Report

Both Catalytic Steps of Nuclear Pre-mRNA Splicing Are Reversible

See allHide authors and affiliations

Science  27 Jun 2008:
Vol. 320, Issue 5884, pp. 1782-1784
DOI: 10.1126/science.1158993

Abstract

Nuclear pre–messenger RNA (pre-mRNA) splicing is an essential processing step for the production of mature mRNAs from most eukaryotic genes. Splicing is catalyzed by a large ribonucleoprotein complex, the spliceosome, which is composed of five small nuclear RNAs and more than 100 protein factors. Despite the complexity of the spliceosome, the chemistry of the splicing reaction is simple, consisting of two consecutive transesterification reactions. The presence of introns in spliceosomal RNAs of certain fungi has suggested that splicing may be reversible; however, this has never been demonstrated experimentally. By using affinity-purified spliceosomes, we have shown that both catalytic steps of splicing can be efficiently reversed under appropriate conditions. These results provide considerable insight into the catalytic flexibility of the spliceosome.

To examine the potential reversibility of splicing, we used a dominant-negative mutant of Prp22, a DExD/H-box RNA helicase. Prp22 is required for the second catalytic step and the release of mature message (fig. S1) (1, 2). The prp22S635A mutant retains its adenosine triphosphatase activity and its ability to promote the catalysis of the second transesterification step but is defective in releasing mRNA from the spliceosome; it remains associated with the spliceosome after completion of splicing (3). We purified the V5-tagged version of the mutant protein and added to in vitro splicing assays. As expected, the spliceosome purified with antibody to V5 contained primarily spliced products (lariat intron and mature mRNA) with a small amount of splicing intermediates (Fig. 1A, lane 2). The availability of these essentially homogeneous postcatalytic spliceosomes allowed us to explore the possibility that the splicing reactions could be reversed.

Fig. 1.

Reverse splicing of both catalytic steps. (A and B) Splicing was carried out in the presence of recombinant V5-tagged prp22S635A protein (lane 1). The spliceosome precipitated by antibody to V5 (lane 2) was then incubated at 25°C under various conditions. (A) 10 mM Tris-HCl at pH 6.8, 7.4, 8.0, or 8.8, 4 mM MgCl2 without (lanes 3 to 6) or with (lanes 7 to 10) 150 mM KCl for 1 hour. (B) 10 mM Tris-HCl, pH 8.8, 4 mM MgCl2, 150 mM KCl for 3, 5, 10, 20, 40, or 60 min (lanes 3 to 8), or without KCl for 3, 5, 10, 20, 40, or 60 min (lanes 10 to 15), then the addition of 150 mM KCl for an additional 3, 5, and 10 min (lanes 16 to 18). (C) Splicing was carried out with ACAC pre-mRNA, using extracts with V5-tagged Prp22 (lane 1). The spliceosome precipitated by antibody to V5 (lane 2) was then incubated at 25°C in 10 mM Tris-HCl at pH 6.8, 7.4, 8.0, or 8.8, 4 mM MgCl2 without (lanes 3 to 6) or with (lanes 7 to 10) 150 mM KCl for 1 hour. R, reaction; SP, spliceosome.

After trying different incubation conditions, we observed, in the absence of monovalent ions, what appeared to be efficient reversal of the second step (the R2 reaction), that is, a dramatic reduction in spliced products accompanied by a corresponding increase in splicing intermediates (lanes 3 to 6). To confirm that this was indeed the case, we performed a number of analyses. Primer extension probing the 5′ end of the intron and branch point indicated that the accumulated splicing intermediates did in fact result from the faithful reversal at the second transesterification reaction. Moreover, sequence analysis of cloned reverse transcription polymerase chain reaction products confirmed the restoration of the authentic 3′ splice junction (table S1). The reverse reaction was precise, with no errors detected in the 28 clones analyzed (table S1), and was independent of pH over the range of 6.8 to 8.8 (Fig. 1A, lanes 3 to 6). These results unambiguously indicated that reversal of splicing occurred in the absence of monovalent ions. The reverse reaction was strongly inhibited by an addition of 150 mM KCl (lanes 7 to 10).

Kinetic analysis revealed that the R2 reaction was remarkably efficient (Fig. 1B, lanes 9 to 15). We then tested whether R2 was itself “reversible.” We incubated the purified spliceosome in the absence of KCl for 1 hour and then added KCl. Strikingly, the splicing intermediates were rapidly converted to spliced products (the F2 reaction) (lanes 16 to 18). These results indicate that the purified spliceosomes were in a dynamic state, with the forward reaction strongly favored in the presence of KCl, whereas the reverse reaction was strongly favored in the absence of KCl.

In these analyses, we noticed that when the purified spliceosomes were incubated in the presence of KCl, a small amount of intact pre-mRNA was also generated, indicating reversal of the first step (the R1 reaction). Since KCl also promoted F2, we speculated that R1 might be in competition with F2, and the efficiency of R1 would increase if F2 were blocked. We then prepared spliceosomes arrested after the first transesterification reaction using pre-mRNA with 3′ splice site mutations (ACAC) that prevent exon ligation (4). Spliceosomes formed on the mutant substrate still bind V5-tagged Prp22 but could not produce spliced products (fig. S1), and the affinity-purified spliceosome contained largely splicing intermediates (Fig. 1C, lane 2). Indeed, we found that reversal of the step one reaction was also highly efficient, but in this case required KCl (lanes 7 to 10). The R1 reaction was remarkably precise, with no error detected at the 5′ splice site in 33 clones analyzed (table S1) (5). Under these conditions, we also observed the completion of the second transesterification reaction to a certain degree (lanes 7 to 10), yet with a high error rate (tables S1 and S2) (5), consistent with a previous report (6). The F2 reaction appeared to be less precise than the reverse reaction because errors were also detected with the wild-type substrate. In this analysis, spliceosomes containing spliced products were incubated in the absence of KCl to drive R2 followed by the addition of KCl to drive F2, and the yielded mRNA was analyzed. Three out of 43 clones analyzed had products of aberrant ligation (tables S1 and S2) (5).

We then examined the divalent cation requirements for the reverse reactions. We found that 4 mM Ca2+ and Mn2+, but not Zn2+, supported R1 in the presence of KCl (Fig. 2A, lanes 8 to 10), but at a lower efficiency than Mg2+ (lane 7). Mn2+ efficiently promoted F2 but with relaxed specificity, particularly in the absence of KCl (tables S1 and S2, lane 5). In contrast, the R2 reaction had a strict requirement for Mg2+ at a concentration of 4 mM (Fig. 2B, lanes 4 to 7 and 9 to 12). The difference in divalent cation requirements for the two reactions is intriguing, but we currently do not understand the mechanistic reasons for this phenomenon.

Fig. 2.

Divalent cation requirements and the SER reaction. (A) Splicing was carried out with ACAC pre-mRNA using extracts with V5-tagged Prp22 (lane 1). The spliceosome precipitated by antibody to V5 (lane 2) was then incubated at 25°C in 10 mM Tris-HCl, pH 8.8, 4 mM MgCl2, CaCl2, MnCl2, or ZnCl2, without (lanes 3 to 6) or with (lanes 7 to 10) 150 mM KCl for 1 hour. (B to E) Splicing was carried out in the presence of prp22S635A-V5 (lane 1). The spliceosome precipitated by antibody to V5 (lane 2) was then incubated under various conditions. (B) 10 mM Tris-HCl, pH 8.8, 0 or 4 mM MgCl2, CaCl2, MnCl2, or ZnCl2, without (lanes 3 to 7) or with (lanes 8 to 12) 150 mM KCl for 1 hour. (C) Primer extension analysis of the putative E2 fragment to map the cleavage site. (D) 10 mM Tris-HCl, pH 8.8, 150 mM KCl with 0, 0.1 mM, 1 mM, 2 mM, 4 mM, 8 mM, or 20 mM MnCl2 (lanes 3 to 9) for 1 hour. (E) 10 mM Tris-HCl at pH 7.4, 8.0, 8.8, or 9.5, 8 mM MnCl2 without (lanes 3 to 6) or with (lanes 7 to 10) 150 mM KCl for 1 hour. R, reaction; SP, spliceosome.

In the presence of Mn2+, we also observed production of the 5′ exon (E1) and the 3′ exon (E2) (Fig. 2B, lane 11), presumably from cleavage of mRNA at the splice junction, a reaction analogous to spliced exon reopening (SER), known to be catalyzed by Group II intron (7). Precise cleavage at the splice junction was confirmed by primer extension analysis of gel-purified E2 fragment (Fig. 2C). Nevertheless, we cannot exclude the possibility that cleavage occurred on the lariat intron-exon 2 following the R2 reaction. Titration of Mn2+ for SER revealed a requirement for Mn2+ at a concentration of ≥4 mM (Fig. 2D, lanes 7 to 9) and high pH (Fig. 2E, lanes 9 and 10). Lower Mn2+ concentrations or pH promoted R2 instead (Fig. 2D, lanes 4 to 6; Fig. 2E, lanes 7 and 8). These results suggest that spliceosomal introns also catalyze SER.

We have shown that both catalytic steps of pre-mRNA splicing are reversible under appropriate conditions. The reverse reactions were extremely precise, with no errors detected in either catalytic step. This is in contrast to the forward reactions, in which aberrant exon ligation was detectable even with the wild-type substrate. The biological importance of higher fidelity in the reverse reaction remains to be explored.

Reversible splicing has been inferred by the presence of introns in spliceosomal small nuclear RNAs (snRNAs) of certain fungi (8) and the fact that identical transesterification reactions are readily reversible in the context of Group II introns (9, 10). Nevertheless, the reversibility in the context of a fully assembled spliceosome is highly efficient. In this regard, the assembly of the spliceosome is a complex and intricately orchestrated process (11, 12), and each chemical step requires the prior action of specific DExD/H-box RNA helicases to prepare the spliceosome for catalysis (13, 14). Our results indicate that these rearrangements are not required for the reverse reaction, at least for the Prp16-dependent step, given that Prp16 is not retained on the purified spliceosome (fig. S2).

Furthermore, our results indicate that the catalytic center of the spliceosome is highly stable but chemically adaptable. Relatively subtle changes in ionic environment can drive the reaction either forward or in reverse direction, or toward SER, which suggests that large conformational rearrangements are not required between these catalytic steps. These observations are consistent with the two-stage model of spliceosome activity proposed by Query and Konarska. Their model suggests that conformations that favor the first step inhibit the second step and vice versa (15). It will be of considerable interest to probe the snRNA topology in the presence and absence of monovalent ions to obtain insight into the active site of the spliceosome.

We have also shown that the spliceosome could catalyze SER, which has previously been demonstrated in Group II introns, and suggested to represent hydrolytic reversal of the second step (7). In this regard, SER is a nonproductive pathway of the reverse reaction. Our results have not only shown that the spliceosome is able to catalyze hydrolytic reaction but also indicated that SER is conserved between these two types of introns, although the importance of such a non-productive pathway in the splicing reaction is unknown.

Supporting Online Material

www.sciencemag.org/cgi/content/full/320/5884/1782/DC1

Materials and Methods

Figs. S1 and S2

Tables S1 and S2

References

References and Notes

View Abstract

Navigate This Article