Today, it is widely accepted that RNA chain elongation and termination by RNA poly-merase II (pol II) is a complex process that is coordinated with 3′-end processing and polyadenylation of the primary transcript. Just over 20 years ago, however, the identification of the first elongation factor, TFIIS or SII, provided the initial indication that RNA pol II transcription could be regulated at the level of elongation.

Back in 1973, Natori and colleagues identified SII by its ability to stimulate transcription in vitro and to enable pol II to synthesize long transcripts. Yet its mode of action remained unknown until 1992, when three groups added a considerable piece to the puzzle by providing insights into the mechanism of SII activity. Reines, and Izban and Luse, noticed that the addition of SII caused a shortening of transcripts associated with stalled RNA pol II. They found that, in the presence of SII, the RNA pol II complex can serve as a nuclease, cleaving its nascent transcript from the 3′ end. Wang and Hawley also presented evidence to support these observations, and proposed a possible proofreading role for the activity described. Surprisingly, it was noted that this process leaves the pol II complex intact and the remaining transcript can subsequently be elongated. The nuclease activity that is stimulated by SII helps pol II bypass specific blocks to elongation and therefore increases elongation efficiency.

A question that puzzled the community was how 3′-end processing was linked to termination. The connection between these processes was established when it became apparent that polyadenylation and transcription termination were dependent on the same DNA sequences at the 3′ ends of genes. A role for poly(A) site cleavage in termination was first established by two groups — Logan and colleagues, and Connelly and Manley. Based on the hypothesis that polyadenylation must be a prerequisite for RNA pol II termination, because this would ensure that mRNA-coding sequences were completely transcribed before a termination event occurred, they introduced several single-base-pair mutations into the polyadenylation motifs, and showed abrogation of both polyadenylation and termination.

Two models were proposed to explain these results. The first postulated that the emergence of polyadenylation sequences on the RNA triggers a change in the factors associated with the polymerase, which eventually results in termination. The second, also known as the 'torpedo' model, states that cleavage of the transcript is required to trigger termination. Recent evidence supports both models and, until the mystery is solved, only one thing is clear: it ain't over until pol II falls off.