NMR investigations of 15.5K associated protein-RNA complexes

Abstract

The spliceosome, consisting of different snRNPs and numerous non-U snRNP factors, catalyzes the splicing reaction. To activate the spliceosome and enable the first step of splicing to begin, the extensively base-paired U4, U6 snRNAs in the U4/U6 snRNP dissociate from each other. The U4, U6 snRNAs are associated to the common core proteins of the spliceosome namely the Sm and LSm proteins, as well as the specific proteins including 15.5K, hPrp31, and the CypH/hPrp4/hPrp3 protein tri-complex. The knowledge on these proteins is of paramount importance for the understanding of the dramatic structural rearrangement of the U4/U6 snRNA duplex prior to splicing. In this work, effort was made to understand the interactions between two of these specific proteins namely 15.5K and hPrp31 in the context of the U4 snRNP. Using HSQC titrations and cross-saturation experiments on the hPrp31-15.5K-U4 5"-SL and hPrp3178-333 -15.5K-U4 5"-SL complexes, we defined the interaction surface on 15.5K in complex with hPrp31. Combining the NMR and biochemical data, we successfully generated a 3D model of the ternary complex using comparative modelling and the HADDOCK2.0 docking program. From these results, we characterized the Nop domain as a bona fide RNP binding domain. The role of the assembly-initiating 15.5K is, therefore, not restricted to inducing or stabilizing a hPrp31 binding site in the RNA; rather 15.5K itself provides approximately half of the contact surface for the Nop domain of hPrp31. In the docking model, the interaction surfaces on 15.5K and hPrp31 showed charge complement. From our docking model we could also demonstrate that the elongation of stem II in U4 snRNA is highly unfavourable due to the physical barrier provided by the Nop domain. As in box C/D snoRNAs the length of the stem II naturally exceeds the required length, hPrp31 is not recruited into the box C/D snoRNPs. A surface mutant of 15.5K, which contains mutations outside the interaction surface between 15.5K and hPrp31, was previously shown to strongly reduce the binding of hPrp31. Using HSQC titration and RDC refinement, we could demonstrate that the structure of this mutant does not significantly vary from the wild type and therefore, the effect of this 15.5K mutant on hPrp31 binding could arise from subtle changes in the charge property of the interaction surface. Furthermore, the structural changes of 15.5K in complexes with different box C/D snoRNA constructs were studied using HSQC titrations to address the question whether the structure of 15.5K also contributes to the selectivity of these primary RNPs. The chemical shift perturbation data showed that the structure of 15.5K does not differ dramatically in this primary RNPs. Therefore, the selectivity arises primarily from the differences in the RNAs and the secondary binding proteins.