Nature Structural Biology
9, 5 - 7 (2002)
doi:10.1038/nsb0102-5
Hanging on to the branchJonathan P. StaleyJonathan P. Staley is in the Department of Molecular Genetics and Cell Biology, University of Chicago, 920 E. 58th Street, Chicago, IL 60637, U.S.A. jstaley@uchicago.edu The mechanism for a critical step in intron recognition has recently been revealed by the structure of the RNA binding domain of splicing factor 1 bound to intronic RNA.The splicing machinery must accurately recognize genuine intronic substrates or risk catastrophic consequences for the cell. A key determinant of an intron is an adenosine whose 2' hydroxyl serves as the nucleophile that cleaves the 5' exon from the intron1. This adenosine, termed the 'branch site' because of the branched structure that results from 5' splice site cleavage, is defined in part by a conserved, flanking branch site sequence (YNCURAY in mammals; UACUAAC in yeast). The exceptional importance of branch site recognition is highlighted by its extraordinarily dynamic nature — the branch site sequence is bound successively by at least two factors2 and the branch site adenosine by at least three factors3. An outstanding goal in the field has been to determine the mechanism for recognizing the branch site adenosine and its flanking sequence. In the first recognition step, splicing factor 1 (SF1, Ref. 4), also called branch point binding protein (BBP, Ref. 2), binds to this sequence through its STAR (signal transduction and activation of RNA; Ref. 5) domain6,
7. In a recent paper in Science, Sattler, Krämer and coworkers8 reveal by NMR the structural basis for recognition of the branch site sequence by SF1, providing key insights into the mechanism of intron recognition. Their work also provides the first view of the RNA binding mechanism for a STAR member of the ubiquitous KH (hnRNP K homology) family9 of RNA binding proteins.
The branch site sequence is recognized not only by proteins but also by RNA. The spliceosome, the machine that executes intron excision, is composed of five small nuclear RNAs (snRNAs) in addition to >70 proteins1. Similar to the ribosome, the RNA components of the spliceosome play key roles in substrate recognition and catalysis. Because the branch site sequence is complementary to a conserved region of U2 snRNA, U2 was readily identified as the first candidate for a factor that recognizes the branch site sequence. The functional importance of this complementarity was easily confirmed genetically through a compensatory analysis10,
11. Importantly, the nucleophilic branch site adenosine does not base pair with U2 but rather bulges out of the recognition helix12, indicating that some other factor(s) recognize this critical nucleotide. Moreover, the branch site sequence is required for spliceosome assembly even before U2 is required, further implicating an additional factor in branch site sequence recognition13,
14. Through a clever genetic screen, Rosbash and colleagues2,
15 identified the yeast orthologue of SF1 as the early-acting factor that recognizes both the branch site sequence and the branch site adenosine.
Getting a grip on the branch
Whittling away the domains of SF1 has revealed that its STAR domain is necessary for specific binding to the branch site sequence6,
7, although a zinc knuckle domain contributes to the stability of RNA binding7. The STAR members5 of the KH family are characterized by a single, large KH domain, termed 'maxi-KH' domain, flanked by an N-terminal extension termed QUA1 (Quaking homology), and a C-terminal extension, termed QUA2. In their recent paper, Liu et al.8 show that the KH-QUA2 region of the SF1 STAR domain is sufficient for binding to the branch site sequence. Their structure of the KH-QUA2 region bound to RNA explores the extent to which a maxi-KH domain resembles a typical KH domain in binding RNA, the structural role the QUA2 domain plays in binding RNA and the mechanism by which SF1 recognizes the branch site sequence — particularly the branch site adenosine.
As expected from sequence conservation, the maxi-KH domain, consisting of a three-stranded  -sheet packed against three  -helices, resembles other KH domains16,
17,
18,
19, except for an enlarged variable loop (Fig. 1). As suggested by mutagenesis20, the maxi-KH domain of SF1 (Ref. 8) and the more typical KH domain, Nova-2 KH3 (Ref. 21), bind RNA in a similar manner (Fig. 1). Two loops, the enlarged variable loop and a conserved GXXG loop, function as a 'molecular vise' to grip the branch site sequence. An exclusively aliphatic / surface, devoid of aromatic side chains, also interacts with the branch site sequence serving as a platform for the vise. Notably, this aliphatic mode of RNA binding is distinct from the aromatic mode of RNA binding, involving stacking interactions, characteristic of the pervasive RNA recognition motif9. The similarity in RNA binding by the KH domains of SF1 and Nova-2 suggests that this aliphatic mechanism of RNA binding extends to all members of the KH family.
 | | |
The structure by Liu et al.8 reveals that the QUA2 domain binds RNA directly and increases the RNA binding capacity of the maxi-KH domain from four nucleotides to six (Fig. 1). The QUA2 domain packs against the maxi-KH domain and extends the RNA binding surface at the 5' end of the substrate. In particular, an amphipathic helix appended to the C-terminus of the maxi-KH domain plays a prominent role in contacting RNA. For example, two leucines, Leu 244 and Leu 247, on the hydrophobic face of the helix form extensive contacts with the CU dinucleotide of the branch site sequence. Consistent with a role for these leucines in binding RNA, mutation to alanine disrupts RNA binding8. Thus, the QUA2 domain not only extends the RNA binding surface of the maxi-KH domain but also reiterates the theme of RNA recognition by aliphatic residues. It will be important to determine if unrelated KH proteins also extend their RNA binding surfaces.
As if paying homage to the spliceosomal snRNAs, SF1 mimics RNA in the manner in which it recognizes the branch site adenosine (Fig. 2). As for the equivalent adenosine in the Nova-2 KH3 protein− RNA complex21, the N1 and N6 of the Watson-Crick face of the branch site adenosine form hydrogen bonds with the backbone carbonyl oxygen and amide of Ile 177, mimicking an interaction with uridine8. An intrastrand interaction between the 2' hydroxyl of the 5' nucleotide and the N7 position of the adenosine completes the definition of an adenosine-specific binding pocket. Interactions with the aliphatic platform and stacking interactions with the 3' base further stabilize SF1 binding to the adenosine (Fig. 2). Consistent with these extensive interactions between SF1 and the branch site adenosine, mutation of this adenosine decreases SF1 binding severely15. Thus, whereas U2 overlooks the branch site adenosine, SF1 clearly recognizes its significance.
| | |
SF1 binds an intron cooperatively with the heterodimeric splicing factor U2AF6,
20,
22. U2AF, an RNA recognition motif-containing protein, binds downstream from the branch site sequence to sequences that contribute to branch site recognition. The cooperativity between SF1 and U2AF arises in part from an interaction between SF1 and the third RNA recognition motif of the U2AF65 subunit6,
22. The structure of SF1 bound to RNA8 suggests an additional contribution to cooperativity. Specifically, the negatively charged phosphate backbone of the branch site sequence remains exposed to solvent, allowing simultaneous interaction with the positively charged N-terminus of U2AF65, which has been shown previously to interact with the branch site sequence23. An important goal is to investigate the structural basis for this cooperativity and to determine the generality for these mechanisms in stabilizing the binding of KH domains to RNA.
Letting go of the branch
Surprisingly, after SF1 binds to the branch site sequence, it must subsequently be displaced to permit splicing. First, the binding of SF1 to the branch site sequence is mutually exclusive with the binding of U2 (refs 22,24), the complementary spliceosomal snRNA. Second, burial of the branch site adenosine within the core of SF1 forbids access of this key nucleophile to the 5' splice site and the catalytic core of the spliceosome. How is the SF1−RNA complex disrupted?
The cooperativity between SF1 and U2AF provides a clue. Splicing factors belonging to the DExD/H-box family of ATPases have long been implicated in remodeling double-stranded RNA in the spliceosome (for review, see ref. 25), but recently they have also been implicated in remodeling protein−RNA complexes26,
27,
28. Notably, the DExD/H-box protein Sub2p has been implicated in displacing Mud2p, the yeast ortholog of U2AF65, from RNA; deletion of the MUD2 gene in vivo bypasses a requirement for the SUB2 gene27. Consistent with this observation, the human ortholog of Sub2p, UAP56, binds to U2AF65 (ref. 29). Thus, UAP56 could displace SF1 from the branch site sequence indirectly by displacing U2AF from downstream sequence elements. The displacement of SF1 and U2AF from an intron promises to serve as a powerful and biologically significant model system for investigating the role for DExD/H box proteins in displacing protein from single-stranded RNA.
Given that U2 replaces SF1 at the branch site sequence, one must wonder what role SF1 plays in splicing. Liu et al.8 propose that SF1 promotes formation of the bulged U2-branch site helix by positioning the branch site adenosine in a 'prebulged' configuration. Interestingly, the spliceosome recognizes not only the branch site sequence in several stages but also the 5' splice site and 3' splice site sequences in several stages. These multiple recognition events may reflect inspections of intron sequence determinants to enhance the specificity of splicing and to ensure the integrity of the spliced message (for example, ref. 30). Elucidating the rationale behind the dynamic design of intron recognition remains a fundamental goal in understanding spliceosome activation.
REFERENCES
- Hastings, M.L. & Krainer, A.R. Curr. Opin. Cell Biol. 13, 302−309 (2001). | Article | PubMed | ISI | ChemPort |
- Abovich, N. & Rosbash, M. Cell 89, 403−412 (1997). | Article | PubMed | ISI | ChemPort |
- Query, C.C., Strobel, S.A. & Sharp, P.A. EMBO J. 15, 1392−1402 (1996). | PubMed | ISI | ChemPort |
- Arning, S., Gruter, P., Bilbe, G. & Krämer, A. RNA 2, 794−810 (1996). | PubMed | ISI | ChemPort |
- Vernet, C. & Artzt, K. Trends Genet. 13, 479−484 (1997). | Article | PubMed | ISI | ChemPort |
- Rain, J.C., Rafi, Z., Rhani, Z., Legrain, P. & Krämer, A. RNA 4, 551−565 (1998). | Article | PubMed | ISI | ChemPort |
- Berglund, J.A., Fleming, M.L. & Rosbash, M. RNA 4, 998−1006 (1998). | Article | PubMed | ISI | ChemPort |
- Liu, Z. et al. Science 294, 1098−1102 (2001). | Article | PubMed | ISI | ChemPort |
- Nagai, K. Curr. Opin. Struct. Biol. 6, 53−61 (1996). | Article | PubMed | ISI | ChemPort |
- Parker, R., Siliciano, P.G. & Guthrie, C. Cell 49, 229−239 (1987). | Article | PubMed | ISI | ChemPort |
- Zhuang, Y. & Weiner, A.M. Genes Dev. 3, 1545−1552 (1989). | PubMed | ISI | ChemPort |
- Query, C.C., Moore, M.J. & Sharp, P.A. Genes Dev. 8, 587−597 (1994). | PubMed | ISI | ChemPort |
- Ruby, S.W. & Abelson, J. Science 242, 1028−1034 (1988). | PubMed | ISI | ChemPort |
- Legrain, P., Seraphin, B. & Rosbash, M. Mol. Cell. Biol. 8, 3755−3760 (1988). | PubMed | ISI | ChemPort |
- Berglund, J.A., Chua, K., Abovich, N., Reed, R. & Rosbash, M. Cell 89, 781−787 (1997). | Article | PubMed | ISI | ChemPort |
- Musco, G. et al. Cell 85, 237−245 (1996). | Article | PubMed | ISI | ChemPort |
- Musco, G. et al. Nature Struct. Biol. 4, 712−716 (1997). | Article | PubMed | ISI | ChemPort |
- Baber, J.L., Libutti, D., Levens, D. & Tjandra, N. J. Mol. Biol. 289, 949−962 (1999). | Article | PubMed | ISI | ChemPort |
- Lewis, H.A. et al. Structure Fold. Des. 7, 191−203 (1999). | Article | PubMed | ISI | ChemPort |
- Peled-Zehavi, H., Berglund, J.A., Rosbash, M. & Frankel, A.D. Mol. Cell. Biol. 21, 5232−5241 (2001). | Article | PubMed | ISI | ChemPort |
- Lewis, H.A. et al. Cell 100, 323−332 (2000). | Article | PubMed | ISI | ChemPort |
- Berglund, J.A., Abovich, N. & Rosbash, M. Genes Dev. 12, 858−867 (1998). | PubMed | ISI | ChemPort |
- Valcárcel, J., Gaur, R.K., Singh, R. & Green, M.R. Science 273, 1706−1709 (1996). | PubMed | ISI | ChemPort |
- Rutz, B. & Seraphin, B. RNA 5, 819−831 (1999). | Article | PubMed | ISI | ChemPort |
- Staley, J.P. & Guthrie, C. Cell 92, 315−326 (1998). | Article | PubMed | ISI | ChemPort |
- Jankowsky, E., Gross, C.H., Shuman, S. & Pyle, A.M. Science 291, 121−125 (2001). | Article | PubMed | ISI | ChemPort |
- Kistler, A.L. & Guthrie, C. Genes Dev. 15, 42−49 (2001). | Article | PubMed | ISI | ChemPort |
- Schwer, B. Nature Struct. Biol. 8, 113−116 (2001). | Article | PubMed | ISI | ChemPort |
- Fleckner, J., Zhang, M., Valcárcel, J. & Green, M.R. Genes Dev. 11, 1864−1872 (1997). | PubMed | ISI | ChemPort |
- Rutz, B. & Seraphin, B. EMBO J. 19, 1873−1886 (2000). | Article | PubMed | ChemPort |
Acknowledgments
I thank P.H. Hutchinson for preparing the figures.
|
