Figure 11.

Hierarchical diagram of structure determination for the U4 core domain. In stage I, the structures of two Sm-protein heterodimers were solved, which defined the Sm fold for the protein family and led to a ring model for the organization of seven Sm proteins in the snRNP core domain (Kambach et al., 1999 ▸). In stage II, structures of the U4 core domain (Leung et al., 2011 ▸) and the U1 snRNP (Pomeranz Krummel et al., 2009 ▸) were solved, with the latter containing the core domain plus two accessory proteins. The protein part of the U4 structure helped to interpret the U1 map (Oubridge et al., 2009 ▸). However, twinning of the U4 crystals prevented the determination of an accurate structure, and the U1 snRNP structure at limited resolution could not resolve the mechanism of 5′ splice site recognition. In stage III, two subcomplexes designed from the 5.5 Å resolution structure of U1 snRNP crystallized well. Of these, the minimal U1 snRNP is closely homologous to the U4 core domain; its structure was determined by molecular replacement from the protein part of the previous U4 structure (U4prev) and the model was completed after density modification (Kondo et al., 2015 ▸). Using the core of minimal U1 snRNP, and a hybrid model from this and U4prev as alternative models for molecular replacement, the U4 core-domain structure was re-refined. An OMIT map from the hybrid model proved that the Sm-site nucleotides are bound in the same manner in the two core domains except at the one position where their sequences differ.