ISFOLD: Structure Prediction of Base Pairs in Non-Helical RNA Motifs from Isostericity Signatures in Their Sequence Alignments

Abstract

The existence and identity of non-Watson-Crick base pairs (bps) within RNA bulges, internal loops, and hairpin loops cannot reliably be predicted by existing algorithms. We have developed the Isfold (Isosteric Folding) program as a tool to examine patterns of nucleotide substitutions from sequence alignments or mutation experiments and identify plausible bp interactions. We infer these interactions based on the observation that each non-Watson-Crick bp has a signature pattern of isosteric substitutions where mutations can be made that preserve the 3D structure. Isfold produces a dynamic representation of predicted bps within defined motifs in order of their probabilities. The software was developed under Windows XP, and is capable of running on PC and MAC with Matlab 7.1 (SP3) or higher. A PC standalone version that does not require Matlab also is available. This software and a user manual are freely available at www.ucsf.edu/frankel/isfold.     

Functional Analysis of the Autographa californica Multiple Nucleopolyhedrovirus GP64 Terminal Fusion Loops and Interactions with Membranes

The Autographa californica multiple nucleopolyhedrovirus (AcMNPV) glycoprotein GP64 is the major envelope protein of the budded virus (BV). GP64 is a class III fusion protein that mediates BV attachment to the cell surface and low-pH-triggered membrane fusion between the BV envelope and the endosome membrane during entry. Class III fusion proteins contain terminal looped structures that are believed to interact with membranes. To examine the functions of 3 loops found at the apex of the GP64 postfusion structure, we generated 2-alanine substitutions that scanned the two so-called fusion loops (loop 1 and loop 2) plus an adjacent loop structure (loop 3) that is closely attached to loop 2 and is also found at the apex of the GP64 postfusion structure. We identified essential residues from Y75 to T86 (loop 1) and N149 to H156 (loop 2) that are required for fusion activity, but no essential residues in loop 3. Further analysis revealed that critical fusion loop residues fall within two groups that are associated with either membrane merger (hemifusion) or fusion pore expansion. We next examined the interactions of soluble GP64 proteins and BV with membranes composed of various phospholipids. BV interacted directly with small unilamellar vesicles (SUVs) comprised of phospholipids phosphatidylcholine and phosphatidic acid (PC/PA) or phosphatidylcholine and phosphatidylserine (PC/PS) under neutral and acidic pH. We also examined the interactions of soluble GP64 constructs containing substitutions of the most hydrophobic residues within each of the two fusion loops. We found that a 2-residue substitution in either single loop (loop 1 [positions 81 and 82] or loop 2 [positions 153 and 154]) was not sufficient to substantially reduce the GP64-liposome interaction, but the same substitutions in both fusion loops severely reduced the GP64-liposome association at neutral pH. These results suggest that critical hydrophobic residues in both fusion loops may be involved in the interaction of GP64 with host cellular membranes and direct GP64-membrane interactions may represent a receptor-binding step prior to a low-pH-triggered conformational change.     

Ribostral: an RNA 3D alignment analyzer and viewer based on basepair isostericities

RNA atomic resolution structures have revealed the existance of different families of basepair interactions, each of which with its own isosteric sub-families. Ribostral (Ribonucleic Structural Aligner) is a user-friendly framework for analyzing, evaluating and viewing RNA sequence alignments with at least one available atomic resolution structure. It is the first of its kind that makes direct and easy- to-understand superposition of the isostericity matrices of basepairs observed in the structure onto sequence alignments, easily indicating allowed and unallowed substitutions at each BP position. Potential mistakes in the alignments can then be corrected using other sequence editing software. Ribostral has been developed and tested under Windows XP, and is capable of running on any PC or MAC platform with MATLAB 7.1 (SP3) or higher installed version. A stand-alone version is also available for the PC platform. AVAILABILITY: http://rna.bgsu.edu/ribostral.     

Efficient construction of long DNA duplexes with internal non-Watson-Crick motifs and modifications

We have developed a semi-synthetic approach for preparing long stretches of DNA (>100 bp) containing internal chemical modifications and/or non-Watson-Crick structural motifs which relies on splint-free, cell-free DNA ligations and recycling of side-products by non-PCR thermal cycling. A double-stranded DNA PCR fragment containing a polylinker in its middle is digested with two restriction enzymes and a small insert ( approximately 20 bp) containing the modification or non-Watson-Crick motif of interest is introduced into the middle. Incorrect products are recycled to starting materials by digestion with appropriate restriction enzymes, while the correct product is resistant to digestion since it does not contain these restriction sites. This semi-synthetic approach offers several advantages over DNA splint-mediated ligations, including fewer steps, substantially higher yields ( approximately 60% overall yield) and ease of use. This method has numerous potential applications, including the introduction of modifications such as fluorophores and cross-linking agents into DNA, controlling the shape of DNA on a large scale and the study of non-sequence-specific nucleic acid-protein interactions.     

The annotation of RNA motifs

The recent deluge of new RNA structures, including complete atomic-resolution views of both subunits of the ribosome, has on the one hand literally overwhelmed our individual abilities to comprehend the diversity of RNA structure, and on the other hand presented us with new opportunities for comprehensive use of RNA sequences for comparative genetic, evolutionary and phylogenetic studies. Two concepts are key to understanding RNA structure: hierarchical organization of global structure and isostericity of local interactions. Global structure changes extremely slowly, as it relies on conserved long-range tertiary interactions. Tertiary RNA-RNA and quaternary RNA-protein interactions are mediated by RNA motifs, defined as recurrent and ordered arrays of non-Watson-Crick base-pairs. A single RNA motif comprises a family of sequences, all of which can fold into the same three-dimensional structure and can mediate the same interaction(s). The chemistry and geometry of base pairing constrain the evolution of motifs in such a way that random mutations that occur within motifs are accepted or rejected insofar as they can mediate a similar ordered array of interactions. The steps involved in the analysis and annotation of RNA motifs in 3D structures are: (a) decomposition of each motif into non-Watson-Crick base-pairs; (b) geometric classification of each basepair; (c) identification of isosteric substitutions for each basepair by comparison to isostericity matrices; (d) alignment of homologous sequences using the isostericity matrices to identify corresponding positions in the crystal structure; (e) acceptance or rejection of the null hypothesis that the motif is conserved.     

Mutational analysis of conserved nucleotides in a self-splicing group I intron.

We have constructed all single base substitutions in almost all of the highly conserved residues of the Tetrahymena self-splicing intron. Mutation of highly conserved residues almost invariably leads to loss of enzymatic activity. In many cases, activity could be regained by making additional mutations that restored predicted base-pairings; these second site suppressors in general confirm the secondary structure derived from phylogenetic data. At several positions, our suppression data can be most readily explained by assuming non-Watson-Crick base-pairings. In addition to the requirements imposed by the secondary structure, the sequence of the intron is constrained by "negative interactions", the exclusion of particular nucleotide sequences that would form undesirable secondary structures. A comparison of genetic and phylogenetic data suggests sites that may be involved in tertiary structural interactions.     

Comparison of the mutations at Hprt exon 3 of T-lymphocytes from B6C3F1 mice and F344 rats exposed by inhalation to 1,3-butadiene or the racemic mixture of 1,2:3,4-diepoxybutane

Experiments were conducted to define the spectra of mutations occurring in Hprt exon 3 of T-cells isolated from spleens of female B6C3F1 mice and F344 rats exposed by inhalation to 1,3-butadiene (BD) or its reactive metabolite, (+/-)-diepoxybutane (DEB). Hprt mutant frequencies (Mfs) in BD-exposed (1250 ppm for 2 weeks or 625 ppm for 4 weeks; 6 h/day, 5 days/week) and DEB-exposed (2 or 4 ppm for 4 weeks or 5 ppm for 6 weeks; 6 h/day, 5 days/week) mice and rats were significantly increased over concurrent control values. Mutant T-cell colonies from control and treated animals were screened for mutations in Hprt exon 3 using PCR amplification of genomic DNA and denaturing gradient gel electrophoresis, followed by sequence analysis. Exon 3 mutations were found at the following frequencies: 20/394 (5%) in control mice, 56/712 (8%) in BD-exposed mice, 59/1178 (5%) in BD-exposed rats, 66/642 (10%) in DEB-exposed mice, and 51/732 (7%) in DEB-exposed rats. Mutations in exposed animals included base substitutions, small deletions (1 to 74 bp), and small insertions (1 to 8 bp), with base substitutions predominating. Among the types of base substitutions observed in mice, the proportions of G.C-->A.T transitions (p=0.035, Fisher's Exact Test) and G.C-->C.G transversions (p=0.05) were significantly different in control vs. BD-exposed animals. Given the small number of exon 3 mutants analyzed, there was a high degree of overlap in the mutational spectra between BD-exposed mice and rats, between BD- and DEB-exposed mice, and between BD- and DEB-exposed rats in terms of the sites with base substitutions, the mutations found at those mutated sites, the relative occurrence of the most frequently observed base substitutions, and the occurrence of a consistent strand bias for the most frequently observed base substitutions. The spectra data suggest that adduction of both G.C and A.T bps is important in the induction of in vivo mutations by BD metabolites in exposed mice and rats.     

Motif prediction in ribosomal RNAs Lessons and prospects for automated motif prediction in homologous RNA molecules

The traditional way to infer RNA secondary structure involves an iterative process of alignment and evaluation of covariation statistics between all positions possibly involved in basepairing. Watson-Crick basepairs typically show covariations that score well when examples of two or more possible basepairs occur. This is not necessarily the case for non-Watson-Crick basepairing geometries. For example, for sheared (trans Hoogsteen/Sugar edge) pairs, one base is highly conserved (always A or mostly A with some C or U), while the other can vary (G or A and sometimes C and U as well). RNA motifs consist of ordered, stacked arrays of non-Watson-Crick basepairs that in the secondary structure representation form hairpin or internal loops, multi-stem junctions, and even pseudoknots. Although RNA motifs occur recurrently and contribute in a modular fashion to RNA architecture, it is usually not apparent which bases interact and whether it is by edge-to-edge H-bonding or solely by stacking interactions. Using a modular sequence-analysis approach, recurrent motifs related to the sarcin-ricin loop of 23S RNA and to loop E from 5S RNA were predicted in universally conserved regions of the large ribosomal RNAs (16S- and 23S-like) before the publication of high-resolution, atomic-level structures of representative examples of 16S and 23S rRNA molecules in their native contexts. This provides the opportunity to evaluate the predictive power of motif-level sequence analysis, with the goal of automating the process for predicting RNA motifs in genomic sequences. The process of inferring structure from sequence by constructing accurate alignments is a circular one. The crucial link that allows a productive iteration of motif modeling and realignment is the comparison of the sequence variations for each putative pair with the corresponding isostericity matrix to determine which basepairs are consistent both with the sequence and the geometrical data.     

A novel RNA structural motif in the selenocysteine insertion element of eukaryotic selenoprotein mRNAs.

In eukaryotes, co-translational insertion of selenocysteine into selenoproteins necessitates the participation of the selenocysteine insertion sequence (SECIS), an element lying in the 3'-untranslated region of selenoprotein mRNAs. We report a detailed experimental study of the secondary structures of the SECIS elements of three selenoprotein mRNAs, the rat and human type I iodothyronine deiodinase (5'DI) and rat glutathione peroxidase (GPx). Based on RNase and chemical probing, a new secondary structure model is established. It is characterized by a stem-loop structure, comprising two helices (I and II) separated by an internal loop, with an apical loop surmounting helix II. Sequence comparisons of 20 SECIS elements, arising from 2 5'DI, 13 GPx, 2 selenoprotein P, and 1 selenoprotein W mRNAs, confirm the secondary structure model. The most striking finding of the experimental study concerns a set of conserved sequences in helix II that interact to form a novel RNA structural motif consisting of a quartet composed of non-Watson-Crick base pairs 5'UGAY3': 5'UGAU3'. The potential for forming the quartet is preserved in 15 SECIS elements, but three consecutive non-Watson-Crick base pairs can nevertheless form in the other five SECIS, the central G.A tandem being invariant in all cases. A 3D model, derived by computer modeling with the use of the solution data, suggests that the base pairing interactions in the G.A tandem are of the type found in GNRA loops. The 3D model displays the quartet lying in an accessible position at the foot of helix II, which is bent at the internal loop, suggesting that the non-Watson-Crick base pair arrangement provides an unusual pattern of chemical groups for putative ligand interaction.     

The chloroplast genome of Phalaenopsis aphrodite (Orchidaceae): comparative analysis of evolutionary rate with that of grasses and its phylogenetic implications

Whether the Amborella/Amborella-Nymphaeales or the grass lineage diverged first within the angiosperms has recently been debated. Central to this issue has been focused on the artifacts that might result from sampling only grasses within the monocots. We therefore sequenced the entire chloroplast genome (cpDNA) of Phalaenopsis aphrodite, Taiwan moth orchid. The cpDNA is a circular molecule of 148,964 bp with a comparatively short single-copy region (11,543 bp) due to the unusual loss and truncation/scattered deletion of certain ndh subunits. An open reading frame, orf91, located in the complementary strand of the rrn23 was reported for the first time. A comparison of nucleotide substitutions between P. aphrodite and the grasses indicates that only the plastid expression genes have a strong positive correlation between nonsynonymous (Ka) and synonymous (Ks) substitutions per site, providing evidence for a generation time effect, mainly across these genes. Among the intron-containing protein-coding genes of the sampled monocots, the Ks of the genes are significantly correlated to transitional substitutions of their introns. We compiled a concatenated 61 protein-coding gene alignment for the available 20 cpDNAs of vascular plants and analyzed the data set using Bayesian inference, maximum parsimony, and neighbor-joining (NJ) methods. The analyses yielded robust support for the Amborella/Amborella-Nymphaeales-basal hypothesis and for the orchid and grasses together being a monophyletic group nested within the remaining angiosperms. However, the NJ analysis using Ka, the first two codon positions, or amino acid sequences, respectively, supports the monocots-basal hypothesis. We demonstrated that these conflicting angiosperm phylogenies are most probably linked to the transitional sites at all codon positions, especially at the third one where the strong base-composition bias and saturation effect take place.     

Structural domains required for Caenorhabditis elegans G protein-coupled receptor kinase 2 (GRK-2) function in vivo

G protein-coupled receptor kinases (GRKs) are key regulators of signal transduction that specifically phosphorylate activated G protein-coupled receptors (GPCRs) to terminate signaling. Biochemical and crystallographic studies have provided great insight into mammalian GRK2/3 interactions and structure. However, despite extensive in vitro characterization, little is known about the in vivo contribution of these described GRK structural domains and interactions to proper GRK function in signal regulation. We took advantage of the disrupted chemosensory behavior characteristic of Caenorhabditis elegans grk-2 mutants to discern the interactions required for proper in vivo Ce-GRK-2 function. Informed by mammalian crystallographic and biochemical data, we introduced amino acid substitutions into the Ce-grk-2 coding sequence that are predicted to selectively disrupt GPCR phosphorylation, Gα(q/11) binding, Gβγ binding, or phospholipid binding. Changing the most amino-terminal residues, which have been shown in mammalian systems to be required specifically for GPCR phosphorylation but not phosphorylation of alternative substrates or recruitment to activated GPCRs, eliminated the ability of Ce-GRK-2 to restore chemosensory signaling. Disrupting interaction between the predicted Ce-GRK-2 amino-terminal α-helix and kinase domain, posited to stabilize GRKs in their active ATP- and GPCR-bound conformation, also eliminated Ce-GRK-2 chemosensory function. Finally, although changing residues within the RH domain, predicted to disrupt interaction with Gα(q/11), did not affect Ce-GRK-2 chemosensory function, disruption of the predicted PH domain-mediated interactions with Gβγ and phospholipids revealed that both contribute to Ce-GRK-2 function in vivo. Combined, we have demonstrated functional roles for broadly conserved GRK2/3 structural domains in the in vivo regulation of organismal behavior.     

Viral escape from antisense RNA

RNA coliphage SP was propagated for several generations on a host expressing an inhibitory antisense RNA complementary to bases 31–270 of the positive-stranded genome. Phages evolved that escaped inhibition. Typically, these escape mutants contained 3–4 base substitutions, but different sequences were observed among different isolates. The mutations were located within three different types of structural features within the predicted secondary structure of SP genomic RNA: (i) hairpin loops; (ii) hairpin stems; and (iii) the 5' region of the phage genome complementary to the antisense molecule. Computer modelling of the mutant genomic RNAs showed that all of the substitutions within hairpin stems improved the Watson–Crick pairing of the stem. No major structural rearrangements were predicted for any of the mutant genomes, and most substitutions in coding regions did not alter the amino acid sequence. Although the evolved phage populations were polymorphic for substitutions, many substitutions appeared independently in two selected lines. The creation of a new, perfect, antisense RNA against an escape mutant resulted in the inhibition of that mutant but not of other escape mutants nor of the ancestral, unevolved phage. Thus, at least in this system, a population of viruses that evolved to escape from a single antisense RNA would require a cocktail of several antisense RNAs for inhibition.     

BiVisu: software tool for bicluster detection and visualization

BiVisu is an open-source software tool for detecting and visualizing biclusters embedded in a gene expression matrix. Through the use of appropriate coherence relations, BiVisu can detect constant, constant-row, constant-column, additive-related as well as multiplicative-related biclusters. The biclustering results are then visualized under a 2D setting for easy inspection. In particular, parallel coordinate (PC) plots for each bicluster are displayed, from which objective and subjective cluster quality evaluation can be performed. Availability: BiVisu has been developed in Matlab and is available at http://www.eie.polyu.edu.hk/~nflaw/Biclustering/.     

ApA Dinucleotide Periodicity in Prokaryote, Eukaryote, and Organelle Genomes

Computer analyses of various genome sequences revealed the existence of certain periodical patterns of adenine–adenine dinucleotides (ApA). For each genome sequence of 13 eubacteria, 3 archaebacteria, 10 eukaryotes, 60 mitochondria, and 9 chloroplasts, we counted frequencies of ApA dinucleotides at each downstream position within 50 bp from every ApA. We found that the complete genomes of all three archaebacteria have clear ApA periodicities of about 10 bps. On the other hand, all of the 13 eubacteria we analyzed were found to have an ApA periodicity of about 11 bp. Similar periodicities exist in the 10 eukaryotes, although higher organisms such as primates tend to have weaker periodic patterns. None of the mitochondria and chroloplasts we analyzed showed an evident periodic pattern. Received: 3 November 1998 / Accepted: 24 March 1999     

Membrane-bound orientation and position of the synaptotagmin C2B domain determined by site-directed spin labeling

Site-directed spin labeling is used to determine the orientation and depth of insertion of the second C2 domain from synaptotagmin I (C2B) into membrane vesicles composed of phosphatidylcholine (PC) and phosphatidylserine (PS). EPR line shapes of spin-labeled mutants located with the Ca(2+)-binding loops of C2B broaden in the presence of Ca(2+) and PC/PS vesicles, indicating that these loops undergo a Ca(2+)-dependent insertion into the membrane interface. Power saturation of the EPR spectra provides a position for each spin-labeled site along the bilayer normal, and these EPR-derived distance constraints, along with a high-resolution structure of the C2B domain, are used to generate a model for the domain orientation and position at the membrane interface. Our data show that the isolated C2B domain from synaptotagmin I penetrates PC/PS membranes, and that the backbone of Ca(2+)-binding loops 1 and 3 is inserted below the level of a plane defined by the lipid phosphates. The side chains of several loop residues are within the bilayer interior, and both Ca(2+)-binding sites are positioned near a plane defined by the lipid phosphates. A Tb(3+)-based fluorescence assay is used to compare the membrane affinity of the C2B domain to that of the first synaptotagmin C2 domain (C2A). Both C2A and C2B bind PC/PS (75:25) membrane vesicles with a micromolar lipid affinity in the presence of metal ion. These results indicate that C2A and C2B have a similar membrane affinity and position when bound to PC/PS (75:25) membrane vesicles. EPR spectroscopy indicates that the C2B domain has different interactions with PC/PS membranes containing 1 mol % phosphatidylinositol 4,5-bisphosphate.     

Use of 8-methoxypsoralen and long wavelength ultraviolet radiation for decontamination of platelet concentrates.

Transmission of viral diseases through blood products remains a problem in transfusion medicine. We have developed a photochemical decontamination system (PCD) for platelet concentrates (PC) utilizing treatment with long wavelength ultraviolet radiation (UVA, 320-400 nm) and 8-methoxypsorlan (8-MOP). This system is capable of inactivating 25-30 logs/hour of bacteria E. coli or S. aureus, 6 logs/hour of bacteriophage fd, 0.9 log/hour of bacteriophage R17, and 1.1 logs/hour of feline leukemia virus (FeLV) in PC. Immediately following 6 hours of PCD treatment, platelet integrity and function of PCD-treated and control PC were equivalent. After overnight storage, PCD-treated and control PC platelet properties were equal, but there was a slight reduction in TXB-2 production of PCD-treated PC compared to controls. Following PCD treatment, PC were stored for 48 to 96 hours. Platelet counts, morphology scores, extracellular LDH levels, aggregation response, dense body (db) content, and alpha granule (alpha g) content of PCD-treated and control PC were comparable. We assessed the ability of the PCD technique to inactivate intracellular and extracellular virus, quantified the degree of DNA adduct formation in contaminating lymphocytes, and measured the inhibition of polymerase chain reaction (PCR) mediated amplification of intracellular DNA. High titers of cell-free murine cytomegalovirus added to human platelet concentrates (final concentration 10(6)) were inactivated by PCD within 30 minutes. Cat renal fibroblasts infected at high levels with feline rhinotracheitis virus (FeRTV) were seeded into PC followed by PCD treatment with inactivation of 4.8 logs of FeRTV within 10 minutes. Purified human lymphocytes were seeded into PC and treated with PCD in the presence of 3H 8-MOP. Six hours of PCD treatment resulted in the formation of 9.3 to 12.8 8-MOP adducts per 1000 base pairs (bp) of DNA. PCR amplification of a 242 bp segment at the HLA-DQ alpha locus was examined. Inhibition of PCR DNA amplification was dependent on the numbers of 8-MOP adducts formed, and no amplification was present when greater than 12 adducts per 1000 bp were formed. These studies indicate that PCD can effectively inactivate high titers of cell-associated and cell-free virus seeded into standard human PC. The efficiency of DNA adduct formation can be quantitated, and the level of 8-MOP adduct formation in lymphocytes contaminating PC is comparable to the level of adduct formation in cellular DNA reported in the absence of platelets.     

An ancient anion‐binding structural module in RNA and DNA helicases

RNA and DNA helicases manipulate or translocate along single strands of nucleic acids by grasping them using a conserved structural motif. We have examined the available crystal structures of helicases of the two principal superfamilies, SF1 and SF2, and observed that the most conserved interactions with the nucleic acid occur between the phosphosugar backbone of a trinucleotide and the three strand‐helix loops within a (β‐strand/α‐helix)₃ structural module. At the first and third loops is a conserved hydrogen‐bonded feature called a thr‐motif, often seen at α‐helical N‐termini, with the threonine as the N‐cap residue. These loops can be aligned with few insertions or deletions, and their main chain atoms are structurally congruent amongst the family members and between the two modules found as tandem pairs in all SF1 and SF2 proteins. The other highly conserved interactions with nucleic acid involve mainchain NH groups, often at the helical N‐termini, interacting with phosphate groups. We comment on how the sequence motifs that are commonly used to identify helicases map to locations on the module and discuss the implications of the conserved orientation of nucleic acid on the surface of the module for directional stepping along DNA or RNA. Proteins 2010. © 2010 Wiley‐Liss, Inc.     

Adenine protonation in domain B of the hairpin ribozyme

Protein enzymes often use ionizable side chains, such as histidine, for general acid-base catalysis because the imidazole pK(a) is near neutral pH. RNA enzymes, on the other hand, are comprised of nucleotides which do not have apparent pK(a) values near neutral pH. Nevertheless, it has been recently shown that cytidine and adenine protonation can play an important role in both nucleic acid structure and catalysis. We have employed heteronuclear NMR methods to determine the pK(a) values and time scales of chemical exchanges associated with adenine protonation within the catalytically essential B domain of the hairpin ribozyme. The large, adenine-rich internal loop of the B domain allows us to determine adenine pK(a) values for a variety of non-Watson-Crick base pairs. We find that adenines within the internal loop have pK(a) values ranging from 4.8 to 5.8, significantly higher than the free mononucleotide pK(a) of 3. 5. Adenine protonation results in potential charge stabilization, hydrogen bond formation, and stacking interactions that are expected to stabilize the internal loop structure at low pH. Fast proton exchange times of 10-50 micros were determined for the well-resolved adenines. These results suggest that shifted pK(a) values may be a common feature of adenines in non-Watson-Crick base pairs, and identify two adenines which may participate in hairpin ribozyme active site chemistry.