View metadata, citation and similar papers at core.ac.uk brought to you by CORE provided by Elsevier - Publisher Connector Current Biology, Vol. 14, 1985–1995, November 23, 2004, 2004 Elsevier Ltd. All rights reserved. DOI 10.1016/j.cub.2004.11.003 Guide RNAs with 5؅ Caps and Novel Box C/D snoRNA-like Domains for Modification of snRNAs in Metazoa Kazimierz T. Tycowski, Alar Aab, duced by small nucleolar RNP particles (snoRNPs). Box and Joan A. Steitz* C/D snoRNPs direct 2Ј-O-methylation, while box H/ACA Department of Molecular Biophysics particles are involved in pseudouridylation (reviewed in and Biochemistry [4]). The RNA components of the snoRNPs select the Howard Hughes Medical Institute sites for modification by engaging in extensive base Yale University School of Medicine pairing interactions with the substrate RNA, while meth- 295 Congress Avenue ylation and pseudouridylation are believed to be carried New Haven, Connecticut 06536 out by a snoRNP core protein specific to each particle class. Box C/D snoRNPs are composed of a small RNA mole- Summary cule, typically 70–90 nt long in vertebrates, and a set of at least four proteins: fibrillarin, which is the methyltrans- Background: Spliceosomal snRNAs and ribosomal ferase [5, 6], Nop56, Nop58, and 15.5 kDa (reviewed RNAs in metazoans contain numerous modified resi- in [4]). All box C/D snoRNAs contain four conserved Ј dues that are functionally important. The most common sequence elements: boxes C and D close to the 5 and Ј modifications are site-specific 2Ј-O-methylation and 3 termini, respectively, and internal copies of each box Ј Ј pseudouridylation, both directed by small ribonucleo- referred to as C and D . The termini usually form a short protein particles. Each particle is composed of a short stem, which together with boxes C and D conform to guide RNA and a set of several proteins. All previously an RNA structural motif termed the kink-turn (K-turn) characterized modification guide RNAs in metazoa are [7], found also in U4 snRNA [8, 9], 23S rRNA [10], and encoded in and processed from introns. prokaryotic mRNAs [11]. The box C/D snoRNA kink-turn Results: We have identified and characterized three motif has been identified as the binding site for the 15.5 novel guide RNAs for conserved 2Ј-O-methylation of U2, kDa protein [7, 12], which is essential for further particle U4, and U12 snRNAs. Two guides, termed mgU2-25/61 assembly, including the binding of fibrillarin [13]. and mgU12-22/U4-8, appear to be independently tran- The spliceosomal snRNPs undergo biogenesis in one scribed as judged by the presence of methylated guano- of two ways. U6 (and most likely U6atac) snRNA is syn- sine caps at their 5Ј ends and upstream promoters simi- thesized by RNA polymerase III (pol III) and is modified lar to those of telomerase RNA. These guide RNAs are and assembled with proteins in the nucleus (reviewed each composed of a canonical box C/D snoRNA and a in [14]). In contrast, U1, U2, U4, and U5 (and most likely novel box C/D snoRNA-like domain, where the CЈ/DЈ U11, U12, and U4atac) snRNAs are made by RNA pol II motif, rather than C/D, can be folded into a conserved and follow a distinct maturation pathway (reviewed in kink-turn structure. The snoRNA-like domains are pre- [14]). These newly transcribed RNAs are exported to dicted to direct 2Ј-O-methylation of invariant G residues the cytoplasm, where they assemble with the Sm core that occupy analogous positions in the U2 and U12 proteins and undergo cap trimethylation. Subsequently, snRNA secondary structures. A third guide, mgU2-19/30 the RNPs are shipped back to the nucleus; there, they RNA, is composed of two canonical box C/D snoRNA are believed to transit through the nucleoplasmic struc- domains encoded within a single intron. tures known as Cajal bodies where they acquire numer- Ј Conclusions: This is the first description in metazoan ous 2 -O-methyl groups and pseudouridines [15]. Addi- cells of 5Ј-capped modification guide RNAs that appear tion of these modifications may trigger further assembly to be independently transcribed. Since plant, yeast, and with snRNP-specific proteins, e.g., the U2-specific SF3a protozoan guide RNAs are mostly independently tran- and SF3b splicing factors [16–18]. Indeed, unmodified scribed, the identification of such RNAs argues that an- U2 snRNA binds the Sm core proteins but fails to assem- cestral metazoans possessed independently tran- ble with the SF3a and SF3b factors [19]. scribed guide RNAs and only later, during the evolution By constructing libraries of small RNAs, several guide of metazoan organisms, did the guide RNA genes shift RNAs for modification of RNA pol II-transcribed spliceo- to introns. somal snRNAs have been recently isolated from verte- brate [20–23], Drosophila [24], and plant [25] species. The vertebrate snRNA guides are more complex than Introduction their counterparts for modifying rRNA; often they consist of two snoRNA domains. These composite guides pos- Stable RNAs in eukaryotic cells are subject to extensive sess either one box C/D and one H/ACA [22] or two box posttranscriptional modification (reviewed in [1]). The H/ACA [23] domains. Others are typical single-domain most common modifications in ribosomal RNAs (rRNA) box C/D or box H/ACA RNAs [20, 22]. Consistent with Ј and small nuclear RNAs (snRNA) are 2 -O-methylation the subnuclear localization of the modification process of selected sugar moieties and conversion of specific [15], these guides are found in Cajal bodies and are often uridines to pseudouridines (reviewed in [2, 3]). Both referred to as Cajal body-specific RNAs or scaRNAs [22, types of modification in rRNA and U6 snRNA are intro- 23]. A tetranucleotide has been identified as a signal responsible for targeting box H/ACA scaRNAs to Cajal *Correspondence: [email protected] bodies [26]. Current Biology 1986 All known metazoan modification guide RNAs are en- the first observation of a doublet box C/D RNA. The coded within introns of either protein coding or noncod- doublet and the mgU2-30 singlet accumulate to equiva- ing genes (reviewed in [27]). Exonucleolytic trimming of lent levels in HeLa cells, while mgU2-19 is less abundant -All three RNAs are precipitable with anti .(%20–%10ف) the spliced and debranched introns gives rise to mature snoRNAs with 5Ј-monophosphate and 3Ј-hydroxyl ter- fibrillarin antibodies (data not shown). mini [28, 29]. A small fraction of yeast and plant snoRNAs The proposed secondary structures of the mgU2-19 exhibit similar gene organization. However, most snoRNAs and mgU2-30 domains in mgU2-19/30 RNA conform to in these organisms are independently transcribed either those of canonical box C/D snoRNAs, with boxes C and as mono- or polycistronic units (reviewed in [30]). In D involved in putative kink-turn formation (Figure 3). yeast, the polycistronic as well as some monocistronic Consistent with this observation, the 5Ј and 3Ј ends of RNAs are processed by RNase III to yield the mature the individual mgU2-19 and mgU2-30 RNAs map to the RNAs [31, 32]. positions common to all box C/D snoRNAs (Figure 3 Here, we describe the identification and characteriza- and data not shown). (The 5Ј ends were mapped by tion of three novel metazoan guide RNAs for 2Ј-O-meth- primer extension, while the 3Ј ends were inferred from ylation of spliceosomal snRNAs. One guide is composed the migration of the RNAs in the gel and the length of of two canonical box C/D snoRNA domains and two the terminal stems.) Unexpectedly, the 5Ј end of the feature a box C/D snoRNA downstream of a novel box doublet mgU2-19/30 extends 28 nts upstream of the C/D snoRNA-like domain. In all cases, the domains are mgU2-19 domain, as determined by primer extension. arranged in tandem. While one guide is encoded in and At present, the functional significance of this extension, appears to be processed from an intron, the two others as well as how it is protected from 5Ј exonucleases, is exhibit novel genomic organization. They are apparently unclear. encoded in intergenic regions and appear to be indepen- dently transcribed. Identification of Guide RNAs with Novel Box C/D snoRNA-like Domains Results Discovery of the doublet organization of mgU2-19/30 RNA prompted us to examine whether recently identified A Doublet Intron-Encoded Box C/D Guide RNA, canonical box C/D RNAs proposed to direct 2Ј-O-meth- mgU2-19/30 ylation of vertebrate RNA pol II-transcribed snRNAs rep- Since the 2Ј-O-methylation patterns of the major resent stable domains of longer RNAs. By constructing spliceosomal snRNAs from several vertebrate species libraries of small RNAs, Huttenhofer et al. [20] identified have been established (compiled in [3, 33]), we predicted three mouse canonical box C/D RNAs, MBII-19, MBII- the sequences of the antisense elements for 2Ј-O-meth- 119, and MBII-382, which they proposed guide the modi- ylation of U1, U2, U4, and U5 snRNAs. We then designed fication of U2, U4, and U2 snRNAs, respectively. Using query sequences for database searches also including a similar approach, Darzacq et al. [22] identified a human boxes C, CЈ, D, and DЈ (see Experimental Procedures MBII-119 homolog termed U91. for details). Using database searches, we have identified the hu- Using the BLAST algorithm, we found an 80 nt long man homolog of MBII-382. Northern blot analysis of sequence in the human database that can specify an MBII-382 and U91 revealed longer RNAs of about 400 RNA containing boxes C, CЈ, D, and DЈ as well as a 10 nucleotides in HeLa cells (Figure 2, lanes 4 and 6).