Functional Association of U2 snRNP with the ATP-Independent Spliceosomal Complex E

Molecular Cell, Vol. 5, 779–787, May, 2000, Copyright 2000 by Cell Press

Functional Association of U2 snRNP with the ATP-Independent Spliceosomal Complex E Rita Das, Zhaolan Zhou, and Robin Reed* Department of Cell Biology Harvard Medical School Boston, Massachusetts 02115

Summary In the current model for spliceosome assembly, U1 snRNP binds to the 5ⴕ splice site in the E complex followed by ATP-dependent binding of U2 snRNP to the branchpoint sequence (BPS) in the A complex. Here we report the characterization of highly purified, functional E complex. We provide evidence that this complex contains functional U2 snRNP and that this snRNP is required for E complex assembly. The BPS is not required for U2 snRNP binding in the E complex. These data suggest a model for spliceosome assembly in which U1 and U2 snRNPs first associate with the spliceosome in the E complex and then an ATP-dependent step results in highly stable U2 snRNP binding to the BPS in the A complex. Introduction During spliceosome assembly, multiple dynamic interactions occur among the five spliceosomal snRNAs (U1, U2, U4, U5, and U6), the approximately 50 spliceosomal proteins, and the pre-mRNA. These interactions take place during assembly of the spliceosomal complexes that form in the temporal order E, A, B, and C. The E complex assembles in the absence of ATP, whereas assembly of the other complexes is ATP dependent. According to the present model for spliceosome assembly, U1 snRNP first binds in the E complex, followed by U2 snRNP binding in the A complex and U4/5/6 snRNP binding in the B complex. Several rearrangements then occur that activate the spliceosome for the two catalytic steps of splicing in the C complex (Burge et al., 1998; Staley and Guthrie, 1998; Reed, 2000). Although a great deal of progress has been made in identifying the critical RNA–RNA and RNA–protein interactions required for splicing, much less is known about the precise timing of these interactions during spliceosome assembly. In addition, little is understood about the factors that mediate the dynamics. Progress in determining the specific timing of events has been hampered, in part because it has not been possible to isolate spliceosomal complexes that are both highly purified and functional. In many of the methods currently used to analyze spliceosomal complexes, high salt or heparin treatment is required. In addition, no single method has been suitable for side-by-side analyses of all of the complexes. Not surprisingly, major differences have been detected in both the protein and snRNA compositions of complexes isolated by different * To whom correspondence should be addressed (e-mail: rreed@ hms.harvard.edu).

methods (e.g., Grabowski and Sharp, 1986; Konarska and Sharp, 1986, 1987; Zillmann et al., 1988; Bennett et al., 1992; Jamison and Garcia-Blanco, 1992; Staknis and Reed, 1994a; Hong et al., 1997; Staley and Guthrie, 1999). As a consequence, it has been difficult to establish the functional significance of many of the factors present in the complexes. Of relevance to the present study, only U1 snRNP is detected as a stable component of highly purified E complex (Bennett et al., 1992; Michaud and Reed, 1993). In contrast, immunoprecipitations carried out under gentle conditions indicate that both U1 and U2 snRNPs are present in the E complex (Hong et al., 1997). Thus, it is not clear when U2 snRNP first functionally joins the spliceosome. Numerous isolation methods have shown that U2 snRNP is tightly bound to the pre-mRNA in the A complex and in the subsequent spliceosomal complexes. An essential base-pairing interaction between U2 snRNA and the branchpoint sequence (BPS) is thought to specify the adenosine that functions as the nucleophile for the first catalytic step of splicing. In addition, interactions among U2 and U6 snRNAs and the premRNA are involved in formation of the catalytic center of the spliceosome (Burge et al., 1998; Staley and Guthrie, 1998). Although these and other important roles of U2 snRNP are well known, the critical issue of when all of these interactions take place has not yet been determined. In this study, we have used a recently established method for isolating spliceosomes that are both highly purified and functional. Our results show that U2 snRNP is functionally associated with the E complex and is also required for its assembly. Thus, in contrast to the current model for the early steps in spliceosome assembly, our data support a model in which both U1 and U2 snRNPs first associate with the spliceosome in the E complex. Results U2 snRNP Is Present in MBP-Purified E Complex In previous studies, functional mammalian spliceosomes were partially purified by gel filtration under conditions compatible with splicing (60 mM salt) (Michaud and Reed, 1991, 1993; Jamison et al., 1992). In contrast, for determining protein compositions, complexes were isolated by gel filtration, treated with high salt (250 mM salt), and purified by biotin-avidin affinity selection (Bennett et al., 1992; Michaud and Reed, 1993; Gozani et al., 1994). Because there are significant differences in the compositions of the complexes isolated by these and other methods, we have now characterized the E complex using a recently developed method for isolating spliceosomes that are both highly purified and functional (see Experimental Procedures; Z. Z. and R. R., unpublished data). In this procedure, spliceosomes are assembled on pre-mRNA that is prebound to the maltose binding protein (MBP). The spliceosomes are then isolated by gel filtration, bound to amylose beads, and gently eluted with maltose. The resulting MBP-purified

Molecular Cell 780

Figure 1. SnRNA and Protein Compositions of Purified E Complex (A) SnRNAs in purified E complex. Total RNA was extracted from the E complex (lane 3), end labeled (lane 2), and fractionated on an 8% polyacrylamide gel. As a marker for the snRNAs, total RNA was extracted from nuclear extract and end labeled (lane 1). RNAs were visualized by phosphorimager analysis. The low level of U5 snRNA detected in the E complex may be the same as the ATP-independent association of U5 snRNP detected previously (Chabot et al., 1985). The significance of this interaction is not known. (B) Native gel analysis of E and A complexes. 32 P-labeled AdML pre-mRNA was incubated in splicing extracts in the absence (lane 1) or presence of ATP (lane 2), and heparin was added prior to loading onto a 1% agarose gel. The bands corresponding to the H, E, and A complexes are indicated. (C) Analysis of proteins in purified E complex. Total protein was prepared from equivalent amounts of purified E and H complexes, separated on a 9% SDS gel, transferred to nitrocellulose, and probed with U1A, U2AF65, U2AF35, and mBBP antibodies as indicated. The smaller bands detected with the U2AF65 and SAP 145 antibodies may be breakdown products. The extra bands detected in nuclear extract with the mBBP antibody may be other forms of this protein (Arning et al., 1996). (D) Same as (C) except blots were probed with antibodies to the U2 snRNP components, SF3a, SF3b (SAP 130 and SAP 145), and B⬘⬘ as indicated.

spliceosomes are active in splicing when incubated in complementing extracts (Z. Z. and R. R., unpublished data, and see below). Significantly, both U1 and U2 snRNAs are detected in the MBP-purified E complex (Figure 1A). Comparison of these snRNAs by ethidium bromide staining and end labeling indicates that they are present in the E complex in about a one to one ratio (data not shown). The presence of U2 snRNA is not due to contaminating A complex, as no A complex is detected in the E complex reactions after heparin treatment and fractionation on a native agarose gel (Figure 1B; note that E and H complexes comigrate under these gel conditions) (Michaud and Reed, 1993; Das and Reed, 1999). Western analysis of the MBP-purified E complex revealed the presence of several proteins expected to be in the E complex, including the U1 snRNP protein U1A, both subunits of U2AF, and the branchpoint binding protein, mBBP/SF1 (referred to hereafter as mBBP) (Bennett et al., 1992; Arning et al., 1996; Berglund et al., 1997). All of these proteins are specifically associated with the E complex, as they were not detected in the hnRNP complex H (Figure 1C). We next asked whether U2 snRNP proteins were present in the MBP-purified E complex. U2 snRNP can be isolated in a 12S and a 17S form (Behrens et al., 1993a, 19993b). The B⬘⬘ protein is a stable component of both forms (Behrens et al., 1993a). In contrast, the two essential multimeric splicing factors, SF3a and SF3b, are present only in the 17S form (Behrens et al., 1993a, 1993b; Brosi et al., 1993a; Staknis and Reed, 1994b; Kramer et al., 1999). SF3a consists of three subunits (spliceosomeassociated proteins [SAPs] 61, 62, and 114), and SF3b consists of four subunits (SAPs 49, 130, 145, and 155) (Brosi et al., 1993b; Das et al., 1999; Kramer et al., 1999).

Significantly, B⬘⬘, as well as SF3a and SF3b, was detected in the MBP-purified E complex (Figure 1D and data not shown; see below for description of the antibody generated against SF3a). None of the U2 snRNP proteins were present in the H complex (Figure 1D). We conclude that 17S U2 snRNP is specifically associated with the E complex. To determine whether the 17S U2 snRNP components were quantitatively associated with the E complex or were only present in a subpopulation of this complex, we used a native gel assay to ask whether antibodies to 17S U2 snRNP can supershift the E complex. For comparison, we also examined the A and B complexes, which are known to contain 17S U2 snRNP. Agarose gels were used for the assays as these gels were recently shown to resolve the ATP-dependent spliceosomal complexes (A, B, and C), as well as the E and H complexes (Das and Reed, 1999). The E complex is not stable in the presence of heparin, whereas the ATPdependent complexes are heparin resistant. For the supershift assay, we first tested the SF3a antibody. An antibody to the catalytic step II protein, hPrp16 (Zhou and Reed, 1998), was used as a negative control. The antibodies were purified under identical conditions and adjusted to equal levels (Figure 2A). As expected, the A and B complexes were supershifted with the SF3a antibody, but not with an equal amount of the hPrp16 antibody (Figure 2B). Significantly, the E complex was also efficiently supershifted with the SF3a antibody, but not with the hPrp16 antibody (Figure 2C). We conclude that SF3a is quantitatively associated with the E complex. In contrast to SF3a, B⬘⬘ is very tightly associated with U2 snRNP (Behrens et al., 1993b). Thus, to determine whether the entire U2 snRNP is likely to be quantitatively

Functional Association of U2 snRNP with E Complex 781

Figure 2. U2 snRNP Is Stoichiometrically Associated with the E and A Complexes (A) Affinity-purified SF3a and hPrp16 antibodies were separated by SDS-PAGE. The arrow indicates the antibody heavy chain. (B and C) The A and B spliceosomal complexes were assembled on 32P-labeled AdML pre-mRNA in presence of ATP. Complexes were incubated without antibody (lanes 1 and 2), with SF3a antibody (lanes 3 and 4), or with hPrp16 antibody (lanes 5 and 6) and fractionated on a native agarose gel. The H, A, and B complexes are indicated. The supershift complexes are detected in the well of the gel. (C) Same as (B) except the E complex was assembled in absence of ATP. The E and H complexes are indicated, and the supershifted complex is detected in the well of the gel. (D) Affinity-purified B⬘⬘ antibody was separated on by SDS-PAGE. The arrows indicate the antibody heavy and light chains. (E) The E complex was assembled on 32Plabeled AdML pre-mRNA in absence of ATP, and complexes were incubated without (lanes 1 and 2) or with the B⬘⬘ antibody and fractionated on a native agarose gel.

associated with the E complex, we carried out the supershift assay using the B⬘⬘ antibody (Figure 2D). As shown in Figure 2E, the E complex is supershifted in a dose-dependent manner by the B⬘⬘ antibody. These data, together with the results in Figure 1, indicate that U2 snRNP is specifically and quantitatively associated with the E complex. The presence of U2 snRNP in the E complex is likely to be general, as the SF3a antibody also quantitatively supershifts the E complex assembled on Ftz pre-mRNA (data not shown). U2 snRNP Associates with the E Complex Independently of the BPS Previous studies have shown that the stable binding of U2 snRNP in the A complex requires the BPS (Champion-Arnaud et al., 1995; Query et al., 1996, 1997). To determine whether the association of U2 snRNP with the E complex is also BPS dependent, we assembled the E complex on a pre-mRNA lacking the BPS. This mutant is unable to form the A complex but forms the E complex efficiently (Champion-Arnaud et al., 1995; Query et al., 1996). Significantly, both U1 and U2 snRNAs were detected in the MBP-purified ⌬BPS E complex (Figure 3A). Moreover, the 17S form of U2 snRNP is present in the ⌬BPS E complex as the subunits of SF3a/b were detected on Western blots of this complex (Figure 3B and data not shown). We conclude that U2 snRNP is associated with the E complex via a BPS-independent interaction. SF3a Is Functional in the E Complex To determine whether U2 snRNP is functionally associated with the E complex, it was first necessary to obtain nuclear extracts specifically lacking U2 snRNP activity. Because this snRNP is so abundant, it is difficult to

completely immunodeplete it and, at the same time, retain a highly active extract (our unpublished data). Oligonucleotide-directed RNase H inactivation of U2 snRNA is not sufficient for similar reasons (our unpublished data). Thus, as an alternative strategy, we raised

Figure 3. U2 snRNP Associates with the E Complex in the Absence of the BPS (A) The E and H complexes were assembled on 32P-labeled AdMLM3⌬BPS pre-mRNA and fractionated by gel filtration, affinity purified by binding to amylose beads, and eluted with maltose. Equal amounts of pre-mRNA were prepared from purified E and H complexes, end labeled with [32P]pCp and RNA ligase, and fractionated on an 8% polyacrylamide gel. The bands corresponding to premRNA and nuclear RNAs are indicated. (B) Western analysis. Total protein was prepared from equivalent amounts of purified E and H complexes and separated on a 9% SDS gel, transferred to nitrocellulose, and probed with the SF3a antibody.

Molecular Cell 782

Figure 4. SF3a Immunodepletion and Reconstitution with Recombinant SF3a (A) Coomassie blue staining of rSF3a complex purified from baculovirus. (B) Western blot of nuclear extract (lanes 1 and 4), mock-depleted extract (lanes 2 and 5), and ⌬SF3a-depleted extract (lanes 3 and 6) probed with SF3a or SAP 155 antibodies as indicated. (C) Immunodepletion/add-back assays of spliceosome assembly. AdML pre-mRNA was incubated in SF3a-depleted (lanes 1–4) or mockdepleted (lanes 5 and 6) extracts for the times indicated. rSF3a (120 ng) was added to the ⌬SF3a extract in lanes 3 and 4. Spliceosomal complexes were analyzed on a 2% native agarose gel. Ori indicates the gel origin. (D) Same as (C) except that splicing products were analyzed on a 13.5% polyacrylamide denaturing gel. Splicing intermediates and products are indicated.

a polyclonal antibody to the 17S U2 snRNP–specific SF3a complex, reasoning that an antibody to the entire complex may be sufficiently high affinity to use for efficient and specific immunodepletions. To raise the antibody, the three recombinant subunits of SF3a were coexpressed in baculovirus. Superose 6 gel filtration revealed that all three proteins were present in a discrete complex in a 1:1:1 stoichiometry (Figure 4A and data not shown). Significantly, a rabbit polyclonal antibody raised against the recombinant SF3a (rSF3a) specifically recognizes all three SF3a subunits on a Western blot of total HeLa cell nuclear extract (Figure 4B, NE). To determine whether the antibodies could be used to prepare a highly active immunodepleted extract, we carried out immunodepletion/reconstitution assays. Little depletion of SF3a or U2 snRNP was detected in nuclear extract under normal splicing conditions (data not shown). However, when the salt in the nuclear extract was raised to 700 mM, efficient depletion of SF3a was observed with the SF3a antibody, but not in the mock control (Figure 4B, lanes 2 and 3). Significantly, other U2 snRNP components, such as SF3b, were not codepleted (e.g., Figure 4B, lane 6). To determine

whether spliceosome assembly is blocked in the ⌬SF3a extract, AdML pre-mRNA was incubated in ⌬SF3a or mock-depleted extracts. As shown in Figure 4C (lanes 1 and 2), A and B complex assembly is blocked in the ⌬SF3a-depleted, but not in the mock-depleted, extract (lanes 5 and 6). Importantly, rSF3a efficiently restores spliceosome assembly in the ⌬SF3a extract (lanes 3 and 4) and in a dose-dependent manner (data not shown). We conclude that SF3a can be depleted from nuclear extract and substituted with rSF3a to regain efficient spliceosome assembly. Splicing is also inhibited in the ⌬SF3a extract but not in the mock-depleted extract (Figure 4D, lanes 3, 4, 7, and 8). Moreover, addition of rSF3a efficiently restores splicing (Figure 4D, lanes 9 and 10). Taken together, these data indicate that nuclear extracts can be specifically depleted of the essential U2 snRNP component, SF3a, and are highly active when complemented with recombinant SF3a. We next asked whether the MBP-purified E complex could be chased to spliced products in the ⌬SF3a extract (Figure 5). MBP-purified A complex, which should contain functional SF3a, was used as a positive control. Both E and A complexes were assembled on AdML-M3

Functional Association of U2 snRNP with E Complex 783

Figure 5. SF3a Is Functionally Associated with the Purified E Complex (A) AdML pre-mRNA (lanes 1 and 2) or AdML-M3 pre-mRNA (lanes 5 and 6) was incubated under standard splicing conditions in nuclear extract. AdML pre-mRNA was incubated in SF3a-depleted extract (lanes 3 and 4). MBP-purified A complex (lanes 7–12) or E complex (lanes 13–18) were incubated under the indicated conditions. (B) AdML pre-mRNA (lane 1) or affinity-purified E complex (lanes 3) was incubated under splicing conditions in U2AF65-depleted extract. Affinity-purified E complex incubated under splicing conditions in the absence of extract is shown in lane 2. Splicing products were separated on 13.5% denaturing polyacrylamide gel. Splicing intermediates and products are indicated.

pre-mRNA, which contains the three hairpins used for the MBP spliceosome purification. AdML pre-mRNA, which lacks these hairpins, was used as a control in some of the assays (see below). As expected, no splicing was observed when naked AdML pre-mRNA (Figure 5, lanes 3 and 4) was incubated in the ⌬SF3a extract for 25⬘ or 50⬘. Likewise, splicing did not occur when either the purified A complex (lanes 11 and 12) or the purified E complex (lanes 17 and 18) were incubated under splicing conditions in the absence of extract. In contrast, splicing intermediates and products were detected when the A complex was incubated in the ⌬SF3a extract (Figure 5, lanes 7 and 8). Significantly, splicing also occurred when the purified E complex was incubated in the ⌬SF3a extract (Figure 5, lanes 13 and 14). One possible interpretation of these data is that the splicing observed with the purified E and A complexes is due to splicing of the pre-mRNA present in these complexes. Alternatively, the SF3a present in these complexes may simply be complementing the ⌬SF3a extract to splice the pre-mRNA. To distinguish between these possibilities, we carried out a mixing experiment

using two different AdML derivatives. The purified E and A complexes were assembled on AdML-M3 pre-mRNA, which contains a longer second exon than AdML premRNA (see Experimental Procedures). The products generated from splicing naked AdML or AdML-M3 premRNA in normal nuclear extract are shown in Figure 5, lanes 1 and 2 and lanes 5 and 6, respectively. Significantly, efficient splicing of only the AdML-M3 was detected when AdML pre-mRNA was mixed with the purified A complex (lanes 9 and 10) or with the purified E complex (lanes 15 and 16). This observation indicates that the SF3a in these complexes is not complementing the ⌬SF3a extract to splice the naked pre-mRNA. We conclude that SF3a is not only a functional component of the A complex, but also of the E complex. The purified E complex can also be chased to spliced products in a U2AF-depleted extract (Figure 5B), indicating that U2AF is a functional component of the E complex. The observation that the pre-mRNA in the E complex is not completely spliced in either the ⌬SF3a or ⌬U2AF extracts may be because a portion of the complex dissociates during purification.

Molecular Cell 784

Figure 6. SF3a Is Required for E Complex Assembly AdML pre-mRNA was incubated in the absence of ATP for the times indicated in SF3a-depleted extract (lanes 1–4) or mock-depleted extract (lanes 5 and 6). rSF3a was added to SF3a-depleted extract in lanes 3 and 4. Reactions were fractionated on a 1.5% native agarose gel. The ⌬SF3a complex and the E and H complexes are indicated.

The data presented above indicate that SF3a is a functional component of the E complex. As SF3a is an essential component of 17S U2 snRNP, and this snRNP is present in the purified E complex (Figure 1), it is likely that the entire U2 snRNP is a functional component of the E complex. To obtain evidence that SF3a (and U2 snRNP) is required for E complex assembly, we investigated complex assembly in the ⌬SF3a extract (Figure 6). When AdML pre-mRNA was incubated in the ⌬SF3a extract, the levels of E complex were significantly decreased. In addition, low levels of a complex (designated the ⌬SF3a complex) that runs with slightly faster mobility than the E complex were reproducibly detected (Figure 6). Significantly, addition of rSF3a to the ⌬SF3a extract restores the E complex (Figure 6). These data indicate that SF3a is required for E complex assembly.

Figure 7. Model for Early Steps in Spliceosome Assembly The tight binding of U1 and U2 snRNPs is indicated by the thickly lined circles, and the loose binding of these snRNPs and U2AF is indicated by the dashed circles. See text for description of model.

Discussion Here we report the characterization of purified functional spliceosomal complex E. In contrast to the current model of spliceosome assembly, which proposes that U2 snRNP first binds in the A complex, our data indicate that U2 snRNP first associates with pre-mRNA during E complex formation. Our evidence that U2 snRNP is a component of the E complex includes the observations that U2 snRNA and the U2 snRNP–specific components, SF3a, SF3b, and B⬘⬘ are detected in the E complex. U2 snRNP is stoichiometrically associated with the E complex because antibodies to B⬘⬘ or SF3a quantitatively shift the E complex in a native gel supershift assay. Finally, U2 snRNP appears to be functionally associated with the E complex, as the E complex can be chased to spliced products in an extract lacking the essential U2 snRNP component SF3a. Moreover, in the absence of SF3a, low levels of a complex slightly smaller than the E complex are detected on a native gel, and the E complex can be restored by addition of rSF3a. Efficient E complex assembly requires both the 5⬘ and 3⬘ splice sites, but not the branch site (Michaud and Reed, 1993; Champion-Arnaud et al., 1995). Our data

indicate that the association of U2 snRNP with the E complex is also not dependent on the branch site. SF3a and SF3b components can be UV cross-linked to the pre-mRNA surrounding the BPS in the A complex (Gozani et al., 1996, 1998). In contrast, there is no detectable cross-linking of these components in the E complex (Gozani et al., 1996, 1998). The association of U2 snRNP is also much tighter in the A complex than in the E complex, as U2 snRNP is only detected in the E complex isolated under gentle conditions. Thus, the most straightforward interpretation of our data is that U2 snRNP first binds loosely in the E complex and then an ATP-dependent event, possibly a conformation change in the snRNP, results in the highly stable binding of U2 snRNP at the branch site in the A complex. A Model for the Early Steps in Spliceosome Assembly A model for the early steps in spliceosome assembly is shown in Figure 7. As has been well established, U1 snRNP first binds tightly to the pre-mRNA in the E complex, and a duplex is formed between U1 snRNA and the 5⬘ splice site. Recent studies in S. cerevisiae indicate

Functional Association of U2 snRNP with E Complex 785

that several U1 snRNP proteins, including U1 70K, interact with the region surrounding the 5⬘ splice site in the commitment complex (the E complex counterpart) (Zhang and Rosbash, 1999). Some of the U1 snRNP proteins are conserved in metazoans raising the possibility that similar interactions may occur in the E complex (Zhang and Rosbash, 1999). In metazoans, the BPS and 3⬘ splice site are recognized by mBBP and the U2AF heterodimer, respectively (Ruskin et al., 1988; Bennett et al., 1992; Kramer, 1992; Zamore et al., 1992; Arning et al., 1996; Abovich and Rosbash, 1997; Berglund et al., 1997), with U2AF65 recognizing the pyrimidine tract and U2AF35 recognizing the AG (Merendino et al., 1999; Wu et al., 1999; Zorio and Blumenthal, 1999). The general features of this model are the same in S. cerevisiae, except that the AG is dispensable for spliceosome assembly in yeast and a U2AF35 homolog does not exist. Previous work showed that the U2AF heterodimer interacts with SAP 155 (Gozani et al., 1998). As SAP 155 is a component of both U2 snRNP and the E complex, this protein may mediate the association of U2 snRNP with the E complex (Figure 7). In addition, fractions of U1 and U2 snRNPs interact with each other in HeLa nuclear extracts (Mattaj et al., 1986). Thus, U1 snRNP may be involved in mediating the association of U2 snRNP with the E complex. Finally, the SR protein family of splicing factors, which promote E complex assembly (Staknis and Reed, 1994a; Kohtz et al., 1994; Zuo and Maniatis, 1996) and interact with exon sequences in the E complex (Chiara et al., 1996), may stabilize the association of U2 snRNP with the E complex. The 5⬘ and 3⬘ ends of the intron are first bridged in the E complex (Seraphin and Rosbash, 1989; Michaud and Reed, 1993). Similar interactions are thought to occur in yeast and metazoans and involve mBBP associating with both U2AF65 and Fbp11 (Fbp11 is the human homolog of the yeast U1 snRNP protein, Prp40, and is a component of the MBP-purified E complex) (M. Bedford, R. D., R. R., and P. Leder, unpublished data; Abovich and Rosbash, 1997; Berglund et al., 1997, 1998). In light of our data, it is also possible that the 5⬘ and 3⬘ ends of metazoan introns are bridged by U1–U2 snRNP interactions. Finally, a network of interactions involving the U1 snRNP protein, U1 70K, SR proteins and U2AF35 may also bridge the 5⬘ and 3⬘ ends of the intron in metazoans (Wu and Maniatis, 1993) (Figure 7). During the ATP-dependent transition from the E to the A complex, U1 snRNP and U2AF become less tightly bound to the pre-mRNA (Figure 7) (Bennett et al., 1992; Michaud and Reed, 1993; Champion-Arnaud et al., 1995; Chiara et al., 1997). It is not known whether U1 snRNP and U2AF completely dissociate during A complex assembly or whether this occurs at a later step (ChampionArnaud et al., 1995; Chiara et al., 1997; Staley and Guthrie, 1999). mBBP is also thought to dissociate from the BPS to allow formation of the U2 snRNA-BPS duplex in the A complex (Figure 7) (Abovich and Rosbash, 1997). In the A complex, U2 snRNP becomes very tightly bound to the pre-mRNA, and SF3a/b subunits interact directly with the pre-mRNA surrounding the BPS (Gozani et al., 1996, 1998). In yeast, U2 snRNP undergoes an ATP-dependent conformational change that exposes the region of U2 snRNA that base pairs with the BPS (O’Day et al., 1996). This change requires the DEAD box

ATPase, Prp5 (O’Day et al., 1996). In metazoans, UAP56 was identified as a DEAD box protein that interacts directly with U2AF65 and is required for A complex assembly (Fleckner et al., 1997). Thus, Prp5 and UAP56 are both candidates for factors that may mediate an ATPdependent conformational change to allow tight binding of U2 snRNP to the BPS in the A complex (Figure 7). Several previous observations are consistent with the conclusion that U2 snRNP first associates with the spliceosome in the E complex. In particular, Query and coworkers (Query et al., 1997) detected ATP-independent binding of U2 snRNP on a minimal RNA substrate containing only the BPS and pyrimidine tract. In addition, Jamison and coworkers detected an ATP-independent complex containing U2 snRNP (Jamison and GarciaBlanco, 1992). In yeast, a mutation that disrupts stable base pairing between U1 snRNA and the 5⬘ splice site allows low levels of ATP-independent U2 snRNP binding (Liao et al., 1992). Whether U2 snRNP is normally associated with the commitment complex remains to be determined. Similarities between Assembly of the Major and Minor Spliceosomes Recently, a rare class of introns was identified that is spliced by the low-abundance minor spliceosome (Tarn and Steitz, 1997; Burge et al., 1998; Wu and Krainer, 1999). The minor spliceosome contains U11 and U12 snRNAs, which are the counterparts of U1 and U2 snRNAs, respectively (Tarn and Steitz, 1996a, 1996b; Hall and Padgett, 1996; Kolossova and Padgett, 1997). Significantly, U11 and U12 snRNPs largely exist in a di-snRNP particle in nuclear extracts (Kolossova and Padgett, 1997), and new studies indicate that U11 and U12 snRNPs bind simultaneously to the 5⬘ splice site and the BPS, respectively, during formation of the minor spliceosomal complex A (Frilander and Steitz, 1999). To date, it has not been possible to detect an ATPindependent minor spliceosomal complex analogous to the E complex (Frilander and Steitz, 1999). Nevertheless, our observation that U1 and U2 snRNPs are both present in the E complex, coupled with previous studies showing that U1 and U2 snRNPs are associated in extracts (Mattaj et al., 1986), raises the possibility that the initial steps in major and minor spliceosome assembly are more similar than previously thought. Indeed, like the subunits of the ribosome, assembly of the major and the minor spliceosome may involve two main steps, the binding of U1/U2 di-snRNP and the binding of U4/5/6 tri-snRNP. Experimental Procedures Plasmids The plasmid encoding wild-type AdML pre-mRNA was described in Michaud and Reed (1993). AdML-M3 pre-mRNA contains three phage R17 MS2 binding sites at the 3⬘ end (Z. Z. and R. R., unpublished data). AdML-M3⌬BPS was constructed from AdML-M3 and LUC pre-mRNA, which lacks the BPS (Champion-Arnaud et al., 1995). AdML and AdML-M3 were linearized with BamHI and XbaI, respectively, for transcription with T7 RNA polymerase. Isolation and Analysis of Functional Spliceosomal Complexes Purification of functional spliceosomal complexes was carried out as described (Z. Z. and R. R., unpublished data). An Adenovirus

Molecular Cell 786

major late pre-mRNA (AdML-M3), which contains three phage R17MS2 coat protein binding sites at the end of exon 2, was incubated with a fusion protein consisting of the MS2 coat protein and the MBP in a buffer containing 20 mM HEPES (pH 7.9), 60 mM NaCl. The MS2/MBP fusion protein was expressed in E. coli and purified by binding to amylose beads according to the manufacturer (NEB). The fusion protein and AdML-M3 pre-mRNA were incubated on ice for 20 min, and the binding was assayed on a 1.5% native agarose gel. Spliceosomes were assembled on the MS2/MBP/AdML-M3 complex using standard conditions and isolated by gel filtration (Bennett et al., 1992). Subsequently, the spliceosomes were affinity selected on amylose beads by rotating for 4 hr at 4⬚ and eluted with 12 mM maltose, 20 mM HEPES (pH 7.9), 60 mM NaCl, 10 mM ␤-mercaptoethanol, 1 mM PMSF. For assembly of the E and H complexes, nuclear extract was depleted of ATP, and the reactions lacked ATP and MgCl2 (Michaud and Reed, 1993) and were incubated at 30⬚C for 25 min. For A/B complex, pre-mRNA was incubated under standard splicing conditions for 10 min at 30⬚C. For Western analysis, total protein was prepared from equivalent amounts of each purified complex, separated by SDS-PAGE, and transferred to nitrocellulose. All rabbit antibodies were used at 1:1000 dilution. Tissue culture supernatant from the B⬘⬘ monoclonal antibody was used undiluted. Secondary antibodies were horseradish peroxidase linked, and the ECL detection system (Amersham) was used. For identification of snRNAs, total RNA was prepared from equivalent amounts of each purified complex and end labeled with [32P]pCp and RNA ligase. Native Gel Supershift Assay SF3a, hPrp16, and B⬘⬘ antibodies were purified by binding to protein A beads and eluted with Tris-glycine (pH 3). For the supershift assay of E and A/B complexes, splicing extracts (25 ␮l) were incubated for an additional 15 min at room temperature with 480 ng and 960 ng of purified SF3a or hPrp16 antibody. The purified B⬘⬘ antibody was used at 100, 200, 400, or 600 ng for supershift of the E complex. Complexes were analyzed on native agarose gels as described (Das and Reed, 1999). Immunodepletion and Reconstitution of SF3a Recombinant His-tagged SF3a was produced using a baculovirus expression system (GIBCO-BRL). SAPs 61, 62, and 114 were expressed separately initially. SF9 cells were then infected with the three viruses, and after 48 hr of infection, cells were harvested and lysed in 50 mM Tris-HCL (pH 8.5), 10 mM 2-mercaptoethanol, 1 mM PMSF, and 1% Triton X-100 at 4⬚C. The SF3a complex was purified on nickel agarose (Qiagen). Rabbit polyclonal antibodies were raised against the recombinant SF3a complex (Covance Research Products, Denver, PA). Immunodepletion of SF3a was carried out according to Zhou and Reed (1998). For reconstitution with recombinant SF3a, 60–120 ng rSF3a were added to 7.5 ␮l of SF3a-depleted extracts in a 25 ␮l splicing reaction. Acknowledgments We are grateful to Rajesh Gaur for U2AF-depleted extracts. We also thank M.-J. Luo, B. Das, and K. N. Clouse for useful discussions and comments on the manuscript. This work was supported by an NIH grant to R. R. Received January 25, 2000; revised March 2, 2000. References Abovich, N., and Rosbash, M. (1997). Cross-intron bridging interactions in the yeast commitment complex are conserved in mammals. Cell 89, 403–412. Arning, S., Gruter, P., Bilbe, G., and Kramer, A. (1996). Mammalian splicing factor SF1 is encoded by variant cDNAs and binds to RNA. RNA 2, 794–810. Behrens, S.E., Galisson, F., Legrain, P., and Luhrmann, R. (1993a). Evidence that the 60-kDa protein of 17S U2 small nuclear ribonucleoprotein is immunologically and functionally related to the

yeast PRP9 splicing factor and is required for the efficient formation of prespliceosomes. Proc. Natl. Acad. Sci. USA 90, 8229–8233. Behrens, S.E., Tyc, K., Kastner, B., Reichelt, J., and Luhrmann, R. (1993b). Small nuclear ribonucleoprotein (RNP) U2 contains numerous additional proteins and has a bipartite RNP structure under splicing conditions. Mol. Cell. Biol. 13, 307–319. Bennett, M., Michaud, S., Kingston, J., and Reed, R. (1992). Protein components specifically associated with prespliceosome and spliceosome complexes. Genes Dev. 6, 1986–2000. Berglund, J.A., Chua, K., Abovich, N., Reed, R., and Rosbash, M. (1997). The splicing factor BBP interacts specifically with the premRNA branchpoint sequence UACUAAC. Cell 89, 781–787. Berglund, J.A., Abovich, N., and Rosbash, M. (1998). A cooperative interaction between U2AF65 and mBBP/SF1 facilitates branchpoint region recognition. Genes Dev. 12, 858–867. Brosi, R., Groning, K., Behrens, S.E., Luhrmann, R., and Kramer, A. (1993a). Interaction of mammalian splicing factor SF3a with U2 snRNP and relation of its 60-kD subunit to yeast PRP9. Science 262, 102–105. Brosi, R., Hauri, H.P., and Kramer, A. (1993b). Separation of splicing factor SF3 into two components and purification of SF3a activity. J. Biol. Chem. 268, 17640–17646. Burge, C.B., Tuschl, T.H., and Sharp, P.A. (1998). Splicing of precursors to mRNAs by the spliceosomes. In The RNA World, Second Edition, R.F. Gesteland, T.R. Cech, and J.F. Atkins, eds. (Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press), pp. 525–560. Chabot, B., Black, D.L., LeMaster, D.M., and Steitz, J.A. (1985). The 3⬘ splice site of pre-messenger RNA is recognized by a small nuclear ribonucleoprotein. Science 230, 1344–1349. Champion-Arnaud, P., Gozani, O., Palandjian, L., and Reed, R. (1995). Accumulation of a novel spliceosomal complex on premRNAs containing branch site mutations. Mol. Cell. Biol. 15, 5750– 5756. Chiara, M., Gozani, O., Bennett, M., Champion-Arnaud, P., Palandjian, L., and Reed, R. (1996). Identification of proteins that interact with exon sequences, splice sites, and the branchpoint sequence during each stage of spliceosome assembly. Mol. Cell. Biol. 16, 3317–3326. Chiara, M.D., Palandjian, L., Feld Kramer, R., and Reed, R. (1997). Evidence that U5 snRNP recognizes the 3⬘ splice site for catalytic step II in mammals. EMBO J. 16, 4746–4759. Das, R., and Reed, R. (1999). Resolution of the mammalian E complex and the ATP-dependent spliceosomal complexes on native agarose mini-gels. RNA 5, 1504–1508. Das, B.K., Xia, L., Palandjian, L., Gozani, O., Chyung, Y., and Reed, R. (1999). Characterization of a protein complex containing spliceosomal proteins SAPs 49, 130, 145, and 155. Mol. Cell. Biol. 19, 6796–6802. Fleckner, J., Zhang, M., Valcarcel, J., and Green, M.R. (1997). U2AF65 recruits a novel human DEAD box protein required for the U2 snRNP–branchpoint interaction. Genes Dev. 11, 1864–1872. Frilander, M.J., and Steitz, J.A. (1999). Initial recognition of U12dependent introns requires both U11/5⬘ splice-site and U12/branchpoint interactions. Genes Dev. 13, 851–863. Gozani, O., Patton, J.G., and Reed, R. (1994). A novel set of spliceosome-associated proteins and the essential splicing factor PSF bind stably to pre-mRNA prior to catalytic step II of the splicing reaction. EMBO J. 13, 3356–3367. Gozani, O., Feld, R., and Reed, R. (1996). Evidence that sequenceindependent binding of highly conserved U2 snRNP proteins upstream of the branch site is required for assembly of spliceosomal complex A. Genes Dev. 10, 233–243. Gozani, O., Potashkin, J., and Reed, R. (1998). A potential role for U2AF-SAP 155 interactions in recruiting U2 snRNP to the branch site. Mol. Cell. Biol. 18, 4752–4760. Grabowski, P.J., and Sharp, P.A. (1986). Affinity chromatography of splicing complexes: U2, U5, and U4 ⫹ U6 small nuclear ribonucleoprotein particles in the spliceosome. Science 233, 1294–1299. Hall, S.L., and Padgett, R.A. (1996). Requirement of U12 snRNA for

Functional Association of U2 snRNP with E Complex 787

in vivo splicing of a minor class of eukaryotic nuclear pre-mRNA introns. Science 271, 1716–1718. Hong, W., Bennett, M., Xiao, Y., Feld Kramer, R., Wang, C., and Reed, R. (1997). Association of U2 snRNP with the spliceosomal complex E. Nucleic Acids Res. 25, 354–361. Jamison, S.F., and Garcia-Blanco, M.A. (1992). An ATP-independent U2 small nuclear ribonucleoprotein particle/precursor mRNA complex requires both splice sites and the polypyrimidine tract. Proc. Natl. Acad. Sci. USA 89, 5482–5486. Jamison, S.F., Crow, A., and Garcia-Blanco, M.A. (1992). The spliceosome assembly pathway in mammalian extracts. Mol. Cell. Biol. 12, 4279–4287. Kohtz, J.D., Jamison, S.F., Will, C.L., Zuo, P., Luhrmann, R., GarciaBlanco, M.A., and Manley, J.L. (1994). Protein–protein interactions and 5⬘-splice-site recognition in mammalian mRNA precursors. Nature 368, 119–124. Kolossova, I., and Padgett, R.A. (1997). U11 snRNA interacts in vivo with the 5⬘ splice site of U12-dependent (AU-AC) pre-mRNA introns. RNA 3, 227–233. Konarska, M.M., and Sharp, P.A. (1986). Electrophoretic separation of complexes involved in the splicing of precursors to mRNAs. Cell 46, 845–855. Konarska, M.M., and Sharp, P.A. (1987). Interactions between small nuclear ribonucleoprotein particles in formation of spliceosomes. Cell 49, 763–774. Kramer, A. (1992). Purification of splicing factor SF1, a heat-stable protein that functions in the assembly of a presplicing complex. Mol. Cell. Biol. 12, 4545–4552. Kramer, A., Gruter, P., Groning, K., and Kastner, B. (1999). Combined biochemical and electron microscopic analyses reveal the architecture of the mammalian U2 snRNP. J. Cell. Biol. 145, 1355–1368. Liao, X.C., Colot, H.V., Wang, Y., and Rosbash, M. (1992). Requirements for U2 snRNP addition to yeast pre-mRNA. Nucleic Acids Res. 20, 4237–4245. Mattaj, I.W., Habets, W.J., and van Vernon, W.J. (1986). Monospecific antibodies reveal details of U2 snRNP structure and interaction between U1 and U2 snRNPs. EMBO J. 5, 997–1002. Merendino, L., Guth, S., Bilbao, D., Martinez, C., and Valcarcel, J. (1999). Inhibition of msl-2 splicing by sex lethal reveals functional interaction between U2AF35 and the 3⬘ splice site AG. Nature 402, 838–841. Michaud, S., and Reed, R. (1991). An ATP-independent complex commits pre-mRNA to the mammalian spliceosome assembly pathway. Genes Dev. 5, 2534–2546. Michaud, S., and Reed, R. (1993). A functional association between the 5⬘ and 3⬘ splice site is established in the earliest prespliceosome complex (E) in mammals. Genes Dev. 7, 1008–1020. O’Day, C.L., Dalbadie-McFarland, G., and Abelson, J. (1996). The Saccharomyces cerevisiae Prp5 protein has RNA-dependent ATPase activity with specificity for U2 small nuclear RNA. J. Biol. Chem. 271, 33261–33267. Query, C.C., Strobel, S.A., and Sharp, P.A. (1996). Three recognition events at the branch-site adenine. EMBO J. 15, 1392–1402. Query, C.C., McCaw, P.S., and Sharp, P.A. (1997). A minimal spliceosomal complex A recognizes the branch site and polypyrimidine tract. Mol. Cell. Biol. 17, 2944–2953. Reed, R. (2000). Mechanisms of fidelity during pre-mRNA splicing. Curr. Opin. Cell Biol. 12, in press. Ruskin, B., Zamore, P.D., and Green, M.R. (1988). A factor, U2AF, is required for U2 snRNP binding and splicing complex assembly. Cell 52, 207–219. Seraphin, B., and Rosbash, M. (1989). Identification of functional U1 snRNA-pre-mRNA complexes committed to spliceosome assembly and splicing. Cell 59, 349–358. Staknis, D., and Reed, R. (1994a). SR proteins promote the first specific recognition of Pre-mRNA and are present together with the U1 small nuclear ribonucleoprotein particle in a general splicing enhancer complex. Mol. Cell. Biol. 14, 7670–7682.

Staknis, D., and Reed, R. (1994b). Direct interactions between premRNA and six U2 small nuclear ribonucleoproteins during spliceosome assembly. Mol. Cell. Biol. 14, 2994–3005. Staley, J.P., and Guthrie, C. (1998). Mechanical devices of the spliceosome: motors, clocks, springs, and things. Cell 92, 315–326. Staley, J.P., and Guthrie, C. (1999). An RNA switch at the 5⬘ splice site requires ATP and the DEAD box protein Prp28p. Mol. Cell 3, 55–64. Tarn, W.Y., and Steitz, J.A. (1996a). Highly diverged U4 and U6 small nuclear RNAs required for splicing rare AT-AC introns. Science 273, 1824–1832. Tarn, W.Y., and Steitz, J.A. (1996b). A novel spliceosome containing U11, U12, and U5 snRNPs excises a minor class (AT-AC) intron in vitro. Cell 84, 801–811. Tarn, W.Y., and Steitz, J.A. (1997). Pre-mRNA splicing: the discovery of a new spliceosome doubles the challenge. Trends Biochem. Sci. 22, 132–137. Wu, Q., and Krainer, A.R. (1999). AT-AC pre-mRNA splicing mechanisms and conservation of minor introns in voltage-gated ion channel genes. Mol. Cell. Biol. 19, 3225–3236. Wu, J.Y., and Maniatis, T. (1993). Specific interactions between proteins implicated in splice site selection and regulated alternative splicing. Cell 75, 1061–1070. Wu, S., Romfo, C.M., Nilsen, T.W., and Green, M.R. (1999). Functional recognition of the 3⬘ splice site AG by the splicing factor U2AF35. Nature 402, 832–835. Zamore, P.D., Patton, J.G., and Green, M.R. (1992). Cloning and domain structure of the mammalian splicing factor U2AF. Nature 355, 609–614. Zhang, D., and Rosbash, M. (1999). Identification of eight proteins that cross-link to pre-mRNA in the yeast commitment complex. Genes Dev. 13, 581–592. Zhou, Z., and Reed, R. (1998). Human homologs of yeast prp16 and prp17 reveal conservation of the mechanism for catalytic step II of pre-mRNA splicing. EMBO J. 17, 2095–2106. Zillmann, M., Zapp, M.L., and Berget, S.M. (1988). Gel electrophoretic isolation of splicing complexes containing U1 small nuclear ribonucleoprotein particles. Mol. Cell. Biol. 8, 814–821. Zorio, D.A., and Blumenthal, T. (1999). Both subunits of U2AF recognize the 3⬘ splice site in Caenorhabditis elegans. Nature 402, 835–838. Zuo, P., and Maniatis, T. (1996). The splicing factor U2AF35 mediates critical protein-protein interactions in constitutive and enhancerdependent splicing. Genes Dev. 10, 1356–1368.