Activation of paternally expressed genes and perinatal death caused by deletion of the Gtl2 gene

Abstract

The Dlk1-Gtl2 imprinting locus is located on mouse distal chromosome 12 and consists of multiple maternally expressed non-coding RNAs and several paternally expressed protein-coding genes. The imprinting of this locus plays a crucial role in embryonic development and postnatal growth. At least one cis-element, the intergenic differentially methylated region (IG-DMR) is required for expression of maternally expressed genes and repression of silenced paternally expressed genes. The mechanism by which the IG-DMR functions is largely unknown. However, it has been suggested that the unmethylated IG-DMR acts as a positive regulator activating expression of non-coding RNAs. Gtl2 is the first non-coding RNA gene downstream of the IG-DMR. Although its in vivo function in the mouse is largely unknown, its human ortholog MEG3 has been linked to tumor suppression in human tumor-derived cell lines. We generated a knockout mouse model, in which the first five exons and adjacent promoter region of the Gtl2 gene were deleted. Maternal deletion of Gtl2 resulted in perinatal death and skeletal muscle defects, indicating that Gtl2 plays an important role in embryonic development. The maternal deletion also completely abolished expression of downstream maternally expressed genes, activated expression of silenced paternally expressed genes and resulted in methylation of the IG-DMR. By contrast, the paternal inherited deletion did not have this effect. These data strongly indicate that activation of Gtl2 and its downstream maternal genes play an essential role in regulating Dlk1-Gtl2 imprinting, possibly by maintaining active status of the IG-DMR.

Fig. 1.

Targeted deletion of the Gtl2 gene. (A) Schematic representation of the Dlk1-Gtl2 locus on mouse chromosome 12. Maternally expressed genes are in red and paternally expressed genes are in blue. Differentially methylated regions are shown as circles. Filled circle, methylated; open circle, unmethylated. (B) The Gtl2 gene consists of ten exons (only six exons are shown). The gene targeting vector contains an 8.0 kb genomic DNA upstream exon 1 (LA) and a 2.0 kb DNA after exon 5 (SA) flanking the neo cassette. After homologous recombination, the targeted region consisting of exons 1-5 of the Gtl2 gene was deleted. P_l and P_s are the probes for detection of long and short arms by Southern blotting, respectively. N1 and A1 are PCR primers for genotyping. (C) Deletion of the targeted region was confirmed by Southern blotting. Genomic DNA was digested with BamHI, and probed with P_l and P_s as indicated.

Fig. 2.

Gtl2 KO/+ embryos display delayed skeletal muscle fiber development. (A-F) Hematoxylin and Eosin (H&E) staining of skeletal muscle from (A) 18.5 dpc wild-type and (B) 18.5 dpc Gtl2 +/KO embryos shows peripherally located nuclei and no central clearing within myofibers. H&E staining of skeletal muscle from (C) 18.5 dpc Gtl2 KO/+ and (D) 18.5 dpc Gtl2 KO/KO embryos display centrally located nuclei and central clearing within myofibers. Similar findings are seen in longitudinal sections (D, inset). H&E staining of skeletal muscle from (E) 15.5 dpc Gtl2^+/+ and (F) 15.5 dpc KO/+ embryos show centrally located nuclei and central clearing within myotubes. Arrows denote central nucleation and central clearing within myofibers. (G) Total numbers of myofibers in multiple fields as well as numbers of myofibers with centrally located nuclei or central clearing in the same fields were counted. Data were obtained from at least three embryos for each genotype from three different litters. Percentages of myofibers with central nucleation or central clearing were calculated in intercostal muscles (left) and diaphragm (right) from embryos at ages 15.5, 16.5 and 18.5 dpc, respectively. Data were represented as mean±s.d. Student's t-test was used to compare values between Gtl2 KO embryos and their litter-matched Gtl2^+/+ embryos. ^*, P<0.05; ^†, P<0.01. Scale bars: 25 μm.

Fig. 3.

Skeletal muscles from Gtl2 KO/+ embryos have central accumulation of glycogen within myofibers. (A) Electron micrographs of limb muscle from 18.5 dpc wild-type embryos show peripherally located nuclei. (B,C) Electron micrographs of limb muscle from 18.5 dpc Gtl2 KO/+ embryos reveal internal nuclei and central washed out areas, and slightly granular material within myofibers. (D) PAS staining of limb muscles from 18.5 dpc Gtl2 KO/+ embryos highlights areas between central nuclei (arrows) that had been clear on H&E staining (Fig. 2D, inset).

Fig. 4.

Expression of imprinted genes in Gtl2 KO embryos. (A) Northern blot analysis of Gtl2 expression. Total RNA was isolated from whole 16.5 dpc embryos with paternal (+/KO) or maternal (KO/+) deletion of the Gtl2 gene and their wild-type littermates (+/+). Each track represents RNA from one embryo. Blots were hybridized with Gtl2 probe P_ex3 or P_ex10. Gapdh was used as a control for equal loading. (B) Northern blot analysis of maternally and paternally expressed genes. (C) Comparison of the imprinted gene expression between embryos carrying maternal (black bar), or paternal (gray bar) Gtl2 deletion and their wild-type littermates (white bar). Values were calculated using control and mutant embryos from multiple litters. (D) Expression of maternally imprinted microRNAs by qRT-PCR. Values from wild-type embryos were designated as 100%, against which values from Gtl2 KO/+ and Gtl2 +/KO embryos were normalized. Data were obtained using at least three embryos for each genotype from two different litters and were represented as mean±s.d.

Fig. 5.

Biallelic expression of Dlk1 in Gtl2 KO/+ embryos. Total RNA was isolated from embryos obtained by crossing DBA/2 with Gtl2 +/KO C57BL/6 mice. The region containing the polymorphism was amplified by RT-PCR. The allelic expression of Dlk1 was determined by sequencing PCR products.

Fig. 6.

Expression of imprinted genes in Gtl2 KO placentas. (A) Northern blot analysis of maternally and paternally expressed genes in 18.5 dpc placentas. Each track represents RNA from one embryo. Gapdh was used as the internal control. (B) Comparison of the imprinted gene expression between 18.5 dpc placentas carrying maternal (black bar), or paternal (gray bar) Gtl2 deletions and their wild-type littermates (white bar). Values from wild-type embryos were designated as 100%, against which values from Gtl2 KO/+ and Gtl2 +/KO placentas were normalized. Data were obtained using at least five placentas for each genotype from three different litters and were represented as mean±s.d. The one-sample t-test was used to compare values between Gtl2 KO embryos and their litter-matched Gtl2^+/+ embryos. ^*, P<0.02; ^†, P<0.01.

Fig. 7.

Methylation analysis of IG-DMR, Gtl2-DMR and Dlk1-DMR. (A-C) Genomic DNA from 16.5 dpc embryos carrying paternal Gtl2 deletion (+/KO) and maternal Gtl2 deletion (KO/+) as well as their wild-type littermates (+/+) were digested with (A) StuI, (B) BsrI or (C) NheI. The fragmented DNAs were further digested with methylation-insensitive MspI, or methylation-sensitive HpaII or HhaI. The digested DNAs were analyzed by Southern hybridization with probes specific against (A) IG-DMR, (B) Gtl2-DMR and (C) Dlk1-DMR, respectively. Restriction enzymes: M, MspI; Hp, HpaII; Hh, HhaI. (D) IG-DMR bisulfite sequencing analysis. Each horizontal bar represents a sequence from each clone. Circles indicate CpG sites. White and black circles designate unmethylated and methylated CpGs, respectively.

Fig. 8.

Methylation analysis of IG-DMR and neo expression in female germ cells. (A) Bisulfite sequencing analysis of IG-DMR in oocytes. The region analyzed is the same as the IG-DMR in Fig. 7A. Each horizontal bar represents a sequence from each clone. Circles indicate CpG sites. White and black circles designate unmethylated and methylated CpGs, respectively. (B) The Neo expression was detected by RT-PCR in embryos and oocytes as indicated. β-actin was used as the internal control.