Transcription termination by RNA polymerase III: uncoupling of polymerase release from termination signal recognition

Abstract

Xenopus RNA polymerase III specifically initiates transcription on poly(dC)-tailed DNA templates in the absence of other class III transcription factors normally required for transcription initiation. In experimental analyses of transcription termination using DNA fragments with a 5S rRNA gene positioned downstream of the tailed end, only 40% of the transcribing polymerase molecules terminate at the normally efficient Xenopus borealis somatic-type 5S rRNA terminators; the remaining 60% read through these signals and give rise to runoff transcripts. We find that the nascent RNA strand is inefficiently displaced from the DNA template during transcription elongation. Interestingly, only polymerases synthesizing a displaced RNA terminate at the 5S rRNA gene terminators; when the nascent RNA is not displaced from the template, read-through transcripts are synthesized. RNAs with 3' ends at the 5S rRNA gene terminators are judged to result from authentic termination events on the basis of multiple criteria, including kinetic properties, the precise 3' ends generated, release of transcripts from the template, and recycling of the polymerase. Even though only 40% of the polymerase molecules ultimately terminate at either of the tandem 5S rRNA gene terminators, virtually all polymerases pause there, demonstrating that termination signal recognition can be experimentally uncoupled from polymerase release. Thus, termination is dependent on RNA strand displacement during transcription elongation, whereas termination signal recognition is not. We interpret our results in terms of a two-step model for transcription termination in which polymerase release is dependent on the fate of the nascent RNA strand during transcription elongation.