COSTBI-455; NO OF PAGES 7 Ribozymes William G Scott The structural molecular biology of ribozymes took another [4,5] 70S ribosomes have now provided us with many great leap forward during the past two years. Before ribozymes new structural insights. However, the main focus of this were discovered in the early 1980s, all enzymes were thought review will be to survey the many advances made in to be proteins. No detailed structural information on ribozymes understanding the structural basis of ribozyme catalysis in became available until 1994. Now, within the past two years, other RNA systems. In reviewing these advances, it near atomic resolution crystal structures are available for seems noteworthy that many of the most important ones almost all of the known ribozymes. The latest additions include emerged from laboratories that are not primarily known as ribonuclease P, group I intron structures, the ribosome (the structural biology research groups, but rather are labora- peptidyl transferase appears to be a ribozyme) and several tories that focus primarily on the RNA molecules them- smaller ribozymes, including a Diels–Alderase, the glmS selves. Examples include Harry Noller’s ribosome group, ribozyme and a new hammerhead ribozyme structure that Norm Pace’s RNase P group, Tom Cech’s intron group reconciles 12 years of discord. Although not all ribozymes are and Scott Strobel’s RNA chemistry group. The blurring of metalloenzymes, acid-base catalysis appears to be a universal such distinctions will probably be a growing trend in the property shared by all ribozymes as well as many of their future of macromolecular structural studies and is an protein cousins. encouraging development. Addresses Department of Chemistry and Biochemistry, and The Center for the All ribozymes were believed originally to be metalloen- 2+ Molecular Biology of RNA, 228 Sinsheimer Laboratories, University of zymes, requiring Mg or other divalent metal ions for both California at Santa Cruz, Santa Cruz, CA 95064, USA folding and catalysis. A ‘two-metal mechanism’ had been proposed [6] in which hydrated Mg2+ ions played the roles Corresponding author: Scott, William G ([email protected]) of general acids and bases. This prediction appears to have been correct for the group I intron (the specific context of Current Opinion in Structural Biology 2007, 17:1–7 the original proposal), as has now been revealed by Stahley and Strobel [7] (described below), but perhaps the most This review comes from a themed issue on Nucleic acids striking result is that it does not generalize to all ribozyme Edited by Dinshaw J Patel and Eric Westhof systems. Acid-base catalysis appears to be a catalytic strategy so fundamental that it occurs in both protein and RNA enzymes; in many cases, it seems that the 0959-440X/$ – see front matter RNA itself, rather than acting as a passive scaffold for # 2007 Elsevier Ltd. All rights reserved. metal ion binding, is an active participant in acid-base catalysis in the sense that nucleotide functional groups, DOI 10.1016/j.sbi.2007.05.003 rather than metal complexes, often mimic the roles that amino acids play in the active sites of protein enzymes. Several of the small self-cleaving RNAs as a consequence Introduction do not strictly require divalent metal ions for catalysis [8] Ribozymes are enzymes whose catalytic centers are com- and no divalent metal ions have yet been observed in the posed entirely of RNA and therefore do not require active site of the peptidyl transferase, the ribozyme that is proteins for catalysis (although many exist naturally as embedded in the ribosome. RNA–protein complexes). Their discovery in the early 1980s provided a potential logical escape from the con- Ribonuclease P undrum that plagues evolutionary molecular biology: Ribonuclease P (RNase P) was the first true RNA enzyme which came first, information-encoding nucleic acids or identified [9]. An RNA–protein complex, the catalytic proteins — including the enzymes needed to replicate subunit of bacterial RNase P is composed entirely of nucleic acids? If RNA can do both, it is a potential self- RNA (and it is thought that this is the case with the replicase and the ‘RNA world’ hypothesis is based on the eukaryotic version as well). It processes precursor tRNAs assumption that pre-biotic self-replicating molecules and other RNAs required for cellular metabolism. were composed of RNA. Although structural fragments of RNase P have been previously elucidated, two structures of the entire cata- Perhaps the most important (apparent) ribozyme is the lytic RNA subunit finally appeared simultaneously in ribosome, as the 23S RNA appears to comprise most if not 2005, one from Thermotoga maritima at 3.85 A˚ resolution all of the active site [1,2]; near atomic resolution struc- [10] and the other from Bacillus stearothermophilus at tures of the Escherichia coli [3] and Thermus thermophilus 3.3 A˚ resolution [11]. Both ribozyme structures were www.sciencedirect.com Current Opinion in Structural Biology 2007, 17:1–7 Please cite this article in press as: Scott WG, Ribozymes, Curr Opin Struct Biol (2007), doi:10.1016/j.sbi.2007.05.003 COSTBI-455; NO OF PAGES 7 2 Nucleic acids obtained in the absence of bound substrate and protein eral insertions that are characteristic of phage introns form non-catalytic subunit, so active site interactions between a ring that completely envelops the active site (Figure 2a). enzyme and substrate can only be inferred. The B. stearothermophilus fold is depicted in Figure 1. A model A Diels–Alderase ribozyme for the structure of eukaryotic RNase P has also been Possibly every crystallographer’s favorite reaction from proposed based on these crystal structures [12]. organic chemistry is the Diels–Alder reaction, as it reveals how orbital symmetry dictates reactivity between a diene Group I intron and a dienophile. Proteins and RNAs have both been The group I intron was the first catalytic RNA discovered evolved in vitro that enzymatically catalyze this reaction. [13] and, in 1986, it was demonstrated to be a true enzyme It is remarkable that RNA, originally thought to be [14]. Four group I intron crystal structures are now incapable of enzymatic catalysis and then thought to available, two of which fall into the current review period. catalyze primarily phosphodiester reactions, is capable Following publication of the Azoarcus [15] and Tetrahy- of catalyzing carbon–carbon bond formation. The protein mena [16] group I intron ribozyme structures in 2004, two Diels–Alderase is a catalytic antibody whose structure is particularly noteworthy structures have appeared. In Jan- known. We now have the structure of a Diels–Alder uary 2005, the structure of a phage Twort group I ribo- ribozyme in both the unbound and enzyme–product zyme–product complex [17] appeared (Figure 2a), complex states (Figure 3a), revealing that the ribozyme followed in September by a second structure of the uses a combination of proximity, spatial complementarity Azoarcus group I intron ribozyme [7], which revealed and electronic effects (Figure 3b) to activate stereoselec- evidence of a two-metal-ion mechanism in ribozyme tive catalysis, reminiscent of the protein Diels–Alderase catalysis (Figure 2b,c). [19]. The folds of the various group I introns are quite similar The glmS ribozyme [18], permitting comparisons between molecular species. The glmS ribozyme is a recently discovered ribozyme that The first Azoarcus structure was in a pre-catalytic state, in is unique in the world of naturally occurring ribozymes in which both exons (the substrate of the reaction in which two respects. First, it is a ribozyme that is also a ribos- adjacent exons are spliced as the intron excises itself) witch. Second, the regulatory effector of the ribozyme, were present. The Tetrahymena group I intron structure glucosamine-6-phosphate (GlcN6P), is actually a func- represents a state in which the 30-terminal v-guanosine tional group that binds to the ribozyme active site and and a metal ion are present in the active site. The newer participates in the acid-base catalysis of RNA self-clea- structures complement these two states with an enzyme– vage. The glmS ribozyme is derived from a self-cleaving product complex, and a complex in which all substrate, RNA sequence found in the 50-untranslated region (50- ribozyme functional groups and predicted metal ions are UTR) of the glmS message; it cleaves itself, inactivating present in the active site. These are observed as a cluster the message, when the cofactor GlcN6P binds. GlcN6P of two metal ions, each of which is coordinated to no less production is thus regulated in many Gram-positive bac- than five phosphate or ribose oxygens in the active site. teria via this ribozyme-mediated negative-feedback The Twort ribozyme in addition reveals how the periph- mechanism. Figure 1 Stereo view of the fold of the RNase P RNA, as determined from B. stearothermophilus. The 50 end of the molecule is in blue and the 30 end is in red. Current Opinion in Structural Biology 2007, 17:1–7 www.sciencedirect.com Please cite this article in press as: Scott WG, Ribozymes, Curr Opin Struct Biol (2007), doi:10.1016/j.sbi.2007.05.003 COSTBI-455; NO OF PAGES 7 Ribozymes Scott 3 Figure 2 Group I intron structures. (a) Cartoon representation of the fold of the Twort group I intron. The 50 end is dark blue and the 30 end is red. Several observed Mg2+ ions are depicted as grey spheres. (b) The previously reported Azoarcus group I intron structure (grey cartoon) shares many structural features with the Twort group I intron. The new Azoarcus structure reveals interactions between two Mg2+ ions (cyan) at the active site and the 50 end of the 30 intron (blue), the 30 end of the 50 intron (red) and the v-guanosine cofactor (green).