Somatic Cell and Molecular Genetics, Vol. 16, No. 1, t990, pp. 79-84
Stability of DNA Methylation of X-Chromosome Genes during Aging F. Pagani, 1 D. Toniolo, 2 and C. Vergani 3 ~Fondazione Rivetti, Laboratorio di Biochimica e Biologia Molecolare, viate Monte Nero 32, 20135 Milan, Italy; 2Istituto di Genetiea Biochimica ed Evoluzionistica del CNR, Via Abbiategrasso 207, 27100 Pavia, Italy; and Slstituto di Medicina Interna, Universith degli Studi di Milano, via Pace 9, 20122 Milan, Italy" Received 25 September 1989--Final 27 October 1989
Abstract--The stability of DNA methylation during aging was assessed in two groups of young (5-20 years old) and old (85-95 years old) women in DNA from blood teukocytes. Three X-linked genes were investigated. Two, G6PD and GdX, are located on Xq28, on the inactivated portion of the X chromosome: demethylation o f specific regions of both genes was shown previously to be directly correlated with gene reactivation. The third, MIC2, is located on the pseudoautosomal region of the X chromosome and escapes X inactivation. The 5' region of the G6PD and GdX genes and the body of the G6PD, GDX, and MIC2 genes were analyzed with specific DNA probes. No age-related changes in methylation pattern were detected. We can conclude therefore that the methylation pattern of the three X-linked genes is stable during aging in female leukocytes and that a high rate of age-related reactivation of X-linked genes may not be a feature of a l l X-linked loci.
INTRODUCTION
has been reported for the alfafetoprotein gene in old mouse liver (9), and many of the genes studied do not modify their methylation pattern during aging in vivo (10, 11). Several lines of evidence suggest that DNA methylation plays a role in the maintenance of the dosage compensation of X-linked genes, which is achieved through the inactivation of one of the two X chromosomes in each female cell (12). The patterns of DNA methylation do differ for some loci on the active and inactive X chromosome: a correlation has been observed between methylation of CpG islands at the 5' region of some X-linked genes and X inactivation (13-16). Wareham et al. (17) have proposed that X-chromosome reactivation occurs during aging: their studies of female mice heterozygous for a deficiency of ornithine carbamoyl-
DNA methylation has been implicated in the regulation of gene expression of eukaryotes (1). Several experiments indicate that DNA methylation inhibits the expression of viral and eukaryotic genes and that 5azacytidine, an analog of cytosine which causes DNA demethylation, is able to switch on the expression of many genes (2-4). It has been proposed that cellular aging is associated with changes in DNA methylation (5, 6). Methylation of intracisternal A particle DNA in mouse liver has been shown to decrease about fivefold between 6 and 26 months of age (7). A study of the endogenous provirus of mouse mammary tumor virus has shown a decrease in methylation during aging in the mouse spleen (8). However, hypermethylation 79
0740-7750/90/0100-0079506.00/0 © 1990 Plenum Publishing Corporation
80
transferase (OCT) have shown that the number of enzyme-positive hepatocytes increases about 50-fold in mice older than 1 year. In contrast, Migeon et al. (11) have reported a low reactivation frequency of the human X-linked hypoxanthine phosphoribosyltransferase (HPRT) gene in fibroblast clones obtained from women heterozygous at that locus. Furthermore Migeon et al. (11) also reported no age-related changes in DNA methylation of the HPRT locus in blood leukocytes from normal subjects. At the glucose 6-phosphate dehydrogenase (G6PD) locus of the X chromosome, on Xq28, three CpG islands clustered within a DNA region of 100 kb have been identified (18). All three CpG islands are located in the 5' flanking region of housekeeping genes, G6PD, GdX and P3 (t 9), and they are methylated on the inactive X, unmethylated on the active X (14), and demethylated when the locus on the inactive is reexpressed (18, 20). We have made a detailed analysis of the methylation pattern of the G6PD and GDX genes in blood leukocytes of young and old women, as well as of a third gene MIC2, located in the short arm of the X chromosome and which escapes inactivation. Our results show that, in blood leukocytes, the methylation pattern of these genes is stable during aging in vivo. MATERIALS AND METHODS
DNA Analysis. DNA analysis was performed on healthy subjects, 12 young (5-20 years old) and 12 old women (85-95 years old). Blood was collected in 0.1% EDTA, and blood cells were lysed to extract the DNA according to the method of Kunkel et al. (21). DNA (10 gg) was digested with 40 units of each restriction enzyme under the conditions recommended by the manufacturer (Boehringer Mannheim, Mannheim, F.R.G.) and fractionated by electrophoresis in 1-1.5% agarose gel and transferred to Z-bind membrane (Bio-Rad, Richmond, California) by Southern blotting (22). The filters were pre-
Pagani et al.
hybridized at 65°C in 5×SSPE, 0.5% SDS, 100 #g/m1 sonicated heat denaturated salmon sperm DNA, 5 × Denhart's solution, for 5-6 h. Hybridization was performed in the same solution containing 1 × 106 cpm/ml of probe labeled with 32p using the multiprime DNA labeling system (Amersham Inst., Amersham, U.K.). The filters were then washed to a final stringency of 0.3 × SSC at 65°C and autoradiographed with X-ray films and intensifying screens for 1-5 days (23). Densitometric scanning of autoradiographic films was made using a Beckman DU8 Spectrophotometer. The extent of methylation was calculated for each DNA sample as the percentage of the total absorbance in the area of the peaks corresponding to the active chromosome bands: the 0.6-kb band for pTb4-21, the 0.5-kb band for pTb4-24, and the 0.9-kb band for pGD1.4. The values obtained for each DNA sample were pooled for each group and the mean _+ SD was calculated. Statistical analysis was performed using chi-square analysis. DNA Probes. For analysis of the G6PD gene, the filters were hybridized to pTb4-21 and pTb4-24, probes for the 5' flanking region of the gene and pTV3-A, which corresponds to most of the G6PD coding exons to the 3' end (18). For the analysis of the GDX gene, the filters were hybridized to the probes pGD1.4 and pT4.2 Eco/A. pGD1.4 has been previously reported (t4) and corresponds to the 5' region of the gene; pT4.2 Eco/A is a cDNA subclone corresponding to the last exon of the GDX gene (24). For the MIC2 gene, the analysis was performed with the p2B probe (a generous gift of Drs. P. Goodfellow, and C. Mondetlo), which corresponds to the first intron of the gene (25). RESULTS
Methylation of CpG Islands. The analysis was first carried out on the CpG islands at the 5' region of G6PD and GDX genes that have been shown previously to have a methyl-
Stability of DNA Methylation
81
Fig. 1. Methylationof the CpG island of the G6PDgene: Southern blot analysis of leukocyteDNA of young (Y) and old (O) women.All digestionsare HpaII + BamHI exceptin lane 1, which is BamHt only. Lane 2 is a male DNA. The probesused were pTb4-21 for panel A and pTb4-24for panel B, tion pattern correlated with their localization on the active or on the inactive X. Double digestions with BamHI and HpalI of D N A samples followed by hybridization with the probe pTb4-21 (Fig. 1A) showed in males (lane 2) only a band of about 0.6 kb. In female DNAs a larger band of about 0.9 kb was also present because of no cleavage of HpaII sites between the two BamHI sites. A band of about 0.5 kb was observed in male D N A digested with B a m H I / H p a I I and probed with pTb4-24 (Fig. 1B, lane 2); in females larger bands appeared corresponding to partial cleavage of t t p a l I sites in this region, a result of methylation of CpG sites in the CpG island (Fig. 1B). The methylation pattern of the 5' region of the GDX gene was studied in D N A samples digested with PstI and HpaII followed by hybridization with the pGD 1.4 probe (Fig. 2). Male D N A showed a band of about 0.9 kb; female DNAs also showed larger bands of 1.3 and 1.4 kb, and, in some DNAs, minor bands of intermediate length. The D N A bands common to female and male samples, detected with pTb4-21, pTb424, and pGD1.4 probes, have been shown previously to originate from the demethylation of CpG islands of the active gene, and all the larger band(s) originate from the methylated CpG island of the inactive chromosome
(18). We have analyzed D N A of t2 women 5-20 years old and 12 women 85-95 years old; no significant difference in the relative amount of active and inactive chromosome bands was detectable between the two groups analyzed. To search for minor differences, densitometric scanning was performed for each D N A sample and the percent of methylation quantified (Table 1). In the old women group a slight increase in the percent of methylation was detected with this method; however, this difference is not significant. Methylation of the Body of the G6PD and GdX Genes and Flanking Regions. pTV3-A and pT4.2 Eco/A probes were used to analyze the methylation stability of internal D N A regions of the G6PD and GDX genes,
Fig. 2. Methylationof the CpG island of the GDXgene: Southern blot analysis of leukocyteDNA of young (Y) and old (O) women. All digestions are HpaII + PstI except in lane I, which is PstI only. Lane 2 is a male DNA. The probe used was pGD1.4.
82
Pagani et al.
Table 1. Percentage of methylation in young and old women, a Methylation (%) Probes
Young
Old
Genes
pTb4-21 pTb4-24 pDG 1.4
47 ± 4 48 +_ 5 48 _+ 5
48 ± 9 51 ± 6 49 ± 7
G6PD G6PD GDX
aValues are calculated as described in Materials and Methods and are expressed as mean ± SD.
respectively. Neither region analyzed showed methylation differences that could be related to X chromosome inactivation (18). Digestion of the same DNA samples described above with HpalI and EcoRI showed a heterogeneous methylation pattern in both young and old women. The restriction fragments detected by pTV3-A ranged from 1.6 to 5 kb (data not shown) and those detected by pT4.2Eco/A from 1.1 to 7 kb (Fig. 3). Although the intensity of the bands was heterogeneous among DNA samples, no
Fig. 3. Methylation pattern of the body of the GdX gene: Southern blot analysis of leukocyte DNA of young (Y) and old (O) women. All digestions were EcoRI + HpaII except in lane I, which is EcoRI only. Lane 2 is a male DNA. The probe used is pT4.2Eco/A.
Fig. 4. Methylation pattern of the MIC2 gene: Southern blot analysis of leukocyte D N A of young (Y) and old (O) women digested with HpalI and probed with p2B. Lane 1 is a male D N A sample.
changes in band intensity or apparent shift toward higher or lower molecular weight bands were observed between young and old groups. Methylation of the MIC2 Gene. We have analyzed a third DNA region of the X chromosome, the M1C2 gene located on the short arm of the X chromosome. The MIC2 gene is not subjected to X chromosome inactivation; however, the methylation of three HpaII sites in the first intron of the gene was previously reported to be lower on the active X compared to the inactive X (25). This difference in methylation was detected only in blood cells, and variability was observed between different tissues and cell types (25), indicating that methylation of this region of the X chromosome is maintained in a less stringent way. For hybridization experiments, the p2B probe was used, from the first intron of the gene (25). The DNAs were digested with HpaII only: they all showed the expected pattern with bands ranging from 1 to 4.7 kb, similar in all DNAs analyzed. The 4.7-kb band is less represented in male DNA compared to the lower bands (Fig. 4, lane I). This is probably caused by heterogeneity of methylation of Y chromosome sequences (25). No age-related changes in band intensity or apparent shift toward higher or lower molecu-
Stability of DNA Methylation
lar weight bands were observed in female DNAs analyzed. DISCUSSION We report here the analysis of the methylation of three X-linked genes in DNA from blood leukocytes of two groups of young (5-20 years old) and old (85-95 years old) women. Two of the genes, G6PD and GdX, are located on a region of the long arm of the X chromosome that is inactivated in females; the third, MIC2, is located in the short arm and escapes inactivation. The pattern of D N A methylation was compared using the methylation-sensitive restriction enzyme HpaII. In our analysis the intensity of bands corresponding to active genes was compared to that of the rest of the bands by eye or was quantified by densitometric scanning of the autoradiagraphic film (Table 1). No significant change in methylation was observed in any instance. The D N A methylation pattern of the 5' region of G6PD and Gdx genes is highly conserved, and their methylated state is important for the maintenance of the inactive state. However, we have also analyzed the methylation of the body of the G6PD, GdX, and M I C 2 genes, which in all cases studied is much more variable. As expected, the bodies of these three genes showed a heterogeneous methylation pattern in all DNAs, but no specific differences were observed between DNA samples of young and old females. Thus, we can conclude that D N A methylation in aging cells is very stable, at least for the three different X-linked genes studied. The lack of any decrease in the extent of methylation of the CpG islands at the 5' of the G6PD and G d X genes strongly supports the idea that age-related reactivation does not occur in vivo. However, we cannot exclude the possibility that it occurs at a very low rate, not giving rise to methylation differences detectable by our analysis. An alternative possibility is that significant demethylation occurs during aging, but it results in a cellular phenotype that causes cell death or negative selection.
83
The genes we have analyzed and the H P R T gene studied by Migeon et al. (11),
which do not seem to reactivate during aging, are housekeeping genes with CpG islands in their 5' regions. Warehan et al. (t7) have reported a different behavior of the O C T gene. This gene, however, is not a housekeeping gene, does not have CpG islands, or any significant differences in methylation of the active vs the inactive X-chromosome allele (26). We can infer that any of the above differences may be responsible for the different behavior of the O C T gene. Moreover, age-related reactivation of X-linked genes may occur only in certain regions of the X-chromosome and/or housekeeping genes with methylated CpG islands may be more stable in their inactive state. The study of genes with CpG islands in the vicinity of the O C T gene could distinguish between the two possibilities. However, tissue and species specific differences may also be important in age-related reactivation rates. ACKNOWLEDGMENTS We thank Drs. P. Goodfellow and C. Mondello for the gift of the p2B probe. We also thank Mrs. G. Lovati for typing the manuscript. This work was supported in part by a grant from the Associazione Italiana per la Ricerca sul Cancro to D.T. L I T E R A T U R E CITED 1. Cedar,H. (1988). Cell53:3-4. 2. Jones,P.A., and Taylor S.M. (1980). Celt 20:8593. 3. Cooper,D.N. (1983:).Hum. Genet. 64:315-333. 4. Venolia,L., Gartler, S.M, Wassman, E.R., Yen, P., Mohandas, T., and Shapiro, L.J. (1982). Proc. Natl. Acad. Sci. U.S.A. 79:2352-2354. 5. Holliday,R. (1987). Nature 327:661-662. 6. Mays-Hoopes,L.L. (1985). In Molecular Biology of Aging, (eds.) Sohal, R.S., Birnbaum, L.S., and Cutler, R.G. (Raven Press, New York), pp. 49-65. 7. Mays-Hoopes,L.L., Brown,A., and Huang, R.C.C. (1983). MoL Cell. Biol, 3:1371-1380. 8. Etkind, P.R., and Sarkar, N.H. (1983). J. Virol. 45:1t4-123. 9. Papacostantinou,J., and Church, W.K. (1984).Age '7:152-154.
84
10. Vijg, J., Uitterlinden, A.G., and Knook, D.L. (1984). In Topics in Aging Research in Europe, VoL 1, (ed.) Vanbezooijen, C.F.A. (Eurage, Rijswijk), pp. 49-58. 11. Migeon, B.R., Axelman, J., and Beggs, A.H. (1988). Nature 335:93-96. 12. Lyon, M.F. (1972). Biol. Rev. 47:1-35. 13. Wolf, S.F., Jolly, D.J., Lunnen, K.D., Friedmann, T., and Migeon, B.R. (1984). Proc. Natl. Acad. Sci. U.S.A. 81:2806-2810. 14. Toniolo, D., D'Urso, M., Martini, G., Persico, M., Tufano, V., Battistuzzi, G., and Luzzato, L. (1984). EMBO J. 9:1987-1995. 15. Yen, P.N., Patel, P., Chinault, A.C., Mahandas, T., and Shapiro, L.J. (1984). Proc. NatL Aead. ScL U.S.A. 81:1759-1763. 16. Riggs, A.D., Singer-Sam, J., and Keith, D.H. (1985). In Biology and Biochemistry o f DNA Methylation, (eds.) Razin, A., and Cantoni, G.L. (Alan R. Liss, New York), pp. 211-222. 17. Wareham, K.A., Lyon, M.F., Glenister, P.H., and Williams, E.D. (1987). Nature 327:725-727.
Pagani et al.
18. Toniolo, D., Martini, G., Migeon, B.R., and Dono, R. (1988). EMBO J. 7:401-406. 19. Alealay, M., and Toniolo, D. (1989). Nucleic Acids Res. 16:9527-9543. 20. Wolf, S.F., Dintzis, S., Toniolo, D., Persico, G., Lunnen, K.D., Axelman, J., and Migeon, B.R. (1984). Nucleic Acids Res. 12:9333-9348. 21. Kunkel, H.M., Smith, K,D., Boyer, S.H., Digamber, S., Borgaonkar, S.S., Miller, V., Breg, W.R., Jones, H.W., and Rary, J.M. (1977). Proc. Natl. Acad. Sci. U.S.A. 74:1245-1249. 22. Southern, E.M. (t975). J. Mol. BioL 98:503-505. 23. Maniatis, T., Fritsch, E.F., and Sambrook, J. (1982). Molecular Cloning: A Laboratory Manual, (Cold Spring Harbor Laboratory, Cold Spring Harbor, New York). 24. Toniolo, D., Persico, M.G., and Alcalay, M. (1988). Proc. Natl. Acad. Sci. U.S.A. 8:851-855. 25. Mondello, C., Goodfellow, P.J., and Goodfellow, P.N. (1988). Nucleic Acids Res. 16:6813-6824. 26. Mullins, L.J., Veres, G., Caskey, T., and Chapman, V. (1987). Mol. Cell. Biol. 7:39t6-3922.
Stability of DNA Methylation of X-Chromosome Genes during Aging F. Pagani, 1 D. Toniolo, 2 and C. Vergani 3 ~Fondazione Rivetti, Laboratorio di Biochimica e Biologia Molecolare, viate Monte Nero 32, 20135 Milan, Italy; 2Istituto di Genetiea Biochimica ed Evoluzionistica del CNR, Via Abbiategrasso 207, 27100 Pavia, Italy; and Slstituto di Medicina Interna, Universith degli Studi di Milano, via Pace 9, 20122 Milan, Italy" Received 25 September 1989--Final 27 October 1989
Abstract--The stability of DNA methylation during aging was assessed in two groups of young (5-20 years old) and old (85-95 years old) women in DNA from blood teukocytes. Three X-linked genes were investigated. Two, G6PD and GdX, are located on Xq28, on the inactivated portion of the X chromosome: demethylation o f specific regions of both genes was shown previously to be directly correlated with gene reactivation. The third, MIC2, is located on the pseudoautosomal region of the X chromosome and escapes X inactivation. The 5' region of the G6PD and GdX genes and the body of the G6PD, GDX, and MIC2 genes were analyzed with specific DNA probes. No age-related changes in methylation pattern were detected. We can conclude therefore that the methylation pattern of the three X-linked genes is stable during aging in female leukocytes and that a high rate of age-related reactivation of X-linked genes may not be a feature of a l l X-linked loci.
INTRODUCTION
has been reported for the alfafetoprotein gene in old mouse liver (9), and many of the genes studied do not modify their methylation pattern during aging in vivo (10, 11). Several lines of evidence suggest that DNA methylation plays a role in the maintenance of the dosage compensation of X-linked genes, which is achieved through the inactivation of one of the two X chromosomes in each female cell (12). The patterns of DNA methylation do differ for some loci on the active and inactive X chromosome: a correlation has been observed between methylation of CpG islands at the 5' region of some X-linked genes and X inactivation (13-16). Wareham et al. (17) have proposed that X-chromosome reactivation occurs during aging: their studies of female mice heterozygous for a deficiency of ornithine carbamoyl-
DNA methylation has been implicated in the regulation of gene expression of eukaryotes (1). Several experiments indicate that DNA methylation inhibits the expression of viral and eukaryotic genes and that 5azacytidine, an analog of cytosine which causes DNA demethylation, is able to switch on the expression of many genes (2-4). It has been proposed that cellular aging is associated with changes in DNA methylation (5, 6). Methylation of intracisternal A particle DNA in mouse liver has been shown to decrease about fivefold between 6 and 26 months of age (7). A study of the endogenous provirus of mouse mammary tumor virus has shown a decrease in methylation during aging in the mouse spleen (8). However, hypermethylation 79
0740-7750/90/0100-0079506.00/0 © 1990 Plenum Publishing Corporation
80
transferase (OCT) have shown that the number of enzyme-positive hepatocytes increases about 50-fold in mice older than 1 year. In contrast, Migeon et al. (11) have reported a low reactivation frequency of the human X-linked hypoxanthine phosphoribosyltransferase (HPRT) gene in fibroblast clones obtained from women heterozygous at that locus. Furthermore Migeon et al. (11) also reported no age-related changes in DNA methylation of the HPRT locus in blood leukocytes from normal subjects. At the glucose 6-phosphate dehydrogenase (G6PD) locus of the X chromosome, on Xq28, three CpG islands clustered within a DNA region of 100 kb have been identified (18). All three CpG islands are located in the 5' flanking region of housekeeping genes, G6PD, GdX and P3 (t 9), and they are methylated on the inactive X, unmethylated on the active X (14), and demethylated when the locus on the inactive is reexpressed (18, 20). We have made a detailed analysis of the methylation pattern of the G6PD and GDX genes in blood leukocytes of young and old women, as well as of a third gene MIC2, located in the short arm of the X chromosome and which escapes inactivation. Our results show that, in blood leukocytes, the methylation pattern of these genes is stable during aging in vivo. MATERIALS AND METHODS
DNA Analysis. DNA analysis was performed on healthy subjects, 12 young (5-20 years old) and 12 old women (85-95 years old). Blood was collected in 0.1% EDTA, and blood cells were lysed to extract the DNA according to the method of Kunkel et al. (21). DNA (10 gg) was digested with 40 units of each restriction enzyme under the conditions recommended by the manufacturer (Boehringer Mannheim, Mannheim, F.R.G.) and fractionated by electrophoresis in 1-1.5% agarose gel and transferred to Z-bind membrane (Bio-Rad, Richmond, California) by Southern blotting (22). The filters were pre-
Pagani et al.
hybridized at 65°C in 5×SSPE, 0.5% SDS, 100 #g/m1 sonicated heat denaturated salmon sperm DNA, 5 × Denhart's solution, for 5-6 h. Hybridization was performed in the same solution containing 1 × 106 cpm/ml of probe labeled with 32p using the multiprime DNA labeling system (Amersham Inst., Amersham, U.K.). The filters were then washed to a final stringency of 0.3 × SSC at 65°C and autoradiographed with X-ray films and intensifying screens for 1-5 days (23). Densitometric scanning of autoradiographic films was made using a Beckman DU8 Spectrophotometer. The extent of methylation was calculated for each DNA sample as the percentage of the total absorbance in the area of the peaks corresponding to the active chromosome bands: the 0.6-kb band for pTb4-21, the 0.5-kb band for pTb4-24, and the 0.9-kb band for pGD1.4. The values obtained for each DNA sample were pooled for each group and the mean _+ SD was calculated. Statistical analysis was performed using chi-square analysis. DNA Probes. For analysis of the G6PD gene, the filters were hybridized to pTb4-21 and pTb4-24, probes for the 5' flanking region of the gene and pTV3-A, which corresponds to most of the G6PD coding exons to the 3' end (18). For the analysis of the GDX gene, the filters were hybridized to the probes pGD1.4 and pT4.2 Eco/A. pGD1.4 has been previously reported (t4) and corresponds to the 5' region of the gene; pT4.2 Eco/A is a cDNA subclone corresponding to the last exon of the GDX gene (24). For the MIC2 gene, the analysis was performed with the p2B probe (a generous gift of Drs. P. Goodfellow, and C. Mondetlo), which corresponds to the first intron of the gene (25). RESULTS
Methylation of CpG Islands. The analysis was first carried out on the CpG islands at the 5' region of G6PD and GDX genes that have been shown previously to have a methyl-
Stability of DNA Methylation
81
Fig. 1. Methylationof the CpG island of the G6PDgene: Southern blot analysis of leukocyteDNA of young (Y) and old (O) women.All digestionsare HpaII + BamHI exceptin lane 1, which is BamHt only. Lane 2 is a male DNA. The probesused were pTb4-21 for panel A and pTb4-24for panel B, tion pattern correlated with their localization on the active or on the inactive X. Double digestions with BamHI and HpalI of D N A samples followed by hybridization with the probe pTb4-21 (Fig. 1A) showed in males (lane 2) only a band of about 0.6 kb. In female DNAs a larger band of about 0.9 kb was also present because of no cleavage of HpaII sites between the two BamHI sites. A band of about 0.5 kb was observed in male D N A digested with B a m H I / H p a I I and probed with pTb4-24 (Fig. 1B, lane 2); in females larger bands appeared corresponding to partial cleavage of t t p a l I sites in this region, a result of methylation of CpG sites in the CpG island (Fig. 1B). The methylation pattern of the 5' region of the GDX gene was studied in D N A samples digested with PstI and HpaII followed by hybridization with the pGD 1.4 probe (Fig. 2). Male D N A showed a band of about 0.9 kb; female DNAs also showed larger bands of 1.3 and 1.4 kb, and, in some DNAs, minor bands of intermediate length. The D N A bands common to female and male samples, detected with pTb4-21, pTb424, and pGD1.4 probes, have been shown previously to originate from the demethylation of CpG islands of the active gene, and all the larger band(s) originate from the methylated CpG island of the inactive chromosome
(18). We have analyzed D N A of t2 women 5-20 years old and 12 women 85-95 years old; no significant difference in the relative amount of active and inactive chromosome bands was detectable between the two groups analyzed. To search for minor differences, densitometric scanning was performed for each D N A sample and the percent of methylation quantified (Table 1). In the old women group a slight increase in the percent of methylation was detected with this method; however, this difference is not significant. Methylation of the Body of the G6PD and GdX Genes and Flanking Regions. pTV3-A and pT4.2 Eco/A probes were used to analyze the methylation stability of internal D N A regions of the G6PD and GDX genes,
Fig. 2. Methylationof the CpG island of the GDXgene: Southern blot analysis of leukocyteDNA of young (Y) and old (O) women. All digestions are HpaII + PstI except in lane I, which is PstI only. Lane 2 is a male DNA. The probe used was pGD1.4.
82
Pagani et al.
Table 1. Percentage of methylation in young and old women, a Methylation (%) Probes
Young
Old
Genes
pTb4-21 pTb4-24 pDG 1.4
47 ± 4 48 +_ 5 48 _+ 5
48 ± 9 51 ± 6 49 ± 7
G6PD G6PD GDX
aValues are calculated as described in Materials and Methods and are expressed as mean ± SD.
respectively. Neither region analyzed showed methylation differences that could be related to X chromosome inactivation (18). Digestion of the same DNA samples described above with HpalI and EcoRI showed a heterogeneous methylation pattern in both young and old women. The restriction fragments detected by pTV3-A ranged from 1.6 to 5 kb (data not shown) and those detected by pT4.2Eco/A from 1.1 to 7 kb (Fig. 3). Although the intensity of the bands was heterogeneous among DNA samples, no
Fig. 3. Methylation pattern of the body of the GdX gene: Southern blot analysis of leukocyte DNA of young (Y) and old (O) women. All digestions were EcoRI + HpaII except in lane I, which is EcoRI only. Lane 2 is a male DNA. The probe used is pT4.2Eco/A.
Fig. 4. Methylation pattern of the MIC2 gene: Southern blot analysis of leukocyte D N A of young (Y) and old (O) women digested with HpalI and probed with p2B. Lane 1 is a male D N A sample.
changes in band intensity or apparent shift toward higher or lower molecular weight bands were observed between young and old groups. Methylation of the MIC2 Gene. We have analyzed a third DNA region of the X chromosome, the M1C2 gene located on the short arm of the X chromosome. The MIC2 gene is not subjected to X chromosome inactivation; however, the methylation of three HpaII sites in the first intron of the gene was previously reported to be lower on the active X compared to the inactive X (25). This difference in methylation was detected only in blood cells, and variability was observed between different tissues and cell types (25), indicating that methylation of this region of the X chromosome is maintained in a less stringent way. For hybridization experiments, the p2B probe was used, from the first intron of the gene (25). The DNAs were digested with HpaII only: they all showed the expected pattern with bands ranging from 1 to 4.7 kb, similar in all DNAs analyzed. The 4.7-kb band is less represented in male DNA compared to the lower bands (Fig. 4, lane I). This is probably caused by heterogeneity of methylation of Y chromosome sequences (25). No age-related changes in band intensity or apparent shift toward higher or lower molecu-
Stability of DNA Methylation
lar weight bands were observed in female DNAs analyzed. DISCUSSION We report here the analysis of the methylation of three X-linked genes in DNA from blood leukocytes of two groups of young (5-20 years old) and old (85-95 years old) women. Two of the genes, G6PD and GdX, are located on a region of the long arm of the X chromosome that is inactivated in females; the third, MIC2, is located in the short arm and escapes inactivation. The pattern of D N A methylation was compared using the methylation-sensitive restriction enzyme HpaII. In our analysis the intensity of bands corresponding to active genes was compared to that of the rest of the bands by eye or was quantified by densitometric scanning of the autoradiagraphic film (Table 1). No significant change in methylation was observed in any instance. The D N A methylation pattern of the 5' region of G6PD and Gdx genes is highly conserved, and their methylated state is important for the maintenance of the inactive state. However, we have also analyzed the methylation of the body of the G6PD, GdX, and M I C 2 genes, which in all cases studied is much more variable. As expected, the bodies of these three genes showed a heterogeneous methylation pattern in all DNAs, but no specific differences were observed between DNA samples of young and old females. Thus, we can conclude that D N A methylation in aging cells is very stable, at least for the three different X-linked genes studied. The lack of any decrease in the extent of methylation of the CpG islands at the 5' of the G6PD and G d X genes strongly supports the idea that age-related reactivation does not occur in vivo. However, we cannot exclude the possibility that it occurs at a very low rate, not giving rise to methylation differences detectable by our analysis. An alternative possibility is that significant demethylation occurs during aging, but it results in a cellular phenotype that causes cell death or negative selection.
83
The genes we have analyzed and the H P R T gene studied by Migeon et al. (11),
which do not seem to reactivate during aging, are housekeeping genes with CpG islands in their 5' regions. Warehan et al. (t7) have reported a different behavior of the O C T gene. This gene, however, is not a housekeeping gene, does not have CpG islands, or any significant differences in methylation of the active vs the inactive X-chromosome allele (26). We can infer that any of the above differences may be responsible for the different behavior of the O C T gene. Moreover, age-related reactivation of X-linked genes may occur only in certain regions of the X-chromosome and/or housekeeping genes with methylated CpG islands may be more stable in their inactive state. The study of genes with CpG islands in the vicinity of the O C T gene could distinguish between the two possibilities. However, tissue and species specific differences may also be important in age-related reactivation rates. ACKNOWLEDGMENTS We thank Drs. P. Goodfellow and C. Mondello for the gift of the p2B probe. We also thank Mrs. G. Lovati for typing the manuscript. This work was supported in part by a grant from the Associazione Italiana per la Ricerca sul Cancro to D.T. L I T E R A T U R E CITED 1. Cedar,H. (1988). Cell53:3-4. 2. Jones,P.A., and Taylor S.M. (1980). Celt 20:8593. 3. Cooper,D.N. (1983:).Hum. Genet. 64:315-333. 4. Venolia,L., Gartler, S.M, Wassman, E.R., Yen, P., Mohandas, T., and Shapiro, L.J. (1982). Proc. Natl. Acad. Sci. U.S.A. 79:2352-2354. 5. Holliday,R. (1987). Nature 327:661-662. 6. Mays-Hoopes,L.L. (1985). In Molecular Biology of Aging, (eds.) Sohal, R.S., Birnbaum, L.S., and Cutler, R.G. (Raven Press, New York), pp. 49-65. 7. Mays-Hoopes,L.L., Brown,A., and Huang, R.C.C. (1983). MoL Cell. Biol, 3:1371-1380. 8. Etkind, P.R., and Sarkar, N.H. (1983). J. Virol. 45:1t4-123. 9. Papacostantinou,J., and Church, W.K. (1984).Age '7:152-154.
84
10. Vijg, J., Uitterlinden, A.G., and Knook, D.L. (1984). In Topics in Aging Research in Europe, VoL 1, (ed.) Vanbezooijen, C.F.A. (Eurage, Rijswijk), pp. 49-58. 11. Migeon, B.R., Axelman, J., and Beggs, A.H. (1988). Nature 335:93-96. 12. Lyon, M.F. (1972). Biol. Rev. 47:1-35. 13. Wolf, S.F., Jolly, D.J., Lunnen, K.D., Friedmann, T., and Migeon, B.R. (1984). Proc. Natl. Acad. Sci. U.S.A. 81:2806-2810. 14. Toniolo, D., D'Urso, M., Martini, G., Persico, M., Tufano, V., Battistuzzi, G., and Luzzato, L. (1984). EMBO J. 9:1987-1995. 15. Yen, P.N., Patel, P., Chinault, A.C., Mahandas, T., and Shapiro, L.J. (1984). Proc. NatL Aead. ScL U.S.A. 81:1759-1763. 16. Riggs, A.D., Singer-Sam, J., and Keith, D.H. (1985). In Biology and Biochemistry o f DNA Methylation, (eds.) Razin, A., and Cantoni, G.L. (Alan R. Liss, New York), pp. 211-222. 17. Wareham, K.A., Lyon, M.F., Glenister, P.H., and Williams, E.D. (1987). Nature 327:725-727.
Pagani et al.
18. Toniolo, D., Martini, G., Migeon, B.R., and Dono, R. (1988). EMBO J. 7:401-406. 19. Alealay, M., and Toniolo, D. (1989). Nucleic Acids Res. 16:9527-9543. 20. Wolf, S.F., Dintzis, S., Toniolo, D., Persico, G., Lunnen, K.D., Axelman, J., and Migeon, B.R. (1984). Nucleic Acids Res. 12:9333-9348. 21. Kunkel, H.M., Smith, K,D., Boyer, S.H., Digamber, S., Borgaonkar, S.S., Miller, V., Breg, W.R., Jones, H.W., and Rary, J.M. (1977). Proc. Natl. Acad. Sci. U.S.A. 74:1245-1249. 22. Southern, E.M. (t975). J. Mol. BioL 98:503-505. 23. Maniatis, T., Fritsch, E.F., and Sambrook, J. (1982). Molecular Cloning: A Laboratory Manual, (Cold Spring Harbor Laboratory, Cold Spring Harbor, New York). 24. Toniolo, D., Persico, M.G., and Alcalay, M. (1988). Proc. Natl. Acad. Sci. U.S.A. 8:851-855. 25. Mondello, C., Goodfellow, P.J., and Goodfellow, P.N. (1988). Nucleic Acids Res. 16:6813-6824. 26. Mullins, L.J., Veres, G., Caskey, T., and Chapman, V. (1987). Mol. Cell. Biol. 7:39t6-3922.
