Novel RNA-binding properties of Pop3p support a role for eukaryotic RNase P protein subunits in substrate recognition.

Ribonuclease P (RNase P) catalyzes the 5'-end maturation of transfer RNA molecules. Recent evidence suggests that the eukaryotic protein subunits may provide substrate-binding functions (True, H. L., and Celander, D. W. (1998) J. Biol. Chem. 273, 7193-7196). We now report that Pop3p, an essential protein subunit of the holoenzyme in Saccharomyces cerevisiae, displays novel RNA-binding properties. A recombinant form of Pop3p (H6Pop3p) displays a 3-fold greater affinity for binding pre-tRNA substrates relative to tRNA products. The recognition sequence for the H6Pop3p-substrate interaction in vitro was mapped to a 39-nucleotide long sequence that extends from position -21 to +18 surrounding the natural processing site in pre-tRNA substrates. H6Pop3p binds a variety of RNA molecules with high affinity (K(d) = 16-25 nm) and displays a preference for single-stranded RNAs. Removal or modification of basic C-terminal residues attenuates the RNA-binding properties displayed by the protein specifically for a pre-tRNA substrate. These studies support the model that eukaryotic RNase P proteins bind simultaneously to the RNA subunit and RNA substrate.

Probing RNA structures with hydroxyl radicals.

Fe(II)-EDTA can be used to conveniently generate hydroxyl radicals to promote cleavage of RNA at nucleotide resolution. Two procedures are described, involving the generation of free radicals from solvated molecular oxygen and from hydrogen peroxide added to the RNA solution. Unlike other footprinting reagents, hydroxyl radicals cleave the sugar-phosphate backbone at every residue and thus provide uniform cleavage in a given RNA secondary structure. Because some positions become protected by tertiary folding, this reagent is useful for monitoring the global folding of RNA at equilibrium.

Translationally repressive RNA structures monitored in vivo using temperate DNA bacteriophages.

The RNA challenge phage system enables genetic selection of proteins with RNA-binding activity in bacteria. These phages are modified versions of the temperate DNA bacteriophage P22 in which post-transcriptional regulatory events control the developmental fate of the phage. The system was originally developed to identify novel RNA ligands that display reduced affinity for the R17/MS2 coat protein, as well as to select for suppressor coat proteins that recognize mutant RNA ligands. During the course of evaluating whether the HIV-1 Rev protein could direct lysogen development for bacteriophage derivatives that encode Rev response element (RRE) RNA sequences, two examples of RRE RNA ligands that interfere with challenge phage development were identified. In the phage examples described, RRE RNA secondary structure prevents Ant protein biosynthesis and lytic development. Phage lysogen formation occurs efficiently in recipient cells, independent of the expression status of the Rev protein or trans-acting competitor RRE RNA ligands. These studies provide the first example whereby RNA challenge phages may be applied to study RNA folding events and RNA structural interactions in an in vivo context.

Protein components contribute to active site architecture for eukaryotic ribonuclease P.

In eukaryotes, ribonuclease P (RNase P) requires both RNA and protein components for catalytic activity. The eukaryotic RNase P RNA, unlike its bacterial counterparts, does not possess intrinsic catalytic activity in the absence of holoenzyme protein components. We have used a sensitive photoreactive cross-linking assay to explore the substrate-binding environment for different eukaryotic RNase P holoenzymes. Protein components from the Tetrahymena thermophila and human RNase P holoenzymes form specific products in photoreactions containing [4-thio]-uridine-labeled pre-tRNAGln. The HeLa RNase P RNA in neither the presence nor the absence of holoenzyme protein components formed cross-link products to the pre-tRNAGln probe. Parallel photo-cross-linking experiments with the Escherichia coli RNase P holoenzyme revealed that only the bacterial RNase P RNA forms specific substrate photoadducts. A protein-rich active site for the eukaryotic RNase P represents one unique feature that distinguishes holoenzyme organization between bacteria and eukaryotes.

Site-specific phosphorylation of the human immunodeficiency virus type-1 Rev protein accelerates formation of an efficient RNA-binding conformation.

Phosphorylation is important in the regulation of many cellular processes, yet the precise role of protein phosphorylation for many RNA-binding protein substrates remains obscure. In this report, we demonstrate that phosphorylation of a recombinant human immunodeficiency virus type-1 Rev protein promotes rapid formation of an efficient RNA-binding state. The apparent dissociation constant for ligand binding is enhanced 7-fold for the protein following phosphorylation; however, phosphate addition leads to a 1. 6-fold decrease in RNA ligand-protein complex stability. RNA ligand binding stimulates slow formation of an equally competent binding state for the unphosphorylated protein, indicating that the addition of phosphate or ligand binding promotes a similar conformational change in Rev. Phosphorylation directly alters the conformation of Rev, as revealed by modification experiments that monitor the solvent accessibility of cysteines in the protein. These biochemical properties are attributed to the addition of phosphate at one of two serine residues (Ser-54 or Ser-56) that lie within the multimerization domain adjacent to the RNA-binding helix. Glutaraldehyde-mediated cross-linking experiments revealed that phosphorylation of Rev does not affect Rev multimerization activity. The Rev protein from the less pathogenic HIV-2 isolate lacks this phosphorylation site in the amino acid sequence; thus, the described biochemical properties of the phosphorylated protein may contribute to Rev activity and possibly to HIV-1 virulence during natural infection.

Direct genetic selection of two classes of R17/MS2 coat proteins with altered capsid assembly properties and expanded RNA-binding activities.

RNA challenge phages are derivatives of bacteriophage P22 that enable direct genetic selection for a specific RNA-protein interaction. The bacteriophage P22 R17 encodes a wild-type R17 operator site and undergoes lysogenic development following infection of susceptible bacterial strains that express the R17/MS2 coat protein. A P22 R17 derivative with an OcRNA site (P22 R17 [A(-10)U]) develops lytically following infection of these strains. RNA challenge phages can be used to isolate second-site coat protein suppressors that recognize an OcRNA sequence by selecting for lysogens with a P22 R17 [Oc] phage derivative. The bacteriophage derivative P22 R17 [A(-10)U] was used in one such scheme to isolate two classes of genes that encode R17 coat proteins with altered capsid assembly properties and expanded RNA-binding characteristics. These mutations map outside the RNA-binding surface and include amino acid substitutions that interfere with interactions between coat protein dimers in the formation of the stable phage capsid. One class of mutants encodes substitutions at the highly conserved first and second positions of the mature coat protein. N-terminal sequence analysis of these mutants reveals that coat proteins with substitutions only at position 1 are defective in post-translational processing of the initiator methionine. All selected proteins possess expanded RNA-binding properties since they direct efficient lysogen formation for P22 R17 and P22 R17 [A(-10)U]; however, bacterial strains that express the protein mutants remain sensitive to lytic infection by other P22 R17 [Oc] bacteriophages. The described selection strategy provides a novel genetic approach to dissecting protein structure within RNA-binding proteins.

Functional recognition of fragmented operator sites by R17/MS2 coat protein, a translational repressor.

The R17/MS2 coat protein serves as a translational repressor of replicase by binding to a 19 nt RNA hairpin containing the Shine-Dalgarno sequence and the initiation codon of the replicase gene. We have explored the structural features of the RNA operator site that are necessary for efficient translational repression by the R17/MS2 coat protein in vivo . The R17/MS2 coat protein efficiently directs lysogen formation for P22 R17 , a bacteriophage P22 derivative that carries the R17/MS2 RNA operator site within the P22 phage ant mRNA. Phages were constructed that contain fragmented operator sites such that the Shine-Dalgarno sequence and the initiation codon of the affected gene are not located within the RNA hairpin. The wild-type coat protein directs efficient lysogen formation for P22 phages that carry several fragmented RNA operator sites, including one in which the Shine-Dalgarno sequence is positioned 4 nt outside the coat protein binding site. Neither the wild-type R17/MS2 coat protein nor super-repressor mutants induce lysogen formation for a P22 phage encoding an RNA hairpin at a distance of 9 nt from the Shine-Dalgarno sequence, implying that a discrete region of biological repression is defined by the coat protein-RNA hairpin interaction. The assembly of RNA species into capsid structures is not an efficient means whereby the coat protein achieves translational repression of target mRNA transcripts. The R17/MS2 coat protein exerts translational regulation that extends considerably beyond the natural biological RNA operator site structure; however, the coat protein still mediates repression in these constructs by preventing ribosome access to linear sequence determinants of the translational initiation region by the formation of a stable RNA secondary structure. An efficient translational regulatory mechanism in bacteria appears to reside in the ability of proteins to regulate RNA folding states for host cell and phage mRNAs.

Efficient modification of RNA by porphyrin cation photochemistry: monitoring the folding of coaxially stacked RNA helices in tRNA(Phe) and the human immunodeficiency virus type 1 rev response element RNA.

Coaxially stacked RNA helices are a determined of RNA tertiary structure, but their presence is rarely detected using conventional chemical modification methods. In this report we describe a porphyrin ion photoreaction that enables one to monitor RNA stacking interactions and the folding of coaxially stacked RNA helices. The porphyrin cations meso-tetrakis(4-N-methylpyridyl)porphine, meso-tetrakis-(para-N-trimethylanilinium)porphine, and meso-tetrakis(2-N-methylpyridyl)porphine were used to characterize tRNA(Phe) and the human immunodeficiency virus type-I Rev response element RNA. Nucleosides at the bases of contiguous RNA helices in each RNA are efficiently modified by the porphyrin cations following irradiation of porphyrin-RNA mixtures. These photomodifications are markedly reduced for RNA equilibrated in ionic buffers that lead to enhanced stabilization of coaxially stacked helices. The porphyrin cation photoreaction specifically modifies G18, G20, and G34 in the tRNA folding produced by Mg(II). These nucleobases are exposed to solvent in the native tRNA structure and thus available to stack with solvent-borne porphyrin molecules. The describe porphyrin cation photochemical method provides a novel approach to study the solvent accessibility of nucleobases in RNA structure and to monitor the folding of coaxially stacked helices in RNA.

Ribonuclease P of Tetrahymena thermophila.

Ribonuclease P (RNase P) is responsible for the generation of mature 5' termini of tRNA. The RNA component of this complex encodes the enzymatic activity in bacteria and is itself catalytically active under appropriate conditions in vitro. The role of the subunits in eucaryotes has not yet been established. We have partially purified RNase P activity from the ciliate protozoan Tetrahymena thermophila to learn more about the biochemical characteristics of RNase P from a lower eucaryote. The Tetrahymena RNase P displays a pH optimum and temperature optimum characteristic of RNase P enzymes isolated from other organisms. The Km of the T. thermophila enzyme for pre-tRNAGln is 1.6 x 10(-7)M, which is comparable to the values reported for other examples of RNase P. The Tetrahymena RNase P is a ribonucleoprotein complex, as supported by its sensitivity to micrococcal nuclease and proteinase K. The buoyant density of the enzyme in Cs2SO4 is 1.42 g/ml, which suggests that the RNA component of the Tetrahymena enzyme comprises a significantly greater percentage of the holoenzyme than that determined for RNase P of other Eucarya or Archaea. The holoenzyme has a requirement for divalent cations displaying characteristics that are unique for RNase P but closely resemble preferences reported for the Tetrahymena group I intron RNA. Puromycin inhibits pre-tRNA processing by the Tetrahymena complex, and implications of the similarities between recognition of tRNA by ribosomal components and RNase P are discussed.

Improved method for selecting RNA-binding activities in vivo.

RNA challenge phages are modified versions of bacteriophage P22 that allow one to select directly for a specific RNA-protein interaction in vivo. The original construction method for generating a bacteriophage that encodes a specific RNA target requires two homologous recombination reactions between plasmids and phages in bacteria. An improved method is described that enables one to readily construct RNA challenge phages through a single homologous recombination reaction in vivo. We have applied the new method to construct a derivative of P22R17, an RNA challenge phage that undergoes lysogenic development in bacterial cells that express the bacteriophage R17/MS2 coat protein.

Direct genetic selection for a specific RNA-protein interaction.

The decision between lytic and lysogenic development of temperate DNA bacteriophages is determined largely by transcriptional regulation through DNA-binding proteins. To determine whether a heterologous RNA-binding activity could control the developmental fate of a DNA bacteriophage, a derivative of P22 was constructed in which the chosen developmental pathway is regulated by an RNA-binding molecule interacting with its RNA target site located in a phage mRNA. In the example presented, lysogenic development of the phage relies upon R17 coat protein expression in the susceptible host cell and the availability of a suitable coat protein binding site encoded by the phage genome. Through the analysis of phage mutants that are able to grow lytically in susceptible cells that express the coat protein, additional insights were obtained regarding the specific interaction of the R17 coat protein with its RNA binding site. This study also suggests a novel and extremely sensitive strategy for selecting RNA-binding activities in vivo.

Mutations at the guanosine-binding site of the Tetrahymena ribozyme also affect site-specific hydrolysis.

Self-splicing group I introns use guanosine as a nucleophile to cleave the 5' splice site. The guanosine-binding site has been localized to the G264-C311 base pair of the Tetrahymena intron on the basis of analysis of mutations that change the specificity of the nucleophile from G (guanosine) to 2AP (2-aminopurine ribonucleoside) (F. Michel et al. (1989) Nature 342, 391-395). We studied the effect of these mutations (G-U, A-C and A-U replacing G264-C311) in the L-21 ScaI version of the Tetrahymena ribozyme. In this enzymatic system (kcat/Km)G monitors the cleavage step. This kinetic parameter decreased by at least 5 x 10(3) when the G264-C311 base pair was mutated to an A-U pair, while (kcat/Km)2AP increased at least 40-fold. This amounted to an overall switch in specificity of at least 2 x 10(5). The nucleophile specificity (G > 2AP for the G-C and G-U pairs, 2AP > G for the A-U and A-C pairs) was consistent with the proposed hydrogen bond between the nucleotide at position 264 and N1 of the nucleophile. Unexpectedly, the A-U and A-C mutants showed a decrease of an order of magnitude in the rate of ribozyme-catalyzed hydrolysis of RNA, in which H2O or OH- replaces G as the nucleophile, whereas the G-U mutant showed a decrease of only 2-fold. The low hydrolysis rates were not restored by raising the Mg2+ concentration or lowering the temperature. In addition, the mutant ribozymes exhibited a pattern of cleavage by Fe(II)-EDTA indistinguishable from that of the wild type, and the [Mg2+]1/2 for folding of the A-U mutant ribozyme was the same as that of the wild type. Therefore the guanosine-binding site mutations do not appear to have a major effect on RNA folding or stability. Because changing G264 affects the hydrolysis reaction without perturbing the global folding of the RNA, we conclude that the catalytic role of this conserved nucleotide is not limited to guanosine binding.

Folding of group I introns from bacteriophage T4 involves internalization of the catalytic core.

Fe(II)-EDTA, a solvent-based cleavage reagent that distinguishes between the inside and outside surfaces of a folded RNA molecule, has revealed some of the higher-order folding of the group IB intron from Tetrahymena thermophila pre-rRNA. This reagent has now been used to analyze the bacteriophage T4 sunY and td introns, both of which are members of the group IA subclass. Significant portions of the phylogenetically conserved secondary structure are protected from Fe(II)-EDTA cleavage. However, the P4 secondary structure element, which is substantially protected in the Tetrahymena intron, is available for cleavage in the two T4 introns. We conclude that a family of catalytic RNAs (ribozymes) that possess similar secondary structures and have similar activities fold into similar but nonidentical tertiary structures that nevertheless serve to internalize portions of the catalytic center. Furthermore, comparison of cleavage patterns of the sunY and td intron RNAs indicates that conserved nucleotides outside as well as within the catalytic core participate in the tertiary structure.

Cloning and expression of genes for the Oxytricha telomere-binding protein: specific subunit interactions in the telomeric complex.

Telomeres of Oxytricha nova macronuclear chromosomes consist of a repeated T4G4 sequence, single-stranded at the 3' terminus, bound by a heterodimeric protein. The cloning of genes for the two polypeptides and their separate expression in E. coli have enabled evaluation of their individual contributions to DNA binding. The 56 kd alpha subunit binds single-stranded DNA by itself, one polypeptide per T4G4 block; multiple subunits can coat a (T4G4)n multimer. The derived amino acid sequence of alpha does not reveal any known DNA-binding motif, so it appears to represent a novel type of DNA-binding protein. The previously cloned 41 kd beta subunit does not by itself protect DNA from methylation, but is required along with alpha to recreate the pattern of methylation protection indicative of telomeres in vivo. The unusual ability of the protein to engage in two different interactions with the same telomeric DNA sequence might provide the versatility necessary for diverse telomere functions.

Visualizing the higher order folding of a catalytic RNA molecule.

The higher order folding process of the catalytic RNA derived from the self-splicing intron of Tetrahymena thermophila was monitored with the use of Fe(II)-EDTA-induced free radical chemistry. The overall tertiary structure of the RNA molecule forms cooperatively with the uptake of at least three magnesium ions. Local folding transitions display different metal ion dependencies, suggesting that the RNA tertiary structure assembles through a specific folding intermediate before the catalytic core is formed. Enzymatic activity, assayed with an RNA substrate that is complementary to the catalytic RNA active site, coincides with the cooperative structural transition. The higher order RNA foldings produced by Mg(II), Ca(II), and Sr(II) are similar; however, only the Mg(II)-stabilized RNA is catalytically active. Thus, these results directly demonstrate that divalent metal ions participate in general folding of the ribozyme tertiary structure, and further indicate a more specific involvement of Mg(II) in catalysis.

Iron(II)-ethylenediaminetetraacetic acid catalyzed cleavage of RNA and DNA oligonucleotides: similar reactivity toward single- and double-stranded forms.

Fe(II)-EDTA catalyzes the cleavage of nucleic acids with little or no base-sequence specificity. We have now studied the preference of this reagent in catalyzing the cleavage of single- versus double-stranded nucleic acid structures. Three RNA and two DNA molecules, each expected to contain both single- and double-stranded regions, were synthesized and their structures characterized by enzymatic digestion using secondary structure specific nucleases. Fe(II)-EDTA catalyzed nearly uniform strand scission along the entire length of each molecule; no correlation with secondary structure was observed. The homopolymer sequence dA30:dT30, embedded in a mixed-sequence context to promote exact register of the homopolymer tract, was cleaved to an extent similar to that of flanking sequences. The reactions were relatively insensitive to K+, Na+, and Mg2+ in the range 10-100 mM and were quenched by Tris-HCl buffer. We conclude that the Fe(II)-EDTA-catalyzed strand scission reaction does not discriminate between typical single- and double-stranded regions, which simplifies the interpretation of experiments in which the reaction is used to probe the tertiary structure of RNA molecules [Latham, J. A., & Cech, T. R. (1989) Science 245, 276-282].

Two versions of the gene encoding the 41-kilodalton subunit of the telomere binding protein of Oxytricha nova.

Macronuclear chromosomes of the ciliated protozoan Oxytricha nova terminate with a single-stranded (T4G4)2 overhang. The (T4G4)2 telomeric overhang is tenaciously bound by a protein heterodimer. We have cloned and sequenced the gene encoding the 41-kDa subunit of this telomere binding protein. The predicted amino acid sequence comprises two distinct regions, a carboxyl-terminal two-thirds that is 23% lysine and bears similarity to histone H1 and an amino-terminal one-third containing a hydrophobic stretch of about 15 amino acids. Two macronuclear versions of the gene differ in nucleotide sequence at several positions, but the derived polypeptides differ only at a single position, Ser-110 or Ala-110. Both versions harbor a small intron. The existence of this intron demonstrates that, despite the elimination of 95% of the micronuclear genome from the developing macronucleus, at least some noncoding DNA is retained during macronuclear development of hypotrichous ciliates.

Regulatory elements within the murine leukemia virus enhancer regions mediate glucocorticoid responsiveness.

Enhancer elements within nonleukemogenic (Akv) and T-cell leukemogenic (SL3-3) murine leukemia viruses demonstrate strong cell type preference in transcriptional activity. These transcription elements are additionally regulated by the synthetic glucocorticoid dexamethasone, and this pattern of regulation varies according to cell type. The sequences required for dexamethasone regulation for both Akv and SL3-3 are shown to include a 17-nucleotide consensus sequence previously termed the glucocorticoid response element (GRE). Although the GREs are identical for both viral enhancers, the sequences surrounding these elements differ, as does the spatial arrangement of the GRE sequences with respect to one another. It is proposed that the spatial arrangement of the GREs, as well as their precise sequence context, determines the difference in the response to dexamethasone of the enhancers in different cell types.