Evolution of the WANCY region in amniote mitochondrial DNA

In most vertebrate mitochondrial genomes, the site for initiation of light-strand replication, OL, is found within a cluster of five transfer RNA (tRNA) genes (tRNA(Trp), tRNA(Ala), tRNA(Asn), tRNA(Cys), and tRNA(Tyr)). This region and part of the adjacent cytochrome c oxydase subunit I (COI) gene were sequenced for two crocodilian, two turtle, and one snake species and for Sphenodon punctatus; part of the adjacent nicotinamide adenine dinucleotide dehydrogenase subunit 2 (ND2) gene was also sequenced for the crocodilian and turtle species. All had the typical vertebrate gene order. The turtles and the snake have a lengthy noncoding sequence between the tRNA(Asn) and tRNA(Cys) genes that we assumed to be homologous to the mammalian OL. The crocodilians and Sphenodon lack such a sequence, a condition they share with birds. Most proposed phylogenies for the amniotes require that OL at this position was lost at least twice during their diversification or was evolved independently more than once. Within the five tRNA genes, frequencies of substitutions are much higher in loops than in stems. Many loops vary dramatically in size among the species; in the most extreme case, the D-arm of the Sphenodon tRNA(Cys) is a "D-arm replacement" loop of seven nucleotides. Frequency of transitions in stems is relatively uniform across tRNAs, but frequency of transversions varies greatly. Mismatches in stems are infrequent, and their relative frequency in a specific tRNA is unrelated to the frequency of substitution in the corresponding gene. Several features of mammalian mitochondrial tRNAs are conserved in WANCY tRNAs throughout amniotes. The inferred initiation codon for COI is GTG in crocodilians, turtles, and the snake, a condition they share with fishes, certain amphibians, and birds. TTG appears to be the initiation codon for COI in Sphenodon; if correct, this would be a novel initiation codon for vertebrate mitochondrial DNA. Phylogenetic analyses of the inferred amino acid sequences of ND2 and COI support the sister-group relationship of birds and crocodilians and suggest that mammals are an early derived lineage within the amniotes.      

Aquifex aeolicus tRNA (Gm18) methyltransferase has unique substrate specificity. TRNA recognition mechanism of the enzyme

Transfer RNA (guanosine-2')-methyltransferase (Gm-methylase) catalyzes the transfer of a methyl group from S-adenosyl-l-methionine to 2'-OH of G18 in the D-loop of tRNA. Based on their mode of tRNA recognition, Gm-methylases can be divided into the following two types: type I having broad specificity toward the substrate tRNA, and type II that methylates only limited tRNA species. Protein synthesized by in vitro cell-free translation revealed that Gm-methylase encoded in the Aquifex aeolicus genome is a novel type II enzyme. Experiments with chimeric tRNAs and mini- and micro-helix RNAs showed that the recognition region of this enzyme is included within the D-arm structure of tRNALeu and that a bulge is essentially required. Variants of tRNALeu, tRNASer, and tRNAPhe revealed that a combination of certain base pairs in the D-stem is strongly recognized by the enzyme, that 4 bp in the D-stem enhance methyl acceptance activity, and that the Py16Py17G18G19 sequence is important for efficient methyl transfer. The methyl acceptance activities of all the A. aeolicus tRNA genes, which can be classified into 14 categories on the basis of their D-arm structure, were tested. The results clearly showed that the substrate recognition mechanism elucidated by the variant experiments was applicable to their native substrates.     

A protein extension to shorten RNA: elongated elongation factor-Tu recognizes the D-arm of T-armless tRNAs in nematode mitochondria

Nematode mitochondria possess extremely truncated tRNAs. Of 22 tRNAs, 20 lack the entire T-arm. The T-arm is necessary for the binding of canonical tRNAs and EF (elongation factor)-Tu (thermo-unstable). The nematode mitochondrial translation system employs two different EF-Tu factors named EF-Tu1 and EF-Tu2. Our previous study showed that nematode Caenorhabditis elegans EF-Tu1 binds specifically to T-armless tRNA. C. elegans EF-Tu1 has a 57-amino acid C-terminal extension that is absent from canonical EF-Tu, and the T-arm-binding residues of canonical EF-Tu are not conserved. In this study, the recognition mechanism of T-armless tRNA by EF-Tu1 was investigated. Both modification interference assays and primer extension analysis of cross-linked ternary complexes revealed that EF-Tu1 interacts not only with the tRNA acceptor stem but also with the D-arm. This is the first example of an EF-Tu recognizing the D-arm of a tRNA. The binding activity of EF-Tu1 was impaired by deletion of only 14 residues from the C-terminus, indicating that the C-terminus of EF-Tu1 is required for its binding to T-armless tRNA. These results suggest that C. elegans EF-Tu1 recognizes the D-arm instead of the T-arm by a mechanism involving its C-terminal region. This study sheds light on the co-evolution of RNA and RNA-binding proteins in nematode mitochondria.     

Mitochondrial tRNAs as light strand replication origins: Similarity between anticodon loops and the loop of the light strand replication origin predicts initiation of DNA replication

Stem-loop hairpins formed by mitochondrial light strand replication origins (OL) and by heavy strand DNA coding for tRNAs that form OL-like structures initiate mitochondrial replication. The loops are recognized by one of the two active sites of the vertebrate mitochondrial gamma polymerase, which are homologuous to the active sites of class II amino-acyl tRNA synthetases. Therefore, the polymerase site recognizing the OL loop could recognize tRNA anticodon loops and sequence similarity between anticodon and OL loops should predict initiation of DNA replication at tRNAs. Strengths of genome-wide deamination gradients starting at tRNA genes estimate extents by which replication starts at that tRNA. Deaminations (A→G and C→T) occur proportionally to time spent single stranded by heavy strand DNA during mitochondrial light strand replication. Results show that deamination gradients starting at tRNAs are proportional to sequence similarity between OL and tRNA loops: most for anticodon-, least D-, intermediate for TψC-loops, paralleling tRNA synthetase recognition interactions with these tRNA loops. Structural and sequence similarities with regular OLs predict OL function, loop similarity is dominant in most tRNAs. Analyses of sequence similarity and structure independently substantiate that DNA sequences coding for mitochondrial tRNAs sometimes function as alternative OLs. Pathogenic mutations in anticodon loops increase similarity with the human OL loop, non-pathogenic polymorphisms do not. Similarity/homology alignment hypotheses are experimentally testable in this system.     

Yeast mitochondrial methionine initiator tRNA: characterization and nucleotide sequence.

Two methionine tRNAs from yeast mitochondria have been purified. The mitochondrial initiator tRNA has been identified by formylation using a mitochondrial enzyme extract. E. coli transformylase however, does not formylate the yeast mitochondrial initiator tRNA. The sequence was determined using both 32P-in vivo labeled and 32P-end labeled mt tRNAf(Met). This tRNA, unlike N. crassa mitochondrial tRNAf(Met), has two structural features typical of procaryotic initiator tRNAs: (i) it lacks a Watson-Crick base-pair at the end of the acceptor stem and (ii) has a T-psi-C-A sequence in loop IV. However, both yeast and N. crassa mitochondrial initiator tRNAs have a U11:A24 base-pair in the D-stem unlike procaryotic initiator tRNAs which have A11:U24. Interestingly, both mitochondrial initiator tRNAs, as well as bean chloroplast tRNAf(Met), have only two G:C pairs next to the anticodon loop, unlike any other initiator tRNA whatever its origin. In terms of overall sequence homology, yeast mitochondrial tRNA(Met)f differs from both procaryotic or eucaryotic initiator tRNAs, showing the highest homology with N. crassa mitochondrial initiator tRNA.     

Mitochondrial import of only one of three nuclear-encoded glutamine tRNAs in Tetrahymena thermophila.

The mitochondrial genome of Tetrahymena does not appear to encode enough tRNAs to perform mitochondrial protein synthesis. It has therefore been proposed that nuclear-encoded tRNAs are imported into the mitochondria. T.thermophila has three major glutamine tRNAs: tRNA(Gln)(UUG), tRNA(Gln)(UUA) and tRNA(Gln)(CUA). Each of these tRNAs functions in cytosolic translation. However, due to differences between the Tetrahymena nuclear and mitochondrial genetic codes, only tRNA(Gln)(UUG) has the capacity to function in mitochondrial translation as well. Here we show that approximately 10-20% of the cellular complement of tRNA(Gln)(UUG) is present in mitochondrial RNA fractions, compared with 1% or less for the other two glutamine tRNAs. Furthermore, this glutamine tRNA is encoded only by a family of nuclear genes, the sequences of several of which are presented. Finally, when marked versions of tRNA(Gln)(UUG) and tRNA(Gln)(UUA) flanked by identical sequences are expressed in the macronucleus, only the former undergoes mitochondrial import; thus sequences within tRNA(Gln)(UUG) direct import. Because tRNA(Gln)(UUG) is a constituent of mitochondrial RNA fractions and is encoded only by nuclear genes, and because ectopically expressed tRNA(Gln)(UUG) fractionates with mitochondria like its endogenous counterpart, we conclude that it is an imported tRNA in T.thermophila.     

Intricacies and surprises of nuclear-mitochondrial co-evolution

In this issue of the Biochemical Journal, Watanabe and colleagues disclose another fascinating facet of the mitochondrial protein synthesis machinery: one of the two nematode mitochondrial elongation factors Tu, EF-Tu1, specifically recognizes the D-arm of T-armless tRNAs via a 57-amino-acid C-terminal extension that compensates for the reduction in tRNA structure. This principle provides a paradigm for the evolutionary events thought to have ignited the transition from an ancient 'RNA world' to the 'protein world' of today.     

Mechanisms of molecular recognition of tRNAs by aminoacyl-tRNA synthetases.

Interactions of Escherichia coli isoleucyl- and glutamyl-tRNA synthetases and their cognate tRNAs were analyzed by phosphate-alkylation mapping with N-nitroso-N-ethylurea and/or by 1H-NMR analysis. When E. coli tRNA(Ile) was bound with isoleucyl-tRNA synthetase, many of the phosphate groups in the anticodon loop and stem and in the D-stem were protected from alkylation. This result is consistent with that of analysis of imino proton resonances due to the secondary and tertiary base pairs. These analyses also suggested that the L-shaped tertiary structure of tRNA(Ile) is distorted upon complex formation with IleRS because of disruption of some tertiary base pairs. In the case of E. coli tRNA(Glu), several phosphate groups in the D-stem and the variable loop were significantly protected by the cognate synthetase. These results indicate that the two tRNAs, unlike other tRNAs studied so far, have some of the "identity determinants" in the D-stem and/or in the anticodon stem.     

Structural compensation in an archaeal selenocysteine transfer RNA.

A new type of structural compensation between the lengths of two perpendicularly oriented RNA double helices was found in the archaeal selenocysteine tRNA from Methanococcus jannascii. This tRNA contains only four base-pairs in the T-stem, one base-pair less than in all other cytosolic tRNAs. Our analysis shows that such a T-stem in an otherwise normal tRNA cannot guarantee the formation of the normal interactions between the D and T-loops. The absence of these interactions would affect the juxtaposition of the two tRNA helical domains, potentially damaging the tRNA function. In addition to the short T-stem, this tRNA possesses another unprecedented feature, a very long D-stem consisting of seven base-pairs. Taken as such, a seven base-pair D-stem will also disrupt the normal interaction between the D and T-loops. On the other hand, the presence of the universal nucleotides in both the D and T-loops suggests that these loops probably interact with each other in the same way as in other tRNAs. Here, we demonstrate that the short T-stem and the long D-stem can naturally compensate each other, thus providing the normal D/T interactions. Molecular modeling has helped suggest a detailed scheme of mutual compensation between these two unique structural aspects of the archaeal selenocysteine tRNA. In the light of this analysis, other structural and functional characteristics of the selenocysteine tRNAs are discussed.     

Molecular Phylogenetics, tRNA Evolution, and Historical Biogeography in Anguid Lizards and Related Taxonomic Families

Phylogenetic relationships among lizards of the families Anguidae, Anniellidae, Xenosauridae, and Shinisauridae are investigated using 2001 aligned bases of mitochondrial DNA sequence from the genes encoding ND1 (subunit one of NADH dehydrogenase), tRNA(Ile), tRNA(Gln), tRNA(Met), ND2, tRNA(Trp), tRNA(Ala), tRNA(Asn), tRNA(Cys), tRNA(Tyr), and COI (subunit I of cytochrome c oxidase), plus the origin for light-strand replication (O(L)) between the tRNA(Asn) and the tRNA(Cys) genes. The aligned sequences contain 1013 phylogenetically informative characters. A well-resolved phylogenetic hypothesis is obtained. Because monophyly of the family Xenosauridae (Shinisaurus and Xenosaurus) is statistically rejected, we recommend placing Shinisaurus in a separate family, the Shinisauridae. The family Anniellidae and the anguid subfamilies Gerrhonotinae and Anguinae each form monophyletic groups receiving statistical support. The Diploglossinae*, which appears monophyletic, is retained as a metataxon (denoted with an asterisk) because its monophyly is statistically neither supported nor rejected. The family Anguidae appears monophyletic in analyses of the DNA sequence data, and statistical support for its monophyly is provided by reanalysis of previously published allozymic data. Anguid lizards appear to have had a northern origin in Laurasia. Taxa currently located on Gondwanan plates arrived there by dispersal from the north in two separate events, one from the West Indies to South America and another from a Laurasian plate to Morocco. Because basal anguine lineages are located in western Eurasia and Morocco, formation of the Atlantic Ocean (late Eocene) is implicated in the separation of the Anguinae from its North American sister taxon, the Gerrhonotinae. Subsequent dispersal of anguine lizards to East Asia and North America appears to have followed the Oligocene drying of the Turgai Sea. The alternative hypothesis, that anguine lizards originated in North America and dispersed to Asia via the Bering land bridge with subsequent colonization of Europe and Morocco, requires a phylogenetic tree seven steps longer than the most parsimonious hypothesis. North African, European, and West Asian anguines were isolated from others by the rapid uplift of Tibet in the late Oligocene to Miocene. Phylogenetic analysis of evolutionary changes in the gene encoding tRNA(Cys) suggests gradual reduction of dihydrouridine (D) stems by successive deletion of bases in some lineages. This evolutionary pattern contrasts with the one observed for parallel elimination of the D-stem in mitochondrial tRNAs of eight other reptile groups, in which replication slippage produces direct repeats. An unusual, enlarged TpsiC (T) stem is inferred for tRNA(Cys) in most species.     

The relation between hairpin formation by mitochondrial WANCY tRNAs and the occurrence of the light strand replication origin in Lepidosauria

Mitochondrial light strand DNA replication is initiated at light strand replication origins (OLs), short stem-loop hairpins formed by the heavy strand DNA. OL-like secondary structures are also formed by heavy strand DNA templating for the five tRNAs adjacent to OLs, the WANCY tRNA cluster. We tested whether natural OL absence associates with greater capacities for formation of OL-like structures by WANCY tRNA genes. Using lepidosaurian taxa (Sphenodon, lizards and amphisbaenids), we compared WANCY tRNA capacities to form OL-like structures between 248 taxa possessing an OL with 131 taxa without OL (from different families). On average, WANCY tRNA genes form more OL-like structures in the absence of a regular OL than in its presence. Formation of OL-like structures by WANCY tRNAs follows hierarchical patterns that may reduce competition between the tRNA's translational function and its secondary OL function: the rarer the tRNA's cognate amino acid, the greater the capacity to form OL-like structures. High OL-forming capacities for neighboring tRNAs are avoided. Because OL absence usually occurs in taxa with reduced genomes, increased formation of OL-like structures by WANCY tRNAs might result from selection for greater metabolic efficiency. Further analyses suggest that OL loss is one of the latest steps in genome reduction, and promotes the increase in formation of OL-like structures by WANCY tRNA genes in Lepidosauria.     

Transfer RNA genes in the cap-oxil region of yeast mitochondrial DNA.

A cytoplasmic "petite" (rho-) clone of Saccharomyces cerevisiae has been isolated and found through DNA sequencing to contain the genes for cysteine, histidine, leucine, glutamine, lysine, arginine, and glycine tRNAs. This clone, designated DS502, has a tandemly repeated 3.5 kb segment of the wild type genome from 0.7 to 5.6 units. All the tRNA genes are transcribed from the same strand of DNA in the direction cap to oxil. The mitochondrial DNA segment of DS502 fills a sequence gap that existed between the histidine and lysine tRNAs. The new sequence data has made it possible to assign accurate map positions to all the tRNA genes in the cap-oxil span of the yeast mitochondrial genome. A detailed restriction map of the region from 0 to 17 map units along with the locations of 16 tRNA genes have been determined. The secondary structures of the leucine and glutamine tRNAs have been deduced from their gene sequences. The leucine tRNA exhibits 64% sequence homology to an E. coli leucine tRNA.     

A correlation between N2-dimethylguanosine presence and alternate tRNA conformers.

Even though the evolutionary conservation of the cloverleaf model is strongly suggestive of powerful constraints on the secondary structure of functional tRNAs, some mitochondrial tRNAs cannot be folded into this form. From the optimal base pairing pattern of these recalcitrant tRNAs, structural correlations between the length of the anticodon stem and the lengths of connector regions between the two helical domains, formed by the coaxial stacking of the anticodon and D-stems and the acceptor and T-stems, have been derived and used to scan the tRNA and tRNA gene database. We show here that some cytosolic tRNA gene sequences that are compatible with the cloverleaf model can also be folded into patterns proposed for the unusual mitochondrial tRNAs. Furthermore, the ability to be folded into these atypical structures correlates in the mature RNA sequences with the presence of dimethylguanosine, whose role may be to prevent the unusual mitochondrial tRNA pattern folding.     

Improved systematic tRNA gene annotation allows new insights into the evolution of mitochondrial tRNA structures and into the mechanisms of mitochondrial genome rearrangements

Transfer RNAs (tRNAs) are present in all types of cells as well as in organelles. tRNAs of animal mitochondria show a low level of primary sequence conservation and exhibit 'bizarre' secondary structures, lacking complete domains of the common cloverleaf. Such sequences are hard to detect and hence frequently missed in computational analyses and mitochondrial genome annotation. Here, we introduce an automatic annotation procedure for mitochondrial tRNA genes in Metazoa based on sequence and structural information in manually curated covariance models. The method, applied to re-annotate 1876 available metazoan mitochondrial RefSeq genomes, allows to distinguish between remaining functional genes and degrading 'pseudogenes', even at early stages of divergence. The subsequent analysis of a comprehensive set of mitochondrial tRNA genes gives new insights into the evolution of structures of mitochondrial tRNA sequences as well as into the mechanisms of genome rearrangements. We find frequent losses of tRNA genes concentrated in basal Metazoa, frequent independent losses of individual parts of tRNA genes, particularly in Arthropoda, and wide-spread conserved overlaps of tRNAs in opposite reading direction. Direct evidence for several recent Tandem Duplication-Random Loss events is gained, demonstrating that this mechanism has an impact on the appearance of new mitochondrial gene orders.     

Methylation of an adenosine in the D-loop of specific transfer RNAs from yeast by a procaryotic tRNA (adenine-1) methyltransferase.

tRNA (adenine-1) methyltransferase occurs in Bacillus subtilis. Eucaryotic tRNAThr and tRNATyr from yeast in which 1-methyladenosine (m1A) is already present in the TpsiC loop, can be methylated in vitro with S-adenosylmethionine and B. subtilis extracts. Each of the specific tRNAs accepts 1 mol of methyl groups per mol tRNA. The enzyme transforms into m1A the 3'-terminal adenylic acid residue of the dihydrouridine loop, a new position for a modified adenosine residue in tRNA. Both tRNAs have the sequence Py-A-A-G-G-C-m2(2)G in the D-loop and D-stem region. Other tRNAs with the same sequence in this region also serve as substrates for the tRNA (adenine-1) methyltransferase.