Conservation of Dnmt1o cytosine methyltransferase in the marsupial Monodelphis domestica
5
LETTER Conservation of Dnmt1o Cytosine Methyltransferase in the Marsupial Monodelphis domestica Feng Ding, 1 Carol Patel, 1 Sarayu Ratnam, 2 John R. McCarrey, 3 and J. Richard Chaillet 1,2 * 1 Department of Pediatrics, University of Pittsburgh, Pittsburgh, Pennsylvania 2 Department of Human Genetics, University of Pittsburgh, Pittsburgh, Pennsylvania 3 Department of Biology, University of Texas at San Antonio, San Antonio, Texas Received 28 January 2003; Accepted 21 May 2003 Summary: Imprinted genes have been identified in both eutherian mammals and in marsupials. In eutherian spe- cies, there is a conservation of the imprinting process, both in terms of the genes imprinted and the epigenetic inheritance mechanism. In the mouse, the inheritance of gametic methylation patterns depends on an oocyte- derived isoform of the Dnmt1 (cytosine-5)-methyltrans- ferase protein, Dnmt1o, which functions during pre- implantation development to maintain methylation patterns on imprinted alleles. To determine if this com- ponent of genomic imprinting is also found in marsupi- als, Dnmt1 isoforms were examined in somatic cells and germ cells of the South American opossum Monodelphis domestica. There is a Dnmt1o protein in Monodelphis oocytes that is synthesized, as in the mouse, from a different transcript than the somatic Dnmt1 protein. Thus, an essential component of imprinting in eutherian mammals is found in a marsupial species, suggesting that marsupials and eutherian mammals imprint their genes with the same methylation-dependent mecha- nism. genesis 36:209 –213, 2003. © 2003 Wiley-Liss, Inc. Key words: methylation; imprinting; marsupial; epigenetic; opossum Genomic imprinting is an epigenetic gene regulatory process that distinguishes alleles of genes by their paren- tal origin (Solter, 1988; Bartolomei and Tilghman, 1997). As a consequence of the epigenetic modification of pa- rental alleles, one allele of an imprinted gene is active and the other is transcriptionally silent. Many of the important features of imprinting have been deduced from studies of mouse and human imprinted genes. Comparison of imprinting in these two eutherian mam- mals has shown that among the genes that are imprinted in the mouse, most are also imprinted in humans. A notable exception is the insulin-like growth factor II receptor (Igf2r) gene, which is imprinted in mice, but is not always imprinted in humans (Killian et al., 2001). Moreover, the genome-wide organization of imprinted genes is similar in the mouse and human. In particular, the arrangements of imprinted and nonimprinted genes within the gene clusters that contain most imprinted genes are conserved from the mouse to the human (Reik and Walter, 1998). The relationship between genomic imprinting and DNA cytosine methylation is also conserved in the mouse and human. For mouse imprinted genes and their human homologs, there is an association between paren- tal allele-specific expression and parental allelic differ- ences in DNA methylation (Bartolomei and Tilghman, 1997). The difference in parental methylation is estab- lished during gametogenesis and is then perpetuated after fertilization by the actions of the Dnmt1 gene products, the Dnmt1o and Dnmt1 proteins. The M r 175,000 Dnmt1o form of the Dnmt1 (cytosine-5)-meth- yltransferase is synthesized and stored in the maturing oocyte (Mertineit et al., 1998). Following fertilization, this stored form is used to maintain DNA methylation patterns on alleles of imprinted genes (Howell et al., 2001). Dnmt1o probably maintains methylation at only the fourth S phase of embryogenesis because it is only found in the nucleus at the 8-cell stage of preimplanta- tion development (Carlson et al., 1992; Howell et al., 2001). Following implantation of the embryo into the uterus, Dnmt1o disappears and the larger M r 190,000 Dnmt1 protein is expressed in the somatic cells (Trasler et al., 1996). Here, Dnmt1 maintains methylation pat- terns on imprinted genes for the duration of the mouse’s life (Li et al., 1993). Thus, Dnmt1o and Dnmt1 are essential components of a conserved imprinting process that maintains parental epigenetic information in the form of parental allele-specific DNA methylation pat- terns. The primary structure of Dnmt1 is the same as Dnmt1o except for an additional 118 amino acids at its * Correspondence to: J. Richard Chaillet, Department of Pediatrics, Uni- versity of Pittsburgh, Pittsburgh, PA 15213. E-mail: [email protected] Contract grant sponsor: the NIH (to JRC). DOI: 10.1002/gene.10215 © 2003 Wiley-Liss, Inc. genesis 36:209 –213 (2003)
Transcript of Conservation of Dnmt1o cytosine methyltransferase in the marsupial Monodelphis domestica
LETTER
Conservation of Dnmt1o Cytosine Methyltransferase in theMarsupial Monodelphis domesticaFeng Ding,1 Carol Patel,1 Sarayu Ratnam,2 John R. McCarrey,3 and J. Richard Chaillet1,2*1Department of Pediatrics, University of Pittsburgh, Pittsburgh, Pennsylvania2Department of Human Genetics, University of Pittsburgh, Pittsburgh, Pennsylvania3Department of Biology, University of Texas at San Antonio, San Antonio, Texas
Received 28 January 2003; Accepted 21 May 2003
Summary: Imprinted genes have been identified in botheutherian mammals and in marsupials. In eutherian spe-cies, there is a conservation of the imprinting process,both in terms of the genes imprinted and the epigeneticinheritance mechanism. In the mouse, the inheritance ofgametic methylation patterns depends on an oocyte-derived isoform of the Dnmt1 (cytosine-5)-methyltrans-ferase protein, Dnmt1o, which functions during pre-implantation development to maintain methylationpatterns on imprinted alleles. To determine if this com-ponent of genomic imprinting is also found in marsupi-als, Dnmt1 isoforms were examined in somatic cells andgerm cells of the South American opossum Monodelphisdomestica. There is a Dnmt1o protein in Monodelphisoocytes that is synthesized, as in the mouse, from adifferent transcript than the somatic Dnmt1 protein.Thus, an essential component of imprinting in eutherianmammals is found in a marsupial species, suggestingthat marsupials and eutherian mammals imprint theirgenes with the same methylation-dependent mecha-nism. genesis 36:209–213, 2003.© 2003 Wiley-Liss, Inc.
Key words: methylation; imprinting; marsupial; epigenetic;opossum
Genomic imprinting is an epigenetic gene regulatoryprocess that distinguishes alleles of genes by their paren-tal origin (Solter, 1988; Bartolomei and Tilghman, 1997).As a consequence of the epigenetic modification of pa-rental alleles, one allele of an imprinted gene is activeand the other is transcriptionally silent. Many of theimportant features of imprinting have been deducedfrom studies of mouse and human imprinted genes.Comparison of imprinting in these two eutherian mam-mals has shown that among the genes that are imprintedin the mouse, most are also imprinted in humans. Anotable exception is the insulin-like growth factor IIreceptor (Igf2r) gene, which is imprinted in mice, but isnot always imprinted in humans (Killian et al., 2001).Moreover, the genome-wide organization of imprintedgenes is similar in the mouse and human. In particular,the arrangements of imprinted and nonimprinted genes
within the gene clusters that contain most imprintedgenes are conserved from the mouse to the human (Reikand Walter, 1998).
The relationship between genomic imprinting andDNA cytosine methylation is also conserved in themouse and human. For mouse imprinted genes and theirhuman homologs, there is an association between paren-tal allele-specific expression and parental allelic differ-ences in DNA methylation (Bartolomei and Tilghman,1997). The difference in parental methylation is estab-lished during gametogenesis and is then perpetuatedafter fertilization by the actions of the Dnmt1 geneproducts, the Dnmt1o and Dnmt1 proteins. The Mr
175,000 Dnmt1o form of the Dnmt1 (cytosine-5)-meth-yltransferase is synthesized and stored in the maturingoocyte (Mertineit et al., 1998). Following fertilization,this stored form is used to maintain DNA methylationpatterns on alleles of imprinted genes (Howell et al.,2001). Dnmt1o probably maintains methylation at onlythe fourth S phase of embryogenesis because it is onlyfound in the nucleus at the 8-cell stage of preimplanta-tion development (Carlson et al., 1992; Howell et al.,2001). Following implantation of the embryo into theuterus, Dnmt1o disappears and the larger Mr 190,000Dnmt1 protein is expressed in the somatic cells (Trasleret al., 1996). Here, Dnmt1 maintains methylation pat-terns on imprinted genes for the duration of the mouse’slife (Li et al., 1993). Thus, Dnmt1o and Dnmt1 areessential components of a conserved imprinting processthat maintains parental epigenetic information in theform of parental allele-specific DNA methylation pat-terns.
The primary structure of Dnmt1 is the same asDnmt1o except for an additional 118 amino acids at its
* Correspondence to: J. Richard Chaillet, Department of Pediatrics, Uni-versity of Pittsburgh, Pittsburgh, PA 15213.
E-mail: [email protected] grant sponsor: the NIH (to JRC).
DOI: 10.1002/gene.10215
© 2003 Wiley-Liss, Inc. genesis 36:209–213 (2003)
amino terminus (Mertineit et al., 1998). The mouseDnmt1o protein is translated from an oocyte-specificmRNA whose synthesis is initiated from an oocyte-spe-cific promoter (Mertineit et al., 1998; Ratnam et al.,2002). The first exon of this Dnmt1o transcript (exon1o) is distinct from the first exon of the somatic Dnmt1transcript (exon 1s). Exon 1o and exon 1s are splicedinto a common Dnmt1 exon 2. The initiation codon forthe Dnmt1 protein is in exon 1s and the initiation codonfor the Dnmt1o protein is in common exon 4 (at theposition of the first internal methionine codon for theDnmt1 protein).
Two genes that are known to be imprinted in themouse are also imprinted in marsupials. The insulin-likegrowth factor II (Igf2) gene is paternally expressed inthe South American opossum Monodelphis domestica(O’Neill et al., 2000), and the insulin-like growth factor IIreceptor (Igf2r) gene is maternally expressed in theAmerican opossum Didelphis virginiana (Killian et al.,2000). Igf2r is not imprinted in monotremes (Killian etal., 2000). These observations are consistent with theevolution of imprinting after the split of Prototheria(monotremes) from Metatheria (marsupials) and Eu-theria (placental mammals), but before the split between
Metatheria and Eutheria. Despite the conservation ofparental allele-specific expression in marsupials and eu-therian mammals, parental allele-specific methylationpatterns have not been described for marsupial im-printed genes. Indeed, the Igf2r differentially methylateddomain (DMD) within intron 2 of the mouse Igf2r geneis not found in intron 2 of the Didelphis virginiana Igf2rgene (Killian et al., 2000). These observations raise thepossibility that marsupials control gene imprinting by anepigenetic mechanism that does not utilize DNA meth-ylation. Therefore, to better define the role of DNAmethylation in genomic imprinting in marsupials, weinvestigated the conservation of expression of theDnmt1o form of the Dnmt1 (cytosine-5)-methyltrans-ferase in the South American opossum, Monodelphisdomestica.
To determine if there is a marsupial Dnmt1o form ofthe Dnmt1 (cytosine-5)-methyltransferase, we first deter-mined the sequence of the mRNA encoding the Dnmt1protein by cloning and sequencing a series of partialcDNAs isolated from a Monodelphis spleen cDNA li-brary. The open reading frame of the MonodelphisDnmt1 mRNA predicts a protein of 1,514 amino acids(Mr 171,149). As expected, much of the primary struc-
FIG. 1. Comparison of amino termini of Dnmt1 proteins from eight different species. The sequence alignment was performed using theClustaw Format program. The approximate positions of the first internal methionine of Dnmt1 proteins from human, rat, mouse, opossum(Monodelphis domestica), chicken, fish (Danio rerio), and urchin (Paracentrotus lividus) are shown by the arrow (position 122 in the humanDnmt1 protein sequence). Dnmt1 from frog (Xenopus laevis) does not have an internal methionine near this position. Black shadingrepresents amino acid identities in the aligned sequences and gray shading represents amino acid similarities.
210 DING ET AL.
ture of the Monodelphis Dnmt1 protein is conservedamong Dnmt1 proteins from numerous invertebrate andvertebrate species (Fig. 1, and data not shown). Withrespect to a possible Monodelphis Dnmt1o protein,there is an internal methionine at position 132 of theMonodelphis Dnmt1 protein. This is the first internalmethionine in the amino terminal portion of the Dnmt1protein, and it is found in a region homologous to theamino terminus of the mouse Dnmt1o protein. More-over, the mRNA sequence surrounding the putative Mo-nodelphis Dnmt1o initiation codon (agagAUGg) is anexcellent hypothetical translation initiation sequence(Kozak, 1987). From this analysis, the predicted size of aMonodelphis Dnmt1o protein is 1,383 amino acids (Mr
156,296). We conclude from this that the primary struc-ture of the Monodelphis Dnmt1 protein is consistentwith the possible existence of a truncated version of theprotein, one that is homologous to the mouse Dnmt1oprotein.
We explored the possibility that Monodelphis oocytescontain a Dnmt1o protein that is homologous to the
mouse Dnmt1o protein by examining Monodelphis ova-ries for the presence of transcripts homologous to themouse Dnmt1o transcript. Using an oligonucleotidecomplementary to the region of the opossum Dnmt1mRNA that is homologous to exon 4 of mouse Dnmt1,we performed 5� RACE on mRNA isolated from Mono-delphis spleen and ovary. Multiple 5� RACE productswere cloned and sequenced from both tissues. Althoughthe 3� ends of the 5� RACE products from ovary andspleen were identical, the 5� sequences of the 5� RACEproducts from the ovary were different from the spleen-derived 5� RACE products. The longest 5� RACE prod-ucts from Monodelphis ovary and spleen were com-pared to each other and to the 5� ends of the mouseDnmt1 and Dnmt1o mRNA sequences (Fig. 2). Interest-ingly, the 5�end of the region of similarity between theopossum spleen and ovary mRNAs coincides with the 5�end of the mouse exon 2, which is the 5� most commonexon found in the mouse Dnmt1o and Dnmt1 tran-scripts (Mertineit et al., 1998). Thus, as in the mouse,there appear to be alternative first exons in the Mono-
FIG. 2. Sequence comparison of the 5� ends of Dnmt1 transcripts derived from Monodelphis ovary (Op_O), Monodelphis spleen (Op_S),mouse spleen (Mouse_S), and mouse oocyte (Mouse_O). mRNA sequence is shown as the sense DNA strand. The initiation codons (ATG)for the four transcripts are highlighted in black. Nucleotide identities among all four sequences, starting at the 5� end of exon 2, are indicatedby gray shading. Nucleotide identities between exon 1 of Op_S and exon 1 of mouse_S are also indicated by gray shading.
211MONODELPHIS DnmT1 METHYLTRANSFERASE
delphis Dnmt1 gene. Because mouse Dnmt1o tran-scripts are found in the germ cells of ovaries (Ratnam etal., 2002), we also conclude that the distinct Monodel-phis ovary Dnmt1 mRNA is found in female germ cells.
To identify Dnmt1 proteins in Monodelphis domes-tica tissues, we probed immunoblots of protein extractsfrom various organs with three different polyclonal an-tisera. The rabbit anti-Dnmt1 Path52 antiserum (Carlsonet al., 1992) did not bind to opossum Dnmt1 protein(data not shown). However, two antisera that we re-cently developed identified specific proteins in opossumtissues (Fig. 3a). The rabbit anti-Dnmt1 UPT82 antiserumidentifies an Mr 175,000 protein in extracts from opos-sum ovary and spleen; this antiserum recognizesepitopes in the amino-terminal 118-amino acid domainof the Mr 190,000 form of mouse Dnmt1 (Ratnam et al.,2002). The chicken anti-Dnmt1 UPTC21 antiserum (Rat-nam et al., 2002), which recognizes epitopes commonto Dnmt1 and Dnmt1o, also identifies the Mr 175,000protein in opossum ovary and spleen. The observedmass of Mr 175,000 is consistent with the predicted size(1,514 amino acids) and mass (Mr 171,149) of the Mo-nodelphis Dnmt1 protein. We conclude from this immu-noblot analysis that both the UPT82 and the UPTC21antisera specifically identify the opossum Dnmt1 pro-tein.
In addition to the Mr 175,000 protein, the UPTC21antiserum identifies a minor ovary protein of Mr 160,000that is not found in the spleen extract (Fig. 3a). Becausethe observed mass is approximately the same as the massof the predicted Monodelphis Dnmt1o protein (Mr
156,296), we wondered whether the Mr 160,000 proteinwas the Monodelphis Dnmt1o protein. Based on knowl-edge of expression of Dnmt1 proteins in the mouseovary, where the Mr 190,000 Dnmt1 form is expressed insomatic cells and the Mr 175,000 Dnmt1o form is exclu-sively expressed in oocytes (Ratnam et al., 2002), wewould expect that UPTC21 would detect only an Mr
160,000 protein in purified Monodelphis oocytes. In-deed, in a protein extract from 60 opossum oocytes,UPTC21 binds only to an Mr 160,000 protein, and therewas no evidence of binding to an Mr 175,000 protein(Fig. 3b). The specificity of UPTC21 for the MonodelphisMr 160,000 and Mr 175,000 Dnmt1 proteins was dem-onstrated by passing UPTC21 through a column withsaturating amounts of bound mouse Dnmt1 fragment(amino acids 636–1,108 in a 6-histidine-Dnmt1 fusionpeptide). UPTC21 treated in this way did not identify theMr 160,000 and the Mr 175,000 proteins in an extractfrom adult ovaries (data not shown). We conclude thatMonodelphis has an Mr 175,000 Dnmt1 protein in so-matic cells and an Mr 160,000 Dnmt1o protein in oo-cytes.
The conservation of expression of the Dnmt1o form ofthe Dnmt1 (cytosine-5)-methyltransferase enzyme in oo-cytes of the mouse and Monodelphis suggests thatDnmt1o is used in both species for the same purpose.Because the only known function of mouse Dnmt1o is tomaintain DNA methylation patterns on parental alleles of
imprinted genes during preimplantation development,opossum Dnmt1o would also be used to maintain im-printed methylation patterns. This is the first indication
FIG. 3. Immunoblot analysis of Monodelphis Dnmt1 proteins. a:Immunoblots of protein extracts probed with either the rabbitUPT82 antibody (left panel) or the chicken UPTC21 antibody (rightpanel). Extracts were made from Monodelphis spleen (Spl) andMonodelphis ovary (Ov). Protein sizes were determined by compar-ison to protein size standards (not shown) run on the same gel. b:Immunoblot of protein extract of 60 Monodelphis oocytes isolatedby hand from the ovaries of adults. For comparison, immunoblots ofextracts from mouse ES cells (M) and Monodelphis spleen (Spl) areshown. The ES cells are a mutant line that express approximatelyequivalent amounts of the mouse Mr 190,000 Dnmt1 and the mouseMr 175,000 Dnmt1o proteins (Ding and Chaillet, 2002). Normal EScells express only the Mr 190,000 Dnmt1 protein.
212 DING ET AL.
that DNA methylation probably does play a role in mar-supial imprinting, specifically through a maternal effectmechanism that maintains gametic methylation patternsin the early marsupial embryo.
Moore and Haig (1991) proposed that genomic im-printing evolved as a consequence of a parental struggleover maternal resources for the embryo. If this hypoth-esis is true, in oviparous species such as frogs andchicken, where zygotic gene expression has no controlover the consumption of maternal resources, thereshould be no genomic imprinting. Consistent with thisidea, genomic imprinting has not been described inchicken (O’Neill et al., 2000), and Xenopus eggs do notexpress a Dnmt1o protein, which is an essential compo-nent of imprinting in the mouse. Rather, Xenopus eggsexpress the same Dnmt1 protein that is expressed insomatic cells (Shi et al., 2001). Thus, the mechanism ofalternative first exon usage to encode for the two Dnmt1isoforms most likely evolved after the evolutionary diver-gence of amphibians and mammals, but before the splitbetween Metatheria and Eutheria.
MATERIALS AND METHODS
Two spleens and two ovaries for protein extracts wereobtained from two Monodelphis females (5 and 6months of age). The tissue was frozen in liquid nitrogen,ground to a fine powder, and dissolved in SDS samplebuffer. RNA for 5� RACE analysis was extracted from theother two ovaries by homogenizing the ovaries and ex-tracting the RNA with Ultraspec RNA solution (BioTecx,Houston, TX). Oocytes were obtained from all eightovaries of four adult Monodelphis females by pulling theovarian tissue between the two blades of a pair of for-ceps and collecting the released oocytes. Only oocytesfree or nearly free of granulosa cells were collected. Alloocytes collected in this manner, regardless of size, werepooled and frozen at –80°C. The entire collection wasthawed and immediately suspended in SDS protein gelloading buffer just prior to immunoblot analysis. Immu-noblot assays were performed using the UPTC21chicken antimouse Dnmt1 polyclonal antiserum or theUPT82 rabbit antimouse polyclonal antiserum, both at1:1,000 dilutions (Ratnam et al., 2002). UPTC21 wasaffinity purified on a column containing a 6-histidine-Dnmt1 (amino acids 636–1,108) fusion protein, andUPT82 was purified on a column containing a 6-histidine-Dnmt1 (amino acids 1–118) fusion protein. The anti-chicken IgY-HRP secondary antibody (Jackson Immuno-Research, West Grove, PA) was used at a 1:50,000dilution. ECL detection kit was from Amersham-Pharma-cia (Piscataway, NJ).
A complete Monodelphis Dnmt1 cDNA sequence wasobtained by screening a spleen cDNA library (Stratagene,La Jolla, CA). The initial screening of the library used a1.5-kb AccI fragment of the mouse Dnmt1 cDNA, in aregion of significant sequence conservation. Sequence of
the entire Monodelphis Dnmt1 gene (GenBank acces-sion number AF527541) was obtained by screening thelibrary again with Monodelphis Dnmt1 cDNA fragmentsobtained from the initial library screening.
5� RACE was performed on mRNA isolated from twoovaries using the Marathon cDNA kit (ClonTech, PaloAlto, CA) and oligonucleotides OPMT2 (CTCCAGTT-GCTCCTTCTGACTTCC) and OPMT3 (GTAACCCTT-GAGCTAGCTGAAG). 5� RACE cDNA fragments werethen cloned into a plasmid vector and sequenced. Align-ment of mouse and opossum Dnmt1 nucleotide se-quences was done with AssemblyLIGN. The derivedamino acid sequence of mouse and opossum Dnmt1were compared and aligned using the Clustaw Formatprogram of MacVector with identity scoring matrix.
ACKNOWLEDGMENT
We thank Dr. John Harder for help with preliminaryaspects of this work.
LITERATURE CITED
Bartolomei M, Tilghman SM. 1997. Genomic imprinting in mammals.Annu Rev Genet 31:493–525.
Carlson LL, Page AW, Bestor TH. 1992. Properties and localization ofDNA methyltransferase in preimplantation mouse embryos: impli-cations for genomic imprinting. Genes Dev 6:2536–2541.
Ding F, Chaillet JR. 2002. In vivo stabilization of the Dnmt1 (cytosine-5)-methyltransferase protein. Proc Natl Acad Sci USA 99:14861–14866.
Howell CY, Bestor TH, Ding F, Latham KE, Mertineit C, Trasler JM,Chaillet JR. 2001. Genomic imprinting disrupted by a maternaleffect mutation in the Dnmt1 gene. Cell 104:829–838.
Killian JK, Byrd JC, Jirtle JV, Munday BL, Stoskopf MK, MacDonald RG,Jirtle RL. 2000. M6P/IGF2R imprinting evolution in mammals. MolCell 5:707–716.
Killian JK, Nolan CM, Wylie AA, Li T, Vu TH, Hoffman AR, Jirtle RL.2001. Divergent evolution in M6P/IGF2R imprinting from theJurassic to the Quaternary. Hum Mol Genet 10:1721–1728.
Kozak M. 1987. An analysis of 5�-noncoding sequences from 699vertebrate messenger RNAs. Nucleic Acids Res 15:8125–8148.
Li E, Beard C, Jaenisch R. 1993. Role for DNA methylation in genomicimprinting. Nature 366:362–365.
Mertineit C, Yoder JA, Taketo T, Laird DW, Trasler JM, Bestor TH. 1998.Sex-specific exons control DNA methyltransferase in mammaliangerm cells. Development 125:889–897.
Moore T, Haig D. 1991. Genomic imprinting in mammalian develop-ment: a parental tug-of-war. Trends Genet 7:1–4.
O’Neill MJ, Ingram RS, Vrana PB, Tilghman SM. 2000. Allelic expres-sion of Igf2 in marsupials and birds. Dev Genes Evol 210:18–20.
Ratnam S, Mertineit C, Ding F, Howell CY, Clarke HJ, Bestor TH,Chaillet JR, Trasler JM. 2002. Dynamics of Dnmt1 methyltrans-ferase expression and intracellular localization during oogenesisand preimplantation development. Dev Biol 245:304–314.
Reik W, Walter J. 1998. Imprinting mechanisms in mammals. CurrOpin Genet Dev 8:154–164.
Shi L, Suetake I, Kawakami T, Aimoto S, Tajima S. 2001. Xenopus eggsexpress an identical DNA methyltransferase, Dnmt1, to somaticcells. J Biochem 130:359–366.
Solter D. 1988. Differential imprinting and expression of maternal andpaternal genomes. Annu Rev Genet 22:127–146.
Trasler JM, Trasler DG, Bestor TH, Li E, Ghibu F. 1996. DNA methyl-transferase in normal and Dnmtn/Dnmtn mouse embryos. DevDyn 206:239–247.
213MONODELPHIS DnmT1 METHYLTRANSFERASE
