The entry point of this project was a clinical study that aimed to identify the genetic factors predisposing children to develop thrombosis. At that time, a common mutation of the coagulation factor II gene (also referred to as prothrombin) was found, which affects about 2% of the North-West European population. This mutation increases prothrombin plasma levels approximately 1.5-fold and the risk of thrombosis approximately 5-fold. Interestingly, this mutation affects the 3’ terminal nucleotide of the mRNA exactly at the site where the pre-mRNA is cleaved and polyadenylated. This location suggested to us that altered RNA processing might be involved in causing the observed increase of prothrombin gene expression by an unidentified molecular mechanism. In our subsequent functional analysis of this mutation we discovered an increase of 3‘end processing efficiency as a novel principle for the pathogenesis of a hereditary disorder (Gehring et al. 2001).
Further studies identified that inefficient 3’ end prothrombin mRNA processing is physiological and has been conserved in evolution. Moreover, we discovered that the prothrombin 3’ end processing signal is characterized by a delicate balance of positive and negative signals, which can be disturbed by a number of clinically relevant gain-of-function mutations in this area (Danckwardt et al. 2004, 2006a & 2006b). One of these mutations represents the first example of a pathologically relevant mutation affecting the CstF binding site in the 3’ flanking sequence of a human gene (Danckwardt et al., 2004).
One important component of the unusual architecture of the prothrombin 3’ end processing site is an upstream sequence element (USE) that is characterized by overlapping conserved 3’ UTR sequence motifs and which is located at a critical distance to the 3’ end processing signal within the 3’ UTR. We have subsequently discovered prothrombin to represent the prototype of a novel class of genes, which is characterized by low efficiency 3’ end processing signals that are balanced by an activating USE. We therefore addressed the mechanism of USE function and identified proteins that specifically interact with the USE to promote 3' end formation. Interestingly, these factors included splicing proteins (U2AF35 and U2AF65, hnRNPI) that also promote 3' end formation via USEs contained in a variety of other RNAs, indicating a broader functional role of this type of 3’ end processing in many other mammalian genes (Danckwardt et al 2007 & 2008).