Group II introns encode proteins with reverse transcriptase activity. These proteins also promote RNA splicing (maturase activity) and then, with the excised intron, form a site-specific DNA endonuclease that promotes intron mobility by reverse splicing into DNA followed by target DNA-primed reverse transcription. Here, we used an Escherichia coli expression system for the Lactococcus lactis group II intron Ll.LtrB to show that the intron-encoded protein (LtrA) alone is sufficient for maturase activity, and that RNP particles containing only the LtrA protein and excised intron RNA have site-specific DNA endonuclease and target DNA-primed reverse transcriptase activity. Detailed analysis of the splicing reaction indicates that LtrA is an intron-specific splicing factor that binds to unspliced precursor RNA with a K(d) of

The retrohoming of the yeast mtDNA intron aI1 occurs by a target DNA-primed reverse transcription (TPRT) mechanism in which the intron RNA reverse splices directly into the recipient DNA and is then copied by the intron-encoded reverse transcriptase. Here, we carried out biochemical characterization of the intron-encoded reverse transcriptase and site-specific DNA endonuclease activities required for this process. We show that the aI1 reverse transcriptase has high TPRT activity in the presence of appropriate DNA target sites, but differs from the closely related reverse transcriptase encoded by the yeast aI2 intron in being unable to use artificial substrates efficiently. Characterization of TPRT products shows that the fully reverse spliced intron RNA is an efficient template for cDNA synthesis, while reverse transcription of partially reverse spliced intron RNA is impeded by the branch point. Novel features of the aI1 reaction include a prominent open-circular product in which cDNAs are incorporated at a nick at the antisense-strand cleavage site. The aI1 endonuclease activity, which catalyzes the DNA cleavage and reverse splicing reactions, is associated with ribonucleoprotein particles containing the intron-encoded protein and the excised intron RNA. As shown for the aI2 endonuclease, both the RNA and protein components are used for DNA target site recognition, but the aI1 protein has less stringent nucleotide sequence requirements for the reverse splicing reaction. Finally, perhaps reflecting this relaxed target specificity, in vitro experiments show that aI1 can reverse splice directly into ectopic mtDNA transposition sites, consistent with the previously suggested possibility that this mechanism is used for ectopic transposition of group II introns in vivo.

The mobile group II intron of Lactococcus lactis, Ll.LtrB, provides the opportunity to analyze the homing pathway in genetically tractable bacterial systems. Here, we show that Ll.LtrB mobility occurs by an RNA-based retrohoming mechanism in both Escherichia coli and L. lactis. Surprisingly, retrohoming occurs efficiently in the absence of RecA function, with a relaxed requirement for flanking exon homology and without coconversion of exon markers. These results lead to a model for bacterial retrohoming in which the intron integrates into recipient DNA by complete reverse splicing and serves as the template for cDNA synthesis. The retrohoming reaction is completed in unprecedented fashion by a DNA repair event that is independent of homologous recombination between the alleles. Thus, Ll.LtrB has many features of retrotransposons, with practical and evolutionary implications.

The Lactococcus lactis group II intron Ll.ltrB is similar to mobile yeast mtDNA group II introns, which encode reverse transcriptase, RNA maturase, and DNA endonuclease activities for site-specific DNA insertion. Here, we show that the Lactococcal intron can be expressed and spliced efficiently in Escherichia coli. The intron-encoded protein LtrA has reverse transcriptase and RNA maturase activities, with the latter activity shown both in vivo and in vitro, a first for any group II intron-encoded protein. As for the yeast mtDNA introns, the DNA endonuclease activity of the Lactococcal intron is associated with RNP particles containing both the intron-encoded protein and the excised intron RNA. Also, the intron RNA cleaves the sense-strand of the recipient DNA by a reverse splicing reaction, whereas the intron-encoded protein cleaves the antisense strand. The Lactococcal intron endonuclease can be obtained in large quantities by coexpression of the LtrA protein with the intron RNA in E. coli or reconstituted in vitro by incubating the expressed LtrA protein with in vitro-synthesized intron RNA. Furthermore, the specificity of the endonuclease and reverse splicing reactions can be changed predictably by modifying the RNA component. Expression in E. coli facilitates the use of group II introns for the targeting of specific foreign sequences to a desired site in DNA.

The Neurospora crassa mitochondrial tyrosyl-tRNA synthetase (CYT-18 protein) functions in splicing the N. crassa mitochondrial large rRNA intron by stabilizing the catalytically active structure of the intron core. Here, a comprehensive study of N. crassa mtDNA group I introns identified two additional introns, cob-I2 and the ND1 intron, that are dependent on CYT-18 for splicing in vitro and in vivo. The other seven N. crassa mtDNA group I introns are not CYT-18-dependent and include five that self-splice and two that do not splice under any conditions examined. Some of these introns may require maturases or other proteins for efficient splicing. All but one of the non-CYT-18-dependent introns contain large peripheral extensions of the P5 stem, related to the P5abc structure that blocks CYT-18 binding to the Tetrahymena large rRNA intron. The remaining non-CYT-18-dependent intron, cob-I1, contains a long, peripheral extension of the P9 stem, denoted P9.1, which also impedes CYT-18 binding. Detailed analysis of the CYT-18-dependent ND1 intron showed that two 3' splice sites are used in vitro and in vivo. The proximal, alternative 3' splice site brings the intron open reading frame, which potentially encodes a mobility endonuclease, in frame with the upstream exon, possibly providing a means of expression. Considered together, our results show that group I introns in N. crassa mitochondria use a variety of strategies involving different proteins and/or RNA structures to assist splicing, and they support the hypothesis that CYT-18 and the peripheral RNA structure P5abc are alternative evolutionary adaptations for stabilizing the active structure of the intron core.

The Mauriceville mitochondrial retroplasmid of Neurospora encodes a novel reverse transcriptase that initiates cDNA synthesis at a 3' tRNA-like structure of the plasmid transcript, either de novo (i.e. without a primer) or by using the 3' OH group of a DNA primer. Both the de novo and primer-mediated initiations involve recognition of structural features at the 3' end of the retroplasmid transcript, which ends with a 3' CCACCA. Here, detailed biochemical characterization of the retroplasmid reverse transcriptase shows that the 3' CCA of the plasmid transcript is the major structural feature recognized by the reverse transcriptase for both the de novo and primer-mediated initiations. Complementarity between the DNA primer and RNA template is not required for the primer-mediated initiation, although short (1 to 3 nt) base-pairing interactions can influence both the efficiency and site of initiation near the 3' end of the transcript. Single nucleotide changes in the 3' CCA lead to less efficient initiation in the upstream CCA with an increased propensity to add extra "non-coded" nucleotides to the 5' end of the cDNA during de novo initiation or to the 3' end of the primer during primer-mediated initiation. Secondary structure features upstream of the 3' CCA also influence the efficiency of initiation, but are not stringently required in vitro. Finally, we find that the retroplasmid reverse transcriptase does not efficiently use DNA primers that are base-paired to internal positions in the RNA template, nor does it use analogs of natural substrates used by non-long terminal repeat retrotransposon or retroviral reverse transcriptases. Our results indicate that the retroplasmid reverse transcriptase is uniquely adapted to initiate cDNA synthesis by recognizing a 3' CCA sequence. The ability to recognize a specific template sequence is common for RNA polymerases, but unprecedented for a reverse transcriptase.

The Mauriceville retroplasmid of Neurospora mitochondria encodes a novel reverse transcriptase that initiates cDNA synthesis de novo (i.e., without a primer) at the 3' CCA of the plasmid transcript's 3' tRNA-like structure (H. Wang and A. M. Lambowitz, Cell 75:1071-1081, 1993). Here, we show that the plasmid reverse transcriptase also initiates cDNA synthesis de novo at the 3' end of tRNAs, leading to synthesis of a full-length cDNA copy of the tRNA. The use of tRNA templates in vivo was suggested previously by the structure of suppressive mutant plasmids that have incorporated mitochondrial tRNA sequences (R. A. Akins, R. L. Kelley, and A. M. Lambowitz, Cell 47:505-516, 1986). The in vitro experiments show that efficient de novo initiation on tRNA templates requires an unpaired 3' CCA and occurs predominantly opposite position C-2 of the 3' CCA sequence, the same position as in the plasmid transcript. In other reactions, the plasmid reverse transcriptase synthesizes cDNA dimers by template switching between two tRNA templates and initiates at an internal position in a tRNA by using the 3' end of the tRNA as a primer. Finally, we show that template switching between the tRNA and the plasmid transcript in vitro gives rise to hybrid cDNAs of the type predicted to be intermediates in the generation of the suppressive mutant plasmids. The ability of the plasmid reverse transcriptase to initiate at the 3' end of tRNAs presumably reflects the recognition of structural features similar to those of the 3' tRNA-like structure of the plasmid transcript. The recognition of tRNAs or tRNA-like structures as templates for cDNA synthesis may be characteristic of primitive reverse transcriptases that evolved from RNA-dependent RNA polymerases.

The mobile group II introns aI1 and aI2 of yeast mtDNA encode endonuclease activities that cleave intronless DNA target sites to initiate mobility by target DNA-primed reverse transcription. For aI2, sense-strand cleavage occurs mainly by a partial reverse splicing reaction, whereas for aI1, complete reverse splicing occurs, leading to insertion of the linear intron RNA into double-stranded DNA. Here, we show that aI1 homing and reverse splicing depend on the EBS1 (RNA)/IBS1(DNA) pairing and that target specificity can be changed by compensatory changes in the target site and the donor intron. Using well-marked strains to follow coconversion of flanking DNA, we show that homing occurs by both RT-dependent and -independent pathways. Remarkably, in most RT-dependent events, the reverse spliced intron is the initial template for first-strand cDNA synthesis.

Group II introns use intron-encoded reverse transcriptase, maturase and DNA endonuclease activities for site-specific insertion into DNA. Remarkably, the endonucleases are ribonucleoprotein complexes in which the excised intron RNA cleaves the sense strand of the recipient DNA by reverse splicing, while the intron-encoded protein cleaves the antisense strand. Here, studies with the yeast group II intron aI2 indicate that both the RNA and protein components of the endonuclease contribute to recognition of an approximately 30 bp DNA target site. Our results lead to a model in which the protein component first recognizes specific nucleotides in the most distal 5' exon region of the DNA target site (E2-21 to -11). Binding of the protein then leads to DNA unwinding, enabling the intron RNA to base pair to a 13 nucleotide DNA sequence (E2-12 to E3+1) for reverse splicing. Antisense-strand cleavage requires additional interactions of the protein with the 3' exon DNA (E3+1 to +10). Our results show how enzymes can use RNA and protein subunits cooperatively to recognize specific sequences in double-stranded DNA.

The Neurospora crassa mitochondrial tyrosyl-tRNA synthetase (CYT-18 protein) functions in splicing group I introns by promoting the formation of the catalytically active structure of the intron RNA. Previous studies showed that CYT-18 binds with high affinity to the P4-P6 domain of the catalytic core and that there is some additional contribution to binding from the P3-P9 domain. Here, quantitative binding assays with deletion derivatives of the N. crassa mitochondrial large rRNA intron showed that at least 70% of the binding energy can be accounted for by the interaction of CYT-18 with the P4-P6 domain. Within this domain, P4 and P6 are required for high affinity CYT-18 binding, while the distal elements P5 and P6a may contribute indirectly by stabilizing the correct structure of the binding site in P4 and P6. CYT-18 binds to a small RNA corresponding to the isolated P4-P6 domain, but not to a permuted version of this RNA in which P4-P6 is a continuous rather than a stacked helix. Iterative in vitro selection experiments with the isolated P4-P6 domain showed a requirement for base-pairing to maintain helices P4, P6 and P6a, but indicate that P5 is subject to fewer constraints. The most strongly conserved nucleotides in the selections were clustered around the junction of the P4-P6 stacked helix, with ten nucleotides (J3/4-2,3, P4 bp -1 and 3, and P6 bp -1 and 2) found invariant in the context of the wild-type RNA structure. In vitro mutagenesis confirmed that replacement of the wild-type nucleotides at J3/4-2 and 3 or P4 bp-3 markedly decreased CYT-18 binding, reflecting either base specific contacts or indirect readout of RNA structure by the protein. Our results suggest that a major function of CYT-18 is to promote assembly of the P4-P6 domain by stabilizing the correct geometry at the junction of the P4-P6 stacked helix. The relatively large number of conserved nucleotides at the binding site suggests that the interaction of CYT-18 with group I introns is unlikely to have arisen by chance and could reflect either an evolutionary relationship between group I introns and tRNAs or interaction with a common stacked-helical structural motif that evolved separately in these RNAs.

Some group II introns are mobile elements as well as catalytic RNAs. Introns aI1 and aI2 found in the gene COX1 in yeast mitochondria encode reverse transcriptases which promote site-specific insertion of the intron into intronless alleles ('homing'). For aI2 this predominantly occurs by reverse transcription of unspliced precursor RNA at a break in double-strand DNA made by an endonuclease encoded by the intron. The aI2 endonuclease involves both the excised intron RNA, which cleaves the DNA's sense strand by partial reverse splicing; and the intron-encoded reverse transcriptase which cleaves the anti-sense strand. Here we show that aI1 encodes an analogous endonuclease specific for a different target site compatible with the different exon-binding sequences of the intron RNA. Over half of aI1 undergoes complete reverse splicing in vitro, thus integrating linear intron RNA directly into the DNA. This unprecedented reaction has implications for both intron mobility and evolution, and potential genetic engineering applications.

The Neurospora crassa mitochondrial tyrosyl-tRNA synthetase (CYT-18 protein) functions in splicing group I introns. We have used chemical-structure mapping and footprinting to investigate the interaction of the CYT-18 protein with the N. crassa mitochondrial large subunit ribosomal RNA (mt LSU) and ND1 introns, which are not detectably self-splicing in vitro. Our results show that both these non-self-splicing introns form most of the short range pairings of the conserved group I intron secondary structure in the absence of CYT-18, but otherwise remain largely unfolded, even at high Mg2+ concentrations. The binding of CYT-18 promotes the formation of the extended helical domains P6a-P6-P4-P5 (P4-P6 domain) and P8-P3-P7-P9 (P3-P9 domain) and their interaction to form the catalytic core. In iodine-footprinting experiments, CYT-18 binding results in the protection of regions of the phosphodiester backbone expected for tertiary folding of the catalytic core, as well as additional protections that may reflect proximity of the protein. In both introns, most of the putative CYT-18 protection sites are in the P4-P6 domain, the region of the SU intron previously shown to bind CYT-18 as a separate RNA molecule, but additional sites are found in the other major helical domain in P8 and P9 in both introns and in L9 and P7.1/P7.1a in the mt LSU intron. Protease digestion of the CYT-18/intron RNA complexes results in the loss of CYT-18-induced RNA tertiary structure and splicing activity. Considered together with previous studies, or results suggest that CYT-18 binds initially to the P4-P6 region of group I introns to form a scaffold for the assembly of the P3-P9 domain, which may contain additional binding sites for the protein. A three-dimensional model structure of the CYT-18 binding site in group I introns indicates that CYT-18 interacts with the surface of the catalytic core on the side opposite the active-site cleft and may primarily recognize a specific three-dimensional geometry of the phosphodiester backbone of group I introns.

The Neurospora crassa mitochondrial (mt) tyrosyl-tRNA synthetase (CYT-18 protein) functions in splicing group I introns, in addition to aminoacylating tRNA(Tyr). Here, we compared the CYT-18 binding sites in the N. crassa mt LSU and ND1 introns with that in N. crassa mt tRNA(Tyr) by constructing three-dimensional models based on chemical modification and RNA footprinting data. Remarkably, superimposition of the CYT-18 binding sites in the model structures revealed an extended three-dimensional overlap between the tRNA and the group I intron catalytic core. Our results provide insight into how an RNA-splicing factor can evolve from a cellular RNA-binding protein. Further, the structural similarities between group I introns and tRNAs are consistent with an evolutionary relationship and suggest a general mechanism for the evolution of complex catalytic RNAs.

The Neurospora crassa mitochondrial tyrosyl-tRNA synthetase, the CYT-18 protein, functions in splicing group I introns by promoting the formation of the catalytically active structure of the intron RNA. The group I intron catalytic core is thought to consist of two extended helical domains, one formed by coaxial stacking of P5, P4, P6, and P6a (P4-P6 domain) and the other consisting of P8, P3, P7, and P9 (P3-P9 domain). To investigate how CYT-18 stabilizes the active RNA structure, we used an Escherichia coli genetic assay based on the phage T4 td intron to systematically test the ability of CYT-18 to compensate for structural defects in three key regions of the catalytic core: J3/4 and J6/7, connecting regions that form parts of the triple-helical-scaffold structure with the P4-P6 domain, and P7, a long-range base-pairing interaction that forms the guanosine-binding site and is part of the P3-P9 domain. Our results show that CYT-18 can suppress numerous mutations that disrupt the J3/4 and J6/7 nucleotide-triple interactions, as well as mutations that disrupt base-pairing in P7. CYT-18 suppressed mutations of phylogenetically conserved nucleotide residues at all positions tested, except for the universally conserved G-residue at the guanosine-binding site. Structure mapping experiments with selected mutant introns showed that the CYT-18-suppressible J3/4 mutations primarily impaired folding of the P4-P6 domain, while the J6/7 mutations impaired folding of both the P4-P6 and P3-P9 domains to various degrees. The P7 mutations impaired the formation of both P7 and P3, thereby grossly disrupting the P3-P9 domain. The finding that the P7 mutations also impaired formation of P3 provides evidence that the formation of these two long-range pairings is interdependent in the td intron. Considered together with previous work, the nature of mutations suppressed by CYT-18 supports a model in which CYT-18 helps assemble the P4-P6 domain and then stabilizes the two major helical domains of the catalytic core in the correct relative orientation to form the intron's active site.

Mobile group II introns encode reverse transcriptases and insert site specifically into intronless alleles (homing). Here, in vitro experiments show that homing of the yeast mtDNA group II intron aI2 occurs by reverse transcription at a double-strand break in the recipient DNA. A site-specific endonuclease cleaves the antisense strand of recipient DNA at position +10 of exon 3 and the sense strand at the intron insertion site. Reverse transcription of aI2-containing pre-mRNA is primed by the antisense strand cleaved in exon 3 and results in cotransfer of the intron and flanking exon sequences. Remarkably, the DNA endonuclease that initiates homing requires both the aI2 reverse transcriptase protein and aI2 RNA. Parallels in their reverse transcription mechanisms raise the possibility that mobile group II introns were ancestors of nuclear non-long terminal repeat retrotransposons and telomerases.

The mobility (homing) of the yeast mitochondrial DNA group II intron al2 occurs via target DNA-primed reverse transcription at a double-strand break in the recipient DNA. Here, we show that the site-specific DNA endonuclease that makes the double-strand break is a ribonucleoprotein complex containing the al2-encoded reverse transcriptase protein and excised al2 RNA. Remarkably, the al2 RNA catalyzes cleavage of the sense strand of the recipient DNA, while the al2 protein appears to cleave the antisense strand. The RNA-catalyzed sense strand cleavage occurs via a partial reverse splicing reaction in which the protein component stabilizes the active intron structure and appears to confer preference for DNA substrates. Our results demonstrate a biologically relevant ribozyme reaction with a substrate other than RNA.

The Neurospora CYT-18 protein, the mitochondrial tyrosyl-tRNA synthetase, functions in the splicing of group I introns. Here, bacterially expressed CYT-18 protein, purified by a new procedure involving polyethyleneimine precipitation to remove tightly bound nucleic acids, was used to characterize properties pertinent to RNA splicing. Analytical ultracentrifugation and other methods showed that the CYT-18 protein is an asymmetric homodimer. The measured frictional ratio, f/fo = 1.55, corresponds to an axial ratio of 10 for a prolate ellipsoid or 12 for an oblate ellipsoid. Like bacterial TyrRSs, the CYT-18 protein exhibits half-sites reactivity, each homodimer having one active site for tyrosyl adenylation and RNA splicing. The splicing activity of CYT-18 was unaffected by aminoacylation substrates at concentrations used in aminoacylation reactions, whereas the TyrRS activity was inhibited by physiological concentrations of the splicing cofactor GTP, as well as CTP or UTP, or by low concentrations of a group I intron RNA. Kinetic measurements suggest that the binding of CYT-18 to a group I intron substrate is a two-step process, with an initial biomolecular step that is close to diffusion limited (3.24 +/- 0.03 x 10(7) M-1s-1) followed by a slower conformational change (0.54 +/- 0.07 s-1). After CYT-18 binding, splicing occurs at a rate of 0.0025 s-1, within 6-fold of the rate of self-splicing of the Tetrahymena large rRNA intron in vitro. The Kd for the complex between the CYT-18 protein and a group I intron substrate, calculated from koff/kon, was < 0.3 pM, substantially lower than determined by presumed equilibrium measurements [Guo, Q., & Lambowitz, A. M. (1992) Genes Dev. 6, 1357-1372]. As a result of this tight binding, the CYT-18 protein functions stoichiometrically in in vitro splicing reactions due to its extremely slow dissociation from the excised intron RNA. The very tight binding of the CYT-18 protein to the intron RNA raises the possibility that specific mechanisms exist for dissociating the protein from the excised intron in vivo.

Group II introns aI1 and aI2 of the yeast mitochondrial COXI gene are mobile elements that encode an intron-specific reverse transcriptase (RT) activity. We show here that the introns of Saccharomyces cerevisiae ID41-6/161 insert site specifically into intronless alleles. The mobility is accompanied by efficient, but highly asymmetric, coconversion of nearby flanking exon sequences. Analysis of mutants shows that the aI2 protein is required for the mobility of both aI1 and aI2. Efficient mobility is dependent on both the RT activity of the aI2-encoded protein and a separate function, a putative DNA endonuclease, that is associated with the Zn2+ finger-like region of the intron reading frame. Surprisingly, there appear to be two mobility modes: the major one involves cDNAs reverse transcribed from unspliced precursor RNA; the minor one, observed in two mutants lacking detectable RT activity, appears to involve DNA level recombination. A cis-dominant splicing-defective mutant of aI2 continues to synthesize cDNAs containing the introns but is completely defective in both mobility modes, indicating that the splicing or the structure of the intron is required. Our results demonstrate that the yeast group II intron aI2 is a retroelement that uses novel mobility mechanisms.

The mitochondria of certain natural isolates of Neurospora contain both the Varkud plasmid, which encodes a reverse transcriptase, and a small unrelated RNA (VS RNA) that performs RNA-mediated self-cleavage and ligation reactions. Here, we show that VS RNA is transcribed from a VS plasmid DNA template by the Neurospora mitochondrial RNA polymerase using a promoter located immediately upstream of the RNA self-cleavage site that generates monomeric transcripts. VS RNA is then reverse transcribed by the Varkud plasmid reverse transcriptase to yield a full-length (-) strand cDNA, a predicted replication intermediate. Combined with previous genetic evidence, our results indicate that the VS plasmid replicates by reverse transcription as a satellite of the Varkud plasmid. This mode of replication, unprecedented for a satellite RNA, likely reflects the promiscuity of the Varkud plasmid reverse transcriptase, which does not require a specific primer to initiate cDNA synthesis. Our findings indicate how primitive reverse transcriptases with similar relaxed specificity could have facilitated the evolution of new retroelements.

The Mauriceville plasmid and the closely related Varkud plasmid of Neurospora spp. are retroelements that propagate in mitochondria. Replication appears to occur by a novel mechanism in which a monomer-length plasmid transcript having a 3' tRNA-like structure ending in CCA is reverse transcribed to give a full-length minus-strand cDNA beginning at or near the 3' end of the RNA. Here, we show that the plasmids are transcribed in vitro by the Neurospora mitochondrial RNA polymerase, with the major in vitro transcription start site approximately 260 bp upstream of the 5' end of the plasmid transcript. The location of the transcription start site suggests that the monomer-length transcripts are generated by transcription around the plasmid combined with a site-specific RNA cleavage after the 3'-CCA sequence. The 5' ends of minus-strand cDNAs in ribonucleoprotein particles were analyzed to obtain insight into the mechanism of initiation of reverse transcription in vivo. A major class of minus-strand cDNAs begins opposite C2 of the 3'-CCA sequence, the same site used for de novo initiation of cDNA synthesis by the plasmid reverse transcriptase in vitro. A second class of minus-strand cDNAs begins with putative primer sequences that correspond to cDNA copies of the plasmid or mitochondrial transcripts. These findings are consistent with the possibility that the plasmid reverse transcriptase initiates minus-strand cDNA synthesis in vivo both by de novo initiation and by a novel template-switching mechanism in which the 3' OH of a previously synthesized cDNA is used to prime the synthesis of a new minus-strand cDNA directly at the 3' end of the plasmid transcript.