Expression of eukaryotic protein-coding genes proceeds through multiple steps, including transcription, addition of a guanosine cap to the 5’-end of the nascent messenger RNA, splicing, cleavage of the mRNA precursor (pre-mRNA) to define the 3’ end of the transcript and in most cases, addition of a poly(A) tail. Like alternative splicing, 3' end processing of a given pre-mRNA can occur at different sites. This gives rise to mRNAs with different coding regions or 3' untranslated regions (UTRs) with different sets of binding sites for regulatory factors such as microRNAs. Recently, a global tendency towards the use of proximal polyadenylation sites has been reported in dividing compared to resting cells and in cancer cells relative to their normal counterparts, highlighting the importance of polyadenylation for cell physiology.

In collaboration with the group of Professor Mihaela Zavolan at the Biozentrum, we have generated the first combined maps of binding sites of core cleavage and polyadenylation factors and of the poly(A) sites that are used in a human cell type. In addition, we have identified factors whose binding is most informative for the location of the cleavage site. We further investigated the change in poly(A) site selection induced by the knockdown of CF I_m68 and CstF-64, two core components that were also reported previously to affect poly(A) site choice, and we found that the knockdown of CF I_m68 results in a global shift towards proximal poly(A) sites. Thus, changes in relative abundance of a single 3’ end processing factor can modulate the length of 3’ untranslated regions across the transcriptome and suggested a mechanism behind the previously observed increase in tumor cell invasiveness upon CF I_m68 knockdown. We now plan to investigate in detail the mechanism by which reduced levels of CF I_m68 results in the use of proximal poly(A) sites. Perhaps the concentration of 3’ end processing factors is limiting in dividing compared to resting cells. We will attempt to measure the mRNA as well as protein levels of several cleavage factors such as CF I_m and CstF in resting and in actively dividing cells but also in tumor tissue and tumor cell lines. Transcript levels will be determined by quantitative real time PCR or Northern blotting and protein levels by Western blotting or immunostaining. Preliminary data indicate that immortalized cells such as HEK293 have a high concentration of CF I_m68 in the cytoplasm compared to the nucleus. The reason for this is unknown but regulation of nuclear import could be a way to keep the concentration of this factor low in the nucleus. A further line of investigation will be to determine whether modifications such as phosphorylation or methylation influence the nuclear import of these factors.

The second project we pursued in the previous years concerns the cleavage and polyadenylation factor CPF of yeast. In collaboration with the group of Professor Andreas Engel we have determined the molecular mass of this complex and the stoichiometry of its subunits by scanning transmission electron microscopy. Professor Holger Stark (Max-Planck-Institute for Biophysical Chemistry, Göttingen) has established the three-dimensional structure of the complex at a resolution of 25Å from negatively stained preparations and by single-particle cryo-electron microscopy. CPF has a complex asymmetric architecture in which an outer protein wall surrounds a large inner cavity. Moreover, three GFP-tagged subunits could be located within the structure and the X-ray structure of poly(A) polymerase could be fitted to the complex. Professor Stark is now planning to fit the X-ray structure of DDB1 (DNA damage binding protein 1) to the CPF structure. Mammalian CPSF160 and yeast YHH1 are homologues of DDB1 over the entire protein sequence.