The application of cluster analysis in the intercomparison of loop structures in RNA

We have developed a computational approach for the comparison and classification of RNA loop structures. Hairpin or interior loops identified in atomic resolution RNA structures were intercompared by conformational matching. The root-mean-square deviation (RMSD) values between all pairs of RNA fragments of interest, even if from different molecules, are calculated. Subsequently, cluster analysis is performed on the resulting matrix of RMSD distances using the unweighted pair group method with arithmetic mean (UPGMA). The cluster analysis objectively reveals groups of folds that resemble one another. To demonstrate the utility of the approach, a comprehensive analysis of all the terminal hairpin tetraloops that have been observed in 15 RNA structures that have been determined by X-ray crystallography was undertaken. The method found major clusters corresponding to the well-known GNRA and UNCG types. In addition, two tetraloops with the unusual primary sequence UMAC (M is A or C) were successfully assigned to the GNRA cluster. Larger loop structures were also examined and the clustering results confirmed the occurrence of variations of the GNRA and UNCG tetraloops in these loops and provided a systematic means for locating them. Nineteen examples of larger loops that closely resemble either the GNRA or UNCG tetraloop were found in the large ribosomal RNAs. When the clustering approach was extended to include all structures in the SCOR database, novel relationships were detected including one between the ANYA motif and a less common folding of the GAAA tetraloop sequence.     

Automated classification of RNA 3D motifs and the RNA 3D Motif Atlas

The analysis of atomic-resolution RNA three-dimensional (3D) structures reveals that many internal and hairpin loops are modular, recurrent, and structured by conserved non-Watson–Crick base pairs. Structurally similar loops define RNA 3D motifs that are conserved in homologous RNA molecules, but can also occur at nonhomologous sites in diverse RNAs, and which often vary in sequence. To further our understanding of RNA motif structure and sequence variability and to provide a useful resource for structure modeling and prediction, we present a new method for automated classification of internal and hairpin loop RNA 3D motifs and a new online database called the RNA 3D Motif Atlas. To classify the motif instances, a representative set of internal and hairpin loops is automatically extracted from a nonredundant list of RNA-containing PDB files. Their structures are compared geometrically, all-against-all, using the FR3D program suite. The loops are clustered into motif groups, taking into account geometric similarity and structural annotations and making allowance for a variable number of bulged bases. The automated procedure that we have implemented identifies all hairpin and internal loop motifs previously described in the literature. All motif instances and motif groups are assigned unique and stable identifiers and are made available in the RNA 3D Motif Atlas (http://rna.bgsu.edu/motifs), which is automatically updated every four weeks. The RNA 3D Motif Atlas provides an interactive user interface for exploring motif diversity and tools for programmatic data access.     

Frequent use of the same tertiary motif by self-folding RNAs.

We have identified an 11 nucleotide RNA motif, [CCUAAG...UAUGG], that is extraordinarily abundant in group I and group II self-splicing introns at sites known, or suspected from co-variation analysis, to interact with hairpin terminal loops with a GNRA consensus sequence. Base substitution experiments using a ribozyme-substrate system derived from a group I intron reveal that this motif interacts preferentially with GAAA terminal loops and binds them with remarkable affinity, compared with other known combinations of GNRA loops and matched targets. A copy of the [CCUAAG...UAUGG] motif which is present in domain I of many group II introns is shown to interact with the GAAA terminal loop that caps domain V. This is the first interaction to be identified between these two domains, whose mutual recognition is known to be necessary and sufficient for group II ribozymic activity. We conclude that interaction of [CCUAAG...UAUGG] with GAAA loops is one of the most common solutions used by nature to solve the problem of compacting and bringing together RNA structural domains.     

Predicting U-turns in ribosomal RNA with comparative sequence analysis

The U-turn is a well-known RNA motif characterized by a sharp reversal of the RNA backbone following a single-stranded uridine base. In experimentally determined U-turn motifs, the nucleotides 3' to the turn are frequently involved in tertiary interactions, rendering this motif particularly attractive in RNA modeling and functional studies. The U-turn signature is composed of an UNR sequence pattern flanked by a Y:Y, Y:A (Y=pyrimidine) or G:A base juxtaposition. We have identified 33 potential UNR-type U-turns and 25 related GNRA-type U-turns in a large set of aligned 16 S and 23 S rRNA sequences. U-turn candidates occur in hairpin loops (34 times) as well as in internal and multi-stem loops (24 times). These are classified into ten families based on loop type, sequence pattern (UNR or GNRA) and the nature of the closing base juxtaposition. In 13 cases, the bases on the 3' side of the turn, or on the immediate 5' side, are involved in tertiary covariations, making these sites strong candidates for tertiary interactions.     

A thermodynamic study of unusually stable RNA and DNA hairpins.

About 70% of the RNA tetra-loop sequences identified in ribosomal RNAs from different organisms fall into either (UNCG) or (GNRA) families (where N = A, C, G, or U; and R = A or G). RNA hairpins with these loop sequences form unusually stable tetra-loop structures. We have studied the RNA hairpin GGAC(UUCG)GUCC and several sequence variants to determine the effect of changing the loop sequence and the loop-closing base pair on the thermodynamic stability of (UNCG) tetra-loops. The hairpin GGAG(CUUG)CUCC with the conserved loop G(CUUG)C was also unusually stable. We have determined melting temperatures (Tm), and obtained thermodynamic parameters for DNA hairpins with sequences analogous to stable RNA hairpins with (UNCG), C(GNRA)G, C(GAUA)G, and G(CUUG)C loops. DNA hairpins with (TTCG), (dUdUCG), and related sequences in the loop, unlike their RNA counterparts, did not form unusually stable hairpins. However, DNA hairpins with the consensus loop sequence C(GNRA)G were very stable compared to hairpins with C(TTTT)G or C(AAAA)G loops. The C(GATA)G and G(CTTG)C loops were also extra stable. The relative stabilities of the unusually stable DNA hairpins are similar to those observed for their RNA analogs.     

Evidence that folding of an RNA tetraloop hairpin is less cooperative than its DNA counterpart

Hairpin secondary structural elements play important roles in the folding and function of RNA and DNA molecules. Previous work from our lab on small DNA hairpin loop motifs, d(cGNAg) and d(cGNABg) (where B is C, G, or T), showed that folding is highly cooperative and obeys indirect coupling, consistent with a concerted transition. Herein, we investigate folding of the related, exceptionally stable RNA hairpin motif, r(cGNRAg) (where R is A or G). Previous NMR characterization identified a complex network of seven hydrogen bonds in this loop. We inserted three carbon (C3) spacers throughout the loop and found coupling between G1 of the loop and the CG closing base pair, similar to that found in DNA. These data support a GNRA motif being expandable at any position but before the G. Thermodynamic measurements of nucleotide-analogue-substituted oligonucleotides revealed pairwise-coupling free energies ranging from weak to strong. When coupling free energies were remeasured in the background of changes at a third site, they remained essentially unchanged even though all of the sites were coupled to each other. This type of coupling, referred to as "direct", is peculiar to the RNA loop. The data suggest that, for small stable loops, folding of RNA obeys a model with nearest-neighbor interactions, while folding of DNA follows a more concerted process in which the stabilizing interactions are linked through a conformational change. The lesser cooperativity in RNA loops may provide a more robust loop that can withstand mutations without a severe loss in stability. These differences may enhance the ability of RNA to evolve.     

Intermolecular interaction between a branching ribozyme and associated homing endonuclease mRNA

RNA tertiary interactions involving docking of GNRA (N; any base; R; purine) hairpin loops into helical stem structures on other regions of the same RNA are one of the most common RNA tertiary interactions. In this study, we investigated a tertiary association between a GAAA hairpin tetraloop in a small branching ribozyme (DiGIR1) and a receptor motif (HEG P1 motif) present in a hairpin structure on a separate mRNA molecule. DiGIR1 generates a 2', 5' lariat cap at the 5' end of its downstream homing endonuclease mRNA by catalysing a self-cleavage branching reaction at an internal processing site. Upon release, the 5' end of the mRNA forms a distinct hairpin structure termed HEG P1. Our biochemical data, in concert with molecular 3D modelling, provide experimental support for an intermolecular tetraloop receptor interaction between the L9 GAAA in DiGIR1 and a GNRA tetraloop receptor-like motif (UCUAAG-CAAGA) found within the HEG P1. The biological role of this interaction appears to be linked to the homing endonuclease expression by promoting post-cleavage release of the lariat capped mRNA. These findings add to our understanding of how protein-coding genes embedded in nuclear ribosomal DNA are expressed in eukaryotes and controlled by ribozymes.     

A selection system for functional internal ribosome entry site (IRES) elements: analysis of the requirement for a conserved GNRA tetraloop in the encephalomyocarditis virus IRES.

Picornavirus internal ribosome entry site (IRES) elements direct cap-independent internal initiation of protein synthesis within mammalian cells. These RNA elements (about 450 nt) contain extensive secondary structure including a hairpin loop with a conserved GNRA motif. Such loops are important in RNA-RNA and RNA-protein interactions. Plasmids that express dicistronic mRNAs of the structure GUS/IRES/HOOK have been constructed. The HOOK sequence encodes a cell-surface-targeted protein (sFv); the translation of this open reading frame within mammalian cells from these dicistronic mRNAs requires a functional IRES element. Cells that express the sFv can be selected from nonexpressing cells. A pool of up to 256 mutant encephalomyocarditis virus IRES elements was generated by converting the wild-type hairpin loop sequence (GCGA) to NNNN. Following transfection of this pool of mutants into COS-7 cells, plasmids were recovered from selected sFv-expressing cells. These DNAs were amplified in Escherichia coli and transfected again into COS-7 cells for further cycles to enrich for plasmids encoding functional IRES elements. The sequence of individual selected IRES elements was determined. All functional IRES elements had a tetraloop with a 3' terminal A residue. Optimal IRES activity, assayed in vitro and within cells, was obtained from plasmids encoding an IRES with the hairpin loop sequence fitting a RNRA consensus. In contrast, IRES elements containing YCYA tetraloops were severely defective.     

A conserved motif in group IC3 introns is a new class of GNRA receptor

Terminal tetraloops consisting of GNRA sequences are often found in biologically active large RNAs. The loops appear to contribute towards the organization of higher order RNA structures by forming specific tertiary interactions with their receptors. Group IC3 introns which possess a GAAA loop in the L2 region often have a phylogenetically conserved motif in their P8 domains. In this report, we show that this conserved motif stands as a new class of receptor that distinguishes the sequences of GNRA loops less stringently than previously known receptors. The motif can functionally substitute an 11 nt motif receptor in the Tetrahymena ribozyme. Its structural and functional similarity to one class of synthetic receptors obtained from in vitro selection is observed.     

The lonepair triloop: a new motif in RNA structure

The lonepair triloop (LPTL) is an RNA structural motif that contains a single ("lone") base-pair capped by a hairpin loop containing three nucleotides. The two nucleotides immediately outside of this motif (5' and 3' to the lonepair) are not base-paired to one another, restricting the length of this helix to a single base-pair. Four examples of this motif, along with three tentative examples, were initially identified in the 16S and 23S rRNAs with covariation analysis. An evaluation of the recently determined crystal structures of the Thermus thermophilus 30S and Haloarcula marismortui 50S ribosomal subunits revealed the authenticity for all of these proposed interactions and identified 16 more LPTLs in the 5S, 16S and 23S rRNAs. This motif is found in the T loop in the tRNA crystal structures. The lonepairs are positioned, in nearly all examples, immediately 3' to a regular secondary structure helix and are stabilized by coaxial stacking onto this flanking helix. In all but two cases, the nucleotides in the triloop are involved in a tertiary interaction with another section of the rRNA, establishing an overall three-dimensional function for this motif. Of these 24 examples, 14 occur in multi-stem loops, seven in hairpin loops and three in internal loops. While the most common lonepair, U:A, occurs in ten of the 24 LPTLs, the remaining 14 LPTLs contain seven different base-pair types. Only a few of these lonepairs adopt the standard Watson-Crick base-pair conformations, while the majority of the base-pairs have non-standard conformations. While the general three-dimensional conformation is similar for all examples of this motif, characteristic differences lead to several subtypes present in different structural environments. At least one triloop nucleotide in 22 of the 24 LPTLs in the rRNAs and tRNAs forms a tertiary interaction with another part of the RNA. When a LPTL containing the GNR or UYR triloop sequence forms a tertiary interaction with the first (and second) triloop nucleotide, it recruits a fourth nucleotide to mediate stacking and mimic the tetraloop conformation. Approximately half of the LPTL motifs are in close association with proteins. The majority of these LPTLs are positioned at sites in rRNAs that are conserved in the three phylogenetic domains; a few of these occur in regions of the rRNA associated with ribosomal function, including the presumed site of peptidyl transferase activity in the 23S rRNA.     

Structural organization of a viral IRES depends on the integrity of the GNRA motif

Little is known about the tertiary structure of internal ribosome entry site (IRES) elements. The central domain of foot-and-mouth disease (FMDV) IRES, named 3 or I, contains a conserved GNRA motif, essential for IRES activity. We have combined functional analysis with RNA probing to define its structural organization. We have found that a UNCG motif does not functionally substitute the GNRA motif; moreover, binding of synthetic GNRA stem-loops to domain 3 was significantly reduced in RNAs bearing UCCG or GUAG substitutions. The apical region of domain 3 consists of a four-way junction where residues of the GNRA tetraloop are responsible for the organization of the adjacent stem-loops, as deduced from ribonucleases and dimethyl sulfate accessibility. A single A-to-G substitution in the fourth position of this motif led to a strong RNA reorganization, affecting several nucleotides away in the secondary structure of domain 3. The study of mutants bearing UNCG or GUAG tetraloops revealed lack of protection to chemical attack in native RNA at specific nucleotides relative to the parental GUAA, suggesting that the GNRA motif dictates the organization and stability of domain 3. This effect is likely mediated by the interaction with distant residues. Therefore, the GNRA motif plays a crucial role in the organization of IRES structure with important consequences on activity.     

Conserved structural motifs located in distal loops of aphthovirus internal ribosome entry site domain 3 are required for internal initiation of translation.

A comparison of picornavirus internal ribosome entry site (IRES) secondary structures revealed the existence of conserved motifs located on loops. We have carried out a mutational analysis to test their requirement for IRES-driven translation. The GUAA sequence, located in the aphthovirus 3A loop, did not tolerate substitutions that disrupt the GNRA motif. Interestingly, this motif was found at similar positions in all picornavirus IRESs, suggesting that it may form part of a tertiary-structure element. The RAAA tetranucleotide located in the 3B loop was conserved only in cardiovirus and aphthovirus. A mutational analysis of the RAAA motif revealed that activities of 3B loop mutants correlated with both the presence of a sequence close to CAAA at the new 3B loop and the absence of reorganization of the 3B and 3C stem-loops. In support of this conclusion, insertion of a large number of nucleotides close to the 3B loop, which was predicted to reorganize the 3B-3C stem-loop structure, led to defective IRES elements. We conclude that the aphthovirus IRES loops located at the most distal part of domain 3, which carries GNRA and RAAA motifs, are essential for IRES function.     

NMR structure of a 4 x 4 nucleotide RNA internal loop from an R2 retrotransposon: identification of a three purine-purine sheared pair motif and comparison to MC-SYM predictions

The NMR solution structure is reported of a duplex, 5'GUGAAGCCCGU/3'UCACAGGAGGC, containing a 4 × 4 nucleotide internal loop from an R2 retrotransposon RNA. The loop contains three sheared purine-purine pairs and reveals a structural element found in other RNAs, which we refer to as the 3RRs motif. Optical melting measurements of the thermodynamics of the duplex indicate that the internal loop is 1.6 kcal/mol more stable at 37°C than predicted. The results identify the 3RRs motif as a common structural element that can facilitate prediction of 3D structure. Known examples include internal loops having the pairings: 5'GAA/3'AGG, 5'GAG/3'AGG, 5'GAA/3'AAG, and 5'AAG/3'AGG. The structural information is compared with predictions made with the MC-Sym program.     

A test of the model to predict unusually stable RNA hairpin loop stability

To investigate the accuracy of a model [Giese et al., 1998, Biochemistry37:1094-1100 and Mathews et al., 1999, JMol Biol 288:911-940] that predicts the stability of RNA hairpin loops, optical melting studies were conducted on sets of hairpins previously determined to have unusually stable thermodynamic parameters. Included were the tetraloops GNRA and UNCG (where N is any nucleotide and R is a purine), hexaloops with UU first mismatches, and the hairpin loop of the iron responsive element, CAGUGC. The experimental values for the GNRA loops are in excellent agreement (deltaG degrees 37 within 0.2 kcal/mol and melting temperature (TM) within 4 degrees C) with the values predicted by the model. When the UNCG hairpin loops are treated as tetraloops, and a bonus of 0.8 kcal/mol included in the prediction to account for the extra stable first mismatch (UG), the measured and predicted values are also in good agreement (deltaG degrees 37 within 0.7 kcal/mol and TM within 3 degrees C). Six hairpins with unusually stable UU first mismatches also gave good agreement with the predictions (deltaG degrees 37 within 0.5 kcal/mol and TM within 8 degrees C), except for hairpins closed by wobble base pairs. For these hairpins, exclusion of the additional stabilization term for UU first mismatches improved the prediction (AG degrees 37 within 0.1 kcal/mol and TM within 3 degrees C). Hairpins with the iron-responsive element loop were not predicted well by the model, as measured deltaG degrees 37 values were at least 1 kcal/mol greater than predicted.     

Clustering RNA structural motifs in ribosomal RNAs using secondary structural alignment

RNA structural motifs are the building blocks of the complex RNA architecture. Identification of non-coding RNA structural motifs is a critical step towards understanding of their structures and functionalities. In this article, we present a clustering approach for de novo RNA structural motif identification. We applied our approach on a data set containing 5S, 16S and 23S rRNAs and rediscovered many known motifs including GNRA tetraloop, kink-turn, C-loop, sarcin-ricin, reverse kink-turn, hook-turn, E-loop and tandem-sheared motifs, with higher accuracy than the state-of-the-art clustering method. We also identified a number of potential novel instances of GNRA tetraloop, kink-turn, sarcin-ricin and tandem-sheared motifs. More importantly, several novel structural motif families have been revealed by our clustering analysis. We identified a highly asymmetric bulge loop motif that resembles the rope sling. We also found an internal loop motif that can significantly increase the twist of the helix. Finally, we discovered a subfamily of hexaloop motif, which has significantly different geometry comparing to the currently known hexaloop motif. Our discoveries presented in this article have largely increased current knowledge of RNA structural motifs.     

Stability of single-nucleotide bulge loops embedded in a GAAA RNA hairpin stem

Forty-six RNA hairpins containing combinations of 3' or 5' bulge loops and a 3' or 5' fluorescein label were optically melted in 1 M NaCl, and the thermodynamic parameters ΔH°, ΔS°, ΔG°(37), and T(M) for each hairpin were determined. The bulge loops were of the group I variety, in which the identity of the bulge is known, and the group II variety, in which the bulged nucleotide is identical to one of its nearest neighbors, leading to ambiguity as to the exact position of the bulge. The fluorescein label at either the 3' end or 5' end of the hairpin did not significantly influence the stability of the hairpin. As observed with bulge loops inserted into a duplex motif, the insertion of a bulge loop into the stem of a hairpin loop was destabilizing. The model developed to predict the influence of bulge loops on the stability of duplex formation was extended to predict the influence of bulge loops on hairpin stability. Specifically, the influence of the bulge is related to the stability of the hairpin stem distal from the hairpin loop.     

Rules for RNA recognition of GNRA tetraloops deduced by in vitro selection: comparison with in vivo evolution.

Terminal loops with a GNRA consensus sequence are a prominent feature of large self-assembling RNA molecules. In order to investigate tertiary interactions involving GNRA loops, we have devised an in vitro selection system derived from a group I ribozyme. Two selections, destined to isolate RNA sequences that would recognize two of the most widespread loops (GUGA and GAAA), yielded variants of previously identified receptors for those loops, and also some yet unrecognized, high-affinity binders with novel specificities towards members of the GNRA family. By taking advantage of available crystal structures, we have attempted to rationalize these results in terms of RNA-RNA contacts and to expose some of the structural principles that govern GNRA loop-mediated tertiary interactions; the role of loop nucleotide 2 in ensuring specific recognition by receptors is emphasized. More generally, comparison of the products of in vitro and natural selection is shown to provide insights into the mechanisms underlying the in vivo evolution of self-assembling RNA molecules.     

UNAC tetraloops: to what extent do they mimic GNRA tetraloops?

The structures of four small RNAs each containing a different version of the UNAC loop were determined in solution using NMR spectroscopy and restrained molecular dynamics. The UMAC tetraloops (where M is A or C) exhibited a typical GNRA fold including at least one hydrogen bond between the first U and fourth C. In contrast, UGAC and UUAC tetraloops have a different orientation of the first and fourth residues, such that they do not closely mimic the GNRA fold. Although the UMAC tetraloops are excellent structural mimics of the GNRA tetraloop backbone, sequence comparisons typically do not reveal co‐variation between the two loop types. The limited covariation is attributed to differences in the location of potential hydrogen bond donors and acceptors as a result of the replacement of the terminal A of GNRA with C in the UMAC version. Thus, UMAC loops do not readily form the common GNRA tetraloop‐receptor interaction. The loop at positions 863‐866 in E. coli 16S ribosomal RNA appears to be a major exception. However, in this case the GNRA loop does not in fact engage in the usual base to backbone tertiary interactions. In summary, UMAC loops are not just an alternative sequence version of the GNRA loop family, but instead they expand the types of interactions, or lack thereof, that are possible. From a synthetic biology perspective their inclusion in an artificial RNA may allow the establishment of a stable loop structure while minimizing unwanted long range interactions or permitting alternative long‐range interactions. © 2012 Wiley Periodicals, Inc. Biopolymers 97: 617–628, 2012.     

Identifying novel sequence variants of RNA 3D motifs

Predicting RNA 3D structure from sequence is a major challenge in biophysics. An important sub-goal is accurately identifying recurrent 3D motifs from RNA internal and hairpin loop sequences extracted from secondary structure (2D) diagrams. We have developed and validated new probabilistic models for 3D motif sequences based on hybrid Stochastic Context-Free Grammars and Markov Random Fields (SCFG/MRF). The SCFG/MRF models are constructed using atomic-resolution RNA 3D structures. To parameterize each model, we use all instances of each motif found in the RNA 3D Motif Atlas and annotations of pairwise nucleotide interactions generated by the FR3D software. Isostericity relations between non-Watson–Crick basepairs are used in scoring sequence variants. SCFG techniques model nested pairs and insertions, while MRF ideas handle crossing interactions and base triples. We use test sets of randomly-generated sequences to set acceptance and rejection thresholds for each motif group and thus control the false positive rate. Validation was carried out by comparing results for four motif groups to RMDetect. The software developed for sequence scoring (JAR3D) is structured to automatically incorporate new motifs as they accumulate in the RNA 3D Motif Atlas when new structures are solved and is available free for download.