Although the synthesis of RNA from a DNA template is (and must be) a generally very stable process to enable transcription of kilobase transcripts, it has long been known that during initial transcription of the first 8–10 bases of RNA complexes are relatively unstable, leading to the release of short abortive RNA transcripts. A wealth of structural data in the past decade has led to specific mechanistic models elaborating an earlier “stressed intermediate” model for initial transcription. In this study, we test fundamental predictions of each of these models in the simple model enzyme T7 RNA polymerase. Nicking or gapping the nontranscribed template DNA immediately upstream of the growing hybrid yields no systematic reduction in abortive falloff, demonstrating clearly that compaction or “scrunching” of this DNA is not a source of functional instability. Similarly, transcription on DNA in which the nontemplate strand in the initially transcribed region is either mismatched or removed altogether leads to at most modest reductions in abortive falloff, indicating that expansion or “scrunching” of the bubble is not the primary driving force for abortive cycling. Finally, energetic stress derived from the observed steric clash of the growing hybrid against the N-terminal domain contributes at most mildly to abortive cycling, as the addition of steric bulk (additional RNA bases) at the upstream end of the hybrid does not lead to predicted positional shifts in observed abortive patterns. We conclude that while structural changes (scrunching) clearly occur in initial transcription, stress from these changes is not the primary force driving abortive cycling.

A recent model for the mechanism of intrinsic transcription termination involves dissociation of the RNA from forward-translocated (hypertranslocated) states of the complex [Yarnell WS, Roberts JW (1999) Science, 284:611–615]. The current study demonstrates that halted elongation complexes of T7 RNA polymerase in the absence of termination signals can also dissociate via a forward-translocation mechanism. Shortening of the downstream DNA or the introduction of a stretch of mismatched DNA immediately downstream of the halt site reduces a barrier to forward translocation and correspondingly reduces the lifetime of halted complexes. Conversely, introduction of a cross-link downstream of the halt site increases the same barrier and leads to an increase in complex lifetime. Introduction of a mismatch within the bubble reduces a driving force for forward translocation and correspondingly increases the lifetime of the complex, but only for mismatches at the upstream edge of the bubble, as predicted by the model. Mismatching only the two most upstream of the eight bases in the bubble provides a maximal increase in complex stability, suggesting that dissociation occurs primarily from early forward-translocated states. Finally, addition in trans of an oligonucleotide complementary to the nascent RNA just beyond the hybrid complements the loss of driving force derived from placement of a mismatch within the bubble, confirming the expected additivity of effects. Thus, forward translocation is likely a general mechanism for dissociation of elongation complexes, both in the presence and absence of intrinsic termination signals.