Efficient RNA 2'-O-methylation requires juxtaposed and symmetrically assembled archaeal box C/D and C'/D' RNPs

Abstract

Box C/D ribonucleoprotein (RNP) complexes direct the nucleotide-specific 2'-O-methylation of ribonucleotide sugars in target RNAs. In vitro assembly of an archaeal box C/D sRNP using recombinant core proteins L7, Nop56/58 and fibrillarin has yielded an RNA:protein enzyme that guides methylation from both the terminal box C/D core and internal C'/D' RNP complexes. Reconstitution of sRNP complexes containing only box C/D or C'/D' motifs has demonstrated that the terminal box C/D RNP is the minimal methylation-competent particle. However, efficient ribonucleotide 2'-O-methylation requires that both the box C/D and C'/D' RNPs function within the full-length sRNA molecule. In contrast to the eukaryotic snoRNP complex, where the core proteins are distributed asymmetrically on the box C/D and C'/D' motifs, all three archaeal core proteins bind both motifs symmetrically. This difference in core protein distribution is a result of altered RNA-binding capabilities of the archaeal and eukaryotic core protein homologs. Thus, evolution of the box C/D nucleotide modification complex has resulted in structurally distinct archaeal and eukaryotic RNP particles.

Fig. 1. Archaeal sR8 sRNP assembly requires a defined order of core protein addition and forms symmetric RNP complexes on the terminal box C/D core and internal C′/D′ motifs. (A) The folded structure of M.jannaschii sR8 sRNA with terminal box C/D core and internal C′/D′ motifs is based upon the consensus structure of the snoRNA box C/D motif implied from the crystal structure of the15.5 kDa binding site on the U4 snRNA (Vidovic et al., 2000; Watkins et al., 2000). Lower panel: sR8 sRNP complexes were assembled by incubating L7, Nop56/58 and fibrillarin core proteins, either individually or in different combinations, with 5′-radiolabeled sR8 sRNA. Assembled complexes were resolved on 4% native polyacrylamide gels and RNPs were visualized by autoradiography. Migration positions of the partially assembled (RNP I and RNP II) and complete (RNP III) RNP complexes are indicated. (B and C) The terminal box C/D core and internal C′/D′ RNA half-molecules are derived from the wild-type sR8 full-length sRNA. Lower panels: terminal box C/D core and internal C′/D′ RNPs were assembled by incubating 5′-radiolabeled RNA with the indicated sRNP core proteins. Migration positions of the partial (RNP I and RNP II) and fully assembled RNPs (RNP III) are indicated. The slower migrating complex in lane 2 of (C) is non-specific L7 binding to the RNA at elevated L7 concentrations.

Fig. 2. Nop56/58 and fibrillarin interact via protein–protein interactions and can bind the internal C′/D′ motif in the absence of core protein L7. (A) Nop56/58 and fibrillarin core proteins bind specifically to the C′/D′ motif in the absence of L7 core protein. Nop56/58 and fibrillarin were incubated with radiolabeled box C/D or C′/D′ RNA. Assembled RNP complexes were resolved on a native polyacrylamide gel and visualized by autoradiography. Competition experiments included a 1000-fold molar excess of non-radiolabeled box C/D or C′/D′ RNA. (B) Nop56/58 interacts with fibrillarin. sRNP core proteins were incubated in equimolar concentrations. Twenty percent of the protein mixture was removed as the applied sample (A) and the remainder applied to affinity resin. Bound proteins were eluted (E), resolved on 14% SDS–polyacrylamide gels and visualized by Coomassie Blue staining. Incubated protein combinations are indicated above each gel lane. (C and D) Pulldown analysis demonstrates the presence of L7 in the C′/D′ RNP complex. His-tagged fibrillarin, Nop56/58 and L7 were incubated with the C/D and C′/D′ RNAs in various combinations as indicated and fibrillarin was affinity-selected via the His tag. Co-isolated sRNP core proteins (upper panels of C and D) and RNAs (lower panels of C and D) were resolved on polyacrylamide gels and visualized by Coomassie Blue and EtBr staining, respectively.

Fig. 3. In vitro assembled sR8 sRNP guides site-specific 2′-O-methylation from both terminal box C/D and internal C′/D′ RNPs. Efficient methylation requires juxtapositioning of both RNP complexes in the full-length sRNA. (A) Schematic presentation of M.jannaschii sR8 base paired with D and D′ target RNAs. D-CH₃ and D′-CH₃ target RNAs possessing a previously 2′-O-methylated sugar at the designate nucleotide are illustrated above. (B) The assembled sR8 sRNP complex guides site-specific methylation of both the D and D′ target RNAs. Assembled sR8 sRNP was incubated at 70°C with the indicated target RNAs (in parentheses) and [³H]SAM. At various times, aliquots of the reaction were collected and TCA precipitated, and [³H]methyl incorporation was measured by scintillation counting. (C) Methylation of D and D′ RNA substrates demonstrates site-specific 2′-O-methylation at designate nucleotides of the target RNAs. Target RNAs in various combinations indicated above the gel were resolved by electrophoresis and 2′-O-methylated RNAs revealed by radiography. (D) Efficient methylation requires juxtapositioned RNPs. RNP complexes assembled upon the box C/D and C′/D′ halfmer RNAs were incubated with their respective target RNAs and assayed for methylation. sR8 (D) target RNA (solid square) is the control level of methylation for the box C/D core RNP when positioned in the full-length sR8 sRNP.

Fig. 4. Mutation of either the terminal box C/D core or internal C′/D′ motifs in full-length sR8 affects guided 2′-O-methylation of both RNP complexes. (A) Schematic presentation of sR8 sRNA with various mutations in the box C, D, C′ and D′ sequences. For simplicity, the complete terminal helix is not shown but it is present in all mutant sRNAs. (B) Box C/D or C′/D′ mutations affect methylation activity of both RNPs. Methylation efficiencies of wild-type sR8, halfmer RNAs and the full-length mutants are reported as total picomoles of methyl incorporation into the target RNAs in 1 h. Methylation of the control D-CH₃ or D′-CH₃ target is subtracted as background. Numbers in parentheses indicate the percentage of methylation with respect to activity of the respective RNPs in the full-length sR8 sRNA. (C) RNP complexes are assembled on sR8 box C/D and C′/D′ RNA mutants. sR8 RNAs containing the individual mutants illustrated in (A) were radiolabeled and incubated with the sRNP core proteins as indicated above each gel lane. Assembled complexes were resolved on native polyacrylamide gels and visualized by autoradiography.

Fig. 5. Core protein L7 exhibits cooperative binding to archaeal sR8 sRNA. (A) L7 protein binds the box C/D and C′/D′ motifs of sR8 sRNA. Increasing concentrations of L7 were incubated with radiolabeled sR8. Assembled RNP complexes were resolved on native polyacrylamide gels and visualized by autoradiography. The slowest migrating L7 RNP (lane 2) is observed only at excess L7 concentrations and represents non-specific L7 binding. (B) L7 association with the sR8 box C/D and C′/D′ motifs is cooperative. Increasing concentrations of L7 were incubated with radiolabeled sR8, assembled RNP complex blotted to nitrocellulose membranes and assembled RNP quantified using a PhosphorImager. The fraction of RNA bound in the RNP complex is plotted as a function of L7 concentration. L7 binding data are also presented as a Hill plot (inset). (C and D) L7 binds to both the terminal box C/D core and internal C′/D′ halfmer RNAs. Increasing concentrations of L7 were incubated with radiolabeled terminal box C/D (C) or internal C′/D′ halfmers (D). Assembled RNP complexes were resolved on native polyacrylamide gels and visualized by autoradiography.

Fig. 6. Archaeal L7 binds both the C/D and C′/D′ motifs, while eukaryotic 15.5 kDa protein binds only the terminal box C/D core motif. (A) Increasing amounts of M.jannaschii L7 protein were incubated with radiolabeled human U15 snoRNA. Assembled RNP complexes were resolved on a native polyacrylamide gel and visualized by autoradiography. (B) Archaeal sRNP core protein L7 exhibits cooperativity in binding to U15 snoRNA. Binding analyses were performed as detailed in Figure 5. (C) Eukaryotic snoRNP core protein 15.5 kDa binds only the U15 box C/D core motif. Increasing amounts of mouse 15.5 kDa protein were incubated with radiolabeled human U15 snoRNA. Assembled RNP complexes were resolved on a native polyacrylamide gel and visualized by radiography. (D) The fraction of U15 snoRNA bound in an RNP complex is plotted as a function of 15.5 kDa concentration. Increasing amounts of mouse 15.5 kDa core protein were incubated with radiolabeled full-length archaeal sR8 sRNA (E) and halfmer RNAs possessing either the terminal box C/D (F) or internal C′/D′ (G) motifs. Assembled RNP complexes were resolved on native polyacrylamide gels and visualized by radiography.

Fig. 7. Archaeal and eukaryotic box C/D RNP complexes exhibit symmetric versus asymmetric distribution of the RNP core proteins bound to sRNAs and snoRNAs, respectively. The illustrated structure of the eukaryotic snoRNP is based on the work of Cahill et al. (2002) and Szewczak et al. (2002).