Short homologous sequences are strongly associated with the generation of chimeric RNAs in eukaryotes

Chimeric RNAs have been reported in varieties of organisms and are conventionally thought to be produced by trans-splicing of two or more distinct transcripts. Here, we conducted a large-scale search for chimeric RNAs in the budding yeast, fruit fly, mouse, and human. Thousands of chimeric transcripts were identified in these organisms except in yeast, in which five chimeric RNAs were observed. RT-PCR experiments for a sample of yeast and fly chimeric transcripts using specific primers show that about one-third of these chimeric RNAs can be reproduced. The results suggest that at least a considerable amount of chimeric RNAs is unlikely from aberrant transcription or splicing, and thus formation of chimeric RNAs is probably a widespread process and can greatly contribute to the complexity of the transcriptome and proteome of organisms. However, only a small fraction (<20%) of these chimeric RNAs has GU-AG at the junction sequences which fits the classical trans-splicing model. In contrast, we observed that about half of the chimeric RNAs have short homologous sequences (SHSs) at the junction sites of the source sequences. Our sequence mutation experiments in yeast showed that disruption of SHSs resulted in the disappearance of the corresponding chimeric RNAs, suggesting that SHSs are essential for generating this kind of chimeric RNA. In addition to the classical trans-splicing model, we propose a new model, the transcriptional slippage model, to explain the generation of those chimeric RNAs synthesized from templates with SHSs.