The human retinoblastoma gene is imprinted

Abstract

Genomic imprinting is an epigenetic process leading to parent-of-origin-specific DNA methylation and gene expression. To date, approximately 60 imprinted human genes are known. Based on genome-wide methylation analysis of a patient with multiple imprinting defects, we have identified a differentially methylated CpG island in intron 2 of the retinoblastoma (RB1) gene on chromosome 13. The CpG island is part of a 5'-truncated, processed pseudogene derived from the KIAA0649 gene on chromosome 9 and corresponds to two small CpG islands in the open reading frame of the ancestral gene. It is methylated on the maternal chromosome 13 and acts as a weak promoter for an alternative RB1 transcript on the paternal chromosome 13. In four other KIAA0649 pseudogene copies, which are located on chromosome 22, the two CpG islands have deteriorated and the CpG dinucleotides are fully methylated. By analysing allelic RB1 transcript levels in blood cells, as well as in hypermethylated and 5-aza-2'-deoxycytidine-treated lymphoblastoid cells, we have found that differential methylation of the CpG island skews RB1 gene expression in favor of the maternal allele. Thus, RB1 is imprinted in the same direction as CDKN1C, which operates upstream of RB1. The imprinting of two components of the same pathway indicates that there has been strong evolutionary selection for maternal inhibition of cell proliferation.

Figure 1. Identification of a novel putative imprinted locus.

(A) Heatmap of the Infinium HumanMethylation27 BeadChip (Illumina) for the RB1 gene. The CpG sites representing CpG 85 show about 50% methylation in DNA from blood of the parents but are hypomethylated in DNA from blood of the patient. In all samples, CpG 42 is methylated and CpG 106 is unmethylated. Target ID of the CpG sites representing CpG 85: cg19427472, cg13431205, cg03085377, cg18481241; CpG 42: cg19447496, cg19296958; CpG 106: cg24937706, cg10552385, cg17055959. (B) Schematic representation of the 5′-region of the RB1 locus (not drawn to scale) and location of CpG islands (green boxes). Regular exons are shown in blue whereas the new exon 2B is shown in light blue. Open lollipops, unmethylated CpGs; filled lollipops, methylated CpGs; black arrows, transcription start sites. (C) Exon connection PCR. M, DNA length standard; +, with RT; −, without RT; arrowhead indicates the RT–PCR product that was used as template for sequencing; red arrows, location of RT–PCR primers.

Figure 2. Structure of KIAA0649 and processed pseudogenes in human (position numbers according to hg19, UCSC), chimpanzee (panTro2, UCSC), and rhesus (rheMac2, UCSC).

In the human genome, two of the four small CpG islands in exon 4 of KIAA0649 (CpG 19 and CpG 17) correspond to CpG 85 in the chromosome 13 copy. The other two (CpG 26 and CpG 19) correspond to CpG 42. The figure also shows the similarities and differences between the situation in humans, chimpanzee and rhesus. Owing to gaps in the chimpanzee and rhesus genome sequences, the picture may not be complete. Light green boxes, CpG islands <300bp; dark green boxes, CpG islands >300bp; arrows, orientation of transcription.

Figure 3. Analysis of CpG 85 and the 2B-transcript.

(A) Methylation analysis of CpG 85 by DNA cloning and sequencing. A total of 12 CpG sites within the CpG island were analyzed. Clones from a normal control (blood DNA) were derived from almost fully methylated or unmethylated sequences. Almost all clones from blood DNA from two patients with a deletion of the maternal RB1 allele were derived from unmethylated sequences, whereas clones obtained from three patients with a paternally derived RB1 deletion were derived from almost fully methylated sequences. Each block of clones represents an individual. Open circles, unmethylated CpGs; filled circles, methylated CpGs. (B) Allelic expression analysis of the 2B-transcript in blood of a patient heterozygous for a rare variant in exon 3 inherited from the mother. Sequencing of RT–PCR products obtained with primers in exon 2B and exon 3 only showed the paternally derived C allele.

Figure 4. Expression profile of the 2B-transcript.

The 2B-transcript is expressed in all of the tissues studied.

Figure 5. Allelic expression imbalance of the RB1 gene.

Plot of the ratio of allelic expression as determined by SNaPshot primer extension on RT–PCR products obtained from RNA from blood of 14 individuals from 7 families informative for expressed variants (Table 1). The primer extension assay for the variant in exon 3 (family A) only detects the regular transcript whereas the assays for the variants downstream of exon 3 (families B to G) detect transcripts initiated in exon 2B in addition to regular transcripts. Of note, direction and extent of skewing in family A are not different from that in the other families and, therefore, the relative abundance of 2B-transcripts compared to regular transcripts is likely to be low. For each sample 3–5 independent experiments were performed. The top and bottom of the means diamonds represent the 95% confidence intervals for the means. Squares, male individuals; circles, female individuals; filled symbols, bilateral retinoblastoma; half-filled symbols, unilateral retinoblastoma; open symbols, unaffected. Asterisk marks individuals in whom parental origin of alleles is unknown.

Figure 6. Treatment of lymphoblastoid cells (LCs) with the demethylation drug 5-aza-2′-deoxycytidine (AzadC).

(A) Methylation analysis of CpG 85 by methylation-specific PCR (MS-PCR) and quantification of allelic expression imbalance of the RB1 gene. The top chart shows the methylation status of CpG 85 in blood, mock-treated and AzadC-treated LCs. The percentage of MS-PCR products specific for methylated and unmethylated alleles is indicated by black and grey bars, respectively. The bottom plot shows the ratio of allelic expression as determined by SNaPshot primer extension on RT–PCR products obtained from RNA. For each sample 3–8 independent experiments were performed. The top and bottom of the means diamonds represent the 95% confidence interval for the means. In family H, we could not investigate allelic RB1 expression in blood, because we did not have RNA from fresh blood. In this family, a male patient with unilateral retinoblastoma (HII-1) inherited the rare variant from his unaffected mother and transmitted it to his unaffected daughter (HIII-1). (B) Electropherograms of SNaPshot primer extension on RT–PCR products specific for the 2B-transcript. Black and red peaks correspond to C and T alleles, respectively. In A III-1, the C allele is known to be of paternal origin. Numbers next to peaks indicate peak areas. Numbers below electropherograms with two peaks show the ratios of peak areas (T-allele/C-allele).