Role for piRNAs and noncoding RNA in de novo DNA methylation of the imprinted mouse Rasgrf1 locus

Abstract

Genomic imprinting causes parental origin-specific monoallelic gene expression through differential DNA methylation established in the parental germ line. However, the mechanisms underlying how specific sequences are selectively methylated are not fully understood. We have found that the components of the PIWI-interacting RNA (piRNA) pathway are required for de novo methylation of the differentially methylated region (DMR) of the imprinted mouse Rasgrf1 locus, but not other paternally imprinted loci. A retrotransposon sequence within a noncoding RNA spanning the DMR was targeted by piRNAs generated from a different locus. A direct repeat in the DMR, which is required for the methylation and imprinting of Rasgrf1, served as a promoter for this RNA. We propose a model in which piRNAs and a target RNA direct the sequence-specific methylation of Rasgrf1.

Fig. 1

DNA methylation of the imprinted Rasgrf1 DMR is impaired in piRNA pathway mutants. (A) Methylation status of a portion of the Rasgrf1 DMR [corresponding to a part of region K in (B)] in spermatogonia from Mili, MitoPLD, and Miwi2 mutants revealed by bisulfite sequencing. Open and filled circles represent unmethylated and methylated cytosines, respectively. A result on 12.5-dpc wild-type whole-embryo DNA is also shown (control, 50% methylation expected). Percentage of methylation is shown below each panel. (B) Methylation status across the Rasgrf1 DMR in MitoPLD mutants. The green and gray bars show the gametic DMRs (regions differentially methylated between oocyte and sperm) (30), of which the green one is important for imprinting. Positions of CpGs are indicated by short vertical bars. Repeat sequences identified by RepeatMasker are shown as colored boxes: green, LINE; purple, short interspersed nucleotide element (SINE); white, simple repeat; gray, LTR; light gray, DNA transposon. A direct repeat consisting of 40 copies of a 41-oligomer (18) and a copy of RMER4B solo LTR are also indicated.

Fig. 2

Decreased piRNA levels in piRNA pathway mutants. (A) Size profiles are shown for small RNAs from wild-type (control), Mili^−/−, MitoPLD^−/−, and Miwi2^−/− testes at 16.5 dpc. The peak at 22 nt represents miRNAs, and the other at 24 to 30 nt represents piRNAs. (B) Relative levels of Mili-bound (26-nt) and Miwi2-bound (29-nt) piRNAs compared with wild-type. The levels of miRNA were used to normalize the piRNA levels.

Fig. 3

An RMER4B sequence in the Rasgrf1 DMR is targeted by piRNAs. (A) Numbers and locations of small RNAs from MitoPLD^+/+ (left) and MitoPLD^−/− testes (right) at 16.5 dpc mapping to the Rasgrf1 DMR (correspond to mm9 chr9: 89769659–89779158). Results for small RNAs with a unique hit (top) and those with mismatches (up to two mismatches including indels) are shown (bottom). (B) Detailed mapping of piRNAs in the RMER4B copy of the Rasgrf1 DMR. piRNAs are represented by bars colored according to the number of sequence reads obtained. Plus strand hits are shown above the sequence, and minus strand hits are shown below it. Mismatches are represented by black portions of the bars. Unique hit piRNAs are indicated.

Fig. 4

The Rasgrf1 direct repeat serves as a promoter for pit-RNA. (A) Detection of pit-RNA by reverse transcription polymerase chain reaction (RT-PCR) in germ cells from 16.5-dpc testes. Oct4 is a marker for germ cells. (B) Detection of a pit-RNA cleavage site by modified RACE. The numbers of cDNA ends obtained by sequencing the recovered RACE products are also shown (right). (C) Transcription start-site analysis by 5′ RACE. The number of cDNA ends determined by sequencing the RACE products is shown with an arrow in the schematic representation of the region (bottom). The positions of the PCR primers for the analyses are also illustrated. (D) Quantitative PCR analysis of pit-RNA in wild-type and MitoPLD^−/− testes at 16.5 dpc. Error bars represent standard deviation (n = 3). The level of Oct4 mRNA was used as a reference. (E) Quantitative PCR analysis of pit-RNA in wild-type and direct repeat mutant testes at 16.5 dpc. Error bars represent standard deviation (n = 3). The level of Oct4 mRNA was used as a reference.