Higher-order RNA Structures Research Articles

Mutational and Structural Analysis of Stem-loop IIIc of the Hepatitis C Virus and GB Virus B Internal Ribosome Entry Sites

Translation of the open reading frames (ORF) of the hepatitis C virus (HCV) and closely related GB virus B (GBV-B) genomes is driven by internal ribosome entry site (IRES) elements located within the 5' non-translated RNA. The functioning of these IRES elements is highly dependent on primary and higher order RNA structures. We present here the solution structures of a common, critical domain within each of these IRESs, stem-loop IIIc. These ten-nucleotide hairpins have nearly identical sequences and similar overall tertiary folds. The final refined structure of each shows a stem with three G:C base-pairs and a novel tetraloop fold. Although the bases are buckled, the first and fourth nucleotides of both tetraloops form a Watson-Crick type base-pair, while the apical nucleotides are located in the major groove where they adopt C(2)-endo sugar puckering with B-form geometry. No hydrogen bonding interactions were observed involving the two apical residues of the tetraloop. Stability of the loops appears to be derived primarily from the stacking of bases, and the hydrogen bonding between the fourth and seventh residues. Mutational analysis shows that the primary sequence of stem-loop IIIc is important for IRES function and that the stem and first and fourth nucleotides of the tetraloop contribute to the efficiency of internal ribosome entry. Base-pair formation between these two positions is essential. In contrast, the apical loop nucleotides differ between HCV and GBV-B, and substitutions in this region of the hairpin are tolerated without major loss of function.

The stem loop II within the 5' nontranslated region of clinical coxsackievirus B3 genomes determines cardiovirulence phenotype in a murine model.

Coxsackievirus B3 (CVB3) is a principal viral cause of acute myocarditis in humans and has been implicated in the pathogenesis of dilated cardiomyopathy. It has been demonstrated elsewhere that, for 2 wild-type CVB3 strains, the primary molecular determinant of cardiovirulence phenotype localizes to the 5' nontranslated region (5' NTR). Here in this study, through construction of CVB3 chimeras, the predicted stem loop (SL) II within the 5' NTR has been identified as a primary viral determinant of cardiovirulence. Replication assays in cardiac-derived murine fibroblasts revealed that only cardiovirulent strains grow robustly at 37 degrees C, whereas all virus strains replicate efficiently in HeLa cells. Computational analyses of RNA secondary structure suggest that the predicted SLII of the noncardiovirulent isolate differs significantly from that of myocarditic strains. These results indicate that, for clinical CVB3 isolates, the major determinant of cardiovirulence localizes to the predicted SLII. The identity of this higher-order RNA structure may be essential for productive infection of cardiac cells.

Biological significance of a human enterovirus B-specific RNA element in the 3' nontranslated region.

The secondary structures predicted for the enteroviral 3' nontranslated region (3'NTR) all seem to indicate a conformation consisting of two (X and Y) hairpin structures. The higher-order RNA structure of the 3'NTR appears to exist as an intramolecular kissing interaction between the loops of these two hairpin structures. The enterovirus B-like subgroup possesses an additional stem-loop structure, domain Z, which is not present in the poliovirus-like enteroviruses. It has been suggested that the Z domain originated from a burst of short sequence repetitions (E. V. Pilipenko, S. V. Maslova, A. N. Sinyakov, and V. I. Agol, Nucleic Acids Res. 20:1739-1745, 1992). However, no functional features have yet been ascribed to this enterovirus B-like-specific RNA element in the 3'NTR. In this study, we tested the functional characteristics and biological significance of domain Z. A mutant of the cardiovirulent coxsackievirus group B3 strain Nancy which completely lacked the Z domain and which therefore acquired enterovirus C-like secondary structures exhibited a wild-type growth phenotype, as determined by single-cycle growth analysis with BGM cells. This result proves that the Z domain is virtually dispensable for viral growth in tissue cultures. Partial distortion of the Z domain structure resulted in a disabled virus with reduced growth kinetics, probably due to alternative conformations of the overall structure of the domain. Infection of mice showed that the recombinant coxsackievirus group B3 mutant which completely lacked the Z domain was less virulent. Pancreatic tissues from mice infected with wild-type virus and recombinant virus were equally affected. However, the heart tissue from mice infected with the recombinant virus showed only slight signs of myocarditis. These results suggest that the enterovirus B-like-specific Z domain plays a role in coxsackievirus-induced pathogenesis.

A conserved RNA structure element involved in the regulation of bacterial riboflavin synthesis genes

Large-scale sequencing of bacterial genomes has opened a new era in computational genomics. Gene complements are successfully analysed on the protein level. However, it has been noted that regulatory sites are less conserved than genes 1 Overbeek R. et al. The use of gene clusters to infer functional coupling. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 2896-2901 Crossref PubMed Scopus (979) Google Scholar , although it is possible to use genomic comparisons in order to predict gene regulation at the level of DNA ( 2 Mironov A.A. et al. Computer analysis of transcription regulatory patterns in completely sequenced bacterial genomes. Nucleic Acids Res. 1999; 27: 2981-2989 Crossref PubMed Scopus (120) Google Scholar ) and RNA ( 3 Vitreschak A. et al. Computer analysis of regulatory patterns in complete bacterial genomes. Translation initiation of the ribosomal protein operons. Biophysics. 1999; (in press) Google Scholar , 4 Huynen M.A. Bork P. Measuring genome evolution. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 5849-5856 Crossref PubMed Scopus (356) Google Scholar ). Indeed, comparative analysis is one of the standard methods to predict the secondary structure of RNA. It has been used to determine the spatial structure of stable RNAs ( 5 Winker S. et al. Structure detection through automated covariance search. Comput. Appl. Biosci. 1990; 6: 365-371 PubMed Google Scholar , 6 Gutell R. et al. Identifying constraints on the higher-order structure of RNA: continued development and application of comparative sequence analysis methods. Nucleic Acids Res. 1992; 20: 5785-5795 Crossref PubMed Scopus (228) Google Scholar , 7 Eddy S.R. Durbin R. RNA sequence analysis using covariance models. Nucleic Acids Res. 1994; 22: 2079-2088 Crossref PubMed Scopus (572) Google Scholar ), and to analyse regulatory RNAs, mostly in viral genomes 8 Le S-Y. Zuker M. Common structures of the 5′ non-coding RNA in enteroviruses and rhinoviruses. Thermodynamical stability and statistical significance. J. Mol. Biol. 1990; 216: 729-741 Crossref PubMed Scopus (83) Google Scholar , 9 Le S.Y. et al. Unusual folding regions and ribosome landing pad within hepatitis C virus and pestivirus RNAs. Gene. 1995; 154: 137-143 Crossref PubMed Scopus (43) Google Scholar . In such studies, either the common fold is selected among many predicted suboptimal folds for a set of RNAs ( 8 Le S-Y. Zuker M. Common structures of the 5′ non-coding RNA in enteroviruses and rhinoviruses. Thermodynamical stability and statistical significance. J. Mol. Biol. 1990; 216: 729-741 Crossref PubMed Scopus (83) Google Scholar ), or analysis of complementary substitutions in aligned sequences is used to construct a single conserved structure 7 Eddy S.R. Durbin R. RNA sequence analysis using covariance models. Nucleic Acids Res. 1994; 22: 2079-2088 Crossref PubMed Scopus (572) Google Scholar , 9 Le S.Y. et al. Unusual folding regions and ribosome landing pad within hepatitis C virus and pestivirus RNAs. Gene. 1995; 154: 137-143 Crossref PubMed Scopus (43) Google Scholar . Non-viral regulatory RNA elements, such as iron-responsive elements 10 Dandekar T. et al. Systematic genomic screening and analysis of mRNA in untranslated regions and mRNA precursors: combining experimental and computational approaches. Bioinformatics. 1999; 14: 271-278 Crossref Scopus (25) Google Scholar , and sites that regulate the initiation of translation in operons of ribosomal proteins 3 Vitreschak A. et al. Computer analysis of regulatory patterns in complete bacterial genomes. Translation initiation of the ribosomal protein operons. Biophysics. 1999; (in press) Google Scholar , 4 Huynen M.A. Bork P. Measuring genome evolution. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 5849-5856 Crossref PubMed Scopus (356) Google Scholar , 11 Keener J. Nomura M. Neidhardt F.C. E. coli and Salmonella. ASM Press, 1996: 1417-1431 Google Scholar often involve conserved structural elements and conserved nucleotides.

Binding of the influenza virus NS1 protein to model genome RNAs.

We have previously shown that the influenza virus NS1 protein exhibits stable binding to a model mini vRNA which is stronger than its binding to dsRNA. In this study, we confirmed that the binding depended on a higher-order structure of the model RNA, probably the panhandle structure formed by pairing between the 5'- and 3'-terminal common sequences. The formation of the structure was enhanced by the NS1 protein itself. We propose that an A bulge in a stretch of double helix results in a binding site with greater affinity for the NS1 protein.

Position dependence of functional hairpins important for human immunodeficiency virus type 1 RNA encapsidation in vivo.

At least two hairpins in the 5' untranslated leader region, stem-loops 1 and 3 (SL1 and SL3), contribute to human immunodeficiency virus type 1 RNA encapsidation in vivo. We used a competitive assay, which measures the relative encapsidation efficiency of mutant viral RNA in the presence of competing wild-type RNA, to compare the contributions of SL1, SL3, and two adjacent secondary structures, SL2 and SL4, to encapsidation. SL2 is not required for RNA encapsidation, while SL1, SL3, and SL4 all contribute approximately equally to encapsidation. To determine whether these hairpins function in a position-dependent manner, we interchanged the positions of two of these stem-loop structures. This resulted in substantial diminution of encapsidation, indicating that the secondary structures that comprise E, the encapsidation signal, function only in their correct contexts. Mutation of nucleotides flanking SL1 and SL3 had little effect on encapsidation. We also showed that SL1, while present on both genomic and subgenomic viral RNAs, nonetheless contributes to selective encapsidation of genomic RNA. Taken together, these data are consistent with the formation of a higher-order RNA structure, partially composed of SL1, SL3, and SL4, that functions to effect concurrent encapsidation of full-length RNA and exclusion of subgenomic RNA. Finally, it has been reported that E is required for efficient translation of Gag mRNA in vivo. However, we have found that a variety of mutants, including a mutant lacking the entire region encompassing SL1, SL2, and SL3, still produce RNAs that are efficiently translated. These data indicate that E is unlikely to contribute to efficient Gag mRNA translation in vivo.

Kissing of the two predominant hairpin loops in the coxsackie B virus 3' untranslated region is the essential structural feature of the origin of replication required for negative-strand RNA synthesis.

Higher-order RNA structures in the 3' untranslated region (3'UTR) of enteroviruses are thought to play a pivotal role in viral negative-strand RNA synthesis. The structure of the 3'UTR was predicted by thermodynamic calculations using the STAR (structural analysis of RNA) computer program and experimentally verified using chemical and enzymatic probing of in vitro-synthesized RNA. A possible pseudoknot interaction between the 3D polymerase coding sequence and domain Y and a "kissing" interaction between domains X and Y was further studied by mutational analysis, using an infectious coxsackie B3 virus cDNA clone (domain designation as proposed by E. V. Pilipenko, S. V. Maslova, A. N. Sinyakov, and V.I. Agol (Nucleic Acids Res. 20:1739-1745, 1992). The higher-order RNA structure of the 3'UTR appeared to be maintained by an intramolecular kissing interaction between the loops of the two predominant hairpin structures (X and Y) within the 3'UTR. Disturbing this interaction had no effect on viral translation and processing of the polyprotein but exerted a primary effect on viral replication, as was demonstrated in a subgenomic coxsackie B3 viral replicon, in which the capsid P1 region was replaced by the luciferase gene. Mutational analysis did not support the existence of the pseudoknot interaction between hairpin loop Y and the 3D polymerase coding sequence. Based on these experiments, we constructed a three-dimensional model of the 3'UTR of coxsackie B virus that shows the kissing interaction as the essential structural feature of the origin of replication required for its functional competence.

Probing the higher order structure of RNA with peroxonitrous acid

Potassium peroxinitrite (ONOOK) and [Fe(EDTA)] 2− were used to analyze the influence of chemically entirely different hydroxyl radical sources on tRNA cleavage profiles. [Fe(EDTA)] 2− gives rise to hydroxyl radicals via a Fenton-like reaction during the oxidation of chelated Fe 2+, while ONOOK generates hydroxyl radicals via its conjugate acid (ONOOK) when adding a stable alkaline solution of ONOOK in samples buffered at neutral pH. [Fe(EDTA)] 2− is known to induce oxidative strand scission at sugar moieties thought to be solvent accessible, while those residues located in the ‘inside’ of structured RNAs are protected. Although ONOOH is neutral and significantly smaller than the metal complex, both reagents generate the same protection pattern on tRNAs, suggesting that access of the commonly formed hydroxyl radical, rather than access of its source, is the determining factor when probing the higher order structure of RNA. Strong difference in reactivity is only seen at the modified 2-thiouridine S34 of tRNA Lys3 which shows hyperreactivity towards ONOOK treatment. This particular reaction may require interaction between the peroxonitrite anion and the thiocarbonyl group of the base, since hyperreactivity is not observed when probing the dethiolated tRNA Lys3.

The cardiovirulent phenotype of coxsackievirus B3 is determined at a single site in the genomic 5' nontranslated region.

We report the construction of chimeric coxsackievirus B3 (CVB3) strains in which sequences of an infectious cDNA copy of a noncardiovirulent CVB3 genome were replaced by the homologous sequences from a cardiovirulent CVB3 genome to identify which of 10 predicted genetic sites determine cardiovirulence. Cardiovirulent phenotype expression was consistently linked to nucleotide 234 (U in cardiovirulent CVB3 and C in avirulent CVB3) in the 5' nontranslated region. Reconstructions of the parental noncardiovirulent CVB3 genome from chimeras restored the noncardiovirulent phenotype when tested in mice. Inoculation of severe combined immunodeficient (scid) mice with the noncardiovirulent CVB3 strain resulted in massive cardiomyocyte necrosis in all animals. Sequence analysis of viral genomes isolated from twelve scid mouse hearts showed that only nucleotide position 234 was different (a C-->U transition) from that in the input parental noncardiovirulent CVB3 genome. Higher-order RNA structures predicted by two different algorithms did not demonstrate an obvious local effect caused by the C-->U change at nucleotide 234. Initial studies of parental and chimeric CVB3 replication in primary cultures of fetal murine heart fibroblasts and in adult murine cardiac myocytes demonstrated that viral RNA transcriptional efficiency is approximately 10-fold lower for noncardiovirulent CVB3 than for cardiovirulent CVB3. CVB3 did not shut off protein synthesis in murine cardiac fibroblasts, nor were levels of viral protein synthesis significantly different as a function of viral phenotype. Taken together, these data support a significant role for determination of the CVB3 cardiovirulence phenotype by nucleotide 234 in the 5' nontranslated region, possibly via a transcriptional mechanism.

A guanosine quadruplex and two stable hairpins flank a major cleavage site in insulin-like growth factor II mRNA.

Insulin-like growth factor II (IGF-II) mRNAs are cleaved by an endonucleolytic event in a conserved part of their 3' untranslated region that is predicted to exhibit a complex higher-order RNA structure. In the present study, we have examined the putative secondary structures of in vitro transcripts from the conserved part of human and rat mRNAs by enzymatic and chemical probing. The results show that the cleavage site is situated between two highly structured domains. The upstream domain consists of two large hairpins, whereas the downstream domain is guanosine-rich. The guanosine-rich domain adopts a compact unimolecular conformation in Na+ or K+ but not in Li+, and it completely arrests reverse transcription in K+ but only partially in Na+, indicating the presence of an intramolecular guanosine quadruplex. The flanking higher-order structures may ensure that the cleavage site is not sequestered in stable RNA structures, thus allowing interactions with RNA or proteins at posttranscriptional stages of IGF-II expression.

An independently folding domain of RNA tertiary structure within the Tetrahymena ribozyme.

The Tetrahymena thermophila pre-rRNA contains a 413-nucleotide self-splicing group I intron. This intron has been converted into a sequence-specific endonuclease or ribozyme. A 160-nucleotide portion of the ribozyme consisting of both highly conserved sequence elements (P4 and P6) and nonconserved peripheral extensions (P5abc and P6ab) was synthesized as a separate molecule. Solvent-based Fe(II)-EDTA, a probe that monitors higher-order RNA structure, revealed a protection pattern that was a large subset of that observed in the whole ribozyme. Data from dimethyl sulfate modification and partial digestion with nucleases were also consistent with maintenance of the proper secondary and tertiary structure in the shortened RNA molecule. Thus, this 160-nucleotide molecule (P4-P6 RNA) is an independently folding domain of RNA tertiary structure. A series of mutations and deletions were made within the P4-P6 domain to further dissect its tertiary structure. Fe(II)-EDTA and dimethyl sulfate analysis of these mutants revealed that the domain consists of two substructures, a localized subdomain involving the characteristic adenosine-rich bulge in P5a, and a subdomain-stabilized structure involving long-range interactions. Therefore, like some proteins, the intron RNA is modular, containing a separable domain and subdomain of tertiary structure.

Simian immunodeficiency virus RNA is efficiently encapsidated by human immunodeficiency virus type 1 particles.

Packaging of retroviral RNA is attained through the specific recognition of a cis-acting encapsidation site (located near the 5' end of the viral RNA) by components of the Gag precursor protein. Human immunodeficiency virus type 1 (HIV-1) and simian immunodeficiency virus (SIV) are two lentiviruses that lack apparent sequence similarity in their putative encapsidation regions. We used SIV vectors to determine whether HIV-1 particles can recognize the SIV encapsidation site and functionally propagate SIV nucleic acid. SIV nucleic acid was replicated by HIV-1 proteins. Thus, efficient lentivirus pseudotyping can take place at the RNA level. Direct examination of the RNA contents of virus particles indicated that encapsidation of this heterologous RNA is efficient. Characterization of deletion mutants in the untranslated leader region of SIV RNA indicates that only a very short region at the 5' end of the SIV RNA is needed for packaging. Comparison of this region with the corresponding region of HIV-1 reveals that both are marked by secondary structures that are likely to be similar. Thus, it is likely that a similar higher-order RNA structure is required for encapsidation.

A native RNA secondary structure controls alternative splice-site selection and generates two human growth hormone isoforms.

Consensus sequences at the splice donor, splice acceptor, and lariat branch point regions are necessary but insufficient determinants of splice-site selection in nuclear precursor mRNAs. Sequences outside of these regions can have a significant effect on the utilization of splice sites. Although the mode of action of such sequences is undefined in most cases, higher order RNA structures have been suggested as a potential contributor to splice-site selection. During a detailed analysis of the splicing patterns of the human growth hormone transcript, we located 2 bases in the vicinity of the exon 3 major splice-acceptor site (B) which facilitate the utilization of a competing downstream acceptor (B'). The effects of a series of site-specific mutations on the splicing pattern demonstrate that these 2 bases function by stabilizing a specific stem-loop structure in the native transcript. This defined secondary structure selectively encompasses the upstream B splice-acceptor site together with its lariat branch point region. Increasing the predicted stability of this stem by point mutations results in a corresponding shift in splicing towards the alternative B' splice-acceptor site. These results indicate that a specific secondary structure within the native human growth hormone transcript controls the relative utilization of two competing splice-acceptor sites with the consequent generation of two functionally distinct hormone isoforms.

An RNA pseudoknot and an optimal heptameric shift site are required for highly efficient ribosomal frameshifting on a retroviral messenger RNA.

Synthesis of the pol gene products of most retroviruses requires ribosomes to shift frame once or twice in the -1 direction while translating gag-pol mRNA. The viral signals for frameshifting include a heptanucleotide sequence on which the shift occurs and higher-order RNA structure just downstream of the shift site. We have made site-directed mutations in two stems (S1 and S2) of a putative RNA pseudoknot that begins 7 nucleotides 3' of the previously identified shift site (A AAA AAC) in the gag-pro region of mouse mammary tumor virus (MMTV) RNA. The mutants confirm the predicted structure, show that loss of either S1 or S2 impairs frameshifting, and exclude alternative RNA structures as significant for frameshifting. The importance of the MMTV pseudoknot has been further demonstrated by showing that shift sites from two other retroviruses function more efficiently in the position of the MMTV site than in their native contexts. However, the MMTV pseudoknot cannot promote detectable frameshifting in the absence of a recognizable upstream shift site. In addition, the species of tRNA that reads the second codon in the shift site appears to be a critical determinant, since changing the 7th nucleotide in the MMTV gag-pro shift site from C to A, U, or G severely impairs frameshifting.