We study the stochastic folding kinetics of RNA sequences into secondary structures with a new algorithm based on the formation, dissociation, and the shifting of individual base pairs. We discuss folding mechanisms and the correlation between the barrier structure of the conformational landscape and the folding kinetics for a number of examples based on artificial and natural sequences, including the influence of base modification in tRNAs.

Generally, mRNAs that prematurely terminate translation are abnormally low in abundance. In the case of mammalian cells, nonsense codons most often mediate a reduction in the abundance of newly synthesized, nucleus-associated mRNA by a mechanism that is not well understood. With the aim of defining cis-acting sequences that are important to the reduction process, the effects of particular β-globin gene rearrangements on the metabolism of β-globin mRNAs harboring one of a series of nonsense codons have been assessed. Results indicate that nonsense codons located 54 bp or more upstream of the 3′-most intron, intron 2, reduce the abundance of nucleus-associated mRNA to 10–15% of normal without altering the level of either of the two introns within pre-mRNA. The level of cytoplasmic mRNA is also reduced to 10–15% of normal, indicating that decay does not take place once the mRNA is released from an association with nuclei into the cytoplasm. A nonsense codon within exon 2 that does not reduce mRNA abundance can be converted to the type that does by (1) inserting a sufficiently large in-frame sequence immediately upstream of intron 2 or (2) deleting and reinserting intron 2 a sufficient distance downstream of its usual position. These findings indicate that only those nonsense codons located more than 54 bp upstream of the 3′-most intron reduce β-globin mRNA abundance, which is remarkably consistent with which nonsense codons within the triosephosphate isomerase (TPI) gene reduce TPI mRNA abundance. We propose that the 3′-most exon–exon junction of β-globin mRNA and, possibly, most mRNAs is marked by the removal of the 3′-most intron during pre-mRNA splicing and that the “mark” accompanies mRNA during transport to the cytoplasm. When cytoplasmic ribosomes terminate translation more than 54 nt upstream of the mark during or immediately after transport, the mRNA is subjected to nonsense-mediated decay. The finding that deletion of β-globin intron 2 does not appreciably alter the effect of any nonsense codon on β-globin mRNA abundance suggests that another cis-acting sequence functions in nonsense-mediated decay comparably to intron 2, at least in the absence of intron 2, possibly as a fail-safe mechanism. The analysis of deletions and insertions indicates that this sequence resides within the coding region and can be functionally substituted by intron 2.

A number of mitochondrial (mt) tRNAs have strong structural deviations from the classical tRNA cloverleaf secondary structure and from the conventional L-shaped tertiary structure. As a consequence, there is a general trend to consider all mitochondrial tRNAs as “bizarre” tRNAs. Here, a large sequence comparison of the 22 tRNA genes within 31 fully sequenced mammalian mt genomes has been performed to define the structural characteristics of this specific group of tRNAs. Vertical alignments define the degree of conservation/variability of primary sequences and secondary structures and search for potential tertiary interactions within each of the 22 families. Further horizontal alignments ascertain that, with the exception of serine-specific tRNAs, mammalian mt tRNAs do fold into cloverleaf structures with mostly classical features. However, deviations exist and concern large variations in size of the D- and T-loops. The predominant absence of the conserved nucleotides G18G19 and T54T55C56, respectively in these loops, suggests that classical tertiary interactions between both domains do not take place. Classification of the tRNA sequences according to their genomic origin (G-rich or G-poor DNA strand) highlight specific features such as richness/poorness in mismatches or G-T pairs in stems and extremely low G-content or C-content in the D- and T-loops. The resulting 22 “typical” mammalian mitochondrial sequences built up a phylogenetic basis for experimental structural and functional investigations. Moreover, they are expected to help in the evaluation of the possible impacts of those point mutations detected in human mitochondrial tRNA genes and correlated with pathologies.

Introns have typically been discovered in an ad hoc fashion: introns are found as a gene is characterized for other reasons. As complete eukaryotic genome sequences become available, better methods for predicting RNA processing signals in raw sequence will be necessary in order to discover genes and predict their expression. Here we present a catalog of 228 yeast introns, arrived at through a combination of bioinformatic and molecular analysis. Introns annotated in the Saccharomyces Genome Database (SGD) were evaluated, questionable introns were removed after failing a test for splicing in vivo, and known introns absent from the SGD annotation were added. A novel branchpoint sequence, AAUUAAC, was identified within an annotated intron that lacks a six-of-seven match to the highly conserved branchpoint consensus UACUAAC. Analysis of the database corroborates many conclusions about pre-mRNA substrate requirements for splicing derived from experimental studies, but indicates that splicing in yeast may not be as rigidly determined by splice-site conservation as had previously been thought. Using this database and a molecular technique that directly displays the lariat intron products of spliced transcripts (intron display), we suggest that the current set of 228 introns is still not complete, and that additional intron-containing genes remain to be discovered in yeast. The database can be accessed at http://www.cse.ucsc.edu/research/compbio/yeast_introns.html.

Catalytic RNAs are often regarded as molecular fossils from the RNA World, yet it is usually difficult to get more specific information about their evolution. Here we have investigated the coevolution of group II intron RNA structures with their intron-encoded reverse transcriptases (RTs). Unlike group I introns, there has been no obvious reshuffling between intron RNA structures and ORFs. Of the six classes of intron structures that encode ORFs, three are conventional forms of group II A1, B1, and B2 secondary structures, whereas the remaining classes are bacterial, are possibly associated with the most primitive ORFs, and have unusual features and hybrid features of group IIA and group IIB intron structures. Based on these data, we propose a new model for the evolution of group II introns, designated the retroelement ancestor hypothesis, which predicts that the major RNA structural forms of group II introns developed through coevolution with the intron-encoded protein rather than as independent catalytic RNAs, and that most ORF-less introns are derivatives of ORF-containing introns. The model is supported by the distribution of ORF-containing and ORF-less introns, and by numerous examples of ORF-less introns that contain ORF remnants.

The RNA-binding/dimerization domain of the NS1 protein of influenza A virus (73 amino acids in length) exhibits a novel dimeric six-helical fold. It is not known how this domain binds to its specific RNA targets, one of which is double-stranded RNA. To elucidate the mode of RNA binding, we introduced single alanine replacements into the NS1 RNA-binding domain at specific positions in the three-dimensional structure. Our results indicate that the dimer structure is essential for RNA binding, because any alanine replacement that causes disruption of the dimer also leads to the loss of RNA-binding activity. Surprisingly, the arginine side chain at position 38, which is in the second helix of each monomer, is the only amino-acid side chain that is absolutely required only for RNA binding and not for dimerization, indicating that this side chain probably interacts directly with the RNA target. This interaction is primarily electrostatic, because replacement of this arginine with lysine had no effect on RNA binding. A second basic amino acid, the lysine at position 41, which is also in helix 2, makes a strong contribution to the affinity of binding. We conclude that helix 2 and helix 2′, which are antiparallel and next to each other in the dimer conformation, constitute the interaction face between the NS1 RNA-binding domain and its RNA targets, and that the arginine side chain at position 38 and possibly the lysine side chain at position 41 in each of these antiparallel helices contact the phosphate backbone of the RNA target.

RNA interference (RNAi) is an evolutionarily conserved eukaryotic adaptive response that leads to the specific degradation of target mRNA species in response to cellular exposure to homologous double-stranded RNA molecules. Here, we have analyzed the subcellular location at which RNA degradation occurs in human cells exposed to double-stranded short interfering RNAs. To unequivocally determine whether a given mRNA is subject to degradation in the cytoplasm, the nucleus, or both, we have used the retroviral Rev/RRE system to control whether target mRNAs remain sequestered in the nucleus or are exported to the cytoplasm. In the absence of export, we found that the nuclear level of the RRE-containing target mRNA was not affected by activation of RNAi. In contrast, when nuclear export was induced by expression of Rev, cytoplasmic target mRNAs were effectively and specifically degraded by RNAi. Curiously, when the target mRNA molecule was undergoing active export from the nucleus, induction of RNAi also resulted in a reproducible approximately twofold drop in the level of target mRNA present in the nuclear RNA fraction. As this same mRNA was entirely resistant to RNAi when sequestered in the nucleus, this result suggests that RNAi is able to induce degradation of target mRNAs not only in the cytoplasm but also during the process of nuclear mRNA export. Truly nucleoplasmic mRNAs or pre-mRNAs are, in contrast, resistant to RNAi.

Sequence elements that can function as internal ribosome entry sites (IRES) have been identified in 5′ noncoding regions of certain uncapped viral and capped cellular mRNA molecules. However, it has remained largely unknown whether IRES elements are functional when located in their natural capped mRNAs. Therefore, the polysomal association and translation of several IRES-containing cellular mRNAs was tested under conditions that severely inhibited cap-dependent translation, that is, after infection with poliovirus. It was found that several known IRES-containing mRNAs, such as BiP and c-myc, were both associated with the translation apparatus and translated in infected cells when cap-dependent translation of most host-cell mRNAs was blocked, indicating that the IRES elements were functional in their natural mRNAs. Curiously, the mRNAs that encode eukaryotic initiation factor 4GI (eIF4GI) and 4GII (eIF4GII), two proteins with high identity and similar functions in the initiation of cap-dependent translation, were both associated with polysomes in infected cells. The 5′-end sequences of eIF4GI mRNA were isolated from a cDNA expression library and shown to function as an internal ribosome entry site when placed into a dicistronic mRNA. These findings suggest that eIF4G proteins can be synthesized at times when 5′ cap-dependent mRNA translation is blocked, supporting the notion that eIF4G proteins are needed in both 5′ cap-independent and 5′ cap-dependent translational initiation mechanisms.

Many viruses have overlapping genes and/or regions in which a nucleic acid signal is embedded in a coding sequence. To search for dual-use regions in the hepatitis C virus (HCV), we developed a facile computer-based sequence analysis method to map dual-use regions in coding sequences. Eight diverse full-length HCV RNA and polyprotein sequences were aligned and analyzed. A cluster of unusually conserved synonymous codons was found in the core-encoding region, indicating a potential overlapping open reading frame (ORF). Four peptides (A1, A2, A3, and A4) representing this alternate reading frame protein (ARFP), two others from the HCV core protein, and one from bovine serum albumin (BSA) were conjugated to BSA and used in western blots to test sera for specific antibodies from 100 chronic HCV patients, 44 healthy controls, and 60 patients with non-HCV liver disease. At a 1:20,000 dilution, specific IgGs to three of the four ARFP peptides were detected in chronic HCV sera. Reactivity to either the A1 or A3 peptides (both ARFP derived) was significantly associated with chronic HCV infection, when compared to non-HCV liver disease serum samples (10/100 versus 1/60; p < 0.025). Antibodies to A4 were not detected in any serum sample. Our western blot assays confirmed the presence of specific antibodies to a new HCV antigen encoded, at least in part, in an alternate reading frame (ARF) overlapping the core-encoding region. Because this novel HCV protein stimulates specific immune responses, it has potential value in diagnostic tests and as a component of vaccines. This protein is predicted to be highly basic and may play a role in HCV replication, pathogenesis, and carcinogenesis.

Myotonic dystrophy is caused by an expanded CTG repeat in the 3′ untranslated region of the DM protein kinase (DMPK) gene. The expanded repeat triggers the nuclear retention of mutant DMPK transcripts, but the resulting underexpression of DMPK probably does not fully account for the severe phenotype. One proposed disease mechanism is that nuclear accumulation of expanded CUG repeats may interfere with nuclear function. Here we show by thermal melting and nuclease digestion studies that CUG repeats form highly stable hairpins. Furthermore, CUG repeats bind to the dsRNA-binding domain of PKR, the dsRNA-activated protein kinase. The threshold for binding to PKR is ∼15 CUG repeats, and the affinity increases with longer repeat lengths. Finally, CUG repeats that are pathologically expanded can activate PKR in vitro. These results raise the possibility that the disease mechanism could be, in part, a gain of function by mutant DMPK transcripts that involves sequestration or activation of dsRNA binding proteins.

A computer program, OligoWalk, is reported that predicts the equilibrium affinity of complementary DNA or RNA oligonucleotides to an RNA target. This program considers the predicted stability of the oligonucleotide-target helix and the competition with predicted secondary structure of both the target and the oligonucleotide. Both unimolecular and bimolecular oligonucleotide self structure are considered with a user-defined concentration. The application of OligoWalk is illustrated with three comparisons to experimental results drawn from the literature.

The activity of the SR protein family of splicing factors in constitutive or alternative splicing requires direct interactions with the pre-mRNA substrate. Thus it is important to define the high affinity targets of the various SR species and to evaluate their ability to discriminate between defined RNA targets. We have analyzed the binding specificity of the 30-kDa SR protein 9G8, which contains a zinc knuckle in addition to the RNA binding domain (RBD). Using a SELEX approach, we demonstrate that 9G8 selects RNA sequences formed by GAC triplets, whereas a mutated zinc knuckle variant selects different RNA sequences, centered around a (A/U)C(A/U)(A/U)C motif, indicating that the zinc knuckle is involved in the RNA recognition specificity of 9G8. In contrast, SC35 selects sequences composed of pyrimidine or purine-rich motifs. Analyses of RNA–protein interactions with purified recombinant 30-kDa SR proteins or in nuclear extracts, by means of UV crosslinking and immunoprecipitation, demonstrate that 9G8, SC35, and ASF/SF2 recognize their specific RNA targets with high specificity. Interestingly, the RNA sequences selected by the mutated zinc knuckle 9G8 variant are efficiently recognized by SRp20, in agreement with the fact that the RBD of 9G8 and SRp20 are similar. Finally, we demonstrate the ability of 9G8 and of its zinc knuckle variant, or SRp20, to act as efficient splicing transactivators through their specific RNA targets. Our results provide the first evidence for cooperation between an RBD and a zinc knuckle in defining the specificity of an RNA binding domain.

The mRNA surveillance system is known to rapidly degrade aberrant mRNAs that contain premature termination codons in a process referred to as nonsense-mediated decay. A second class of aberrant mRNAs are those wherein the 3′ UTR is abnormally extended due to a mutation in the polyadenylation site. We provide several observations that these abnormally 3′-extended mRNAs are degraded by the same machinery that degrades mRNAs with premature nonsense codons. First, the decay of the 3′-extended mRNAs is dependent on the same decapping enzyme and 5′-to-3′ exonuclease. Second, the decay is also dependent on the proteins encoded by the UPF1, UPF2, and UPF3 genes, which are known to be specifically required for the rapid decay of mRNAs containing nonsense codons. Third, the ability of an extended 3′ UTR to trigger decay is prevented by stabilizing sequences within the PGK1 coding region that are known to protect mRNAs from the rapid decay induced by premature nonsense codons. These results indicate that the mRNA surveillance system plays a role in degrading abnormally extended 3′ UTRs. Based on these results, we propose a model in which the mRNA surveillance machinery degrades aberrant mRNAs due to the absence of the proper spatial arrangement of the translation-termination codon with respect to the 3′ UTR element as defined by the utilization of a polyadenylation site.

In this study we demonstrate that the HIV-1 leader RNA exists in two alternative conformations, a branched structure consisting of several well-known hairpin motifs and a more stable structure that is formed by extensive long-distance base pairing. The latter conformation was first identified as a compactly folded RNA that migrates unusually fast in nondenaturing gels. The minimally required domains for formation of this conformer were determined by mutational analysis. The poly(A) and DIS regions of the leader are the major determinants of this RNA conformation. Further biochemical characterization of this conformer revealed that both hairpins are disrupted to allow extensive long-distance base pairing. As the DIS hairpin is known to be instrumental for formation of the HIV-1 RNA dimer, the interplay between formation of the conformer and dimerization was addressed. Formation of the conformer and the RNA dimer are mutually exclusive. Consequently, the conformer must rearrange into a branched structure that exposes the dimer initiation signal (DIS) hairpin, thus triggering formation of the RNA dimer. This structural rearrangement is facilitated by the viral nucleocapsid protein NC. We propose that this structural polymorphism of the HIV-1 leader RNA acts as a molecular switch in the viral replication cycle.

In eukaryotic cells, efficient translation of most cellular mRNAs requires the synergistic interplay between the m7GpppN cap structure and the poly(A) tail during initiation. We have developed and characterized a cell-free system from human HeLa cells that recapitulates this important feature, displaying more than one order of magnitude of translational synergism between the cap structure and the poly(A) tail. The stimulation of cap-dependent translation by the poly(A) tail is length-dependent, but not mediated by changes in mRNA stability. Using this system, we investigated the effect of the poly(A) tail on the translation of picornaviral RNAs, which are naturally polyadenylated but initiate translation via internal ribosome entry sites (IRESs). We show that translation driven by the IRESs of poliovirus (PV), encephalomyocarditis virus (EMCV), and hepatitis A virus is also significantly augmented by a poly(A) tail, ranging from an approximately 3-fold stimulation for the EMCV-IRES to a more than 10-fold effect for the PV IRES. These results raise interesting questions concerning the underlying molecular mechanism(s). The cell-free system described here should prove useful in studying these questions as well as providing a general biochemical tool to examine the translation initiation pathway in a more physiological setting.

RNA affinity tags would be very useful for the study of RNAs and ribonucleoproteins (RNPs) as a means for rapid detection, immobilization, and purification. To develop a new affinity tag, streptavidin-binding RNA ligands, termed “aptamers,” were identified from a random RNA library using in vitro selection. Individual aptamers were classified into two groups based on common sequences, and representative members of the groups had sufficiently low dissociation constants to suggest they would be useful affinity tools. Binding of the aptamers to streptavidin was blocked by presaturation of the streptavidin with biotin, and biotin could be used to dissociate RNA/streptavidin complexes. To investigate the practicality of using the aptamer as an affinity tag, one of the higher affinity aptamers was inserted into RPR1 RNA, the large RNA subunit of RNase P. The aptamer-tagged RNase P could be specifically isolated using commercially available streptavidin-agarose and recovered in a catalytically active form when biotin was used as an eluting agent under mild conditions. The aptamer tag was also used to demonstrate that RNase P exists in a monomeric form, and is not tightly associated with RNase MRP, a closely related ribonucleoprotein enzyme. These results show that the streptavidin aptamers are potentially powerful tools for the study of RNAs or RNPs.

The eukaryotic nucleolus contains a large number of small nucleolar RNAs (snoRNAs) that are involved in pre-ribosomal RNA (pre-rRNA) processing. The H box/ACA-motif (H/ACA) class of snoRNAs has recently been demonstrated to function as guide RNAs targeting specific uridines in the pre-rRNA for pseudouridine (Ψ) synthesis. To characterize the protein components of this class of snoRNPs, we have purified the snR42 and snR30 snoRNP complexes by anti-m3G-immunoaffinity and Mono-Q chromatography of Saccharomyces cerevisiae extracts. Sequence analysis of the individual polypeptides demonstrated that the three proteins Gar1p, Nhp2p, and Cbf5p are common to both the snR30 and snR42 complexes. Nhp2p is a highly basic protein that belongs to a family of putative RNA-binding proteins. Cbf5p has recently been demonstrated to be involved in ribosome biogenesis and also shows striking homology with known prokaryotic Ψ synthases. The presence of Cbf5p, a putative Ψ synthase in each H/ACA snoRNP suggests that this class of RNPs functions as individual modification enzymes. Immunoprecipitation studies using either anti-Cbf5p antibodies or a hemagglutinin-tagged Nhp2p demonstrated that both proteins are associated with all H/ACA-motif snoRNPs. In vivo depletion of Nhp2p results in a reduction in the steady-state levels of all H/ACA snoRNAs. Electron microscopy of purified snR42 and snR30 particles revealed that these two snoRNPs possess a similar bipartite structure that we propose to be a major structural determining principle for all H/ACA snoRNPs.

Conversion of tRNA precursors to their mature forms requires the action of both endo- and exoribonucleases. Although studies over many years identified the endoribonuclease, RNase P, and several exoribonucleases as the enzymes responsible for generating the mature 5′ and 3′ termini, respectively, of Escherichia coli tRNAs, relatively little is known about how tRNAs are separated from long multimeric or multifunction transcripts, or from long leader and trailer sequences. To examine this question, the tRNA products that accumulate in mutant strains devoid of multiple exoribonucleases plus one or several endoribonucleases were analyzed by northern analysis. We find that the multifunction tyrT transcript, which contains two tRNA1Tyr sequences separated by a 209-nt spacer region plus a downstream mRNA, is cleaved at three sites in the spacer region by the endoribonuclease, RNase E. When both RNase E and RNase P are absent, a product containing both tRNAs accumulates. Two multimeric tRNA transcripts, those for tRNA Arg-His-Leu-Pro and tRNA Gly-Cys-Leu also require RNase E for maturation. For the former transcript, products with long 3′ extensions on tRNAArg, tRNAHis, and tRNAPro, as well as the primary transcript, accumulate in the absence of RNase E. For the latter transcript, RNase E cleaves downstream of each tRNA. Little processing of either multimeric transcript occurs in the absence of both RNase E and RNase P. These data indicate that RNase E is a major contributor to the initial processing of E. coli tRNA transcripts, providing substrates for final maturation by RNase P and the 3′ exoribonucleases. Based on this new information, a detailed model for tRNA maturation is proposed.

Several genes encoding putative RNA:5-methylcytidine-transferases (m5C-transferases) from different organisms, including yeast, have been identified by sequence homology with the recently identified 16S rRNA:m5C967-methyltransferase (gene SUN) from Escherichia coli. One of the yeast ORFs (YBL024w) was amplified by PCR, inserted in the expression vector pET28b, and the corresponding protein was hyperexpressed in E. coli BL21(DE3). The resulting N-terminally His6-tagged recombinant Ybl024p was purified to apparent homogeneity by one-step affinity chromatography on Ni2+-NTA-agarose column. The activity and substrate specificity of the purified Ybl024p were tested in vitro using T7 transcripts of different yeast tRNAs as substrates and S-adenosyl-L-methionine as a donor of the methyl groups. The results indicate that yeast ORF YBL024w encodes S-adenosyl-L-methionine-dependent tRNA: m5C-methyltransferase that is capable of methylating cytosine to m5C at several positions in different yeast tRNAs and pre-tRNAs containing intron. Modification of tRNA occurs at all four positions (34, 40, 48, and 49) at which m5C has been found in yeast tRNAs sequenced so far. Disruption of the ORF YBL024w leads to the complete absence of m5C in total yeast tRNA. Moreover no tRNA:m5C-methyltransferase activity towards all potential m5C methylation sites was detected in the extract of the disrupted yeast strain. These results demonstrate that the protein product of a single gene is responsible for complete m5C methylation of yeast tRNA. Because this newly characterized multisite-specific modification enzyme Ybl024p is the fourth tRNA-specific methyltransferase identified in yeast, we suggest designating it as TRM4, the gene corresponding to ORF YBL024w.

We have identified a homolog of the ADAR (adenosine deaminases that act on RNA) class of RNA editases from Drosophila, dADAR. The dADAR locus has been localized to the 2B6–7 region of the X chromosome and the complete genomic sequence organization is reported here. dADAR is most homologous to the mammalian RNA editing enzyme ADAR2, the enzyme that specifically edits the Q/R site in the pre-mRNA encoding the glutamate receptor subunit GluR-B. Partially purified dADAR expressed in Pichia pastoris has robust nonspecific A-to-I deaminase activity on synthetic dsRNA substrates. Transcripts of the dADAR locus originate from two regulated promoters. In addition, alternative splicing generates at least four major dADAR isoforms that differ at their amino-termini as well as altering the spacing between their dsRNA binding motifs. dADAR is expressed in the developing nervous system, making it a candidate for the editase that acts on para voltage-gated Na+ channel transcripts in the central nervous system. Surprisingly, dADAR itself undergoes developmentally regulated RNA editing that changes a conserved residue in the catalytic domain. Taken together, these findings show that both transcription and processing of dADAR transcripts are under strict developmental control and suggest that the process of RNA editing in Drosophila is dynamically regulated.