Organization and Parent-of-Origin-Specific Methylation of Imprinted Peg3 Gene on Mouse Proximal Chromosome 7

Genomics 63, 333–340 (2000) doi:10.1006/geno.1999.6103, available online at http://www.idealibrary.com on

Organization and Parent-of-Origin-Specific Methylation of Imprinted Peg3 Gene on Mouse Proximal Chromosome 7 Li-Lan Li,* ,1 Irene Yuk-yee Szeto,* Bruce M. Cattanach,† Fumitoshi Ishino,‡ and M. Azim Surani* ,2 *Wellcome/CRC Institute of Cancer and Developmental Biology, University of Cambridge, Cambridge CB2 1QR, United Kingdom; †Mammalian Genetics Unit, Medical Research Council, Harwell, Didcot, Oxon OX11 ORD, United Kingdom; and ‡Gene Research Centre, Tokyo Institute of Technology, Yokohama, Japan Received October 28, 1999; accepted December 17, 1999

Peg3 is the first imprinted gene to be identified on mouse proximal chromosome 7; the human PEG3 homologue is on chromosome 19q13.4. Peg3 encodes a C 2H 2-type zinc finger protein that is expressed only from the paternal allele in embryos and adult brain. The gene has been shown to regulate maternal behavior and offspring growth and has been implicated in the TNF–NF␬B signal pathway. Here we show that Peg3 consists of nine exons spanning 26 kb. The 5ⴕ region of the gene contains a region rich in repeated sequences and a CpG island. Analysis of expressed sequence tags revealed a transcript present upstream of the island and on the strand opposite to Peg3. These structural features and DNA sequences are conserved in mouse and human. The 5ⴕ region of Peg3 is preferentially methylated on the inactive maternal allele, as shown by comparing embryos with paternal (PatDp. prox7) and maternal (MatDp.prox7) duplication of proximal chromosome 7. Recently, a new maternally expressed Zim1 gene located downstream of Peg3 was identified, which suggested that another imprinted cluster is present on proximal chromosome 7. © 2000 Academic Press

INTRODUCTION

Paternal and maternal genomes contribute unequally to mammalian development (McGrath and Solter, 1984; Surani et al., 1984) through parent-oforigin-dependent monoallelic expression of imprinted genes. The underlying mechanism involves heritable germ-line-specific epigenetic modifications (imprints) Sequence data from this article have been deposited with the GenBank/EMBL Data Libraries under Accession Nos. AF105262– AF105266. 1 Present address: Department of Genetics and Development, Columbia University, New York, NY 10032. 2 To whom correspondence should be addressed at Wellcome/CRC Institute of Cancer and Developmental Biology, University of Cambridge, CB2 1QR, UK. Telephone: (01223)-334136. Fax: (01223)334182. E-mail: [email protected].

that mark the parental origin of these genes. These imprints subsequently affect transcription during embryogenesis and in adults. Several structural features are apparently shared by imprinted genes (Reik and Walter, 1998). These genes often have few and small introns (Hurst et al., 1996). They are often present as a cluster of disparate genes within certain chromosomal domains. Some of these genes encode untranslated mRNAs (Pfeifer and Tilghman, 1994). Imprinted genes also usually contain CpG-rich regions that are associated with direct tandem repeats (Neumann et al., 1995) and are often subject to parental allele-specific differential methylation. Demethylation by a targeted mutation of the Dnmt1 gene abolished allele-specific expression of several imprinted genes (Li et al., 1993), but not all tested genes were affected (Caspary et al., 1998). It has not been possible so far to define a set of sequences and/or a unique structure that confers imprinted gene expression. We previously identified Peg3 as the first imprinted gene on mouse proximal chromosome 7 that is expressed from the paternal allele in embryonic mesoendodermal tissues, the hypothalamus, and the adult brain (Kuroiwa et al., 1996; Relaix et al., 1996; Kim et al., 1997; Li et al., 1999). Peg3 was also isolated independently by differential screening for genes involved in myogenesis (Relaix et al., 1996). The Peg3 protein contains 12 C 2H 2-type zinc-finger motifs and two proline/acidic amino acid-rich repeat domains (Kuroiwa et al., 1996; Relaix et al., 1996), suggesting its DNA- and protein-binding capacities, respectively. The protein was recently implicated as a partner for TRAF2 in the TNF–NF␬B signaling pathway controlling cell proliferation, cell differentiation, and apoptosis (Relaix et al., 1998). Targeted mutation of Peg3 resulted in abnormal maternal behavior and growth retardation (Li et al., 1999). Genetic studies previously showed that MatDp.prox7 3 caused neonatal lethality and intrauter3 Abbreviations used: dpc, day postcoitum; EST, expressed sequence tag; ISH, in situ hybridization; MatDp.prox7, maternal du-

333

0888-7543/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.

334

LI ET AL.

ine growth retardation, while PatDp.prox7 may be associated with reduced postnatal viability, growth, and behavior phenotypes (Searle and Beechey, 1990; Cattanach et al., 1992). Peg3 is likely to contribute in part to the MatDp.prox7 phenotypes. The human PEG3 homologue was localized to the syntenic region on chromosome 19q13.4 (Kim et al., 1997), but its imprinting status remains unknown. We report here the characterization of the Peg3 gene with special attention to its 5⬘ region. We also examined its epigenetic modifications associated with parental origin by analyzing MatDp.prox7 and PatDp.prox7 embryos. MATERIALS AND METHODS Reverse transcription-polymerase chain reaction (RT-PCR) and Northern analysis. Total RNA was prepared from adult brain of 129/Sv mouse strain using the TRIzol reagent (Gibco BRL). The reverse transcription reaction was performed using random hexamers as primers with Superscript II (Gibco BRL). The RT product (1/20 vol of the reaction), genomic DNA (5 ng), and/or plasmid DNA (5 pg) were used for PCR under the following conditions: 3 min of denaturation at 93°C, 32 amplification cycles (30 s at 93°C, 30 s at 64°C, and 2 min at 72°C), and 5 min of extension at 72°C. The primer pairs used (Fig. 1A) were P1, tgtgggtgtgaaagatcgtgtgtc; R1, tcttccaactctccaggacacttc; F1, acgtgaacagagtttacctgctgg; R2, cacaggaaatactagaacagctatc; F2, actgtccccaccatccaaagatag; and R3, ctcactgtaattccgccacaactg. One RT-PCR product (b) was labeled with [␣- 32P]dCTP using the random primers DNA labeling kit (Gibco BRL) and hybridized with a Northern blot (Clontech) as described by the manufacturer. The membrane was washed sequentially in 2⫻ SSC/0.1% SDS at room temperature, 0.2⫻ SSC/0.1% SDS at 65°C, and 0.1⫻ SSC/ 0.1% SDS at 65°C. Isolation of Peg3 genomic clones. Peg3 cDNA clones CS46 and p3-17 (Kuroiwa et al., 1996) were used as probes to screen phage genomic libraries ␭K0, ␭PS (Nehls et al., 1994), and ␭2001 (provided by A. Smith) of 129/Sv mouse strain DNA. After three rounds of screening, several positive clones were identified and characterized by restriction mapping. Three overlapping clones were further analyzed and used in this study (Fig. 1B): pB and pS were obtained from the ␭K0 and ␭PS libraries, respectively, and excised in plasmid form after infection of Cre recombinase-expressing Escherichia coli strain BNN132 (Nehls et al., 1994), and 1A was isolated from the ␭2001 library and after digestion with XhoI (the cloning site in the ␭2001 library) and EagI. The fragments 5⬘ and 3⬘ of the EagI site were subcloned into Bluescript II KS(⫹) as two plasmids, 1A-5 and 1A-3. In situ hybridization. MatDp.prox7 and normal littermate embryos generated by using the T9H translocation breakpoint (Searle and Beechey, 1990) were provided by J. Jones. A Peg3 cDNA clone, CS46 (Kuroiwa et al., 1996), was used for in vitro transcription with T7 and T3 RNA polymerase to prepare digoxigenin–UTP-labeled antisense and sense probes (Boehringer Mannheim). In situ hybridization (ISH) on sections was carried out as described previously (Kikyo et al., 1997). Methylation analysis. Genomic DNA was isolated from embryos of 129/Sv mouse strain using the proteinase K/SDS lysis solution as described (Hogan et al., 1994). Probes used for methylation analysis (Fig. 4A, gray boxes 1–5) were as follows: Probe 1, a 1.3-kb NheI fragment isolated from 1A-3; probe 2, a 4.8-kb ApaI–NheI fragment isolated from 1A-3; probe 3, a 5⬘ 2.8-kb fragment of pS purified following KpnI and NotI digestion; probe 4, a 4.5-kb XbaI fragment

plication and paternal deficiency of proximal chromosome 7; ORF, open reading frame; PatDp.prox7, paternal duplication and maternal deficiency of proximal chromosome 7; UTR, untranslated region.

isolated from pB; and probe 5, the 3.8-kb cDNA clone CS46. Five to ten micrograms of DNA was digested with appropriate enzymes, subjected to agarose gel electrophoresis, and transferred to Hybond-N membrane (Amersham). Probe labeling, blot hybridization, and washing were performed as described for Northern analysis. DNA sequencing and sequence analysis. Plasmid DNAs or PCRamplified fragments were sequenced using the ABI Prism DNA sequencing kit (Perkin–Elmer). The 6-kb 5⬘ flanking sequence of Peg3 was determined from 1A-5, the 5⬘ end of 1A-3, and a PCR product amplified from 1A and connecting these two clones. The most 3⬘ 2.5-kb sequence of PEG3 cDNA (GenBank Accession No. AB006625) is homologous to 11 mouse ESTs (Accession Nos. AA408083, AA098260, AA596814, AA415217, W34606; AA003064, AA124394, AA657156; AA241970, AA408082, AA172673, AA726627, AA409557), which form three contigs of 1.0, 0.7, and 0.7 kb size, respectively.

RESULTS

Gene structure of the mouse Peg3 gene. The Peg3 transcript is approximately 9 kb based on Northern analysis (Kuroiwa et al., 1996; Relaix et al., 1996). We previously obtained a 5.9-kb Peg3 cDNA (p3-17 and CS46), encoding a 1571-aa open reading frame (ORF) and 0.37-kb 5⬘ and 1-kb 3⬘ untranslated regions (UTR) (Kuroiwa et al., 1996; GenBank Accession No. AF038939). An additional 2.7-kb 3⬘ cDNA sequence was cloned by comparison to the human PEG3 homologue, which encodes a similar-sized transcript and shares significant similarity at both the amino acid and the nucleotide level (Kim et al., 1997). One 7-kb PEG3 cDNA clone (GenBank Accession No. AB006625) is 2.5 kb longer than the mouse clone, CS46, at the 3⬘ UTR. A TBLASTN search of the dbEST GenBank using this sequence detected 11 mouse ESTs, which form three contigs (see Material and Methods). Using primers specific to CS46 and these contigs, we amplified specific products (a– c) from adult mouse brain cDNA by PCR (Fig. 1A, left), showing that they are a contiguous transcript. Northern analysis using the product b detected a transcript of the expected size in embryos (Fig. 1A, right), confirming that they are 3⬘ UTR of Peg3. The previously reported cDNA, together with these RTPCR products and ESTs, forms an 8.7-kb sequence, which terminates in a poly(A) tract. We were unable to detect further 5⬘ cDNA clones by rapid amplification of cDNA ends (data not shown). These results indicated that the 8.7-kb cDNA sequence was nearly full length because its size is very close to the estimated size of Peg3 mRNA. Peg3 genomic clones (pB, pS, 1A-5, and 1A-3) were isolated from phage libraries, and a restriction enzyme map of this locus was determined (Figs. 1B and 4A). For the exon–intron organization of the Peg3 gene, we first performed PCR on the cDNA clones (or RT products) and genomic clones (or genomic DNA) using primers designed from the cDNA sequence (Fig. 1A, left; data not shown). This analysis indicated the presence of one large exon at the 3⬘ end of the gene. The precise exon–intron boundary was subsequently determined by sequencing genomic clones. The transcribed 8.7-kb sequence is divided into nine

335

METHYLATION OF IMPRINTED Peg3 GENE

FIG. 1. Organization of the Peg3 gene. (A) Cloning of Peg3 3⬘ UTR. Left, RT-PCR analysis using primers specific to previously reported Peg3 cDNA (P1) and/or EST contigs (F1–F2, R1–R3). cDNA was synthesized from adult brain RNA with (RT ⫹) or without (RT ⫺) reverse transcriptase. Genomic DNA and a Peg3 genomic clone, positive control and negative (N) control for PCR. Right, Northern analysis of poly(A) ⫹ RNA (2 ␮g/lane) from 13.5-dpc mouse embryos with the RT-PCR product b. (B) Exon–intron organization of Peg3. Boxes 1 to 9 represent exons and black areas indicate the coding region. cDNA clones are shown above the gene map and genomic clones below the gene map. Dashed lines indicate splicing of Peg3 transcript. The CpG island and an EST oriented on the opposite strand are indicated. The precise location of exon 2 has not been determined.

exons, with eight small exons of 90 –160 bp followed by one large exon of 7.9 kb, and spans 26 kb of the genome (Fig. 1B, Table 1; designated exons 1 to 9). All the exon–intron junctions conform to the AG/GT

rules (Breathnach and Chambon, 1981). The Peg3 ORF begins within exon 3 and terminates within exon 9 (nucleotides 371–5086 in the 8.7-kb cDNA), approximately 90% of which (including the se-

TABLE 1 The Exon–Intron Structure of the Peg3 Gene Exon

1 2 3 4 5 6 7 8 9

Exon size (bp) 1 170 ..........AGCCCTTGCT..(170)..GCTCCGGGAGgtgagtcagc 171 255 ttctccatagCCCTACCTTC..(85)...GCTGGTCCAGgtgagtgata 256 386 ttggttgcagGCAGGCCTTC..(131)..CATCACGAAGgtgaggcact 387 476 gtcttctcagACGACACCAA..(90)...GCTTTTGGCAgtgagtatca 477 560 acctgcctagGTGAGCGAGA..(84)...CCCAGAAGCAgtaagttgtc 561 664 tccttcctagGGCTGCCTCA..(104)..TCGATCCCAGgtatgcccag 665 767 gatgttttagGATGCCGAGT..(103)..CTCTCGCTGGgtaaagacct 768 860 ataactatagGAGTCCAGCT..(93)...ACCAGCCGAGgtgagtgaca 861 ttctttccagGTCTCAAACC..(⬎7900)

Intron size (kb)

Accession No.

AF105262 11.415 AF105263 0.54

AF105264

0.60

AF105264

1.00

AF105264

3.29

AF105265

0.73

AF105266

0.12

AF105266 AF105266

Note. Exon and intron sequences are in uppercase and lowercase letters, respectively. Numbers shown above the exon sequence represent the nucleotide number in the 8.7-kb Peg3 cDNA (Accession No. AF038939). The sizes of the introns were deduced from the locations of exons in the genomic region determined by restriction mapping and/or Southern analysis using exon-specific oligonucleotides as probes.

336

LI ET AL.

quences encoding all zinc finger motifs and amino acid repeats) resides in the last exon. Features of Peg3 5⬘ region. We sequenced the 6-kb 5⬘ region of the Peg3 gene comprising 2.8 kb of 5⬘ flanking sequences, exon 1, and 3.1 kb of intron 1. Sequence analysis revealed that this region shares several features with the reverse sequence of nucleotides 160367–166367 in a human BAC clone (GenBank Accession No. AC006115) (Fig. 2), which we conclude represents the PEG3 gene. First, the 5⬘ 3-kb region corresponding to Peg3 exon 1 and its upstream region showed an overall 70% sequence similarity to the human counterpart (Fig. 2A). High homology was apparent in stretches of sequences including those in exon 1. Several potential transcription factor binding sites were present in both the mouse and the human genes, including one for SOX-5, three for NFAT, one for CREBP1cJUN sites located upstream of exon 1, one for AP2, and one for DELTAEF1 sites present within exon 1. Sequence conservation suggests that this region contains a promoter, and here we referred to the human counterpart as the PEG3 5⬘ region. Second, the region 3⬘ of this conserved element is GC rich. CpGplot analysis revealed a 1.1-kb region in the Peg3 gene (nucleotides ⫺42 to ⫹1077 relative to the transcription start site) and a 2.3-kb region in the PEG3 gene (reverse of AC006115, nucleotides 161305–163624) as CpG islands (Bird, 1986). (Fig. 2B). Third, a region closely associated with the CpG island is rich in repetitive sequences (Figs. 2A and 2C). The Peg3 intron 1 region 1.4 –3.2 kb downstream of the transcription start site contains direct repeats (five 61-bp, eight 30-bp, and eight 24-bp elements) and two copies of a 72-bp repeat at the 3⬘ end, which all have ⬎80% internal sequence homology. There are two complete and five partial copies of a 108-bp element in a 1.1-kb region (reverse of AC006115, nucleotides 161667–162767) within the PEG3 CpG island. The 72- and 61-bp repeats are unique to the Peg3 gene and the 108-bp repeat to the PEG3 gene. The 30- and 24-bp repeats are present in six and three copies, respectively, in the PEG3 gene. Although these two repeats partially match to some sequences in the genome, they were found only as single copies. The reiteration of these sequences on both Peg3/PEG3 genes appears not to be by chance. Finally, sequence similarity searches detected an EST located on the opposite strand and upstream of Peg3

exon 1 and the human corresponding region (Fig. 2A). The mouse EST (Accession Nos. W49208, W90868) of 440 bp contains 104 bp identical to the Peg3 sequence. The 450-bp human EST (Accession Nos. HS201228 and HS403225) is almost identical to the PEG3 sequence. No repeats were detected within these homologous regions. Parental-origin-specific differential methylation. Peg3 mRNA is expressed from the paternal allele in embryonic mesoendodermal tissues (Kuroiwa et al., 1996; Li et al., 1999; see Fig. 3, left). When we examined MatDp.prox7 embryos at 13.5 dpc by ISH, no Peg3 transcript was detectable from the maternal alleles (Fig. 3, right). Since DNA methylation is an important epigenetic modification involved in gene silencing (Li et al., 1993), we used DNA from normal, MatDp.prox7, and PatDp.prox7 embryos at 13.5 dpc to search for parental-origin-specific methylation of Peg3. Figure 4A summarizes the methylation pattern of the Peg3 locus obtained by using Southern hybridization and methylation-sensitive enzymes (HpaII, EagI, PvuI, SacII, and ClaI). The Peg3 5⬘ 6-kb region contains 11 HpaII, 1 EagI, 1 PvuI, and 2 tightly linked SacII sites. Methylation of these sites was investigated in a 20-kb KpnI fragment using probe 1 (Fig. 4B). In normal embryos, this fragment was partially cleaved by HpaII into very small pieces. Complete digestion of this fragment by the methylation-insensitive isoschizomer, MspI, demonstrated that the HpaII sites in this region were partially methylated. Similarly, the EagI, PvuI, and SacII sites within the CpG island were partially methylated, where ⬃50% of this fragment was cut to 9 kb. This partial methylation of the region was allele-specific. The fragment from MatDp.prox7 embryos remained largely uncut by these enzymes, showing that these sites were methylated when maternally inherited. The paternal copies in PatDp.prox7 embryos were sensitive to these enzymes and hence unmethylated. Three HpaII sites (Hp1–3) were mapped within the region encoding the 5⬘ UTR of Peg3. They were analyzed in a 12.2-kb EcoRI fragment in normal embryos with probe 2 (Fig. 4B). MspI digestion yielded two small bands of 1.7 and 4.2 kb, but, upon HpaII digestion, a 7.5-kb and the uncut 12.2-kb band were detected. Therefore the Hp2 and Hp3 sites were methylated, and the distal Hp1 site was partially methylated.

FIG. 2. Characteristics of the Peg3 and PEG3 5⬘ regions. (A) Comparison of the Peg3 (GenBank Accession No. AF105262) and PEG3 (reverse of GenBank Accession No. AC006115, nucleotides 160367–166367) 5⬘ nucleotide sequences using the Compare program of GCG package. The position of Peg3 exon 1 (open box, 1), the reverse strand of ESTs [open boxes, W49208 (⫺)/W90868 (⫺) and HS201228 (⫺)/HS403225 (⫺)], CpG islands (shaded areas), and tandem direct repeat-rich regions (rectangular boxes) are indicated. The human region showing strong homology to Peg3 exon 1 is referred to as PEG3 putative exon 1 (dash lined open box, 1). Arrowheads, transcription direction. (B) CpGplot analysis of the Peg3 (left) and PEG3 (right) 5⬘ sequences. A CpG island spans the region ⫺0.042 to ⫹1.077 kb relative to the transcription start site of Peg3 [CG obs/exp (black line) ⬎0.6 (0.63), C%⫹G% (gray line) ⬎50% (60%), length ⬎1 kb (1.1 kb)] and the region (reverse of GenBank Accession No. AC006115, nucleotides 161305–163624) in the PEG3 gene [CG obs/exp ⬎0.6 (0.62), C%⫹G% ⬎50% (64%), length ⬎1 kb (2.3 kb)]. (C) Self-alignment of the Peg3 (left) and PEG3 (right) 5⬘ sequences by the Compare program (window 25, stringency 20). The repeat regions are enlarged below to show size and copy number of repetitive sequences. The sequence that is present in more than 50% of copies at that position is shown, and a dash (–) indicates no sequence present in ⬎50%, as determined by ClustalW (1.7). The repetitive sequences conserved between human and mouse genes are in bold.

METHYLATION OF IMPRINTED Peg3 GENE

337

338

LI ET AL.

FIG. 3. Absence of Peg3 expression in MatDp.prox7 embryos. ISH analysis of 13.5-dpc normal (left) and MatDp.prox7 (right) embryos with a Peg3 antisense RNA probe. fs, facial skeleton; h, heart; hy, hypothalamus; li, liver; nt, neural tube; p, pituitary gland; t, tongue; vc, vertebral cartilage.

Like the CpG island region, the Hp1 site was methylated on the maternal allele, as shown by the fact that HpaII digestion of the KpnI fragment from PatDp.prox7 embryos produced a 6.6-kb product, while the same fragment from MatDp.prox7 embryos remained uncut, when detected with probe 3 (Fig. 4B). This site may represent the 3⬘ boundary of the differentially methylated region. The region spanning the coding region and 3⬘ UTR of Peg3 was methylated on both parental alleles. Two HpaII sites (Hp4 and 5) and one ClaI site located together in exon 9 were all methylated, as indicated by resistance to digestion of the 11.1-kb EcoRI fragment by HpaII (but not MspI) and ClaI using probe 4 (Fig. 4B). A similar methylation pattern was observed for three HpaII sites (Hp6 –9) further at the 3⬘ end, as revealed in KpnI/HpaII or KpnI/MspI digestion using probe 5 (Fig. 4B). DISCUSSION

We have characterized the Peg3 gene structure and compared it to the human homologue, which showed that certain structural features of Peg3 are shared with other imprinted genes. Peg3 encodes an 8.7-kb transcript consisting of nine exons that spreads over 26 kb of the genome. Its ORF of 4.7 kb spans from exons 3 to 9. Relaix et al. (1996) obtained an 8.4-kb DNA sequence (GenBank Accession No. U48804) from 5⬘ genomic and 3⬘ cDNA clones of Pw1. This sequence was found to represent the Peg3 genomic sequence spanning from exon 6 to the coding region of exon 9. Based on RT-PCR analysis, Relaix et al. (1996) suggested that the Peg3/Pw1 gene was intronless and encoded mainly by this sequence. However, we obtained further 5⬘ (p3-17) and 3⬘ (3⬘ UTR) cDNA clones from the adult brain, which is the same tissue analyzed by Relaix et al. (1996). These exons are

part of the Peg3 transcript, because a promoterless ␤geo cassette inserted within exon 5 showed an expression pattern similar to that of Peg3 (Li et al., 1999). Furthermore, there is a high degree of homology between the 3⬘ UTRs of Peg3 and PEG3. These results show that Peg3 is clearly not intronless. Although the gene structure of PEG3 has not yet been fully characterized, our comparative analysis indicates evolutionary conservation of this gene at the level of protein, nucleotide sequence, and gene organization. Previously, Kim et al. (1997) reported conservation of the 3⬘ region of the gene in mice and human, with 83% amino acid similarity and the absence of introns. We detected significant sequence homology between Peg3 exons 1, 2, 5, 8, and 9 with distinct regions in the PEG3 BAC clone (GenBank Accession No. AC006115). There were also striking sequence and organizational similarities between the 5⬘ regions of both genes: the presence of a CpG island, closely associated direct repeats, several conserved potential transcription factor binding sites, and an EST located on the opposite strand upstream of the island. The strong evolutionary conservation is indicative of important functions for the gene. Indeed, we recently showed that the mouse gene played a role in the regulation of maternal nurturing behavior (Li et al., 1999). Hurst et al. (1996) previously proposed that imprinted genes typically have few and small introns based on a comparison of 16 imprinted genes with 90 control genes. Peg3 has at least eight introns with an average size of 2211 bp. Thus, the organization of this gene does not agree with that view. However, the number of imprinted genes examined so far is still relatively small to reach a definitive conclusion concerning common features of imprinted genes. We showed that Peg3 was differentially methylated at the 5⬘ CpG island region. The two maternal copies were methylated in MatDp.prox7 embryos in which Peg3 was silent, consistent with the role of DNA methylation as an epigenetic modification involved in gene silencing (Li et al., 1993). We detected direct repeats associated with the 5⬘ CpG island of Peg3/PEG3 genes, as observed in several imprinted genes (Neumann et al., 1995). Some direct repeats are unique to Peg3 and to PEG3, but two repeats of short stretches of sequences were found in both genes. The functional significance of these sequences remains to be tested, although it is suggested that the secondary structure of repetitive elements may have a role in imprinting. While direct repeats are present in the cis elements required for imprinting of the Igf2r, H19, and RSVIgmyc transgenes (Reik and Walter, 1998), short direct repeats in the H19 gene do not play a role in imprinting (Stadnick et al., 1999). Arrays of multicopy transgenes are known to induce gene silencing in several organisms, including mammals (Garrick et al., 1998; reviewed in Henikoff, 1998). While short direct repeats found in the imprinted loci may not be equivalent to

METHYLATION OF IMPRINTED Peg3 GENE

339

FIG. 4. Parental-origin-specific methylation of the Peg3 locus. (A) Parental-origin-specific methylation of the Peg3 locus as determined by Southern analysis in (B). The region between the two KpnI (K*) sites contains several other KpnI sites that have not been mapped. N, NheI; R, EcoRI; Xb, XbaI; Xh, XhoI. The positions of methylation-sensitive restriction enzymes (C, ClaI; E, EagI; Hp, HpaII; P, PvuI; S, SacII) and probes 1–5 (gray boxes) used for Southern analysis are indicated. Open circles indicate no methylation at the site and solid circles represent methylation. Sites showing no methylation difference between parental alleles are indicated as circles on the lines; circles above the line represent paternal origin (P) and circles below the line, maternal origin (M). (B) Southern analysis. Genomic DNA (from normal, MatDp.prox7, or PatDp.prox7 embryos at 13.5 dpc) was digested with restriction enzymes and hybridized with probes as indicated. Digestion of DNA by the methylation-insensitive enzyme (M, MspI) served as a control.

these repeats, it is possible that these elements may tag imprinted loci and cooperate with differential DNA methylation and/or other unidentified factors to achieve imprinted gene expression. Sequence analysis revealed an EST present on the opposite strand and immediately upstream of the Peg3 and PEG3 genes. Our preliminary experiments suggest the existence of this mouse transcript. Kim et al. (1999) recently identified an imprinted Krup-

pel-type zinc-finger gene, Zim1/ZIM1, located downstream of Peg3/PEG3. Zim1 is transcribed from the opposite parental allele and in the direction opposite to Peg3. Thus, there are at least two imprinted genes on the mouse proximal chromosome 7. Clusters of imprinted genes are also seen on the mouse central and distal chromosome 7 and proximal chromosome 17. There is evidence for coordinate regulation of imprinting of these gene clusters (Leighton et al.,

340

LI ET AL.

1995; Wutz et al., 1997; Yang et al., 1998). In this context, the finding of a potential transcript associated with Peg3/PEG3 is of importance. Further studies will determine the structure and imprinted status of this mouse transcript and demonstrate whether there is a mechanistic link between imprinting of Peg3 and the neighboring genes. ACKNOWLEDGMENTS The authors thank Dr. J. Jones for preparing MatDp.prox7 and PatDp.prox7 embryos, Dr. T. Kohda for sharing information on PEG 3⬘ cDNA and homologous mouse ESTs, and Drs. J. F.-X. Ainscough, R. John, and T. H. Bestor for their valuable comments on the manuscript. L.-L. Li is very grateful to Dr. T. H. Bestor for his support while writing the manuscript in his lab. The work was supported by grants from the Wellcome Trust and Human Frontier Science Program to M.A.S.

REFERENCES Bird, A. P. (1986). CpG-rich islands and the function of DNA methylation. Nature 321: 209 –213. Breathnach, R., and Chambon, P. (1981). Organization and expression of eucaryotic split genes coding for proteins. Annu. Rev. Biochem. 50: 349 –383. Caspary, T., Cleary, M. A., Baker, C. C., Guan, X. J., and Tilghman, S. M. (1998). Multiple mechanisms regulate imprinting of the mouse distal chromosome 7 gene cluster. Mol. Cell. Biol. 18: 3466 – 3474. Cattanach, B. M., Barr, J. A., Evans, E. P., Burtenshaw, M., Beechey, C. V., Leff, S. E., Brannan, C. I., Copeland, N. G., Jenkins, N. A., and Jones, J. (1992). A candidate mouse model for Prader–Willi syndrome which shows an absence of Snrpn expression. Nat. Genet. 2: 270 –274. Garrick, D., Fiering, S., Martin, D. I., and Whitelaw, E. (1998). Repeat-induced gene silencing in mammals. Nat. Genet. 18: 56 –59. Henikoff, S. (1998). Conspiracy of silence among repeated transgenes. BioEssays 20: 532–535. Hogan, B., Beddington, R., Costantini, F., and Lacy, E. (1994). “Manipulating the Mouse Embryo: A Laboratory Manual,” Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Hurst, L. D., McVean, G., and Moore, T. (1996). Imprinted genes have few and small introns. Nat. Genet. 12: 234 –237. Kikyo, N., Williamson, C. M., John, R. M., Barton, S. C., Beechey, C. V., Ball, S. T., Cattanach, B. M., Surani, M. A., and Peters, J. (1997). Genetic and functional analysis of neuronatin in mice with maternal or paternal duplication of distal Chr 2. Dev. Biol. 190: 66 –77. Kim, J., Ashworth, L., Branscomb, E., and Stubbs, L. (1997). The human homologue of a mouse-imprinted gene, Peg3, maps to a zinc finger gene-rich region of human chromosome 19q13.4. Genome Res. 7: 532–540.

Kim, J., Lu, X., and Stubbs, L. (1999). Zim1, a maternally expressed mouse kruppel-type zinc-finger gene located in proximal chromosome 7. Hum. Mol. Genet. 8: 847– 854. Kuroiwa, Y., Kaneko Ishino, T., Kagitani, F., Kohda, T., Li, L.-L., Tada, M., Suzuki, R., Yokoyama, M., Shiroishi, T., Wakana, S., Barton, S. C., Ishino, F., and Surani, M. A. (1996). Peg3 imprinted gene on proximal chromosome 7 encodes for a zinc finger protein. Nat. Genet. 12: 186 –190. Leighton, P. A., Ingram, R. S., Eggenschwiler, J., Efstratiadis, A., and Tilghman, S. M. (1995). Disruption of imprinting caused by deletion of the H19 gene region in mice. Nature 375: 34 –39. Li, E., Beard, C., and Jaenisch, R. (1993). Role of DNA methylation in genomic imprinting. Nature 366: 362–365. Li, L.-L., Keverne, E. B., Aparicio, S. A., Ishino, F., Barton, S. C., and Surani, M. A. (1999). Regulation of maternal behavior and offspring growth by paternally expressed Peg3. Science 284: 330 – 333. McGrath, J., and Solter, D. (1984). Completion of mouse embryogenesis requires both the maternal and paternal genomes. Cell 37: 179 –183. Nehls, M., Messerle, M., Sirulnik, A., Smith, A. J., and Boehm, T. (1994). Two large insert vectors, ␭PS and ␭KO, facilitate rapid mapping and targeted disruption of mammalian genes. Biotechniques 17: 770 –775. Neumann, B., Kubicka, P., and Barlow, D. P. (1995). Characteristics of imprinted genes. Nat. Genet. 9: 12–13. Pfeifer, K., and Tilghman, S. M. (1994). Allele-specific gene expression in mammals: The curious case of the imprinted RNAs. Genes Dev. 8: 1867–1874. Reik, W., and Walter, J. (1998). Imprinting mechanisms in mammals. Curr. Opin. Genet. Dev. 8: 154 –164. Relaix, F., Wei, X. J., Wu, X. W., and Sassoon, D. A. (1998). Peg3/Pw1 is an imprinted gene involved in the TNF-NF kappa B signal transduction pathway. Nat. Genet. 18: 287–291. Relaix, F., Weng, X., Marazzi, G., Yang, E., Copeland, N., Jenkins, N., Spence, S. E., and Sassoon, D. (1996). Pw1, a novel zinc finger gene implicated in the myogenic and neuronal lineages. Dev. Biol. 177: 383–396. Searle, A. G., and Beechey, C. V. (1990). Genome imprinting phenomena on mouse chromosome 7. Genet. Res. 56: 237–244. Stadnick, M. P., Pieracci, F. M., Cranston, M. J., Taksel, E., Thorvaldsen, J. L., and Bartolomei, M. S. (1999). Role of a 461-bp G-rich repetitive element in H19 transgene imprinting. Dev. Genes Evol. 209: 239 –248. Surani, M. A. H., Barton, S. C., and Norris, M. L. (1984). Development of reconstituted mouse eggs suggests imprinting of the genome during gametogenesis. Nature 308: 548 –550. Wutz, A., Smrzka, O. W., Schweifer, N., Schellander, K., Wagner, E. F., and Barlow, D. P. (1997). Imprinted expression of the Igf2r gene depends on an intronic CpG island. Nature 389: 745–749. Yang, T., Adamson, T. E., Resnick, J. L., Leff, S., Wevrick, R., Francke, U., Jenkins, N. A., Copeland, N. G., and Brannan, C. I. (1998). A mouse model for Prader–Willi syndrome imprintingcentre mutations. Nat. Genet. 19: 25–31.