Mutagenesis vol. 18 no. 6 pp. 511-519,
November 2003
© 2003 UK Environmental Mutagen Society/Oxford University Press
UVA induces C
T transitions at methyl-CpG-associated dipyrimidine sites in mouse skin epidermis more frequently than UVB
Department of Cell Biology, Graduate School of Medicine, Tohoku University, Sendai, Japan
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionReferences |
|---|
We studied the kinetics of mutation induction in skin epidermis and dermis of UVA-irradiated transgenic MutaTM mice and analyzed the sequence changes in 80 lacZ transgene mutants from the irradiated epidermis. The mutant frequency increased linearly in both the epidermis and dermis up to 240 kJ/m2 UVA, twice as efficiently in the epidermis as in the dermis, without provoking any inflammatory reactions in the exposed skin. The 83 mutations detected in the UVA-exposed epidermis were dominated by C
T transitions (88%), found almost exclusively at dipyrimidine sites, and specified by four occurrences of CCTT tandem substitutions, suggesting that UV-specific photoproducts induced in DNA have a major role in the genotoxicity. No TG transversions, which have been considered as a UVA signature mutation, and few mutations suggesting the relevance of oxidative damage were recovered in the present study. An analysis of the bases adjacent to the mutated cytosines revealed that the 3'-cytosine of dipyrimidine sites is the preferred target of UVA-induced CT transition. Moreover, CT transitions were induced at dipyrimidine sites associated with CpG much more frequently by UVA than by UVB, forming hotspots at several of these sites. These results suggest that UVA contributes more to the formation of recurrent or hotspot mutations at methylated CpG sites in the mammalian genome than UVB, since methylation of the CpG motif is observed entirely in the lacZ transgenes and is known to enhance the formation of cyclobutane pyrimidine dimers by longer wavelength UV. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
The solar radiation reaching the Earth’s surface includes UV light consisting of the shorter wavelength component UVB (290–320 nm) and the longer component UVA (320–400 nm). The UVB component is known to induce specific DNA damage, like cyclobutane pyrimidine dimers (CPD) and pyrimidine–pyrimidone (6–4) photoproducts (64PP) efficiently in the skin genome (Ley et al., 1977
; Sutherland et al., 1980; Mitchell et al., 1990; Chadwick et al., 1995; Qin et al., 1995) and is primarily responsible for sunlight-induced human skin cancer (Setlow, 1974; De Gruijl and Van der Leun, 1994). UVA induces these UV-specific DNA lesions less efficiently, although its fraction of solar UV is larger than that of UVB (>10-fold), but some contribution to skin carcinogenesis has still been suggested (Stary et al., 1997; De Gruijl, 2002). Risk assessment of UVA is also important in terms of public health since commercially available UVA sources have been widely popularized by the cosmetic, health, sanitation and advertising industries.
Evidence for the relevance of UV-specific DNA damage in UVB-induced skin cancer has been obtained through analyses of mutations in the tumor suppressor gene p53 from human non-melanoma skin cancers (Brash et al., 1991; Ziegler et al., 1993; Daya-Grosjean et al., 1995; Nataraj et al., 1995) and from mouse skin squamous cell carcinomas experimentally induced by UVB (Kress et al., 1992; Kanjilal et al., 1993; Dumaz et al., 1997). In these studies, p53 mutations were detected in more than half of the tumors, and CT transitions at dipyrimidine sites, where pyrimidines adjoin each other in the DNA sequence, were predominant. CCTT tandem substitutions were also specifically induced. These observations indicated that cytosine-containing CPD and/or 64PP induced by UVB are the major causal DNA lesions for the mutations, since dipyrimidines in DNA could be targeted by these UV-specific types of DNA damage.
On the other hand, the contribution of the UVA component to skin cancer induction by sunlight is not so remarkable and is estimated to represent 10–20% of the solar carcinogenic dose (Kelfkens et al., 1990). It has been proved experimentally that UVA alone actually has the ability to induce skin cancer in mouse epidermis (Sterenborg and Van der Leun, 1990; Kelfkens et al., 1991; De Laat et al., 1997), although induction was much less efficient, e.g. four orders of magnitude less efficient at wavelengths longer than 340 nm compared with that at 293 nm, where the maximum carcinogenicity was observed (De Gruijl et al., 1993). Much attention has been paid to what kind of DNA damage caused by UVA irradiation is responsible for the genotoxicity (Stary et al., 1997; De Gruijl, 2002). UVA is known to produce reactive oxygen species (ROS) in irradiated cells which could damage the cellular genome (Tyrrell, 2000). The production of 8-hydroxyguanine, one of the major types of oxidative DNA base damage, was actually reported in skin epidermis of mice irradiated with near UV (Hattori-Nakakuki et al., 1994). However, the formation of CPD and/or 64PP is still detectable at physiologically applicable UVA doses not only in naked or cellular DNA (Tyrrell, 1973; Matsunaga et al., 1991) but also in skin tissue (Freeman et al., 1987; Ley and Fourtanier, 2000), although the direct absorption of photon energy by DNA is extremely low in the wavelength range of UVA. Studies on UVA-induced mutation in mammalian cells and skin have actually caused controversy, since their reported mutation spectra differed remarkably (Drobetsky et al., 1995; Robert et al., 1996; Van Kranen et al., 1997; Persson et al., 2002). Some studies favored a contribution of UV-specific DNA photoproducts (Robert et al., 1996; Van Kranen et al., 1997) whereas others supported a major role of ROS (Persson et al., 2002).
In the present study, we have analyzed the kinetics of mutation induction in the epidermis and dermis of UVA-exposed skin using transgenic MutaTM mice, which have the bacterial lacZ gene on a 
phage shuttle vector as a reporter of mutations (Gossen et al., 1989). We determined the sequence changes in mutant lacZ genes recovered from the epidermis of irradiated skin and analyzed the mutation data obtained in terms of mutation-causing DNA damage, adjacent base effects and frequencies of incidence at specific base positions in the gene. The observed UVA-induced mutation spectrum was also compared with that of UVB-induced mutations that we reported previously (Ikehata et al., 2003). We found that the mutation induced by our UVA source is highly UV specific and occurs even more frequently at 5'-CG-3' dinucleotide (CpG) motif-associated dipyrimidine sites than UVB-induced mutation, suggesting the importance of CpG methylation in UVA-induced skin carcinogenesis in mammals.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Mouse treatment
MutaTM mice (Covance Research Products, Denver, PA) aged 8–12 weeks were shaved and depilated with depilatory cream on their dorsal surface 3 days prior to exposure under anesthesia to UVA from broad wavelength range fluorescent FL20.BLB lamps (peak emission 352 nm; Toshiba, Tokyo, Japan) with filtration through Mylar polyester film (Du Pont, Wilmington, DE), which cuts out wavelengths <310 nm (Figure 1). The UVA dosimetry was performed with a UVX radiometer equipped with a UVX-36 sensor (UVP, San Gabriel, CA). Four weeks after depilation, the mice were depilated again and killed 3 days later. The irradiated area (4 x 6 cm) of skin was excised and separated into the epidermis and dermis with thermolysin (Sigma, St Louis, MO). From each tissue, genomic DNA was extracted as described previously (Ikehata et al., 2001
). Animal husbandry and experiments were conducted according to the Guidelines for Animal Welfare and Experimentation at Tohoku University.
|
Mutation assay
The lacZ transgenes were recovered from the epidermal and dermal genomic DNA as phage particles with in vitro
packaging extract as described (Gunther et al., 1993) and evaluated for mutant frequency (MF) with the phenyl-ß-D-galactopyranoside (P-gal) positive selection method as described previously (Ikehata and Ono, 2002). Briefly, the packaged phages were dispensed into several tubes at an appropriate volume, then absorbed to galE-defective Escherichia coli strain C supplied by the manufacturer (Covance Research Products) by mixing with 0.5 ml of the bacterial suspension and incubation at room temperature for 20 min. A portion of the absorption mixture in each tube was transferred to another tube containing 0.25 ml of the same bacterial suspension, mixed quickly with 6 ml of molten top agar solution (LB medium containing 10 mM MgSO4 and 0.75% agar) and plated onto a 10 cm dish containing 6 ml of 1.5% agar in LB medium in the bottom for phage titer determination. The remaining absorption mixture was used for selection of lacZ mutant phages by adding 6 ml of top agar together with 0.06 ml of 1 M P-gal and plating immediately after the titer plating. After incubation at 37°C overnight, the numbers of plaques appearing on the plates were scored and MFs were calculated by dividing the number of P-gal-resistant plaques by that of the total phage deduced from scoring the titer plates.
The mean MF values of the lacZ gene in unirradiated mouse skin were 1.18 (± 0.30) x 10–4 for the epidermis (n = 14) and 1.38 (± 0.34) x 10–4 for the dermis (n = 16), as reported previously (Ikehata and Ono, 2002). In the present paper these values were used as the background MF and were subtracted from the observed MF values.
DNA sequencing
Mutations in the mutant lacZ genes isolated in the mutation assay were determined by sequencing the whole region containing the entire lacZ coding sequence with an ABI PrismTM 377 DNA sequencer using a BigDyeTM terminator cycle sequencing kit (Applied Biosystems, Foster City, CA) as described previously (Ikehata et al., 2003).
Statistical analysis
The significance of differences between the mutation spectra was tested with the Monte Carlo algorithm of Adams and Skopek (1987) using a computer program developed by Cariello et al. (1994).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Mutation expression time
Previously we studied the time-dependent induction of lacZ transgene mutations in mouse dorsal skin epidermis and dermis exposed to a single dose (0.5 kJ/m2) of UVB and found that MFs are fully expressed 3–7 days after irradiation (Ikehata and Ono, 2002
). We checked the expression efficiency in the same tissues for UVA-induced mutations at 1 or 4 weeks after irradiation at 120 kJ/m2 from Mylar filtered UVA lamps. The MFs observed 7 and 28 days after irradiation were similar in both the epidermis and dermis (data not shown), indicating that the MF is fully expressed within 1 week after UVA exposure.
Dose–response kinetics of mutation induction
We investigated the dose-dependent induction of lacZ mutations in epidermis and dermis 4 weeks after exposing the mouse dorsal skin to a single dose of 60, 120, 180 or 240 kJ/m2 UVA filtered through a Mylar film. For each dose point, four or five mice were used. MFs increased nearly linearly in both the epidermis and dermis at rates of 5.8 (± 1.3) x 10–9 and 2.4 (± 0.9) x 10–9 per J/m2, respectively, as the UVA dose increased (Figure 2A, closed and open circles, respectively), showing that mutation induction was twice as efficient in the epidermis as in the dermis.
|
We also checked the effectiveness of filtration through a Mylar film of the light emitted from our UVA lamps, which cuts the irradiance by 60 and 96% at 315 and 310 nm, respectively (see Figure 1). One mouse was used for each dose and irradiated with 80, 160 or 240 kJ/m2 light from the UVA lamps without filtration and MF in the epidermis was scored 4 weeks later (Figure 2B, triangles). The MF increased with dose nearly linearly, and induction was roughly twice as efficient as that with filtered UVA (Figure 2B, circles), although only 2% of the total irradiance was removed by filtration. This result shows that filtration with a Mylar film is effective in reducing the UVB component contaminating our UVA source.
Interestingly, we observed inflammation in epidermis exposed to non-filtered UVA at high doses of 160 and 240 kJ/m2, and even partial ulceration at 240 kJ/m2. A similar inflammatory reaction was observed for UVB-exposed mouse epidermis (Ikehata and Ono, 2002). On the other hand, filtered UVA induced hardly any visible alterations on the surface of irradiated skin up to 240 kJ/m2.
DNA sequencing analysis of lacZ mutants
We analyzed sequence changes in the lacZ transgene of 80 mutants recovered from the epidermis of four mice (20 mutants/mouse, Table I) irradiated with 240 kJ/m2 Mylar filtered UVA, which increased the MF of the lacZ transgene in the epidermis from the background level of 1.2 x 10–4 to 13.0 x 10–4 (see Figure 2A). In all the mutants analyzed, sequence changes were detected within the coding region. Base substitutions were the major type of mutation, being observed in 77 mutants, including four tandem base substitutions and two multiple base substitutions (Tables I and II). One of the multiple substitution mutants contained two base substitutions (Table I, mouse 2, positions 1809 and 3067) and the other contained three (mouse 4, positions 1371, 1627 and 2374), and in both mutants each pair of base changes occurred at sites separated by at least 250 nt, favoring the independent occurrence of each mutation in these multiple mutants. Frameshifts were detected in the other three mutants (Tables I and II).
|
|
Among the 77 base change mutants, 80 base substitutions (76 single and four tandem) were observed in total (Table II). The most frequent, predominant type was a single C
T transition, observed in 91% of the substitutions (73/80, Table III). A unique tandem substitution, CCTT, was also recovered independently in four mutants (Tables I and III). For the other types of base substitutions, two CA transversions and a TA transversion occurred (Table III).
|
The C
T transitions observed in UVA-irradiated epidermis occurred almost exclusively at dipyrimidine sites (71/73, Table III), suggesting that UVA induces DNA damage preferably at these sites. In addition, the CCTT tandem transitions detected in the present study are known as a signature mutation of UV insult. These results indicate that UV-specific DNA damage, CPD and/or 64PP, was mostly responsible for the mutations induced by our UVA source.
Mutation distribution in the lacZ gene
Figure 3 summarizes the distribution of the UVA-induced epidermal mutations detected in the lacZ coding sequence in comparison with that of UVB-induced epidermal ones (Ikehata et al., 2003). Thirty-nine sites of UVA-induced mutations are shown in the lower half of Figure 3 and 49 sites of UVB-induced mutations in the upper half.
|
For the UVA-induced distribution, three hotspots (positions 1187, 1627 and 2392, numbering starting from the first nucleotide of the initiation codon of the lacZ transgene) were evident and all occurred at dipyrimidine sites associated with CpG, which is the target motif of DNA methylation observed in the vertebrate genome (Razin and Cedar, 1984
). Ten other relatively infrequent recurrent sites (positions 616, 928, 1072, 1090, 1388, 1579, 2374, 2713, 2728 and 2818) were detected, all of which were at dipyrimidine sites and seven of which were also associated with the CpG motif (marked with lollipops in Figure 3). For the UVB-induced distribution, most of the relatively frequent recurrent sites (positions 928, 1072, 1187, 1627 and 2392) occurred at CpG-associated dipyrimidine (PyCpG) sites, and these sites coincided with the UVA-induced recurrent sites, although some of them were less remarkable than the UVA-induced ones (positions 1187, 1627 and 2392). In addition, only half of the recurrent sites detected in the UVB-induced distribution (7/13) were in a CpG motif whereas 10 of the 13 recurrent sites with UVA were. These observations suggest that UVA induces mutations more preferably at sites associated with the CpG motif than UVB does.
Influence of adjacent bases on UV-specific CT mutations
To examine the preferred DNA sequence contexts for UVA-induced CT base substitution, all 1235 cytosines in dipyrimidine sites of both strands of the lacZ coding region were listed and classified by their adjacent bases as shown in Table IV. Among these cytosine sites, there are 791 sites where transitions of the cytosine produce amino acid substitutions or protein synthesis terminations (mutable sites, Table IV). Although only about 10% or less (mutability in context, Table IV) of these mutable sites produced mutations in the present study (detected sites), probably because of the relatively small size of the data set (71 mutations), the differences in the mutability among sequence contexts indicated that UVA-induced CT mutations exclusively preferred the 3'-cytosine of dipyrimidines, especially that of 5'-TC-3'. The contexts containing 5'-TC-3' showed mutabilities of 0.03–0.14, whereas the contexts containing the sequence CC showed mutabilities of 0.01–0.05, except for 5'-ACC-3' and 5'-CCT-3', where no mutations were recovered. Moreover, in those CC containing contexts where CT mutations were detected, the 3'-C was more mutable than the 5'-C (0.02–0.05 versus 0.01–0.03) and most of the 5'-C mutations were more likely to be attributable to the neighboring 5'-TC-3' dipyrimidine (5'-TCC-3') than to CC itself. No mutations were recovered in the contexts containing 5'-CT-3' dipyrimidine except for 5'-TCT-3', which includes the most mutable 5'-TC-3', again indicating the poor mutability of 5'-C.
|
C
T mutations were frequently recovered in 5'-TCG-3' sequence contexts about five times per detected site (average recurrence, Table IV). These CT mutations were recovered repeatedly at 8 of the 10 detected sites with a 5'-TCG-3' context (see Table I). Especially, 19 and 10 mutants with the mutation at positions 1187 and 1627, respectively, were recovered in total from all four animals and seven mutants from three animals at position 2392 (Table I and Figure 3), suggesting that cytosines in 5'-TCG-3' context could be highly vulnerable to UVA-induced mutation in some situations. |
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
MF induction kinetics with UVA dose
MFs increased dose-dependently with increasing UVA exposure in both the epidermis and dermis at rates of 5.8 x 10–9 and 2.4 x 10–9 per J/m2, respectively (see Figure 2A). These induction rates for the epidermis and dermis are about 600- and 300-fold less efficient, respectively, than those with UVB exposure, which are 3.4 x 10–6 per J/m2 for epidermis and 8 x 10–7 per J/m2 for dermis (Ikehata and Ono, 2002
). The difference in the induction efficiency between epidermis and dermis is relatively smaller for UVA than for UVB (2- versus 4-fold), reflecting the fact that UVA penetrates through the epidermis more easily than UVB (Bruls et al., 1984; Sterenborg and van der Leun, 1988). Although UVA is present at 10-fold the level of UVB in solar UV (Stary et al., 1997), these results would indicate that the contribution of UVA to solar UV mutagenesis is still smaller (1/60 for epidermis and 1/30 for dermis) than that of UVB.
We found that UVA increases the MF in both mouse skin epidermis and dermis nearly proportionally to the irradiated dose up to 240 kJ/m2, irrespective of Mylar filtration. In a previous study we showed that the dose-dependent increase in MF is suddenly suppressed in the epidermis at doses >0.5 kJ/m2 UVB (Ikehata and Ono, 2002). These observations suggest that the suppression of mutation induction observed in the epidermis does not occur with UV wavelengths longer than 320 nm. Since the MFs induced with Mylar filtered UVA were not higher than that with 0.5 kJ/m2 UVB (1.7 x 10–3; see Figure 2A) (Ikehata and Ono, 2002), it is possible that the UVA doses used in the present study were not large enough to reveal the suppression of mutation induction. However, unfiltered UVA increased the MF dose-dependently up to 2.7 x 10–3 without any acute suppression beyond the saturation level observed with UVB (Ikehata and Ono, 2002), although it induced inflammation in the exposed skin area at doses of 160 and 240 kJ/m2, as did UVB at doses high enough for the mutation induction suppression (Ikehata and Ono, 2002).
UVA-induced mutation spectrum
UVA is known to induce oxidative damage in DNA as well as UV-specific damage (Kielbassa et al., 1997). The UVA-induced mutations determined in the present study were, however, highly UV specific and contained only two CA transversions, the mutation most indicative of ROS. In a study with human cultured cells transfected with a shuttle vector containing the lacZ' gene as a mutation reporter, the dominance of UV-specific CT transitions at dipyrimidine sites in a UVA-induced mutation spectrum was also reported (Robert et al., 1996). In another study with mammalian cultured cells using the endogenous aprt gene as a mutational target, although the UV-specific CT mutations still represented a considerable population, the most dominant mutation was TG transversion, which was such a unique type of mutation that its occurrence could be regarded as a fingerprint of UVA mutagenesis (Drobetsky et al., 1995). However, no TG transversions were detected in the present study or in the study mentioned above (Robert et al., 1996). In an analysis of p53 mutations in murine skin tumors induced experimentally by UVA, all the mutations detected were UV-specific transitions, although the incidence of p53 mutations was quite low (6/42) compared with that with UVB (Van Kranen et al., 1997). Thus, most studies on UVA mutation spectra suggest that UVA genotoxicity stems from the UV photoproducts CPD and/or 64PP, produced specifically in the DNA by direct absorption of the photon energy. Persson et al. (2002), however, reported that all three p53 mutations detected among 37 single cells from UVA-exposed human skin were CA transversions, suggesting the involvement of oxidative damage.
Since UV light filtered through Mylar film still contains a small residual component of UVB in the range 310–320 nm, it is conceivable that a major part of our UVA mutation spectrum resulted from such contaminating UVB wavelengths. The UVA source used in the present study contained 6 and 13% of the maximum (352 nm) energy fluence at 310 and 320 nm, respectively. The Mylar film allows transmission of 4, 70 and 82% of 310, 320 and 352 nm UV, respectively (Figure 1). The efficiencies of photochemical production of CPD in DNA at 310 and 320 nm are about 350- and 15-fold, respectively, that at 352 nm (Kielbassa et al., 1997). The effectiveness of CPD production resulting from multiplication of the energy fluence ratio, the filter transmittance and the photochemical reaction efficiency is not greatly different (<2-fold) between 310, 320 and 352 nm. On the other hand, the efficiency of production of 8-hydroxyguanine at wavelengths of 310–320 nm is 30- to 300-fold lower than that of CPD, although it is still 10- to 20-fold less efficient at 330–360 nm and always lower up to nearly 400 nm (Kielbassa et al., 1997). Actually, Yoon et al. (2000) failed to detect oxidized base damage in naked or cellular DNA exposed to simulated sunlight, which included all UVA wavelengths as well as some longer UVB wavelengths. Thus, the dominance of the UV-specific mutation in UVA-induced mutation spectra accords well with the photochemistry of the production of CPD and 8-hydroxyguanine. Hence, the contribution of contaminating UVB, if any, would not be so remarkable as to distort the validity of our UVA spectrum.
Comparison of UVA-induced, UVB-induced and background mutations
We previously reported the UVB-induced and background mutation spectra in mouse skin epidermis and dermis (Ikehata et al., 2003) (see Tables II and III). The overall spectra of the UVA-induced, UVB-induced and background mutations in mouse epidermis were quite similar, as shown in Table II, consisting of 95–97% base substitutions and a few frameshifts. However, statistical comparison of the base substitution spectra showed significant differences between UVA- and UVB-induced or between UVA-induced and background mutations (P < 0.01 for either; see Table III).
The base substitution spectra of UVA and the background were different with respect to the ratio of CT transitions (91 versus 62%; see Table III) and in the CT frequency at dipyrimidine sites (0.97 versus 0.62; see Tables I and III). These differences probably reflect a difference in the causative molecular changes in the DNA. As discussed before, the spectrum for UVA strongly suggests that UV-specific base photoproducts like CPD and 64PP are primarily responsible for the DNA damage leading to mutation induced by UVA. On the other hand, several different alterations in DNA could be involved in the background mutations. The major base substitution (CT transition) occurred at dipyrimidine sites at a frequency less than the statistically expected level (0.75) if it was occurring at random, suggesting no relevance of UV-specific DNA damage, but occurred almost exclusively at CpG sites. Because the CpG sites in the lacZ transgenes of MutaTM mice have been shown to be fully methylated (Ikehata et al., 2000), CT transitions in the background mutation spectrum likely originated from spontaneous deamination of methylated cytosines in CpG motifs, as suggested before (Coulondre et al., 1978; Rideout et al., 1990). CA (GT) transversions were also recovered relatively frequently in the background mutations (Table III), suggesting the involvement of oxidative damage. In addition, comparison of the mutation distributions in the lacZ gene revealed that the hotspot distributions were different between UVA-induced and background mutations, with three hotspot sites for UVA (positions 1187, 1627 and 2392) but one prominent hotspot of TC (AG) transitions at position 625 for the background (Ikehata et al., 2003).
In a comparison of the UVA- and the UVB-induced spectra of base substitutions, CT transitions induced by UVA were distinguished from those induced by UVB by the ratio of occurrence at CpG sites (75 versus 40%; see Table III), which are methylated in the lacZ transgenes (Ikehata et al., 2000). Although it is known that methylation strongly enhances CPD production by UVB at CpG sites (Drouin and Therrien, 1997; Tommasi et al., 1997), this reaction can be stimulated by sunlight more than by UVB (Tommasi et al., 1997). These studies suggest that UV wavelengths longer than UVB, i.e. UVA, which is also present in sunlight, can produce more CPD at dipyrimidine sites associated with methylated CpG (Py-mCpG) in the lacZ transgene, consistent with the higher ratio of UVA-induced CT occurrences at CpG sites in our study. Actually, UVA and UVB had similar distributions with respect to position of the frequently recurrent CT mutations, which were located mainly at PyCpG sites, although some of those sites appeared as prominent hotspots only for UVA (see Figure 3).
Another difference in the mutation specificity between UVA and UVB was evident from the adjacent base analyses of the UV-specific CT transitions (see Table IV) (Ikehata et al., 2003). In the case of UVA, CT mutations preferred the 3'-cytosine of affected CC dipyrimidines, as mentioned before, whereas such a preference could not be found in the analysis for UVB (Ikehata et al., 2003). Although the underlying reason for the UVA preference for 3'-cytosines is unknown, it might reflect a difference in the efficiency of 64PP and/or its Dewar valence isomer production between UVA and UVB, since these photoproducts have a directivity in DNA sequences (Taylor and Cohrs, 1987).
The strand bias of occurrence of CT mutations in the lacZ transgene should also be noted. For the background mutation spectrm, CT transitions showed no strand preference, as described previously (Ikehata et al., 2003). On the other hand, the same mutations in UVB- and UVA-irradiated epidermis occurred on the non-coding strand (GA) at 3- and 2-fold more sites, respectively, than on the coding strand (CT) (see Table I) (Ikehata et al., 2003). It is unlikely that these mutation strand biases are due to strand-specific DNA repair, such as transcription-coupled repair (Madhani et al., 1986), since transgenes of bacterial origin cannot be transcribed in mammalian cells. It is more likely that the difference between DNA strands during replication, i.e. the leading strand and the lagging strand, influences the mutation occurrence. Translesional synthesis at sites of UV-induced DNA damage might introduce mutations preferably on one of the daughter strands if the replication fork always proceeds in one direction through the region of the genome where the transgenes are integrated. Actually, unequal fidelity in DNA replication between the leading and lagging strands has been reported (Fijalkowska et al., 1998).
UVA enhances mutation recurrence at methylated CpG sites
CPD formation in DNA due to UVB and solar UV is enhanced at PyCpG sites by methylation of the CpG motif (Drouin and Therrien, 1997; Tommasi et al., 1997). Accordingly, in mammalian mutation studies with transgenes that were shown to be methylated, some of the PymCpG sites were found to be relatively frequent sites of UVB- or solar UV-induced CT mutations (You et al., 1999; You and Pfeifer, 2001; Ikehata et al., 2003).
In our study, we have shown that UVA induces CT transitions at PymCpG sites more frequently than UVB. These transitions in the UVA-exposed epidermis recurred at 12 sites in the lacZ transgene, and 10 of these sites (83%) were associated with a CpG motif, whereas only 58% of the recurrent sites in the UVB-irradiated epidermis (7/12) coincided with a CpG motif (Ikehata et al., 2003). Moreover, some of the CT recurrent sites with a CpG motif (positions 1187, 1627 and 2392) appeared as hotspots much more prominently in the mutation distribution with UVA than with UVB. These results indicate that UVA enhances CT mutation recurrence at PymCpG sites more strongly than UVB and could produce remarkable hotspots. Actually, six of seven CT transitions detected in mouse skin cancer induced experimentally by UVA were found at the same single site in the p53 gene, which is in a PymCpG motif, and which is also the most prominent hotspot in UVB-induced skin cancer (Dumaz et al., 1997; Van Kranen et al., 1997).
Origin of UV-specific CT mutation at methylated CpG
We showed that CT mutations recurred frequently at PymCpG sites on irradiation with UVB and UVA. These results are consistent with the hypothesis that CpG methylation could result in UVB- and solar UV-induced mutation hotspots through the acceleration of CPD formation at those sites (Drouin and Therrien, 1997; Tommasi et al., 1997). However, UV can also induce 64PP and its Dewar valence isomer as DNA damage at dipyrimidine sites, which have the ability to induce base substitutions (LeClerc et al., 1991), especially 3'CT at 5'-TC-3' and CC dipyrimidines (Glickman et al., 1986; Horsfall and Lawrence, 1994). The preference of UVA-induced CT transition for the 3'-cytosine of dipyrimidines (see Table IV) might suggest some involvement of these UV photoproducts. Nevertheless, since dipyrimidines containing a methylated cytosine are known to be refractory to the UV-evoked production of 64PP (Glickman et al., 1986; Pfeifer et al., 1991), the relevance of 64PP and the Dewar isomer should be limited. The production of either photoproduct was actually hardly detectable in naked or cellular DNA and in skin tissue after solar UV irradiation (Qin et al., 1996; Yoon et al., 2000). It was also shown in a study using CPD and 64PP photolyases that CPD, but not 64PP, is overwhelmingly the major damage causing CT transitions at dipyrimidine sites in heavily methylated transgenes from UVB-exposed repair-proficient mammalian cells (You et al., 2001).
Although CPDs are induced more frequently by UVB and solar UV than by UVC at PymCpG sites (Drouin and Therrien, 1997; Tommasi et al., 1997), this alone would not explain the frequent CT mutation occurrence at these sites with UVB, UVA and solar UV irradiation (present study; Dumaz et al., 1997; Van Kranen et al., 1997; You et al., 1999; You and Pfeifer, 2001; Ikehata et al., 2003), since dipyrimidine sites unassociated with a CpG motif still form CPD nearly as frequently with UVB and solar UV as with UVC (Drouin and Therrien, 1997; Tommasi et al., 1997). It is known that CPD formation at cytosine-containing dipyrimidines promotes deamination of the constituent cytosine (Tessman et al., 1994; Barak et al., 1995; Tu et al., 1998; Burger et al., 2003). The resulting uracil residues of the deaminated CPD are thought to induce CT mutations as a template in the following translesional DNA synthesis (Bridges, 1992; Horsfall et al., 1997). If methylation of the C5 position of cytosine stimulates deamination in CPD, frequent CT recurrence would be expected preferably at PymCpG sites. Rapid photochemical deamination of the methylated cytosine with the formation of CPD was actually shown in a study using 5-methylcytosine-containing dinucleotide monophosphates (Douki and Cadet, 1994). Consequently, enhanced formation of 5-methylcytosine-containing CPD by longer wavelength UV and the simultaneous promotion of 5-methylcytosine deamination can stimulate occurrence of CT transitions at PymCpG sites synergistically, resulting in the appearance of more remarkable mutation hotspots for longer UV, as shown in the present study.
Thus, PymCpG sites could be a frequently recurrent site for CT transitions, whereas such frequent mutation occurrences were confined exclusively to 5'-TCG-3' sites rather than 5'-CCG-3' sites in our studies with UVB and UVA (see Table IV) (Ikehata et al., 2003). A similar preference of hotspots or frequently recurrent sites for 5'-TCG-3' was observed in methylated transgenes from solar UV-irradiated murine cells (You and Pfeifer, 2001) and in the p53 gene from skin cancers induced in UVB- and UVA-exposed mice (Dumaz et al., 1997; Van Kranen et al., 1997). However, in the p53 gene from human skin cancers, CT mutation hotspots appeared preferably at 5'-CCG-3' sites rather than at 5'-TCG-3' sites (Ziegler et al., 1993). Although the reason for the difference in the hotspot site preference between mouse and human is obscure, phenotypic selection of cancer cells in human oncogenesis might have a tendency to miss p53 mutants with mutations at 5'-TCG-3' sites. This should be elucidated in future studies.
|
Acknowledgements |
|---|
We thank Ms Y.Shono and Ms Y.Ikeda for experimental assistance, Ms S.Kikuchi for manuscript preparation and Mr B.Bell for reading the manuscript. This work was supported by a Grant-in-Aid for Scientific Research (C) from the Japan Society for the Promotion of Science and by the Toyota High-Tech Research Grant Program.
|
Notes |
|---|
1To whom correspondence should be addressed at: Division of Genome and Radiation Biology, Department of Cell Biology, Graduate School of Medicine, Tohoku University, Seiryo-machi 2-1, Aoba-ku, Sendai 980-8575, Japan. Tel: +81 22 717 8134; Fax: +81 22 717 8136; Email: ikehata{at}mail.tains.tohoku.ac.jp
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
-
Adams,W.T. and Skopek,T.R. (1987) Statistical test for the comparison of samples from mutational spectra. J. Mol. Biol., 194, 391–396.[CrossRef][ISI][Medline]
Barak,Y., Cohen-Fix,O. and Livneh,Z. (1995) Deamination of cytosine-containing pyrimidine photodimers in UV-irradiated DNA. J. Biol. Chem., 270, 24174–24179.
Brash,D.E., Rudolph,J.A., Simon,J.A., Lin,A., McKenna,G.J., Baden,H.P., Halperin,A.J. and Pontén,J. (1991) A role for sunlight in skin cancer: UV-induced p53 mutations in squamous cell carcinoma. Proc. Natl Acad. Sci. USA, 88, 10124–10128.
Bridges,B.A. (1992) Mutagenesis after exposure of bacteria to ultraviolet light and delayed photoreversal. Mol. Gen. Genet., 233, 331–336.[ISI][Medline]
Bruls,W.A.G., Slaper,H., Van der Leun,J.C. and Berrens,L. (1984) Transmission of human epidermis and stratum corneum as a function of thickness in the ultraviolet and visible wavelengths. Photochem. Photobiol., 40, 485–494.[ISI][Medline]
Burger,A., Fix,D., Liu,H., Hay,J. and Bockrath,R. (2003) In vivo deamination of cytosine-containing cyclobutane pyrimidine dimers in E. coli: a feasible part of UV-mutagenesis. Mutat. Res., 522, 145–156.[ISI][Medline]
Cariello,N.F., Piegorsch,W.W., Adams,W.T. and Skopek,T.R. (1994) Computer program for the analysis of mutational spectra: application to p53 mutations. Carcinogenesis, 15, 2281–2285.
Chadwick,C.A., Potten,C.S., Nikaido,O., Matsunaga,T., Proby,C. and Young,A.R. (1995) The detection of cyclobutane thymine dimers, (6–4) photolesions and the Dewar photoisomers in sections of UV-irradiated human skin using specific antibodies and the demonstration of depth penetration effects. J. Photochem. Photobiol. B, 28, 163–170.[CrossRef][Medline]
Coulondre,C., Millar,J.H., Farabaugh,P.J. and Gilbert,W. (1978) Molecular basis of base substitution hotspots in Escherichia coli. Nature, 274, 775–780.[CrossRef][Medline]
Daya-Grosjean,L., Dumaz,N. and Sarasin,A. (1995) The specificity of p53 mutation spectra in sunlight induced human cancers. J. Photochem. Photobiol. B, 28, 115–124.[CrossRef][Medline]
De Gruijl,F.R. (2002) Photocarcinogenesis: UVA vs. UVB radiation. Skin Pharmacol. Appl. Skin Physiol., 15, 316–320.[CrossRef][ISI][Medline]
De Gruijl,F.R., Sterenborg,H.J.C.M., Forbes,P.D., Davies,R.E., Cole,C., Kelfkens,G., Van Weelden,H., Slaper,H. and Van der Leun,J.C. (1993) Wavelength dependence of skin cancer induction by ultraviolet irradiation of albino hairless mice. Cancer Res., 53, 53–60.
De Gruijl,F.R. and Van der Leun,J.C. (1994) Estimate of the wavelength dependency of ultraviolet carcinogenesis in humans and its relevance to the risk assessment of a stratospheric ozone depletion. Health Phys., 67, 319–325.[ISI][Medline]
De Laat,A., Van der Leun,J.C. and De Gruijl,F.R. (1997) Carcinogenesis induced by UVA (365-nm) radiation: the dose-time dependence of tumor formation in hairless mice. Carcinogenesis, 18, 1013–1020.
Douki,T. and Cadet,J. (1994) Formation of cyclobutane dimers and (6–4) photoproducts upon far-UV photolysis of 5-methylcytosine-containing dinucleotide monophosphates. Biochemistry, 33, 11942–11950.[CrossRef][Medline]
Drobetsky,E.A., Turcotte,J. and Châteauneuf,A. (1995) A role for ultraviolet A in solar mutagenesis. Proc. Natl Acad. Sci. USA, 92, 2350–2354.
Drouin,R. and Therrien,J.-P. (1997) UVB-induced cyclobutane pyrimidine dimer frequency correlates with skin cancer mutational hotspots in p53. Photochem. Photobiol., 66, 719–726.[ISI][Medline]
Dumaz,N., Van Kranen,H.J., De Vries,A., Berg,R.J.W., Wester,P.W., Van kreijl,C.F., Sarasin,A., Daya-Grosjean,L. and De Gruijl,F.R. (1997) The role of UV-B light in skin carcinogenesis through the analysis of p53 mutations in squamous cell carcinomas of hairless mice. Carcinogenesis, 18, 897–904.
Fijalkowska,I.J., Jonczyk,P., Tkaczyk,M.M., Bialoskorska,M. and Schaaper,R.M. (1998) Unequal fidelity of leading and lagging strand DNA replication on the Escherichia coli chromosome. Proc. Natl Acad. Sci. USA, 95, 10020–10025.
Freeman,S.E., Gange,R.W., Sutherland,J.C., Matzinger,E.A. and Sutherland,B.M. (1987) Production of pyrimidine dimers in DNA of human skin exposed in situ to UVA radiation. J. Invest. Dermatol., 88, 430–433.[CrossRef][ISI][Medline]
Gossen,J.A., De Leeuw,W.J.F., Tan,C.H.T., Zwarthoff,E.C., Berends,F., Lohman,P.H.M., Knook,D.L. and Vijg,J. (1989) Efficient rescue of integrated shuttle vectors from transgenic mice: a model for studying mutations in vivo. Proc. Natl Acad. Sci. USA, 86, 7971–7975.
Glickman,B.W., Schaaper,R.M., Haseltine,W.A., Dunn,R.L. and Brash,D.E. (1986) The C-C (6–4) UV photoproduct is mutagenic in Escherichia coli. Proc. Natl Acad. Sci. USA, 83, 6945–6949.
Gunther,E.J., Murray,N. and Glazer,P.M. (1993) High efficiency, restriction-deficient in vitro packaging extracts for bacteriophage lambda DNA using a new E.coli lysogen, Nucleic Acids Res., 21, 3903–3904.
Hattori-Nakakuki,Y., Nishigori,C., Okamoto,K., Imamura,S., Hiai,H. and Toyokuni,S. (1994) Formation of 8-hydroxy-2'-deoxyguanosine in epidermis of hairless mice exposed to near-UV. Biochem. Biophys. Res. Commun., 201, 1132–1139.[CrossRef][ISI][Medline]
Horsfall,M.J. and Lawrence,C.W. (1994) Accuracy of replication past the T-C (6–4) adduct. J. Mol. Biol., 235, 465–471.[CrossRef][ISI][Medline]
Horsfall,M.J., Borden,A. and Lawrence,C.W. (1997) Mutagenic properties of the T-C cyclobutane dimer. J. Bacteriol., 179, 2835–2839.
Ikehata,H. and Ono,T. (2002) Mutation induction with UVB in mouse skin epidermis is suppressed in acute high-dose exposure. Mutat. Res., 508, 41–47.[ISI][Medline]
Ikehata,H., Takatsu,M., Saito,Y. and Ono,T. (2000) Distribution of spontaneous CpG-associated G:CA:T mutations in the lacZ gene of MutaTM mice: effects of CpG methylation, the sequence context of CpG sites and severity of mutations on the activity of the lacZ gene product. Environ. Mol. Mutagen., 36, 301–311.[CrossRef][ISI][Medline]
Ikehata,H., Aiba,S., Ozawa,H. and Ono,T. (2001) Thermolysin improves mutation analysis in skin epidermis from ultraviolet light-irradiated MutaTM mouse. Environ. Mol. Mutagen., 38, 55–58.[CrossRef][ISI][Medline]
Ikehata,H., Masuda,T., Sakata,H. and Ono,T. (2003) Analysis of mutation spectra in UVB-induced mouse skin epidermis and dermis: frequent occurrence of CT transition at methylated CpG-associated dipyrimidine sites. Environ. Mol. Mutagen., 41, 280–292.[CrossRef][ISI][Medline]
Kanjilal,S., Pierceall,W.E., Cummings,K.K., Kripke,M.L. and Ananthaswamy,H.N. (1993) High frequency of p53 mutations in ultraviolet radiation-induced murine skin tumors: evidence for strand bias and tumor heterogeneity. Cancer Res., 53, 2961–2964.
Kelfkens,G., De Gruijl,F.R. and Van der Leun,J.C. (1990) Ozone depletion and increase in annual carcinogenic ultraviolet dose. Photochem. Photobiol., 52, 819–823.[ISI][Medline]
Kelfkens,G., De Gruijl,F.R. and Van der Leun,J.C. (1991) Tumorigenesis by short-wave ultraviolet A: papillomas versus squamous cell carcinomas. Carcinogenesis, 12, 1377–1382.
Kielbassa,C., Roza,L. and Epe,B. (1997) Wavelength dependence of oxidative DNA damage induced by UV and visible light. Carcinogenesis, 18, 811–816.
Kress,S., Sutter,C., Strickland,P.T., Mukhtar,H., Schweizer,J. and Schwarz,M. (1992) Carcinogen-specific mutational pattern in the p53 gene in ultraviolet B radiation-induced squamous cell carcinomas of mouse skin. Cancer Res., 52, 6400–6403.
LeClerc,J.E., Borden,A. and Lawrence,C.W. (1991) The thymine-thymine pyrimidine-pyrimidone(6–4) ultraviolet light photoproduct is highly mutagenic and specifically induces 3' thymine-to-cytosine transitions in Escherichia coli. Proc. Natl Acad. Sci. USA, 88, 9685–9689.
Ley,R.D. and Fourtanier,A. (2000) UVAI-induced edema and pyrimidine dimers in murine skin. Photochem. Photobiol., 72, 485–487.[CrossRef][ISI][Medline]
Ley,R.D., Sedita,B.A., Grube,D.D. and Fry,R.J.M. (1977) Induction and persistence of pyrimidine dimers in the epidermal DNA of two strains of hairless mice. Cancer Res., 37, 3243–3248.
Madhani,H.D., Bohr,V.A. and Hanawalt,P.C. (1986) Differential DNA repair in transcriptionally active and inactive proto-oncogenes: c-abl and c-mos. Cell, 45, 417–423.[CrossRef][ISI][Medline]
Matsunaga,T., Hieda,K. and Nikaido,O. (1991) Wavelength dependent formation of thymine dimers and (6–4) photoproducts in DNA by monochromatic ultraviolet light ranging from 150 to 365 nm. Photochem. Photobiol., 54, 403–410.[ISI][Medline]
Mitchell,D.L., Cleaver,J.E. and Epstein,J.H. (1990) Repair of pyrimidine(6–4)pyrimidone photoproducts in mouse skin. J. Invest. Dermatol., 95, 55–59.[CrossRef][ISI][Medline]
Nataraj,A.J., Trent,J.C.,II and Ananthaswamy,H.N. (1995) p53 gene mutations and photocarcinogenesis. Photochem. Photobiol., 62, 218–230.[ISI][Medline]
Persson,Å.E., Edström,D.W., Bäckvall,H., Lundeberg,J., Pontén,F., Ros,A.-M. and Williams,C. (2002) The mutagenic effect of ultraviolet-A1 on human skin demonstrated by sequencing the p53 gene in single keratinocytes. Photodermatol. Photoimmunol. Photomed., 18, 287–293.[CrossRef][ISI][Medline]
Pfeifer,G.P., Drouin,R., Riggs,A.D. and Holmquist,G.P. (1991) In vivo mapping of a DNA adduct at nucleotide resolution: detection of pyrimidine (6–4) pyrimidone photoproducts by ligation-mediated polymerase chain reaction. Proc. Natl Acad. Sci. USA, 88, 1374–1378.
Qin,X., Zhang,S., Oda,H., Nakatsuru,Y., Shimizu,S., Yamazaki,Y., Nikaido,O. and Ishikawa,T. (1995) Quantitative detection of ultraviolet light-induced photoproducts in mouse skin by immunohistochemistry. Jpn. J. Cancer Res., 86, 1041–1048.[CrossRef][ISI]
Qin,X., Zhang,S., Zarkovic,M., Nakatsuru,Y., Shimizu,S., Yamazaki,Y., Oda,H., Nikaido,O. and Ishikawa,T. (1996) Detection of ultraviolet photoproducts in mouse skin exposed to natural sunlight. J. Cancer Res., 87, 685–690.
Razin,A. and Cedar,H. (1984) DNA methylation in eukaryotic cells. Int. Rev. Cytol., 92, 159–185.[ISI][Medline]
Rideout,W.M.,III, Coetzee,G.A., Olumi,A.F. and Jones,P.A. (1990) 5-Methylcytosine as an endogenous mutagen in the human LDL receptor and p53 genes. Science, 249, 1288–1290.
Robert,C., Muel,B., Benoit,A., Dubertret,L., Sarasin,A. and Stary,A. (1996) Cell survival and shuttle vector mutagenesis induced by ultraviolet A and ultraviolet B radiation in a human cell line. J. Invest. Dermatol., 106, 721–728.[CrossRef][ISI][Medline]
Setlow,R.B. (1974) The wavelengths in sunlight effective in producing skin cancer: a theoretical analysis. Proc. Natl Acad. Sci. USA, 71, 3363–3366.
Stary,A., Robert,C. and Sarasin,A. (1997) Deleterious effects of ultraviolet A radiation in human cells. Mutat. Res., 383, 1–8.[ISI][Medline]
Sterenborg,H.J.C.M. and Van der Leun,J.C. (1988) Change in epidermal transmission due to UV-induced hyperplasia in hairless mice: a first approximation of the action spectrum. Photodermatology, 5, 71–82.[ISI][Medline]
Sterenborg,H.J.C.M. and Van der Leun,J.C. (1990) Tumorigenesis by a long wavelength UV-A source. Photochem. Photobiol., 51, 325–330.[ISI][Medline]
Sutherland,B.M., Harber,L.C. and Kochevar,I. (1980) Pyrimidine dimer formation and repair in human skin. Cancer Res., 40, 3181–3185.
Taylor,J.-S. and Cohrs,M.P. (1987) DNA, light and Dewar pyrimidinones: the structure and biological significance of TpT3. J. Am. Chem. Soc., 109, 2834–2835.
Tessman,I., Kennedy,M.A. and Liu,S.-K. (1994) Unusual kinetics of uracil formation in single and double-stranded DNA by deamination of cytosine in cyclobutane pyrimidine dimers. J. Mol. Biol., 235, 807–812.[CrossRef][ISI][Medline]
Tommasi,S., Denissenko,M.F. and Pfeifer,G.P. (1997) Sunlight induces pyrimidine dimers preferentially at 5-methylcytosine bases. Cancer Res., 57, 4727–4730.
Tu,Y., Dammann,R. and Pfeifer,G.P. (1998) Sequence and time-dependent deamination of cytosine bases in UVB-induced cyclobutane pyrimidine dimers in vivo. J. Mol. Biol., 284, 297–311.[CrossRef][ISI][Medline]
Tyrrell,R.M. (1973) Induction of pyrimidine dimers in bacterial DNA by 365 nm radiation. Photochem. Photobiol., 17, 69–73.[ISI][Medline]
Tyrrell,R.M. (2000) Role for singlet oxygen in biological effects of ultraviolet A radiation. Methods Enzymol., 319, 290–296.[ISI][Medline]
Van Kranen,H.J., De Laat,A., Van de Ven,J., Wester,P.W., De Vries,A., Berg,R.J.W., Van Kreijl,C.F. and De Gruijl,F.R. (1997) Low incidence of p53 mutations in UVA (365-nm)-induced skin tumors in hairless mice. Cancer Res., 57, 1238–1240.
Yoon,J.-H., Lee,C.-S., O’Connor,T.R., Yasui,A. and Pfeifer,G.P. (2000) The DNA damage spectrum produced by simulated sunlight. J. Mol. Biol., 299, 681–693.[CrossRef][ISI][Medline]
You,Y.-H. and Pfeifer,G.P. (2001) Similarities in sunlight-induced mutational spectra of CpG-methylated transgenes and the p53 gene in skin cancer point to an important role of 5-methylcytosine residues in solar UV mutagenesis. J. Mol. Biol., 305, 389–399.[CrossRef][ISI][Medline]
You,Y.-H., Li,C. and Pfeifer,G.P. (1999) Involvement of 5-methylcytosine in sunlight-induced mutagenesis. J. Mol. Biol., 293, 493–503.[CrossRef][ISI][Medline]
You,Y.-H., Lee,D.-H., Yoon,J.-H., Nakajima,S., Yasui,A. and Pfeifer,G.P. (2001) Cyclobutane pyrimidine dimers are responsible for the vast majority of mutations induced by UVB irradiation in mammalian cells. J. Biol. Chem., 276, 44688–44694.