Mit1/Lb9 and Copg2, new members of mouse imprinted genes closely linked to Peg1/Mest

Abstract

Two mouse genes, Mit1/Lb9 and Copg2, linked to Peg1/Mest on mouse chromosome 6, were identified to be imprinted maternally and paternally, respectively. Mit1/Lb9 encoding untranslated transcripts resides within the intron 20 of Copg2. The gene is maternally imprinted in adult mouse brain, partially imprinted in other tissues. Copg2 consists of 24 exons within the >40 kb genomic region, being expressed ubiquitously in mouse tissues with a partial imprinting pattern in embryos, neonates, and adult brain in contrast to maternally imprinted human COPG2. In addition, we identified an antisense transcript of Copg2, Copg2AS, which overlaps 3′-UTRs of Copg2 and Peg1/Mest. The Copg2AS transcript is maternally imprinted in embryos, neonates, and adult tissues.

1 Introduction

Genomic imprinting is an epigenetic mechanism in mammals that directs differential silencing of a gene depending on the parental allelic origin in the offspring's soma [1, 2]. More than two dozens of human and mouse genes have been identified to be imprinted through various efforts including genome-wide systematic searches for monoallelically expressed genes [3-9]. The clustering behavior of imprinted genes was exploited in a strategic screening for the identification of new imprinted genes. The increased wealth of genomic and EST database enabled us to ambulate around four genomic regions each of which contains only a single previously identified imprinted gene. We have identified two mouse imprinted genes, Mit1/Lb9 and Copg2, closely linked to Peg1/Mest. The expression profiles and imprinting patterns of the genes were determined.

2 Materials and methods

2.1 Determination of imprinting status

Human genes and ESTs closely linked to the known imprinted genes were chosen from Human GeneMap’99 (http://www.ncbi.nlm.nih.gov/genemap). Mouse homologues of the selected human genes were identified from the GenBank database using BLAST programs. Polymorphisms of the candidate genes were used to physically discriminate one of parental alleles from the other in the F1 hybrids crossed between two inbred mouse strains, C57BL/6J (a strain of Mus musculus domesticus) and KJR/Msf (a strain of Mus musculus molossinus). Restriction fragment length polymorphism (RFLP) analyses of RT-PCR products were carried out to determine the allele-specific expression of the genes in the (KJR/Msf×C57BL/6J)F1 and the (C57BL/6J×KJR/Msf)F1 hybrids. Total RNAs were isolated using Tri reagent (Molecular Research Center). Reverse transcription was performed with SuperScript II Reverse Transcriptase (GibcoBRL) using random hexamer for Mit1/Lb9 or strand-specific primers for Copg2 and Copg2AS. The PCR primers 4–5 (5′-AACAAAACTAGCTTTACTTGAGAG-3′) and 4–32 (5′-ATCTGTAACTGTAACCCTGGGTCG-3′) were used for amplification of the Mit1/Lb9 cDNA. The primer 4–32 ends at one nucleotide upstream to the polymorphic site and contains a C residue rather than a native G residue as the 3′-penultimate nucleotide. Therefore, the PCR product of the KJR/Msf allele would be digested by TaqI restriction enzyme while that of the C57BL/6J allele would not (see Fig. 4A). The PCR products were digested with TaqI restriction enzyme and analyzed on a 5% polyacrylamide gel. PCR amplifications were carried out for 33 cycles in 50 μl of 10 mM Tris–HCl (pH 8.3), 40 mM KCl, 1.5 mM MgCl₂, 1 mM DTT, 0.5 μg/ml acetylated BSA, and 200 μM of each dNTP with 50 pmol of each primer and 2.5 units of Taq DNA polymerase at 94°C for 30 s, at 55°C for 30 s, and at 72°C for 30 s. The PCR primers G2-F1 (5′-GCTGCTGCCTGGGAAGAGGT-3′) and G2-PF (5′-GAAGAAACTCTAAAGCTCATGCTC-3′) located in exon 20 and exon 24, respectively, were used for amplification of the Copg2 cDNA. To amplify the Copg2AS cDNA, the PCR primers G2-PR (5′-GGCCCTGCGCCAGCAGATCAAACA-3′) and Peg1-3UF (5′-CCTAAGAGCAAATGGTGCTG-3′) located in exon 24 of Copg2 and exon 12 of Peg1/Mest, respectively, were used. The PCR amplifications were performed under the same conditions as above except that the annealing temperature was 62°C. Genomic DNA was amplified using G2-PR and G2-PF primers. The PCR products of Copg2 and Copg2AS were digested with HhaI and BsmAI restriction enzymes, respectively and analyzed on a 20% polyacrylamide gel.

Figure 1

Determination of imprinting status of Mit1/Lb9. A: Expected restriction DNA fragments using mismatch PCR-mediated site-directed mutagenesis. The polymorphic site between C57BL/6J and KJR/Msf is indicated in boldface. The mismatched base in the primer 4–32 is shown in lowercase. Polymorphic TaqI site created in the KJR/Msf allele is underlined. B: Imprinting status of Mit1/Lb9 in adult tissues. RNAs derived from the (C57BL/6J×KJR/Msf)F1 hybrid (C×M) and the (KJR/Msf×C57BL/6J)F1 hybrid (M×C) were used for RT-PCR amplification. PCR products were digested with TaqI and separated on a 5% polyacrylamide gel. C: Imprinting status of Mit1/Lb9 in mouse embryos and neonates. C57BL, C57BL/6J allele; Mol, KJR/Msf allele; E17.5H, head of E17.5; E17.5B, body of E17.5; NeoH, head of neonate; NeoB, body of neonate.

2.2 Characterization of cDNA and genomic DNA clones

Six cDNA clones of Mit1/Lb9 and three cDNA clones of Copg2 were isolated from mouse brain cDNA library and testis cDNA library, respectively. Six genomic clones for Mit1/Lb9 and Copg2 were isolated from mouse 129 genomic DNA library. Library screening was carried out according to the standard protocol. The nucleotide sequences were determined using a TaqTrak Sequencing System (Promega) and a Bigdye Terminator cycle sequencing kit (Perkin-Elmer).

2.3 Southern blot and Northern blot analyses

Fifteen micrograms of mouse genomic DNA were digested with StuI, BglII, PvuII, HindIII, EcoRI, XbaI, or DraI, electrophoresed on a 0.8% agarose gel, and transferred to a Hybond-N⁺ membrane (Amersham) in the Southern hybridization experiment. Thirty micrograms of total RNAs were electrophoresed on a 1% agarose gel and transferred to a Hybond-N⁺ membrane for the Northern hybridization. The [α-³²P]dCTP random primed Mit1/Lb9 and Copg2 cDNAs were used as the probes. Hybridization was carried out at 65°C using the QuickHyb solution (Stratagene).

3 Results

3.1 Isolation and characterization of Mit1/Lb9 and Copg2

We investigated parent of origin-specific expression patterns of 12 mouse genes of which human homologues are closely linked to Peg1/Mest [10], PEG3 [11], PEG5/NNAT [12], and MEG1/GRB10 [13]. It has not been investigated whether the genomic regions of those imprinted genes harbor additional imprinted genes yet. The allelic expression patterns of the 12 candidates were determined in the F1 hybrids of C57BL/6J (a strain of Mus musculus domesticus) and KJR/Msf (a strain of Mus musculus molossinus). Mouse homologues of UBE2H, IRF5, KIAA0265, Cda15a02, and CALU linked to Peg1/Mest, WI-8030, EMAP2, AAD23609.1, and CGI-146 linked to PEG3, RBL1 linked to PEG5/NNAT, and TTC4 linked to MEG1/GRB10 were ascertained to be expressed biallelically in adult mouse brain, heart, lung, and kidney (data not shown). However, a gene closely linked to Peg1/Mest, a maternally imprinted gene on mouse chromosome 6 [3], was identified to be imprinted. It was found that the mouse cDNA sequence (GenBank accession no. AA611551), a homologue of a human clone 23582 (GenBank accession no. AF038190), was maternally imprinted in mouse brain (see below). The gene was designated as Mit1 (Mest-linked imprinted transcript 1; GenBank accession no. AF217545). It shares 67.3% nucleotide sequence identity with human clone 23582. Subsequently, cDNA and genomic DNA clones of Mit1 were isolated and characterized. Sequencing analyses showed that Mit1 is located within the intron 20 of an adjacent gene, Copg2 (Fig. 1), a homologue of coatomer protein complex subunit γ (γ-COP) [21] (Fig. 2A). Copg2 shares 80.8% amino acid sequence identity with mouse Copg1 (GenBank accession no. AF187079) (Fig. 1) and 97.6% amino acid sequence identity with human COPG2 [9]. A CpG island was found in the 5′-region of Copg2 (Fig. 2A). The 3′-UTR of Copg2 overlapped at its terminal 52 nucleotides with the 3′-UTR of Peg1/Mest (Fig. 2A). The overlapping is conserved in mouse, human [9], and rat (GenBank accession nos. AI408151 and AA851833). The exon–intron boundaries of mouse Copg2 and human COPG2 are conserved. Database analyses of the cDNA and genomic DNA sequences of Mit1 revealed that the previously identified Lb9 cDNA (GenBank accession no. U20263) was located at the upstream of Mit1 (Fig. 2A). RT-PCR experiments using the PCR primers, one of which annealed to the Lb9 cDNA and the other to the Mit1 cDNA, showed both sequences were within the same gene transcript (data not shown).

Figure 2

Comparison of deduced amino acid sequences between mouse Copg1 and mouse Copg2. Identical amino acids are shaded.

Figure 3

Genomic organization of Mit1/Lb9 and Copg2. A: Genomic structure of Mit1/Lb9 and Copg2. The protein coding regions are indicated by solid boxes. Open boxes represent 3′-UTRs of Copg2 and Peg1/Mest. Mit1 cDNA and Lb9 cDNA are indicated by hatched boxes. Introns of which the sizes were determined are represented by horizontal lines. Transcriptional orientations of Mit1/Lb9, Copg2, and Peg1/Mest are indicated by arrows. Translation start codon (ATG), stop codon (TAA), and CpG island of Copg2 are indicated. B: Southern blot analysis of Mit1/Lb9. The size markers are indicated in kb.

Mit1/Lb9 is a single copy gene as shown by Southern blot analysis (Fig. 2B). The transcription products of Mit1/Lb9 seem to be highly heterogeneous since multiple transcripts were detected in Northern blot analysis (Fig. 3A). In addition, we could find six different forms of the Lb9 cDNA (GenBank accession nos. U20262–U20267) from the GenBank database. It is likely that the observed multiple Mit1/Lb9 transcripts are generated by alternative splicing and/or differential usage of promoters and polyadenylation sites. It is of interest that no considerable open reading frame could be predicted from the sequences of Mit1/Lb9 cDNAs. It seems that Mit1/Lb9 functions at the RNA level as H19 [14], Igf2rAS [15], IPW [16], or LIT1 [17] does.

Figure 4

Expression analyses of Mit1/Lb9 and Copg2. Total RNAs from different tissues of mouse were blotted and hybridized with Mit1/Lb9 (A) or Copg2 (B) cDNA as a probe. Reprobing of the same blot with the β-actin probe was used for the control.

3.2 Expression of Mit1/Lb9 and Copg2

The expression patterns of Mit1/Lb9 and Copg2 were determined by Northern blot analyses with total RNAs isolated from various tissues (Fig. 3). The transcription level of Mit1/Lb9 varied among different tissues (Fig. 3A). The expression of Mit1/Lb9 was highest in brain. The major transcript, about 7 kb in length, was also expressed in heart, lung, and muscle, while the expression in other tissues was negligible. On the other hand, Copg2 was ubiquitously expressed in most mouse tissues with the highest expression in testis where the Mit1/Lb9 was hardly expressed (Fig. 3B). The major transcript of Copg2 was about 3 kb in length.

3.3 Imprinting status of Mit1/Lb9 and Copg2

The imprinting status of Mit1/Lb9 was determined using a single base polymorphism between mouse strains, C57BL/6J and KJR/Msf. Since the polymorphism did not provide any available RFLP site, mismatch PCR-mediated site-directed mutagenesis [18] was used for the discrimination of the expressed alleles (see Section 2 and Fig. 4A). The single base polymorphism and the mismatch PCR primer introduced a TaqI restriction enzyme site in the RT-PCR product of the KJR/Msf allele, but not in that of the C57BL/6J allele. The KJR/Msf allele was exclusively expressed in brain of the (C57BL/6J×KJR/Msf)F1 hybrid while the C57BL/6J allele was expressed in the reciprocal (KJR/Msf×C57BL/6J)F1 hybrid (Fig. 4B). However, relaxed imprinting patterns were observed in heart, lung and muscle. The expression level of the C57BL/6J allele was higher than that of the KJR/Msf allele in these tissues of the (KJR/Msf×C57BL/6J)F1 hybrid. In the reciprocal (C57BL/6J×KJR/Msf)F1 hybrid, both alleles were expressed at comparable levels in the same tissues (Fig. 4B). The imprinting status of Mit1/Lb9 in mouse embryos and neonates was investigated to determine the developmental stage at which the imprinting of Mit1/Lb9 is established (Fig. 4C). Mit1/Lb9 was apparently imprinted in E13.5 embryos, E17.5 embryos, and in neonates. However, E11.5 embryos exhibited a relaxed imprinting pattern. The result suggests that the imprinted expression pattern of Mit1/Lb9 is established at around E13.5.

The four-nucleotide length polymorphism between C57BL/6J and KJR/Msf was used to determine the imprinting status of Copg2 (Fig. 5A). We initially obtained inconsistent imprinting patterns with different primer sets. Such an inconsistency raised the possibility for the existence of the antisense transcript of Copg2. We prepared strand-specific RT products as PCR templates to distinguish Copg2 from its antisense transcript, Copg2AS (Fig. 5A). In contrast to human COPG2 reported to be maternally imprinted in several human fetal tissues [9], mouse Copg2 is paternally imprinted. Majority of the Copg2 transcripts was expressed from the maternal allele in brain (Fig. 5B). Similar imprinting patterns of Copg2 were also observed in mouse embryos and neonates (Fig. 5B). Copg2AS was maternally imprinted in tested adult tissues, embryos, and neonates (Fig. 5C). Copg2AS spans 3′-UTR of Peg1/Mest and 3′-UTR of Copg2. It is unclear that Copg2AS is a part of the heterogeneous Mit1/Lb9 transcripts or an uncharacterized Peg1/Mest transcript.

Figure 5

Determination of imprinting status of Copg2 and Copg2AS. A: Nucleotide sequences of the region containing a length polymorphism between C57BL/6J and KJR/Msf. Annealing sites of primers G2-F1, G2-PF, G2-PR, and Peg1-3UF are indicated. Strand-specific reverse transcription was performed using primer G2-PF for Copg2 or G2-PR for Copg2AS. The RT-PCR product of Copg2 transcript was amplified with G2-F1 and G2-PF and digested with HhaI. The reverse-transcribed product of Copg2AS transcript was amplified with Peg1-3UF and G2-PR and digested with BsmAI. Digested DNA fragments were separated on a 20% polyacrylamide gel. Expected restriction DNA fragments are represented. Transcripts of Copg2, Copg2AS, and Peg1/Mest are indicated by arrows. B: Imprinting status of Copg2 in adult tissues and in mouse embryos and neonates. C: Imprinting status of Copg2AS in adult tissues and in mouse embryos and neonates. Imprinting status was determined in the (C57BL/6J×KJR/Msf)F1 hybrid (C×M) and the (KJR/Msf×C57BL/6J)F1 hybrid (M×C). C57BL, C57BL/6J allele; Mol, KJR/Msf allele; C, C57BL/6J; M, KJR/Msf; G, genomic DNA of the (C57BL/6J×KJR/Msf)F1 hybrid; E17.5H, head of E17.5; E17.5B, body of E17.5; NeoH, head of neonate; NeoB, body of neonate.

4 Discussion

We have demonstrated that Copg2 is parternally imprinted in mouse brain and the whole embryo in a relaxed manner. The relaxed imprinting pattern of Copg2 shown in this study can be ascribed to the partial imprinting in whole tissues or the tissue-specific imprinting in the whole embryo since we used total RNA of the whole embryos to determine the imprinting status of Copg2. The imprinting strengths of Mit1/Lb9 in heart, lung, and muscle are different between (C57BL/6J×KJR/Msf)F1 hybrid and (KJR/Msf×C57BL/6J)F1 hybrid. The inconsistency can be explained by enhanced lability of imprinted expression pattern under some epigenetic environments including strain-specific genetic background as demonstrated in the imprinting pattern of mouse K_vlqt1 [19].

It is striking that mouse Copg2 is paternally imprinted in contrast to human orthologue COPG2 which was reported to be maternally imprinted in several human fetal tissues [9]. Alternatively, putative antisense transcripts of COPG2 might have obscured the imprinting status of human COPG2 since Blagitko et al. used random hexamer primers for reverse transcription to determine the allelic expression of COPG2 [9].

The allele-specific expression is correlated with dynamic change in the allele-specific methylation at CpG island(s) [20]. The CpG island of Copg2 is a good candidate target for the allele-specific differential methylation. The methylation status of the CpG island is currently being investigated.

Maternal uniparental disomy (UPD) of proximal region of mouse chromosome 6 is associated with embryonic lethality [20]. The overexpression of Copg2 or the absence of Mit1/Lb9 or Copg2AS transcript in the maternal UPD embryo might have caused the embryonic lethality. The sub-proximal end of mouse chromosome 6 shows a conserved synteny with human chromosomal region 7q31–35 [3, 10]. The maternal UPD of human chromosome 7 has been reported in approximately 10% of Silver-Russell syndrome (SRS) cases [21, 22], suggesting that imprinted gene(s) on human chromosome 7 is responsible for SRS phenotype in the UPD patients [23]. The imprinted COPG2 and/or its antisense transcripts including human orthologue of mouse Mit1/Lb9 can be candidate genes for the SRS.

Acknowledgements

This work was supported by a grant from Korea Science and Engineering Foundation (grant no. 1999-1-210-001-3).