A new class of non-coding RNAs associated with 3’ untranslated regions of mRNAs

The importance of non-coding RNAs (ncRNAs) in controlling gene expression is becoming increasingly evident. However, except for some well characterized examples, such as miRNAs, Xist and Air, the function of most non-coding transcripts is still to be determined. Moreover, while small regulatory RNAs can be relatively easily classified on the basis of their length, secondary structure, and biochemical pathway, the classification of long “mRNA-like” ncRNAs has been problematic. Here we identify a large class of non-coding transcripts that originate within the 3’UTR of at least one third of all genes in the mouse genome. We have several lines of evidence from genome-wide bioinformatic analyses (EST coverage, CAGE data, chromatin state maps of active promoters) and from invitro studies (strand-specific RT-PCR, 5’RACE, Northern blot) showing that these 3’UTR-associated ncRNAs (uaRNAs) can be either linked or transcribed separately to the upstream protein-coding sequences. In addition, expression profiles obtained by custom-designed microarrays on three different developmental systems (myoblast differentiation, male gonadal ridge formation, embryonic stem cell differentiation) showed that uaRNA expression is highly regulated and tissue-specific, and might be either concordant or discordant with respect to the upstream coding region depending on the cell type and on the developmental stage. This observation is confirmed by in-situ hybridization experiments, which evidenced that uaRNA and the associated coding transcript might have different subcellular locations. Our results highlight a further level of complexity at 3’UTRs, suggesting the presence of new regulatory mechanisms that control gene expression during embryonic development. Our data have also important implications for the design of in-situ hybridization and microarray probes as well as for the interpretation of gene expression dat