Cytosine photoproduct-DNA glycosylase in Escherichia coli and cultured human cells

Ultraviolet irradiation produces novel endonuclease III-sensitive cytosine photoproducts at dipyrimidine sites

Ultraviolet light irradiation of DNA in vitro and in vivo induces cyclobutane dimers, (6-4) pyrimidine-pyrimidone photoproducts and a variety of minor products. Using a defined DNA fragment, we have identified two classes of sites that can be cleaved by Escherichia coli endonuclease III: single cytosines whose heat lability corresponds to that of cytosine hydrates and more heat-stable dipyrimidines containing cytosine. The dipyrimidine products are induced at sites suggestive of (6-4) photoproducts but are not recognized as (6-4) photoproducts by radioimmunoassay. Use of oligonucleotides containing a single cyclobutane thymine dimer, a (6-4) photoproduct or the Dewar photoisomer of the (6-4) photoproduct also indicated that these products are not substrates for endonuclease III. We have therefore identified a minor UV photoproduct that has the same sequence specificity as the two major dipyrimidine photoproducts; it may be a minor isomer, a unique derivative or an oxidative lesion confined to dipyrimidine sites. Its biological significance is not yet known but may be masked by the preponderance of major products at the same sites. Its occurrence at the particular site in dipyrimidine sequences involved in the mutagenic action of UV photoproducts suggests that it may play a role in generating C to T transitions that are common UV-induced mutations.

Excision of ultraviolet-induced photoproducts of 5-methylcytosine from DNA

Transition mutations at DNA 5-methylcytosines, congregated at CpG islands, are implicated in the etiogenesis of human diseases. Formation of 5-methylcytosine hydrate (5-methyl-6-hydroxy-5,6-dihydrocytosine) by hydration of the 5,6 double bond of 5-methylcytosine has been suggested as an intermediate in a possible mechanism of deamination to thymine. Ultraviolet irradiation of DNA yields pyrimidine hydrates, which are removed by repair glycosylases. We have identified 5-methylcytosine photoproducts following their excision from DNA by E. coli endonuclease III. Poly(dG-[3H]5-medC):poly(dG-[3H]5-medC) was irradiated and reacted with the enzyme. Radiolabeled photoproduct releases were directly proportional to irradiation doses and enzyme concentrations. These were identified as cis-thymine hydrate (6-hydroxy-5,6-dihydrothymine) and trans-thymine hydrate. Recovery of thymine hydrates is consistent with hydration of pyrimidines. Subsequent heating (which converts thymine hydrates to thymines) and chemical sequencing of an irradiated, 3' end-labeled, synthetic DNA strand demonstrated the appearance of thymine at the 5-methylcytosine site. These results demonstrate a mechanism for deamination of DNA 5-methylcytosine via hydration of the 5,6 double bond, putatively yielding 5-methylcytosine hydrate; this deaminates to thymine hydrate, and loss of water yields thymine formation at the 5-methylcytosine site. Identification of these DNA 5-methylcytosine modified moieties indicates a possible molecular mechanism for the frequent transition mutations found at CpG loci.

Enzymic removal of 5-methylcytosine from DNA by a human DNA-glycosylase

DNA 5-methylcytosine is a major factor in the silencing of mammalian genes; it is involved in gene expression, differentiation, embryogenesis and neoplastic transformation. A decrease in DNA 5-methylcytosine content is associated with activation of specific genes. There is much evidence indicating this to be an enzymic process, with replacement of 5-methylcytosine by cytosine. We demonstrate here enzymic release of 5-methylcytosines from DNA by a human 5-methylcytosine-DNA glycosylase activity, which affords a possible mechanism for such replacement. This activity generates promutagenic apyrimidinic sites, which can be related to the high frequency of mutations found at DNA 5-methylcytosine loci. The recovery of most released pyrimidines as thymines indicates subsequent deamination of free 5-methylcytosines by a 5-methylcytosine deaminase activity. This prevents possible recycling of 5-methylcytosine into replicative DNA synthesis via a possible 5-methyl-dCTP intermediate synthesized through the pyrimidine salvage pathway. Taken together, these findings indicate mechanisms for removal of 5-methylcytosines from DNA, hypermutability of DNA 5-methylcytosine sites, and exclusion of 5-methylcytosines from DNA during replication.

In vivo repair of cytosine hydrates in DNA of cultured human lymphoblasts

Ultraviolet irradiation of DNA in vitro results in the production of a wide variety of pyrimidine base alterations, including cytosine hydrates. Enzymes that initiate the repair of monomeric pyrimidine damage have been identified in both bacterial and mammalian systems; however, the in vivo formation and repair of cytosine photohydrates has not been demonstrated in cellular DNA. Using Escherichia coli endonuclease III as a damage-specific probe, we have shown that ring-saturated pyrimidines are formed in cultured human cells by irradiation with broad-spectrum UV light. In addition, these types of base damage are removed from the DNA of human lymphoblasts within 5 h following the irradiation. Analysis of the action spectrum for the formation of cytosine hydrates in DNA reveals that these photoproducts are formed most efficiently by irradiation in the range of 255-265 nm light, coinciding with the wavelengths that are maximally absorbed by the DNA bases.

Properties and biological functions of the NTH and FPG proteins of Escherichia coli: two DNA glycosylases that repair oxidative damage in DNA

Oxidative damage to DNA is one of the most important causes of spontaneous mutations and may play a role in aging and related diseases, such as cancer, in humans. Oxidative damage results from the attack of biomolecules by free radicals and reactive oxygen species formed as byproducts of normal cell metabolism or during oxidative stress. To counteract the lethal and mutagenic effects of oxidative lesions in DNA, cells have developed defence strategies including DNA repair systems. In Escherichia coli, the repair of oxidized bases in DNA is mostly mediated by the base excision repair pathway. The first step in this DNA repair pathway is catalysed either by the NTH protein which excises oxidized pyrimidines or by the FPG protein which excises oxidized purines. The nucleotide excision repair pathway mediated by the UvrABC complex may also play a role when the DNA glycosylases are inactive or saturated. This review summarizes the structural and catalytic properties of the NTH and FPG proteins of Escherichia coli and presents evidence to indicate that these two enzymes constitute an important component of the cellular defence against oxidative stress in prokaryotes and eukaryotes.

The distribution of UV damage in the lacI gene of Escherichia coli: correlation with mutation spectrum

We have determined the UV (254 nm) damage distribution in the transcribed and non-transcribed strands of the i-d region of the Escherichia coli lacI gene. The locations of replication blocking lesions were revealed as termination sites of T7 DNA polymerase and/or T4 DNA polymerase 3'-5' exonuclease. Termination products, i.e. both cyclobutane pyrimidine dimers and 6-4 photoproducts, were resolved and analysed on an automated DNA sequencer. These two major photoproducts are not randomly distributed along the gene, but occur in clusters, in longer runs of pyrimidines. We also have compared the UV damage distribution with the previously reported UV-induced base substitutions in the same region. Mutational hotspots, in both repair-deficient and repair-proficient strains of E. coli, are all located in stretches of pyrimidines, and thus correlate well with the distribution of photolesions. One mutational hotspot in the wild-type strain may reflect the high frequency of closely opposed lesions. To explain the other mutational hotspots, we propose that the repair of UV lesions is impaired due to the local conformation of the DNA, which might deviate from the B-form. This hypothesis is supported by the excess of mutational hotspots in pyrimidine runs in the Uvr+ strain compared to Uvr-. Runs of pyrimidines thus represent both damage- and mutation-prone sequences following UV treatment.

Stability of DNA thymine hydrates

Pyrimidine hydrates are products of ultraviolet irradiation of DNA. We have already demonstrated the formation of both cis-thymine hydrate and trans-thymine hydrate (6-hydroxy-5,6-dihydrothymine) in irradiated poly(dA-dT):poly(dA-dT). These are released from DNA as free bases by bacterial or human glycosylases. Thymine hydrate stabilities were studied in irradiated DNA substrates using purified E. coli endonuclease III as a reagent for their removal. After irradiation, substrate poly(dA-dT):poly(dA-dT), radiolabeled in thymine, was incubated at 50, 60, 70 or 80 degrees C, cooled, and then reacted with the enzyme under standard conditions. Thymine hydrates were assayed by enzymic release of labeled material into the ethanol-soluble fraction. Their identities were confirmed by high performance liquid chromatography. The decay of thymine hydrates in heated DNA followed first-order kinetics with a k = 2.8 x 10(-5)/sec at 80 degrees C. These hydrates were also detected in lesser quantities in the unirradiated, control substrate. Extrapolation from an Arrhenius plot yields an estimated half-life of 33.3 hours at 37 degrees C for DNA thymine hydrates. Such stability, together with their formation in unirradiated DNA, suggest thymine hydrates to be formed under physiological conditions and to be sufficiently stable in DNA to be potentially genotoxic. This necessitates their constant removal from DNA by the excision-repair system.

Repair of DNA damage induced by reactive oxygen species

DNA repair limits the mutagenic, and thereby the carcinogenic, effect of DNA modifications. Free radicals, particularly reactive oxygen species, induce all forms of DNA damage, including base modifications, base free sites, strand breakage, and cross-links. These lesions are repaired by a variety of enzymes of partly overlapping substrate specificity, some of which may be induced.

Yields of radiation-induced base products in DNA: effects of DNA conformation and gassing conditions

Gas chromatography-mass spectrometry with selected-ion monitoring was used to measure the yields of radiation-induced base products in aqueous solutions of native or heat-denatured DNA irradiated in the dose range 20-100 Gy. These DNA solutions were saturated with nitrous oxide, nitrogen, air or 20% oxygen in nitrous oxide during irradiation. The products measured were as follows: 5,6-dihydrothymine; 5-hydroxy-5,6-dihydrothymine; 5,6-dihydrothymine (thymine glycol); 5-hydroxy-5,6-dihydrocytosine; 5,6-dihydroxy-5,6- dihydrocytosine (cytosine glycol); 4,6-diamino-5-formamidopyrimidine; 7,8-dihydro-8-oxoadenine (8-hydroxyadenine); 2,6-diamino-4-hydroxy-5- formamidopyrimidine; and 7,8-dihydro-8-oxoguanine (8-hydroxyguanine). In oxygenated solutions, 5,6-dihydrothymine, 5-hydroxy-5,6-dihydrothymine and 5-hydroxy-5,6-dihydrocytosine were not formed. The yields of all products, other than 5,6-dihydrothymine, were greater in irradiated DNA samples from N2O-saturated solutions than from N2-saturated solutions. In N2-saturated solutions the yield of 8-hydroxyadenine was low and 8-hydroxyguanine was undetectable. Yields of pyrimidine products in heat-denatured DNA were greater than those in native DNA using all types of gases. However, the effects of DNA conformation on the yields of purine products were dependent on the type of gas used to saturate the irradiated DNA solutions. Yields of formamidopyrimidines were generally lower in solutions of DNA irradiated in the native than in the heat-denatured conformation. In air-saturated solutions of DNA, yields of 8-hydroxypurines were not influenced greatly by DNA conformation. In DNA solutions saturated with N2O/O2, 8-hydroxypurine formation was more favourable in the heat-denatured conformation than in the native conformation. On the other hand, in deoxygenated solutions, formation of 8-hydroxypurines was favoured in the native conformation. Data indicate that DNA conformation and the type of gas used to saturate the irradiated solutions have a profound influence on yields of base products in DNA.

Excision of cytosine hydrates from Z-DNA

Ultraviolet irradiation of DNA produces cytosine hydrate, released as a free base by E. coli endonuclease III. Cytosine hydrate excision was investigated by assaying photoproduct release from cytosine-radiolabeled, irradiated poly(dG-dC):poly(dG-dC). Conformational shifts between B-DNA and Z-DNA were affected by heating the polymer in either nickel chloride or cobaltous chloride, and were determined by circular dichroism. Rates of enzymic cytosine hydrate release did not differ between the different substrate conformations. Irradiation of left-handed poly(dG-dC):poly(dG-dC) resulted in cytosine hydrate formation. Therefore, neither formation nor enzymic excision of ultraviolet-induced cytosine hydrates are substantially affected by these DNA conformational states.

Free-radical reactions induced by OH-radical attack on cytosine-related compounds: a study by a method combining ESR, spin trapping and HPLC

Free-radical reactions induced by OH-radical attack on cytosine-related compounds were investigated by a method combining ESR, spin trapping with 2-methyl-2-nitrosopropane and high-performance liquid chromatography (HPLC). Cytidine, 2'-deoxycytidine, cytidine 3'-monophosphate, cytidine 5'-monophosphate, 2'-deoxycytidine 5'-monophosphate and their derivatives, of which 5,6-protons at the base moiety were replaced by deuterons, and polycytidylic acid (poly(C] were employed as samples. OH radicals were generated by X-irradiating an N2O-saturated aqueous solution. Five spin adducts were separated by HPLC. Examination of them by ESR spectroscopy and UV photospectrometry showed that spin adducts assigned to C5 and C6 radicals due to OH addition to the 5,6 double-bond, a deaminated form of the spin adduct derived from a C5 radical due to the cyclization reaction between C5' of the sugar and C6 of the base, and a spin adduct assigned to the C4' radical due to H abstraction by OH radicals were produced. From these results the sites of OH-radical attack and the subsequent radical reactions in cytosine-related compounds were clarified.

Glycosylases that excise modified DNA pyrimidines in young and senescent human WI-38 fibroblasts

Cellular DNA is continuously subject to damages by both endogenous and exogenous oxidizing agents. Excision repair in human cells is initiated by DNA glycosylases which remove oxidized bases from DNA. 5-Hydroxymethyluracil-DNA glycosylase excises 5-hydroxymethyluracil from DNA. A different enzyme has glycosylic activity against many ring-saturated DNA pyrimidines. Levels of these enzymes were examined in WI-38 fibroblasts of different culture ages. All glycosylases were assayed by measurements of direct release of modified free bases from their respective DNA substrates. Levels of 5-hydroxymethyluracil-DNA glycosylase were reduced in aging cells. Specific activities of the glycosylase that releases ring-saturated pyrimidines and of uracil-DNA glycosylase were not substantially altered in senescent cells. Therefore, although aging cells might have reduced excision of DNA 5-hydroxymethyluracil, there is no overall age-dependent decrease of DNA glycosylase activities.

Differential cell cycle modulation of human DNA glycosylases against oxidized pyrimidines

Cellular DNA is continuously subject to damages by both endogenous and exogenous oxidizing agents. Excision repair of oxidized bases in human cells is initiated by DNA glycosylases which remove them from DNA. 5-Hydroxymethyluracil-DNA glycosylase excises 5-hydroxymethyluracil from DNA. A different enzyme, termed a redoxyendonuclease, has glycosylase activity against many modified DNA pyrimidines. The regulation of these enzymes in proliferating human cells was examined. Both glycosylases were assayed in serum-stimulated WI-38 cells by measurements of direct release of modified free bases from their respective DNA substrates. There was no significant variation of 5-hydroxymethyluracil-DNA glycosylase activity during the cell cycle. However, the glycosylic activity of the redoxyendonuclease was stimulated with DNA synthesis. This activity again increased at the beginning of a second cell cycle. Therefore, the glycosylases that initiate excision repair of oxidized DNA are subject to different controls during the cell cycle.

A human endonuclease incises ultraviolet-irradiated DNA at cytosines and oxidized DNA at thymines

Both ultraviolet irradiation and oxidation of DNA produce a variety of pyrimidine base damages. A human endonuclease recognizes such altered bases on these DNA substrates. This human endonuclease incises ultraviolet-irradiated DNA exclusively at sites of photochemically modified cytosines. The precise sites of incision by the human enzyme were determined by DNA sequencing. Chemically oxidized DNA was incised exclusively at thymine loci. The degree of enzymic cleavage at cytosine photoproducts was identical at each site. However, the extent of incision at selected oxidized thymine residues varied within the DNA sequence. These results indicate that the distribution of thymine oxidative modifications is influenced by the neighboring DNA bases.