- Email: [email protected]
DNA methylation in genomic imprinting
Mutation Research 386 Ž1997. 131–140
DNA methylation in genomic imprinting Benjamin Tycko
)
Department of Pathology, Columbia UniÕersity College of Physicians and Surgeons, 630 W 168th Street, New York, NY 10032, USA
Abstract As a reversible epigenetic modification which can affect gene expression, DNA methylation has been an attractive candidate for the biochemical mechanism of genomic imprinting. Many correlations in mice and humans link allele-specific DNA methylation to the allele-restricted RNA expression which is the hallmark of imprinted genes. Moreover, abnormal DNA methylation accompanies the pathological functional imprinting of certain human genes on chromosome 11p15.5 in Wilms’ tumors and in the Beckwith-Weidemann syndrome and on chromosome 15q11-13 in the Prader-Willi and Angelman syndromes. A role for DNA methylation in maintaining the transcriptional silence of imprinted alleles at some loci has been supported by pharmacological manipulation with 5-aza-2X-deoxycytidine and by experiments with methyltransferase deletion mice. Gametic differences in DNA methylation could also account for the initiation of imprints, but this remains unproven. Comprehensive physical models for the role of DNA methylation in imprinting must account not only for local allele-restricted gene expression but also for the existence of large chromosomal domains containing multiple coordinately imprinted genes. Keywords: DNA methyltransferase; Genomic imprinting; Chromosome 11p15.5; Wilms’ tumor
1. A priori considerations Genomic imprinting is the reversible molecular marking of the genome prior to fertilization such that maternal and paternal alleles of certain genes are differentially expressed in various tissues in the offspring. The transcriptionally silenced allele is usually referred to as the functionally imprinted allele, although the primary imprint might in some cases be carried by the active allele. Imprinting is fully reversible Ža paternally imprinted gene is reactivated if passed by a female to the next generation. and therefore, barring speculative scenarios such as precisely reversible DNA inversions, the imprinting pro-
)
E-mail: [email protected]
cess must be epigenetic. Any candidate biochemical modification of the DNA or chromatin which is invoked to explain imprinting must therefore meet four criteria: Ži. it must be reversible; Žii. it must be able to silence gene expression; Žiii. it must be stably propagated in dividing somatic cells; and Živ. it must differ in the male and female gametes. As documented in the other articles in this issue, DNA methylation at cytosine residues within CpG dinucleotides amply satisfies each of the first three criteria. It is also well established that the two types of gametes contain DNA with markedly different patterns of CpG-methylation w1–5x. While the bulk of these gametic methylation differences are expected to be erased in a genome-wide wave of demethylation in the preimplantation embryo w4–6x, allelic methylation differences might be preserved at some
1383-5742r97r$17.00 Copyright q 1997 Elsevier Science B.V. All rights reserved. PII S 1 3 8 3 - 5 7 4 2 Ž 9 6 . 0 0 0 4 9 - 8
132
B. Tyckor Mutation Research 386 (1997) 131–140
critical sites, perhaps corresponding to important transcriptional regulatory sequences of imprinted genes. 2. DNA methylation in imprinted transgenes Mammalian genomic imprinting was first recognized in pronuclear transplantation studies which showed abnormal embryonic development of uniparental conceptuses in mice w7–9x, but the first evidence for an association of imprinting with allele-specific DNA methylation came from the analysis of certain transgenic mouse strains. In an experimental paradigm which has seen numerous subsequent applications, genomic DNA from various tissues of these mice was digested with an infrequently cutting non-methylation-sensitive restriction enzyme either with or without a second restriction enzyme which was frequently cutting but methylation-sensitive. The methylation status of the transgene could then be visualized by Southern blotting using the transgene as a probe – digestion of the primary high molecular weight band to smaller fragments by the methylation-sensitive enzyme indicated hypomethylation at CpG-containing restriction sites, while resistance to digestion indicated hypermethylation. The key finding was that the methylation status of the transgene differed markedly depending on the sex of the transmitting parent w10–13x. In two of these transgenic lines the parent-of-origin-specific DNA methylation could be correlated with RNA expression from the transgene and in one line the gene was even found to reactivate on passage through a parent of the opposite sex w13x. The transgene carried by this strain, which contains Myc-IgA fusion sequences, has subsequently been systematically modified in an effort to define the minimal sequences needed for imprinting – a region sufficient for imprinting of this transgene consists of a non-methylatable simple-sequence repeat derived from the IgA switch region w14x. We have found more than 50 contiguous repeats of a similar pentanucleotide repeat sequence ŽTrAGGGC. n , which is identical to the human immunoglobulin switch sequence repeats, located about 7 kb downstream of the 3X border of the endogenous imprinted human H19 gene ŽT. Crenshaw and B. Tycko, unpublished observations.. Cross-species comparisons are in progress to clarify
the functional significance, if any, of these sequences. An interesting recent variation on the theme of genomic imprinting affecting transgenic loci has been the description of naturally occurring mouse mutations resulting from insertion of a retroviral-like intracisternal A-particle ŽIAP. sequence upstream of the agouti gene, in which the resulting agouti overexpression is found to be silenced only after paternal transmission, correlating with CpG-hypermethylation of the IAP long terminal repeat sequence w15,16x. This highlights an aspect of the transgenic mouse system for studying imprinting which may be a source of artifact or alternatively may be indicating something fundamental about the imprinting machinery – that is the similarity of transgene DNA to invasive viral DNA. Some investigators have proposed that it is a viral-like structure, with small direct repeat sequences, which targets loci for imprinting by the same cellular defense system, prominently including the DNA methyltransferase enzyme, which they posit has evolved to inactivate invasive viral DNA w17–19x. In this theory it is further suggested that the repetitive sequence ‘imprinting box’ need not itself be methylatable, but rather must be embedded in a CpG-rich sequence context. Repetitive sequences may function to promote non-B-form DNA which can be a preferred substrate of the DNA methyltransferase in vitro Žw20x, M. Turker and T. Bestor, this volume., but alternatively they could provide binding sites for repressive chromosomal proteins, as in telomeric silencing mediated by RAP1-SIR protein complexes in yeast w21x. As mentioned above, short direct repeats are indeed found in or near several endogenous imprinted genes, but whether this association is statistically or functionally significant remains to be seen. 3. DNA methylation of endogenous imprinted genes With the identification of native imprinted genes in mice w22–24x and subsequently in humans w25–31x an obvious question was whether DNA methylation would correlate with functional imprinting. This was addressed by a modification of the above Southern blotting strategy in which the non-methylation-sensitive restriction enzyme was chosen to recognize a
B. Tyckor Mutation Research 386 (1997) 131–140
restriction fragment length polymorphism which distinguished the two parental alleles or by sensitive assays in which PCR analysis was coupled to predigestion of genomic DNA with methylation-sensitive restriction enzymes. In general, the answers have been affirmative. In late embryonic and post-embryonic tissues the inactive alleles of several imprinted genes, including the mouse and human H19 genes w27,32–35x, the mouse Igf2r gene w36x, the mouse and human Igf2 genes w33,37,38x, the human SNRPN gene w39x, the mouse Xist gene w40,41x and, in less extensive analyses the human ZNF127 gene w42x and the mouse U2af w43x and Kip2 genes w44x are more heavily methylated, particularly in their 5X promoter and first exon regions but at some sights in the body of the gene as well. These data are consistent with a role for DNA methylation in maintaining the transcriptionally silent state of the imprinted alleles of these genes – a role which is also supported by functional data Žsee below.. However, not all functionally imprinted genes have revealed impressive patterns of allele-specific DNA methylation. Notably, while two differentially methylated CpGs were identified, the mouse Igf2 gene was not found to be heavily methylated at numerous other assayable CpG sites in its 5X end w37x. It may be that for some imprinted genes the critical maintenance elements are outside of this region. Indeed, in the case of Igf2 there is good evidence, based on perturbations of imprinting in human tumors Žsee below. and in H19-deletion mice w45x, for long-range control of its expression Õia ‘competition’ for occupancy of a downstream enhancer element which it shares with the H19 gene. Igf2 and H19 are oppositely imprinted and according to this model the functional repression of the maternal Igf2 allele may be maintained not by hypermethylation of its promoter region but rather by the exclusion of this region from interaction with the downstream H19 enhancer element; the converse situation applies on the paternal chromosome, with the H19 promoter excluded from the enhancer which is occupied by the active Igf2 promoter. In this model dense CpG-methylation of the H19 promoter is sufficient to account for maintenance of the functional imprinting of both H19 and Igf2. Recent evidence for large domains of coordinate imprinting Žsee below. suggests that yet longer-range mecha-
133
nisms of gene regulation will have to be included in complete models of imprinting. Another imprinted gene which does not show extensive hypermethylation of the repressed allele is human KIP2. While the possibility of interspersed sites of allele-specific methylation has not been ruled out, numerous CpG dinucleotides throughout this gene and its 5X upstream region are unmethylated on both alleles w46x. Interestingly, the imprinting of this gene is ‘leaky’ in that there is significant low-level transcription from the repressed Žpaternally imprinted. allele in many fetal and adult tissues w46–48x. Thus, extensive DNA hypermethylation may be essential for the maintenance of strong but not weak transcriptional repression at imprinted loci. As an interesting practical point, allele-specific DNA methylation at imprinted loci has recently been used as a marker for the identification and cloning of novel imprinted genes w43x. In the restriction landmark genomic scanning ŽRLGS. method two-dimensional gel electrophoresis is used to resolve thousands of restriction fragments which have been isotopically end-labeled after digestion of genomic DNA with an infrequently cutting methylation-sensitive restriction enzyme such as NotI. By comparing the gel patterns obtained from reciprocal F1 mouse crosses it is possible to detect spots which are parent-of-origin-specific. The DNAs in such spots are considered as candidates for being located in or close to imprinted genes. At least one imprinted gene, U2af, was found in the initial application of this method w43x. 4. DNA methylation in human diseases with altered functional imprinting Certain human diseases arise at least in part from the disruption of normal genomic imprinting. The Prader-Willi and Angelman syndromes ŽPWS and AS. are mental retardation syndromes with distinct phenotypes which are in many cases caused by deletions of DNA on chromosome 15q11-13. A role for genomic imprinting in producing the distinct phenotypes was raised when it was found that the deleted DNA in the two syndromes was of opposite parental origin: in each case of PWS the deletion had occurred on the paternal chromosome 15, while for each case of AS it had occurred on the maternal
134
B. Tyckor Mutation Research 386 (1997) 131–140
homologue w49x. Additional evidence for opposite imprinting in the two syndromes was the finding that cases of PWS with maternal disomy for the entire chromosome 15 are fairly frequent w50x and that rare cases of AS can be caused by paternal disomy of this same chromosome w51x. Mapping of the minimal deleted regions has shown that PWS and AS are caused by losses of two very closely linked but distinct genes, or most likely a gene cluster for PWS, which are oppositely imprinted w52,53x. Three maternally imprinted genes in the PWS minimal deleted region, SNRPN w30,31x, ZNF127 w54x, IPW w55x and two less well characterized maternally imprinted transcripts, PAR1 and PAR5 w56x have been identified to date. Which if any of these genes is essential for the PWS clinical phenotype is not yet resolved, but patients with PWS who are of the non-deletion type and who are genotypically biparental nevertheless frequently show a bimaternal pattern of allelic DNA methylation on Southern blot analysis with several different probes spaced throughout the chromosome 15q11-q13 region, i.e., a bimaternal ‘epigenotype’ w42x. In fact, a finding of this pattern, or indeed the simple finding of biallelic SNRPN hypermethylation with a single probe is considered as molecular support for a diagnosis of PWS w39x. The AS gene has yet to be identified but in a ‘mirror image’ pattern to PWS, even genotypically biparental patients with AS can manifest a bipaternal methylation epigenotype at certain chromosome 15q11-q13 markers w39,42x. In a major recent advance, rare but potentially highly informative PWS kindreds have recently been identified in which the disease is caused by small Žless than 60 kb. DNA deletions at the upstream border of SNRPN. In these cases there is altered functional and methylation imprinting not only of SNRPN but also of the other imprinted genes and altered methylation of anonymous DNA markers over more than a megabase of DNA, suggesting that the deletions have disrupted an ‘imprinting center’ w56,57x. According to a chromatin accessibility model described below, cis-acting DNA sequence elements of this type might be analogous to insulators or chromatin boundary elements w58x. Wilms’ tumors ŽWTs. and certain other embryonal tumors such as embryonal rhabdomyosarcomas show frequent loss of heterozygosity for markers on
chromosome 11p15.5. A role for imprinted genes in the etiology of these tumors is strongly suggested by the fact that the lost alleles are always maternal in origin Žreviewed in w59x.. Similarly, the BeckwithWeidemann syndrome ŽBWS. of tissue overgrowth and increased predisposition to embryonal tumors has been linked to chromosome 11p15.5 in some families and maternal transmission of the syndrome suggests a role for imprinted genes Žreviewed in w60x.. Disruption of functional and methylation imprinting of loci on chromosome 11p15.5 in WTs and in some cases of BWS was initially characterized by analysis of two imprinted genes which map to this region, IGF2 and H19. A large proportion of the subset of WTs which have retained 11p15.5 heterozygosity, as well as some but not all patients with BWS, show biallelic expression of IGF2 in tissues which should normally show monoallelic Žimprinted. expression w26,61x. This group of WTs, representing about 30% of cases overall, are uniformly characterized by very low or absent expression of H19, accompanied by biallelic hypermethylation of this gene, and this correlates with pathological biallelic expression of IGF2 w62–64x. The complete inactivation of H19 in WTs associated with biallelic hypermethylation is unlikely to be an epiphenomenon of the neoplastic state since it is highly site-specific Žlocally restricted to the body and promoter region of the H19 gene. w63x and since H19 is also biallelically hypermethylated in nonneoplastic tissues of a small subset of BWS patients w38x as well as in the non-neoplastic kidney parenchyma Žin addition to the neoplastic tissue. in a small subset of sporadic WT patients w63x. Based on the enhancer competition model for interaction of the IGF2 and H19 loci the loss of H19 expression in these tissues, either through a primary pathway of hypermethylation or through some other mechanism with secondary hypermethylation, allows interaction of the IGF2 promoter with the downstream shared enhancer element and thereby causes the observed biallelic IGF2 expression. Indeed, this precise scenario is what has been observed in H19 deletion mice which express IGF2 biallelically w45x. However, the situation in the WTs may not be quite so simple. The fact that the Mash-2 and Kip-2 genes, which map to the region of mouse chromosome 7 which is syntenic with human chromosome 11p15.5,
B. Tyckor Mutation Research 386 (1997) 131–140
are both paternally imprinted w44,65x has suggested that the domain of imprinting in this region may extend well beyond the local circuit of IGF2 r H19. A ubiquitously expressed gene, L23MRP, which maps just 40 kb downstream of human H19 is not subject to functional imprinting in fetal and adult human tissues w66x. If imprinting has not somehow ‘skipped over’ this gene then the interval between H19 and L23MRP may contain the downstream Žtelomeric. boundary of the chromosome 11p15.5 imprinted domain. In the opposite Žcentromeric. direction from H19, the chromosome 11p15.5 KIP2 gene maps an unknown but most likely small distance upstream of IGF2 w67x. There is a reproducible allelic expression bias at this gene locus consistent with ‘partial’ functional imprinting w46–48x, and most WTs with loss of H19 expression show a more variable but nonetheless definite reduction in KIP2 expression w46x. In fact, most WTs can be placed into one of three categories: Ži. those with 11p15.5 LOH and reduced expression of both H19 and KIP2 Ž45% of cases.; Žii. those with retention of heterozygosity, H19 hypermethylation, and reduced expression of both H19 and KIP2 Žabout 30% of cases.; and Žiii. those with normal methylation of H19 and high expression of both H19 and KIP2 Žabout 25% of cases.. Thus, the disruption of imprinting in WTs may be similar to that seen in PWSrAS in the sense that both are ‘domain effects’. The two systems appear to differ in one fundamental respect: the PWSrAS disruption of imprinting phenotype is only manifested after germline transmission of the abnormal chromosome while the WT phenotype originates in somatic tumor precursor cells. At this point we do not understand the molecular basis for the disruption of imprinting and it is not clear whether the two types of pathological imprinting, lack of appropriate imprinting in gametogenesis and somatic disruption of imprinting, will be mechanistically related.
5. Pharmacological and genetic manipulation of DNA methylation and imprinting Consistent with a role for CpG methylation in the maintenance of imprinting, the drug 5-azacytidine Žaza-C., which is a covalent inhibitor of the cytosine
135
Fig. 1. Partial erasure of functional imprinting of the human H19 gene by exposure of tumor cells to 5-azacytidine. A malignant rhabdoid tumor cell line, SM, which expresses low but detectable levels of H19 RNA from a single allele, was exposed to the indicated Žmicromolar. concentrations of the drug for a period of 2 weeks. Reverse transcription-PCR with H19 primers followed by digestion of the PCR products with RsaI ŽZhang and Tycko w25x. shows that the imprinted H19 allele Žlower band. was partially reactivated by this treatment. Parallel analysis of genomic DNAs showed partial erasure of allele-specific CpG-methylation of H19 DNA Ždata not shown.. Cell growth was drastically inhibited after 1 week, probably reflecting changes in the expression of numerous genes regulated by DNA methylation. G s genomic PCR control showing the expected two alleles of equal intensity; the allelic bands are at a higher molecular weight because of the presence of two small introns.
DNA methyltransferase enzyme Žreviewed in w68x., can at least partially erase the functional imprinting of the H19 gene in human tumor cell lines, including a malignant rhabdoid tumor line ŽFig. 1. and an embryonal rhabdomyosarcoma line w46x and in the rhabdomyosarcoma line the increased expression of H19 induced by aza-C was accompanied by the predicted reciprocal down-modulation of IGF2 mRNA w46x. Mice with germline insertional mutagenesis of the methyltransferase gene have decreased global DNA methylation and show embryonic lethality w69x and the homozygous embryos are also deficient in the maintenance of functional imprinting, as indicated by their biallelic expression of H19 and reduced expression of IGF2 w70x. These observations support a functional role for DNA methylation in the maintenance of repression of the imprinted allele of at least some imprinted genes and also suggest the intriguing possibility of using pharmacological agents analogous to aza-C to treat diseases associated with abnormalities of functional imprinting. Indeed, aza-C has been used with some success to treat patients with severe refractory thalassemias, in which the drug causes reexpression
136
B. Tyckor Mutation Research 386 (1997) 131–140
of fetal hemoglobin w71x, but since aza-C is also mutagenic in animals it is not an appropriate drug for treatment of lower-risk conditions like WT or PWSrAS. Clearly, a worthwhile problem for study is the design of non-mutagenic agents capable of promoting DNA demethylation.
6. DNA methylation in the establishment of imprints The findings of allele-specific DNA methylation at imprinted loci in late somatic tissues and the disruption of this functional imprinting by alterations in DNA methylation do not have direct bearing on the question of the role of methylation in establishment of the imprint. A cautionary note is sounded by evidence from the analysis of the process of X-chromosome inactivation, in which widespread DNA methylation follows rather than precedes the initial phase of chromatin-mediated gene silencing w72x. It may be that the initial parental imprint consists of a sperm- or oocyte-specific chromatin feature which subsequently initiates a series of molecular events in cells of the early embryo eventuating in allele-specific DNA methylation and allele-restricted transcription. To date there seems to be no direct evidence bearing on this possibility. In contrast, the DNA methylation status of imprinted genes in the earliest stages of embryonic development has been actively investigated by several groups and the data are consistent with the possibility of a primary function for gametic methylation in establishing the functional imprint. Both the mouse and human H19 genes are more heavily methylated in sperm than in oocytes w27,32– 35x. For the mouse H19 gene PCR assays have indicated that at least one HpaII site roughly 700 bp upstream of the first exon maintains its paternalspecific methylation through the blastocyst stage of development, resisting the genome-wide preimplantation wave of demethylation w35x. Similarly, for the mouse Igf2r gene two HpaII sites in an intron showed maternal-specific methylation which was preserved through early embryonic development w36x. Since these sites were hypermethylated on the active Žmaternal. Igf2r allele it was proposed that some cis-acting imprinting signal sequences might exist on the active rather than the silent allele. Similar studies
have suggested the presence of several CpG sites in the 5X promoter region of the mouse Xist gene which are methylated in oocytes but not in sperms and which remain methylated on the maternal X chromosome in early development, perhaps accounting for the selective expression of Xist RNA from the paternal chromosome in early extraembryonic tissues w40,41x. Functional studies using precise germline deletions or mutations could in principal clarify the role of these or other candidate imprinting signals, but such experiments might be fatally compromised by concomitant effects on binding of transcription factors.
7. DNA methylation and domain effects in imprinting How can a role for DNA methylation in the establishment of imprints be reconciled with the evidence for broad domains of coordinate imprinting? One concise model has been proposed by Paldi et al. based on their study of sex-specific meiotic recombination rates in imprinted chromosomal regions w73x. This study, which expanded on the early observations of Thomas and Rothstein w74x, found that the rates of recombination in the IGF2 r H19 region of chromosome 11p15.5 and in the SNRPN region of chromosome 15q11-q13 were much higher in male than in female meioses, while the sex-specific recombination rates in non-imprinted regions tended to be less biased. Based on these data they illustrated an accessibility model for imprinting similar to that originally suggested by Thomas and Rothstein and shown in modified form in Fig. 2. The high rates of recombination in male gametogenesis were taken to imply an open chromatin structure of the chromosomal regions which were destined for imprinting, while the lower rates of recombination in oogenesis were interpreted as indicating a compacted or inaccessible chromatin structure. The open male structure would allow modification of multiple ‘imprinting boxes’, each associated with a single imprinted gene, either by CpG methylation or by binding of a repressive chromatin protein or both, while the compacted female structure would prevent such modifications by masking the imprinting boxes. Chromosomal boundary elements active in the developing gametes
B. Tyckor Mutation Research 386 (1997) 131–140
137
tent with the evidence for disruption of gametic imprinting by DNA deletions in PWSrAS, though analysis of the DNA sequences which have been deleted in these kindreds might well hold some surprises. How it might apply to somatic disruption of imprinting in WTs is less clear. While deletions in cis-acting regulatory sequences may be awaiting discovery in these tumors, alternative explanations for conversion to the bipaternal chromosome 11p15.5 epigenotype involving direct allelic interactions in tumor precursor cells, such as transfer of the paternal methylation imprint to critical sites on the maternal homologue via the action of the DNA methyltransferase on transiently formed heteroduplex DNA w20x, should also be considered.
8. Conclusions
Fig. 2. The accessibility model for genomic imprinting Žmodified from w73x and w74x..
might determine the boundaries of these chromatin domains and chromatin compaction might be mediated by proteins binding to repetitive DNA elements. Subsequent somatic development might homogenize the chromatin structures and DNA methylation patterns of the two homologues to a large extent, but the methylation imprints at critical sites Žin H19 and SNRPN as well as in other nearby genes. would still be carried by the paternal homologue. What might define the extents of imprinted domains? One possibility is that there might be chromosomal boundary elements, DNA sequences which can insulate active or ‘open’ chromatin from adjacent inactive or ‘closed’ chromatin, flanking imprinted domains. Such elements have been visualized bordering transcriptional ‘puffs’ in Drosophila polytene chromosomes by virtue of their binding to at least one specific chromatin protein w58x. For the chromosome 11p15.5 domain one such boundary might lie between H19 and L23MRP w66,75x. Whether this model will be useful as a framework for explaining the disruption of imprinting in human diseases remains to be seen. It appears to be consis-
DNA methylation at cytosine residues is essential for functional imprinting of at least a subset of imprinted mammalian genes and pharmacological manipulation of DNA methylation might ultimately be useful in correcting pathological imprinting. On the other hand, while the hypothesis that DNA methylation is the biochemical correlate of the primary imprint is attractive a priori and is consistent with all available data, it nonetheless remains unproven. Future progress in understanding the biochemical basis for imprinting may depend on a better understanding of DNA and chromatin structure in the developing gametes and its influence on the methylation patterns of gamete DNA.
Acknowledgements This work was supported by grants CA60765 from the N.I.H. and JFRA-482 from the A.C.S. The experiment shown in Fig. 1 was done by Dr. Thomas Moulton in the author’s laboratory.
References w1x Monk, M., Boubelik, M. and S. Lehnert Ž1987. Temporal and regional changes in DNA methylation in the embryonic, extraembryonic, and germ cell lineages during mouse development, Development, 99, 371–382.
138
B. Tyckor Mutation Research 386 (1997) 131–140
w2x Sanford, J.P., Clark, H.J., Chapman, V.M. and J. Rossant Ž1987. Differences in DNA methylation during oogenesis and spermatogenesis and their persistence during early embryogenesis in the mouse, Genes Dev., 1, 1039–1046. w3x Driscoll, D.J. and B.R. Migeon Ž1990. Sex difference in methylation of single-copy genes in human meiotic germ cells: implications for X chromosome inactivation, parental imprinting, and origin of CpG mutations, Somatic Cell Mol. Genet., 16, 267–282. w4x Howlett, S.K. and W. Reik Ž1991. Methylation levels of maternal and paternal genomes during preimplantation development, Development, 113, 119–127. w5x Kafri, T., Ariel, M., Brandeis, M., Shemer, R., Urven, L., McCarrey, J., Cedar, H. and A. Razin Ž1992. Developmental pattern of gene-specific DNA methylation in the mouse embryo and germ line, Genes Dev., 6, 705–714. w6x Frank, D., Keshet, I., Shani, M., Levine, A., Razin, A. and H. Cedar Ž1991. Demethylation of CpG islands in embryonic cells, Nature, 351, 239–241. w7x Surani, M.A.H., Barton, S.C. and M.L. Norrris Ž1984. Development of reconstituted mouse eggs suggests imprinting of the genome during gametogenesis, Nature, 308, 548–550. w8x Barton, S.C., Surani, M.A.H. and Norris, M.L. Ž1984. Role of paternal and maternal genomes in mouse development, Nature, 311, 374–376. w9x McGrath, J. and D. Solter Ž1984. Completion of mouse embryogenesis requires both the maternal and paternal genomes, Cell, 37, 179–183. w10x Reik, W., Collick, A., Norris, M.L., Barton, S.C. and M.A. Surani Ž1987. Genomic imprinting determines methylation of parental alleles in transgenic mice, Nature, 328, 248–251. w11x Sapienza, C., Peterson, A.C., Rossant, J. and R. Balling Ž1987. Degree of methylation of transgenes is dependent on gamete of origin, Nature, 328, 251–254. w12x Hadchouel, M., Farza, H., Simon, D., Tiollais, P. and C. Pourcel Ž1987. Maternal inhibition of hepatitis B surface antigen gene expression in transgenic mice correlates with de novo methylation, Nature, 329, 454–456. w13x Swain, J.L., Stewart, T.A. and P. Leder Ž1987. Parental legacy determines methylation and expression of an autosomal transgene: A molecular mechanism for parental imprinting, Cell, 50, 719–727. w14x Chaillet, R., Bader, D.S. and P. Leder Ž1995. Regulation of genomic imprinting by gametic and embryonic processes, Genes Dev., 9, 1177–1187. w15x Michaud, E.J., van Vugt, M.J., Bultman, S.J., Sweet, H.O., Davisson, M.T. and R.P. Woychik Ž1994. Differential expression of a new dominant agouti allele ŽAiapy. is correlated with methylation state and is influenced by parental lineage, Genes Dev., 8, 1463–1472. w16x Duhl, D.M.J., Vrieling, H., Miller, K.A., Wolff, G.L. and G.S. Barsh Ž1994. Neomorphic agouti mutations in obese yellow mice, Nature Genet., 8, 59–65. w17x Bestor, T.H. Ž1990. DNA methylation: evolution of a bacterial immune function into a regulator of genome structure and activity in higher eukaryotes, Phil. Trans. Royal Soc. Lond. B326, 179–187.
w18x Barlow, D.P. Ž1993. Methylation and imprinting: from host defense to gene regulation?, Science, 260, 309–10. w19x Neumann, B., Kubicka, P. and D.P. Barlow Ž1995. Characteristics of imprinted genes, Nature Genet., 9, 12–13. w20x Bestor, T.H. and B. Tycko Ž1996. Creation of genomic methylation patterns. Nature Genet., 12, 363–367. w21x Moretti, P., Freeman, K., Coodly, L. and D. Shore Ž1994. Evidence that a complex of SIR proteins interacts with the silencer and telomere-binding protein RAP1, Genes Dev., 8, 2257–2269. w22x DeChiara, T.M., Robertson, E.J. and A. Efstratiadis Ž1991. Paternal imprinting of the mouse insulin-like growth factor II gene, Cell, 64, 849–859. w23x Barlow, D,P, Stoger, R., Hermann, B.G., Saito, K. and N. Schweifer Ž1991. The mouse insulin-like growth factor type-2 receptor is imprinted and closely linked to the Tme locus, Nature, 349, 84–87. w24x Bartolomei, M.S., Zemel, S. and S.M. Tilghman Ž1991. Parental imprinting of the mouse H19 gene, Nature, 351, 153–155. w25x Zhang, Y. and B. Tycko Ž1992. Monoallelic expression of the human H19 gene. Nature Genet., 1, 40–44. w26x Rainier, S., Johnson, L., Dobry, C.J., Ping, A.J., Grundy, P.E. and A.P. Feinberg Ž1993. Relaxation of imprinted genes in human cancer, Nature, 362, 747–749. w27x Zhang, Y., Shields, T., Crenshaw, T., Hao, Y., Moulton, T. and B. Tycko Ž1993. Imprinting of human H19: allelespecific CpG methylation, loss of the active allele in Wilms’ tumor and potential for somatic allele switching, Am. J. Hum. Genet., 53, 113–124. w28x Ohlsson, R., Nystrom, A., Pfeifer-Ohlsson, S., Tohonen, V., Hedbourg, F., Schofield, P., Flam, F. and T.J. Ekstrom Ž1993. IGF2 is parentally imprinted during human embryogenesis and in the Beckwith-Wiedemann syndrome, Nature Genet., 4, 94–97. w29x Giannoukakis, N., Deal, C., Paquette, J., Goodyer, C.G. and C. Polychronakos Ž1993. Parental genomic imprinting of the human IGF2 gene, Nature Genet., 4, 98–101. w30x Leff, S.E., Brannan, C.I., Reed, M.L., Ozcelik, T., Francke, U., Copeland, N.J and N.A. Jenkins Ž1992. Maternal imprinting of the mouse Snrpn gene and conserved linkage homology with the human Prader-Willi syndrome region, Nature Genet., 2, 259–264. w31x Glenn, C.C., Porter, K.A., Jong, M.T.C., Nicholls, R.D. and D.J. Driscoll Ž1993. Functional imprinting and epigenetic modification of the human SNRPN gene, Hum. Mol. Genet., 2, 2001–2005. w32x Ferguson-Smith, A.C., Sasaki, H., Cattanach, B.M. and M.A. Surani Ž1993. Parental-origin-specific epigenetic modification of the mouse H19 gene, Nature, 362, 751–755. w33x Brandeis, M., Kafri, T., Ariel, M., Chaillet, J.R., McCarrey, J., Razin, A. and H. Cedar Ž1993. The ontogeny of allelespecific methylation associated with imprinted genes in the mouse, EMBO J., 12, 3669–3677. w34x Bartolomei, M.S., Webber, A.L., Brunkow, M.E. and S.M. Tilghman Ž1993. Epigenetic mechanisms underlying the imprinting of the mouse H19 gene, Genes Dev., 7, 1663–1673.
B. Tyckor Mutation Research 386 (1997) 131–140 w35x Tremblay, K.D., Saam, J.R., Ingram, R.S., Tilghman, S.M. and M.S. Bartolomei Ž1995. A paternal-specific methylation imprint marks the alleles of the mouse H19 gene, Nature Genet., 9, 407–413. w36x Stoger, R., Kubicka, P., Liu, C.-G., Kafri, T., Razin, A., Cedar, H. and D.P. Barlow Ž1993. Maternal-specific methylation of the imprinted Igf2r locus identifies the expressed locus as carrying the imprinting signal, Cell, 73, 61–72. w37x Koide, T., Ainscough, J., Wijgerde, M. and M.A. Surani Ž1994. Comparative analysis of the Igf-2r H19 imprinted domain: identification of a highly conserved intergenic DNase I hypersensitive region, Genomics, 24, 1–8. w38x Reik, W., Brown, K.W., Schneid, H., Le Bouc, Y, Bickmore W. and E.R. Maher Ž1995. Imprinting mutations in the Beckwith-Wiedemann syndrome suggested by an altered imprinting pattern in the IGF2-H19 domain, Hum. Mol. Genet., 4, 2379–2385. w39x Glenn, C.C., Saitoh, S., Jong, M.T.C., Filbrandt, M.M., Surti, U., Driscoll, D.J. and R.D. Nicholls Ž1996. Gene structure, DNA methylation and imprinted expression of the human SNRPN gene. Am. J. Hum. Genet., 58, 335–346. w40x Ariel, M., Robinson, E., McCarrey, J.R. and H. Cedar, Ž1995. Gamete-specific methylation correlates with imprinting of the murine Xist gene, Nature Genet., 9, 312–315. w41x Zuccotti, M. and M. Monk Ž1995. Methylation of the mouse Xist gene in spern and eggs correlates with imprinted Xist expression and paternal X-inactivation, Nature Genet., 9, 316–320. w42x Glenn, C., Nicholls, R.D., Robinson, W.P., Saitoh, S., Niikawa, N., Schinzel, A. Horsthemke, B. and D. Driscoll Ž1993. Modification of 15q11-q13 DNA methylation imprints in unique Angelman and Prader-Willi patients, Hum. Mol. Genet., 2, 1377–1382. w43x Hayashizaki, Y., Shibata, H., Hirotsune, S., Sugino, H., Okazaki, Y., Sasaki, N., Hirose, K., Imoto, H., Okuizumi, H., Muramatsu, M., Komatsubara, H., Shiroishi, T, Moriwaki, K, Katsuki, M., Hatano, N., Sasaki, H., Ueda, T., Mise, N., Takagi, N., Plass, C. and V.M. Chapman Ž1994. Identification of an imprinted U2af binding protein related sequence on mouse chromosome 11 using the RLGS method, Nature Genet., 6, 33–40. w44x Hatada, I. and T. Mukai Ž1995. Genomic imprinting of p57Kip2, a cyclin-dependent kinase inhibitor, in mouse, Nature Genet., 11, 204–206. w45x Leighton, P.A., Ingram, R.S., Eggenschwiler, J., Efstratiatis, A. and S.M. Tilghman Ž1995. Disruption of imprinting caused by deletion of the H19 gene region in mice. Nature, 375, 34–39. w46x Chung, W.-Y., Yuan, L., Feng, L., Hensle, T. and B. Tycko Ž1996. Chromosome 11p15.5 regional imprinting: comparative analysis of KIP2 and H19 in human tissues and Wilms’ tumors. Hum. Mol. Genet., 8, 1101–1108. w47x Kondo, M., Matsuoka, S., Uchida, K., Osada, H., Nagatake, M., Takagi, K., Harper, J.W., Takahashi, T., Elledge, S.J. and T. Takahashi Ž1996. Selective maternal allele loss in human lung cancers of the maternally expressed p57KIP2 gene at 11p15.5. Oncogene, 12, 1365–1368.
139
w48x Matsuoka, S., Thompson, J.S., Edwards, M.C., Barletta, J.M., Grundy, P., Kalikin, L.M., Harper, J.W., Elledge, S.J. and A.P. Feinberg Ž1996. Imprinting of the gene encoding a human cyclin-dependent kinase inhibitor, p57KIP2, on chromosome 11p15. Proc. Natl. Acad. Sci. USA, 93, 3026–3030. w49x Knoll, J.H.M., Nicholls, R.D., Magenis, R.E., Graham, J.M. Jr, Lalande, M. and S.A. Latt Ž1989. Angelman and PraderWilli syndromes share a common chromosome 15 deletion but differ in parental origin of the deletion., Am. J. Med. Genet., 32, 285–290. w50x Nicholls, R.D., Knoll, J.H.M., Butler, M.G., Karam, S. and M. Lalande Ž1989. Genetic imprinting suggested by maternal heterodisomy in non-deletion Prader-Willi syndrome, Nature, 342, 281–285. w51x Malcolm, S., Clayton-Smith, J,. Nichols, M., Robb, S., Webb, T., Armour, J.A.L., Jeffreys, A.J. and M.E. Pembrey Ž1991. Uniparental paternal disomy in Angelman’s syndrome, Lancet, 337, 694–697. w52x Wagstaff, J., Knoll, J.H.M., Glatt, K.A., Shugart, Y.Y., Sommer, A. and M. Lalande Ž1992. Maternal but not paternal transmission of 15q11-13-linked nondeletion Angleman syndrome leads to phenotypic expression, Nature Genet., 1, 291–294 w53x Buiting, K. Dittrich, B., Gross, S., Greger, V., Lalande, M., Robinson, W., Mutirangura, A. et al. Ž1993. Molecular definition of the Prader-Willi syndrome chromosome region and orientation of the SNRPN gene, Hum. Mol. Genet., 2, 1991– 1994. w54x Jong, M.T.C., Carey, A.H., Stewart, C.L., Rinchik, E.M., Glenn, C.C., Driscoll, D.J. and R.D. Nicholls Ž1993. The ZNF127 gene encodes a novel C3HC4 zinc-finger protein and its expression is regulated by genomic imprinting, Am. J. Hum. Genet., 53, Abstr. 697. w55x Wevrick, R., Kerns, J.A. and U. Francke Ž1994. Identification of a novel paternally expressed gene in the Prader-Willi syndrome region, Hum. Mol. Genet., 3, 1877–1882. w56x Sutcliffe, J.S., Nakao, M., Christian, S,. Orstavik, K.H., Tommerup, N., Ledbetter, D.H. and A.L. Beaudet Ž1994. Deletions of a differentially methylated CpG island at the SNRPN gene define a putative imprinting control region, Nature Genet., 8, 52–58. w57x Buiting, K., Saitoh, S., Gross, S., Dittrich, B., Schwartz, S., Nicholls, R.D. and B. Horsthemke Ž1995. Inherited microdeletions in the Angelman and Prader-Willi syndromes define an imprinting centre on human chromosome 15, Nature Genet., 9, 395–400. w58x Zhao, K., Hart, C.M. and U.K. Laemmli Ž1995. Visualization of chromosomal domains with boundary element-associated factor BEAF-32, Cell, 81, 879–889. w59x Tycko, B. Ž1996. Genomic imprinting in Wilms’ tumor. Med. Ped. Oncol., in press. w60x Mannens, M., Hoovers, J.M., Redeker, E., Verjaal, M., Feinberg, A.P., Little, P. et al. Ž1994. Parental imprinting of human chromosome region 11p15.3-pter involved in the Beckwith-Wiedemann syndrome and various human neoplasia, Eur. J. Hum. Genet., 2, 3–23. w61x Weksberg, R., Shen, D.R., Fei, Y.L., Song, Q.L. and J.
140
w62x
w63x
w64x
w65x
w66x
w67x
B. Tyckor Mutation Research 386 (1997) 131–140 Squire Ž1993. Disruption of insulin-like growth factor 2 imprinting in Beckwith-Wiedemann syndrome, Nature Genet., 5, 143–150. Steenman, M.J.C., Rainier, S., Dobry, C.J., Grundy, P., Horon, I.L. and A.P. Feinberg Ž1994. Loss of imprinting of IGF2 is linked to reduced expression and abnormal methylation of H19 in Wilms’ tumor, Nature Genet., 7, 433–439. Moulton, T., Crenshaw, T., Hao, Y., Moosikasuwan, J., Lin, N., Dembitzer, F., Hensle, T., Weiss, L., McMorrow, L., Loew, T., Kraus, W., Gerald, W. and B. Tycko Ž1994. Epigenetic lesions at the H19 locus in Wilms’ tumor patients, Nature Genet., 7, 440–447. Taniguchi, T., Sullivan, M.J., Osamu, O. and A. Reeve Ž1995. Epigenetic changes encompassing the IGF2r H19 locus associated with relaxation of IGF2 imprinting and silencing of H19 in Wilms tumor, Proc. Natl. Acad. Sci. USA, 92, 2159–2163. Guillemot, F., Caspary, T., Tilghman, S.M., Copeland, N.G., Gilbert, D.J., Jenkins, N.A. anderson, D.J.., Joyner, A.L., Rossant, J. and A. Nagy Ž1995. Genomic imprinting of Mash2, a mouse gene required for trophoblast development, Nature Genet., 9, 235–241. Tsang, P., Gilles, F., Yuan, L., Kuo, Y-H., Lupu, F., Samara, G., Moosikasuwan, J., Goy, A., Zelenetz, A.D., Selleri, L. and B. Tycko Ž1995. A novel L23-related gene 40 Kb downstream of the imprinted H19 gene is biallelically expressed in mid-fetal and adult human tissues, Hum. Mol. Genet., 4, 1499–1507. Hoovers, J.M.N., Kalikin, L.M., Johnson, L.A., Alders, M.,
w68x w69x
w70x
w71x
w72x
w73x
w74x w75x
Redeker, B., Law, D.J., Bliek, J, Steenman, M., Benedict, M. Wiegant, J., Lengauer, C. Taillon-Miller P., Schlessinger, D., Edwards, M.C., Elledge, S.J., Ivens, A., Westerveld, A., Little, P. Mannens, M. and A. Feinberg Ž1995. Multiple genetic loci within 11p15 defined by Beckwith-Weidemann syndrome rearrangement breakpoints and subchromosomal transferable fragments, Proc. Natl. Acad. Sci. USA, 92, 12456–12460. Riggs, A.D. and P.A. Jones Ž1983. 5-methylcytosine, gene regulation and cancer, Adv. Cancer Res., 40, 1–28. Li, E., Bestor, T.H. and R. Jaenisch Ž1992. Targeted mutation of the DNA methyltransferase gene results in embryonic lethality, Cell, 69, 915–926. Li, E., Beard C. and R. Jaenisch. Ž1993. Role for DNA methylation in genomic imprinting, JOURNAL, 366, 362– 365. Lowrey, C.H. and A.W. Nienhuis Ž1993. Treatment with azacytidine of patients with end-stage beta-thalassemia, New Engl. J. Med., 329, 845–848. Lock, L.F., Takagi, N. and G.R. Martin Ž1987. Methylation of the Hprt gene on the inactive X occurs after chromosome inactivation, Cell, 48, 39–46. Paldi, A., Gyapay, G. and J. Jami Ž1995. Imprinted chromosomal regions of the human genome display sex-specific meiotic recombination frequencies, Curr. Biol., 5, 1030– 1035. Thomas, B.J. and R. Rothstein Ž1991. Sex, maps, and imprinting, Cell, 64, 1–3. Yuan, L. et al. Ž1996. Human Mol. Genet., 5, 1931–1937.
DNA methylation in genomic imprinting Benjamin Tycko
)
Department of Pathology, Columbia UniÕersity College of Physicians and Surgeons, 630 W 168th Street, New York, NY 10032, USA
Abstract As a reversible epigenetic modification which can affect gene expression, DNA methylation has been an attractive candidate for the biochemical mechanism of genomic imprinting. Many correlations in mice and humans link allele-specific DNA methylation to the allele-restricted RNA expression which is the hallmark of imprinted genes. Moreover, abnormal DNA methylation accompanies the pathological functional imprinting of certain human genes on chromosome 11p15.5 in Wilms’ tumors and in the Beckwith-Weidemann syndrome and on chromosome 15q11-13 in the Prader-Willi and Angelman syndromes. A role for DNA methylation in maintaining the transcriptional silence of imprinted alleles at some loci has been supported by pharmacological manipulation with 5-aza-2X-deoxycytidine and by experiments with methyltransferase deletion mice. Gametic differences in DNA methylation could also account for the initiation of imprints, but this remains unproven. Comprehensive physical models for the role of DNA methylation in imprinting must account not only for local allele-restricted gene expression but also for the existence of large chromosomal domains containing multiple coordinately imprinted genes. Keywords: DNA methyltransferase; Genomic imprinting; Chromosome 11p15.5; Wilms’ tumor
1. A priori considerations Genomic imprinting is the reversible molecular marking of the genome prior to fertilization such that maternal and paternal alleles of certain genes are differentially expressed in various tissues in the offspring. The transcriptionally silenced allele is usually referred to as the functionally imprinted allele, although the primary imprint might in some cases be carried by the active allele. Imprinting is fully reversible Ža paternally imprinted gene is reactivated if passed by a female to the next generation. and therefore, barring speculative scenarios such as precisely reversible DNA inversions, the imprinting pro-
)
E-mail: [email protected]
cess must be epigenetic. Any candidate biochemical modification of the DNA or chromatin which is invoked to explain imprinting must therefore meet four criteria: Ži. it must be reversible; Žii. it must be able to silence gene expression; Žiii. it must be stably propagated in dividing somatic cells; and Živ. it must differ in the male and female gametes. As documented in the other articles in this issue, DNA methylation at cytosine residues within CpG dinucleotides amply satisfies each of the first three criteria. It is also well established that the two types of gametes contain DNA with markedly different patterns of CpG-methylation w1–5x. While the bulk of these gametic methylation differences are expected to be erased in a genome-wide wave of demethylation in the preimplantation embryo w4–6x, allelic methylation differences might be preserved at some
1383-5742r97r$17.00 Copyright q 1997 Elsevier Science B.V. All rights reserved. PII S 1 3 8 3 - 5 7 4 2 Ž 9 6 . 0 0 0 4 9 - 8
132
B. Tyckor Mutation Research 386 (1997) 131–140
critical sites, perhaps corresponding to important transcriptional regulatory sequences of imprinted genes. 2. DNA methylation in imprinted transgenes Mammalian genomic imprinting was first recognized in pronuclear transplantation studies which showed abnormal embryonic development of uniparental conceptuses in mice w7–9x, but the first evidence for an association of imprinting with allele-specific DNA methylation came from the analysis of certain transgenic mouse strains. In an experimental paradigm which has seen numerous subsequent applications, genomic DNA from various tissues of these mice was digested with an infrequently cutting non-methylation-sensitive restriction enzyme either with or without a second restriction enzyme which was frequently cutting but methylation-sensitive. The methylation status of the transgene could then be visualized by Southern blotting using the transgene as a probe – digestion of the primary high molecular weight band to smaller fragments by the methylation-sensitive enzyme indicated hypomethylation at CpG-containing restriction sites, while resistance to digestion indicated hypermethylation. The key finding was that the methylation status of the transgene differed markedly depending on the sex of the transmitting parent w10–13x. In two of these transgenic lines the parent-of-origin-specific DNA methylation could be correlated with RNA expression from the transgene and in one line the gene was even found to reactivate on passage through a parent of the opposite sex w13x. The transgene carried by this strain, which contains Myc-IgA fusion sequences, has subsequently been systematically modified in an effort to define the minimal sequences needed for imprinting – a region sufficient for imprinting of this transgene consists of a non-methylatable simple-sequence repeat derived from the IgA switch region w14x. We have found more than 50 contiguous repeats of a similar pentanucleotide repeat sequence ŽTrAGGGC. n , which is identical to the human immunoglobulin switch sequence repeats, located about 7 kb downstream of the 3X border of the endogenous imprinted human H19 gene ŽT. Crenshaw and B. Tycko, unpublished observations.. Cross-species comparisons are in progress to clarify
the functional significance, if any, of these sequences. An interesting recent variation on the theme of genomic imprinting affecting transgenic loci has been the description of naturally occurring mouse mutations resulting from insertion of a retroviral-like intracisternal A-particle ŽIAP. sequence upstream of the agouti gene, in which the resulting agouti overexpression is found to be silenced only after paternal transmission, correlating with CpG-hypermethylation of the IAP long terminal repeat sequence w15,16x. This highlights an aspect of the transgenic mouse system for studying imprinting which may be a source of artifact or alternatively may be indicating something fundamental about the imprinting machinery – that is the similarity of transgene DNA to invasive viral DNA. Some investigators have proposed that it is a viral-like structure, with small direct repeat sequences, which targets loci for imprinting by the same cellular defense system, prominently including the DNA methyltransferase enzyme, which they posit has evolved to inactivate invasive viral DNA w17–19x. In this theory it is further suggested that the repetitive sequence ‘imprinting box’ need not itself be methylatable, but rather must be embedded in a CpG-rich sequence context. Repetitive sequences may function to promote non-B-form DNA which can be a preferred substrate of the DNA methyltransferase in vitro Žw20x, M. Turker and T. Bestor, this volume., but alternatively they could provide binding sites for repressive chromosomal proteins, as in telomeric silencing mediated by RAP1-SIR protein complexes in yeast w21x. As mentioned above, short direct repeats are indeed found in or near several endogenous imprinted genes, but whether this association is statistically or functionally significant remains to be seen. 3. DNA methylation of endogenous imprinted genes With the identification of native imprinted genes in mice w22–24x and subsequently in humans w25–31x an obvious question was whether DNA methylation would correlate with functional imprinting. This was addressed by a modification of the above Southern blotting strategy in which the non-methylation-sensitive restriction enzyme was chosen to recognize a
B. Tyckor Mutation Research 386 (1997) 131–140
restriction fragment length polymorphism which distinguished the two parental alleles or by sensitive assays in which PCR analysis was coupled to predigestion of genomic DNA with methylation-sensitive restriction enzymes. In general, the answers have been affirmative. In late embryonic and post-embryonic tissues the inactive alleles of several imprinted genes, including the mouse and human H19 genes w27,32–35x, the mouse Igf2r gene w36x, the mouse and human Igf2 genes w33,37,38x, the human SNRPN gene w39x, the mouse Xist gene w40,41x and, in less extensive analyses the human ZNF127 gene w42x and the mouse U2af w43x and Kip2 genes w44x are more heavily methylated, particularly in their 5X promoter and first exon regions but at some sights in the body of the gene as well. These data are consistent with a role for DNA methylation in maintaining the transcriptionally silent state of the imprinted alleles of these genes – a role which is also supported by functional data Žsee below.. However, not all functionally imprinted genes have revealed impressive patterns of allele-specific DNA methylation. Notably, while two differentially methylated CpGs were identified, the mouse Igf2 gene was not found to be heavily methylated at numerous other assayable CpG sites in its 5X end w37x. It may be that for some imprinted genes the critical maintenance elements are outside of this region. Indeed, in the case of Igf2 there is good evidence, based on perturbations of imprinting in human tumors Žsee below. and in H19-deletion mice w45x, for long-range control of its expression Õia ‘competition’ for occupancy of a downstream enhancer element which it shares with the H19 gene. Igf2 and H19 are oppositely imprinted and according to this model the functional repression of the maternal Igf2 allele may be maintained not by hypermethylation of its promoter region but rather by the exclusion of this region from interaction with the downstream H19 enhancer element; the converse situation applies on the paternal chromosome, with the H19 promoter excluded from the enhancer which is occupied by the active Igf2 promoter. In this model dense CpG-methylation of the H19 promoter is sufficient to account for maintenance of the functional imprinting of both H19 and Igf2. Recent evidence for large domains of coordinate imprinting Žsee below. suggests that yet longer-range mecha-
133
nisms of gene regulation will have to be included in complete models of imprinting. Another imprinted gene which does not show extensive hypermethylation of the repressed allele is human KIP2. While the possibility of interspersed sites of allele-specific methylation has not been ruled out, numerous CpG dinucleotides throughout this gene and its 5X upstream region are unmethylated on both alleles w46x. Interestingly, the imprinting of this gene is ‘leaky’ in that there is significant low-level transcription from the repressed Žpaternally imprinted. allele in many fetal and adult tissues w46–48x. Thus, extensive DNA hypermethylation may be essential for the maintenance of strong but not weak transcriptional repression at imprinted loci. As an interesting practical point, allele-specific DNA methylation at imprinted loci has recently been used as a marker for the identification and cloning of novel imprinted genes w43x. In the restriction landmark genomic scanning ŽRLGS. method two-dimensional gel electrophoresis is used to resolve thousands of restriction fragments which have been isotopically end-labeled after digestion of genomic DNA with an infrequently cutting methylation-sensitive restriction enzyme such as NotI. By comparing the gel patterns obtained from reciprocal F1 mouse crosses it is possible to detect spots which are parent-of-origin-specific. The DNAs in such spots are considered as candidates for being located in or close to imprinted genes. At least one imprinted gene, U2af, was found in the initial application of this method w43x. 4. DNA methylation in human diseases with altered functional imprinting Certain human diseases arise at least in part from the disruption of normal genomic imprinting. The Prader-Willi and Angelman syndromes ŽPWS and AS. are mental retardation syndromes with distinct phenotypes which are in many cases caused by deletions of DNA on chromosome 15q11-13. A role for genomic imprinting in producing the distinct phenotypes was raised when it was found that the deleted DNA in the two syndromes was of opposite parental origin: in each case of PWS the deletion had occurred on the paternal chromosome 15, while for each case of AS it had occurred on the maternal
134
B. Tyckor Mutation Research 386 (1997) 131–140
homologue w49x. Additional evidence for opposite imprinting in the two syndromes was the finding that cases of PWS with maternal disomy for the entire chromosome 15 are fairly frequent w50x and that rare cases of AS can be caused by paternal disomy of this same chromosome w51x. Mapping of the minimal deleted regions has shown that PWS and AS are caused by losses of two very closely linked but distinct genes, or most likely a gene cluster for PWS, which are oppositely imprinted w52,53x. Three maternally imprinted genes in the PWS minimal deleted region, SNRPN w30,31x, ZNF127 w54x, IPW w55x and two less well characterized maternally imprinted transcripts, PAR1 and PAR5 w56x have been identified to date. Which if any of these genes is essential for the PWS clinical phenotype is not yet resolved, but patients with PWS who are of the non-deletion type and who are genotypically biparental nevertheless frequently show a bimaternal pattern of allelic DNA methylation on Southern blot analysis with several different probes spaced throughout the chromosome 15q11-q13 region, i.e., a bimaternal ‘epigenotype’ w42x. In fact, a finding of this pattern, or indeed the simple finding of biallelic SNRPN hypermethylation with a single probe is considered as molecular support for a diagnosis of PWS w39x. The AS gene has yet to be identified but in a ‘mirror image’ pattern to PWS, even genotypically biparental patients with AS can manifest a bipaternal methylation epigenotype at certain chromosome 15q11-q13 markers w39,42x. In a major recent advance, rare but potentially highly informative PWS kindreds have recently been identified in which the disease is caused by small Žless than 60 kb. DNA deletions at the upstream border of SNRPN. In these cases there is altered functional and methylation imprinting not only of SNRPN but also of the other imprinted genes and altered methylation of anonymous DNA markers over more than a megabase of DNA, suggesting that the deletions have disrupted an ‘imprinting center’ w56,57x. According to a chromatin accessibility model described below, cis-acting DNA sequence elements of this type might be analogous to insulators or chromatin boundary elements w58x. Wilms’ tumors ŽWTs. and certain other embryonal tumors such as embryonal rhabdomyosarcomas show frequent loss of heterozygosity for markers on
chromosome 11p15.5. A role for imprinted genes in the etiology of these tumors is strongly suggested by the fact that the lost alleles are always maternal in origin Žreviewed in w59x.. Similarly, the BeckwithWeidemann syndrome ŽBWS. of tissue overgrowth and increased predisposition to embryonal tumors has been linked to chromosome 11p15.5 in some families and maternal transmission of the syndrome suggests a role for imprinted genes Žreviewed in w60x.. Disruption of functional and methylation imprinting of loci on chromosome 11p15.5 in WTs and in some cases of BWS was initially characterized by analysis of two imprinted genes which map to this region, IGF2 and H19. A large proportion of the subset of WTs which have retained 11p15.5 heterozygosity, as well as some but not all patients with BWS, show biallelic expression of IGF2 in tissues which should normally show monoallelic Žimprinted. expression w26,61x. This group of WTs, representing about 30% of cases overall, are uniformly characterized by very low or absent expression of H19, accompanied by biallelic hypermethylation of this gene, and this correlates with pathological biallelic expression of IGF2 w62–64x. The complete inactivation of H19 in WTs associated with biallelic hypermethylation is unlikely to be an epiphenomenon of the neoplastic state since it is highly site-specific Žlocally restricted to the body and promoter region of the H19 gene. w63x and since H19 is also biallelically hypermethylated in nonneoplastic tissues of a small subset of BWS patients w38x as well as in the non-neoplastic kidney parenchyma Žin addition to the neoplastic tissue. in a small subset of sporadic WT patients w63x. Based on the enhancer competition model for interaction of the IGF2 and H19 loci the loss of H19 expression in these tissues, either through a primary pathway of hypermethylation or through some other mechanism with secondary hypermethylation, allows interaction of the IGF2 promoter with the downstream shared enhancer element and thereby causes the observed biallelic IGF2 expression. Indeed, this precise scenario is what has been observed in H19 deletion mice which express IGF2 biallelically w45x. However, the situation in the WTs may not be quite so simple. The fact that the Mash-2 and Kip-2 genes, which map to the region of mouse chromosome 7 which is syntenic with human chromosome 11p15.5,
B. Tyckor Mutation Research 386 (1997) 131–140
are both paternally imprinted w44,65x has suggested that the domain of imprinting in this region may extend well beyond the local circuit of IGF2 r H19. A ubiquitously expressed gene, L23MRP, which maps just 40 kb downstream of human H19 is not subject to functional imprinting in fetal and adult human tissues w66x. If imprinting has not somehow ‘skipped over’ this gene then the interval between H19 and L23MRP may contain the downstream Žtelomeric. boundary of the chromosome 11p15.5 imprinted domain. In the opposite Žcentromeric. direction from H19, the chromosome 11p15.5 KIP2 gene maps an unknown but most likely small distance upstream of IGF2 w67x. There is a reproducible allelic expression bias at this gene locus consistent with ‘partial’ functional imprinting w46–48x, and most WTs with loss of H19 expression show a more variable but nonetheless definite reduction in KIP2 expression w46x. In fact, most WTs can be placed into one of three categories: Ži. those with 11p15.5 LOH and reduced expression of both H19 and KIP2 Ž45% of cases.; Žii. those with retention of heterozygosity, H19 hypermethylation, and reduced expression of both H19 and KIP2 Žabout 30% of cases.; and Žiii. those with normal methylation of H19 and high expression of both H19 and KIP2 Žabout 25% of cases.. Thus, the disruption of imprinting in WTs may be similar to that seen in PWSrAS in the sense that both are ‘domain effects’. The two systems appear to differ in one fundamental respect: the PWSrAS disruption of imprinting phenotype is only manifested after germline transmission of the abnormal chromosome while the WT phenotype originates in somatic tumor precursor cells. At this point we do not understand the molecular basis for the disruption of imprinting and it is not clear whether the two types of pathological imprinting, lack of appropriate imprinting in gametogenesis and somatic disruption of imprinting, will be mechanistically related.
5. Pharmacological and genetic manipulation of DNA methylation and imprinting Consistent with a role for CpG methylation in the maintenance of imprinting, the drug 5-azacytidine Žaza-C., which is a covalent inhibitor of the cytosine
135
Fig. 1. Partial erasure of functional imprinting of the human H19 gene by exposure of tumor cells to 5-azacytidine. A malignant rhabdoid tumor cell line, SM, which expresses low but detectable levels of H19 RNA from a single allele, was exposed to the indicated Žmicromolar. concentrations of the drug for a period of 2 weeks. Reverse transcription-PCR with H19 primers followed by digestion of the PCR products with RsaI ŽZhang and Tycko w25x. shows that the imprinted H19 allele Žlower band. was partially reactivated by this treatment. Parallel analysis of genomic DNAs showed partial erasure of allele-specific CpG-methylation of H19 DNA Ždata not shown.. Cell growth was drastically inhibited after 1 week, probably reflecting changes in the expression of numerous genes regulated by DNA methylation. G s genomic PCR control showing the expected two alleles of equal intensity; the allelic bands are at a higher molecular weight because of the presence of two small introns.
DNA methyltransferase enzyme Žreviewed in w68x., can at least partially erase the functional imprinting of the H19 gene in human tumor cell lines, including a malignant rhabdoid tumor line ŽFig. 1. and an embryonal rhabdomyosarcoma line w46x and in the rhabdomyosarcoma line the increased expression of H19 induced by aza-C was accompanied by the predicted reciprocal down-modulation of IGF2 mRNA w46x. Mice with germline insertional mutagenesis of the methyltransferase gene have decreased global DNA methylation and show embryonic lethality w69x and the homozygous embryos are also deficient in the maintenance of functional imprinting, as indicated by their biallelic expression of H19 and reduced expression of IGF2 w70x. These observations support a functional role for DNA methylation in the maintenance of repression of the imprinted allele of at least some imprinted genes and also suggest the intriguing possibility of using pharmacological agents analogous to aza-C to treat diseases associated with abnormalities of functional imprinting. Indeed, aza-C has been used with some success to treat patients with severe refractory thalassemias, in which the drug causes reexpression
136
B. Tyckor Mutation Research 386 (1997) 131–140
of fetal hemoglobin w71x, but since aza-C is also mutagenic in animals it is not an appropriate drug for treatment of lower-risk conditions like WT or PWSrAS. Clearly, a worthwhile problem for study is the design of non-mutagenic agents capable of promoting DNA demethylation.
6. DNA methylation in the establishment of imprints The findings of allele-specific DNA methylation at imprinted loci in late somatic tissues and the disruption of this functional imprinting by alterations in DNA methylation do not have direct bearing on the question of the role of methylation in establishment of the imprint. A cautionary note is sounded by evidence from the analysis of the process of X-chromosome inactivation, in which widespread DNA methylation follows rather than precedes the initial phase of chromatin-mediated gene silencing w72x. It may be that the initial parental imprint consists of a sperm- or oocyte-specific chromatin feature which subsequently initiates a series of molecular events in cells of the early embryo eventuating in allele-specific DNA methylation and allele-restricted transcription. To date there seems to be no direct evidence bearing on this possibility. In contrast, the DNA methylation status of imprinted genes in the earliest stages of embryonic development has been actively investigated by several groups and the data are consistent with the possibility of a primary function for gametic methylation in establishing the functional imprint. Both the mouse and human H19 genes are more heavily methylated in sperm than in oocytes w27,32– 35x. For the mouse H19 gene PCR assays have indicated that at least one HpaII site roughly 700 bp upstream of the first exon maintains its paternalspecific methylation through the blastocyst stage of development, resisting the genome-wide preimplantation wave of demethylation w35x. Similarly, for the mouse Igf2r gene two HpaII sites in an intron showed maternal-specific methylation which was preserved through early embryonic development w36x. Since these sites were hypermethylated on the active Žmaternal. Igf2r allele it was proposed that some cis-acting imprinting signal sequences might exist on the active rather than the silent allele. Similar studies
have suggested the presence of several CpG sites in the 5X promoter region of the mouse Xist gene which are methylated in oocytes but not in sperms and which remain methylated on the maternal X chromosome in early development, perhaps accounting for the selective expression of Xist RNA from the paternal chromosome in early extraembryonic tissues w40,41x. Functional studies using precise germline deletions or mutations could in principal clarify the role of these or other candidate imprinting signals, but such experiments might be fatally compromised by concomitant effects on binding of transcription factors.
7. DNA methylation and domain effects in imprinting How can a role for DNA methylation in the establishment of imprints be reconciled with the evidence for broad domains of coordinate imprinting? One concise model has been proposed by Paldi et al. based on their study of sex-specific meiotic recombination rates in imprinted chromosomal regions w73x. This study, which expanded on the early observations of Thomas and Rothstein w74x, found that the rates of recombination in the IGF2 r H19 region of chromosome 11p15.5 and in the SNRPN region of chromosome 15q11-q13 were much higher in male than in female meioses, while the sex-specific recombination rates in non-imprinted regions tended to be less biased. Based on these data they illustrated an accessibility model for imprinting similar to that originally suggested by Thomas and Rothstein and shown in modified form in Fig. 2. The high rates of recombination in male gametogenesis were taken to imply an open chromatin structure of the chromosomal regions which were destined for imprinting, while the lower rates of recombination in oogenesis were interpreted as indicating a compacted or inaccessible chromatin structure. The open male structure would allow modification of multiple ‘imprinting boxes’, each associated with a single imprinted gene, either by CpG methylation or by binding of a repressive chromatin protein or both, while the compacted female structure would prevent such modifications by masking the imprinting boxes. Chromosomal boundary elements active in the developing gametes
B. Tyckor Mutation Research 386 (1997) 131–140
137
tent with the evidence for disruption of gametic imprinting by DNA deletions in PWSrAS, though analysis of the DNA sequences which have been deleted in these kindreds might well hold some surprises. How it might apply to somatic disruption of imprinting in WTs is less clear. While deletions in cis-acting regulatory sequences may be awaiting discovery in these tumors, alternative explanations for conversion to the bipaternal chromosome 11p15.5 epigenotype involving direct allelic interactions in tumor precursor cells, such as transfer of the paternal methylation imprint to critical sites on the maternal homologue via the action of the DNA methyltransferase on transiently formed heteroduplex DNA w20x, should also be considered.
8. Conclusions
Fig. 2. The accessibility model for genomic imprinting Žmodified from w73x and w74x..
might determine the boundaries of these chromatin domains and chromatin compaction might be mediated by proteins binding to repetitive DNA elements. Subsequent somatic development might homogenize the chromatin structures and DNA methylation patterns of the two homologues to a large extent, but the methylation imprints at critical sites Žin H19 and SNRPN as well as in other nearby genes. would still be carried by the paternal homologue. What might define the extents of imprinted domains? One possibility is that there might be chromosomal boundary elements, DNA sequences which can insulate active or ‘open’ chromatin from adjacent inactive or ‘closed’ chromatin, flanking imprinted domains. Such elements have been visualized bordering transcriptional ‘puffs’ in Drosophila polytene chromosomes by virtue of their binding to at least one specific chromatin protein w58x. For the chromosome 11p15.5 domain one such boundary might lie between H19 and L23MRP w66,75x. Whether this model will be useful as a framework for explaining the disruption of imprinting in human diseases remains to be seen. It appears to be consis-
DNA methylation at cytosine residues is essential for functional imprinting of at least a subset of imprinted mammalian genes and pharmacological manipulation of DNA methylation might ultimately be useful in correcting pathological imprinting. On the other hand, while the hypothesis that DNA methylation is the biochemical correlate of the primary imprint is attractive a priori and is consistent with all available data, it nonetheless remains unproven. Future progress in understanding the biochemical basis for imprinting may depend on a better understanding of DNA and chromatin structure in the developing gametes and its influence on the methylation patterns of gamete DNA.
Acknowledgements This work was supported by grants CA60765 from the N.I.H. and JFRA-482 from the A.C.S. The experiment shown in Fig. 1 was done by Dr. Thomas Moulton in the author’s laboratory.
References w1x Monk, M., Boubelik, M. and S. Lehnert Ž1987. Temporal and regional changes in DNA methylation in the embryonic, extraembryonic, and germ cell lineages during mouse development, Development, 99, 371–382.
138
B. Tyckor Mutation Research 386 (1997) 131–140
w2x Sanford, J.P., Clark, H.J., Chapman, V.M. and J. Rossant Ž1987. Differences in DNA methylation during oogenesis and spermatogenesis and their persistence during early embryogenesis in the mouse, Genes Dev., 1, 1039–1046. w3x Driscoll, D.J. and B.R. Migeon Ž1990. Sex difference in methylation of single-copy genes in human meiotic germ cells: implications for X chromosome inactivation, parental imprinting, and origin of CpG mutations, Somatic Cell Mol. Genet., 16, 267–282. w4x Howlett, S.K. and W. Reik Ž1991. Methylation levels of maternal and paternal genomes during preimplantation development, Development, 113, 119–127. w5x Kafri, T., Ariel, M., Brandeis, M., Shemer, R., Urven, L., McCarrey, J., Cedar, H. and A. Razin Ž1992. Developmental pattern of gene-specific DNA methylation in the mouse embryo and germ line, Genes Dev., 6, 705–714. w6x Frank, D., Keshet, I., Shani, M., Levine, A., Razin, A. and H. Cedar Ž1991. Demethylation of CpG islands in embryonic cells, Nature, 351, 239–241. w7x Surani, M.A.H., Barton, S.C. and M.L. Norrris Ž1984. Development of reconstituted mouse eggs suggests imprinting of the genome during gametogenesis, Nature, 308, 548–550. w8x Barton, S.C., Surani, M.A.H. and Norris, M.L. Ž1984. Role of paternal and maternal genomes in mouse development, Nature, 311, 374–376. w9x McGrath, J. and D. Solter Ž1984. Completion of mouse embryogenesis requires both the maternal and paternal genomes, Cell, 37, 179–183. w10x Reik, W., Collick, A., Norris, M.L., Barton, S.C. and M.A. Surani Ž1987. Genomic imprinting determines methylation of parental alleles in transgenic mice, Nature, 328, 248–251. w11x Sapienza, C., Peterson, A.C., Rossant, J. and R. Balling Ž1987. Degree of methylation of transgenes is dependent on gamete of origin, Nature, 328, 251–254. w12x Hadchouel, M., Farza, H., Simon, D., Tiollais, P. and C. Pourcel Ž1987. Maternal inhibition of hepatitis B surface antigen gene expression in transgenic mice correlates with de novo methylation, Nature, 329, 454–456. w13x Swain, J.L., Stewart, T.A. and P. Leder Ž1987. Parental legacy determines methylation and expression of an autosomal transgene: A molecular mechanism for parental imprinting, Cell, 50, 719–727. w14x Chaillet, R., Bader, D.S. and P. Leder Ž1995. Regulation of genomic imprinting by gametic and embryonic processes, Genes Dev., 9, 1177–1187. w15x Michaud, E.J., van Vugt, M.J., Bultman, S.J., Sweet, H.O., Davisson, M.T. and R.P. Woychik Ž1994. Differential expression of a new dominant agouti allele ŽAiapy. is correlated with methylation state and is influenced by parental lineage, Genes Dev., 8, 1463–1472. w16x Duhl, D.M.J., Vrieling, H., Miller, K.A., Wolff, G.L. and G.S. Barsh Ž1994. Neomorphic agouti mutations in obese yellow mice, Nature Genet., 8, 59–65. w17x Bestor, T.H. Ž1990. DNA methylation: evolution of a bacterial immune function into a regulator of genome structure and activity in higher eukaryotes, Phil. Trans. Royal Soc. Lond. B326, 179–187.
w18x Barlow, D.P. Ž1993. Methylation and imprinting: from host defense to gene regulation?, Science, 260, 309–10. w19x Neumann, B., Kubicka, P. and D.P. Barlow Ž1995. Characteristics of imprinted genes, Nature Genet., 9, 12–13. w20x Bestor, T.H. and B. Tycko Ž1996. Creation of genomic methylation patterns. Nature Genet., 12, 363–367. w21x Moretti, P., Freeman, K., Coodly, L. and D. Shore Ž1994. Evidence that a complex of SIR proteins interacts with the silencer and telomere-binding protein RAP1, Genes Dev., 8, 2257–2269. w22x DeChiara, T.M., Robertson, E.J. and A. Efstratiadis Ž1991. Paternal imprinting of the mouse insulin-like growth factor II gene, Cell, 64, 849–859. w23x Barlow, D,P, Stoger, R., Hermann, B.G., Saito, K. and N. Schweifer Ž1991. The mouse insulin-like growth factor type-2 receptor is imprinted and closely linked to the Tme locus, Nature, 349, 84–87. w24x Bartolomei, M.S., Zemel, S. and S.M. Tilghman Ž1991. Parental imprinting of the mouse H19 gene, Nature, 351, 153–155. w25x Zhang, Y. and B. Tycko Ž1992. Monoallelic expression of the human H19 gene. Nature Genet., 1, 40–44. w26x Rainier, S., Johnson, L., Dobry, C.J., Ping, A.J., Grundy, P.E. and A.P. Feinberg Ž1993. Relaxation of imprinted genes in human cancer, Nature, 362, 747–749. w27x Zhang, Y., Shields, T., Crenshaw, T., Hao, Y., Moulton, T. and B. Tycko Ž1993. Imprinting of human H19: allelespecific CpG methylation, loss of the active allele in Wilms’ tumor and potential for somatic allele switching, Am. J. Hum. Genet., 53, 113–124. w28x Ohlsson, R., Nystrom, A., Pfeifer-Ohlsson, S., Tohonen, V., Hedbourg, F., Schofield, P., Flam, F. and T.J. Ekstrom Ž1993. IGF2 is parentally imprinted during human embryogenesis and in the Beckwith-Wiedemann syndrome, Nature Genet., 4, 94–97. w29x Giannoukakis, N., Deal, C., Paquette, J., Goodyer, C.G. and C. Polychronakos Ž1993. Parental genomic imprinting of the human IGF2 gene, Nature Genet., 4, 98–101. w30x Leff, S.E., Brannan, C.I., Reed, M.L., Ozcelik, T., Francke, U., Copeland, N.J and N.A. Jenkins Ž1992. Maternal imprinting of the mouse Snrpn gene and conserved linkage homology with the human Prader-Willi syndrome region, Nature Genet., 2, 259–264. w31x Glenn, C.C., Porter, K.A., Jong, M.T.C., Nicholls, R.D. and D.J. Driscoll Ž1993. Functional imprinting and epigenetic modification of the human SNRPN gene, Hum. Mol. Genet., 2, 2001–2005. w32x Ferguson-Smith, A.C., Sasaki, H., Cattanach, B.M. and M.A. Surani Ž1993. Parental-origin-specific epigenetic modification of the mouse H19 gene, Nature, 362, 751–755. w33x Brandeis, M., Kafri, T., Ariel, M., Chaillet, J.R., McCarrey, J., Razin, A. and H. Cedar Ž1993. The ontogeny of allelespecific methylation associated with imprinted genes in the mouse, EMBO J., 12, 3669–3677. w34x Bartolomei, M.S., Webber, A.L., Brunkow, M.E. and S.M. Tilghman Ž1993. Epigenetic mechanisms underlying the imprinting of the mouse H19 gene, Genes Dev., 7, 1663–1673.
B. Tyckor Mutation Research 386 (1997) 131–140 w35x Tremblay, K.D., Saam, J.R., Ingram, R.S., Tilghman, S.M. and M.S. Bartolomei Ž1995. A paternal-specific methylation imprint marks the alleles of the mouse H19 gene, Nature Genet., 9, 407–413. w36x Stoger, R., Kubicka, P., Liu, C.-G., Kafri, T., Razin, A., Cedar, H. and D.P. Barlow Ž1993. Maternal-specific methylation of the imprinted Igf2r locus identifies the expressed locus as carrying the imprinting signal, Cell, 73, 61–72. w37x Koide, T., Ainscough, J., Wijgerde, M. and M.A. Surani Ž1994. Comparative analysis of the Igf-2r H19 imprinted domain: identification of a highly conserved intergenic DNase I hypersensitive region, Genomics, 24, 1–8. w38x Reik, W., Brown, K.W., Schneid, H., Le Bouc, Y, Bickmore W. and E.R. Maher Ž1995. Imprinting mutations in the Beckwith-Wiedemann syndrome suggested by an altered imprinting pattern in the IGF2-H19 domain, Hum. Mol. Genet., 4, 2379–2385. w39x Glenn, C.C., Saitoh, S., Jong, M.T.C., Filbrandt, M.M., Surti, U., Driscoll, D.J. and R.D. Nicholls Ž1996. Gene structure, DNA methylation and imprinted expression of the human SNRPN gene. Am. J. Hum. Genet., 58, 335–346. w40x Ariel, M., Robinson, E., McCarrey, J.R. and H. Cedar, Ž1995. Gamete-specific methylation correlates with imprinting of the murine Xist gene, Nature Genet., 9, 312–315. w41x Zuccotti, M. and M. Monk Ž1995. Methylation of the mouse Xist gene in spern and eggs correlates with imprinted Xist expression and paternal X-inactivation, Nature Genet., 9, 316–320. w42x Glenn, C., Nicholls, R.D., Robinson, W.P., Saitoh, S., Niikawa, N., Schinzel, A. Horsthemke, B. and D. Driscoll Ž1993. Modification of 15q11-q13 DNA methylation imprints in unique Angelman and Prader-Willi patients, Hum. Mol. Genet., 2, 1377–1382. w43x Hayashizaki, Y., Shibata, H., Hirotsune, S., Sugino, H., Okazaki, Y., Sasaki, N., Hirose, K., Imoto, H., Okuizumi, H., Muramatsu, M., Komatsubara, H., Shiroishi, T, Moriwaki, K, Katsuki, M., Hatano, N., Sasaki, H., Ueda, T., Mise, N., Takagi, N., Plass, C. and V.M. Chapman Ž1994. Identification of an imprinted U2af binding protein related sequence on mouse chromosome 11 using the RLGS method, Nature Genet., 6, 33–40. w44x Hatada, I. and T. Mukai Ž1995. Genomic imprinting of p57Kip2, a cyclin-dependent kinase inhibitor, in mouse, Nature Genet., 11, 204–206. w45x Leighton, P.A., Ingram, R.S., Eggenschwiler, J., Efstratiatis, A. and S.M. Tilghman Ž1995. Disruption of imprinting caused by deletion of the H19 gene region in mice. Nature, 375, 34–39. w46x Chung, W.-Y., Yuan, L., Feng, L., Hensle, T. and B. Tycko Ž1996. Chromosome 11p15.5 regional imprinting: comparative analysis of KIP2 and H19 in human tissues and Wilms’ tumors. Hum. Mol. Genet., 8, 1101–1108. w47x Kondo, M., Matsuoka, S., Uchida, K., Osada, H., Nagatake, M., Takagi, K., Harper, J.W., Takahashi, T., Elledge, S.J. and T. Takahashi Ž1996. Selective maternal allele loss in human lung cancers of the maternally expressed p57KIP2 gene at 11p15.5. Oncogene, 12, 1365–1368.
139
w48x Matsuoka, S., Thompson, J.S., Edwards, M.C., Barletta, J.M., Grundy, P., Kalikin, L.M., Harper, J.W., Elledge, S.J. and A.P. Feinberg Ž1996. Imprinting of the gene encoding a human cyclin-dependent kinase inhibitor, p57KIP2, on chromosome 11p15. Proc. Natl. Acad. Sci. USA, 93, 3026–3030. w49x Knoll, J.H.M., Nicholls, R.D., Magenis, R.E., Graham, J.M. Jr, Lalande, M. and S.A. Latt Ž1989. Angelman and PraderWilli syndromes share a common chromosome 15 deletion but differ in parental origin of the deletion., Am. J. Med. Genet., 32, 285–290. w50x Nicholls, R.D., Knoll, J.H.M., Butler, M.G., Karam, S. and M. Lalande Ž1989. Genetic imprinting suggested by maternal heterodisomy in non-deletion Prader-Willi syndrome, Nature, 342, 281–285. w51x Malcolm, S., Clayton-Smith, J,. Nichols, M., Robb, S., Webb, T., Armour, J.A.L., Jeffreys, A.J. and M.E. Pembrey Ž1991. Uniparental paternal disomy in Angelman’s syndrome, Lancet, 337, 694–697. w52x Wagstaff, J., Knoll, J.H.M., Glatt, K.A., Shugart, Y.Y., Sommer, A. and M. Lalande Ž1992. Maternal but not paternal transmission of 15q11-13-linked nondeletion Angleman syndrome leads to phenotypic expression, Nature Genet., 1, 291–294 w53x Buiting, K. Dittrich, B., Gross, S., Greger, V., Lalande, M., Robinson, W., Mutirangura, A. et al. Ž1993. Molecular definition of the Prader-Willi syndrome chromosome region and orientation of the SNRPN gene, Hum. Mol. Genet., 2, 1991– 1994. w54x Jong, M.T.C., Carey, A.H., Stewart, C.L., Rinchik, E.M., Glenn, C.C., Driscoll, D.J. and R.D. Nicholls Ž1993. The ZNF127 gene encodes a novel C3HC4 zinc-finger protein and its expression is regulated by genomic imprinting, Am. J. Hum. Genet., 53, Abstr. 697. w55x Wevrick, R., Kerns, J.A. and U. Francke Ž1994. Identification of a novel paternally expressed gene in the Prader-Willi syndrome region, Hum. Mol. Genet., 3, 1877–1882. w56x Sutcliffe, J.S., Nakao, M., Christian, S,. Orstavik, K.H., Tommerup, N., Ledbetter, D.H. and A.L. Beaudet Ž1994. Deletions of a differentially methylated CpG island at the SNRPN gene define a putative imprinting control region, Nature Genet., 8, 52–58. w57x Buiting, K., Saitoh, S., Gross, S., Dittrich, B., Schwartz, S., Nicholls, R.D. and B. Horsthemke Ž1995. Inherited microdeletions in the Angelman and Prader-Willi syndromes define an imprinting centre on human chromosome 15, Nature Genet., 9, 395–400. w58x Zhao, K., Hart, C.M. and U.K. Laemmli Ž1995. Visualization of chromosomal domains with boundary element-associated factor BEAF-32, Cell, 81, 879–889. w59x Tycko, B. Ž1996. Genomic imprinting in Wilms’ tumor. Med. Ped. Oncol., in press. w60x Mannens, M., Hoovers, J.M., Redeker, E., Verjaal, M., Feinberg, A.P., Little, P. et al. Ž1994. Parental imprinting of human chromosome region 11p15.3-pter involved in the Beckwith-Wiedemann syndrome and various human neoplasia, Eur. J. Hum. Genet., 2, 3–23. w61x Weksberg, R., Shen, D.R., Fei, Y.L., Song, Q.L. and J.
140
w62x
w63x
w64x
w65x
w66x
w67x
B. Tyckor Mutation Research 386 (1997) 131–140 Squire Ž1993. Disruption of insulin-like growth factor 2 imprinting in Beckwith-Wiedemann syndrome, Nature Genet., 5, 143–150. Steenman, M.J.C., Rainier, S., Dobry, C.J., Grundy, P., Horon, I.L. and A.P. Feinberg Ž1994. Loss of imprinting of IGF2 is linked to reduced expression and abnormal methylation of H19 in Wilms’ tumor, Nature Genet., 7, 433–439. Moulton, T., Crenshaw, T., Hao, Y., Moosikasuwan, J., Lin, N., Dembitzer, F., Hensle, T., Weiss, L., McMorrow, L., Loew, T., Kraus, W., Gerald, W. and B. Tycko Ž1994. Epigenetic lesions at the H19 locus in Wilms’ tumor patients, Nature Genet., 7, 440–447. Taniguchi, T., Sullivan, M.J., Osamu, O. and A. Reeve Ž1995. Epigenetic changes encompassing the IGF2r H19 locus associated with relaxation of IGF2 imprinting and silencing of H19 in Wilms tumor, Proc. Natl. Acad. Sci. USA, 92, 2159–2163. Guillemot, F., Caspary, T., Tilghman, S.M., Copeland, N.G., Gilbert, D.J., Jenkins, N.A. anderson, D.J.., Joyner, A.L., Rossant, J. and A. Nagy Ž1995. Genomic imprinting of Mash2, a mouse gene required for trophoblast development, Nature Genet., 9, 235–241. Tsang, P., Gilles, F., Yuan, L., Kuo, Y-H., Lupu, F., Samara, G., Moosikasuwan, J., Goy, A., Zelenetz, A.D., Selleri, L. and B. Tycko Ž1995. A novel L23-related gene 40 Kb downstream of the imprinted H19 gene is biallelically expressed in mid-fetal and adult human tissues, Hum. Mol. Genet., 4, 1499–1507. Hoovers, J.M.N., Kalikin, L.M., Johnson, L.A., Alders, M.,
w68x w69x
w70x
w71x
w72x
w73x
w74x w75x
Redeker, B., Law, D.J., Bliek, J, Steenman, M., Benedict, M. Wiegant, J., Lengauer, C. Taillon-Miller P., Schlessinger, D., Edwards, M.C., Elledge, S.J., Ivens, A., Westerveld, A., Little, P. Mannens, M. and A. Feinberg Ž1995. Multiple genetic loci within 11p15 defined by Beckwith-Weidemann syndrome rearrangement breakpoints and subchromosomal transferable fragments, Proc. Natl. Acad. Sci. USA, 92, 12456–12460. Riggs, A.D. and P.A. Jones Ž1983. 5-methylcytosine, gene regulation and cancer, Adv. Cancer Res., 40, 1–28. Li, E., Bestor, T.H. and R. Jaenisch Ž1992. Targeted mutation of the DNA methyltransferase gene results in embryonic lethality, Cell, 69, 915–926. Li, E., Beard C. and R. Jaenisch. Ž1993. Role for DNA methylation in genomic imprinting, JOURNAL, 366, 362– 365. Lowrey, C.H. and A.W. Nienhuis Ž1993. Treatment with azacytidine of patients with end-stage beta-thalassemia, New Engl. J. Med., 329, 845–848. Lock, L.F., Takagi, N. and G.R. Martin Ž1987. Methylation of the Hprt gene on the inactive X occurs after chromosome inactivation, Cell, 48, 39–46. Paldi, A., Gyapay, G. and J. Jami Ž1995. Imprinted chromosomal regions of the human genome display sex-specific meiotic recombination frequencies, Curr. Biol., 5, 1030– 1035. Thomas, B.J. and R. Rothstein Ž1991. Sex, maps, and imprinting, Cell, 64, 1–3. Yuan, L. et al. Ž1996. Human Mol. Genet., 5, 1931–1937.