Fig. 1.

Exogenously expressed input chimeric RNAs induce the expression of endogenous fusion transcripts. (A, Upper) Schematics of the designed input RNAs containing complete TMPRSS2 exon-1 (78 nt, uc002yzj.3) and ERG exon-4 (218 nt, uc021wjd.1), expressed in the sense or antisense orientation from the CMV or U6 promoters. pA, poly-A signal; ts, transcriptional stop “TTTTTT” for U6 promoter. (Lower) RT-PCR detection of induced fusion transcript (Upper gel) and input RNA (Lower gel). Antisense short fusion RNAs (lanes 3 and 4), but not sense short fusion RNAs (lanes 1 and 2), induced a band of fusion transcript. For negative controls, transfection with a parental plasmid expressing mCherry sequence (lane 5), DHT treatment without plasmid transfection (lane 6), cells without transfection and DHT treatment (lane 7), and PCR without cDNA (lane 8) all resulted in the absence of endogenous fusion transcripts. M, DNA markers. (B) Length and positional effect of antisense input RNAs. (Upper) Antisense input RNAs with 75 nt (blue) targeting ERG exon-4 and varying lengths (82, 67, 52, and 33 nt, green) targeting TMPRSS2. Dashed line links ERG and TMPRSS2 sequence in the input RNA and contains no sequence. A 20-nt mutation (black line) was introduced to the input RNAs to discern input RNAs from the induced fusion transcript (SI Appendix, Fig. S1B). (Lower) RT-PCR detection of induced fusion transcript (Top), input RNA (Middle), or detection of induced fusion transcript using a different primer pair (Bottom). (C) Expression of the corresponding sense input RNAs (Lower) all failed to induce the fusion transcript (Upper, lanes 1–4). (D) Induction by antisense-5 occurred at physiologically relevant DHT concentrations as low as 20 nM. Three-rounds of nested PCR were performed to reveal the lowest amount of DHT required. (E) Antisense-5 led to clear induction while antisense-3 and -7 did not, indicating that it is not the length of input RNAs but targeted regions that is critical.