Structural insights into intron catalysis and dynamics during splicing

The group II intron ribonucleoprotein is an archetypal splicing system with numerous mechanistic parallels to the spliceosome, including excision of lariat introns1,2. Despite the importance of branching in RNA metabolism, structural understanding of this process has remained elusive. Here we present a comprehensive analysis of three single-particle cryogenic electron microscopy structures captured along the splicing pathway. They reveal the network of molecular interactions that specifies the branchpoint adenosine and positions key functional groups to catalyse lariat formation and coordinate exon ligation. The structures also reveal conformational rearrangements of the branch helix and the mechanism of splice site exchange that facilitate the transition from branching to ligation. These findings shed light on the evolution of splicing and highlight the conservation of structural components, catalytic mechanism and dynamical strategies retained through time in premessenger RNA splicing machines.

In this manuscript the authors present three novel, high resolution cryo-EM structures of a pre-mRNA group II intron bound by the maturase protein at different stages of the splicing reaction, namely at the pre-branching, pre-ligation and post-ligation stages.The pre-branching structure reveals for the first time how domain 1 of the group II intron, together with the maturase protein, align the branch helix (D6) for step 1 of the splicing reaction in which a branched intron intermediate is formed.It elucidates the molecular interactions involving D1 nucleotides and residues of the thumb/DBD domains of the maturase protein that mediate positioning of D6.The authors carry out mutational analyses that support the functional importance of these nucleotides/amino acids, as they are shown to lead to branching defects.This structure also reveals a role for the 5'ss in docking D6 and bringing the branchpoint close to the 5'ss, which is a prerequisite for the subsequent branching reaction.Finally, the pre-branching structure elucidates the structural basis for recognition of the branchpoint adenosine.Comparisons with branchpoint recognition in the spliceosome intriguingly show that both splicing systems employ the same molecular recognition pattern, revealing a novel mechanistic parallel between the spliceosome and group II introns.The authors next compare the cryo-EM structures of pre-branching and pre-and post-ligation complexes, and thereby uncover the local conformational dynamics of the branch point adenosine during splicing.In addition, they show that D6 undergoes a major structural rearrangement after branching, swinging downward by ca 90 degrees.This repositioning removes the 5'ss and branched adenosine from the active site, and thereby allows binding instead of the 3'ss, which is a prerequisite for exon ligation.Comparisons with the published cryo-EM structures of B*, C and C* spliceosomal complexes intriguingly demonstrate an analogous movement of the U2/branch helix, which also allows docking of the 3'ss into the spliceosome's active site.
The cryo-EM structures presented in this manuscript also allow direct visualization of the two-metal ion mechanism for group II intron branching and confirm that the same active site with the same catalytic ion conformation is also used for both branching and exon ligation.They thus reveal that both the group II intron splicing machinery and the spliceosome swap the bound splice sites between the first and second catalytic steps without disrupting the catalytic core.
Finally, structural comparisons with the yeast spliceosome also reveal some intriguing differences in the roles of proteins involved in branching, and indicate that the multiple functions of the maturase protein in group II intron branching, have been fragmented and are carried out by multiple proteins (i.e., PRP8 and YJU2) in the spliceosome, which allows for a more highly regulated process.
In summary, this manuscript presents a plethora of new information about the molecular mechanisms of group II intron splicing and, in combination with previously reported group II structures, allows a complete structural and mechanistic view throughout the group II intron splicing lifecycle.Furthermore, the structural analyses presented in this manuscript provide for the first time molecular insights into the sequential structural rearrangements that are required for the branching reaction during group II splicing, as well as and for the selection of the splice sites for subsequent exon ligation.They also elucidate additional, previously unknown molecular recognition mechanisms and functionally significant structural dynamics that are evolutionarily conserved between the group II intron splicing machinery and the spliceosome, thereby strengthening the conclusion that the spliceosome has evolved from pre-mRNA group II introns.

Referee #2:
The manuscript, "Structural insights into intron catalysis and dynamics during splicing", by Xu et al. represents a significant advance in understanding the catalytic mechanism of a group II intron, particularly the new cryo-EM structure of an intron maturase complex captured just prior to the branching reaction.The structures of the same complex remodeled for exon ligation and a product complex mimic enable the authors to model the atomic interactions and rearrangements that accompany the splicing reaction.They describe structural parallels with the same reaction in pre-mRNA splicing by the spliceosome, which further reinforces support for a common evolutionary splicing ancestor.Beyond the group II intron research field, the work will be of broad interest to scientists interested in splicing mechanism, evolution of splicing, RNA structure, and RNA/protein interactions.I have no major concerns with the data and its interpretation.Minor concerns can be addressed by simple revision of some text and figures.

Specific critiques:
1.The abbreviations Pre-1F, Pre-2F and Post-2F are not intuitive.Why "F"? Avoiding the alphabet soup of the spliceosome complexes with names like, for example, "pre-branching state" or "preligation complex" would be most helpful to readers.
2. The authors should take care to emphasize that current group II introns and the spliceosome are thought to share a common ancestor molecule.Although that molecule likely more closely resembles an autocatalyzing intron, statements like "living fossil" confuses the evolutionary trajectory.
3. While the ∆G86 and ∆C601 data in Fig. 2d supports the importance of both residues for branching, the argument that the G86A/C601U mutation demonstrates the importance of base pairing is overreaching.Deletion of nucleotides may have other structural effects relative to simply changing the base.
4. The large rearrangement of helix D6 following branching is striking.While Fig. 3 and the supplemental movies help to visualize the changes, it would also be nice to have a movie that highlights the interactions that must be broken to achieve the rearrangement.Can the authors further speculate what might drive this arrangement in relation to the correlating change in the spliceosome? 5.All yellow labels in figures need to be changed to black or something with enough contrast to be legible.
6.When comparing the group II intron to spliceosome structures, it would be helpful if the position of intron nucleotides relative to the branch point adenosine could also be indicated.The spliceosome intron nucleotide numbering does not make this comparison intuitive.There are also labeling errors in Ext.Data Fig. 5.5 for both U2 snRNA and the intron.Labeling the snRNA outright will help with confusion between the second U of the intron and the snRNA.This manuscript reports cryoEM structures of a group IIC intron in complex with the associated maturase protein.Group II introns are ancestral prototypes for spliceosomes, offering a fundamental model for understanding RNA splicing and the evolutionary progression of spliceosomes from retroelements.Prior endeavours to capture Group II intron complexes that elucidate the branching pathway had proven ineffective.Additionally, the role of the maturase in facilitating branching remained elusive.Both of these aspects are of paramount importance, as branching is evolutionary conserved, and the maturase stands as a vestige of the spliceosome protein PRP8.
Here, changes in the preparative biochemistry, particularly the incubation with the maturase that promotes branching, and the replacement of Mg2+ with Ca2+ enabled stalling precursor and branching intermediates for cryoEM reconstruction.

Key findings include:
The mechanism by which the maturase favors branching instead of hydrolysis, by holding D6 in an orientation that brings the branch point adenosine (bpA) to the catalytic center.
The branching mechanism involves recognizing and positioning the bpA, and using a lysine finger (K361), which binds the 5'SS to juxtapose the nucleophile with the scissile phosphate.
The substrates-exchange mechanism, by a large-scale swinging of D6 that pulls the lariat and replaces it with the 3'SS.
Novel lines of evidence for the common ancestry of spliceosomes and group II introns, with respect to the dynamics of RNA-RNA and RNA-protein contacts during branching.Some findings are entirely unexpected, while others elucidate mechanisms proposed decades ago.The data are of high quality, and the findings hold significance for a wide readership interested in RNA metabolism and evolutionary biology.Considering these merits, I recommend the publication of this work in Nature after addressing the specific points outlined below.

Specific comments
1.The mechanistic comparison shows that the UP and DOWN positions of the D6 intron domain have equivalents in the branch duplex positions in B* and C complexes (Fig. 5b,c).The D6 domain is kept in the UP position by extensive interactions with D1 and the maturase domains (Fig. 2).Which spliceosome components interact with the branch duplex in B* and C complexes to stabilize the UP and DOWN positions, other than the Thumb-DBD of the maturase?An analysis might extend the insight into the evolution of spliceosomes, where the absence of D1 can be compensated by stagespecific proteins and PRP8 domains absent in the maturase.
2. Is it conceivable that K361 of the maturase complements the catalytic site by neutralizing charges developed during the phosphoryl transfer, upon branching?Can minor local rearrangements possibly place K361 in the appropriate distances and geometry to influence catalysis?3. The structure-guided mutations in the maturase have a drastic effect on lariat formation (Fig. 2f, lanes 2-4).It is important to check experimentally that the mutations do not disrupt the native fold of the maturase.4. Fig. 1d requires a more detailed description, possibly in the results section or the legend.As the maturase is present in lanes 3-5, the difference between the three samples is unclear.I understand the post-ligation RNP from lane 4 contains an oligonucleotide equivalent to the ligated exon.Are there any other differences?Besides, an additional band below the precursor is not indicated.5.The gels from Fig. 2d,f look rather different than 1d, although both variants denote splicing gels.The linear fragment and the linear-3'exon are not indicated in 1d. 6. Fig. 2a looks overcrowded in the central area.Perhaps it will be easier to discern the various elements by removing the density (or by other means).The bpA should be indicated for the sake of orientation.I suggest indicating how the detailed views from Fig. 2b,e relate to the overview from 2a.
7. Labeling the domains in the figures that show details of interacting residues, such as Figs 2b,e, 3ac, facilitates the analysis (e.g., D6, exon, U2 snRNA, etc).10.To better appreciate the maturase location and impact on the RNA structure, I would indicate its location (possibly as a shadow) in Fig. 3f or a similar figure from the Ext.Data.11.A comparative view between pre-1F and the spliceosome, as in Fig. 5a bottom is required, especially since the parallels between the branching in the two splicing systems are a highlight of the paper.
12. The branch helix is not labeled in Fig. 5 and Ext.Data Fig. 8.Only the U2 snRNA is indicated, leading to the potential confusion that lone U2 is the equivalent of D6.

7.
The resolution of Ext.Data Figs 2 and 3 needs to be improved.Currently, none of the graph labels in panels b-d are legible.Referee #3: 8. A figure showing clearly the contacts between D6 and the maturase contacts relative to the bpA (A632) would be helpful.The ED Fig.8ais useful, but insufficient.9. Label the conformational states in Fig 3e.
Fig. 3d might require optimization for conveying a message, as the indicated changes are hard to follow.