.

Highly efficient base excision repair (BER) in human and rat male germ cells.

The quality of germ cell DNA is critical for the fate of the offspring, yet there is limited knowledge of the DNA repair capabilities of such cells. One of the main DNA repair pathways is base excision repair (BER) which is initiated by DNA glycosylases that excise damaged bases, followed by incision of the generated abasic (AP) sites. We have studied human and rat methylpurine-DNA glycosylase (MPG), uracil-DNA glycosylase (UNG), and the major AP endonuclease (HAP1/APEX) in male germ cells. Enzymatic activities and western analyses indicate that these enzymes are present in human and rat male germ cells in amounts that are at least as high as in somatic cells. Minor differences were observed between different cellular stages of rat spermatogenesis and spermiogenesis. Repair of methylated DNA was also studied at the cellular level using the Comet assay. The repair was highly efficient in both human and rat male germ cells, in primary spermatocytes as well as round spermatids, compared to rat mononuclear blood cells or hepatocytes. This efficient BER removes frequently occurring DNA lesions that arise spontaneously or via environmental agents, thereby minimising the number of potential mutations transferred to the next generation.

Reconstitution of the base excision repair pathway for 7,8-dihydro-8-oxoguanine with purified human proteins.

In mammalian cells, repair of the most abundant endogenous premutagenic lesion in DNA, 7,8-dihydro-8-oxoguanine (8-oxoG), is initiated by the bifunctional DNA glycosylase OGG1. By using purified human proteins, we have reconstituted repair of 8-oxoG lesions in DNA in vitro on a plasmid DNA substrate containing a single 8-oxoG residue. It is shown that efficient and complete repair requires only hOGG1, the AP endonuclease HAP1, DNA polymerase (Pol) beta and DNA ligase I. After glycosylase base removal, repair occurred through the AP lyase step of hOGG1 followed by removal of the 3'-terminal sugar phosphate by the 3'-diesterase activity of HAP1. Addition of PCNA had a slight stimulatory effect on repair. Fen1 or high concentrations of Pol beta were required to induce strand displacement DNA synthesis at incised 8-oxoG in the absence of DNA ligase. Fen1 induced Pol beta strand displacement DNA synthesis at HAP1-cleaved AP sites differently from that at gaps introduced by hOGG1/HAP1 at 8-oxoG sites. In the presence of DNA ligase I, the repair reaction at 8-oxoG was confined to 1 nt replacement, even in the presence of high levels of Pol beta and Fen1. Thus, the assembly of all the core proteins for 8-oxoG repair catalyses one major pathway that involves single nucleotide repair patches.

Repair of 8-oxodeoxyguanosine lesions in mitochondrial dna depends on the oxoguanine dna glycosylase (OGG1) gene and 8-oxoguanine accumulates in the mitochondrial dna of OGG1-defective mice.

Mitochondria are not only the major site for generation of reactive oxygen species, but also one of the main targets of oxidative damage. One of the major products of DNA oxidation, 8-oxodeoxyguanosine (8-oxodG), accumulates in mitochondrial DNA (mtDNA) at levels three times higher than in nuclear DNA. The main pathway for the repair of 8-oxodG is the base excision repair pathway initiated by oxoguanine DNA glycosylase (OGG1). We previously demonstrated that mammalian mitochondria from mice efficiently remove 8-oxodG from their genomes and isolated a protein from rat liver mitochondria with 8-oxoguanine (8-oxodG) DNA glycosylase/apurinic DNA lyase activity. In the present study, we demonstrated that the mitochondrial 8-oxodG DNA glycosylase/apurinic DNA lyase activity is the mitochondrial isoform of OGG1. Using mouse liver mitochondria isolated from ogg1(-/-) mice, we showed that the OGG1 gene encodes for the mitochondrial 8-oxodG glycosylase because these extracts have no incision activity toward an oligonucleotide containing a single 8-oxodG DNA base lesion. Consistent with an important role for the OGG1 protein in the removal of 8-oxodG from the mitochondrial genome, we found that mtDNA isolated from liver from OGG1-null mutant animals contained 20-fold more 8-oxodG than mtDNA from wild-type animals.

Identification of cytidine deaminase as inhibitor of granulocyte-macrophage colony formation.

Mature human blood granulocytes produce regulatory factors that inhibit colony formation by human and murine granulocyte-macrophage colony-forming cells (GM-CFC). The inhibition of GM-CFC by granulocyte extract (GRE) was strongly enhanced by the addition of thymidine (3 to 6 x 10(-5) M for human cells) and by the presence of fetal calf serum (FCS) in the growth medium. Deoxycytidine and deoxyuridine produced effects similar to those of thymidine, but at higher concentrations (2 to 4 x 10(-4) M). It was further observed that GRE prevented the antiproliferative effects of cytosine arabinoside (Ara-C) and azadeoxycytidine, suggesting that GRE contained cytidine deaminase (CDD) activity, since CDD is known to abolish the effects of these nucleoside analogs. Accordingly, the GRE was tested for and shown to contain an enzymatic activity that converted deoxycytidine to deoxyuridine, confirming the presence of CDD activity in GRE. The GM-CFC inhibition factor was found to copurify with CDD activity during three succeeding chromatographic separations, indicating that CDD was indeed the inhibiting factor itself. This conclusion was further substantiated by gel filtration experiments demonstrating a molecular weight (MW) of approximately 50 kd, which corresponds to the MW previously published for CDD activity. Furthermore, addition of tetrahydrouridine (THU), a known specific inhibitor of CDD, abolished the suppressive effect of GRE on GM-CFC, which independently confirmed the identification of CDD as an inhibitor of GM-CFC. The growth-regulating property of CDD could be explained by depletion of deoxycytidine nucleotides necessary for DNA synthesis or by a direct effect of CDD binding to specific receptors on progenitor cells.

Different efficiencies of the Tag and AlkA DNA glycosylases from Escherichia coli in the removal of 3-methyladenine from single-stranded DNA.

Escherichia coli possesses two different DNA repair glycosylases, Tag and AlkA, which have similar ability to remove the alkylation product 3-methyladenine from double-stranded DNA. In this study we show that these enzymes have quite different activities for the excision of 3-methyladenine from single-stranded DNA, AlkA being 10-20 times more efficient than Tag. We propose that AlkA and perhaps other glycosylases as well may have an important role in the excision of base damage from single-stranded regions transiently formed in DNA during transcription and replication.

Overexpression of endonuclease III protects Escherichia coli mutants defective in alkylation repair against lethal effects of methylmethanesulphonate.

Endonuclease III of Escherichia coli is normally involved in the repair of oxidative DNA damage. Here, we have investigated a possible role of EndoIII in the repair of alkylation damage because of its structural similarity to the alkylation repair enzyme 3-methyladenine DNA glycosylase II. It was found that overproduction of EndoIII partially relieved the alkylation sensitivity of alkA mutant cells. Site-directed mutagenesis to make the active site of EndoIII more similar to AlkA (K120W) had an adverse effect on the complementation and the mutant protein apparently inhibited repair by competing for the substrate without base release. These results suggest that EndoIII might replace AlkA in some aspect of alkylation repair, although high expression levels are needed to produce this effect.

Two separable protein species which both restore uvrABC endonuclease activity in extracts from uvrC mutated cells.

Two different protein species which both complement the detective repair endonuclease (uvrABC endonuclease) in uvrC mutated cells have been detected. These proteins have quite different chromatographic properties and were easily separated by ion exchange chromatography. One has affinity for DEAE cellulose and co-cromatographs with the uvrB protein. The other has strong affinity for phosphocellulose and appears to be the uvrC protein itself. The uvrB associated uvrC+ activity is absent from both uvrC and uvrB mutated cells, indicating that this species result from an interaction between uvrB+ and uvrC+ functions at the protein level.

Cloning and expression of a neuronal rat brain glutamate transporter.

Glutamate is the major excitatory transmitter in the mammalian central nervous system. Glutamate transporters, which keep the extracellular glutamate concentration low, are required both for normal brain function and for protecting neurons against harmful glutamatergic overstimulation. We have isolated the cDNA for a rat brain glutamate transporter (REAAC1) which has 90% amino acid and 86% nucleotide identity to the rabbit EAAC1. When REAAC1 was expressed in HeLa cells using a recombinant vaccinia-T7 virus expression system, a sodium dependent glutamate uptake was observed. The affinity of the carrier to various substrates was typical of brain "high affinity' glutamate uptake: threo-3-hydroxyaspartate, (R)-aspartate, (S)-glutamate and (S)-trans-pyrrolidine-2,4-dicarboxylic acid were strong inhibitors, but not (R)-glutamate or gamma-aminobutyrate. High resolution, non-radioactive in situ hybridization histochemistry in rat brain revealed the mRNA in several types of glutamatergic as well as non-glutamatergic neurons, but not in glial cells.

Production, isolation and purification of bacteriocins expressed by two strains of Neisseria meningitidis.

The systemic Neisseria meningitidis strain P241 and the healthy pharyngeal carrier strain BT878 produce bacteriocin-like substances during growth. A method has been devised for obtaining the active substances in solution. The activity was recovered by freeze-thaw extraction of dialyzed Todd-Hewitt agar medium into which the bacteriocins had diffused during growth of the producer strains. The bacteriocins were purified more than 50-fold by ammonium-sulphate precipitation and hydrophobic interaction chromatography. They are quite stable to heat and remain active 100% after 30 min at 100 degrees C. However, the protein nature of the bacteriocins has been confirmed by their sensitivity to alpha-chymotrypsin. Gel filtration indicated an Mr of 100-110 kDa, whereas SDS-polyacrylamide gel electrophoresis produced a common band by Coomassie staining corresponding to an Mr of 47-48 kDa, suggesting a dimer form of the active protein component.

Brain hypoplasia caused by exposure to trichlorfon and dichlorvos during development can be ascribed to DNA alkylation damage and inhibition of DNA alkyltransferase repair.

Treatment of pregnant guinea pigs with trichlorfon causes cerebellar hypoplasia in offspring. The most sensitive period for treatment is days 42-47 of gestation, which coincides with the rapid brain growth spurt and with the development of cerebellar granule cells. When rat granule cells were exposed in vitro to trichlorfon and dichlorvos for 24 hours they died, whereas trichloroethanol had no effect. When the cells were exposed to trichlorfon and dichlorvos for 3 hours, only dichlorvos was lethal indicating that the metabolite dichlorvos was more potent than trichlorfon itself. Cultured cerebellar granule cells were also found to be quite sensitive to other DNA-alkylating agents such as methylazoxymethanol and methylmethane sulphonate and to O6-benzylguanine; a potent and specific inhibitor of the DNA alkyltransferase involved in the repair of DNA alkylation damage. The organophosphorous compounds were also found to cause inhibition of the alkyltransferase and the lethal effects of the tested compounds on granule cell culture correlated well with the potency of inhibition. In a bacterial test system for monitoring alkylation effects on the DNA, dichlorvos was demonstrated to have a strong DNA alkylation effect. These results suggest that alkylation of DNA and inhibition of its repair can contribute to the brain hypoplasia observed after exposure to trichlorfon and dichlorvos during brain development.

5-Formyluracil and its nucleoside derivatives confer toxicity and mutagenicity to mammalian cells by interfering with normal RNA and DNA metabolism.

Oxidation of the methyl group of thymine yields 5-(hydroxymethyl)uracil (5-hmU) and 5-formyluracil (5-foU) as major products. Whereas 5-hmU appears to have normal base pairing properties, the biological effects of 5-foU are rather poorly characterised. Here, we show that the colony forming ability of Chinese hamster fibroblast (CHF) cells is greatly reduced by addition of 5-foU, 5-formyluridine (5-foUrd) and 5-formyl-2'-deoxyuridine (5-fodUrd) to the growth medium. There are no toxic effects of 5-fodUrd on cells defective in thymidine kinase or thymidylate synthetase, suggesting that the toxicity may be caused by 5-fodUrd phosphorylation and subsequent inhibition of thymidylate synthetase. Whereas 5-fodUrd was the most effective 5-foU derivative causing cell growth inhibition, the corresponding ribonucleoside 5-foUrd was more effective in inhibiting [3H]uridine incorporation in non-dividing rat nerve cells in culture, suggesting that 5-foUrd exerts its toxicity through interference with RNA rather than DNA synthesis. Addition of 5-foU and 5-fodUrd was also found to promote mutagenicity at the hypoxanthine-guanine phosphoribosyltransferase (HPRT) locus of CHF cells; 5-fodUrd being three orders of magnitude more potent than 5-foU. In contrast, neither 5-hmU nor 5-(hydroxymethyl)-2'-deoxyuridine induced HPRT mutations. The mutation induction indicates that 5-foU will be incorporated into DNA and has base pairing properties different from that of thymine. These results suggest that 5-foU residues, originating from incorporation of oxidised bases, nucleosides or nucleotides or by oxidation of DNA, may contribute significantly to the damaging effects of oxygen radical species in mammalian cells.

Strand cleavage at psoralen adducts and pyrimidine dimers in DNA caused by interaction between semi-purified uvr+ gene products from Escherichia coli.

Partially purified extracts of Escherichia coli containing either uvrA+ or a mixture of uvrB+ and uvrC+ gene products were tested for an endonuclease activity on DNA treated with 8-methoxypsoralen plus 360-nm light. Neither of these fractions was active alone. The combined fractions, however, caused extensive strand cleavage of the psoralen-treated DNA. The endonuclease activity was dependent upon addition of ATP and Mg2+ to the reaction mixtures, and hence appeared similar to the UV-endonuclease activity previously shown to be reconstituted from the same fractions. It is concluded that the uvr+ gene products in these fractions interact to cause breakage of both psoralen-treated and UV-irradiated DNA. An examination of the dose-dependence relationship of the break formation in psoralen-treated DNA revealed that the enzyme acts upon psoralen mono-adducts. By varying the experimental conditions to increase the ratio of interstrand cross-links to mono-adducts it was found that the enzyme also acts upon cross-links, but with lower efficiency than for mono-adducts. Further studies of break formation in UV-irradiated DNA showed that elimination of pyrimidine dimers by treatment with photoreactivating enzyme in the light resulted in a loss of endonuclease-sensitive sites. This shows directly that pyrimidine dimers are the lesions recognized by the complemented uvr+ gene products in UV-irradiated DNA. For comparison, another endonuclease acting at pyrimidine dimers in DNA, the Micrococcus luteus UV-endonuclease, was also tested with psoralen-treated DNA, but no activity was observed. This and other data indicate that the repair endonuclease encoded by the uvr+ genes in E. coli is basically different from the other dimer-specific endonucleases previously characterized.

Cellular effects of 5-formyluracil in DNA.

5-Formyluracil is a major oxidation product of thymine, formed in DNA in yields comparable to that of 8-oxo-7,8-dihydroguanine by exposure to gamma-irradiation. Whereas the repair pathways for removal and the biological effects of persisting 8-oxo-7,8-dihydroguanine are much elucidated, much less attention has been paid to the cellular implications of 5-formyluracil in DNA. Here we review the present state of knowledge in this important area within research on oxidative DNA damage.

Strand-break formation in DNA modified by benzo[alpha]pyrene diolepoxide. Quantitative cleavage by Escherichia coli uvrABC endonuclease.

Covalently closed circular plasmid DNA was modified by benzo[alpha]pyrene diolepoxide and incubated with partially purified fractions of the Escherichia coli uvr+ gene products. Strand breaks were introduced into the modified DNA by the uvrABC endonuclease; on average, one break was formed for each bound benzo[alpha]pyrene residue in the DNA. These results are direct evidence that benzo[alpha]pyrene adducts in DNA are acted upon by the same repair enzyme as those that handle UV-induced lesions in DNA.

Cell-cycle regulation, intracellular sorting and induced overexpression of the human NTH1 DNA glycosylase involved in removal of formamidopyrimidine residues from DNA.

Endonuclease III (Nth) of Escherichia coli is a DNA glycosylase essential for the removal of oxidised pyrimidine base residues from DNA. Several eukaryotic homologues have recently been identified and shown to have biochemical properties similar to those of Nth. However, some of the eukaryotic counterparts also appear to remove imidazole ring-opened purine residues (faPy), a property not shared by the enzymes of bacterial origin. Here, we show that the human enzyme also possesses efficient faPy DNA glycosylase activity as indicated both from studies of the purified protein and induced overexpression of the human NTH1 cDNA in HeLa cells. We constructed green fluorescent protein-tagged hNTH1 fusion proteins to study the cellular localisation of hNTH1 and found strong and exclusive sorting to the nucleus. Studies with synchronised cells showed that the expression of hNTH1 is regulated during the cell cycle with increased transcription during early and mid S-phase.

DNA damage induced by 3-chloro-4-(dichloromethyl)-5-hydroxy-2[5H]-furanone (MX) in HL-60 cells and purified DNA in vitro.

Chlorinated tap water often contains 3-chloro-4-(dichloromethyl)-5-hydroxy-2[5H]-furanone (MX), which is a potent directly acting bacterial mutagen. We have investigated the induction of DNA damage by MX in a promyelocytic human leukaemia cell line (HL-60 cells). Exposure of HL-60 cells to 100-300 microM MX resulted in increased levels of DNA single-strand breaks and/or alkali-labile sites (SSBs) as detected by alkaline filter elution. When adding inhibitors of DNA break repair (AraC plus hydroxyurea), increased levels of DNA SSBs were observed at very low concentrations (1-3 microM) of MX, as observed by both alkaline filter elution and the single-cell gel electrophoresis assay. Increased DNA SSBs could also be observed if DNA repair inhibitors were added immediately after exposure to 10 microM MX, indicating that low concentrations of MX cause a relatively stable modification of DNA that may be recognized and incised by DNA repair enzyme activities. Further studies with DNA break repair inhibitors indicated that HL-60 cells exposed to 10 microM MX for 1 h repaired 50% of their initial DNA damage during a 2-h period and the repair appeared to be complete at 22 h. Analysis of MX-treated DNA by sequencing methods indicated that MX preferentially reacts with guanines in DNA.

Mutations induced by 5-formyl-2'-deoxyuridine in Escherichia coli include base substitutions that can arise from mispairs of 5-formyluracil with guanine, cytosine and thymine.

5-Formyluracil (5-foU) is a major oxidation product of thymine formed in yields comparable to that of 8-oxoguanine in DNA by ionizing radiation. Whereas the mutagenic effects of 8-oxoguanine are well understood, the investigation of the biological implications of 5-foU has so far been limited. Here we demonstrate that 5-formyl-2'-deoxyuridine (5-fodUrd) supplied to the growth medium of Escherichia coli induces several base substitutions at different frequencies at position 461 in the lacZ gene in the following order: A.T-->G.C>G.C-->A.T>G.C-->T.A>>A.T-->T.A>A.T-->C.G. No induction of G.C-->C.G transversions was observed. It is inferred that 5-fodUrd will be incorporated into the DNA during cell growth, forming mispairs with guanine, cytosine and thymine during replication. It, thus, appears that cell growth in the presence of 5-fodUrd may represent a good model for elucidating the cellular effects of 5-foU residues in DNA.