Enzymatic DNA Methylation

Wavelength- and Tissue-dependent Variations in the Mutagenicity of Cyclobutane Pyrimidine Dimers in Mouse Skin

The cyclobutane pyrimidine dimer (CPD) is a main mutagenic photolesion in DNA produced by UVR. We previously studied the wavelength-dependent kinetics of mutation induction efficiency using monochromatic UVR sources and transgenic mice developed for mutation assay and established the action spectra of UVR mutagenicity in the mouse epidermis and dermis. Here, we further established the action spectra of CPD and pyrimidine(6-4)pyrimidone photoproduct formation in the same tissues and in naked DNA using the same sources and mouse strain. Quantitative ELISA helped us estimate the photolesion formation efficiencies on a molecule-per-nucleotide basis. Using these action spectra, we confirmed that the UVR mutation mostly depends on CPD formation. Moreover, the mutagenicity of a CPD molecule (CPD mutagenicity) was found to vary by wavelength, peaking at approximately 313 nm in both the epidermis and dermis with similar wavelength-dependent patterns. Thus, the CPD formation efficiency is a main determinant of UVR mutagenicity in mouse skin, whereas a wavelength-dependent variation in the qualitative characteristics of CPD molecules also affects the mutagenic consequences of UVR insults. In addition, the CPD mutagenicity was always higher in the epidermis than in the dermis, suggesting different cellular responses to UVR between the two tissues irrespective of the wavelength.

Mechanistic considerations on the wavelength-dependent variations of UVR genotoxicity and mutagenesis in skin: the discrimination of UVA-signature from UV-signature mutation

Ultraviolet radiation (UVR) predominantly induces UV-signature mutations, C → T and CC → TT base substitutions at dipyrimidine sites, in the cellular and skin genome. I observed in our in vivo mutation studies of mouse skin that these UVR-specific mutations show a wavelength-dependent variation in their sequence-context preference. The C → T mutation occurs most frequently in the 5'-TCG-3' sequence regardless of the UVR wavelength, but is recovered more preferentially there as the wavelength increases, resulting in prominent occurrences exclusively in the TCG sequence in the UVA wavelength range, which I will designate as a "UVA signature" in this review. The preference of the UVB-induced C → T mutation for the sequence contexts shows a mixed pattern of UVC- and UVA-induced mutations, and a similar pattern is also observed for natural sunlight, in which UVB is the most genotoxic component. In addition, the CC → TT mutation hardly occurs at UVA1 wavelengths, although it is detected rarely but constantly in the UVC and UVB ranges. This wavelength-dependent variation in the sequence-context preference of the UVR-specific mutations could be explained by two different photochemical mechanisms of cyclobutane pyrimidine dimer (CPD) formation. The UV-signature mutations observed in the UVC and UVB ranges are known to be caused mainly by CPDs produced through the conventional singlet/triplet excitation of pyrimidine bases after the direct absorption of the UVC/UVB photon energy in those bases. On the other hand, a novel photochemical mechanism through the direct absorption of the UVR energy to double-stranded DNA, which is called "collective excitation", has been proposed for the UVA-induced CPD formation. The UVA photons directly absorbed by DNA produce CPDs with a sequence context preference different from that observed for CPDs caused by the UVC/UVB-mediated singlet/triplet excitation, causing CPD formation preferentially at thymine-containing dipyrimidine sites and probably also preferably at methyl CpG-associated dipyrimidine sites, which include the TCG sequence. In this review, I present a mechanistic consideration on the wavelength-dependent variation of the sequence context preference of the UVR-specific mutations and rationalize the proposition of the UVA-signature mutation, in addition to the UV-signature mutation.

Quantitative analysis of UV photolesions suggests that cyclobutane pyrimidine dimers produced in mouse skin by UVB are more mutagenic than those produced by UVC

The amount of photolesions produced in DNA after exposure to physiological doses of ultraviolet radiation (UVR) can be estimated with high sensitivity and at low cost through an immunological assay, ELISA, which, however, provides only a relative estimate that cannot be used for comparisons between different photolesions such as cyclobutane pyrimidine dimer (CPD) and pyrimidine(6-4)pyrimidone photoproduct (64PP) or for analysis of the genotoxicity of photolesions on a molecular basis. To solve this drawback of ELISA, we introduced a set of UVR-exposed, calibration DNA whose photolesion amounts were predetermined and estimated the absolute molecular amounts of CPDs and 64PPs produced in mouse skin exposed to UVC and UVB. We confirmed previously reported observations that UVC induced more photolesions in the skin than UVB at the same dose, and that both types of UVR produced more CPDs than 64PPs. The UVR protection abilities of the cornified and epidermal layers for the lower tissues were also evaluated quantitatively. We noticed that the values of absorbance obtained in ELISA were not always proportional to the molecular amounts of the lesion, especially for CPD, cautioning against the direct use of ELISA absorbance data for estimation of the photolesion amounts. We further estimated the mutagenicity of a CPD produced by UVC and UVB in the epidermis and dermis using the mutation data from our previous studies with mouse skin and found that CPDs produced in the epidermis by UVB were more than two-fold mutagenic than those by UVC, which suggests that the properties of CPDs produced by UVC and UVB might be different. The difference may originate from the wavelength-dependent methyl CpG preference of CPD formation. In addition, the mutagenicity of CPDs in the dermis was lower than that in the epidermis irrespective of the UVR source, suggesting a higher efficiency in the dermis to reduce the genotoxicity of CPDs produced within it. We also estimated the minimum amount of photolesions required to induce the mutation induction suppression (MIS) response in the epidermis to be around 15 64PPs or 100 CPDs per million bases in DNA as the mean estimate from UVC and UVB-induced MIS.

In Vivo Spectrum of UVC-induced Mutation in Mouse Skin Epidermis May Reflect the Cytosine Deamination Propensity of Cyclobutane Pyrimidine Dimers

Although ultraviolet radiation (UVR) has a genotoxicity for inducing skin cancers, the skin may tolerate UVC component because the epidermal layer prevents this short wavelength range from passing through. Here, UVC genotoxicity for mouse skin was evaluated in terms of DNA damage formation and mutagenicity. UVC induced UVR photolesions and mutations remarkably in the epidermis but poorly in the dermis, confirming the barrier ability of the epidermis against shorter UVR wavelengths. Moreover, the epidermis itself responded to UVC mutagenicity with mutation induction suppression, which suppressed the mutant frequencies to a remarkably low, constant level regardless of UVC dose. The mutation spectrum observed in UVC-exposed epidermis showed a predominance of UV-signature mutation, which occurred frequently in 5'-TCG-3', 5'-TCA-3' and 5'-CCA-3' contexts. Especially, for the former two contexts, the mutations recurred at several sites with more remarkable recurrences at the 5'-TCG-3' sites. Comparison of the UVC mutation spectrum with those observed in longer UVR wavelength ranges led us to a mechanism that explains why the sequence context preference of UV-signature mutation changes according to the wavelength, which is based on the difference in the mCpG preference of cyclobutane pyrimidine dimer (CPD) formation among UVR ranges and the sequence context-dependent cytosine deamination propensity of CPD.

Remarkable induction of UV-signature mutations at the 3'-cytosine of dipyrimidine sites except at 5'-TCG-3' in the UVB-exposed skin epidermis of xeroderma pigmentosum variant model mice

The human POLH gene is responsible for the variant form of xeroderma pigmentosum (XP-V), a genetic disease highly susceptible to cancer on sun-exposed skin areas, and encodes DNA polymerase η (polη), which is specialized for translesion DNA synthesis (TLS) of UV-induced DNA photolesions. We constructed polη-deficient mice transgenic with lacZ mutational reporter genes to study the effect of Polh null mutation (Polh(-/-)) on mutagenesis in the skin after UVB irradiation. UVB induced lacZ mutations with remarkably higher frequency in the Polh(-/-) epidermis and dermis than in the wild-type (Polh(+/+)) and heterozygote. DNA sequences of a hundred lacZ mutants isolated from the epidermis of four UVB-exposed Polh(-/-) mice were determined and compared with mutant sequences from irradiated Polh(+)(/)(+) mice. The spectra of the mutations in the two genotypes were both highly UV-specific and dominated by C→T transitions at dipyrimidines, namely UV-signature mutations. However, sequence preferences of the occurrence of UV-signature mutations were quite different between the two genotypes: the mutations occurred at a higher frequency preferentially at the 5'-TCG-3' sequence context than at the other dipyrimidine contexts in the Polh(+/+) epidermis, whereas the mutations were induced remarkably and exclusively at the 3'-cytosine of almost all dipyrimidine contexts with no preference for 5'-TCG-3' in the Polh(-/-) epidermis. In addition, in Polh(-/-) mice, a small but remarkable fraction of G→T transversions was also observed exclusively at the 3'-cytosine of dipyrimidine sites, strongly suggesting that these transversions resulted not from oxidative damage but from UV photolesions. These results would reflect the characteristics of the error-prone TLS functioning in the bypass of UV photolesions in the absence of polη, which would be mediated by mechanisms based on the two-step model of TLS. On the other hand, the deamination model would explain well the mutation spectrum in the Polh(+/+) genotype.

Solar-UV-signature mutation prefers TCG to CCG: extrapolative consideration from UVA1-induced mutation spectra in mouse skin

UVA1 exerts its genotoxicity on mammalian skin by producing cyclobutane pyrimidine dimers (CPDs) in DNA and preferentially inducing solar-UV-signature mutations, C → T base substitution mutations at methylated CpG-associated dipyrimidine (Py-mCpG) sites, as demonstrated previously using a 364 nm laser as a UVA1 source and lacZ-transgenic mice that utilize the transgene as a mutational reporter. In the present study, we confirmed that a broadband UVA1 source induced the same mutation profiles in mouse epidermis as the UVA1 laser, generalizing the previous result from a single 364 nm to a wider wavelength range of UVA1 (340-400 nm). Combined with our previous data on the mutation spectra induced in mouse epidermis by UVB, UVA2 and solar UVR, we proved that the solar-UV-signature mutation is commonly observed in the wavelength range from UVB to UVA, and found that UVA1 induces this mutation more preferentially than the other shorter wavelength ranges. This finding indicates that the solar-UV-signature mutation-causing CPDs, which are known to prefer Py-mCpG sites, could be produced with the energy provided by the longer wavelength region of UVR, suggesting a photochemical reaction through the excitation of pyrimidine bases to energy states that can be accomplished by absorption of even low-energy UVR. On the other hand, the lower proportions of solar-UV-signature mutations observed in the mutation spectra for UVB and solar UVR indicate that the direct photochemical reaction through excited singlet state of pyrimidine bases, which can be accomplished only by high-energy UVR, is also involved in the mutation induction at those shorter wavelengths of UVR. We also found that the solar-UV signature prefers 5'-TCG-3' to 5'-CCG-3' as mutational target sites, consistent with the fact that UVA induces CPDs selectively at thymine-containing dipyrimidine sites and that solar UVR induces them preferably at Py-mCpG sites. However, the mutation spectrum in human p53 gene from non-melanoma skin cancers shows the opposite preference for 5'-CCG-3' sites. This apparent discrepancy in the site preference seems to result from the lack of 5'-TCG-3' sites mutable to missense mutations on the nontranscribed strand of human p53 gene, which should be evolutionally acquired under selective pressure from the sun.

UVA1 genotoxicity is mediated not by oxidative damage but by cyclobutane pyrimidine dimers in normal mouse skin

UVA1 induces the formation of 8-hydroxy-2'-deoxyguanosines (8-OH-dGs) and cyclobutane pyrimidine dimers (CPDs) in the cellular genome. However, the relative contribution of each type of damage to the in vivo genotoxicity of UVA1 has not been clarified. We irradiated living mouse skin with 364-nm UVA1 laser light and analyzed the DNA damage formation and mutation induction in the epidermis and dermis. Although dose-dependent increases were observed for both 8-OH-dG and CPD, the mutation induction in the skin was found to result specifically from the CPD formation, based on the induced mutation spectra in the skin genome: the dominance of C --> T transition at a dipyrimidine site. Moreover, these UV-specific mutations occurred preferentially at the 5'-TCG-3' sequence, suggesting that CpG methylation and photosensitization-mediated triplet energy transfer to thymine contribute to the CPD-mediated UVA1 genotoxicity. Thus, it is the CPD formation, not the oxidative stress, that effectively brings about the genotoxicity in normal skin after UVA1 exposure. We also found differences in the responses to the UVA1 genotoxicity between the epidermis and the dermis: the mutation induction after UVA1 irradiation was suppressed in the dermis at all levels of irradiance examined, whereas it leveled off from a certain high irradiance in the epidermis.

Significance of CpG methylation for solar UV-induced mutagenesis and carcinogenesis in skin

Mutations detected in the p53 gene in human nonmelanoma skin cancers show a highly UV-specific mutation pattern, a dominance of C --> T base substitutions at dipyrimidine sites plus frequent CC --> TT tandem substitutions, indicating a major involvement of solar UV in the skin carcinogenesis. These mutations also have another important characteristic of frequent occurrences at CpG dinucleotide sites, some of which actually show prominent hotspots in the p53 gene. Although mammalian solar UV-induced mutation spectra were studied intensively in the aprt gene using rodent cultured cells and the UV-specific mutation pattern was confirmed, the second characteristic of the p53 mutations in human skin cancers had not been reproduced. However, studies with transgenic mouse systems developed thereafter for mutation research, which harbor methyl CpG-abundant transgenes as mutation markers, yielded complete reproductions of the situation of the human skin cancer mutations in terms of both the UV-specific pattern and the frequent occurrence at CpG sites. In this review, we evaluate the significance of the CpG methylation for solar UV mutagenesis in the mammalian genome, which would lead to skin carcinogenesis. We propose that the UV-specific mutations at methylated CpG sites, C --> T transitions at methyl CpG-associated dipyrimidine sites, are a solar UV-specific mutation signature, and have estimated the wavelength range effective for the solar-UV-specific mutation as 310-340 nm. We also recommend the use of methyl CpG-enriched sequences as mutational targets for studies on solar-UV genotoxicity for human, rather than conventional mammalian mutational marker genes such as the aprt and hprt genes.

Control of methylation spreading in synthetic DNA sequences by the murine DNA methyltransferase

Methylation spreading, which involves a propensity for the mammalian DNA-(cytosine-5)-methyltransferase to de novo methylate cytosine-guanine dinucleotides (CpGs) near pre-existing 5-methylcytosine bases, has been implicated in the control of numerous biological processes. We have assessed methylation spreading by the murine DNA methyltransferase in vitro using synthetic copolymers and oligonucleotides which differ only in their methylation state. Double-stranded oligonucleotides were found to undergo higher levels of de novo methylation overall than otherwise identical single-stranded oligonucleotides. This difference reflects the greater number of de novo methylatable cytosine bases in double-stranded than single-stranded sequences. All tested oligonucleotides containing pre-existing 5-methyl-cytosine(s) were de novo methylated at several fold the rates of non-methylated controls. No mammalian proteins besides the DNA methyltransferase were required for this observed enhancement of de novo methylation. Studies using oligonucleotides differing in patterns of pre-methylation showed that methylation spreading can be initiated by hemimethylated or duplex methylated CpGs indicating that recognition of 5-methylcytosine by the enzyme is sufficient to stimulate methylation spreading. Double and single-stranded oligonucleotides with several bases between CpGs underwent considerably more de novo methylation per CpG than sequences containing sequential uninterrupted methylatable sites. Spacing preferences by the DNA methyltransferase were also observed in hemimethylated oligonucleotides, suggesting that this is a general property of the enzyme. Although methylation spreading outside of CpG dinucleotides was relatively rare, single-stranded DNA incurred higher levels of de novo methylation at sites other than CpG as compared to double-stranded DNA. This indicates less specificity of methylation spreading in single-stranded sequences. Finally, enhanced de novo methylation in the presence of fully methylated CpG sites in double-stranded oligonucleotides was not as high as the rates of methylation of hemimethylated CpGs in otherwise identical oligonucleotides. These studies provide further elucidation of the mechanisms and regulation of the methylation spreading process and its potential role in the biological processes it influences.

Repression of transient expression by DNA methylation in transcribed regions of reporter genes introduced into cultured human cells

We developed a convenient method to methylate all CpG dinucleotides in both strands in a selected region of a plasmid, and investigated the effect of DNA methylation in the transcribed regions of reporter genes on the transient expression in HeLa cells. In a construct containing the chloramphenicol acetyltransferase (CAT) gene linked to the SV40 early promoter, methylation in the CAT structural gene repressed CAT activity. Methylation in the transcribed region of the Escherichia coli lacZ gene driven by the human cytomegalovirus immediate early promoter also inhibited expression of beta-galactosidase activity. These results suggest that methylation in the transcribed region as well as promoter methylation may affect transcription.

Mechanisms for methylation-mediated gene silencing and aging

Gene silencing is often mediated by CpG methylation of key protein binding sites within gene regulatory sequences (GRSs). An aging mechanism is proposed based on this gene-silencing phenomenon whereby accumulation over time of methylation within GRSs contributes to cellular senescence. The proposed molecular mechanism for age-related gene silencing is the spreading of methylation through the regulatory sequences of genes resulting in progressive reduction of gene transcription. There is considerable experimental evidence for methylation spreading and its role in gene silencing, but the mechanism responsible for this process has not been elucidated. A four-step mechanism is proposed whereby an original methylation occurs, methyltransferase (MTase) molecules progressively move 5' to 3' from this site, neighboring CpG dinucleotides become methylated, and diminished gene expression ensues. Over time, this process may lead to widespread gene silencing in diverse dividing and nondividing cell types contributing to aging of the organism.

Correlation between DNA methylation and expression of O6-methylguanine-DNA methyltransferase gene in cultured human tumor cells

Approximately 20% of human tumor cell strains are deficient in a DNA repair protein, O6-methylguanine-DNA methyltransferase (MGMT), and are called Mer- strains. In an attempt to determine the molecular basis for the extinction of MGMT expression in Mer- human cells, the distribution of DNA methylation sites in and around the exon sequences of the repair gene was compared in 6 Mer+ (repair-proficient) and 12 Mer- cell lines. Southern blot analysis of the genomic DNA digested with isoschizomeric restriction endonucleases MspI and HpaII to detect 5-methylcytosine in CCGG sequences indicated that the DNA of all the Mer+ cells but of none of the Mer- cells is heavily methylated in the exon-containing regions. The methylation pattern contradicts the general belief that inactive genes are hypermethylated compared to hypomethylation of transcriptionally active genes. It appears that the regulation of the MGMT gene in human cells is much more complex than simply dictated by its methylation level.

Methylated cytosine level in human liver DNA does not decline in aging process

In order to ascertain a generality of the age-dependent decrease in DNA methylation level among different mammalian species, methylated cytosine contents in human liver and spleen DNA at different ages have been determined using high performance liquid chromatography (HPLC). Unexpectedly, the liver DNA revealed no appreciable decline with age while the spleen DNA showed a slight reduction. It indicates that a decrease of methylation level in genomic DNA is not a common denominator of age-related changes in mammals.

Comparison of age-associated changes of c-myc gene methylation in liver between man and mouse

Age-associated changes in DNA methylation of the c-myc gene in liver were compared between humans and mice which have large differences in aging rate. Although the overall methylation profiles of the gene and their age-related changes were found to be similar, the rate of change was much slower in humans than in mice. It suggests that the age-associated alteration of c-myc gene methylation is closely related to biological aging rather than to chronological aging.

In vivo footprint and methylation analysis by PCR-aided genomic sequencing: comparison of active and inactive X chromosomal DNA at the CpG island and promoter of human PGK-1

The promoter region of the X-linked human phosphoglycerate kinase-1 (PGK-1) gene is a CpG island, similar to those often found near autosomal genes. We used ligation-mediated polymerase chain reaction (PCR) for a genomic sequencing study in which 450 bp of the human PGK-1 promoter region was analyzed for the presence of in vivo protein footprints and cytosine methylation at all CpG sites. A technique was devised to selectively visualize the DNA of the inactive X chromosome (Xi), even in the presence of the active X chromosome (Xa). We found that the human Xa in both normal male lymphocytes and hamster-human hybrids is completely unmethylated at all 120 CpG sites. In contrast, 118 of the CpG sites are methylated on the human Xi in hamster-human hybrids. The Xi in normal female lymphocytes is also highly methylated, but some GCG or CGC trinucleotides partially escape methylation; all other CpGs are fully methylated. In vivo footprinting studies with dimethylsulfate (DMS) revealed eight regions of apparent protein-DNA contacts on the Xa. Four of the footprints contained the consensus sequence of the binding site for transcription factor Sp1. The other regions include potential binding sites for transcription factors ATF, NF1, and a CCAAT-binding protein. The Xi did not show any specifically protected sequences, and with the exception of four hyperreactive sites, the in vivo DMS reactivity profile of Xi DNA was very similar to that of purified, linear Xi DNA. The implications of these findings with regard to the maintenance of methylation-free islands, X chromosome inactivation, and the chromatin structure of facultative heterochromatin are discussed.

Genomic sequencing and methylation analysis by ligation mediated PCR

Genomic sequencing permits studies of in vivo DNA methylation and protein-DNA interactions, but its use has been limited because of the complexity of the mammalian genome. A newly developed genomic sequencing procedure in which a ligation mediated polymerase chain reaction (PCR) is used generates high quality, reproducible sequence ladders starting with only 1 microgram of uncloned mammalian DNA per reaction. Different sequence ladders can be created simultaneously by inclusion of multiple primers and visualized separately by rehybridization. Relatively little radioactivity is needed for hybridization and exposure times are short. Methylation patterns in genomic DNA are readily detectable; for example, 17 CpG dinucleotides in the 5' region of human X-linked PGK-1 (phosphoglycerate kinase 1) were found to be methylated on an inactive human X chromosome, but unmethylated on an active X chromosome.