Long-range DNase I hypersensitivity mapping reveals the imprinted Igf2r and Air promoters share cis-regulatory elements

Abstract

Epigenetic mechanisms restrict the expression of imprinted genes to one parental allele in diploid cells. At the Igf2r/Air imprinted cluster on mouse chromosome 17, paternal-specific expression of the Air noncoding RNA has been shown to silence three genes in cis: Igf2r, Slc22a2, and Slc22a3. By an unbiased mapping of DNase I hypersensitive sites (DHS) in a 192-kb region flanking Igf2r and Air, we identified 21 DHS, of which nine mapped to evolutionarily conserved sequences. Based on the hypothesis that silencing effects of Air would be directed towards cis regulatory elements used to activate genes, DHS are potential key players in the control of imprinted expression. However, in this 192-kb region only the two DHS mapping to the Igf2r and Air promoters show parental specificity. The remaining 19 DHS were present on both parental alleles and, thus, have the potential to activate Igf2r on the maternal allele and Air on the paternal allele. The possibility that the Igf2r and Air promoters share the same cis-acting regulatory elements, albeit on opposite parental chromosomes, was supported by the similar expression profiles of Igf2r and Air in vivo. These results refine our understanding of the onset of imprinted silencing at this cluster and indicate the Air noncoding RNA may specifically target silencing to the Igf2r promoter.

Figure 1.

Overview of investigated region in the mouse chromosome 17 bA cluster. (A) The chromosome map spans 513 kb of mouse chromosome 17 band A and contains seven known transcripts (http://www.ensembl.org/Mus_musculus) (Lyle et al. 2000). The expression pattern in mouse embryo fibroblasts (MEFs) is shown for the maternal and paternal chromosome. Note that the 108-kb Air ncRNA overlaps the 5′ part of the Igf2r transcript by 30 kb and the 3′ part of the Mas1 transcript by 1 kb. Black boxes indicate expressed genes, striped boxes indicate imprinted silencing that affects one parental allele, white boxes indicate tissue-specific silencing that affects both alleles, arrows indicate transcriptional orientation. The Plg and Mas1 genes mark the known limits of this imprinted cluster. Note that Slc22a2 and Slc22a3 show imprinted expression in the embryonic placenta but are not expressed in MEFs. Parental-specific DNA methylation is present on the Igf2r and Air promoters. (Black circle) DNA-methylated CpG island promoter; (white circle) unmethylated CpG island promoter. Note that the Slc22a2 promoter is a weak CpG island (white diamond). (B) Modified ENSEMBL map (http://www.ensembl.org/Mus_musculus), including custom tracks (available on request) of the region including the Slc22a1, Igf2r, Air, and Mas1 genes. (Contigs row) The ENSEMBL contig view is shown, where the continuous sequence coverage stops in the 5′ region of Mas1. Note that two restriction fragments end in this region and their sizes were predicted from AJ249895. (ENSEMBL genes row) The ENSEMBL genes of this area are indicated; note that in this build Slc22a1 is annotated as “Ensembl novel trans” and the Air ncRNA is not annotated. (Restriction fragments row) Three significant differences between the public mouse sequence (C57Bl6 strain) and the sequence of the mouse strains used here (129Sv and FVB) were detected in three restriction fragments (black boxes): A—A polymorphic BclI fragment that contains a 6-kb LINE insertion only in the 129Sv mouse strain (Supplemental Fig.1); B—a 29.5-kb and C—a 6.8-kb insertion, shown in an EcoRI and an EcoRV fragment, that were not detected in any genomic DNA (see Supplemental Fig.1) and are false insertions in the ENSEMBL sequence. The positions of all other tested restriction fragments are shown as light gray boxes. (Gaps row) The gaps in the region coverage; (Probes row) DNA blot probes used for the DHS assays; (ECRs row) location of evolutionarily conserved regions (see also Figs. 2 and 3); (DHS row) the position of the identified DHS. Note that the only gap in the coverage of the region consists of a 100% interspersed repeat sequence of 1.1 kb.

Figure 2.

Figure 2.

DHS map of the Igf2r/Air cluster. All identified DHS were shown at their correct position within the genomic organization of the investigated region. Because of its length, this region is presented in two parts. (A) The region from Slc22a1 to the Igf2r promoter; (B) the region from the Igf2r promoter to Mas1. Note that blots from DHS12 at the Igf2r promoter are shown in A (BglII fragments) and B (EcoRI fragments) to show the overlap between the two maps. The PipMaker plot (Schwartz et al. 2000) above the DNA blots shows the position and transcriptional orientation of genes by numbered exons (black boxes) and horizontal arrows on top of the sequence, together with the gene name. Symbols indicating positions of repeats identified by RepeatMasker (A.F.A. Smit, R. Hubley and P. Green, unpubl.) and CpG islands are shown (see legend). Evolutionary conservation to the human sequence between 50% and 100% is shown below by short horizontal bars in the PipMaker plot. ECRs defined as >70% identity over >50 bp, excluding exons and interspersed repeats are indicated by gray vertical bars within the PipMaker plot. Sequences not investigated for DHS, as well as erroneously assembled sequence insertions (Supplemental Fig.1), are indicated by hatched boxes. The strain 129Sv specific LINE insertion (Supplemental Fig.1) is indicated by a white box above the PipMaker plot in B; numbers in this box refer to sequence AJ249895. Note that the 12-kb BclI fragment shown for DHS17 and DHS18 in B is created by this LINE insertion, as the ENSEMBL m33 sequence predicts a 19-kb BclI fragment (see Fig. 1 and Supplemental Table 1). The mapped positions of identified DHS are indicated by vertical arrows together with the name of the DHS with closely related DHS combined under one name. For each identified DHS, a representative set of DNA blots is shown from MEF cell lines [NIH3T3 wild type (+/+), MEF-B1: paternal T^hp deletion (+/Thp), MEF-F: maternal T^hp deletion (Thp/+)], from CCE embryonic stem cells, and from spleen and brain of an adult FVB mouse. DNase I concentrations are indicated by a black triangle above the blot; the two lanes on the right side of each blot contain nuclei treated only with incubation buffer at 0°C and 37°C. The enzyme used for restriction digests and the hybridization probe are indicated for each blot (location of used restriction fragments and probes are listed in Supplemental Table 1). The DHS were visible on DNA blots as additional bands to the expected restriction fragment in lanes with DNase I treated DNA. The NIH3T3 blot of DHS8 shows an exception where two restriction fragments were visible in all lanes, which is likely due to a cell line-specific polymorphism. Strong DHS bands are indicated by horizontal arrows and weak ones by asterisks on the left of each DNA blot. The sizes of the shown fragments are indicated on the right in kilobase pairs.

Figure 3.

Overview of DHS and evolutionary conservation. The position and regulation of the 21 DHS identified in the 192-kb Mas1-Slc22a1 interval are shown relative to mouse-human evolutionarily conserved regions (ECRs). (A) Numbered vertical arrows indicate the position of all identified DHS relative to the gene map and ECRs; “X” indicates regulated DHS absent in the tested sample; a blank space indicates the sample was not tested. For each cell line (ES, MEF) and organ (spleen, brain), all identified DHS were shown on one line. Open arrows represent single or multiple DHS at regions free of interspersed repeats but >2 kb distant from ECRs (DHS 1-2, 4, 9-10, 17-18, 21); black arrows mark DHS located at an interspersed repeat (DHS 3, 8, 11, 13-15, 19); shaded arrows show repeat-free DHS located within 2 kb of an ECR (DHS5-7, 12, 16, 20). The parental-specific DHS12 at the Igf2r promoter and DHS9 at the Air promoter are marked by an asterisk. Note that DHS11 is most likely a cell line-specific DHS. The lack of DHS between DHS19 and DHS20 is due to incorrect sequence insertions presented in Figure 1. (B) The positions and transcriptional orientation of the genes are indicated by horizontal arrows; solid circles at the start of the arrow indicate CpG island promoters; the Mas1 promoter has not yet been identified. Below the gene map, ECRs, selected from the PipMaker output (defined in Fig. 2), are indicated by filled circles. Gray circles indicate PipMaker ECRs associated with a DHS that were also shown below as conservation peaks in the output of the VISTA program (Mayor et al. 2000). Black circles indicate PipMaker ECRs not shown in the VISTA output. The VISTA plots in the bottom panel display the percent identity of the human and mouse genomic sequence relative to the mouse sequence but only values above 50% are shown. Conserved exons are indicated by black shading and the conservation peaks, corresponding to the PipMaker ECRs, in gray. The location of the Igf2r and Air promoter are indicated with arrows; note that the Air promoter lacks sequence conservation.

Figure 4.

Relative expression levels of Igf2r and Air. Quantitative real-time PCR analysis of Igf2r and Air expression. Total RNA from organs of four different adult male C57Bl6/CBA mice (sample C was obtained by pooling organs of two individuals) and 15.5-dpc embryos was reverse transcribed and analyzed for Air (shaded bars) and Igf2r (open bars) expression by qPCR using Taqman probes. Expression data were normalized to Cyclophilin A mRNA, and Igf2r and Air expression in the heart of sample B was set to 100 (marked by an asterisk). The numbers on the x-axis indicate the relative amounts of Igf2r and Air in different tissues but do not make a comparison between Igf2r and Air (see Supplemental Table 2 for the data used). The standard deviation representing three technical replicates for each sample is indicated. (E15.5 +/+) Wild-type embryo at 15.5 dpc. The E15.5 +/T^hp embryo littermate that has a paternal deletion of the proximal Igf2r/Air imprinted cluster (Barlow et al. 1991) was used as a control for background levels of Air RNA.