UCLA

A major focus of my studies is to understand how mRNA transcription is functionally interconnected, or coupled, to 3'-end processing (cleavage and polyadenylation) of mRNAs . A product of this coupling effect is that transcription can affect and enhance 3'-end processing. To understand how transcription is able to affect 3'-end processing, I investigated a few mechanisms that have been suggested in the literature to be the primary modes of coupling 3'-end processing to transcription: recruitment of processing factors to the promoter, cotranscriptional recruitment of processing factors during gradual extrusion of the poly(A) signal from the polymerase, and phosphorylation of C-terminal domain (CTD) of RNA polymerase II (Chapter 3). Using an in vitro system that can support 3'-end processing that is functionally coupled to transcription, I tested the extent to which each of these mechanisms contribute to coupling in our in vitro system. I demonstrate that recruitment of processing factors to the promoter or to the transcription elongation complex prior to transcription of the poly(A) signal is not necessary for coupled 3'-end processing in our in vitro system. I found that assembly of cleavage and polyadenylation factors during gradual extrusion of the poly(A) signal from the polymerase is not important for coupling. I also show that the phosphates put on the C-terminal domain (CTD) of the transcribing polymerase at or shortly after the time of transcription initiation are unimportant for coupling in our system. Surprisingly, nothing that happens at the promoter is necessary to direct coupling since transcription elongation complexes (TEC) initiated on oligo(dT) tailed templates can still engage in 3'-end processing that is an order of magnitude more efficient than processing uncoupled from transcription. Therefore, these results argue that the ability to couple 3'-end processing to transcription is intrinsic to the structure of the TEC. Consistent with this, I found that coupling is inhibited if the RNA tether which connects the RNA transcript to the polymerase is severed. Based on these results, we propose that coupling is somehow mediated by having the poly(A) signal held close to the CTD via an intact RNA tether in the context of the TEC.

Another aspect of this coupling phenomenon was demonstrated by Blencowe's group in which they reported that transcriptional activators can contribute to 3'-end cleavage in vivo. It is intriguing that transcriptional activators that are recruited to the promoter can affect a process that occurs at the other end of the gene. To understand this further, I investigated the effect of transcriptional activators on 3'-end processing in vitro (Chapter 5). I found that while transcriptional activators can dramatically enhance transcription, they did not alter 3'-end processing levels significantly in a coupled in vitro system. However, I did observe that transcription initiates randomly in the absence of a transcriptional activator to direct promoter-specific transcription. At face value, these results suggest that the mechanism by which transcriptional activators affect 3'-end processing cannot be supported in our in vitro system. However, my results do suggest a possibility that transcriptional activators may affect cleavage levels by redirecting initiation from cryptic promoters to promoter-specific initiation.

It has been suggested that coupling of 3'-end processing to transcription may be in part mediated by speeding up the assembly of the cleavage and polyadenylation apparatus (CPA) at the poly(A) signal. One way that this is thought to be achieved is by early recruitment of processing factors to the promoter during initiation and its subsequent transfer to the elongating polymerase so that assembly can begin as soon as the poly(A) signal is transcribed. While this may be the case, very little is known regarding the actual assembly process of the CPA and the rate-limiting step that may be subject to regulation. Therefore, I sought to dissect the assembly pathway of the CPA in hopes of revealing how 3'-end processing may be regulated by transcription. I decided to begin by resolving whether ATP is a required cofactor for 3'-end processing since there have been conflicting reports in the literature. Interestingly, I found that ATP is indeed a required cofactor for 3'-end cleavage in our in vitro system since withholding ATP can inhibit cleavage (Chapter 4). I found that this inhibition is reversible since cleavage resumes upon re-addition of ATP. I also demonstrate that a partial CPA can form in the absence of ATP and this partial apparatus consists of at least CPSF and CstF, the core cleavage and polyadenylation complexes. Based on preliminary data, the rate-limiting step of the cleavage pathway occurs during or after the ATP-requiring step. These results suggest that ATP may play an important role in regulation of 3'-end cleavage. The significance of an ATP requirement for 3'-end cleavage in our in vitro system will be discussed.

An interesting product of the interconnection between 3'-end processing to transcription is the possibility that there may be a coordinated surveillance mechanism that can assess the quality of the transcript being made and decides the fate of the transcript by directing it to process or to degrade. Here, we studied poly(A)-dependent pausing in vitro, which has been proposed as a surveillance checkpoint, and poly(A)-dependent degradation of unprocessed transcripts from weak poly(A) signals (Chapter 2). We confirm directly, by measuring the length of RNA within isolated transcription complexes, that a newly transcribed poly(A) signal reduces the rate of elongation by RNA polymerase II, resulting in the accumulation of these complexes downstream of the poly(A) signal. We then show that if the RNA in these elongation complexes contain a functional but unprocessed poly(A) signal, degradation of the transcripts ensues. We propose that during normal 3'-end processing, a decision is made whether to process or to degrade. In the case of weak poly(A) signals, where cleavage at the poly(A) site is slow, the default pathway to degradation predominates.