Abstract
Poly(A) signals located at the 3′ end of eukaryotic genes drive cleavage and polyadenylation at the same end of pre-mRNA. Although these sequences are expected only at the 3′ end of genes, we found that strong poly(A) signals are also predicted within the 5′ untranslated regions (UTRs) of many Drosophila melanogaster mRNAs. Most of these 5′ poly(A) signals have little influence on the processing of the endogenous transcripts, but they are very active when placed at the 3′ end of reporter genes. In investigating these unexpected observations, we discovered that both these novel poly(A) signals and standard poly(A) signals become functionally silent when they are positioned close to transcription start sites in either Drosophila or human cells. This indicates that the stage when the poly(A) signal emerges from the polymerase II (Pol II) transcription complex determines whether a putative poly(A) signal is recognized as functional. The data suggest that this mechanism, which probably prevents cryptic poly(A) signals from causing premature transcription termination, depends on low Ser2 phosphorylation of the C-terminal domain of Pol II and inefficient recruitment of processing factors.
ACKNOWLEDGMENTS
We thank Matthias Soller, Alicia Hidalgo, and Steve Dove and their lab members for useful discussions and for sharing equipment, Domenico Libri and Steve Buratowski for valuable discussions and experimental suggestions, and Yi Jin Liew for bioinformatics help. We also thank Domenico Libri, Bob Michell, and Nick Proudfoot for critically reading the manuscript.
This study was supported by a Royal Society University Research Fellowship and a Wellcome Trust project grant to S.B. and by a Dorothy Hodgkin Postgraduate Award to J.G.
Supplemental material for this article may be found at http://dx.doi.org/10.1128/MCB.00919-10.