Research Article

Structure of a yeast step II catalytically activated spliceosome

See allHide authors and affiliations

Science  13 Jan 2017:
Vol. 355, Issue 6321, pp. 149-155
DOI: 10.1126/science.aak9979

Poised for the second step of splicing

In eukaryotes, noncoding sequences in transcribed precursor mRNA are cut out by a dynamic macromolecular machine, the spliceosome. This involves two sequential reactions. The first cuts one end of the noncoding intron and loops it back on itself to form an intron lariat, and the next excises the intron and ligates the coding mRNA. Insights into the first step of splicing have come from the structures of two intermediates: the Bact complex, which is primed for catalysis, and the C complex, which is formed after the first splicing reaction. Yan et al. now report a high-resolution structure of the step II catalytically activated spliceosome (the C* complex). This structure shows conformational changes that position catalytic motifs to accomplish the second splicing reaction.

Science, this issue p. 149

Abstract

Each cycle of precursor messenger RNA (pre-mRNA) splicing comprises two sequential reactions, first freeing the 5′ exon and generating an intron lariat–3′ exon and then ligating the two exons and releasing the intron lariat. The second reaction is executed by the step II catalytically activated spliceosome (known as the C* complex). Here, we present the cryo–electron microscopy structure of a C* complex from Saccharomyces cerevisiae at an average resolution of 4.0 angstroms. Compared with the preceding spliceosomal complex (C complex), the lariat junction has been translocated by 15 to 20 angstroms to vacate space for the incoming 3′-exon sequences. The step I splicing factors Cwc25 and Yju2 have been dissociated from the active site. Two catalytic motifs from Prp8 (the 1585 loop and the β finger of the ribonuclease H–like domain), along with the step II splicing factors Prp17 and Prp18 and other surrounding proteins, are poised to assist the second transesterification. These structural features, together with those reported for other spliceosomal complexes, yield a near-complete mechanistic picture on the splicing cycle.

The removal of an intron and joining of two neighboring exons in a precursor messenger RNA (pre-mRNA) are executed in two sequential steps, each involving an SN2-type transesterification reaction (13). In the first step, the ribose 2′-OH of an adenine nucleotide in the intron branch-point sequence (BPS) initiates a nucleophilic attack on the phosphate group at the 5′ end of the 5′ splice site (5′SS), freeing the 5′ exon and generating an intron lariat–3′ exon intermediate. In the second step, the ribose 3′-OH at the 3′ end of the 5′ exon executes a nucleophilic attack on the phosphate group at the 5′ end of the 3′ exon, ligating the two exons (1).

These reactions are catalyzed by a dynamic spliceosome that exists in seven distinct structural states: the B, Bact, B*, C, C*, P, and ILS (intron lariat spliceosome) complexes (4, 5). Except for the precatalytic B complex, these spliceosomes share a stable structural core, which consists of U5 small nuclear ribonucleoprotein particle (snRNP), U6 small nuclear RNA (snRNA), the 5′-end portion of U2 snRNA, and about a dozen proteins of the nineteen complex (NTC) and NTC-related complex (NTR) (611). Interconversion of these spliceosomal complexes is driven by eight conserved DExD/H-box adenosine triphosphatases (ATPases)/helicases (12). In the B to Bact transition, the U1 and U4 snRNPs are dissociated by the ATPase Brr2 (1316), and the NTC and NTR are recruited to help formation of an active site. The DEAH-box ATPase Prp2 converts the Bact complex to the catalytically activated B* complex where the first-step reaction occurs (1719), resulting in the formation of the catalytic step I spliceosome (the C complex). Another DEAH-box ATPase Prp16 catalyzes the conversion of the C complex to the step II catalytically activated C* complex where the second-step transesterification takes place (18, 20, 21). The ligated exon in the postcatalytic P complex is released through the action of the DEAH-box ATPase Prp22 (22), generating the ILS complex.

The spliceosome is a protein-directed metalloribozyme (6, 7, 2325). The splicing active site, comprising the intramolecular stem loop (ISL) of U6 snRNA and the associated magnesium (Mg2+) ions, loop I of U5 snRNA, and helix I of the U2/U6 duplex, is nestled in a catalytic cavity in the central component Prp8 (6, 7). Conserved nucleotides in the ISL coordinate Mg2+ ions (3, 6, 7, 26). Throughout both catalytic steps, nucleotides at the 3′ end of the 5′ exon are recognized by U5 loop I, and the BPS and 5′SS are anchored by U2 and U6 snRNA, respectively. The conformation and the atomic details of the spliceosomal active site were first revealed by the cryo–electron microscopy (cryo-EM) structure of the ILS complex from Schizosaccharomyces pombe at 3.6 Å (6, 7). The overall conformations of the active site and U5 and U6 snRNAs from the S. pombe ILS complex (6, 7) are exactly preserved in both the Bact complex (9, 10) and the C complex (8, 11) from Saccharomyces cerevisiae.

Mechanistic insights into the first-step splicing reaction have been revealed by the structures of the Bact and C complexes (811). Here, we report the structure of the C* complex at an average resolution of 4.0 Å, which unveils features of the spliceosome just preceding the second-step transesterification.

Overall structure

As previously reported (9), the NTC component Cef1 was affinity-tagged for the purification of a variety of spliceosomal complexes from S. cerevisiae (fig. S1). The sample was imaged on a Titan Krios microscope. A total of 12,142 micrographs yielded 761,767 particles (9). These particles were subjected to two-dimensional (2D) and 3D classifications, resulting in the reconstruction of the Bact complex at 3.5 Å resolution by 84,486 particles (9) and the C complex at 3.4 Å by 161,066 particles (8). We sorted the remaining particles using a similar strategy (figs. S2 and S3) and determined the structure of a step II catalytically activated spliceosome (C* complex) at an average resolution of 4.0 Å on the basis of the gold-standard Fourier shell correlation criteria (figs. S4 and S5 and tables S1 to S4). The final reconstruction used 27,558 particles. The EM density maps in the core regions of the C* complex display an actual resolution of about 3.5 Å (figs. S6 and S7). The maps are mostly at 3.5 to 4.5 Å in the central body, 5 to 10 Å in the surrounding regions, and 10 to 30 Å in the periphery (table S2). The quality of the maps, together with information on the C complex (8), allow atomic modeling in the central body and docking of known structures in the peripheral regions.

The atomic model of the C* complex from S. cerevisiae contains 9565 amino acids from 35 proteins and 362 RNA nucleotides (nt), with a combined molecular mass of ~1.1 MDa (Fig. 1A and tables S1 to S4). Of the 362 nucleotides, 311 are assigned to three snRNAs (U2, U5, and U6) and 51 to the intron lariat and the 5′ exon. Among the modeled amino acids, 6838 have side chains and the rest were built as poly-Ala. Different from the C complex (8), the NTC component Isy1 and the step I splicing factors Cwc25 and Yju2 are no longer present in the active site of the C* complex; instead, the step II splicing factors Prp17 and Prp18 are bound (Fig. 1B). In addition, an ATPase/helicase (Prp16 or Prp22) interacts with the Linker and reverse transcriptase (RT) Finger/Palm domains of Prp8; the limited resolution at the peripheral region precludes its conclusive identification.

Fig. 1 Structure of the S. cerevisiae spliceosomal C* complex.

(A) A cartoon representative of the S. cerevisiae C* complex at an average resolution of 4.0 Å. Three views are shown. The protein and RNA components are color-coded and tabulated below the structural images. The structure shown here includes 35 proteins, three snRNAs, a free 5′ exon, and an intron lariat, with a combined molecular weight of ~1.1 MDa. (B) Three close-up views of the protein and RNA components at and surrounding the active site of the S. cerevisiae C* complex. All structural images were created using PyMol (47).

Similar to the C complex (8, 11), the C* complex has an irregular appearance, with the NTC components Syf1/Clf1/Prp19/Snt309/Cef1 arching over the central body (Fig. 1A). Within the central body, Prp8, Snu114, three proteins from NTC (Cef1, Clf1, and Syf2), six components from NTR (Bud31, Cwc2, Cwc15, Ecm2, Prp45, and Prp46), and four splicing factors (Cwc21, Cwc22, Prp17, and Prp18) are organized around the active site (Fig. 1B). The 5′ end of the 5′ exon and the 3′ end of the linear portion of the intron lariat emanate from the active-site assembly. Similar to the Bact and the C complexes (811), an extended sequence from the splicing factor Cwc21 stabilizes the 5′ exon through direct interactions; these interactions are buttressed by the Switch loop of Prp8 and the C-terminal MA3 domain of Cwc22. The MA3 domain of Cwc22 interacts with the Linker domain of Prp8, whereas the N-terminal MIF4G domain of Cwc22 simultaneously contacts Snu114 and the extended loop from the Prp8 N-domain that forms a lasso over Snu114 (6) (Fig. 1B).

The RNA elements

The EM density maps for the RNA elements (fig. S7), together with information on the C complex (8, 11), allowed us to build an atomic model for U2 snRNA (nt 1 to 52, 110 to 123, and 1096 to 1120), U5 snRNA (nt 28 to 55, 60 to 127, and 163 to 183), U6 snRNA (nt 1 to 103), a free 5′ exon, and an intron lariat (Fig. 2A). Among the 51 modeled nucleotides of pre-mRNA, 13 are attributed to the 5′ exon and 38 to the intron lariat. Similar to that in the C complex (8, 11), the 5′ exon remains anchored to loop I of U5 snRNA (fig. S7F), and the lariat junction is held in place through duplex formation of 5′SS with U6 snRNA and BPS with U2 snRNA (Fig. 2, A and B, and fig. S7G). The observed base-pairing interactions involving the snRNAs are consistent with biochemical characterizations (2735).

Fig. 2 Arrangement of the RNA elements in the S. cerevisiae C* complex.

(A) An overall cartoon representation of the RNA map. The catalytic center comprises the ISL of U6 snRNA (green) and the associated Mg2+ ions (red spheres), loop I of U5 snRNA (orange), and helix I of the U2/U6 duplex (marine/green). Preceding the step II reaction, the 5′ exon is anchored to loop I, and the lariat junction has moved away from the 3′ end of the 5′ exon to vacate space for 3′SS–3′-exon. The disordered RNA sequences are represented by dotted lines. (B) Structural comparison of the overall RNA map between the C and C* complexes. The RNA elements in the C* complex are color-coded as shown, and those in the C complex are in gray. (C) The overall structures of U5 and U6 snRNAs in the C* complex are identical to those in the C complex. (D) Structural comparison of U2 snRNA between the C and C* complexes. Nucleotides 1 to 30 of U2 snRNA in the C* complex are structurally similar to those in the C complex; other nucleotides display quite different conformations. (E) Structural comparison of the lariat junction and the 5′ exon between the C and C* complexes. The 5′ exon and 13 contiguous intron nucleotides downstream of the third nucleotide of 5′SS (GUAUGU) are structurally indistinguishable between the C and C* complexes.

The structures of U5 and U6 snRNA, along with nucleotides 1 to 30 of U2 snRNA that form duplexes with U6 snRNA (i.e., helices I and II), remain nearly identical between the C and C* complexes (Fig. 2, C and D). The 5′ exon and a stretch of 13 contiguous intron nucleotides, which begins at the third nucleotide of the 5′SS (GUAUGU), are structurally indistinguishable between the C and C* complexes (Fig. 2E). In sharp contrast, the majority of U2 snRNA sequences exhibit drastic structural changes (Fig. 2D). The lariat junction in the C* complex has been translocated by ~15 to 20 Å compared with that in the C complex (Fig. 2E). The phosphate group of the guanine nucleotide at the 5′ end of the 5′SS in the C* complex is 19 Å away from that in the C complex. The intron sequences just upstream of BPS undergo even more drastic movements, with these nucleotides flipping as a whole by as much as 55 to 65 Å. These movements are likely indispensable for accommodation of the 3′ exon and the preceding intron 3′SS into the active site.

The splicing active site

The detailed structural features observed in the C complex (8) are mostly preserved in the C* complex, including the approximate locations of five metal ions in the active site (Fig. 3A). Two of the five metals have been interpreted as catalytic Mg2+ ions M1 and M2 (Fig. 3B). During the first transesterification, M1 stabilizes the leaving group—the 3′-OH of the 3′-end nucleotide in the 5′ exon, whereas M2 activates the nucleophile —the 2′-OH of an invariant adenine nucleotide in the BPS (3). The roles of M1 and M2 reciprocate in the second transesterification, with M1 activating the nucleophile (3′-OH of the 3′-end nucleotide in the 5′ exon) and M2 stabilizing the leaving group (5′-end phosphate group of the 3′ exon) (3).

Fig. 3 The active site of the S. cerevisiae spliceosomal C* complex.

(A) Structure of the RNA elements in the catalytic center of the spliceosomal C* complex. A comparison of the Mg2+ ions between the C complex (blue spheres) and the C* complex (red spheres) is shown in the right panel. (B) A close-up view on the coordination of the catalytic Mg2+ ions by the ISL of U6 snRNA. M1 is bound by three ligands: the phosphate groups of G78 and U80 from U6 snRNA and 3′-OH of the 3′-end nucleotide in the 5′ exon. M2 is similarly coordinated by two ligands (the phosphate groups of A59 and U80 of U6 snRNA), as in the C complex (8). (C) Proposed accommodation of the 3′ exon in the active site of the C* complex. The location of the 3′ exon (colored gray) is suggested by a continuous stripe of density (fig. S8).

M2 coordination in the C* complex is the same as that in the C complex (8). Because of the translocation of the lariat junction, M1 is only coordinated by three ligands in the C* complex—the phosphates of G78 and U80 from U6 snRNA and the 3′-OH of the 3′-end nucleotide in the 5′ exon—but separated from the fourth ligand in the C complex (the phosphate group of the 5′-end nucleotide in the 5′SS) by a distance of about 21 Å (Fig. 3B and fig. S7D). The vacated space is now available for occupation by the 3′SS–3′ exon, and the nucleophile—the 3′-OH of the 3′-end nucleotide in the 5′ exon—is already activated for the second transesterification. In our cryo-EM map, a nearly continuous stripe of density suggests how the 3′SS–3′-exon sequences may be positioned into the active site for the second-step reaction (Fig. 3C and fig. S8). Unfortunately, the quality of the density disallows RNA nucleotide modeling.

Proteins at the active site

In the C* complex, at least 15 protein components associate with each other and with the RNA elements at or surrounding the active site. Except for the splicing factors Prp17 and Prp18, most of these proteins play a similar or identical role compared with that in the C complex (8). For example, the NTR component Cwc2 recognizes both the intron and U6 snRNA while simultaneously interacting with Ecm2 and Bud31 (Fig. 1B). The NTC component Syf2 closely recognizes the U2/U6 helix II and associates with Prp45 and Clf1.

The step I splicing factors Cwc25 and Yju2 and the NTC component Isy1, which stabilize the splicing active site in the C complex (8), are absent from the active site of the C* complex (Fig. 4A and fig. S9A). In contrast, the step II splicing factor Prp18, absent in the C complex (8), is positioned close to the active site through direct interactions with the ribonuclease H (RNase H)–like domain of Prp8 in the C* complex. Although the RNase H–like domain remains associated with the intron/U2 snRNA duplex in the C to C* transition, its location is changed as a result of spliceosomal remodeling. Similarly, the WD40 domain of Prp17, which is located between the BPS/U2 snRNA duplex and the 5′SS/U6 snRNA duplex, has been translocated by 40 to 80 Å compared with that in the C complex (8) (fig. S9A). Notably, an α-helix and the ensuing loop at the N terminus of Prp17 (residues 51 to 73), which together with Bud31 anchors the 5′ stem loop of U6 snRNA, remain unchanged between the C and C* complexes.

Fig. 4 Protein components at the active site of the C* complex.

(A) The step II splicing factor Prp18 interacts with the RNase H–like domain of Prp8 near the active site of the C* complex. The orientation of Prp18 relative to the RNA elements is shown in the left panel; the relative positioning among Prp18, Prp17, and the RNase H–like domain of Prp8 is illustrated in the middle and right panels. (B) A close-up view on the active site of the C* complex. The 1585 loop (red) is positioned close to the ISL of U6 snRNA and the 3′ end of the 5′ exon. The β finger from the RNase H–like domain of Prp8 is inserted between the 5′SS/U6 and BPS/U2 duplexes right next to the lariat junction. (C) A close-up view on the relative position of the 1585 loop and the β finger of Prp8. This view is related to that in (B) by a 90° rotation around a horizontal axis. (D) Proposed accommodation of the 3′SS–3′ exon sequence. The 3′SS–3′ exon (labeled simply as 3′ exon in the image) is proposed to be sandwiched between the 5′ exon and the 1585 loop in the active site of the C* complex. This view is identical to that in (B). (E) Proposed accommodation of the 3′SS–3′ exon. This view is identical to that in (C). The labeled residues in the β finger are those whose mutations inhibit the step II reaction. A, Ala; T, Thr; V, Val.

Catalytic motifs for step II reaction

Two prominent motifs of Prp8 are located in the active site of the C* complex and likely play a critical role in catalyzing the step II reaction. One motif is the 1585 loop (residues 1585 to 1598 of Prp8) (36), which inserts into the center of the active site and simultaneously interacts with the ISL and the lariat junction (Fig. 4, B and C). The tip of the 1585 loop is positioned close to the 3′ end of the 5′ exon and the catalytic Mg2+ ions. The 3′ exon and the preceding 3′SS are likely placed between the 1585 loop and the 5′ exon (Fig. 4, D and E). This analysis suggests an important role for the 1585 loop in stabilizing the 3′SS–3′ exon for step II transesterification.

The other catalytic motif is a protruding β hairpin, also known as the β finger (37), from the Prp8 RNase H–like domain (Fig. 4, B and C). In previous structures of the spliceosome, the RNase H–like domain has been recognized to associate with a mobile RNA element during the splicing cycle (8, 9). In the C* complex, the β finger interacts with Prp17 and is located right above the lariat junction between the 5′SS/U6 snRNA duplex and the BPS/U2 snRNA duplex (Fig. 4, B and C). The β finger and the 1585 loop are placed on two sides of the lariat junction (Fig. 4, D and E); this organization likely stabilizes the RNA conformation to allow delivery of the 3′SS–3′ exon into the active site. Notably, four missense mutations affecting the β finger (V1860D, T1865K, A1871E, and T1872E) were previously found to negatively affect the step II, but not step I, reaction (37). Of the mutated residues, Val1860, Ala1871, and Thr1872 are located close to the negatively charged backbone of the intron lariat and U6 snRNA (Fig. 4E); the mutation of these residues to negatively charged Asp/Glu likely destabilizes RNA binding by the β finger, hence inhibiting the step II reaction. The mutation of Thr1865 to a bulky Lys residue may cause local structural perturbation, hence slowing the step II reaction.

Spliceosome remodeling during the C to C* transition

Through adenosine triphosphate (ATP) binding and hydrolysis, Prp16 executes the transition from the C to the C* complex (18, 20, 21). Comparison of the overall structures reveals a major structural rearrangement and numerous minor changes throughout the spliceosome (Fig. 5A). The major change involves the intron BPS/U2 snRNA duplex and the surrounding protein components (Fig. 5, A and B). Compared with that in the C complex (8), the lariat junction in the C* complex undergoes a translation of about 15 to 20 Å and the helical axis of the BPS/U2 duplex is rotated by ~90° (Fig. 5B and fig. S9A). The distal portion of the BPS/U2 duplex away from the lariat junction is translocated by 55 to 65 Å in the C to C* transition. Consequently, the step I splicing factors Cwc25 and Yju2 and the NTC component Isy1, which directly contact the BPS/U2 duplex in the C complex (8) (fig. S9A), are dissociated in the C to C* transition. The step II splicing factor Prp17 and the RNase H–like domain of Prp8 move into their new positions in the C* complex, and the step II splicing factor Prp18 is recruited to interact with the RNase H–like domain (fig. S9A). U2 snRNP as a whole is translocated by 60 to 100 Å, resulting in the alteration of the overall spliceosomal appearance (Fig. 5, A and B). There are numerous minor structural changes during the C to C* transition. For example, the 1585 loop of Prp8 simultaneously interacts with Cwc25 and the BPS/U2 duplex in the C complex (8) (fig. S9B). Although the general location of the 1585 loop remains largely unchanged in the C to C* transition, it is inserted into the active site in the C* complex as a result of remodeling and may directly contact the intron–3′ exon (fig. S9B).

Fig. 5 Spliceosomal remodeling during the C to C* transition.

(A) Comparison of the overall structures between the C and C* complexes. Both spliceosomal complexes are displayed in surface representation. Only those components that undergo drastic conformational or structural rearrangements are colored. The long arrow indicates the direction of movement for U2 snRNP in the C to C* transition. (B) The major structural change occurs to the BPS/U2 snRNA duplex. Both spliceosomal complexes are displayed in cartoon representation. All protein components and U5 and U6 snRNAs are shown in gray.

Discussion

To elucidate the mechanism of pre-mRNA splicing, we have determined the near-atomic structures of four spliceosomal complexes: Bact (9), C (8), C* (this study), and ILS (6, 7), the first three of which were derived from the same set of 761,767 particles of endogenous S. cerevisiae spliceosomes. Neither exogenous pre-mRNA nor recombinant protein was used in our preparation of the spliceosomal particles. Therefore, with the caveat that the purification relied on a Cef1 affinity tag, the relative abundance of the various spliceosomal complexes in the sample may faithfully recapitulate that of the unperturbed yeast nucleus. This analysis suggests that the Bact or C complex is considerably more abundant than the C* complex in S. cerevisiae. Curiously, however, we have been unable to observe a substantial population of the P complex, suggesting its transient nature during the S. cerevisiae splicing cycle. Another transient species is the catalytically activated B* complex, whose classification may have been complicated by its structural resemblance to the C complex. We cannot rule out the possibility that a fraction of the particles used for the reconstruction of the C complex actually represent the B* complex.

To date, detailed structural information has been captured on four of seven spliceosomal complexes (611), yielding unprecedented clarity in mechanistic understanding of the two splicing reactions (Fig. 6). The step I reaction cannot proceed in the Bact complex because the reactants—the BPS and the 5′SS—are separated by a distance of about 50 Å (9, 10). The SF3a protein Prp11 and the splicing factor Cwc24 span the gap between BPS and 5′SS. Prp2, located at the 3′ end of the intron, likely pulls the intron, leading to dissociation of the precursor mRNA retention and splicing (RES) complex and the SF3b complex from the intron. These changes trigger dissociation of Prp11/Cwc24 and recruitment of the step I factors Cwc25/Yju2 into the active site, forming the B* complex. The structure of the B* complex, which remains at large, should be nearly identical to that of the C complex (8, 11). The product of the step I reaction is an intron lariat, as observed in the structure of the C complex (8, 11) (Fig. 6).

Fig. 6 A mechanistic model of pre-mRNA splicing.

Structural changes of the RNA elements and key protein components are indicated in this schematic diagram of a cycle of pre-mRNA splicing. During the step I reaction, the splicing factors Cwc25 and Yju2 play a central role in stabilizing the active site conformation through direct interactions with the RNA elements and the associated protein components. During the second-step reaction, the splicing factors Prp18 and Slu7 facilitate the delivery of the 3′ exon into the active site through physical interactions with the RNase H–like domain of Prp8. The RNase H–like domain, instead, stabilizes the active site conformation by placing the β finger into contact with both BPS/U2 and 5′SS/U6 duplexes. The 1585 loop of Prp8, although not depicted in the diagram here, may play a crucial role in step II catalysis through direct interactions with 3′SS–3′ exon.

The active-site RNA conformation in the C complex is stabilized by at least 15 protein components, which, in addition to Prp8, most notably includes the step I splicing factors Cwc25 and Yju2 (8). The transition from the C to C* complex is executed by Prp16, which presumably pulls the 3′ end of the lariat to trigger its translocation (5, 20, 38, 39). Such a movement likely destabilizes the binding of Cwc25/Yju2 and the NTC component Isy1, leading to their dissociation (Fig. 6). Consequently, the step II factors Prp18/Slu7, which help place the 3′SS–3′ exon into the active site (21, 40, 41), move in to interact with the RNase H–like domain of Prp8 near the active site (Fig. 6). We speculate that, driven by the action of Prp16, the RNase H–like domain of Prp8 undergoes a translocation, which helps move the BPS/U2 duplex to its new location in the C* complex and through direct interaction allows Prp18/Slu7 to facilitate delivery of the 3′SS–3′ exon into the active site. Given its location, the 1585 loop of Prp8 may directly bind and deliver the 3′SS–3′ exon into the active site. The step II reaction proceeds in the C* complex, resulting in the postcatalytic P complex (Fig. 6).

In analogy to that between the B* and C complexes, the structure of the P complex should be very similar to that of the C* complex. The dissociation of the ligated exon from the P complex is triggered by Prp22, which grabs onto and pulls the 3′ exon (22, 39). In the S. pombe ILS structure (6), loop I of U5 snRNA is no longer bound to any exon (Fig. 6). The location of the lariat junction is similar between the S. cerevisiae C* complex and the S. pombe ILS complexes, suggesting relatively minor structural rearrangement from the C* to P and to the ILS complex.

The quality of our EM density map allowed unambiguous identification of an ATPase/helicase, but not differentiation between Prp16 and Prp22. Intriguingly, the location of this ATPase/helicase is close to that of Prp16 in the C complex reported by Nagai and colleagues (11) (fig. S10A). Superposition of the stable structural core between the C and C* complexes reveals marked translation as well as rotation of the ATPase/helicase in the C* complex relative to Prp16 in the C complex (11). This analysis seems to favor the assignment of Prp22, although positional shift of spliceosomal components is common.

Previous studies have strongly suggested the existence of two or more distinct stages in the second catalytic step (5, 40, 42, 43), with Prp18 playing an ATP-independent role (42). Thus, it is entirely possible that our structure of the C* complex only represents that of a particular spliceosomal state. Nonetheless, our structure provides a basis for rationalization of the delicate interactions that are required for the step II reaction. The step II factor Prp18 contains a five-helix bundle and interacts with the RNase H–like domain of Prp8 through helices α4 and α5 (fig. S11A). This observation is consistent with the prediction that helices α4 and α5 interact with U5 snRNP, whereas residues from α1 and α2 may bind Slu7 (44). Formation of a stable Prp18/Slu7 heterodimer (21, 41) suggests that the weak EM density located next to α1/α2 of Prp18 might be attributable to Slu7 (fig. S11B). Mutations of several residues on helices α1/α2 of Prp18 (His118, Lys140, Lys141, Arg150, and Arg151) were found to negatively affect the step II reaction (43, 45), presumably due to weakened interactions with Slu7 (fig. S11C). Deletions or mutations of a conserved surface loop between α4 and α5 of Prp18 (Val191-Ala218), which is located close to the RNase H–like domain of Prp8 (fig. S11C), also led to deficiency in the step II reaction (43, 45), presumably through altered interactions with Prp8 and/or the RNA elements. The surface residues 235TETG238 in Prp17 are located next to, and likely interact with, U2 snRNA and the intron lariat (fig. S11D); their replacement by AAAI or EGGL led to a nonfunctional Prp17 (46), likely through altered interactions.

Supplementary Materials

www.sciencemag.org/content/355/6321/149/suppl/DC1

Materials and Methods

Figs. S1 to S11

Tables S1 to S4

References

References and Notes

  1. ACKNOWLEDGMENTS: We thank the Tsinghua University Branch of China National Center for Protein Sciences (Beijing) for the cryo-EM facility and computation support. This work was supported by funds from the Ministry of Science and Technology (2014ZX09507003006) and the National Natural Science Foundation of China (31430020 and 31321062). The atomic coordinates have been deposited in the Protein Data Bank with accession code 5WSG, and the EM maps have been deposited in Electron Microscopy Data Bank with accession code EMD-6684. The authors declare no competing financial interests.
View Abstract

Navigate This Article