U6 snRNA Secondary Structure in Free U6 snRNPs Elizabeth A. Dunn B. Sc., University Of Northern British Columbia Thesis Submitted In Partial Fulfillment Of The Requirements For The Degree Of Master Of Science In Mathematical, Computer, and Physical Sciences (Chemistry) The University of Northern British Columbia July 2009 © Elizabeth A. Dunn, 2009 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1*1 Library and Archives Canada Bibliothgque et Archives Canada Published Heritage Branch Direction du Patrimoine de l'6dition 395 Wellington Street Ottawa ON K1A0N4 Canada 395, rue Wellington Ottawa ON K1A 0N4 Canada Your file Votre reference ISBN: 978-0-494-60837-1 Our file Notre inference ISBN: 978-0-494-60837-1 NOTICE: AVIS: The author has granted a non­ exclusive license allowing Library and Archives Canada to reproduce, publish, archive, preserve, conserve, communicate to the public by telecommunication or on the Internet, loan, distribute and sell theses worldwide, for commercial or non­ commercial purposes, in microform, paper, electronic and/or any other formats. L'auteur a accorde une licence non exclusive permettant a la Bibliotheque et Archives Canada de reproduire, publier, archiver, sauvegarder, conserver, transmettre au public par telecommunication ou par I'lnternet, preter, distribuer et vendre des theses partout dans le monde, a des fins commerciales ou autres, sur support microforme, papier, electronique et/ou autres formats. The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission. L'auteur conserve la propriete du droit d'auteur et des droits moraux qui protege cette these. Ni la these ni des extraits substantiels de celle-ci ne doivent etre imprimes ou autrement reproduits sans son autorisation. In compliance with the Canadian Privacy Act some supporting forms may have been removed from this thesis. Conformement a la loi canadienne sur la protection de la vie privee, quelques formulaires secondaires ont ete enleves de cette these. While these forms may be included in the document page count, their removal does not represent any loss of content from the thesis. Bien que ces formulaires aient inclus dans la pagination, il n'y aura aucun contenu manquant. Canada Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. LAC Condensed Abstract - Elizabeth A. Dunn A critical step in spliceosome assembly is the formation of the U4/U6 di-snRNP particle in which U4 and U6 snRNAs are engaged in an extensive intermolecular base pairing interaction. Little is known about how this particle is formed or its role in pre-mRNA splicing, and current models of U6 snRNA secondary structure in free U6 snRNP do not offer insight into these questions. We have generated a new model of U6 snRNA secondary structure in free U6 snRNPs that is consistent with existing structural and genetic data and we show support for this model through genetic analyses and oligonucleotide accessibility experiments. Our model predicts that catalytically important elements are unavailable for interaction with the pre-mRNA transcript until a large structural re-arrangement exposes these elements. We propose that U4 snRNA is responsible for promoting this re-arrangement through the establishment of an intermolecular interaction between U4 and U6. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. TABLE OF CONTENTS Abstract ii Table of Contents iii List of Tables v List of Figures vi Acknowledgement viii Chapter One - Introduction 1 1.1 Nuclear Pre-mRNA Splicing 1.2 Spliceosome Assembly 1.3 U6 snRNA 1.4 U6 snRNA in Free U6 snRNP 1.5 Motivation for a New Model of U6 snRNA Secondary Structure Chapter Two - Introduction to the Dunn-Rader Model of U6 snRNA Structure in Free U6 snRNP 2.1 The Dunn-Rader Model of U6 snRNA Secondary Structure 2.2 Three Dimensional Modeling of U6 snRNA 2.2.1 Methods 2.2.2 U6 snRNA in Three Dimensions 2.3 Support for the Dunn-Rader Model from the Literature 2.3.1 Chemical Modification of Accessible U6 snRNA Bases 2.3.2 Compensatory Base Mutation 2.3.3 The U6 snRNA 3' Intramolecular Stem/loop (3' ISL) 2.3.4 Genetic Suppression Analysis - A62G and A79G 2.3.5 Exonuclease Degradation of U6 snRNA Chapter Three - Structure Probing by Oligonucleotide Accessibility 3.1 Materials and Methods 3.2 Results and Discussion 1 3 5 9 12 13 13 15 15 15 18 19 24 26 28 31 33 33 36 Chapter Four - Genetic Support for the Lower Central Stem 4.1 Materials and Methods 4.2 Results and Discussion 46 46 50 iii Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter Five - Co-variation of U6 snRNA Secondary Structure and Prp24 Homolog Architecture 5.1 Materials and Methods 5.2 Phylogenetic Conservation of U6 snRNA Secondary Structure 5.3 Evolutionary Relationship Based on U6 snRNA Sequence and Secondary Structure 5.4 Subtle Changes in U6 snRNA Sequence Correlate with Dramatic Changes in Prp24 Homolog Architecture 58 59 60 69 72 Chapter Six - The Dunn-Rader Model of U6 snRNA Secondary Structure Reveals the Primary Function of U4 snRNA in Pre-mRNA Splicing 77 Chapter Seven - Future Directions and Concluding Remarks 82 7.1 U6 snRNP Structure Determination 7.2 Genetic Studies 7.3 Structure Probing 7.4 U6 snRNA Structural Re-arrangements in Pre-mRNA Splicing 7.5 Concluding Remarks Works Cited 82 83 84 85 86 ix iv Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. LIST OF TABLES Table 1: DNA Oligonucleotide Sequence, GC Content, and Melting Temperature. Table 2: Putative Prp24 Homologs. V Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. LIST OF FIGURES Figure 1. Nuclear pre-mRNA splicing is catalyzed by the spliceosome and is achieved through two transesterification reactions. 2 Figure 2. U6 snRNA conformational re-arrangements throughout the splicing cycle. 7 Figure 3. The Dunn-Rader model of U6 snRNA structure in the free U6 snRNP. 14 Figure 4. U6 snRNA can fold into a closed and open conformation without disrupting the base pairing interactions proposed in the Dunn-Rader model. 17 Figure 5. 2D TOCSY NMR spectrum of in vitro transcribed U6 snRNA. 18 Figure 6. Chemical modification of naked U6 snRNA and U6 snRNA in U6 snRNPs. 22 Figure 7. The Karaduman et al. (2006) chemical modification data for in vitro transcribed U6 snRNA and U6 snRNPs mapped to the proposed closed and open Dunn-Rader models. 23 Figure 8. A62G and A79G suppressor mutations mapped onto U6 snRNA secondary structures in the free U6 snRNP, di-snRNP, and active spliceosome. 29 Figure 9. U6 snRNA secondary structure probing by oligonucleotide accessibility. 37 Figure 10. Construction of U6 snRNA central stem double mutants. 52 Figure 11. Growth phenotypes and U4/U6 assembly phenotypes of the central stem double mutants, a wild type control, and a control for cold sensitivity, U6-A62G. 55 Figure 12. U6 snRNA and U6atac sequence alignment. 62 Figure 13. S. cerevisiae and H. sapiens stem/loops A and B can fold into similar three-dimensional structures. 69 Figure 14. Subtle changes in U6 snRNA sequence are accompanied by substantial changes in Prp24 homolog architecture. 71 vi Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 15. U6 snRNA is tightly regulated through riboswitch-like conformational changes during spliceosome assembly and activation. vii Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ACKNOWLEDGEMENT First and foremost I would like to thank my supervisor Dr. Stephen Rader for introducing me to the world of research and providing mentorship and support throughout the duration of my graduate work. I would also like to thank my graduate thesis committee members Dr. Andrea Gorrell, Dr. Daniel Erasmus, and Dr. Saif Zahir for helpful discussions and project direction. Special thanks to Dr. Phillip Johnson for collaborating with us for the NMR experiments and Dr. Fred Damberger for helping us with the interpretation of the NMR data. I would like to thank all past and present members of the Rader Lab for helpful discussions and guidance. There are many others outside of academics that I would like to extend a special thanks to: William Dunn for his support, understanding and encouragement over the last seven years, and my parents, Rick and Kim Chester who have provided encouragement and support from a very young age. I would also like to thank my sisters, Adrianne and Jen, and brother, Chris, for their understanding and entertainment over the years. This work was funded through an NSERC Discovery Grant awarded to Dr. Stephen Rader and UNBC administered research project awards awarded to Elizabeth Dunn. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter One - Introduction A critical step in eukaryotic gene expression is the processing of precursor messenger RNA (pre-mRNA) transcripts to generate mature mRNAs that can subsequently be translated into the appropriate polypeptide chain. One such processing event, known as nuclear pre-mRNA splicing, involves the removal of non-coding RNA and the subsequent joining of the remaining regions of coding RNA. This complex process is catalyzed by the spliceosome, a large ensemble of over 100 proteins and five small nuclear RNAs (snRNA). Although splicing has been studied extensively over the past three decades, very little is known about the function of many of these essential splicing factors. U6 snRNA in particular has been studied considerably and although experimental data suggest that U6 plays a central role throughout the splicing reactions, the mechanism by which it may do so has yet to be elucidated. The major goal of this thesis was to develop a model of the structure of U6 snRNA in its inactive free small nuclear ribonucleoprotein (snRNP) particle as a step toward understanding how U6 becomes activated for splicing. 1.1 Nuclear Pre-mRNA Splicing Splicing of pre-mRNA transcripts was first proposed following the finding that a mature mRNA transcript produced in cells infected by Adenovirus 2 was generated through the joining of several RNA fragments (Berget et al. 1977). Since this discovery, a general mechanism for the excision of intervening sequences (introns) and joining of coding sequences (exons) has been developed (Fig. 1). In the first of two transesterification reactions, the 2' hydroxyl of a bulged adenosine, located at the branch 1 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. I 3' Splice Site I I Mature mRNA + oA, oA_ 5' Splice Site Lariat Intron B C@-@CZ] Pre-mRNA Pre-spliceosome ©@ U4JJJ5 U6 U4IU5 C@-(u2)czi (U6 Recycled snRNPs U61U2 Activated Spliceosome I I I Mature mRNA + oA_ Lariat Intron Figure 1. Nuclear pre-mRNA splicing is catalyzed by the spliceosome and is achieved through two transesterification reactions. The splicing reaction (A) and stepwise spliceosome assembly (B) are shown. Exons are represented by open boxes and the intron is shown with a line containing the branch point adenosine. site in the pre-mRNA intron, reacts with the 5' splice site (Padgett et al. 1984, Konarska et al. 1985). This results in the liberation of the 5' exon concomitantly with the formation of a 2'-5' phosphodiester linkage, which joins the 5' end of the intron to the branch residue (Padgett et al. 1984, Konarska et al. 1985). In the second reaction, the 3' hydroxyl 2 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. of the 5' exon reacts with the 3' splice site, cleaving away the lariat intron and joining the 5' and 3' exons through a 3'-5' phosphodiester linkage (Padgett et al. 1984, Konarska et al. 1985). The inversion of stereochemistry at the chiral phosphate in each step suggests that both of these transesterifications occur through an in line SN2 nucleophilic reaction (Maschhoff & Padgett 1993, Moore & Sharp 1993). 1.2 Spliceosome Assembly Despite the simplicity of the splicing reactions, recognition of the 5' and 3' splice sites and the branch point adenosine to achieve high fidelity splicing is a complex process. While the splice sites and branch site in each pre-mRNA transcript are defined by consensus sequences, these sequences are short and in some cases quite degenerate (Irimia et al. 2007). The length of intervening sequences from one transcript to another also varies considerably, and the presence of alternative splice sites in more complex eukaryotic organisms complicates the process further (Bon et al. 2003, Irimia et al. 2007). Consequently the splicing reactions have been proposed to be facilitated by a massive, and highly dynamic, macromolecular complex known as the spliceosome. The spliceosome is thought to be responsible for a number of events including splice site and branch point recognition and positioning of the pre-mRNA transcript in a suitable orientation for the splicing reactions to take place (Brody & Abelson 1985, Grabowski et al. 1985). The spliceosome is composed of more than one hundred proteins and five snRNAs, Ul, U2, U4, U5, and U6 (Stevens et al. 2002, Reviewed in Jurica & Moore 2003). Each snRNA associates with a specific set of proteins to generate an snRNP particle, and these particles undergo extensive structural re-arrangements 3 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. throughout the splicing cycle (Fortner et al. 1994, Staley & Guthrie 1999, Hilliker et al. 2007, Perriman & Ares 2007). Although there have been reports of a pre-assembled "holospliceosome" containing all five splicing snRNPs (Stevens et al. 2002), the assembly of the spliceosome has traditionally been thought to occur de novo on each new pre-mRNA transcript through the ordered association of the snRNP particles (Fig .1, Bindereif & Green 1987, Cheng & Abelson 1987, Konarska & Sharp 1987). In the traditional view of step-wise assembly of the spliceosome, the U1 and U2 snRNPs first recognize and associate with the pre-mRNA transcript through base paring interactions with the 5' splice site and branch sequence respectively (Fig. 1, Black et al. 1985, Parker et al. 1987). Following formation of this pre-spliceosome complex, the U4, U5 and U6 snRNPs associate with the spliceosome as a preformed triple snRNP complex (Cheng & Abelson 1987, Konarska & Sharp 1987). U4/U6 base pairing is then disrupted and U6 replaces U1 at the 5' splice site (Wassarman & Steitz 1992, Lesser & Guthrie 1993, Staley & Guthrie 1999). At the same time, U4 and U1 become loosely associated with the spliceosome and may even be released while U2, U5 and U6 become tightly associated with the pre-mRNA transcript to form the catalytically active spliceosome (Cheng & Abelson 1987, Konarska & Sharp 1987). A number of conformational re­ arrangements of the active spliceosome then allow the splicing reactions to proceed, resulting in the production of a mature mRNA transcript and the excised lariat intron (Staley & Guthrie 1999, Hilliker et al. 2007). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1.3 U6 snRNA It is not yet clear whether nuclear pre-mRNA splicing is catalyzed by the snRNA or the protein components of the spliceosome, however it is becoming apparent that both the snRNAs and the proteins are required to perform critical functions throughout the splicing cycle (Guthrie 1991, Collins & Guthrie 2000, Valadkhan 2007, Abelson 2008). Many of the proteins serve as a structural scaffold around which the spliceosome can correctly assemble while many others are involved in promoting structural re­ arrangement of the snRNAs (Ohi et al. 2005, Raghunathan & Guthrie 1998b, Staley & Guthrie 1999). However it is largely the snRNAs that recognize and base pair to the splice sites and branch sequence of pre-mRNA transcripts (Black et al. 1985, Parker et al. 1987). The close proximity of U2 and U6 snRNAs to the catalytic center of the spliceosome, and their tight association with the pre-mRNA transcript in functionally activated spliceosomes, has lead to the speculation that splicing may be an RNA catalyzed event (Cech 1986, Madhani & Guthrie 1992). Of the five snRNAs, U6 is the most likely to play a direct role in pre-mRNA splicing catalysis. It is the most highly conserved in size, with essentially all size variation residing in the phylogenetically conserved 5' stem/loop, where stem length can vary by several base pairs (Brow & Guthrie 1988). The size conservation is accompanied by a highly conserved nucleotide sequence that shows approximately 80% sequence identity across the middle third of yeast and human U6 snRNAs, and approximately 60% identity across the entire U6 sequence of these same two species (Brow & Guthrie 1988, Guthrie & Patterson 1988). Steric constraints within the catalytically active spliceosome 5 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. may explain the strict size conservation while the high level of sequence conservation may reflect nucleotide specific requirements during spliceosome assembly and activation, and/or throughout the splicing reactions (Brow & Guthrie 1988). While many mutations throughout U6 snRNA result in only weak conditional growth phenotypes or show no observable deviation from wild type growth at all, one stretch of nucleotides, known as the ACAGAGA box, is unexpectedly intolerant of mutation (Fabrizio & Abelson 1990, Madhani et al. 1990, Wolff et al. 1994). There are a few known examples of variation in this sequence element in which deviation at the second position to a U appears so far to be limited to the Trypanosomatidae family (Xu et al. 1994). Point mutations within this stretch of seven nucleotides result in lethality in vivo and severe inhibition of splicing in vitro with various levels of first and/or second step blocks (Fabrizio & Abelson 1990, Madhani et al. 1990). The ACAGAGA sequence has been shown to genetically interact with, as well as to cross-link to, the 5' splice site region of the pre-mRNA and has also been proposed to play an as of yet undefined role throughout the splicing reactions (Fig. 2, Fabrizio & Ableson 1990, Wassarman & Steitz 1992, Lesser & Guthrie 1993, Kandels-Lewis & Seraphin 1993, Wolff et al. 1994, Kim & Abelson 1996). A second highly conserved sequence, the AGC triad, is located approximately six nucleotides downstream of the ACAGAGA box in mature U6 snRNAs. The AGC triad has been shown to interact with U2 snRNA through the formation of helix lb, a structure that has been proposed to play a role in exon ligation (Fig. 2, Madhani & Guthrie 1992, Hilliker & Staley 2004). Interestingly, group II self-splicing introns also contain an AGC 6 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. U —A U—A G-C A C-G G • U UUUACAAAGAGAUUUAUUUCGUUUU3' *c U—AG/V3AUUUAUUUCGUUUU3' UAUUUCGUUUJ3' Vidaveretal. (1999) Fortneretal. (1994) B U6 Karaduman etal. (2006) A AC U C G-C 5' G —CAUUUGGUC jGVACCGUUUUACMAGteAUUUAUUlXGUULUy A V E. Aauu^AAA c 1*1I I I * iiii iii 3' UUUCCAUAAGGUUUUUAAG *U CUGCCAGACCAA. ..UUAA UUUAAAG - UC« U4 C^f-AG 0 G G uSvG U6 5' G-CAUUUGGUCAAUUUO INTRON 5' G—CAUUUGGUCAAUuUC LARIAT INTERMEDIATE c -G C* A. 5' EXON AGIC G A UC UUACAAAGAGAUUUAUUUCGUUUU3' I IIII II UGFLU G U G A G| 3'EXON GA U GL U2 JC A*CrUUUUA GAUCAVG C G I II II A C U A G A U u C AAAGAGAUUUAUUUCGUUUU 3' HI" M U U U C r GUUUCU C UAAGCA 5' Figure 2. U6 snRNA conformational re-arrangements throughout the splicing cycle. Proposed U6 snRNA secondary structures in free U6 snRNP (A), U4U6 di-snRNA (B), and the first and second step catalytic spliceosome conformations (C). Invariant nucleotides are shown in bold and the ACAGAGA sequence is shaded in gray. 7 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. sequence base paired in a structure that resembles U2-U6 helix I or an extended U6 intramolecular stem/loop (ISL), both of which have been proposed to form near the active site of the spliceosome at different times throughout splicing (Peebles et al. 1995, Madhani & Guthrie 1992, Sashital et al. 2004). The AGC triad has been shown to be essential for catalysis of self-splicing introns and the essential metal ion binding at a bulge in domain V of group II self-splicing introns and in the U6 3' ISL further extend the parallels between these systems (Peebles et al. 1995, Yean et al. 2000). These observations suggest that nuclear pre-mRNA splicing may in fact be catalyzed by RNA through a mechanism similar to that of group II introns (Madhani & Guthrie 1992, Peebles et al. 1995). Like pre-mRNA splicing, group II self splicing is achieved through two transesterification reactions, resulting in the excision of a lariat intron that is formed through a 2'-5' phosphodiester linkage (Peebles et al. 1986, van der Veen et al. 1986, Cech 1986). U6 snRNA is found in three major forms throughout the splicing cycle: intramolecularly base paired in a free U6 snRNP, base paired extensively to U4 snRNA in the U4/U6 di-snRNP, and base paired to the pre-mRNA transcript and U2 snRNA in activated spliceosomes (Fig. 2, Brow & Guthrie 1988, Datta & Weiner 1991, Wu & Manely 1991, Madhani & Guthrie 1992, Field & Friesen 1996). These interactions are mutually exclusive and may reflect temporal regulation of structural re-arrangements during spliceosome assembly and activation since interactions between U6, U2, and the pre-mRNA transcript within the catalytically competent spliceosome may juxtapose important sequence elements for RNA catalyzed splicing (Fig. 2, Madhani & Guthrie 8 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1992, Johnson & Abelson 2001). Thus it may be critical to ensure that one interaction is correctly established prior to engagement in a second interaction. 1.4 U6 snRNA in Free U6 snRNP In its free U6 snRNP form, U6 snRNA is thought to interact with only a small complement of proteins: the precursor RNA processing protein, Prp24, and a group of seven small proteins, known as the Lsm complex (Shannon & Guthrie 1991, Cooper et al. 1995, Seraphin 1995, Stevens et al. 2001, Karaduman et al. 2006, Karaduman et al. 2008). Homologs of Prp24 have been identified in a number of organisms and the strong conservation of repeated RNA recognition motifs (RRMs) and a stretch of 12 highly conserved amino acids at the C-terminus implies that the function has been conserved (Bell et al. 2002, Rader & Guthrie 2002). Indeed, both Prp24 and its human homolog, SART3/pl 10nrb, have been shown to be required under U4/U6 di-snRNP recycling conditions (Raghunathan & Guthrie 1998, Bell et al. 2002), and the highly conserved Cterminal domain has been shown to be involved in promoting formation of this complex in yeast (Rader & Guthrie 2002). The Lsm proteins are thought to form a doughnut shaped complex that interacts with the uridine rich 3' end of U6 snRNA (Achsel et al. 1999, Vidal et al. 1999, Ryan et al. 2002). This complex is required for U6 snRNA stability and has been proposed to play a role in U6 snRNP conformational re-arrangements (Cooper et al. 1995, Mayes et al. 1999). Further, the entirety of this complex has recently been shown to be required for nuclear localization of free U6 snRNP (Spiller et al. 2007). The presence of the Lsm complex appears to enhance Prp24 binding, suggesting that protein association with U6 9 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. snRNA to form free U6 snRNP is a co-operative binding event (Ryan et al. 2002, Rader & Guthrie 2002). Although it will be some time before high resolution structural information for free U6 snRNP is obtained, low resolution electron microscope images are now available (Karaduman et al. 2008). These images have revealed the relative positioning of the two large protein domains and suggest that the Lsm complex interacts with U6 in a fixed orientation relative to Prp24 (Karaduman et al. 2008). Unfortunately it is not possible to visualize U6 snRNA at such low resolution, and there is still some disagreement surrounding the RNA interactions within free U6 snRNP (Ryan et al. 2002, Karaduman et al. 2006, McManus et al. 2007). Understanding the structure of U6 snRNA in the free snRNP will provide insight into the mechanism of U4/U6 di-snRNP formation and ultimately the activation of U6 snRNA for splicing. Several models of S. cerevisiae U6 snRNA secondary structure in free U6 snRNP have been proposed over the last two decades (Fortner et al. 1994, Brow & Vidaver 1995, Vidaver et al. 1999, Karaduman et al. 2006). The first widely considered proposal consisted of a 5' stem/loop, which was phylogenetically conserved, and two other stem structures: the central stem/loop and the 3' ISL (Brow & Guthrie 1988, Fortner et al. 1994). The latter two structures were based on the presence of 3' exonuclease degraded U6 species of length 60 and 90 nucleotides found in yeast sub-cellular extract (Eschenlauer et al. 1993). Fortner et al. (1994) attributed this observation to 3' exonuclease activity that defined the 3' border of the two stem structures. While the 3' 10 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ISL is phylogenetically conserved, the central stem/loop is not and was regarded as phylogenetically unproven (Fortner et al. 1994). Phylogenetic, genetic, and chemical structure probing experiments have provided strong support for both the 5' stem/loop and 3' ISL and consequently these structures have been proposed in all other models of yeast and mammalian secondary structure (Brow & Guthrie 1988, Fortner et al. 1994, Jandrositz & Guthrie 1995, Wolff & Bindereif 1993, Wolff & Bindereif 1995, Karaduman et al. 2006, McManus et al. 2007). In contrast, the stretch of 35 nucleotides located between these stem structures has undergone considerable remodeling ranging from a pseudoknot structure that later evolved into a telestem proposed by Brow and colleagues (Brow & Vidaver 1995, Vidaver et al. 1999), to completely unstructured in order to accommodate binding of Prp24 as proposed by Liihrmann and colleagues (Karaduman et al. 2006). In contrast to the yeast models, the mammalian model had remained unchanged since its first proposal in 1980 until recently. The initial version of mammalian U6 was much more extensively base paired throughout the central region of the molecule with an asymmetric bulge present about mid-way through an extended 3' stem/loop (Epstein et al. 1980). Karaduman et al. (2006) disrupted much of this intramolecular network to propose a mammalian model that is conserved between yeast and mammals, arguing that the high level of sequence conservation should be reflected through similar secondary structures (Karaduman et al. 2006). Like the yeast model, these researchers suggest that the Prp24 homolog binds U6 snRNA at this large bulge. 11 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1.5 Motivation for a New Model of U6 snRNA Secondary Structure In spite of extensive phylogenetic and structural information, there is no widely accepted secondary structure model of U6 snRNA in free U6 snRNP, let alone any understanding of its three dimensional shape. More importantly, these models do not offer compelling explanations of how U6 becomes incorporated into the spliceosome, nor do they explain the functional significance of the large structural re-arrangements in U6 throughout splicing. In particular, they do not address the function of transient association with U4 snRNA, an absolute requirement for U6 snRNA assembly into functional spliceosomes. Further, many of the experimental techniques employed to study U6 snRNA secondary structure have generated ambiguous results that have largely been interpreted in a manner consistent with the accepted model of the time without exploration of other possible interpretations of the data. The main objective of the work described in this thesis was to generate a new model of U6 snRNA secondary structure that is compatible with existing data, and offers some explanation of the functional relevance of the large structural re-arrangements in U6 snRNA during spliceosome assembly and activation, particularly in the function of the transient U4/U6 interaction. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter Two - Introduction to the Dunn-Rader Model of U6 snRNA Structure in Free U6 snRNP Two main objectives in our manual re-modeling of U6 snRNA were i) to minimize the extent of unstructured regions of RNA, and ii) to sequester the catalytically important ACAGAGA sequence within the RNA interaction network. Such sequestration of this potentially catalytic element may prevent U6 from engaging in premature interactions with pre-mRNA transcripts until signaled to do so through an induced conformational change. This chapter introduces the Dunn-Rader model of U6 snRNA secondary structure and assesses the quality of the structure by mapping data from the literature to the model, as well as to previously proposed models for comparison. 2.1 The Dunn-Rader Model of U6 snRNA Secondary Structure In addition to a 5' stem/loop that has been proposed in all other models of U6 snRNA secondary structure, the Dunn-Rader model predicts three additional helical segments which meet at a three-way junction: the central stem, stem/loop A, and stem/ loop B (Fig. 3). While the base pair composition of the central stem is very similar to that proposed by Karaduman et al. (2006), the presence of two additional stems is unique to the Dunn-Rader model. Stem/loops A and B greatly reduce the number of unstructured nucleotides by incorporating residues that have previously been proposed to reside in a large asymmetric bulge into these stems (Fig. 2A, 3). Importantly, the highly conserved ACAGAGA sequence is predicted to base pair within stem A in the Dunn-Rader model while all other models predict that this sequence lies in a large unstructured region comprising up to 23 unpaired nucleotides (Karaduman et al. 2006). The significance of this is discussed in chapter six. 13 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. m B r A C C A U• G A u Stem/loop A G Coaxial Stack A\ . _. . JCA A r I / 90|| A u ioG— C 5'Stem/loop A—U A—U G— C 'A40 \JG-C C U- A r u20 G * U C —G U•G (j — A MBmam Stem/loop [ „ G A 8 0 fl * • / UG A UG c A c JBC u— A U~ A A oG a A Central Stem C-G | | — A u A G-U'°° 30 G • 5'G—CAUUU "o UAUUUCGUULILI3' Figure 3. The Dunn-Rader model of U6 snRNA structure in the free U6 snRNP. (A) U6 snRNA secondary structure. Invariant nucleotides are shown in bold and the ACAGAGA sequence is shaded in gray. The 3' end of 3' exonuclease degradation products are indicated with wedges. The junction between stem/loops A and B (JAB), stem/loop B and the central stem (JBC), and the central stem and stem/loop A (JCA) are indicated. (B) Threedimensional representation of the proposed three-way junction from the front (top) and side (bottom): stem/loop A (red), stem/loop B (yellow), the central stem (blue), and the single-stranded junctions (green). Three-way junctions are common RNA folds characterized by the convergence of three helical segments at a single junction, where two of these helices tend to stack threedimensionally in a nearly coaxial fashion (Lescoute & Westhof 2006). The helices are connected by no more than three single stranded RNA segments, the length of which dictates which two of the three helices likely stack and the relative orientation of the third helix to the stack (Lescoute & Westhof 2006). The Dunn-Rader model predicts single stranded linking segments of lengths eight, four, and one nucleotide for the junction 14 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. between the central stem and stem/loop A (JCA), stem/loop A and stem/loop B (JAB), and stem/loop B and the central stem (JBC) respectively (Fig. 3). Since JCA is larger than JBC, the Lescoute and Westhof (2006) classification scheme suggests that stem/loops A and B may stack coaxially while the central stem bends in toward stem/loop A. 2.2 Three Dimensional Modeling of U6 snRNA 2.2.1 Methods To generate a theoretical three-dimensional representation of our proposed base pairing interactions, the core U6 snRNA nucleotide sequence consisting of the S. cerevisiae residues 30 to 101 (excluding two nucleotides in JAB) was loaded into the RNA modeling program MC-Sym (http://www.major.iric.ca/MC-Pipeline) along with our secondary structure constraints (Parisien & Major 2008). Three-dimensional models were viewed and manipulated in MacPyMOL (DeLano 2008). 2.2.2 U6 snRNA in Three Dimensions MC-Sym is an RNA modeling program that is capable of generating a threedimensional model from a primary nucleotide sequence, which can be supplied with or without secondary structure constraints (Parisien & Major 2008). Small RNA cycles, known as nucleotide cyclic motifs (NCMs), have been extracted from RNA crystal structures to generate a database of RNA tertiary structure building blocks. Since adjacent NCMs share a common border, a three dimensional model can be generated by overlapping the common edge of NCMs selected from the database. Thus an entire threedimensional model can be produced through the step-wise addition of NCMs to a growing chain of building blocks. 15 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. To model our U6 snRNA secondary structure in three dimensions, we supplied MC-Sym with a 5' and 3' truncated U6 snRNA construct consisting of nucleotides 30-101 along with our secondary structure constraints. This construct contains the central stem and stem/loop A and B regions of U6 snRNA and will be referred to as the core domain. While MC-sym generated a number of potential structures, many of these were rejected for one of two reasons: either the path of the theoretical backbone contained bond angles that would produce substantial steric clashing, or the structure of the three-way junction did not agree with predictions based on empirical data collected from high resolution Xray crystal structures. However, as predicted from the lengths of the single stranded junctions connecting the helices, MC-Sym did predict a model with a relatively compact three dimensional structure in which stem/loops A and B stack coaxially, and the central stem bends in toward stem/loop A (Fig. 3). The compact structure was made possible by wrapping JCA around the backside of the molecule (Fig. 3). This model was free of severe steric clashing and was in agreement with empirically derived predictions. A second model generated by MC-Sym takes on a much more open conformation where the base pairing has been maintained, however twisting the central stem, in combination with the flexibility of the eight nucleotide single stranded junction, un-tucks this helix from folding up toward stem/loop A (Fig. 4). Karaduman et al. (2006, 2008) have provided evidence to suggest that binding of U6 snRNA associated proteins induces a conformational change that results in a much more open structure in the presence of Prp24. These authors accommodated this observation by simply disrupting base pairing of their extensively base paired protein-free U6 snRNA model to create a large 16 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. asymmetric bulge in which Prp24 could bind. The three dimensional models depicted here suggest that upon protein binding, U6 snRNA structure may open, however this conformational change does not require disruption of the base pairing interactions. Central Stem Figure 4. U6 snRNA can fold into a closed (left) and open (right) conformation without disrupting the base pairing interactions proposed in the Dunn-Rader model. The 5' end is blue and the 3' end is red. To assess the overall flexibility of U6 snRNA, 2D Total Coherence Transfer Spectroscopy (TOCSY) Nuclear Magnetic Resonance (NMR) was performed in collaboration with Dr. Philip Johnson at York University. The TOCSY-NMR spectrum obtained indicated that there were nine or ten pyrimidines within the core domain of U6 that reside in unstructured regions of the molecule (Fig. 5). While the Karaduman et al. (2006) model of U6 snRNA secondary structure in the free U6 snRNP predicts sixteen pyrimidines in flexible, single-stranded regions of the core domain, the Dunn-Rader and Vidaver et al. (1999) models predict only nine and ten respectively. Thus this experiment revealed that the core domain of U6 snRNA was not as unstructured as Karaduman et al. (2006) propose for the protein bound U6. However, since the TOCSY-NMR experiment was performed on in vitro transcribed U6 in the absence of proteins, it is more 17 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. appropriate to consider the structure proposed by Karaduman et al. (2006) in the absence of protein. Since this structure also predicts nine unstructured pyrimidines, the TOCSYNMR experiment was not capable of distinguishing between the models. 4.8 4.9 5.0 5.' 5.2 5.3 H 5.4 P «- p 5.3 5.6 'm 5.7 5.8 5.9 6.0 6." 6.2 Figure 5. 2D TOCSY NMR spectrum of in vitro transcribed U6 snRNA. Chemical shifts are given on the x and y axes and mobile pyrimidines are indicated in red. 2.3 Support for the Dunn-Rader Model from the Literature An important first step in testing our new model of U6 snRNA secondary structure was to assess the compatibility with experimental results obtained previously by other labs. Extensive chemical structure probing and genetic techniques have been employed to study U6 snRNA conformation and function. Unfortunately, many of these experimental methods have generated results with ambiguous interpretations, and while 18 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. we do not dispute the authenticity of these data, we do question whether the data support various structural elements within the context of a free U6 snRNP as opposed to some other U6 containing complex that forms throughout the splicing cycle. Interestingly, much of the data present in the literature can be explained at least as well by the DunnRader model, and in some cases more thoroughly, than either the Vidaver et al. (1999) or Karaduman et al. (2006) models. Here we present a brief analysis of such data analyzed in the context of each of the proposed free U6 snRNA conformations in order to assess the quality of each model based on known experimental information. 2.3.1 Chemical Modification of Accessible U6 snRNA Bases U6 snRNA secondary structure, both in the absence and presence of protein, has been tested through chemical structure probing by several labs (Fortner et al. 1994, Jandrositz & Guthrie 1995, Karaduman et al. 2006). The principle behind these experiments is that various reagents can chemically modify nucleotides that have accessible Watson/Crick base pairing positions (Stern et al. 1988). During subsequent primer extension reactions, the activity of reverse transcriptase is blocked when the enzyme encounters such a modification, resulting in the production of truncated molecules that can be separated on a gel to identify the modified nucleotides. Thus chemical structure probing can be used to distinguish between nucleotides that are involved in canonical Watson/Crick base pairing, which are strongly protected from chemical modification, and those that are not, which become strongly modified. While these experiments can be very informative, it is important to acknowledge that the experimental conditions greatly influence the results, and more importantly, that 19 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. the assignment of the level of modification to each nucleotide is a subjective process. Interpretation of the results is complicated even further by dynamic processes such as helix breathing where modification signals are weaker and repeated experiments may provide conflicting results. In addition, the protection of a nucleotide from chemical modification in the absence of proteins suggests that the nucleotide is base paired; however, no information is provided as to which nucleotide it is base paired with. Thus the results are ambiguous and should be interpreted with caution since they may support multiple secondary structure predictions as illustrated below. The first chemical structure probing of S. cerevisiae U6 snRNA was performed in vivo by allowing the chemical reagent to diffuse into the cells (Fortner et al. 1994). Consequently the chemical modification pattern obtained was representative of all possible U6 snRNA conformations throughout its lifetime and was therefore not considered here. In contrast, Jandrositz and Guthrie (1995) and Karaduman et al. (2006) applied chemical structure probing to U6 snRNPs partially purified on a glycerol gradient or TAP-tag purified respectively. In addition, these two groups compared the snRNP modification pattern to protein-free U6 obtained either through hot phenol: chloroform extraction of RNA from peak U6 snRNP-containing glycerol gradient fractions, or from in vitro transcribed U6 snRNA respectively. We have mapped the Jandrositz and Guthrie (1995) and Karaduman et al. (2006) chemical modification data in the absence and presence of protein to each of the three competing models of U6 snRNA secondary structure (Fig. 6, 7). Note that while the Vidaver et al. (1999) (VB) and Dunn-Rader (DR) models were considered for both the 20 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. RNA and RNP data, Karaduman et al. (2006) (KL) have proposed different structures in the absence and presence of protein and thus we consider each structure where appropriate. Both sets of data were consistent with the 5' stem/loop proposed in all models and therefore only residues following the 5' stem/loop region were examined here. Since it is not clear whether weak chemical modifications represent nucleotides that are base paired or single stranded, we have limited our analysis to only nucleotides for which there was a strong protection or strong modification in both sets of data. This reduced our analysis to only 13 residues in protein-free U6 snRNA and 22 residues in U6 snRNA in free U6 snRNPs. Of the thirteen nucleotides that met our criteria, eight were consistent with the VB model, 11 were consistent with the DR model, and 12 were consistent with the KL model (Fig 6). It should be noted that Karaduman et al. (2006) have proposed a U*U mismatch as part of their extended stem structure and that such a mismatch would be expected to disrupt the canonical form of the helix (Leontis et al. 2002). Since both of these nucleotides were strongly protected, the number of nucleotides consistent with the data may be artificially high for the KL model if this interaction does not actually occur. The U6 snRNP data set contained one nucleotide out of 22 where the strong modification from the Jandrositz and Guthrie (1995) data was not consistent with the strong protection observed by Karaduman et al. (2006). Of the remaining 21 nucleotides, 16 positions were consistent with the VB model, 16 were consistent with the DR model, and 13 were consistent with the KL model. Three of these nucleotides, A40, A41, and A42, were strongly modified in the protein free data set and strongly protected in the U6 21 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. U6snRNP Naked U6 snRNA Jandrositz | v | D | K I Karaduman Jandrositz ' A26 Karaduman A26 •IW:1 X *** *** U29 G30 G31 U32 G30 G31 X C33 WB1 X^ C33 X A3 5 U36 U37 U38 G39 X x 'x' E3 • • G39 A40 A41 A42 C43 G30 G31 U32 ** X A35 U36 U37 U38 G39 A40 • A • A • • • A • A • A * A A • A • A • A • A A RIM "C48 ! A49 •• 1 rxri KGH rxri MEEM A49 G50 A51 G50 A51 G52 ^•_i P H U54 X A51 G55 A56 • •• • • XM C58 A59 [ • C61 A62 G63 G55 A56 U57 G55 X C58 A59 G6Q I C61 A62 G63 G63 U64 U65 C66 C67 C68 C69 U70 G71 C68 C69 X A • A • A • A * • C66 A • A • WM X C66 A49 G50 G52J G60 x • • • A r>< • A A A • A G63 U64 U65 C66 C68 C69 U70 G71 C72J A76 G77 G78 X | zzliu • • >< >< >< A73 U74 A75 A76 G77 G78 A79 U80 G81 A82 A83 A76 G77 G78 A76 G77 G78 IS U80 G81 A82 G81 A82 A83 C84 C85 G86 X G86 U87 U88 U89 22 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Figure 6. Chemical modification of naked U6 snRNA and U6 snRNA in U6 snRNPs. (A) Data from Jandrositz & Guthrie (1995) and Karaduman et al. (2006) are given on the left and right of each grid respectively. Strongly modified residues are shown in a black box, weakly modified residues in gray, strongly protected residues in white, and residues for which there is no information are replaced by an X. The middle three columns represent the predicted modification (black) or protection (white) pattern for the Vidaver et al. (1999) (V), Dunn-Rader (D), and Karaduman et al. (2006) (K), models of U6 snRNA secondary strucuture. Residues that UV cross-link to Prp24, the Lsm complex, or both are indicated by one, two, or three asterisks respectively (Karaduman et al. 2006, 2008). Residues protected by hydroxyl radical cleavage are shown by small (weak protection) or large (strong protection) circles (Karaduman et al. 2006) or triangles (Ghetti & Abelson 1995). Stem/loop A Stem/loop B Central Stem Figure 7. The Karaduman et al. (2006) chemical modification data for in vitro transcribed U6 snRNA and U6 snRNPs mapped to the proposed closed (A) and open (B) Dunn-Rader models. Strongly modified residues are red, weakly modified are orange, strongly protected are yellow, and no information are green. snRNP data set. Further, treatment of free U6 snRNPs with SDS and proteinase K eliminated the strong protection of these residues (Jandrositz & Guthrie 1995). These observations, combined with strong protection of the RNA backbone from hydroxyl radical cleavage, suggest that residues A40-A42 are in fact single stranded and that protection in the snRNP was due to the presence of protein (Fig. 6, Jandrositz & Guthrie 23 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 1995, Karaduman et al. 2006). This increases the consistency between the snRNP data and the KL and DR models to 16 and 19 respectively. In summary, while each of the models under consideration predicted dramatically different base pairing interactions, the DR model was no less consistent with chemical modification data than either the VB or KL models. In fact, while the consistency between the data and the DR model was approximately the same as that for the KL model when considering protein-free U6 snRNA, consistency between the DR model and the U6 snRNP data was slightly higher than the VB and KL models, providing additional support for the DR model. 2.3.2 Compensatory Base Mutations One region of U6 snRNA secondary structure that has been thoroughly tested through compensatory base mutation is the telestem structure proposed by Vidaver et al. (1999). This structure can be divided into the upper and lower telestem, which are composed of base pairing between nucleotides 40-43 with 86-89, and 36-39 with 92-95 respectively. Ryan et al. (2002) showed through a series of compensatory base mutations that disruption of the lower telestem inhibited Prp24 binding to synthetic U6 snRNA in vitro, and that Prp24 binding could be restored by restoring base pairing in the lower telestem. In contrast, restoration of base pairing in the upper telestem did not restore Prp24 binding. While tri-nucleotide substitution of the 5' strand of the upper telestem inhibited Prp24 binding, tri-nucleotide substitution of the 3' strand of the upper telestem did not affect Prp24 binding, tri-snRNP assembly, or pre-mRNA splicing. Thus Ryan et al. (2002) concluded that base pairing in the upper telestem does not occur during 24 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. spliceosome assembly or splicing and that the lower telestem and residues 40-42 are important for Prp24 binding in vitro. In a slightly more elaborate in vivo experiment, Vidaver et al. (1999) drew essentially the same conclusion as Ryan et al. (2002). These authors had created a cold sensitive mutant, U6-A62G, and isolated cw-acting suppressors in a previous publication (Fortner et al. 1994). Many of these suppressors fell into the telestem region and were predicted to suppress the cold sensitive phenotype of A62G, which was thought to hyperstabilize the 3' ISL (see below), by destabilizing the telestem. Vidaver et al. (1999) reasoned that if this truly was the mechanism of suppression, then introducing a mutation that restored base pairing in the telestem would result in reversion back to a cold sensitive growth phenotype. Consistent with this hypothesis, restoring base pairing at positions 36 and 37 with positions 94 and 95 in the lower telestem reverted the growth phenotype. However, restoration of base pairing between positions 41 and 42 with positions 87 and 88, and between residues 40 and 89 in the upper telestem did not revert the growth phenotype. Thus Vidaver et al. (1999) were able to find in vivo support for the lower telestem, but not the upper telestem. Both the Karaduman et al. (2006) and Dunn-Rader models of U6 snRNA secondary structure propose the lower telestem as the upper half of the central stem, and do not incorporate residues 40-42 into any base pairing interactions. Thus both of these models have maintained the stem structure and adjacent single stranded nucleotides required for Prp24 binding. These structural features are consistent with the strong protection of nucleotides in the 5' strand of the upper central stem as well as with the 25 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. strong modification of residues 40-42 in the absence of Prp24, and the strong protection of these residues in the presence of Prp24 (Fig. 6, Jandrositz & Guthrie 1995, Karaduman et al. 2006). Further, the entire 5' strand of the central stem and residues 40-42 are strongly protected from hydroxyl radical cleavage, suggesting that a protein component of the U6 snRNP does indeed bind at this location (Fig. 6, Ghetti et al. 1995, Karaduman et al. 2006). We conclude that the compensatory analyses described above provide support for at least the upper half of the central stem, but not for the upper telestem proposed in the VB model. 2.3.3 The U6 snRNA 3' Intramolecular Stem/loop (3' ISL) Previously proposed models of U6 snRNA secondary structure in free U6 snRNP suggest that a universally proposed 3' intramolecular stem structure must unwind to allow for interaction with U4 snRNA during spliceosome formation (Fig. 2, Epstein et al. 1980, Fortner et al. 1994, Vidaver et al. 1999, Ryan et al. 2002, Karaduman et al. 2006). This structure is then thought to reform in active spliceosomes where it coordinates a metal ion that has been shown to be important for the splicing reactions (Madhani & Guthrie 1992, Fortner et al. 1994, Yean et al. 2000). Hydroxyl radical cleavage of portions of the 3'ISL that are consistent with the turn of a helix places this stem structure in proximity of other important catalytic elements within the active spliceosome (Rhode et al. 2006). Support for the 3' ISL in the free U6 snRNP comes largely from a genetic analysis of two mutants, A62G and A79G, which were predicted to hyperstabilize the yeast 3' ISL and resulted in the predicted cold sensitivity and U4/U6 di-snRNP assembly defect expected for stem hyperstabilization (Fortner et al. 1994, McManus et al. 2007). The U4/ 26 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. U6 assembly defect, which was also present at 30C where growth was wild type, could be overcome by over-expressing U4, however, the cold sensitive phenotype persisted even when the U4/U6 assembly defect had been alleviated (Fortner et al. 1994). If the cold sensitive growth phenotype were due to a shortage of U4/U6, as predicted by the authors, then restoring U4/U6 levels would allow the yeast to grow in the cold, which was not observed. These observations suggest that destabilization of the 3' ISL is important at a later stage in spliceosome assembly and activation, or during the splicing reactions. Since destabilization of the 3' ISL would be required for interaction with U4 snRNA to generate the di-snRNP particle, these results argue against the presence of the stem in the free U6 snRNP. Similar mutational analyses conducted in a mammalian system also argue strongly for the importance of the 3' ISL following, but not preceding, U4/U6 interaction. Disruption of the C62-G72 base pairing interaction with either a C62G or a G72C point mutation resulted in wild type U4/U6 levels and spliceosome assembly, however splicing was reduced to less than 10% (Wolff & Bindereif 1993). When the double mutation C62G/G72C, which inverted the base pair interaction and should therefore not affect the stability of the 3' ISL, was introduced, U4/U6 and spliceosome assembly levels were again comparable to wild type, however splicing had increased to between 10% and 50% (Wolff & Bindereif 1993). This observation suggested that formation of the C62-G72 base pairing interaction was important in the context of the active spliceosome (Wolff & Bindereif 1993). 27 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. To our knowledge there are currently no unambiguous data available to suggest that the 3' ISL forms in the free U6 snRNP. While the genetic data discussed above highlighted the importance of the formation of this structure in the active spliceosome, it neither supported nor disproved the existence of the structure in free U6 snRNP. Chemical structure probing of U6 snRNA in free U6 snRNPs was just as consistent with the formation of the 3' ISL as they were with the replacement of this structure with stem/ loops A and B as proposed in the Dunn-Rader model (see above). Thus the available genetic data at this time are fully consistent with the base pairing interactions proposed in the Dunn-Rader model. 2.3.4 Genetic Suppression Analysis - A62G and A79G Suppressors of the cold sensitive growth phenotype exhibited by both the A62G and A79G point mutations have been proposed to exert their effect through two different mechanisms: destabilization of the 3' ISL, or destabilization of the telestem or telestem/ Prp24 interactions (Fortner et al. 1994, Vidaver et al. 1999, McManus et al. 2007). We agree with these modes of suppression, however we suggest that suppressors that act through 3' ISL destabilization do so in the context of the active spliceosome rather than in free U6 snRNPs (Fig. 8). Destabilization of this structure may be required during active site structural re-arrangement between splicing steps and/or for spliceosome disassembly. Thus hyperstabilizing the 3' ISL with the original A62G or A79G mutations may stall the active spliceosome and prevent recycling of the spliceosomal components, slowing growth of these strains at low temperature. 28 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ^I^A-A,CU r.% EM-C-G A U-A->G ^U—A C-G A-U I* A-U - JZ G C UA-u Jrlffl u-© F A •A,«CU U ^ A,U a c 77' c, ^ / ® /.G«aUrca/j> u-'./ a G_C~*-AlVOC gI\A i=u "V X A_tV MI_» 1 ®^.U-A->-U A ACAA A G-C^. -A->-G ®' A U U-AGAGAI •AGAGAUUJAUUUCGUUU U3' C G »'« U -•C(G C,G c G u-vu<:A^ffVyjU ->U A ^ * i# cA H-> •A, C*U A—U ^*A U<-G G>A,c IM INTERMEDIATE [u]-4-c— -( G ->C,G IOr A U U A CAAAGAGAUUUAUUUCGUUUU3' U A C U A C AUG IIII U CUC UAAGCA5' U A C U A C A I IIII UG UG AU G AU GL AGL 3'EXON U_A ft1 C,G fu / A" Ai"^0*C gU "J* «_fc.ij RM*Ai^H'CrUUUUis fi UA CAAAGAGAUUUAUUUCGUUUU " <-C 3' RU UG A U C A IIIII v GW M.U A.U 'C G U %uutC II II II' 111**1 GUUUCUC UAAGCA 5' U2 Figure 8. A62G and A79G suppressor mutations mapped onto U6 snRNA secondary structures in the free U6 snRNP (A), di-snRNP (B), and active spliceosome (C). The A62G and A79 G mutations are indicated in a black box. Suppressor mutations are indicated as follows: Bold: suppress A62G, regular font: suppress A79G, bold with an asterisk: suppress A62G but has no affect on A79G, circled bold: suppress both A62G and A79G,square bold: suppress A62G but enhance A79G. 29 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. In the context of the Dunn-Rader model of U6 snRNA secondary structure, A62G may increase the stability of stem/loop A by creating an additional base pair at the top of the stem while A79G would have a relatively mild affect on stem/loop B stability (Fig. 8). Many of the suppressors of both mutations map to the Prp24 binding site located at the upper central stem and residues 40-43, and may act through central stem or central stem/ Prp24 destabilization as proposed by Fortner et al. (1994) and Vidaver et al. (1999). In most cases, suppressors that map outside of this region may provide a slight destabilizing affect on stem/loop A or B. Interestingly, many of the A79G suppressors cluster in the stem/loop B region very close to the original A79G mutation (Fig. 8). Nevertheless, many of these suppressors likely exert their effect at a later stage in the splicing cycle in the context of the 3' ISL. If the sole effect of the A62G and A79G mutations were to hyperstabilize the 3' ISL, then suppressors of one mutation would be expected to also suppress the other. While this was true for four of the A62G suppressors, four others suppressed A62G and had no affect when combined with A79G, and four others suppressed A62G but actually enhanced the A79G phenotype (Fig. 8, McManus et al. 2007). Thus something more is going on with one or both of these mutations beyond 3' ISL hyperstabilization. In the context of the active spliceosome, protonation of A79 in the C67»A79 mismatch facilitates coordination of a catalytically important Mg2+ in the 3' ISL (Huppler et al. 2002). A to C is the only known sequence variation at position 79, and is also the only mutation that would allow for similar protonation and maintenance of almost perfect isostericity with the C67*A79 mismatch (Huppler et al. 2002). Interestingly, the only two 30 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. A79G suppressors identified at position 79 were reversion back to an A, or mutation to a C (McManus et al. 2007). 2.3.5 Exonuclease Degradation of U6 snRNA The 3' end of U6 snRNA is generally protected from 3' exonuclease activity in vivo as a result of binding of the Lsm complex to this region. However, U6 snRNA added to sub-cellular or splicing extracts is typically not protected from this activity and the length of degradation products can provide some insight into the U6 snRNA secondary structure. 3' exonuclease activity is thought to occur progressively from the 3' end of the molecule to the 3' end of a stem structure where nuclease activity becomes blocked (Eschenlauer et al. 1993). Full length U6 transcribed in yeast sub-cellular extract is routinely degraded to a 90-nucleotide fragment, suggesting that a stem structure ends with a base pair containing residue 90 (Margottin et al. 1991, Eschenlauer et al. 1993). This observation is not consistent with any of the proposed U6 snRNA secondary structures in free U6 snRNP, the di-snRNA, or the active spliceosome. Residue 90 is predicted to either lie in a single stranded region, where nuclease activity should continue, or in the middle of a stem structure, where nuclease activity should be blocked several nucleotides earlier, in all but the Dunn-Rader model where nucleotide 90 is found at the 3' edge of stem/loop B (Fig. 2, 3). However, in order for exonuclease activity to continue through to stem/loop B, the central stem would have to be disrupted. It is possible that the in vitro transcription assay produces U6 transcripts that have not yet fully folded, thus disruption of the long-range central stem interaction, if it has formed at 31 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. all, may allow for degradation up to the first stably formed stem, which may be stem/loop B. Ryan et al. (2002) have also observed 3' degradation products of synthetic full length and 3' truncated versions of U6 snRNA reconstituted into U6 snRNPs in U6 snRNA-depleted splicing extract. Full length U6 snRNA was typically degraded to 101 nucleotides while shortened versions consisting of residues 1-94 and 1-91 were degraded to 90 nucleotides, and even shorter versions containing residues 1-88, 1-86, and 1-84 were degraded to 81 nucleotides (Ryan et al. 2002). Degradation of full length U6 to 101 nucleotides is consistent with degradation continuing up to the 3' edge of U2/U6 helix II as suggested by Ryan et al. (2002), but is also consistent with degradation up to the 3' edge of the intramolecular U6 snRNA central stem, proposed in both the Karaduman et al. (2006) and Dunn-Rader models (Fig. 2, 3). Degradation of the 1-94 and 1-91 constructs is consistent with degradation to the 3' edge of stem/loop B in the Dunn-Rader model, since much of the central stem base pairing region has been eliminated in these constructs, allowing for exonuclease activity to proceed to stem/loop B (Fig. 3). A degradation product of 81 nucleotides is not consistent with any of the proposed U6 snRNA secondary structures except for U6 snRNA in the U4/U6 di-snRNP (Fig. 2, Ryan et al. 2002). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter Three - Structure Probing by Oligonucleotide Accessibility Accessible regions of RNA are capable of annealing to short complementary oligonucleotides while regions of RNA that are involved in RNA/RNA or RNA/protein interactions are not. The accessibility of segments of an RNA molecule to a complementary oligonucleotide can therefore indicate a great deal about the overall secondary structure of the molecule. These types of experiments have been employed both in the absence and presence of proteins as well as in both human and yeast splicing systems (Black & Steitz 1986, Blencowe et al. 1989, Fabrizio et al. 1989, Li & Brow 1993, Brow & Vidaver 1995). Here we have tested the Dunn-Rader model of U6 snRNA secondary structure by designing oligonucleotides that were complementary to S. cerevisiae U6 snRNA and examining the ability or inability of these oligonucleotides to anneal to yeast U6 snRNA in a protein-free system. 3.1 Materials and Methods Cold PhenoUChloroform Total RNA Extraction One milliliter of wild type W303 yeast culture, grown overnight to an O D 5 9 5 between one and two, was spun down in a microcentrifuge tube in an Eppendorf Centrifuge 5417C at 13,000 rpm for one minute at room temperature. The cell pellet was washed twice with cold, sterile dftO prior to resuspension in 300(iL chilled, filtered RNA extraction buffer (50mM Tris-HCl pH 7.5, lOOmM NaCl, lOmM EDTA). Acid washed, baked 0.5mm Zirconia/Silica beads (BioSpec Products Inc.) were added to the 200jiL mark and tubes were vortexed for one minute on the maximum setting. Following a five minute incubation on ice, the tubes were vortexed for an additional minute before 33 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. adding 300|iL chilled RNA extraction buffer [60|xL 10% SDS, and 400|xL acid phenol: chloroform (5:1, pH 4) (Ambion)]. The samples were then vortexed for one minute on the highest setting and centrifuged in an Eppendorf Centrifuge 5415D at 13,200 rpm for two minutes at 4C. The aqueous phase was transferred to a tube containing 500|^L cold acid phenol: chloroform (5:1, pH 4) (Ambion), was vortexed for one minute, and centrifuged for two minutes as before. A third phenol: chloroform (5:1, pH 4) (Ambion) extraction was performed followed by an extraction with 500(iL chloroform (Sigma) as described above. The aqueous phase from the chloroform extraction was transferred to a clean eppendorf tube where 40(iL 3M sodium acetate and lmL 100% ethanol was added. Samples were cooled at -80C for at least 20 minutes. Precipitated RNA was pelleted by centrifugation in an Eppendorf Centrifuge 5415D at 13,200 rpm for 20 minutes at 4C. The pellets were washed with 70% ethanol and allowed to air dry for 5-10 minutes prior to resuspension in 30|iL lOmM Tris-HCl, pH 7.5. The final nucleic acid concentration was determined by Nanodrop. 5'End Labeling of DNA Oligonucleotides Fifty picomoles of DNA oligonucleotide were incubated with 2.5|aL of 10X T4 Polynucleotide Kinase Buffer (NEB), 20 units of T4 Polynucleotide Kinase (PNK) (NEB), and 3p.L of [y-32P]-ATP (PerkinElmer) in a final reaction volume of 25(iL. The reactions were incubated for one hour at 37C followed by heat inactivation of the T4 PNK for twenty minutes at 65C. The reaction was diluted to 50|a.L with dFhO and unincorporated [y-32P]-ATP was removed using either an Illustra MicroSpin G25 Column 34 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. or an illustra ProbeQuant G50 Micro Column (GE Healthcare) according to the manufacturer's instructions. Oligonucleotide Accessibility Experiments Twenty micrograms of total RNA was incubated with l|a.L (approximately 1 pmol) of 5' end labeled DNA oligonucleotide and l|o.L of Hybridization Buffer [150mM NaCl, 50mM Tris-HCl pH 7.5, ImM EDTA] at either 4C or 70C for 40 minutes. Reactions that were pre-incubated with unlabeled oligonucleotide were first heat treated at 70C for 5 minutes in the presence of 1(iM unlabeled oligonucleotide and were then allowed to slow cool to room temperature for 15 minutes prior to incubation with the labeled oligonucleotide for 20 minutes. All reactions were moved to ice prior to gel loading and each reaction was loaded into a pre-chilled 9% non-denaturing polyacrylamide gel and the gel was run at 250V for 1 hour. The gel was then exposed to a phosphor-imager screen at -80C and the resulting autoradiogram was visualized and quantified with a Cyclone phosphor imager and OptiQuant Software© (Packard Instruments). Temperature course experiments with oSDR464 were conducted by incubating 7|iT. of 5' end labeled oSDR464 with 100 pmol of in vitro transcribed U6 snRNA (Ambion MEGAshortscript Kit) for 15 minutes at 4C, 10C, 19C, 40C, and 70C. An additional heat treatment was performed in which the incubation was placed at 70C for 5 minutes and then slow cooled to room temperature for 10 minutes. Each reaction was loaded into a pre-chilled 9% non-denaturing polyacrylamide gel and the gel was run at 250V for 1 hour 20 minutes. The gel was exposed to a phosphor imager screen and visualized as above. 35 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Oligonucleotide GC content and melting temperature (Tm) were calculated using the Integrated DNA Technologies (IDT) SciTools Oligo Analyzer 3.1 at http://www.idtdna.com (Owczarzy et al. 2008). Table 1. DNA Oligonucleotide Sequence, GC Content, and Melting Temperature Oligo Number Sequence Length % GC Tm (C) oSDR.180 5 '-AAAACGAAATAAATCTCTTTG-3' 21nts 23.8 44.8 OSDR461 5'-ATCCTTATGCAGGGGAAC-3' 18nts 50 51.6 OSDR464 5 '-CATCTCTGTAT-3' 1 lnts 36.4 24.9 OSDR465 5 '-GATCATCTCTGTATTG-3' 16nts 37.5 39.5 OSDR466 5 '-AATCTCTTTGTAAAAC-3' 16nts 25 36.8 OSDR467 5 '-TTGTTTC AAATTGACC-3' 16nts 31.3 41.1 OSDR494 5 '-CTTCGCGAAC-3' lOnts 60 34.4 3.2 Results and Discussion The Dunn-Rader model of U6 snRNA secondary structure was tested using an oligonucleotide (oligo) accessibility experiment with several oligos that were designed to be complementary to specific regions of U6 snRNA sequence (Fig. 9). The oligos were 5' end labeled in order to detect annealing to U6 snRNA, which would generate a complex that migrated through a non-denaturing polyacrylamide gel with a slower mobility than the oligo alone. Thus we expected to observe a shift in the mobility of oligos that were predicted to anneal to single stranded regions of U6 upon incubation with yeast total RNA, and no shift for oligos designed to anneal to regions of U6 predicted to take part in base pairing interactions. Since the oligos were incubated with total RNA under non-denaturing conditions, some oligos were also predicted to anneal to 36 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. U6 in the U4/U6 complex and therefore we also expected to see oligo shifts of an even larger U4/U6/oligo complex running more slowly than the U6/oligo complex. Labeled Oligo 0SPRI8O OSDR494 OSDR461 Temperature 4 4 4 4 70 4 70 4 4 70 70 70 4 70 4 - 461 - - 4 4 70 4 4 4 70 Total RNA A—U _ Pre-incubation _ 461 - 461 - 10 11 12 13 14 15 16 17 18 1920 - 467 - - - 466 - 21 22 23 24 25 26 OSDR464 uriLu.! ' I A-U OSDR467 A—U U—A U —A oSDR466 U —A J —A U— A G—C A U U U £ Temperature ice 10 U6/oligo Labeled oligo UAUUUCGUUUU3' 19 40 « Labeled oligo 70 AMI! D m— Figure 9. U6 snRNA secondary structure probing by oligonucleotide accessibility. (A) The annealing sites of complementary DNA oligonucleotides are indicated on the Dunn-Rader model of U6 snRNA secondary structure. (B) Oligonucleotide accessibility of U6 snRNA in cold phenol: chloroform extracted yeast total RNA. Reaction conditions are indicated above each lane. (C) Temperature course of oSDR464 annealing with in vitro transcribe U6 snRNA. (D) oSDR466 (blue) and oSDR467 (red) annealing sites mapped onto the MC-Sym three-dimensional model of U6 snRNA secondary structure. 37 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. oSDR.494 was designed to be complementary to the 5' strand of the proposed 5' stem/loop and was used here to visualize inefficient annealing since all evidence suggests that this structure does exist in free U6 and U4/U6 (Fig. 9). oSDR494 did not anneal to free U6 snRNA or the U4/U6 di-snRNA complex at either 4C or following heating at 70C (Fig. 9). This result was not surprising since the estimated melting temperature of the 5' stem/loop is approximately 95C. Even if this structure became destabilized enough at 70C for some annealing of oSDR494, the much stronger intramolecular base pairing interaction would probably out-compete the U6/oligo interaction, knocking the oligo off of the RNA to regenerate the hairpin. The absence of oSDR494 annealing is consistent with the existence of a 5' stem/loop as predicted by all proposed models of U6 snRNA secondary structure. Sequestering the ACAGAGA region in a base pairing interaction is unique to the Dunn-Rader model. All other models of U6 snRNA in free U6 snRNP, as well as an extensively base paired model for protein-free U6 snRNA proposed by Karaduman et al. (2006), predict that this sequence is located in a bulge and should therefore be accessible to a labeled oligo. oSDR464 was designed to be complementary to an eleven nucleotide stretch containing the ACAGAGA sequence and was used here to distinguish between our model, where it would not be expected to anneal, and all other proposed models where it would be expected to anneal. Consistent with only the Dunn-Rader model, oSDR464 did not efficiently anneal to free U6 at 4C, however this oligo also did not anneal any more stably to U6 when samples were heat-treated (Fig. 9). We expected that heat treatment would result in destabilization of the overall U6 snRNA structure, 38 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. allowing for efficient annealing of the labeled oligo. However, since oSDR464 was capable of annealing to in vitro transcribed U6 snRNA, it was clear that the oligo should have annealed to the ACAGAGA sequence had it been accessible as proposed by all other models (Fig. 9). In contrast to U6 in total RNA, annealing of oSDR464 to in vitro transcribed U6 occurred even at low temperatures, and increased with increasing temperature. One explanation for the discrepancy in behavior of the oligo when incubated with yeast total RNA compared to in vitro transcribed U6 is that the RNAs may not fold into the same structure. In vivo, U6 snRNA is bound by the Lsm complex, which is important for U6 snRNA stability, and Prp24, which has been proposed to stabilize base pairing of the long range interaction between nucleotides 35-39 and 92-96 (Vidaver et al. 1999, Cooper et al. 1995). Conducting the RNA extraction on ice stabilized base pairing even when the proteins were removed by phenol: chloroform extraction. In the absence of these proteins, as was the case for in vitro transcribed U6, the long-range central stem interaction may not form, leading to an overall less stable U6 structure. Thus while the in vitro transcribed U6 may still sequester the ACAGAGA sequence through base pairing in stem/loop A, oSDR464 may disrupt stem/loop A base pairing more easily in the absence of longer range RNA/RNA interactions. While the oSDR464 annealing experiments were supportive of only the DunnRader model, the results were somewhat dissatisfying in that we were unable to observe annealing to U6 following heat treatment when incubated with yeast total RNA. We therefore designed a second oligo, oSDR465, which also spanned the ACAGAGA 39 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. sequence but consisted of a two and three nucleotide extension at the 3' and 5' ends of oSRD464 respectively. This oligo also did not anneal to free U6 efficiently at 4C, but did stably associate with U6 following heat treatment at 70C (Fig. 9, compare lanes 16 and 18). Since the melting temperature of the U4/U6 di-snRNA complex is approximately 53C (Brow & Guthrie 1988), the free U6 snRNA signal following heat treatment represents both the free U6 snRNA as well as U6 snRNA that has been released from the U4/U6 di-snRNA complex. Ideally, if a region of U6 snRNA is accessible in both free U6 and the U4/U6 di-snRNA, then the sum of these signals at 4C should be approximately equal to the U6 signal at 70C. If the annealing site is inaccesible, then annealing at low temperature will be relatively inefficient while annealing following denaturation will be much more efficient, resulting in a much greater U6 signal in heat-treated samples than the combined U6 and U4/U6 signals at low temperatures. The oSDR465/free U6 signal was approximately five-fold stronger than the combined U4/U6 and free U6 signals at 4C, indicating that approximately 80% of the available U6 was inaccessible at low temperatures. While the Dunn-Rader model predicts that this annealing site is inaccessible, all other models of yeast secondary structure predict this region to be single stranded. Further, strong chemical modification in the disnRNP indicates that this annealing site is fully accessible in the U4/U6 complex (Jandrositz & Guthrie 1995). Thus the inefficient annealing to free U6 at 4C and the five­ fold increase in signal upon heating strongly support the Dunn-Rader model where heat treatment of total RNA would disrupt base pairing in stem/loop A, releasing the 40 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ACAGAGA sequence and allowing for efficient annealing between oSDR465 and U6. All other proposed models of U6 snRNA secondary structure should have resulted in much more stable association of the oligo at low temperatures. To test the prediction that disruption of stem/loop A would expose the ACAGAGA sequence to allow for stable association of complementary oligos, we attempted to open the structure by pre-incubating the total RNA preparation with unlabeled oSDR461 prior to incubation with either labeled oSDR464 or oSDR465. Labeled oSDR461 alone did not efficiently anneal to U6 unless heat treated, thus pre­ incubations with the unlabeled oligo were heated to 70C and slow cooled to allow annealing prior to the addition of oSDR464 or oSDR465 (Fig. 9, compare lanes 8 and 9). Unfortunately, these experiments were inconclusive since the short heat treatment resulted in some destabilization of the U4/U6 species and it was consequently not clear whether the increase in intensity of the free U6 signal was the result of increased annealing of open free U6, or if the increase was the result of increased free U6 levels due to the destabilization of U4/U6 base pairing. 0SDRI8O, also known as U6 6D in the literature, has been used extensively by many splicing labs as a probe for Northern blot and solution hybridization since it anneals well to the U6 3' tail in both free U6 snRNA and U4/U6 di-snRNA (Li & Brow 1993). When the sample was heat treated at 70C, a much stronger free U6 signal was observed (Fig 9, lane 3). Some of this signal increase was a result of U4/U6 disassembly due to the high temperature, however the U6 signal at 70C was still approximately two-fold higher than the combined U4/U6 and U6 intensities at 4C. 41 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The increased free U6 signal upon heating suggests that the region of U6 that is complementary to 0SDRI8O was moderately inaccessible at lower temperatures. While all models of yeast U6 snRNA secondary structure proposed prior to the Karaduman et al. (2006) and Dunn-Rader models predict the 0SDRI8O annealing site to be single stranded, the central stem in the Karaduman et al. (2006) and Dunn-Rader models predict half of this annealing site to be occluded through base pairing (Fig. 9). Thus a possible explanation for the two-fold increase in signal at 70C is that at lower temperatures 0SDRI8O was capable of annealing to the 3' half of the U6 annealing site, but that this interaction, with an estimated melting temperature of approximately 22C, was relatively unstable and may not have allowed for efficient annealing. Upon heating, denaturation of the central stem may have released the remaining half of the 0SDRI8O annealing site, allowing for much more stable association where the oligo/U6 complex had an estimated melting temperature of approximately 45C. Interestingly, the annealing pattern of oSDR467 matched that observed for 0SDRI8O where some annealing occurred at 4C and an approximately two-fold increase in annealing was observed upon heat treatment (Fig. 9). oSDR467 was designed to anneal to the opposite strand of the central stem, and like 0SDRI8O, extended into a single stranded region 3' of the stem. The annealing pattern observed for oSDR467 can be explained in the same way as that observed for 0SDRI8O; that is, at low temperatures some oligo annealing was observed because the single stranded region of U6 could serve as a nucleation site for annealing, however the interaction was through so few residues that it was relatively unstable. Heating the RNA would have disrupted base pairing in the 42 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. central stem, allowing for more stable association of the oligo. The identical annealing patterns for similarly constructed oligos designed to anneal to opposite strands of the same stem offer strong support for the presence of the stem structure. In contrast to 0 S D R I 8 O and oSDR467, oSDR466, which was complementary to the same strand of the central stem as 0SDRI8O but extended to the 5' side of the stem into the 3' strand of stem/loop B, did not efficiently anneal to U6 at 4C or 70C (Fig. 9). This result was not too surprising since this annealing site in U6 was predicted to be base paired through both the central stem and stem/loop B, and further, when the annealing site was located in the three-dimensionally modeled structure, it became evident that portions of this region of U6 may be buried within the overall U6 structure (Fig. 9). The stem/loop B hairpin itself is quite stable, with an estimated melting temperature above 40C and in the context of the overall three-dimensional structure, may be even higher. Thus cooling on ice following the heat treatment at 70C may result in rapid regeneration of stem/loop B, resulting in the destabilization of any remaining oSDR466/U6 interactions. The difference in annealing observed for 0 S D R I 8 O and oSDR466 brings into question the interpretation of the results of a similar experiment conducted by Brow and Vidaver (1995) using HeLa cell nuclear extract. In this experiment an oligo that was complementary to nucleotides 85-99 of the human U6 snRNA sequence, hU63', showed the same inability to anneal to U6 as oSDR466, even when heat-treated to 90C and slow cooled. This observation has long since been interpreted as support for the proposal that yeast and mammalian U6 snRNA fold into different secondary structures, however, the 43 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. hU63' annealing site did not extend all the way through to the 3' end of human U6 like 0SDRI8O in yeast, but instead left seven nucleotides 3' of the annealing site. oSDR466, while spanning much of the same region of yeast U6 as 0SDRI8O, also did not extend all the way through to the 3' end of the molecule. We have shown here that although oSDR466 was only shifted from the 3' end of the molecule by eleven nucleotides, this shift was enough to generate dramatically different annealing behaviors between the two oligos. We suggest that oSDR466 is a more appropriate candidate for comparison to hU63' since the annealing sites are more similar than that of 0 S D R I 8 O . oSDR466 and oSDR467 were designed to anneal to similar regions of yeast U6 snRNA as two oligos designed by Brow and Vidaver (1995) to anneal to human U6, hU63' and hU6cen2 respectively. While hU63' did not anneal to free U6, hU6cen2 strongly associated with both free U6 and U4/U6. The authors attributed this discrepancy in annealing to an unexplained ability of hU6cen2 to more easily invade the proposed stem structure. However, the similar annealing capabilities observed here for the equivalent oligos in yeast, oSDR466 and oSDR467 respectively, suggest that the human and yeast structures are much more similar in this region than previously thought. Thus we propose that like oSDR466, hU63' was complementary to a region of human U6 that was largely inaccessible while hU6cen2 annealed to a region of human U6 that was at least partially accessible, much like oSDR467. This is consistent with the latest human U6 snRNA secondary structure proposed by Karaduman et al. (2006) where a central stem, similar in composition to the yeast central stem, has been proposed for human U6. 44 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Interestingly, annealing of hU63' to free U6 snRNA can be stimulated by first preincubating the HeLa cell extract with unlabeled hU6cen2 (Brow & Vidaver 1995). We conducted the same experiment using yeast total RNA pre-incubated with unlabeled oSDR467 to determine whether or not annealing of oSDR466 could be similarly encouraged. Although the oSDR466 signals were often quite weak, quantification of the bands revealed similar annealing to U4/U6 and an approximately 1.5 fold increase in annealing to free U6 (Fig. 9). In contrast, pre-incubation of yeast total RNA with oSDR466 prior to incubation with labeled oSDR467 did not stimulate annealing of oSDR467 to either free U6 or U4/U6 (Fig. 9). In conclusion, all of the oligos tested supported the Dunn-Rader model with the most important of those being the ACAGAGA complement, oSDR465, whose inability to anneal at low temperatures was only consistent with the Dunn-Rader model. The two-fold increase in annealing of heat-treated incubations of 0SDRI8O and oSDR467 was consistent with partial occlusion of the complementary U6 sequence through base pairing in the central stem at lower temperatures. The inefficient annealing of oSDR466 at both 4C and in heat-treated incubations was consistent with the complement of this oligo being both base paired and buried within the U6 structure. The comparable behavior of oSDR466 and oSDR467 to two oligos designed to anneal to human U6 snRNA suggested that the yeast and human U6 snRNA secondary structures are more similar than previously thought. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter Four - Genetic Support for the Lower Central Stem The central stem was proposed by Karaduman et al. (2006) and is therefore not unique to the Dunn-Rader model of U6 snRNA secondary structure. Nevertheless, this structural element should be tested experimentally since evidence for its existence is still lacking. Protection of residues from chemical modifying reagents in the 5' strand of the stem suggests that these residues are in fact base paired, however chemical modification data for the 3' strand of the stem is unavailable since the primer extension oligonucleotide anneals to U6 snRNA in this region (Jandrositz & Guthrie 1995, Karaduman et al. 2006). Further, the upper half of the central stem has been supported by compensatory mutation, however the lower half has not been subjected to genetic analysis (Vidaver et al. 1999, Ryan et al. 2002). To test the lower half of the stem, we designed three double mutations in the bottom of the central stem and subsequently evaluated the in vivo effect of the mutations on cell growth, as well as U4/U6 assembly. 4.1 Materials and Methods Generation of Mutant SNR6 Point mutations were introduced into SNR6 via PCR-mediated site directed mutagenesis of pSR85 (pUC + 1.2 Kb EcoRI-PstI genomic fragment containing S. cerevisiae SNR6) using 2.5 units of Pfu Turbo AD DNA polymerase (Stratagene). PCRs were treated with 20 units of Dpnl (NEB) for 7 hours to ensure complete digestion of wild type SNR6 template plasmid. Dpnl digests were subsequently ethanol precipitated and transformed into RbCh competent DH5a by standard methods. Transformations were plated onto LB plates containing 0.05mg/mL carbenicillin and incubated overnight at 46 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 37C. A single colony of each mutant was selected and grown in liquid LB containing 0.05mg/mL carbenicillin overnight at 37C on a shaker at 200rpm. Plasmid DNA was collected from each strain through alkaline lysis and approximately 20|ag was subsequently subjected to restriction enzyme digest with 20 units each of EcoRI (NEB) and SphI (NEB) for 3 hours at 37C. The SNR6 containing EcoRI-SphI fragment was inserted into the yeast vector pSE358, which carries a TRP1 marker, in a standard ligation reaction containing 400 units of T4 DNA ligase (NEB) and a 3:1 molar ratio of insert to vector. Ligation reactions were incubated for 1.5 hours at 37C, heat inactivated for 20 minutes at 65C, and subsequently transformed into RbCh competent DH5a, plated on LB - carbenicillin plates, and incubated as above. A single colony of each mutant was grown up in liquid culture as above and the plasmid was subsequently isolated by Qiagen® plasmid mini-prep according to the Qiagen mini-prep handbook. Two micrograms of each plasmid in a total volume of 20|xL was sent to UNBC sequencing facilities to ensure the presence of the appropriate mutations. PCR Primer Pairs: U6-G30U/G31U 0SDR551 FWD: 5 '-CCCTTCGTGGACATTTTTTCAATTTGAAACAATAC-3' OSDR552 REV: 5'-GTATTGTTTCAAATTGAAAAAATGTCCACGAAGGG-3' U6 - U100C/U101C OSDR547 FWD: 5'-CCGTTTTACAAAGAGACCTATTTCGTTTTTTTTTTATC-3' OSDR548 REV: 5' -GATAAAAAAAAAACGAAATAGGTCTCTTTGTAAAACGG-3' U6 - U100A/U101A oSDR549 FWD: 5'-CCGTTTTACAAAGAGAAATATTTCGTTTTTTTTTTATC-3' OSDR550 REV: 5'-GATAAAAAAAAAACGAAATATTTCTCTTTGTAAAACGG-3' Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Yeast Transformations and Plasmid Shuffle Mutant SNR6 sequences were introduced into the yeast strain YHM1 (genomic deletion of SNR6 covered by a wild type SNR6 carried on a URA3 marked YCp50 plasmid) by the plasmid shuffle technique. Approximately 15|xL ofYHMl was collected from an overnight plate patch and resuspended in 100|jL of TEL [lOmM Tris pH 7.5, 3.33mM EDTA, lOOmM lithium acetate]. Forty micrograms of sheared salmon sperm DNA, approximately 500ng of Qiagen® prepped mutant plasmid, and lmL of PEG-TEL [lOmM Tris pH 7.5, 3.33mM EDTA, lOOmM lithium acetate, 40% PEG-4000] were added to the cells and mixed well. Following an overnight incubation at room temperature, the transformations were heat shocked at 42C for 10 minutes and then centrifuged for 30 seconds at 3000rpm. As much PEG-TEL was removed as possible and cells were resuspended in 200jiL of TE [lOmM Tris pH 7.5, 3.33mM EDTA], The transformations were plated on trp (-) media and incubated at 30C for three days. A single colony of each mutant was selected and streaked onto fresh trp (-) media and incubated at 30C for three days. These plates were then replica plated in the order trp (-), ura (-), 5FOA, and YPD, and all plates were incubated at 30C for three days. A single colony of each mutant that grew on trp (-), 5-FOA and YPD, but not ura (-) was selected and grown overnight to saturation at 30C in 3mL of liquid YPD. Glycerol stocks of each mutant were made by mixing one part of cell culture to two parts of 50% glycerol. Stocks were snap frozen in liquid nitrogen and stored at -80C. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Growth Phenotypes by Dot Dilution Each U6 mutant yeast strain was grown up in 3mL of liquid YPD to an OD595 between 2.0 and 3.0. A five-fold dilution series of eight dilutions was generated with the most concentrated of the set having a concentration of 1 OD/mL. Dilution series were set up in 96 well plates and were plated onto three different YPD plates with a frogger. The plates were then incubated at 18C, 30C, or 37C for three days. Preparation of Yeast Total RNA Each U6 mutant containing yeast strain was grown up from the glycerol stock on YPD plates at 30C for three days. A single colony of each was then grown overnight in liquid YPD at 30C to an OD595 between 1.5 and 2.5. One milliliter of culture was collected for cold phenol: chloroform total RNA extraction (see chapter 3). Two 500|iL aliquots of culture were collected and spun down for temperature shift. One pellet was resuspended in IOOJXL YPD pre-incubated to 18C while the other was resuspended in IOO^IL YPD YPD pre-incubated to 37C. The resuspended cells were used to inoculate 2mL of that had been pre-incubated at 18 or 37C. Temperature shifted cultures were incubated at the appropriate temperature for three hours and then lmL was collected for cold phenol: chloroform total RNA extraction as before. Northern Blots Ten micrograms of total RNA was loaded onto a pre-chilled 4.5% non-denaturing polyacrylamide gel and samples were subjected to electrophoresis in the cold room for 1.5 hours at 250V. The RNA was transferred to Amersham Hybond-N+ nylon transfer membrane (GE Healthcare) for 30 minutes at 450mA in a semi-dry electro-blotter (Owl 49 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Panther Hep-3) and was subsequently cross-linked to the membrane in a UV Stratalinker 1800 (Stratagene) with 120,000 J of ultraviolet radiation. The membrane was blocked with 7mL Rapid-hyb Buffer (GE Healthcare) for 45 minutes at 65C, and blocking continued while the temperature was slowly lowered to approximately 30C (approximately 30 minutes). 5' end labeled U4 probe oligonucleotide (U4 14B) was added and allowed to incubate with the membrane for one hour at 30C. The membrane was washed with 7mL Wash Solution [0.2% SDS, 6X SSC] four times: the first for two minutes, the second for 20 minutes, and the remaining two washes for 40 minutes. The membrane was exposed to a phosphor imager screen overnight and the resulting autoradiogram was visualized and quantified with a Cyclone phosphor imager and OptiQuant Software (Packard Instruments). Membranes were stripped with boiling Northern Strip Solution (0.1X SSC, 0.1% SDS, 40mM Tris pH 7.5) twice for 10 minutes and were re-probed with 5' end labeled U6 probe oligonucleotide (U6 6D) as described above. 4.2 Results and Discussion Genetic manipulations have been used extensively to test predicted base pairing interactions throughout splicing, providing strong support for interactions between U6 and other splicing snRNAs as well as the U6 3' ISL and lower telestem (Madhani & Guthrie 1992, Lesser & Guthrie 1993, Fortner et al. 1994, Vidaver et al. 1999, Ryan et al. 2002). Mutations that increase the stability of a stem are thought to generate a cold sensitive growth phenotype since lowering the temperature further stabilizes the already hyperstabilized structure. Consequently the stem may be more difficult to unwind for 50 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. subsequent structural re-arrangements. Mutations that destabilize a stem are thought to exhibit a heat sensitive growth phenotype since increasing the temperature is predicted to destabilize the structure further, hindering its formation in vivo. The double mutations constructed here were designed to either hyperstabilize (U100C/U101C) or destabilize (G30U/G31U and U100A/U101A) the bottom two central stem G*U wobbles and were therefore predicted to generate a cold or heat sensitive growth phenotype respectively. To test the central stem mutations in vivo, a yeast strain carrying only the mutant U6 snRNA gene, SNR6, was constructed to ensure that any phenotype observed was the result of introducing the mutant sequence. The genomic copy of SNR6 was replaced by LEU2 in the yeast strain employed in this experiment, and since U6 snRNA is an essential gene product, a wild type copy of SNR6 carried on a URA3 marked plasmid was present to cover the genomic deletion. Mutant SNR6 sequences were introduced by plasmid shuffling where the mutant sequences, carried on a TRP1 marked yeast vector, were transformed into the yeast strain (Fig. 10). The presence of TRP1 allowed these yeast strains to grow in media lacking tryptophan and therefore selective pressure to take up the TRP1 marked plasmid, and consequently the mutant SNR6, was applied by growing the strain on trp (-) plates (Fig. 10). No selective pressure was applied to maintain the wild type SNR6 carried on the URA3 marked plasmid and this plasmid was eventually lost. The inability to grow on media lacking uracil and the ability to grow on media containing 5-Fluoroorotic Acid (5-FOA), a chemical that in the presence of the URA3 gene is converted to fluorodeoxyuridine, which is toxic to yeast, confirmed that this plasmid had been successfully shuffled out (Fig. 10). 51 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. A B C C A U •G Stem/loop A A— U G U U A G" cc SNR6 LEU2 TRP J SNR6 L I URA3 ,c A '- c ™G -U 70 A' -G .C I LEU2 A A A C A, J G 5'Stem/loop A 80 1URA31 V . >G A A A A C U C ^ S N R 6 ^SSNR6 Stem/loop B AG-C ' 'GC U—A SNR6 10 G—C U-A A—U U —A A—U A o G Central Stem G-C A A C — G 20 C-G G •U C-G U • G U *- C, • U - •C,A •C,A U —A U <-! G U A U U U C G U U U U 3' 5' G— CAUUU LEU2 URA3 I *SNR6* U^A TRP1 » uH -TRP -URA 5-FOA YPD Figure 10. Construction of U6 snRNA central stem double mutants. (A) The location of the double mutants are shown on the Dunn-Rader model with a shaded box. (B) The plasmid shuffle strategy employed to introduce mutant SNR6 sequences, indicated with asterisks, into the SNR6 genomic deletion yeast strain. (C) The constructed double mutants, G30/G31U, U100A/U101A, and U100C/U101C plated on selective media. 52 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The growth phenotype of the mutant U6 snRNA strains were tested by dot dilution on YPD plates incubated at 18C, 30C, and 37C for three days. All yeast strains tested, including the wild type control and a strain serving as a control for cold sensitivity, U6-A62G, showed wild type growth at 30C (Fig. 11). U100A/U101A also showed wild type growth at 18C and 37C while U100C/U101C showed wild type growth at 37C and severe inhibition of growth at 18C (Fig. 11). Interestingly, G30U/G31U did not exhibit wild type growth at elevated or lowered temperatures. A weak growth defect at 37C and a severe growth defect at 18C were observed for this strain (Fig. 11). The Karaduman et al. (2006) and the Dunn-Rader models of U6 snRNA secondary structure predict that the U100A/U101A and G30U/G31U mutations would destabilize the central stem. Such destabilization may be expected to push the equilibrium between free U6 and U4/U6 toward U4/U6 formation, and since this is the direction that spliceosome assembly proceeds, it was not clear if driving formation of the di-snRNA complex would generate a detectable growth phenotype at elevated temperatures. In fact, the U100A/U101A mutation grew as well as the wild type strain at 37C, suggesting that maintenance of the lower portion of the stem may not be critical for growth if the central stem in fact exists. Additionally, this result indicated that mutating nucleotides U100 and U101 to adenosines did not interfere with any RNA-RNA or RNA-protein interactions that may be important for progression through splicing. In contrast, the weak heat sensitivity of the G30U/G31U mutant observed at 37C was consistent with central stem destabilization. However, since the G30U/G31U and U100A/U101A mutants were predicted to destabilize the same base pairing interactions, 53 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 37C 30C 18C % • • A62G • • U100C/U101C # • UIOOA/UIOIA * m •• G30U/G31U wt G30U U100C A62G G31U U101C wt G30U U100C A62G G31U U101C wt G30U U100C A62G G31U U101C W303 U4/U6 Figure 11. Growth phenotypes and U4/U6 assembly phenotypes of the central stem double mutants, a wild type control, and a control for cold sensitivity, U6-A62G. (A) U6 mutant growth phenotypes were tested by serial dot dilution on YPD plates that were incubated for three days at the temperature indicated. (B) U4/U6 assembly Northern blot probed for U4 snRNA. Total RNA was extracted from the central stem mutants after shifting to the temperature indicated and was separated on a 4.5% non-denaturing polyacrylamide gel. the slightly different growth phenotypes exhibited by these mutants suggest that either the proposed base pairing interaction does not form in vivo, or that G30 and G31 are involved in some additional interaction(s) that is also disrupted by the double mutation. While we cannot rule out the first interpretation, we favor the latter since growth of this mutant 54 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. strain was severely inhibited at 18C and total RNA extracted from this strain grown at 18C, 30C, and 37C revealed a defect in U4/U6 assembly (Fig. 11). The U4/U6 assembly defect was evident by the increased levels of free U4 snRNA and diminished levels of U4/U6 in this strain compared to a wild type strain. These observations suggest that mutation of these nucleotides either increases the stability of free U6 or shifts the equilibrium toward free U6 by destabilizing interactions with U4. The G30U/G31U mutation may destabilize a third phylogenetically conserved helix in U4/U6 that has been proposed at positions 35 to 42 in yeast U6 snRNA (Jakab et al. 1997). Introduction of a tri-nucleotide bulge in U4 snRNA could potentially extend base pairing by an additional eight base pairs to generate a relatively stable helix composed of fourteen base pairs with a di- and tri-nucleotide bulge in the U6 and U4 strands of the helix respectively. This lengthened helix would extend through both G30 and G31, which would be predicted to base pair with U4-C80 and C81. Thus the G30U/ G31U mutation may result in a U4/U6 assembly defect by destabilizing this U4/U6 interaction, preventing efficient annealing of U4 and U6 during di-snRNP formation. Strong protection of G30, G31, and C33 from chemical modification in the U4/U6 disnRNP suggests that this helix may form in yeast, however it is also possible that the protection of these nucleotides may be due to the presence of proteins in the di-snRNP particle (Jandrositz & Guthrie 1995). An alternative explanation for the unexpected G30U/G31U phenotypes is that mutation of these residues may increase the stability of U6 snRNA/protein interactions in free U6 snRNP. G30 has been UV cross-linked to Lsm2 of the Lsm complex and 55 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. neighboring residues U28 and U29 have been UV cross-linked to both Prp24 and Lsm2 (Karaduman et al. 2006, 2008). Thus it is possible that introducing additional uridines may provide a stronger binding site for one or more of these proteins. The Lsm complex in particular is known to bind the uridine rich 3' tail of U6 snRNA and the proximity of this natural Lsm binding site to the lower central stem, where we have introduced additional uridine residues, may result in aberrant Lsm binding. Strengthening the interaction between U6 snRNA and its associated proteins may account for the severe cold sensitive growth phenotype and U4/U6 assembly defect exhibited by this mutation. It is interesting to note that the U4/U6 assembly defect of G30U/G31U was present at 30C where the strain exhibited wild type growth. A similar result was observed for U6-A62G where a U4/U6 assembly defect was present at 18C and 30C while growth was severely inhibited or comparable to wild type at these temperatures respectively (Fortner et al. 1994). Additionally we have shown that although the U4/U6 assembly defect of A62G persisted at 37C, growth was also comparable to wild type at this elevated temperature (Fig. 11). These observations demonstrate that yeast growth phenotypes are not sensitive enough to detect all molecular disruptions, and further, that the mechanisms governing the generation of growth phenotypes are not well understood at this time. In contrast to the destabilizing mutations, U100C/U101C was predicted to hyperstabilize the central stem, which was consistent with the observed severe cold sensitivity at 18C. The central stem is expected to destabilize at some point throughout the splicing cycle to accommodate interactions between U2 and U6 to form U2/U6 helix 56 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. II. However, at low temperatures where central stem destabilization may be less efficient, U2/U6 helix II formation may be impeded. This interpretation is consistent with the observation that the U100C/U101C mutation did not result in a defect in U4/U6 assembly, suggesting that the associated growth defect is generated following U4/U6 disnRNP formation (Fig. 11). Since the U4/U6 interaction region of U6 lies outside of the central stem, disruption of the central stem at an early stage in spliceosome assembly may not be important, however disruption of the central stem following U4/U6 association may be a critical step in U6 activation in functional spliceosomes. Human U6 snRNA has been proposed to carry a long-range intramolecular interaction of the 3' tail and an internal region of U6 through to the di-snRNP where RNase VI cleavages on the 5' strand of the helix lend some support to this proposed base pairing interaction (Brow & Vidaver 1995, Mougin et al. 2002). In fact, Brow and Vidaver (1995) proposed that U2/U6 helix II formation actually stabilizes the U4/U6 interaction by antagonizing reformation of the U6 long-range intramolecular interaction. Thus it is possible that the proposed central stem does not disassemble until U6 and U2 snRNA come into contact, at which time disassembly of the central stem may allow for formation of both the third U4/U6 helix and U2/U6 helix II. In conclusion, we have shown that the double mutations generated in the lower half of the central stem produce growth phenotypes and U4/U6 assembly defects consistent with either disruption or hyperstabilization of this stem. Surprisingly, however, mutations in the predicted central stem region also give rise to unexpected phenotypes that cannot be easily explained by existing models. These phenotypes suggest that the 57 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. central stem, while forming in the free U6 snRNP, may persist through to the di- and trisnRNP complexes until U6 engages in interactions with U2 snRNA. Much of this hypothesis should be tested in a more rigorous manner with additional mutational analyses, including compensatory mutations and isolation and identification of both cis and trans acting suppressors of the observed growth phenotypes. Additionally, pull down assays could be performed to assess the level of U6 snRNA complexed with the U6 snRNP-associated proteins to determine whether or not introduction of these mutations has created additional protein recognition sites as proposed. 58 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Chapter Five - Co-variation of U6 snRNA Secondary Structure and Prp24 Homolog Architecture RNA sequence can change throughout evolution without dramatically altering the RNA secondary or tertiary structure, as long as mutations that occur in regions important to the overall structure are compensated for by additional mutation(s) (Leontis et al. 2002). This principle has opened up an entire field of RNA structure prediction based on the idea that nucleotide co-variation at a base pair position confirms the existence of that base pair, and that co-variation at only a few positions within a putative helix can confirm the existence of that helix (Woese et al. 1983, Gutell et al. 1994). As a testament to the power of comparative sequence analysis, approximately 97% of the predicted 16S and 23S rRNA base pairs have since been confirmed in high resolution crystal structures (Gutell et al. 2002). Early comparative sequence analyses were strictly based on co-variations between the standard Watson-Crick base pair interactions, G-C, A-U, and G*U (Woese et al. 1983). It is now clear that many non-canonical base pairs occupy the same or very similar geometries as the standard base pairs and can thus substitute without substantial perturbation of the RNA structure (Leontis et al. 2002). This chapter examines the conservation of the Dunn-Rader model of U6 snRNA secondary structure throughout evolution by analyzing sequence co-variations within U6 snRNA from more than 70 different organisms, including U6atac from the minor spliceosome of plants and animals. This analysis revealed a pattern to some of the differences in U6 snRNA sequence and may represent an evolving secondary structure. Interestingly, these patterned changes in 59 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. U6 snRNA sequence and/or secondary structure are accompanied by structured differences in the protein domain architecture of Prp24 homologs. 5.1 Materials and Methods U6 snRNA and U6atac sequences were obtained from the NCBI website using the BLAST algorithm and from the Rfam database (Altschul et al. 1990, Gardner et al. 2009). Multiple sequence alignments were generated with CLUSTALW in the software program MEGA 4.0 and the alignment was subsequently optimized by manual inspection (Thompson et al. 1994, Tamura et al. 2007). NCBI accession numbers are given for each sequence in figure 12. U6 snRNA evolutionary trees were constructed in MEGA 4.0 under bootstrap conditions of 1000 replicates where both transitions and transversions were allowed to occur at a uniform rate. Prp24 homologs were obtained from the NCBI website using the BLAST algorithm and protein domains were identified with the Simple Modular Architecture Research Tool (SMART) at http://smart.embl-heidelberg.de/ in normal mode with default settings (Schultz et al. 1998). Homologs used in our analysis are listed below along with a description of the BLAST hit and the NCBI accession number. Table 2. Putative Prp24 Homologs Description Organism Name NCBI Accession # S. cerevisiae Prp24 AY723858 V. polyspora Hypothetical proteinKpol_1069p8 XP_001643044 K. lactis Unnamed protein product XP_454447 M. grisae 70-15 hypothetical protein MGG 04727 XP362282 S. pombe RNA-binding protein Prp24 NP596721 A. fumigatus Af293 pre-mRNA splicing factor Prp24 XPJ749317 60 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Description Organism Name NCBI Accession # N. crassa OR74A hypothetical protein NCU09048 XP958552 A. capsulatus NAml conserved hypothetical protein XP001536207 P. anserina Unnamed protein product XP001903653 D. hansenii CBS767 hypothetical protein DEHA0F08250g XP460681 L. elongisporus NRRL YB-4239 hypothetical protein LELG 00207 XP001527687 Y. lipolytica Hypothetical protein XP504884 P. stipitis CBS 6054 U4/U6 snRNA-associated splicing factor XP_001384763 P. nodorum SN15 hypothetical protein SNOG11646 EAT80690 C. albicans SC5314 hypothetical protein CaO19.10095 XP_718326 H. sapiens Squamous cell carcinoma antigen recognized by T cells NP055521 R. norvegicus Squamous cell carcinoma antigen recognized by T cells NP001100626 D. melanogaster RNA binding protein CAA75535 O. sativa 0s08g0113200 NP_001060838 C. elegans Protein B0035.12, partially confirmed by transcript evidence CAA97405 5.2 Phylogenetic Conservation of U6 snRNA Secondary Structure We have generated a comprehensive alignment of U6 snRNA sequences from 72 different organisms in order to identify positions of genetic co-variation in our proposed stem structures (Fig. 12). Although these sequences span all eukaryotic kingdoms, more than half of the sequences collected are from fungal organisms while the remaining sequences include representatives from other unicellular organisms, the euglenozoa, through to the most complex plants and animals. Analysis of the U6 snRNA sequence from such a diverse group of organisms should allow for the identification of subtle 61 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. '3 88 83 8 t «^ g'g1 § §|§ | || || 3 3 3 3 * glllgiliBalu&ooBoooouoSooooouu&oiDu&uuuo 5 mi s! !!!!!!! sssssss SlllfII ,51 — fo3oSSo3S3SS3SySoSuS5SSS3(S3SSSS5ScSoopoo5 fl g" 5s33333322522525222532222222233322««2««« StS3S4ooooooiiSot)ooooooyooi3(5ooo353S33SSo w— If g f | | « «?> i Iii ii 4 If i$ }$ f 8111 WWB>B>0iC>Wo>^o>&frj?g'i?5pgiJ?iTicn&ilyEn£n§iwcncncncpcpwijititjitj!CT>cncf> 33 ffff - 8 8 8 § 8 8 H-8 8 8 is 8 § § a'8 8 8 8-8 8 8 8 8-8 8 8 8-8 8 8 8 8-8885 iiiiiimiimfmmmmmjjjMlw ;S111II11!1111 I| |IP-S' |l-S'111 -S'-g'-g'11II11IIII -S'-g'-S'-S'-S'-S'-S'-S'-g'-g'-III S' S1 S1III1 ST E 2-3 § ii i § § i .lll-llli; '•gfi II -L= if ill-l-IN-ll 11-311113-13 s i-s 11 l-i nmum-M Hill!! 'liliili • S«i ^^ !iiililill!iiillliii?illl!!!l ^n&j{>&>§i§;§ii>l»ini»8iI>§ii|§iln§i6&oifn&6>i>i>i[§)§*.§>ii§i§>§i ' g,g»"? 7'j i 3 3 3 3 3" | 3 35 Iff lit! :ISI rill .8 fi u ' & D>&&8>&S&§i3i8>SicnenwwenDioiBrwoio)o>cnoiwo>craig>g>''''' fJ 33 Ii Il Il 35 3 > uuo va t SHSIIS'li? ? !3 3!3!3!3 !3 3!U!( !7 iBi U : 3 3 § g 3 § § 3 |tU T> JiT (555raS§,!SS^3o ss1 i i 0 i2S2S3S££°i2£ 0 ,2g 3S6sS£ 6 3 § S55555g5SdS95S15ill5l595Id55edS5Sy55y§5 HI N & 6 o u r> A Q « o 3 M i-i "|S "l ~|3 b 5 D E "|S b 5 S T <* F E I a EL, IHE 9 is sifi§5-2 £! 3 582-SS51S^ 1 g«s 3 i S "S § 2 h rt - 2 I 5 I 2 "3 i !| o ? « u 11