Variable allelic expression of imprinted genes at the Peg13, Trappc9, Ago2 cluster in single neural cells

Abstract

Genomic imprinting is an epigenetic process through which genes are expressed in a parent-of-origin specific manner resulting in mono-allelic or strongly biased expression of one allele. For some genes, imprinted expression may be tissue-specific and reliant on CTCF-influenced enhancer-promoter interactions. The Peg13 imprinting cluster is associated with neurodevelopmental disorders and comprises canonical imprinted genes, which are conserved between mouse and human, as well as brain-specific imprinted genes in mouse. The latter consist of Trappc9, Chrac1 and Ago2, which have a maternal allelic expression bias of ∼75% in brain. Findings of such allelic expression biases on the tissue level raise the question of how they are reflected in individual cells and whether there is variability and mosaicism in allelic expression between individual cells of the tissue. Here we show that Trappc9 and Ago2 are not imprinted in hippocampus-derived neural stem cells (neurospheres), while Peg13 retains its strong bias of paternal allele expression. Upon analysis of single neural stem cells and in vitro differentiated neurons, we find not uniform, but variable states of allelic expression, especially for Trappc9 and Ago2. These ranged from mono-allelic paternal to equal bi-allelic to mono-allelic maternal, including biased bi-allelic transcriptional states. Even Peg13 expression deviated from its expected paternal allele bias in a small number of cells. Although the cell populations consisted of a mosaic of cells with different allelic expression states, as a whole they reflected bulk tissue data. Furthermore, in an attempt to identify potential brain-specific regulatory elements across the Trappc9 locus, we demonstrate tissue-specific and general silencer activities, which might contribute to the regulation of its imprinted expression bias.

Keywords: Ago2; Peg13; Trappc9; allelic expression; genomic imprinting; neural stem cell; neurosphere; single-cell analysis.

FIGURE 1

Analysis of alternative 5′-exons and transcriptional start sites of Trappc9. (A) Scheme of transcript variants 202, 201, 206, 203 and 205 as indicated on ENSEMBL. The alternative exons of variant 201, a1 and 5, are indicated by arrows. (B) Scheme of exons and primers used in RT-PCR (right) on total C57BL/6J RNA from multiple tissues using primer combinations specific for transcript 202 exon 2 or transcript 201 exon a1 with shared downstream primers in exons 5 and 6 (indicated in the scheme with FW and RV annotations). While expression of transcript 202 exons 2–5 and 2–6 (expected: 740 bp and 780 bp, respectively) was confirmed in multiple tissues, no expression involving transcript 201 exon a1 was detectable (expected: 318 bp and 358 bp, respectively). Size markers of 500 and 1,000 bp are indicated. (C) Sequencing of 5′-RACE RCR products using a downstream primer in exon 6 confirmed multiple transcription start sites in 202 exon 1 (indicated by red arrows), while no start sites in the region of 201 exon a1 could be detected. The first codon is underlined in exon 2.

FIGURE 2

Allelic expression in tissues and primary hippocampal neurospheres of hybrid mice. Parental allelic expression was quantified via SNP pyrosequencing of cDNA from tissues or cultured neurospheres (NS) obtained from newborn mice. Average expression (n = 2–3) and example traces are shown. SNP IDs and exon locations are: Peg13 (rs238259968 and rs31423566 located in exon 1), Trappc9 (rs31440851 located in exon 2), Ago2 (rs232384843 located in exon 5) and Kcnk9 (rs225149059 located in exon 1). Points on the bar graphs represent individual pyrosequencing results from reciprocal crosses as shown in the legend. For Peg13 each data point shows the average of the two SNP values in the same sequence read. Representative pyrograms from whole brain samples are shown, indicating the SNP position in the sequence and the quantification of allelic expression.

FIGURE 3

Variable single-cell allele-specific expression of Peg13 analyzed through Sanger sequencing of cDNA SNPs from neural stem (neurosphere) cells and their in vitro differentiated neurons. (A) The mouse cross, cDNA amplicon and SNP locations in exon 1 are shown. B6 SNP variants are indicated in color; primers are underlined. (B) Summary data showing the proportions of cells falling into the five categories of allelic expression. (C) Example single-cell sequence tracks for the three expression categories indicated. Sequence tracks for cells with maternal or paternal bias displayed a major SNP peak for those alleles with a minor overlapping SNP peak for the other allele, respectively. SNP positions are highlighted by arrows.

FIGURE 4

Variable single-cell allele-specific expression of Trappc9 analyzed through Sanger sequencing of a cDNA SNP from neural stem (neurosphere) cells and their in vitro differentiated neurons. (A) The mouse cross, cDNA amplicon and SNP location in exon 7 are shown. B6 SNP variant is indicated in color; primers are underlined. (B) Summary data showing the proportions of cells falling into the five categories of allelic expression. (C) Example single-cell sequence tracks for the three expression categories indicated. Sequence tracks for cells with maternal or paternal bias displayed a major SNP peak for those alleles with a minor overlapping SNP peak for the other allele, respectively. SNP position is highlighted by an arrow.

FIGURE 5

Variable single-cell allele-specific expression of Ago2 analyzed through Sanger sequencing of a cDNA SNP from neural stem (neurosphere) cells and their in vitro differentiated neurons. (A) The mouse cross, cDNA amplicon and SNP location in exon 5 are shown. B6 SNP variant is indicated in color; primers are underlined. (B) Summary data showing the proportions of cells falling into the five categories of allelic expression. (C) Example single-cell sequence tracks for the three expression categories indicated. Sequence tracks for cells with maternal or paternal bias displayed a major SNP peak for those alleles with a minor overlapping SNP peak for the other allele, respectively. SNP position is highlighted by an arrow.

FIGURE 6

Promoter-reporter gene assays indicate silencer elements for Trappc9. (A) Scheme of the murine Peg13—Kcnk9—Trappc9 cluster of imprinted genes. For Trappc9 all introns are shown; for Chrac1, Ago2 and Kcnk9 introns have been omitted. Promoters, transcriptional activity and directions are indicated by arrows. Also indicated are the approximate locations of the candidate regulatory elements (listed in Supplementary Figure S6) across Trappc9 introns. Not to scale. (B) Schematic of the promoter/enhancer constructs used for transfection of primary cells. Enhancers were positioned downstream or upstream of the promoter-reporter gene cassette depending on their relative locations within the Trappc9 gene and in the same orientation. (C) Reporter gene activity of the constructs in cultures of primary hippocampal neurons from newborn mice and primary embryonic fibroblasts. Normalized activity of Firefly luciferase (FLuc) to co-transfected Renilla luciferase (RLuc) is shown. Mean values ± S.E.M. are indicated. *p ≤ 0.05; **p ≤ 0.01.