Open Access

Lorenzo Corsini

• The research presented in this thesis characterizes U2AF homology motifs (UHM) and their interactions with UHM ligand motifs (ULM) in the context of splicing regulation. UHM domains are a subgroup of RNA recognition motifs (RRM) originally discovered in the proteins U2AF65 and U2AF35. Whereas canonical RRMs are usually involved in binding of RNA, UHM domains bind tryptophan containing linear protein motifs (ULM) instead. In the first article, we analyze the complex network of interactions between splicing factors and RNA that initiate the assembly of the spliceosome at the 3´ splice site of an intron. The protein U2AF65 binds a pyrimidine-rich element in introns and recruits U2snRNP by binding its protein component SF3b155. My contribution was to define the binding site of the protein U2AF65 to the intrinsically unstructured N-terminus of the scaffolding protein SF3b155. I could show that the UHM domain of U2AF65 recognizes a ULM in SF3b155, and that this binding site is not overlapping with the binding sites of other splicing factors, like p14, to SF3b155. As the U2AF65-UHM:SF3b155-ULM interaction is mutually exclusive with an interaction between U2AF65-UHM and a ULM in the splicing factor SF1, which was reported to initially recognize the branch point sequence, my results provide the molecular details on how SF3b155 replaces SF1 during spliceosomal reorganizations. In the second article, we show that overexpression of the UHM domain of the splicing factor SPF45 induces exon 6 skipping in the pre-mRNA of Fas (CD95/APO-1). I provide evidence for in vitro binding of SPF45-UHM to ULM sequences in the splicing factors U2AF65, SF1, and SF3b155. I crystallized free and SF3b155-bound SPF45 UHM and solved both structures by X-ray crystallography. The analysis of the complex interface and sequence differences in the ULMs allowed me to design mutations of SPF45-UHM, which selectively inhibit binding to distinct ULMs. After assessing the ULM binding properties in vitro, we could show that the activity of SPF45-UHM in influencing the splicing pattern of Fas relies on interactions with SF3b155 and/or SF1, but that an interaction with U2AF65 is dispensable. A mechanism for the activity of SPF45-UHM could thus be engaging in ULM interactions and thus interfering with the network of interactions that initiate the assembly of the spliceosome at the 3´splice site, as described above. In the third article, we describe an unusual flexible homodimerization mode of the UHM in the splicing factor Puf60, which enables simultaneous interactions with ULM sequences on other splicing factors. I could show that the NMR relaxation properties of Puf60-UHM are inconsistent with a model of a rigid dimer, but rather indicate a dimerization via a flexible linker. I identified a flexible loop in the peptide backbone of Puf60-UHM, and showed that mutiation of acidic residues in this loop impairs the dimerization. To analyze the dimerization interface in further detail, I solved the structure of Puf60-UHM by X-ray crystallography. The acidic residues in the flexible loop of one UHM dimer subunit mediate the dimerization by contacting basic residues on the β-sheet surface of the other dimer subunit. Differences in the four dimer interfaces observed for the eight molecules in the asymmetric unit of the crystal support the model of an undescribed, flexible mode of dimerization, and thus complement the NMR relaxation data. Furthermore, I could show that the Puf60-UHM dimer and U2AF65-UHM contact different ULM sequences on the SF3b155 N-terminus in vitro, thus providing a possible explanation for the mutual cooperative activation of Puf60 and U2AF65 in splicing assays described in the literature. The fourth article is a review about recent research on the recognition of DNA double strand breaks (DSB) by covalent histone modifications. The p53 binding protein 1 (53BP1) is a DSB sensor and a checkpoint protein for mitosis. Recent crystallographic evidence indicates that 53BP1 recognizes DSB sites by binding histone H4 dimetylated at lysine 20 (H4-K20). We provide a comprehensive overview of the atomic resolution structures that revealed how proteins can specifically recognize histone tail modifications, especially methylated lysines, to read the information stored in what is called the histone code.
• Die Möglichkeit, unterschiedliche Exons oder gar unterschiedliche Teil-Exons in die reife mRNA einzubauen (alternatives Spleißen) vervielfacht die genetische Information der DNA in höheren Eukaryonten (Black, 2003). Noch während der Transkription durch RNA-Polymerase II werden Introns und Exons auf dem entstehenden Transkript von Spleiß-Faktoren erkannt. Die 5´-Spleiß-Stelle (SS) eines jeden Introns wird als erstes Sequenz-Element von der Transkriptions- Maschinerie synthetisiert und von U1-snRNP (engl. U1 small nuclear ribonuclear particle) erkannt (Puig et al., 1999; Rosbash and Seraphin, 1991). Der Spleiß-Faktor 1 (SF1) erkennt die Sequenz für die spätere Lariat-Verzweigung (BPS, engl. branch point sequence) (Berglund et al., 1997). SF1 bindet die BPS nur schwach, diese Interaktion wird aber kooperativ durch die simultane Bindung an den Proteinkomplex U2AF verstärkt. U2AF (engl. U2-snRNP auxiliary factor) besteht aus einer großen und einer kleinen Untereinheit, von denen die große (U2AF65) einen Pyrimidin-reichen Abschnitt (Py-Trakt) zwischen BPS und 3´-SS erkennt (Banerjee et al., 2003; Singh et al., 1995; Zamore et al., 1992), während und die kleine Untereinheit, U2AF35, die 3-SS selbst bindet (Berglund et al., 1998; Merendino et al., 1999; Wu and Maniatis, 1993; Zorio and Blumenthal, 1999). Die oben genannten Proteine und U1-snRNP definieren die Intron-Grenzen in einem noch relativ schwach verbundenen Komplex, der als E (engl. early) bezeichnet wird. Komplex E wird in ATP-abhängiger Weise zu Komplex A umstrukturiert, wobei der RNA-Bestandteil von U2snRNP mit der BPS (engl: branch point sequence) Basenpaarungen eingeht (Brow, 2002). SF3b155, ein Protein des U2snRNP, bindet dabei an U2AF65 und verdrängt SF1 (Gozani et al., 1998). Die RNA-Bindedomäne (RRM, engl. RNA recognition motif) von U2AF35 und die dritte RRM von U2AF65 unterscheiden sich von gewöhnlichen RRM-Domänen darin, dass sie statt RNA an Tryptophan-haltige lineare Protein-Sequenzen binden, und werden daher UHM (U2AF Homologie Motif) genannt. Die Proteine SF3b155 und SF1 enthalten solche Tryptophan-haltigen Motive (ULM, UHM Ligand-Motif), mit denen sie an die UHM in U2AF65 binden. Im N-Terminus von U2AF65 befindet sich ebenfalls ein ULM, das an die UHM von U2AF35 bindet und so die zwei Untereinheiten des U2AF-Komplexes verbindet. Interessanterweise befinden sich UHM-Domänen in mehreren Proteinen, die alle mit Spleißen und/oder alternativem Spleißen in Verbindung gebracht wurden. In meiner Arbeit am Europäischen Molekularbiologie Laboratorium habe ich mich damit beschäftigt, wie Proteine, die eine UHM enthalten, mit der Spleiß-Maschinerie interagieren und sie beeinflussen können. Ich habe mich dabei auf die UHM-Domänen in den Spleiß-Faktoren U2AF65, SPF45, und PUF60 konzentriert und deren Bindung an die ULM-Sequenzen in U2AF65, SF1 und SF3b155 untersucht. ...