Report

The Prp19p-Associated Complex in Spliceosome Activation

See allHide authors and affiliations

Science  10 Oct 2003:
Vol. 302, Issue 5643, pp. 279-282
DOI: 10.1126/science.1086602

Abstract

During spliceosome activation, a large structural rearrangement occurs that involves the release of two small nuclear RNAs, U1 and U4, and the addition of a protein complex associated with Prp19p. We show here that the Prp19p-associated complex is required for stable association of U5 and U6 with the spliceosome after U4 is dissociated. Ultraviolet crosslinking analysis revealed the existence of two modes of base pairing between U6 and the 5′ splice site, as well as a switch of such base pairing from one to the other that required the Prp19p-associated complex during spliceosome activation. Moreover, a Prp19p-dependent structural change in U6 small nuclear ribonucleoprotein particles was detected that involves destabilization of Sm-like (Lsm) proteins to bring about interactions between the Lsm binding site of U6 and the intron sequence near the 5′ splice site, indicating dynamic association of Lsm with U6 and a direct role of Lsm proteins in activation of the spliceosome.

Splicing of precursor mRNA (pre-mRNA) takes place on a large dynamic ribonucleoprotein complex, the spliceosome, which assembles through sequential interactions of small nuclear ribonucleoprotein particles (snRNPs), in the order of U1, U2, then the preformed U4/U6-U5 tri-snRNP, and numerous protein factors (1, 2). Subsequent to binding of all snRNPs, a large conformational rearrangement in the spliceosome occurs in which U1 and U4 are released, accompanied by new base-pairing formation between U6 and U2, and between U6 and the 5′ splice site of the pre-mRNA, leading to activation of the spliceosome (3). Factors that mediate such structural rearrangements have not been directly demonstrated, although Brr2p was implicated in the unwinding of U4/U6 (4), and Prp28p in displacing U1 from the 5′ splice site (5, 6).

The Prp19p-associated complex NTC (nineteen complex), consisting of at least eight protein components (711), is associated with the spliceosome simultaneously with or immediately after dissociation of U4 (12, 13). Thus, NTC is likely to play an important role in mediating spliceosome activation. To understand the function of NTC, we first examined whether NTC is required for U4 dissociation.

When extracts prepared from the PRP19-HA (hemagglutinin)–tagged strain were depleted of NTC with antibody to HA, the splicing activity was abolished (Fig. 1A, lane 2) but could be restored upon addition of purified NTC (lane 3). Such NTC-depleted extracts were used to assemble the spliceosome on biotinylated pre-mRNA, and the associated small nuclear RNAs (snRNAs) were analyzed after precipitation with streptavidin Sepharose (Fig. 1B). The ACAC pre-mRNA, which contains a 3′ splice site mutation that blocks the second catalytic step (14), was used in this experiment for accumulation of large amounts of the spliceosome. Northern blotting revealed that the spliceosome formed at 2 mM adenosine triphosphate (ATP) contained no U4, regardless of the presence or absence of NTC (lanes 1 and 2), suggesting that U4 dissociation was NTC-independent. Binding of U4 to the spliceosome was not affected by the absence of NTC, as demonstrated by the splicing reaction performed at 0.05 mM ATP (lanes 3 and 4), under which condition the U4-containing splicing complex accumulated, presumably as a result of insufficient amount of ATP to release U4 (13). These results indicate that NTC is not required for dissociation of U4 from the spliceosome.

Fig. 1.

NTC was not required for U4 dissociation but was required for stable association of U5 and U6 after U4 was dissociated. In all of the experiments in (B) through (D), splicing was carried out with biotinylated ACAC pre-mRNA, and the spliceosome formed was isolated by precipitation with streptavidin Sepharose. RNA was extracted for Northern blotting probed with U1, U2, U4, U5, and U6. (A) Extracts prepared from the PRP19-HA–tagged strain were depleted of NTC with antibody to HA and used for splicing assays using the wild-type pre-mRNA with (lane 3) or without (lane 2) the addition of purified NTC. M, mock-depleted; dNTC, NTC-depleted. (B) Splicing was carried out with NTC-depleted (lanes 2 and 4) or mock-depleted extracts (lanes 1, 3, and 5) in the presence of 0.05 mM (lanes 3 to 5) or 2 mM of ATP (lanes 1 and 2). M, mock-depleted; d, NTC-depleted. (C) Splicing was carried out in PRP19-HA extracts in the presence of 0.05 mM, 0.2 mM, 0.5 mM, 1 mM, and 2 mM ATP, and the reaction mixtures were precipitated with streptavidin Sepharose (lanes 1 to 5), or with antibody to HA (lanes 6 to 10). (D) Splicing was carried out in mock-depleted (lanes 1 to 5) or NTC-depleted extracts (lanes 6 to 10), and the spliceosome was precipitated with streptavidin Sepharose. After washing off unbound materials, the pellet was separated into three fractions: one for total precipitate (T, lanes 1 and 6); to the other two fractions were each added the splicing buffer with/without ATP, followed by incubation at room temperature for 20 min. After removing the supernatant (S), the pellet (P) was further washed, and RNA in the pellet (lanes 2, 4, 7, and 9) and supernatant (lanes 3, 5, 8, and 10) fractions was analyzed by Northern blotting.

Comparison of the NTC-containing spliceosome (NTC-spliceosome) and total spliceosome revealed that the association of NTC with the spliceosome was exclusive of U1 and U4 (Fig. 1C). The spliceosome was formed with biotinylated ACAC pre-mRNA in PRP19-HA extracts with increasing concentrations of ATP, and isolated by precipitation with either streptavidin Sepharose for total spliceosome (lanes 1 to 5) or with antibody to HA for NTC-spliceosome (lanes 6 to 10). Whereas total spliceosome contained U4 only at lower concentrations of ATP (lanes 1 to 3) and lost U4 at higher ATP concentrations (lanes 4 and 5), the NTC-spliceosome contained exclusively no U1 or U4 (lanes 6 to 10), suggesting that stable association of NTC with the spliceosome occurs only after dissociation of U1 and U4, with NTC not likely to be required for dissociation of U1.

Activation of the spliceosome is initiated by dissociation of U1 and U4 from the spliceosome, with subsequent base-pair formation between U2 and U6 and between U6 and the 5′ splice site (3). The involvement of NTC in spliceosome assembly after U1 and U4 are dissociated endows NTC with a possible role in the mediation of new base-pair formation. Reasoning that if new base-pair formation is impeded after U4 is dissociated, binding of U6 on the spliceosome might be unstable, we developed an assay to test stabilization of U6 by NTC (Fig. 1D). Spliceosomes formed on biotinylated ACAC pre-mRNA in the presence of 2 mM ATP were isolated by precipitation with streptavidin Sepharose, then incubated at room temperature after re-addition of the reaction buffer containing or not containing ATP. The RNA dissociated from or retained on the spliceosome was analyzed by Northern blotting. In NTC-depleted extracts, although U5 and U6 co-precipitated with the spliceosome after U4 dissociation (lane 6), both were partially dissociated from the spliceosome upon incubation at room temperature (lanes 7 to 10), with ATP facilitating the dissociation. In mock-depleted extracts, U5 and U6 remained stably bound on the spliceosome even after incubation (lanes 2 to 5). This result indicates that NTC is required for stable association of U5 and U6 with the spliceosome after dissociation of U4.

We examined the mechanism of stabilization of U6 by NTC by ultraviolet (UV) crosslinking to probe interactions between snRNAs and the pre-mRNA using an actin precursor, Ac/Cla, which is truncated of sequences beyond the branch point and thus cannot undergo catalytic reaction but allows assembly of the spliceosome and binding of NTC (15). Two major U6/pre-mRNA crosslinked products, X1 and X2 (Fig. 2A, lane 1), were identified in mock-depleted extracts as judged by their sensitivity to RNase H (ribonuclease H) digestion in the presence of U6- and substrate-specific oligonucleotides (lanes 2 and 3). Formation of X1 was NTC-dependent for its absence in NTC-depleted extracts (lane 5). X2 appeared to consist of two bands, the lower one NTC-dependent and the upper NTC-independent (lane 5).

Fig. 2.

Identification of NTC-dependent and NTC-independent base pairings between U6 and the 5′ splice site by UV crosslinking. Splicing was carried out in mock-depleted or NTC-depleted extracts using labeled Ac/Cla, a truncated actin pre-mRNA with only five bases beyond the branch point. The reaction mixtures were precipitated with antibody to Smd1p first, then irradiated with UV, and then the RNA was extracted for further analysis. (A) Oligonucleotide-directed RNase H digestion with U6-specific (lanes 2 and 6), substrate-specific (lanes 3 and 7), and U2-specific (lanes 4 and 8) primers revealed that X1 and X2 were crosslinked to U6. X1 and X2 could be isolated by selection with 5′-biotinylated oligonucleotide U6-A through precipitation with streptavidin Sepharose followed by gel electrophoresis, and then used for analysis in (B) through (D). Lanes 9 and 10 are bound and unbound materials. (B) RNase H mapping of crosslinking sites on the pre-mRNA with oligonucleotides Pre-I, II, III, and IV. Arrows indicate the fragments upshifted because of the presence of crosslinks. Pairs of fragments generated after RNase H digestion are indicated on the left, with IR representing the right-side fragment from Pre-I, and IL the left-side fragment, and so on. dX2, crosslinked product X2 isolated from splicing reaction carried out in NTC-depleted extracts. (C) Primer extension analysis to determine precise crosslinking sites on the pre-mRNA, with oligonucleotide Pre-V as the primer. The sequence of the 5′ splice site is shown on the left with the crosslinked bases marked by *. (D) RNase H mapping of crosslinking sites on U6 with oligonucleotides U6-A, B, C, and D. (E) Primer extension analysis of X1, X2, and dX2, isolated by selection with 5′-biotinylated Pre-IV, then separated on gels to determine the precise crosslinking sites on U6 with U6-B as a primer. The mapped crosslinking sites are marked on the side of the sequence by *. (F) Two documented models for base-pair interactions between U6 and the 5′ splice site. The conserved ACAGA box is indicated. The zigzag indicates proposed crosslinks identified in this study. Consecutive residues that were crosslinked are circled.

Crosslinking sites were mapped by RNase H digestion of affinity-selected, gel-purified crosslinked products (Fig. 2A, lane 9) with oligonucleotides complementary to different regions of U6 or the actin pre-mRNA (16). As shown in Fig. 2B for the pre-mRNA, the crosslinking site on X2, as well as dX2 isolated by using NTC-depleted extracts, was mapped between oligonucleotides I and II of the pre-mRNA in the 5′ splice site region, judging from the fragments that were up-shifted as a result of crosslinks. The crosslinking site of X1 was mapped between the regions of oligonucleotides I and III, likely in the region of oligonucleotide II, because RNase H digestion was less efficient and yielded no upshifted product in the presence of oligonucleotide II (lane 8). Precise crosslinking positions were determined by primer extension with oligonucleotide V as a primer. U6 formed crosslinks with five uridines at positions 28, 29, 30, 34, and 36 of the intron sequence in X1, with U7 and U9 in X2, and primarily with U9 in dX2 (Fig. 2C).

Crosslinking sites on U6 were mapped with four oligonucleotides, U6-A, B, C, and D (Fig. 2D). The crosslinking site in X1 was mapped to the 3′ end of U6, because RNase H digestion with U6-A removed most of the sequence of the crosslinked U6 (lanes 5 and 6). The crosslinking sites in X2 and dX2 were determined to be between C and D regions because of a dramatic band shift from C to D after RNase H treatment (lanes 13, 14 and 22, 23). X2 appeared to consist of two products distinguished after RNase H digestion in the presence of C or D (lanes 13 and 14), suggesting at least two different sets of crosslinks. In contrast, dX2 comprised only one product corresponding to the upper band of X2 (lanes 22 and 23). To determine the precise crosslinking sites, crosslinked products were selected with 5′-biotinylated oligonucleotide Pre-IV and purified on gels. Primer extension analysis with U6-B revealed crosslinking at AAU at positions 44 to 46 immediately upstream of the conserved ACAGA box and G39 in X2, and in dX2 crosslinking of primarily U36–38, which also weakly appeared in X2 (Fig. 2E). These results suggest that U9 of the pre-mRNA intron may crosslink to U36–38 of U6 in dX2, whereas U7, and possibly also U9, of the intron crosslinks to A44A45U46 of U6 in X2.

Two models of base pairing between U6 and the 5′ splice site have been proposed (Fig. 2F), based on the data of UV crosslinking (A) (17) and mutational analysis (B) (18), respectively. Base pairing was shifted by five bases between the two pairing schemes. Another study by site-specific crosslinking of U6 to the pre-mRNA identified crosslinks that fit better in scheme B (19, 20). Comparison of our crosslinks with these two proposed schemes reveals that the NTC-independent crosslinks, U9 of the intron to U36–38 of U6, fit better in scheme A, and the NTC-dependent crosslinks, U7 and U9 of the intron to A44A45U46 of U6, fit in scheme B. Therefore, our results are consistent with the documented U6/5′ splice site interactions and may not only explain the difference in previously observed interactions of detecting different splicing complexes but also suggest a switch of base pairing between U6 and the 5′ splice site that requires NTC.

The crosslinking site of X1 unexpectedly mapped to the region 3′ of primer A (Fig. 2D, lanes 5 to 9) near the 3′ end of the U6 RNA where Lsm proteins bind (21), suggesting that the 3′ end of U6 interacts with the intron in a region near the 5′ splice site during or after spliceosome activation in an NTC-dependent manner. Lsm proteins, which exhibit marked homology to Sm proteins, also form a doughnut-shaped heteroheptamer structure and bind to the 3′-terminal U-tract of U6 RNA (21, 22). Crosslinking of U6 in this region to the intron sequence in an NTC-dependent manner indicates that the U6 snRNP might undergo structural rearrangement, possibly by destabilization of Lsm binding to U6, to expose the 3′-terminal U-tract for interaction with the intron during spliceosome activation. Association of Lsm with the spliceosome was examined by immunoprecipitation of the spliceosome with antibody to V5, using Lsm3p-V5 tagged extracts and ACAC pre-mRNA for accumulation of large amounts of the activated spliceosome. In contrast to the antibody to Ntc85p (Cef1p) (lanes 5 and 6), the antibody to V5 precipitated the spliceosome at a low concentration of ATP (lane 3) but only a very small amount at a high concentration of ATP (lane 4), suggesting the absence of Lsm3p-V5 on the spliceosome formed at high ATP concentrations (Fig. 3A). To rule out the possibility of epitope blocking, the absence of Lsm3p-V5 from the activated spliceosome was further confirmed by Western blotting of the streptavidin Sepharose–pulled down spliceosome formed on biotinylated ACAC pre-mRNA (Fig. 3B). Whereas U6 was present in greater amounts, the amount of Lsm3p was much lower at high ATP than at low ATP concentrations (lanes 3 and 4). These results indicate that Lsm3p, in association with U6, bound to the spliceosome at early stages during spliceosome assembly but was not tightly associated with the spliceosome after the first catalytic reaction had occurred. The same experiment with Ac/Cla pre-mRNA (Fig. 3C) further suggests that Lsm proteins became loosely bound even before the first catalytic step. Similar results were obtained with Lsm5p and Lsm8p, and also with the CYH2 pre-mRNA (not shown), suggesting that the entire Lsm heptamer might be released during spliceosome activation and that this mechanism for pre-mRNA splicing is not unique to the actin pre-mRNA.

Fig. 3.

Immunoprecipitation analysis revealed NTC-dependent destabilization of Lsm3p during activation of the spliceosome. (A) Extracts prepared from the LSM3-V5 strain were used for the splicing reaction carried out in the presence of 0.1 mM of ATP or 2 mM of ATP. The reaction mixtures were immunoprecipitated with antibodies to V5 (lanes 3 and 4), Ntc85p (lanes 5 and 6), and Smd1p (lanes 7 and 8). A 3′ splice site mutant pre-mRNA ACAC was used to block the second reaction in order to accumulate the spliceosome. One-fourth of the materials used for precipitation are shown in lanes 1 and 2 to reveal the extent of splicing. (B) The splicing reaction was carried out as in A with biotinylated ACAC pre-mRNA, and the spliceosome was pulled down with streptavidin Sepharose. The precipitates were subjected to Western blotting using antibodies to Prp19p and V5, and Northern blotting using U6 as a probe. (C) The same immunoprecipitation experiment as in A except with Ac/Cla pre-mRNA. One-eighth of the materials used for precipitation are shown in lanes 1 and 2 to reveal the extent of splicing. (D) Immunoprecipitation experiment with mock- or NTC-depleted extracts prepared from the PRP19-HA/LSM3-V5 strain with Ac/Cla pre-mRNA. RXN, reaction.

Because NTC was required for stable association of U5 and U6 with the spliceosome after U4 dissociation, it is possible that NTC could also play a role in mediating the structural changes of U6 snRNP to untie Lsm proteins. To address this possibility, NTC was depleted from the PRP19-HA/LSM3-V5 double-tagged extract with antibody to HA for the same immunoprecipitation analysis. Figure 3D shows that indeed the amount of Lsm3p retained on the spliceosome increased substantially at high ATP concentrations in NTC-depleted extracts (lane 4), indicating that NTC is required for destabilization of Lsm3p from the spliceosome during spliceosome activation.

Biochemical and genetic analyses have suggested that Lsm proteins play roles in facilitating conformational rearrangements of U6 snRNP and in promoting U4/U6 snRNP formation during cycling of the spliceosome (21, 23). Our finding that Lsm proteins became loosely associated with the active spliceosome further demonstrated a dynamic association of Lsm proteins with U6 and a direct role for Lsm in spliceosome activation. Thus, despite structural similarity, Lsm proteins may play distinct roles from Sm proteins in the splicing reaction, not only being involved in the assembly of U6 or U4/U6 snRNP but also actively participating in the conformational rearrangement of the spliceosome during its activation.

In this report, we show that NTC was not required for dissociation of U4 but was required for subsequent stabilization of U5 and U6 on the spliceosome. Stabilization of U6 involved remodeling of interactions between U6 and the pre-mRNA, including a switch of base pairings between U6 and the 5′ splice site and the formation of new interactions between the 3′ terminal U-tract of U6 and the intron sequence near the 5′ splice site. The latter interaction might require destabilization of Lsm proteins from U6 to expose its 3′ terminal sequence. In this view, NTC may be involved in multiple steps of structural rearrangement of the spliceosome. At least eight proteins have been identified as components of NTC (811), and several others are possible candidates for NTC components from recent proteomic analyses of the spliceosome and NTC-related complexes (2428). Whether different NTC components function in mediating different aspects of structural change or the entire complex functions as a whole to catalyze all the structural rearrangements remains to be studied.

Supporting Online Material

www.sciencemag.org/cgi/content/full/1086602/DC1

Materials and Methods

References

References and Notes

View Abstract

Navigate This Article