The influence of non-coding RNAs on allele-specific gene expression in mammals

Human Molecular Genetics, 2005, Vol. 14, Review Issue 1 doi:10.1093/hmg/ddi108 R113–R120 The influence of non-coding RNAs on allele-specific gene expression in mammals Michael J. O’Neill* Department of Molecular and Cell Biology, University of Connecticut, Storrs, CT 06235, USA Received January 10, 2005; Revised and Accepted February 15, 2005 INTRODUCTION Despite pronouncements to the contrary, consensus is currently an elusive concept. The difficulty emanates from differences that are likely born of a conflict at the core of our identity. The divisions run deep and wide, from discord between neighbors so near they overlap to antagonism between participants so remote they would seem insignificant to one another. In addition, we ponder the influence of what once were thought merely to be passive purveyors of information. We, of course, are not speaking of our identity in regards to party affiliation, but rather phylogenetic affiliation of our identity as mammals. The conflict is not a political one, but an intragenomic one; the antagonism is not ideological but epigenetic and the influential purveyors not media conglomerates but non-coding RNAs. And to put the metaphor to rest, the lack of consensus stems not from a profusion of irreconcilable cultural mores, but from our inability to arrive at a mechanistic consensus about the role, in mammals, of non-coding RNAs in the regulation of gene expression. This review attempts to gather much of the current knowledge about non-coding RNAs involved in parent-of-origin specific expression at eight heavily studied loci in mammals (Table 1). Each of these loci are subject to genomic imprinting: the epigenetic marking of alleles through differential cytosine methylation or chromatin modifications, which result in allele-specific transcriptional silencing during embryonic development. The limitations of space in the face of the extraordinarily complex nature of transcriptional regulation at these loci, involving far more than the production of non-coding RNAs, insure that this review will unfortunately give short shrift to the brilliant work of many. With apologies to those left out, it is hoped that this review may at least serve as a sort of crib sheet for students in the field. For more thorough reviews about the individual topics, readers are referred to work of Verona et al. (1) and Peters and Beechey (2) for genomic imprinting; Plath et al. (3) and Meard (4) for X-inactivation and Lippman and Martienssen (5) and Lavorgna et al. (6) for non-coding RNA. IGF2/H19 LOCUS This cluster (human 11p15.5 and mouse distal 7) was not only the first imprinted locus to be identified (7), but also provided the first example of a spliced, poly-adenylated non-coding RNA ever identified, H19 (8). This locus features reciprocal imprinted expression of H19 (maternally active) (9) and IGF2 (paternally active) (Fig. 1A). Mutations disrupting imprinted expression of IGF2 underlie a substantial proportion of cases *To whom correspondence should be addressed. Email: # The Author 2005. Published by Oxford University Press. All rights reserved. For Permissions, please email: Current research has revealed that the influence of RNA molecules on gene expression reaches beyond the realm of protein synthesis back into the nucleus, where it not only dictates the transcriptional activity of genes, but also shapes the chromatin architecture of extensive regions of DNA. Non-coding RNA, in the context of this review, refers to transcripts expressed and processed in the nucleus much like any protein coding gene, but lacking an open reading frame and often transcribed antisense to bona fide protein coding genes. In mammals, these types of transcripts are highly coincident with allele-specific silencing of imprinted genes and have a proven role in dosage compensation via X-inactivation. The biochemistry of how noncoding RNAs regulate transcription is the subject of intense research in both prokaryotic and eukaryotic models. Mechanisms such as RNA interference may have deep phylogenetic roots, but their relevance to imprinting and X-inactivation in mammals has not been proven. The remarkable diversity of non-coding transcription associated with parent-of-origin directed gene silencing hints at an equally diverse assortment of mechanisms. R114 Human Molecular Genetics, 2005, Vol. 14, Review Issue 1 Table 1. A partial list of imprinted loci exhibiting non-coding RNA transcription Locus Non-coding RNA Expressed parental allele Silencing targets Igf2/H19 Igf2r H19 Air Maternal Paternal Kcnq1 Dlk/Gtl2 Kcnq1ot Anti-Rtl1 (mir127 and mir136 ) C/D snoRNA genes miRNA genes and Mrg Anti-Dio3 Ube3a-ats and snoRNA genes Nespas 1A Xist Tsix Paternal Maternal Maternal Maternal Maternal Paternal Paternal Paternal Paternal
Maternal
None Igf2r Slc22a2 Slc22a3 Unknown Unknown PWS/AS Gnas Xist Ube3a Nesp Gsa X chr in cis Xist of the congenital growth disorder, Beckwith – Wiedemann syndrome (BWS) in humans (see KCNQ1 locus subsequently) (10,11). A great deal of work over the past 15 years involving genetic manipulation of this locus in mice indicates that the H19 transcript itself has no apparent role in the imprinted expression of its neighboring genes. Most compelling of these experiments was a ‘clean’ knockout of H19 in mice, leaving the promoter and surrounding transcription unit intact but removing the entirety of the RNA coding sequence (12). The clean knockout had no discernable phenotype and no effect on the imprinted expression of Igf2. Several studies (13 – 15) show that the key imprinting control element for this locus is a differentially methylated domain (DMD) 2– 4 kb upstream of the H19 transcription start, the H19-DMD. Four repeat units within the DMD contain closely apposed but separable DNA elements, which attract CpG methylation or bind the chromatin insulator nucleating factor, CTCF (16,17). The presence of methylated CpGs in the DMD on the paternal chromosome prevents the assembly of the insulator and allows enhancers downstream of H19 to interact with the promoter of Igf2, driving its expression in a tissue-specific manner (18). The dispensability of the H19 transcript to the imprinted expression of Igf2 and to the normal mouse development suggests that the RNA may be non-functional. The relatively high sequence conservation of H19 among mammals (77% identity between human and mouse) (8), however, indicates that the gene is subject to purifying selection, a hallmark of functional sequences. In addition to H19, other non-coding RNAs emanating from transcription units in the Igf2/H19 region have been identified (19,20). Some of these are expressed in an imprinted fashion, whereas others are expressed bi-allelically, but their role in the transcriptional activity of this locus is currently unknown. IGF2R/M6PR LOCUS A key element of the growth regulatory axis of IGF2 is the IGF2 receptor (IGF2R/M6PR ), an antagonist of IGF2 mitogenic acti (...truncated)