The Mouse Murr1 Gene Is Imprinted in the Adult Brain, Presumably Due to Transcriptional Interference by the Antisense-Oriented U2af1-rs1 Gene

Mice.

C57BL/6 (B6), PWK mouse strains, and the F₁ progeny obtained by reciprocal crosses between them were used as the sources of total RNAs and genomic DNAs. BPF1 progeny resulted from the cross of B6 females and PWK males. PBF1 progeny resulted from crosses of PWK females and B6 males.

Mice with two paternal copies of proximal chromosome 11, including the Murr1 locus (paternal duplication [PatDp] prox11), were produced by a standard protocol using the mouse reciprocal translocation T(7;11)40Ad and identified by using genetic markers (4).

Allelic expression of the Murr1 and U2af1-rs1 genes.

Total RNAs were isolated with ISOGEN (Nippon Gene) from various tissues of adult F₁ progeny obtained by reciprocal crosses between C57BL/6 and PWK mice. Contaminating chromosomal DNA was removed by DNase I treatment. Reverse transcription was conducted with oligo(dT) primers according to the manufacturer's protocol of the RNA PCR kit (Takara). To exclude the problem of heteroduplex formation skewing the results of restriction endonuclease digestion, hot-stop reverse transcription-PCR (RT-PCR) was employed (40). Thirty cycles of PCR with primers MuF1 (5′-TGGAGGGTGGCAAGTCCCTG-3′) and UMUR6 (5′-GGTAACACCAGTGGGCAAAG-3′) (Fig. 1) were followed by one cycle of amplification with primer UMUR6 labeled with α-³²P. The resulting 686-bp fragment was digested with BspHI and electrophoresed. The intensity of electrophoretic bands was measured by using BAS2000 (Fujifilm). For the analysis of U2af1-rs1 gene, a regular PCR with primers 5′-U2af1rs1 (5′-AGATAACCACGGATACCTGG-3′) and 3′-U2af1rs1 (5′-AGTACATAGGCCTGCCCATG-3′) was performed. The resulting 236-bp PCR product was digested with MspI and electrophoresed.

Quantification of mRNAs by real-time RT-PCR.

A real-time kinetic PCR method using the SYBR Green and Light-Cycler technique (Roche) was employed to quantify the amount of the Murr1 and U2af1-rs1 mRNAs in the brains and livers of different developmental stages (13.5-days-postcoitum [dpc] embryonic, newborn, 2-week-old, and 12-week-old mice). Total RNAs were isolated from the tissues of BPF1 and PBF1 mice by using ISOGEN (Nippon Gene). The RNAs were reverse transcribed by using oligo(dT) primer and then were used as templates in each real-time kinetic PCR containing a primer pair for Murr1, U2af1-rs1, or Gapdh (glyceraldehyde-3-phosphate dehydrogenase) mRNA. The primer pairs used were as follows. For Murr1, MU109 (5′-ACTCTATCCGGAAGTGCCAC-3′) and UMUR5 (5′-AGGCTTGCCATCGACTCTCC-3′) were used; for U2af1-rs1, U2F1 (5′-AGACCTACCAGCAGTTCTTG-3′) and U2R (5′-GGTCACTGGACAGAATTCAC-3′) were used; and for Gapdh, GAPDHR1 (5′-CAGTCTTCTGGGTGGCAGTG-3′) and GAPDHF (5′-GCCAAACGGGTCATCATCTC-3′) were used.

The kinetics of PCR amplification was monitored as the kinetics of enhancement of the fluorescence of SYBR Green (a dye specific for double-stranded DNA) in the reactions. The relative amounts of three mRNAs were obtained by determining the numbers of cycles required to obtain the same levels of fluorescence in each PCR. In order to evaluate the quantification of this assay and to condition the PCRs, pilot experiments were performed by using serially diluted cDNAs of these mRNAs as the templates according to the manufacturer's protocol.

In situ hybridization of Murr1 mRNAs in mouse brain.

The fresh cuts of brains were fixed immediately in paraformaldehyde after removal and sectioned after being embedded in paraffin. The product of RT-PCR with primers Murr1F (5′-TAGCGCAGAACGCCTTTCAC-3′) and Mu3′ (5′-ATCCTTGAAGACTTTCATGC-3′) (Fig. 1) was cloned into the pT7Blue T-vector (Takara) and used as the template for in vitro transcription of antisense and sense single-stranded RNA probes for Murr1. The antisense and sense RNA probes for U2af1-rs1 were prepared in the same manner from the RT-PCR product with primers U2F (5′-GGAAACACGACTTTCCCACG-3′) and U2AluI-R (5′-GCTGTCCTCTTCTCTTCCTC-3′). The digoxigenin-labeled antisense and sense probes were synthesized with T7 RNA polymerase and digoxigenin-UTP according to the manufacturer's protocol (RNA labeling and detection kit; Roche). The transcripts were used as probes at a concentration of 10 ng/μl in hybridization buffer. The hybridized probe was detected with alkaline phosphatase-coupled antibodies to digoxigenin according to the manufacturer's protocol (RNA labeling and detection kit; Roche).

Methylation analyses.

The DNA methylation status of the CpG islands (CGIs) was assessed by sodium bisulfite PCR and sequencing assay as described previously (46). Briefly, genomic DNA from the brain of a C57BL/6 mouse was extracted by standard phenol-chloroform extraction methods and treated with sodium bisulfite prior to PCR with each primer set. PCR products were cloned into pT7Blue T-vector, and each cloned PCR product was amplified by colony PCR using a Takara Ex Taq Kit (Takara Shuzo). The amplified products were subjected to sequencing reaction using BigDye (Applied Biosystems) and DYEnamic ET terminator kits (Amersham Pharmacia Biotech) after treatment with exonuclease I and shrimp alkaline phosphatase (both from Amersham Pharmacia Biotech).

Detection and quantification of the primary transcript of the Murr1 and U2af1-rs1 by strand-specific RT-PCR and semiquantitative RT-PCR.

DNase I-treated total RNAs (0.25 μg) isolated from the brain, liver, heart, or lung of adult mice using ISOGEN (Nippon Gene) were mixed with 5 pmol of either forward primers 1, 3, 4, and 5 or reverse primer 2 (see Fig. 6) in a final volume of 10 μl; the RNAs were then reverse transcribed for 30 min at 55°C, heat inactivated at 99°C for 5 min, and cooled on ice. Following the reverse transcription and denaturation, 33 PCR cycles with four different primer sets corresponding to the primers for reverse transcription were performed, respectively.

Semiquantitative RT-PCR was employed to detect the downstream primary transcript of the Murr1 and U2af1-rs1. The product of reverse transcription with either primer 1 or 3 as described above was subjected to 33 and 36 PCR cycles, respectively, the products of which were verified as linear ranges of the PCR with corresponding primers. After agarose electrophoresis and staining with ethidium bromide, the quantity of the PCR products was analyzed by National Institutes of Health Image software (version 1.62).

Murr1 contains U2af1-rs1 in its first intron.

Murr1 spans 83 kb and consists of three exons and two introns. An imprinted gene, U2af1-rs1, which shows paternal allele-specific expression, is located within the first intron of Murr1 and transcribed in a direction opposite to that of the host gene (Fig. 1) (27). The genomic sequence, including that of the Murr1 gene, is available from DDBJ under the accession numbers AB089806 and NT_039515. This gene encodes a mRNA of approximately 1 kb and is expressed ubiquitously in adult tissues as shown by Northern blotting in Fig. 2. Cloning and sequencing of the cDNA revealed that Murr1 mRNA is 793 bases long, excluding the polyA tail, and encodes a protein of 188 amino acids. The transcriptional initiation site was determined by 5′ rapid amplification of cDNA ends analysis. Several of the largest clones had the same 5′ ends (data not shown). The cDNA sequence has been submitted to the DDBJ database under accession numbers D85430 (27) and AB104816. Although the function of the protein has not yet been analyzed in mice, the homologue in the dog was reported to be involved in copper metabolism (41, 42).

The maternal allele is expressed preferentially in all tissues but predominantly in adult brain.

The existence of an imprinted gene within the Murr1 gene has led us to analyze the allelic expression of Murr1 because many imprinted genes are found to be organized as clusters (3, 10). The gene has been reported to show biallelic expression in the whole body of neonatal mice specimens (27). Tissues of adult mice were examined by using the F₁ progeny of reciprocal crosses between C57BL/6 and PWK mice. We performed hot-stop RT-PCR followed by digestion with BspHI that could distinguish between the Murr1 alleles of each parental strain (Fig. 1). As seen in Fig. 3A, Murr1 showed preferential expression of the maternal allele in all the adult tissues examined. In the brain, predominant expression of the maternal allele was the most significant. In order to know whether this allelic expression also occurs in early developmental stages, we analyzed embryonic and neonatal brains (Fig. 3B). These brains express the Murr1 gene biallelically with moderate preference for the maternal allele, as seen in adult tissues other than the brain. The brain of a 2-week-old mouse showed predominant expression of the maternal allele to the same extent as that found in the brains of 6- and 12-week-old mice.

Because there is a possibility that the paternal expression of U2af1-rs1 affects the allelic expression of its host gene, Murr1, its allelic expression was also examined in the same tissues. In previous reports, U2af1-rs1 was shown to be expressed exclusively from the paternal allele in the early embryo and whole neonate (16, 17). Our analysis showed that the U2af1-rs1 gene was expressed exclusively from the paternal allele in all the adult tissues and in the brains of mice in all the developmental stages so far examined (Fig. 3C and D).

Predominant maternal expression was observed in all regions of adult brain.

Maternal allele-predominant expression of the Murr1 gene in the adult brain may suggest a function of the gene in the development of the brain after birth. In order to gain insight into this function, we determined the regions expressing the gene by RNA in situ hybridization on the adult brains of C57BL/6 mice (Fig. 4a, b, c, f, i, and l). The expression of Murr1 was observed in all the regions of the cerebrum, cerebellum, and medulla oblongata. The hippocampus region of the cerebrum had the highest levels of expression. The expression was seen mainly in neuronal cells.

Maternal allele-predominant expression of Murr1 raises two possibilities for the allelic expression of this gene in adult brain. First, Murr1 may be expressed predominantly from the maternal allele in all the regions of adult brain. Alternatively, the gene may be expressed exclusively from the maternal allele in some regions but biallelically in other regions. Ube3a is an imprinted gene with such a regional-imprinted expression (1). In situ hybridization was performed using adult mouse brain cells with two paternal copies of proximal chromosome 11, including the Murr1 locus [PatDp(prox11)]. Weak but detectable signals for Murr1 mRNA were observed in all the regions and were essentially similar to those of wild-type C57BL/6 brains (Fig. 4d, g, j, and m). There was no region lacking hybridization signal, and the expression pattern of Murr1 appeared to be normal in the PatDp(prox11) brain. This result indicated that the gene was expressed maternal allele predominantly in all the regions expressing the gene.

The CGIs associated with the Murr1 promoter region are not methylated.

Three CGIs have been identified in the region spanning from 10 kb upstream to 10 kb downstream of the Murr1 gene (Fig. 1). CGI1 is located approximately 8 kb upstream from the Murr1 gene, and CGI2 spans from the promoter region to the 5′ part of exon 1. CGI3 is the region spanning from the promoter to the body of the U2af1-rs1 gene. CGI3 has been reported to be methylated only on the maternal allele and has been thought to be involved in the paternal expression of U2af1-rs1 (17). These three CGIs were analyzed for methylation in the brain genome by a sodium bisulfite sequencing method (data not shown). One can expect that the CGIs 1 and/or 2 may be paternally methylated if differential methylation is involved in the maternal expression of Murr1. The CGI3 was found to be differentially methylated in the brain as reported previously. On the other hand, no methylation was found in the other two CGIs, 1 and 2. Nonmethylation of these CGIs suggests that differential methylation within the genome is not involved in the allelic expression of Murr1, in contrast to many of the imprinted genes.

The expression levels of Murr1 and U2af1-rs1 change inversely during brain development.

In the adult brain, U2af1-rs1 and Murr1 show reciprocal allelic expression. The U2af1-rs1 gene is expressed exclusively from the paternal allele, and the Murr1 gene is expressed predominantly from the maternal allele. This fact suggests that there may be some kinds of interaction, such as transcriptional interference, between allelic expressions of these genes. In order to investigate this possibility, the mRNA levels of these genes were quantified in the brain and liver during development by quantitative real-time RT-PCR (Fig. 5).

In the brain (Fig. 5A), Murr1 expression occurs at a relatively high level during the embryonic and neonatal periods. The expression levels decrease considerably between the neonatal stage and 2 weeks after birth, the period during which maternal allele-predominant expression of Murr1 becomes remarkable (Fig. 3B). This overall decrease from the embryo to adult stages is approximately fourfold. In contrast, the expression of U2af1-rs1 continues to increase during development (Fig. 5B). Overall, this increase is approximately two- to threefold. As a result, the ratio of expression of U2af1-rs1 versus that of Murr1 increased about 10-fold during development (Fig. 5C). The final level of U2af1-rs1 expression (0.05) is almost the same as that of Murr1 expression (0.06) in the adult brain.

On the other hand, the expression of both genes decreased simultaneously during development in the liver, in which the maternal allele-predominant expression of Murr1 does not occur (Fig. 5D and E). Relative levels of expression of U2af1-rs1 and Murr1 do not change significantly during development (Fig. 5F). The expression of Murr1 is more than 10-fold higher than that of U2af1-rs1 in both embryonic and adult periods.

Transcription from U2af1-rs1 may act through the promoter region of Murr1.

Antisense orientation of the U2af1-rs1 gene relative to Murr1 suggests the possibility that transcription originating from U2af1-rs1 promoter overlaps the promoter region of the Murr1 gene and interferes with Murr1 transcription. The 3′ end of a mature mRNA is created by the cleavage of the primary transcript. A primary transcript is synthesized beyond the end of the gene, sometimes far more than a distance of 1 kb (29). If this is because of the paternal expression of U2af1-rs1, this interference should occur only on the paternal allele and may cause the reduction of the paternal expression of Murr1. We employed strand-specific and semiquantitative RT-PCR to detect and quantify the transcripts in the region downstream of U2af1-rs1 up to the promoter region of Murr1 (Fig. 6).

In the analysis performed using adult brain RNA, the transcript was detected in the promoter region when primer 1 with sense orientation was used for cDNA synthesis. In contrast, no transcript was detected in this region when primer 2 with antisense orientation was used as the primer for cDNA synthesis (Fig. 6B). This result indicated that the Murr1 promoter region was transcribed in the antisense direction relative to the Murr1 gene. The sequencing of the PCR product revealed that the transcript was expressed exclusively from the paternal allele (Fig. 6D). On the other hand, when liver RNA was used for this analysis, no transcript was detected with either of the primers for cDNA synthesis using PCR with same number of amplification cycles (33 cycles) (Fig. 6B). An antisense transcript was, however, detected in the liver when PCR was done with 36 amplification cycles (data not shown). In the brain, transcripts with antisense orientation relative to Murr1 were also detected at two sites in the first intron downstream of U2af1-rs1. Only the transcript at the site close to U2af1-rs1 was detected in the liver, however. No transcript with antisense orientation was detected at the promoter region of U2af1-rs1 in either the brain or liver (Fig. 6C). These results strongly suggest that transcription from the U2af1-rs1 promoter to the upstream region of the Murr1 gene on the paternal chromosome occurs more abundantly in the brain than in the liver, although we have no direct evidence that these antisense transcripts are parts of one continuous transcript. This finding implies that the transcription of U2af1-rs1 interferes with the expression of the paternal allele of Murr1 in the adult brain. If this is the case, U2af1-rs1 should be expressed in the regions where Murr1 is expressed in the brain. This has been shown to be true by in situ hybridization, in which the expression of U2af1-rs1 observed in the regions was essentially the same as that of Murr1 expression in C57BL/6 brain (Fig. 4e, h, k, and n). We also detected the same antisense transcripts at the Murr1 promoter in lung and kidney cells. Semiquantitative RT-PCR indicated that the primary antisense transcript in the Murr1 promoter region was much more abundant in the adult brain than in the heart, lung, and liver (data not shown).