The evolution of homing endonuclease genes and group I introns in nuclear rDNA

Group I introns are autonomous genetic elements that can catalyze their own excision from pre-RNA. Understanding how group I introns move in nuclear ribosomal (r)DNA remains an important question in evolutionary biology. Two models are invoked to explain group I intron movement. The first is termed homing and results from the action of an intron-encoded homing endonuclease that recognizes and cleaves an intronless allele at or near the intron insertion site. Alternatively, introns can be inserted into RNA through reverse splicing. Here, we present the sequences of two large group I introns from fungal nuclear rDNA, which both encode putative full-length homing endonuclease genes (HEGs). Five remnant HEGs in different fungal species are also reported. This brings the total number of known nuclear HEGs from 15 to 22. We determined the phylogeny of all known nuclear HEGs and their associated introns. We found evidence for intron-independent HEG invasion into both homologous and heterologous introns in often distantly related lineages, as well as the "switching" of HEGs between different intron peripheral loops and between sense and antisense strands of intron DNA. These results suggest that nuclear HEGs are frequently mobilized. HEG invasion appears, however, to be limited to existing introns in the same or neighboring sites. To study the intron-HEG relationship in more detail, the S943 group I intron in fungal small-subunit rDNA was used as a model system. The S943 HEG is shown to be widely distributed as functional, inactivated, or remnant ORFs in S943 introns.     

Functionally distinct nucleic acid binding sites for a group I intron encoded RNA maturase/DNA homing endonuclease

A large number of group I introns encode a family of homologous proteins that either promote intron splicing (maturases) or are site-specific DNA endonucleases that function in intron mobility (a process called "homing"). Genetic studies have shown that some of these proteins have both activities, yet how a single protein carries out both functions remains obscure. The similarity between respective DNA-binding sites and the RNA structure near the 5' and 3' splice sites has fueled speculation that such proteins may use analogous interactions to perform both functions. The Aspergillus nidulans mitochondrial COB group I intron encodes a bi-functional protein, I-AniI, that has both RNA maturase and site-specific DNA endonuclease activities in vitro. Here, we show that I-AniI shows distinctive features of the endonuclease family to which it belongs, including highly specific, tight binding and sequential DNA strand cleavage. Competition experiments demonstrate that I-AniI binds the COB intron RNA even in saturating concentrations of its DNA target site substrate, suggesting that the protein has a separate binding site for RNA. In addition, we provide evidence that two different DNA-binding site mutants of I-AniI have little effect on the protein's RNA maturation activity. Since RNA splicing is likely a secondary adaptation of the protein, these observations support a model in which homing endonucleases may have developed maturase function by utilizing a hitherto "non-functional" protein surface.     

Characterization of products derived from self-splicing of intron aI5 alpha which is located in the mitochondrial COX I gene of Saccharomyces cerevisiae.

We have characterized the in vitro self-splicing of intron aI5 alpha containing precursor RNA from the yeast mitochondrial gene coding for cytochrome oxidase subunit I. This intron follows the rules for group I self-splicing introns and all the characteristic products have been identified. In addition we have detected abnormal RNA products with features that indicate that the self-splicing behaviour of this intron is more complex. Two intron circles are formed by use of a major and minor intron-internal site for circle closure. A cryptic 5'-splice site located in the 3' exon results in guanosine nucleotide mediated opening at a position 30 nt downstream of the normal 3' splice site. The reactions can all be explained on the basis of the "splice guide" model proposed by Davies et al (1982 Nature 300 719-724). Although the sequence motifs at cyclization and splice sites occur more often in this intron, only some of them are allowed to interact with the internal guide sequence, suggesting that both primary structure and spatial folding of the RNA are involved in formation of productive reaction sites.     

Exon sequences distant from the splice junction are required for efficient self-splicing of the Tetrahymena IVS.

The presence of a natural rRNA secondary structure element immediately preceding the 5' splice site of the Tetrahymena IVS can inhibit self-splicing by competing with base pairing between the 5' exon and the guide sequence of the IVS (P1). Formation of this alternative hairpin is preferred in short precursor RNAs, and results in loss of G-addition to the 5' splice site. Pre-rRNAs which contain longer exons of ribosomal sequence, however, splice rapidly. As many as 146 nucleotides of the 5' exon and 86 nucleotides of the 3' exon are required for efficient self-splicing of Tetrahymena precursors. The presence of nucleotides distant from the 5' splice site apparently alters the equilibrium between the alternative hairpins, and promotes formation of active precursors. This effect is dependent on the specific sequences of the ribosomal pre-RNA, since point mutations within this region reduce the rate of splicing as much as 50-fold. This system provides an opportunity to study the way in which long-range interactions can influence splice site selection in a highly structured RNA.     

Selection of the 3′-splice site in group I introns

A model for selection of 3′-splice sites in splicing of RNA precursors containing group I introns is presented. The key feature of this model is a newly identified tertiary interaction between the catalytic core of the intron and the 3′-splice site. This tertiary pairing would bring the 3′-splice site into the core of the intron, which is known to contain RNA sequences and structures essential for catalyzing the splicing reactions. The proposed tertiary interaction can coexist with P10, a pairing between 3′-exon sequences and the ‘internal guide sequence’ near the 5′-end of the intron. The model predicts that three RNA-RNA interactions are important in selection of 3′-splice sites: (i) binding of intron sequences with the core; (ii) pairing of exon sequences with the internal guide sequence; and (iii) binding of the terminal guanosine to an unknown site within the core.     

Structural requirement for Mg2+ binding in the group I intron core

Divalent metal ions are required for splicing of group I introns, but their role in maintaining the structure of the active site is still under investigation. Ribonuclease and hydroxyl radical footprinting of a small group I intron from Azoarcus pre-tRNA(Ile) showed that tertiary interactions between helical domains are stable in a variety of cations. Only Mg(2+), however, induced a conformational change in the intron core that correlates with self-splicing activity. Three metal ion binding sites in the catalytic core were identified by Tb(III)-dependent cleavage. Two of these are near bound substrates in a three-dimensional model of the ribozyme. A third metal ion site is near an A minor motif in P3. In the pre-tRNA, Tb(3+) cleavage was redirected to the 5' and 3' splice sites, consistent with metal-dependent activation of splice site phosphodiesters. The results show that many counterions induce global folding, but organization of the group I active site is specifically linked to Mg(2+) binding at a few sites.     

A sequence-based model accounts largely for the relationship of intron positions to protein structural features

Claims of intron-structure correlations have played a major role in debates surrounding split gene origins. In the formative (as opposed to disruptive or "insertional") model of split gene origins, introns represent the scars of chimaeric gene assembly. When analyzed retrospectively, formative introns should tend to fall between modular units, if such units exist, or at least to exhibit a preference for sites favorable to chimaera formation. However, there is another possible source of preferences: under a disruptive model of split gene origins, fortuitous intron-structure correlations may arise because the gain of introns is biased with respect to flanking nucleotide sequences. To investigate the extent to which a sequence-biased intron gain model may account for the present-day distribution of introns, data on over 10,000 introns in eukaryotic protein-coding genes were integrated with structural data from a set of 1,851 nonredundant protein chains. The positions of introns with respect to secondary structures, solvent accessibility, and so-called "modules" were evaluated relative to the expectations of a null model, a disruptive model based on amino acid frequencies at splice junctions, and a formative model defined relative to these. The null model can be excluded for most structural features and is highly improbable when intron sites are grouped by reading frame phase. Phase-dependent correlations with secondary structure and side-chain surface accessibility are particularly strong. However, these phase-dependent correlations are explained largely by the sequence-based disruptive model.     

Computational prediction of efficient splice sites for trans-splicing ribozymes

Group I introns have been engineered into trans-splicing ribozymes capable of replacing the 3'-terminal portion of an external mRNA with their own 3'-exon. Although this design makes trans-splicing ribozymes potentially useful for therapeutic application, their trans-splicing efficiency is usually too low for medical use. One factor that strongly influences trans-splicing efficiency is the position of the target splice site on the mRNA substrate. Viable splice sites are currently determined using a biochemical trans-tagging assay. Here, we propose a rapid and inexpensive alternative approach to identify efficient splice sites. This approach involves the computation of the binding free energies between ribozyme and mRNA substrate. We found that the computed binding free energies correlate well with the trans-splicing efficiency experimentally determined at 18 different splice sites on the mRNA of chloramphenicol acetyl transferase. In contrast, our results from the trans-tagging assay correlate less well with measured trans-splicing efficiency. The computed free energy components suggest that splice site efficiency depends on the following secondary structure rearrangements: hybridization of the ribozyme's internal guide sequence (IGS) with mRNA substrate (most important), unfolding of substrate proximal to the splice site, and release of the IGS from the 3'-exon (least important). The proposed computational approach can also be extended to fulfill additional design requirements of efficient trans-splicing ribozymes, such as the optimization of 3'-exon and extended guide sequences.     

Minimal catalytic domain of a group I self-splicing intron RNA

The self-splicing intron ribozymes have been regarded as primitive forms of the splicing machinery for eukaryotic pre-mRNAs. The splicing activity of group I self-splicing introns is dependent on an absolutely conserved and exceptionally densely packed core region composed of two helical domains, P3-P7 and P4-P6, that are connected rigidly via base triples. Here we show that a mutant group I intron ribozyme lacking both the P4-P6 domain and the base triples can perform the phosphoester transfer reactions required for splicing at both the 5' and 3' splice sites, demonstrating that the elements required for splicing are concentrated in the stacked helical P3-P7 domain. This finding establishes that the conserved core of the intron consists of two physically and functionally separable components, and we present a model showing the architecture of a prototype of this class of intron and the course of its molecular evolution.     

Distribution of rRNA introns in the three-dimensional structure of the ribosome

More than 1200 introns have been documented at over 150 unique sites in the small and large subunit ribosomal RNA genes (as of February 2002). Nearly all of these introns are assigned to one of four main types: group I, group II, archaeal and spliceosomal. This sequence information has been organized into a relational database that is accessible through the Comparative RNA Web Site (http://www.rna.icmb.utexas.edu/) While the rRNA introns are distributed across the entire tree of life, the majority of introns occur within a few phylogenetic groups. We analyzed the distributions of rRNA introns within the three-dimensional structures of the 30S and 50S ribosomes. Most sites in rRNA genes that contain introns contain only one type of intron. While the intron insertion sites occur at many different coordinates, the majority are clustered near conserved residues that form tRNA binding sites and the subunit interface. Contrary to our expectations, many of these positions are not accessible to solvent in the mature ribosome. The correlation between the frequency of intron insertions and proximity of the insertion site to functionally important residues suggests an association between intron evolution and rRNA function.     

Group II introns deleted for multiple substructures retain self-splicing activity.

Group II introns can be folded into highly conserved secondary structures with six major substructures or domains. Domains 1 and 5 are known to play key roles in self-splicing, while the roles of domains 2, 3, 4, and 6 are less clear. A trans assay for domain 5 function has been developed which indicates that domain 5 has a binding site on the precursor RNA that is not predicted from any secondary structure element. In this study, the self-splicing group II intron 5 gamma of the coxI gene of yeast mitochondrial DNA was deleted for various intron domains, singly and in combinations. Those mutant introns were characterized for self-splicing reactions in vitro as a means of locating the domain 5 binding site. A single deletion of domain 2, 3, 4, or 6 does not block in vitro reactions at either splice junction, though the deletion of domain 6 reduces the fidelity of 3' splice site selection somewhat. Even the triple deletion lacking domains 2, 4, and 6 retains some self-splicing activity. The deletion of domains 2, 3, 4, and 6 blocks the reaction at the 3' splice junction but not at the 5' junction. From these results, we conclude that the binding site for domain 5 is within domain 1 and that the complex of 5' exon, domain 1, and domain 5 (plus short connecting sequences) constitutes the essential catalytic core of this intron.     

DNA sequence and secondary structures of the large subunit rRNA coding regions and its two class I introns of mitochondrial DNA fromPodospora anserina

Summary DNA sequence analysis has shown that the gene coding for the mitochondrial (mt) large subunit ribosomal RNA (rRNA) fromPodospora anserina is interrupted by two class I introns. The coding region for the large subunit rRNA itself is 3715 bp and the two introns are 1544 (r1) and 2404 (r2) bp in length. Secondary structure models for the large subunit rRNA were constructed and compared with the equivalent structure fromEscherichia coli 23S rRNA. The two structures were remarkably similar despite an 800-base difference in length. The additional bases in theP. anserina rRNA appear to be mostly in unstructured regions in the 3 part of the RNA. Secondary structure models for the two introns show striking similarities with each other as well as with the intron models from the equivalent introns inSaccharomyces cerevisiae, Neurospora crassa, andAspergillus nidulans. The long open reading frames in each intron are different from each other, however, and the nucleotide sequence similarity diverges as it proceeds away from the core structure. Each intron is located within regions of the large subunit rRNA gene that are highly conserved in both sequence and structure. Computer analysis showed that the open reading frame for intron r1 contained a common maturase-like polypeptide. The open reading frames of intron r2 apeared to be chimeric, displaying high sequence similarity with the open reading frames in the r1 and ATPase 6 introns ofN. crassa.     

Comparison of fungal mitochondrial introns reveals extensive homologies in RNA secondary structure

The complete sequences of nine Saccharomyces cerevisiae mitochondrial introns, six of which carry long open reading frames, have already been published. We have recently determined the sequence of an intron in the large ribosomal mitochondrial RNA of Kluyveromyces thermotolerans (Jacquier et al., in preparation), which we found to be closely related to its S. cerevisiae counterpart. This latter result prompted us to undertake a systematic search for possible homologous elements in the other, available sequences with the help of an original computer program. A previously unsuspected wealth of evolutionarily conserved sequences and secondary structures was thus uncovered. Seven at least of the available sequences may be folded up into elaborate secondary structure models, the cores of which are nearly identical. These models result in bringing together the exon-intron junctions into relatively close spatial proximity and looping out either all or most of the sequences in open reading frame, when present. These results and their possible implications with respect to the mechanism of splicing are discussed in the light of available genetic and biochemical data.     

Group I-like ribozymes with a novel core organization perform obligate sequential hydrolytic cleavages at two processing sites.

A new category of self-splicing group I introns with conserved structural organization and function is found among the eukaryotic microorganisms Didymium and Naegleria. These complex rDNA introns contain two distinct ribozymes with different functions: a regular group I splicing-ribozyme and a small internal group I-like ribozyme (GIR1), probably involved in protein expression. GIR1 was found to cleave at two internal sites in an obligate sequential order. Both sites are located 3' of the catalytic core. GIR1-catalyzed transesterification reactions could not be detected. We have compared all available GIR1 sequences and propose a common RNA secondary structure resembling that of group I splicing-ribozymes, but with some important differences. The GIR1s lack most peripheral sequence components, as well as a P1 segment, and, at approximately 160-190 nt, they are the smallest functional group I ribozymes known from nature. All GIR1s were found to contain a novel 6-bp pseudoknot (P15) within their catalytic core region. Experimental support of the proposed structure was obtained from the Didymium GIR1 by RNA structure probing and site-directed mutagenesis. Three-dimensional modeling indicates a compactly folded ribozyme with the functionally essential P15 exposed in the cleft between the two principal domains P3-P8 and P4-P6.     

Alternative secondary structures in the 5' exon affect both forward and reverse self-splicing of the Tetrahymena intervening sequence RNA

The natural splice junction of the Tetrahymena large ribosomal RNA is flanked by hairpins that are phylogenetically conserved. The stem immediately preceding the splice junction involves nucleotides that also base pair with the internal guide sequence of the intervening sequence during splicing. Thus, precursors which contain wild-type exons can form two alternative helices. We have constructed a series of RNAs where the stem-loop in the 5' exon is more or less stable than in the wild-type precursor, and tested them in both forward and reverse self-splicing reactions. The presence of a stable hairpin in ligated exon substrates interferes with the ability of the intervening sequence to integrate at the splice junction. Similarly, the presence of the wild-type hairpin in the 5' exon reduces the rate of splicing 20-fold in short precursors. The data are consistent with a competition between unproductive formation of a hairpin in the 5' exon and productive pairing of the 5' exon with the internal guide sequence. The reduction of splicing by a hairpin that is a normal feature of rRNA structure is surprising; we propose that this attenuation is relieved in the natural splicing environment.     

The mechanism of group I self-splicing: an internal guide sequence can be provided in trans.

We have reconstituted a group I self-splicing reaction between two RNA molecules with different functional RNA parts: a substrate molecule containing the 5' splice site and a functional internal guide sequence (IGS), and a ribozyme molecule with core structure elements and splice sites but a mutated IGS. The 5' exon of the substrate molecule is ligated in trans to the 3' exon of the ribozyme molecule, suggesting that the deficient IGS in the ribozyme can be replaced by an externally added IGS present on the substrate molecule. This result is different from catalysis mediated by proteins where it is not possible to dissect the specificity of an enzyme from its catalytic activity.     

Molecular genetics of group I introns: RNA structures and protein factors required for splicing--a review

In vivo and in vitro genetic techniques have been widely used to investigate the structure-function relationships and requirements for splicing of group-I introns. Analyses of group-I introns from extremely diverse genetic systems, including fungal mitochondria, protozoan nuclei, and bacteriophages, have yielded results which are complementary and highly consistent. In vivo genetic studies of fungal mitochondrial systems have served to identify cis-acting sequences within mitochondrial introns, and trans-acting protein products of mitochondrial and nuclear genes which are important for splicing, and to show that some mitochondrial introns are mobile genetic elements. In vitro genetic studies of the self-splicing intron within the Tetrahymena thermophila nuclear large ribosomal RNA precursor (Tetrahymena LSU intron) have been used to examine essential and nonessential RNA sequences and structures in RNA-catalyzed splicing. In vivo and in vitro genetic analysis of the intron within the bacteriophage T4 td gene has permitted the detailed examination of mutant phenotypes by analyzing splicing in vivo and self-splicing in vitro. The genetic studies combined with phylogenetic analysis of intron structure based on comparative nucleotide sequence data [Cech 73 (1988) 259-271] and with biochemical data obtained from in vitro splicing experiments have resulted in significant advances in understanding the biology and chemistry of group-I introns.     

Improving the prediction of RNA secondary structure by detecting and assessing conserved stems

The prediction of RNA secondary structure can be facilitated by incorporating with comparative analysis of homologous sequences. However, most of existing comparative methods are vulnerable to alignment errors and thus are of low accuracy in practical application. Here we improve the prediction of RNA secondary structure by detecting and assessing conserved stems shared by all sequences in the alignment. Our method can be summarized by: 1) we detect possible stems in single RNA sequence using the so-called position matrix with which some possibly paired positions can be uncovered; 2) we detect conserved stems across multiple RNA sequences by multiplying the position matrices; 3) we assess the conserved stems using the Signal-to-Noise; 4) we compute the optimized secondary structure by incorporating the so-called reliable conserved stems with predictions by RNAalifold program. We tested our method on data sets of RNA alignments with known secondary structures. The accuracy, measured as sensitivity and specificity, of our method is greater than predictions by RNAalifold.