Spliceosome complexes remove non-coding sequences from RNA transcripts in two steps. A structure of a spliceosome after the first step reveals active-site interactions and evolutionary constraints on these non-coding regions. See Article p.197
The presence of non-coding sequences called introns in nascent RNA transcripts is a defining characteristic of the genomes of eukaryotic organisms, which include plants, animals and fungi. Intron removal by a spliceosome complex is an essential step in gene expression and regulation. Decades of biochemical and genetic studies have provided detailed insights into the composition of these complexes and the RNA structures within them. However, the dynamic nature of the complexes has hindered efforts at modelling and structure determination at atomic resolution. On page 197, Galej et al.1 present a structure of the catalytic spliceosome at 3.8 ångströms resolution, obtained using single-particle cryo-electron microscopy (cryo-EM). This structure not only provides evidence in support of reported interactions2 that bind and position catalytic metal ions, but also reveals previously unknown molecular features of splicing catalysis.
The spliceosome is a large, dynamic RNA–protein complex that catalyses intron removal in two sequential chemical reactions (Fig. 1). The chemical mechanism of intron removal, as well as the core spliceosomal RNAs and proteins, are highly evolutionarily conserved in most eukaryotes. The first reaction cleaves the nascent transcript at the 5′ end of the intron (the 5′ splice site; 5′SS), causing the intron to form a lasso-shaped, or 'lariat', structure. Compositional and structural changes in the spliceosome then occur, whereupon the second reaction joins together (ligates) the coding exon sequences that flank the intron, simultaneously generating the mature messenger RNA and excising the intron lariat.
Technological and computational advances3 in cryo-EM have led to the structural determination of many spliceosomal complexes within the past year4,5,6,7,8,9. To obtain particles for their study, Galej and co-workers assembled spliceosomes in vitro on RNA substrates that can proceed through the first catalytic step, but not the second one. They then purified the resulting complexes using proteins that interact with the spliceosome only after the first step has occurred. The reported structure (Fig. 2) therefore represents complexes that form immediately after the first catalytic step. This, along with another recently published structure9, is the most relevant structure to splicing catalysis available.
Spliceosomal complexes follow an intricate pathway during assembly, catalysis and recycling, characterized by compositional and structural changes (Fig. 1). Enzymes known as ATPases facilitate many of the transitions between spliceosome complexes. The ATPase Prp2 remodels the fully assembled but inactive complex (Bact) to form the catalytically active complex (C). Another ATPase, Prp16, removes the intron lariat from the active site after the first reaction, and positions the 3′SS near to the 5′SS to allow exon ligation. Once the second reaction has excised the intron, the ATPase Prp22 binds the 3′ exon and moves along the mRNA, thus releasing the mRNA from the spliceosome.
Shi and colleagues8 recently reported a structure in which Prp2 is bound in the Bact form of the spliceosome, whereas, in Galej and co-workers' structure, Prp2 has been replaced by Prp16 in the C complex. These structures are the first visualizations of these two ATPases bound to spliceosomes. Both enzymes are positioned similarly in the overall topology of the spliceosome near the 3′ end of the intron. On the basis of the interactions between Prp16 and the spliceosome observed in their structure, Galej et al. suggest that splicing factors unique to each complex recruit the specific ATPase needed (see Fig. 6 of the paper1). Hydrolysis of ATP molecules by the ATPases could subsequently destabilize the associated splicing factors, allowing the RNA structures in the catalytic core to be remodelled.
The two substrates for the first catalytic step are the 5′SS and an adenosine nucleotide, known as the branch site, within the intron. Although it has long been thought that these two substrates almost certainly interact with each other, to help bring them together as needed for the catalytic step, neither evidence nor models for such an interaction existed. Galej and co-workers' structure reveals intimate interactions between the 5′SS (specifically, its GU sequence, which consists of a guanosine nucleotide next to a uridine nucleotide) and the sequence flanking the branch site; these interactions help to explain the evolutionary conservation of the two sequences. For example, an RNA base triple (a structure analogous to a base pair, but involving three bases) was identified between the uridine of the 5′SS-GU and the helix created by base pairing between the intron sequence flanking the branch site and U2, one of the small nuclear RNAs that forms the spliceosome's active site.This base triple helps to position the 5′SS near the branch-site adenosine, as required for the first catalytic step.
By contrast, in Shi and colleagues' structure8, the 5′SS and branch site are separated by a large distance (approximately 49 Å). The guanosine of the 5′SS-GU is protected by a pocket formed by a protein subunit of the spliceosome and a first-step splicing factor. Analogously, the branch-site adenosine is positioned in a positively charged pocket of another protein subunit (SF3B1, which is highly mutated in human cancers10). These two pockets protect the reactive groups involved in the first catalytic step until the spliceosome has transitioned to a catalytically active conformation.
Galej and colleagues' structure also helps to explain the evolutionary sequence conservation of the branch site–U2 duplex by revealing another base triple interaction between the branch-site adenosine and the intron–U2 RNA helix two nucleotides away. This was presaged in part by interactions observed between the branch-site adenosine and the intron–U2 RNA helix in an RNA-only structure11 previously determined by nuclear magnetic resonance spectroscopy. This base triple positions the reactive hydroxyl group of the branch-site adenosine outward towards the 5′SS.
The structural insights obtained through the identification of hundreds of RNA–protein and protein–protein interactions in the new structures1,8,9 suggest innumerable biochemical and genetic experiments to ascertain which splicing step these interactions contribute most to, and for what intron features they are most important. The stage is now set for the exploration and discovery of many other spliceosome structures. Like the explosion of successes that followed the determination of the ribosome structure12 (the protein-synthesis apparatus), we eagerly await structures not just for normal spliceosome complexes, but also for complexes that include mutations in pre-mRNA substrates or in spliceosomal components, such as those found in many cancers10. The future will allow a more comprehensive picture of the basic mechanisms of splicing catalysis, and of how splice sites are recognized and catalysis is regulated. Other achievements may also include the determination of features vital to the alternative splicing regulation found in complex organisms.Footnote 1
Galej, W. P. et al. Nature 537, 197–201 (2016).
Fica, S. M. et al. Nature 503, 229–234 (2013).
Nogales, E. Nature Methods 13, 24–27 (2016).
Agafonov, D. E. et al. Science 351, 1416–1420 (2016).
Nguyen, T. H. et al. Nature 530, 298–302 (2016).
Wan, R. et al. Science 351, 466–475 (2016).
Yan, C. et al. Science 349, 1182–1191 (2015).
Yan, C., Wan, R., Bai, R., Huang, G. & Shi, Y. Science http://dx.doi/org/10.1126/science.aag0291 (2016).
Wan, R., Yan, C., Bai, R., Huang, G. & Shi, Y. Science http://dx.doi.org/10.1126/science.aag2235 (2016).
Yoshida, K. et al. Nature 478, 64–69 (2011).
Newby, M. I. & Greenbaum, N. L. Nature Struct. Mol. Biol. 9, 958–965 (2002).
Schmeing, T. M. & Ramakrishnan, V. Nature 461, 1234–1242 (2009).