Antisense transcription of the myotonic dystrophy locus yields low-abundant RNAs with and without (CAG)n repeat

The unstable (CTG·CAG)n trinucleotide repeat in the myotonic dystrophy type 1 (DM1) locus is bidirectionally transcribed from genes with terminal overlap. By transcription in the sense direction, the <i>DMPK</i> gene produces various alternatively spliced mRNAs with a (CUG)n repeat in their 3′ UTR. Expression in opposite orientation reportedly yields (CAG)n-repeat containing RNA, but both structure and biological significance of this antisense gene (<i>DM1-AS</i>) are largely unknown. Via a combinatorial approach of computational and experimental analyses of RNA from unaffected individuals and DM1 patients we discovered that <i>DM1-AS</i> spans &gt;6 kb, contains alternative transcription start sites and uses alternative polyadenylation sites up- and downstream of the (CAG)n repeat. Moreover, its primary transcripts undergo alternative splicing, whereby the (CAG)n segment is removed as part of an intron. Thus, in patients a mixture of <i>DM1-AS</i> RNAs with and without expanded (CAG)n repeat are produced. <i>DM1-AS</i> expression appears upregulated in patients, but transcript abundance remains very low in all tissues analyzed. Our data suggest that <i>DM1-AS</i> transcripts belong to the class of long non-coding RNAs. These and other biologically relevant implications for how (CAG)n-expanded transcripts may contribute to DM1 pathology can now be explored experimentally.