Investigation of the 3D structure of the human activated spliceosome by cryo-electron microscopy

The spliceosome is a highly dynamic megadalton ribonucleoprotein (RNP) complex that catalyses the removal of introns from eukaryotic precursor messenger RNA (pre mRNA) in two consecutive transesterification reactions. The spliceosome is assembled de novo on each pre-mRNA intron by the sequential recruitment of five RNA protein complexes (snRNPs) and numerous non snRNP factors in a dynamic manner, driven by numerous DExH/D box ATPases or RNA helicases. After the U1 and U2 snRNPs bind to the 5’ splice site (ss) and branch site (BS) of the pre-mRNA, respectively, the pre formed U4/U6.U5 tri snRNP is recruited to give spliceosomal complex B, which lacks an active site. Catalytic activation of the B complex occurs in a stepwise manner. Initially, Brr2 RNA helicase dissociates the U4/U6 snRNA duplex, allowing U6 snRNA to restructure and to form an intricate U2/U6 RNA network, which is at the heart of the catalytic centre of the spliceosome. During activation, all of the U4/U6 and several U5 proteins are dissociated, while more than 20 new proteins including the Prp19 complex proteins (NTC) are stably integrated into the spliceosome, yielding the Bact complex, which is still pre catalytic. Catalytic activation requires the ATP dependent action of the RNA helicase Prp2, which displaces several proteins and remodels the U2 SF3A and SF3B complex proteins, yielding the B* complex. B* then catalyses the first step of the splicing reaction, generating the C complex. Following an additional restructuring step, the resulting C* complex catalyses the 2nd step of the splicing reaction to form the mRNA product. During the last two years, high-resolution electron cryo-microscopy (cryo-EM) structures have been published for several assembly intermediates of the yeast spliceosome including the B, Bact, C, C* and intron-lariat complexes, providing completely new insight into the complex structure of the yeast spliceosome and its structural dynamics during the catalytic cycle. As of now, only the human C* complex has been investigated by cryo EM. In this work, I have used cryo-EM to investigate the 3D structure of the human Bact complex. Human and yeast activated spliceosomes share a large number of conserved proteins but differ in their protein composition in several aspects. Human Bact contains numerous proteins that are absent in yeast, including numerous peptidyl-prolyl isomerases (PPIases) and the RNA helicase Aquarius (Aqr), which is required for catalytic activation of the human spliceosome in addition to Prp2. On the other hand, proteins conserved between yeast and human are missing from the human Bact spliceosome, raising the possibility that the 3D structure of the human and yeast Bact complexes may differ to some extent. The Bact complex was assembled in HeLa nuclear extracts using a pre mRNA construct, PM5 10, which contained the 5’ exon and an intron that is truncated 10 nucleotides (nts) after the BS, and was affinity-purified for cryo EM analysis. After exhaustive 3D multi-reference refinement (3D classification) of the human Bact particles, two major forms, termed A and B, of the Bact structure were obtained at resolutions of 5.3 Å and 8.1 Å, respectively. While their overall structure is largely similar, forms A and B differ with respect to the presence/absence of several densities, as described below. The structure of the central domain of the human Bact complex, including the catalytic U2/U6 RNP core, is highly conserved between the human and yeast spliceosomes. At the catalytic U2/U6 RNA centre, density for catalytic metal ion 2 (M2) is present but not for M1, indicating that the catalytic centre in this state is not yet functionally active. The 5’ end of the intron is engaged in base-pair interactions with the U6 ACAGA box and an additional ca. 10 nts of U6 snRNA, a distinctive feature not present in yeast spliceosomes. The 5’ss is positioned close to the catalytic centre but the 5’ terminal GU nts of the intron are engaged in protein interactions with Rnf113A (hCwc24) and probably also with U2 SF3A2. The BS forms an extended helix with U2 snRNA, which is clamped between the terminal HEAT repeats of the toroidal HEAT domain of SF3B1. The BS adenosine (BS-A) is occluded in a protein pocket, comprised of C terminal HEAT repeats and the PHF5A (hRds3) and is spatially separated from the catalytic centre by ca. 5 nm. The hPrp2 RNA helicase is bound to the convex side of SF3B1’s HEAT domain, close to the site, where the 3’ end of the intron exits the HEAT domain, but spatially separated from the U2/BS helix by ca. 7 nm. Thus, in a similar way to the scenario proposed for yeast Bact spliceosomes, the ATP dependent Prp2 mediated remodelling may lead to conformational changes in SF3B1’s HEAT domain that liberates the first-step reactants for catalysis. In the human Bact structure, the U2/U6 helix II adopts a significantly different conformation, when compared with the yeast Bact structure and other spliceosomal assembly intermediates including human C*. Moreover, the conserved Syf2 protein, which binds to the base of U2/U6 helix II in yeast Bact, is absent from human Bact, probably because in the latter the U2/U6 helix II is sandwiched between two other proteins. This raises the interesting possibility that in humans the catalytic activation is more complex than in yeast and requires remodelling of U2/U6 helix II and concomitant integration of hSyf2 as a pre requisite for B* complex formation. Another distinguishing feature of the human Bact structure is the existence of an intricate protein protein interaction network that connects the complex of the Aqr helicase and protein Xab2 (hSyf1) to the main body of the spliceosome. Interestingly, all four PPIases present in human Bact are involved in protein protein interactions, indicating that one of their functions appears to be that of serving as bridges between various protein modules of the spliceosome. Finally, while forms A and B of the human Bact structure obtained by 3D classification share most of the structural features described above, they differ in respect of the presence/absence of densities for several protein domains. The most dramatic difference is the absence of density for the large elongated Prp19 helical bundle in form A, while its well defined density is present in form B of the Bact structure. At the same time, density for Ppil1 appears in form B, forming a bridge between the centre of the helical bundle and the central body of Bact. Moreover, the position of the U5 40K WD40 domain also differs in form A and B. Experimental evidence further indicates that the absence of a protein density in one of the forms is not due to the physical absence of the respective proteins, but instead indicates conformational flexibility of the protein domains. Further evidence indicates that form A is a precursor of form B of the Bact structure. Thus, it was possible to capture by cryo EM analysis two conformational states of the human Bact complex that differ in their degree of conformational maturation towards the catalytic activation step. As it is unlikely, that an ATP requiring step is involved in the transition of the Bact complex from form A to B, this suggests that the observed conformational changes of the various protein domains are facilitated by the thermal energy of the system