The complete mitochondrial genome of the mackerel icefish,Champsocephalus gunnari (Actinopterygii: Channichthyidae), with reference to the evolution of mitochondrial genomes in Antarctic notothenioids

The mackerel icefish (Champsocephalus gunnari Lönnberg, 1905) is a ray‐finned fish living in the Southern Ocean around Antarctica. We sequenced the complete mitochondrial (mt) genome of the mackerel icefish and a segment from cytochrome b to the control region (CR) in 32 individuals. The mt genome of the mackerel icefish was rearranged, containing two nicotinamide adenine dinucleotide (reduced form) dehydrogenase subunit 6 (ND6), two tRNA^Glu, and two CRs. However, variations in numbers of ND6 and tRNA^Glu were observed amongst individuals. These variations included type 1 (containing two ND6 and two tRNA^Glu), type 2 (containing one ND6, one incomplete ND6, and one tRNA^Glu), and type 3 (containing one ND6 and one tRNA^Glu). The gene orders of types 1 and 2, and variations in numbers of ND6 and tRNA^Glu were not previously found in any Antarctic notothenioids, whereas type 3 is the same as that of Racovitzia glacialis. Phylogenetic analyses of CR DNA sequences showed that duplicated CRs of the same species formed a monophyletic group, suggesting that duplication of CRs occurred in each species. The frequent duplication of mt genomes in Antarctic notothenioids is an unusual feature in vertebrates. We propose that interspecific hybridization and impairment of mismatch repair might account for the high frequency of gene duplications and rearrangement of mt genomes in Antarctic notothenioids.     

Nucleotide sequence and evolution of coding and noncoding regions of a quail mitochondrial genome

Summary Segments of the Japanese quail mito-chondrial genome encompassing many tRNA and protein genes, the small and part of the large rRNA genes, and the control region have been cloned and sequenced. Analysis of the relative position of these genes confirmed that the tRNA^Glu and ND6 genes in galliform mitochondrial DNA are located immediately adjacent to the control region of the molecule instead of between the cytochrome b and ND5 genes as in other vertebrates. Japanese quail and chicken display another distinctive characteristic, that is, they both lack an equivalent to the light-strand replication origin found between the tRNA^Cys and tRNA^Asn genes in all vertebrate mitochondrial genomes sequenced thus far. Comparison of the protein-encoding genes revealed that a great proportion of the substitutions are silent and involve mainly transitions. This bias toward transitions also occurs in the tRNA and rRNA genes but is not observed in the control region where transversions account for many of the substitutions. Sequence alignment indicated that the two avian control regions evolve mainly through base substitutions but are also characterized by the occurrence of a 57-bp deletion/addition event at their 5′ end. The overall sequence divergence between the two gallinaceous birds suggests that avian mitochondrial genomes evolve at a similar rate to other vertebrate mitochondrial DNAs.     

Complete Mitochondrial DNA Sequence of Conger myriaster (Teleostei: Anguilliformes): Novel Gene Order for Vertebrate Mitochondrial Genomes and the Phylogenetic Implications for Anguilliform Families

The complete nucleotide sequence of the mitochondrial genome was determined for a conger eel, Conger myriaster (Elopomorpha: Anguilliformes), using a PCR-based approach that employs a long PCR technique and many fish-versatile primers. Although the genome [18,705 base pairs (bp)] contained the same set of 37 mitochondrial genes [two ribosomal RNA (rRNA), 22 transfer RNA (tRNA), and 13 protein-coding genes] as found in other vertebrates, the gene order differed from that recorded for any other vertebrates. In typical vertebrates, the ND6, tRNA^Glu, and tRNA^Pro genes are located between the ND5 gene and the control region, whereas the former three genes, in C. myriaster, have been translocated to a position between the control region and the tRNA^Phe gene that are contiguously located at the 5′ end of the 12S rRNA gene in typical vertebrates. This gene order is similar to the recently reported gene order in four lineages of birds in that the latter lack the ND6, tRNA^Glu, and tRNA^Pro genes between the ND5 gene and the control region; however, the relative position of the tRNA^Pro to the ND6–tRNA^Glu genes in C. myriaster was different from that in the four birds, which presumably resulted from different patterns of tandem duplication of gene regions followed by gene deletions in two distantly related groups of organisms. Sequencing of the ND5–cyt b region in 11 other anguilliform species, representing 11 families, plus one outgroup species, revealed that the same gene order as C. myriaster was shared by another 4 families, belonging to the suborder Congroidei. Although the novel gene orders of four lineages of birds were indicated to have multiple independent origins, phylogenetic analyses using nucleotide sequences from the mitochondrial 12S rRNA and cyt b genes suggested that the novel gene orders of the five anguilliform families had originated in a single ancestral species. Received: 13 July 2000 / Accepted: 30 November 2000     

家禽MHC结构研究进展

禽主要组织相容性复合体(Major histocompatibility complex,MHC)的结构与禽病防控、禽免疫学、禽类遗传学研究密切相关。文章对鸡、火鸡、鹌鹑、鸭和鹅的MHC结构方面的研究进展进行了综述,表明其有以下共同特点:都有保守的MHC区域,包括MHC I基因和MHC II基因及一些功能未知基因;基因排列简单而紧凑;MHC I基因内含子的长度都比哺乳动物小;鸡、火鸡、鸭和鹅的MHC I基因组序列都有8个外显子和7个内含子,MHC IIβ基因组序列都有6个外显子和5个内含子;鸡、火鸡和鹌鹑的BG基因结构模式相同;都存在微卫星重复单元。但也存在种属差异:鸡的MHC I基因和MHC II基因是双拷贝,而鸭、鹅和鹌鹑有若干个拷贝;BG基因的拷贝数及其外显子数目不同。对主要家禽MHC结构进行分析比较,将有利于对禽病学及禽免疫遗传学的进究。     

Drosophila mitochondrial DNA: Conserved sequences in the A+T-rich region and supporting evidence for a secondary structure model of the small ribosomal RNA

Summary The sequence of a segment of theDrosophila virilis mitochondrial DNA (mtDNA) molecule that contains the A+T-rich region, the small rRNA gene, the tRNA^f-met, tRNA^gln, and tRNA^ile genes, and portions of the ND2 and tRNA^val genes is presented and compared with the corresponding segment of theD. yakuba mtDNA molecule. The A+T-rich regions ofD. virilis andD. yakuba contain two correspondingly located sequences of 49 and 276/274 nucleotides that appear to have been conserved during evolution. In each species the replication origin of the mtDNA molecule is calculated to lie within a region that overlaps the larger conserved sequence, and within this overlap is found a potential hairpin structure. Substitutions between the larger conserved sequences of the A+T-rich regions, the small mt-rRNA genes, and the ND2 genes are biased in favor of transversions, 71–97% of which are AT changes. There is a 13.8 times higher frequency of nucleotide differences between the 5 halves than between the 3 halves of theD. virilis andD. yakuba small mt-rRNA genes. Considerations of the effects of observed substitutions and deletion/insertions on possible nucleotide pairing within the small mt-rRNA genes ofD. virilis andD. yakuba strongly support the secondary structure model for theDrosophila small mt-rRNA that we previously proposed.     

Maternally inherited hypertension is associated with the mitochondrial tRNA(Ile) A4295G mutation in a Chinese family

Mutations in mitochondrial DNA have been associated with cardiovascular disease. We report here the clinical, genetic, and molecular characterization of one three-generation Han Chinese family with maternally transmitted hypertension. All matrilineal relatives in this family exhibited the variable degree of hypertension at the age at onset of 36 to 56 years old. Sequence analysis of the complete mitochondrial DNA in this pedigree revealed the presence of the known hypertension-associated tRNA^Ile A4295G mutation and 33 other variants, belonging to the Asian haplogroup D4j. The A4295G mutation, which is extraordinarily conserved from bacteria to human mitochondria, is located at immediately 3′ end to the anticodon, corresponding to conventional position 37 of tRNA^Ile. The occurrence of the A4295G mutation in several genetically unrelated pedigrees affected by cardiovascular disease but the absence of 242 Chinese controls strongly indicates that this mutation is involved in the pathogenesis of cardiovascular disease. Of other variants, the tRNA^Glu A14693G and ND1 G11696A mutations were implicated to be associated with other mitochondrial disorders. The A14693G mutation, which is a highly conserved nucleoside at the TψC-loop of tRNA^Glu, has been implicated to be important for tRNA structure and function. Furthermore, the ND4 G11696A mutation was associated with Leber’s hereditary optic neuropathy. Therefore, the combination of the A4295G mutation in the tRNA^Ile gene with the ND4 G11696A mutation and tRNA^Glu A14693G mutation may contribute to the high penetrance of hypertension in this Chinese family.     

Discrepancy in divergence of the mitochondrial and nuclear genomes ofDrosophila teissieri andDrosophila yakuba

Summary Restriction sites were compared in the mitochondrial DNA (mtDNA) molecules from representatives of two closely related species of fruit flies: nine strains ofDrosophila teissieri and eight strains ofDrosophila yakuba. Nucleotide diversities amongD. teissieri strains and amongD. yakuba strains were 0.07% and 0.03%, respectively, and the nucleotide distance between the species was 0.22%. Also determined was the nucleotide sequence of a 2305-nucleotide pari (ntp) segment of the mtDNA molecule ofD. teissieri that contains the noncoding adenine+thymine (A+T)-rich region (1091 ntp) as well as the genes for the mitochondrial small-subunit rRNA, tRNA^f-met, tRNA^gln, and tRNA^ile, and portions of the ND2 and tRNA^val genes. This sequence differs from the corresponding segment of theD. yakuba mtDNA by base substitutions at 0.1% and 0.8% of the positions in the coding and noncoding regions, respectively. The higher divergence due to base substitutions in the A+T-rich region is accompanied by a greater number of insertions/deletions than in the coding regions. From alignment of theD. teissieri A+T-rich sequence with those ofD. yakuba andDrosophila virilis, it appears that the 40% of this sequence that lies adjacent to the tRNA^ile gene has been highly conserved. Divergence between the entireD. teissieri andD. yakuba mtDNA molecules, estimated from the sequences, was 0.3%; this value is close to the value (0.22%) obtained from the restriction analysis, but 10 times lower than the value estimated from published DNA hybridization results. From consideration of the relationships of mitochondrial nucleotide distance and allozyme genetic distance found among seven species of theDrosophila melanogaster subgroup, the mitochondrial nucleotide distance observed forD. teissieri andD. yakuba is anomalously low in relation to the nuclear genetic distance.     

Platyhelminth mitochondrial DNA: Evidence for early evolutionary origin of a tRNAserAGN that contains a dihydrouridine arm replacement loop,and of serine-specifying AGA and AGG codons

Summary The nucleotide sequence of a segment of the mitochondrial DNA (mtDNA) molecule of the liver flukeFasciola hepatica (phylum Platyhelminthes, class Trematoda) has been determined, within which have been identified the genes for tRNA^ala, tRNA^asp, respiratory chain NADH dehydrogenase subunit I (ND1), tRNA^asn, tRNA^pro, tRNA^ile, tRNA^lys, ND3, tRNA^serAGN, tRNA^trp, and cytochromec oxidase subunit I (COI). The 11 genes are arranged in the order given and are all transcribed from the same strand of the molecule. The overall order of theF. hepatica mitochondrial genes differs from what is found in other metazoan mtDNAs. All of the sequenced tRNA genes except the one for tRNA^serAGN can be folded into a secondary structure with four arms resembling most other metazoan mitochondrial tRNAs, rather than the tRNAs that contain a TψC arm replacement loop, found in nematode mtDNAs. TheF. hepatica mitochondrial tRNA^serAGN gene contains a dihydrouridine arm replacement loop, as is the case in all other metazoan mtDNAs examined to date. AGA and AGG are found in theF. hepatica mitochondrial protein genes and both codons appear to specify serine. These findings concerningF. hepatica mtDNA indicate that both a dihydrouridine arm replacement loop-containing tRNA^serAGN gene and the use of AGA and AGG codons to specify serine must first have occurred very early in, or before, the evolution of metazoa.     

Nucleotide sequence of nine protein-coding genes and 22 tRNAs in the mitochondrial DNA of the sea starPisaster ochraceus

Summary We have cloned and sequenced over 9 kb of the mitochondrial genome from the sea starPisaster ochraceus. Within a continuous 8.0-kb fragment are located the genes for NADH dehydrogenase subunits 1, 2, 3, and 4L (ND1, ND2, ND3, and ND4L), cytochrome oxidase subunits I, II, and III (COI, COII, and COIII), and adenosine triphosphatase subunits 6 and 8 (ATPase 6 and ATPase 8). This large fragment also contains a cluster of 13 tRNA genes between ND1 and COI as well as the genes for isoleucine tRNA between ND1 and ND2, arginine tRNA between COI and ND4L, lysine tRNA between COII and ATPase 8, and the serine (UCN) tRNA between COIII and ND3. The genes for the other five tRNAs lie outside this fragment. The gene for phenylalanine tRNA is located between cytochrome b and the 12S ribosomal genes. The genes for tRNA^glu and tRNA^thr are 3 to the 12S ribosomal gene. The tRNAs for histidine and serine (AGN) are adjacent to each other and lie between ND4 and ND5. These data confirm the novel gene order in mitochondrial DNA (mtDNA) of sea stars and delineate additional distinctions between the sea star and other mtDNA molecules.     

Somatic mutations of mitochondrial genome in hepatocellular carcinoma

Somatic mutations have been identified in mitochondrial DNA (mtDNA) of various human primary cancers. However, their roles in the pathophysiology of cancers are still unclear. In our previous study, high frequency of somatic mutations was found in the D-loop region of mtDNA of hepatocellular carcinomas (HCCs). In the present study, we examined 44 HCCs and corresponding non-cancerous liver tissues, and identified 13 somatic mutations in the coding region of mtDNAs from 11 HCC samples (11/44, 25%). Among the 13 mtDNA mutations, six mutations (T6787C, G7976A, A9263G, G9267A, A9545G and A11708G) were homoplasmic while seven mutations (956delC, T1659C, G3842A, G5650A, 11032delA, 12418insA and a 66 bp deletion) were heteroplasmic. Moreover, the G3842A transition created a premature stop codon and the 66 bp deletion could omit 22 amino acid residues in the NADH dehydrogenase (ND) subunit 1 (ND1) gene. The 11032delA and 12418insA could result in frame-shift mutation in the ND4 and ND5 genes, respectively. The T1659C transition in tRNA^Val gene and G5650A in tRNA^Ala gene were reported to be clinically associated with some mitochondrial disorders. In addition, the T6787C (cytochrome c oxidase subunit I, COI), G7976A (COII), G9267A (COIII) and A11708G (ND4) mutations could result in amino acid substitutions in the highly conserved regions of the affected mitochondrial genes. These mtDNA mutations (10/13, 76.9%) have the potential to cause mitochondrial dysfunction in HCCs. Taken these results together, we suggest that there may be a higher frequency of mtDNA mutations in HCC than in normal liver tissues from the same individuals.     

Cloning, Sequencing and Analysis of the 16S-23S rDNA Intergenic Spacers (IGSs) of Two Strains of Vibrio vulnificus

According to the conserved sequences flanking the 3′ end of the 16S and the 5′ end of the 23S rDNAs, PCR primers were designed, and the 16S-23S rDNA intergenic spacers (IGSs) of two strains of Vibrio vulnificus were amplified by PCR and cloned into pGEM-T vector. Different clones were selected to be sequenced and the sequences were analyzed with BLAST and the software DNAstar. Analyses of the IGS sequences suggested that the strain ZSU006 contains five types of polymorphic 16S-23S rDNA intergenic spacers, namely, IGS^GLAV, IGS^GLV, IGS^lA, IGS^G and IGS^A; while the strain CG021 has the same types of IGSs except lacking IGS^A. Among these five IGS types, IGS^GLAV is the biggest type, including the gene cluster of tRNA^Glu - tRNA^Lys - tRNA^Ala - tRNA^Val; IGS^GLV includes that of tRNA^Glu-tRNA^Lys-tRNA^Val; IGS^AG, tRNA^Ala-tRNA^Glu; IGS^IA, tRNA^Ile-tRNA^Ala; IGS^G, tRNA^Glu and IGS^A, tRNA^Ala. Intraspecies multiple alignment of all the IGS sequences of these two strains with those of V. vulnificus ATCC27562 available at GenBank revealed several highly conserved sequence blocks in the non-coding regions flanking the tRNA genes within all of strains, most notably the first 40 and last 200 nucleotides, which can be targeted to design species-specific PCR primers or detection probes. The structural variations of the 16S-23S rDNA intergenic spacers lay a foundation for developing diagnostic methods for V. vulnificus.     

Duplication and Concerted Evolution of the Mitochondrial Control Region in the Parrot Genus Amazona

We report a duplication and rearrangement of the mitochondrialgenome involving the control region of parrots in the genusAmazona. This rearrangement results in a gene order of cytochromeb/tRNA^Thr/pND6/pGlu/CR1/tRNA^Pro/NADH dehydrogenase 6/tRNA^Glu/CR2/tRNA^Phe/12srRNA, where CR1 and CR2 refer to duplicate control regions,and pND6 and pGlu indicate presumed pseudogenes. In contrastto previous reports of duplications involving the control regionsof birds, neither copy of the parrot control region shows anyindications of degeneration. Rather, both copies contain manyof the conserved sequence features typically found in aviancontrol regions, including the goose hairpin, TASs, the F, C,and D boxes, conserved sequence box 1 (CSB1), and an apparenthomolog to the mammalian CSB3. We conducted a phylogenetic analysisof homologous portions of the duplicate control regions from21 individuals representing four species of Amazona (A. ochrocephala,A. autumnalis, A. farinosa, and A. amazonica) and Pionus chalcopterus.This analysis revealed that an individual's two control regioncopies (i.e., the paralogous copies) were typically more closelyrelated to one another than to corresponding segments of otherindividuals (i.e., the orthologous copies). The average sequencedivergence of the paralogous control region copies within anindividual was 1.4%, versus a mean value of 4.1% between controlregion orthologs representing nearest phylogenetic neighbors.No differences were found between the paralogous copies in eitherthe rate or the pattern in which the two copies accumulatedbase pair changes. This pattern suggests concerted evolutionof the two control regions, perhaps through occasional geneconversion events. We estimated that gene conversion eventsoccurred on average every 34,670 ± 18,400 years basedon pairwise distances between the paralogous control regionsequences of each individual. Our results add to the growingbody of work indicating that under some circumstances duplicatedmitochondrial control regions are retained through evolutionarytime rather than degenerating and being lost, presumably dueto selection for a small mitochondrial genome.     

The complete mitochondrial genome of the leafminer <Emphasis Type="Italic">Liriomyza trifolii</Emphasis> (Diptera: Agromyzidae)

Liriomyza trifolii (Diptera: Agromyzidae) is one of the most economically significant pests in the world. In this paper we present sequence data for the complete mitochondrial genome of L. trifolii. The circular genome is 16,141 bp long and contains one encoding region including 37 genes and one non-coding A+T-rich region. Gene numbers and organization is similar to that of the typical insect mitochondrial genomes except that two additional tRNA genes are found in the A+T-rich region (tRNA^Thr and tRNA^Leu(UUR)). All of the protein initiation codons are ATN, except ND1 which begins with GTG and COI which is initiated by the quadruplet ATCA. The 22 tRNA anticodons of L. trifolii match those observed in Drosophila yakuba, and all of tRNAs form the typical cloverleaf structure except for tRNA^Ser(AGN), which has lost the DHU-arm. The A+T-rich region of L. trifolii also contains two previously noted Diperan features—a highly conserved polyT stretch and a (TA)_n stretch.     

Evolution of the noncoding regions inDrosophila mitochondrial DNA: Two intergenic regions

The sequences of the mitochondrial DNA (mtDNA) segment containing the two intergenic regions were determined for six species belonging to theDrosophila immigrans species group and compared to the corresponding segments ofDrosophila species which had been studied previously. We found remarkable differences in the evolutionary rates of the two intergenic regions. The Intergenic I region, which lies between thetRNA ^gln and thetRNA ^ile genes, was found to be highly conserved in terms of both size (30 ntp) and nucleotide sequence among the species studied. In contrast, the sequences of the Intergenic II region, which lies between thetRNA ^f-met and thetRNA ^ile genes, showed considerable variation. The size of the Intergenic II region ranged from 0 to 88 ntp, and accurate alignment was possible only among sequences from geographical strains or very closely related species in thenasuta species subgroup. The observed differences in conservation of the two mtDNA intergenic regions are discussed in light of functional constraints on mtDNA sequences.     

Mitochondrial DNA sequences of greenbug (Homoptera: Aphididae) biotypes

Sequence comparisons were made for 738-bp of mtDNA cloned from seven greenbug, Schizaphis graminum, biotypes (B, C, E, F, G, H and I) obtained from laboratory colonies maintained by USDA-ARS, Stillwater, OK. These sequences include parts of the genes for 16S ribosomal subunit (16S rRNA), tRNA^leu, tRNA^ser, cytochrome b (cytb) and NADH dehydrogenase (ND) subunits one and four. Sequence data revealed considerable variation in 86 (12%) nucleotide sites over the 738-bp sequenced among the seven greenbug biotypes. Nucleotide invariance was observed within the seven greenbug biotypes from both the laboratory colonies and field collected biotype E greenbugs from Kansas, Nebraska, Oklahoma, and Texas.     

The complete mitochondrial genome structure of snow leopard <Emphasis Type="Italic">Panthera uncia</Emphasis>

The complete mitochondrial genome (mtDNA) of snow leopard Panthera uncia was obtained by using the polymerase chain reaction (PCR) technique based on the PCR fragments of 30 primers we designed. The entire mtDNA sequence was 16 773 base pairs (bp) in length, and the base composition was: A—5,357 bp (31.9%); C—4,444 bp (26.5%); G—2,428 bp (14.5%); T—4,544 bp (27.1%). The structural characteristics [0] of the P. uncia mitochondrial genome were highly similar to these of Felis catus, Acinonyx jubatus, Neofelis nebulosa and other mammals. However, we found several distinctive features of the mitochondrial genome of Panthera unica. First, the termination codon of COIII was TAA, which differed from those of F. catus, A. jubatus and N. nebulosa. Second, tRNA^{Ser (AGY)}, which lacked the ‘‘DHU’’ arm, could not be folded into the typical cloverleaf-shaped structure. Third, in the control region, a long repetitive sequence in RS-2 (32 bp) region was found with 2 repeats while one short repetitive segment (9 bp) was found with 15 repeats in the RS-3 region. We performed phylogenetic analysis based on a 3 816 bp concatenated sequence of 12S rRNA, 16S rRNA, ND2, ND4, ND5, Cyt b and ATP8 for P. uncia and other related species, the result indicated that P. uncia and P. leo were the sister species, which was different from the previous findings.     

Nucleotide sequences of transfer RNA genes in the Pisum sativum chloroplast DNA

Summary Eight transfer RNA (tRNA) genes which were previously mapped to five regions of the Pisum sativum (pea) chloroplast DNA (ctDNA) have been sequenced. They have been identified as tRNA^Val(GAC), tRNA^Asn(GUU), tRNA^Arg(ACG), tRNA^Leu(CAA), tRNA^Tyr(GUA), tRNA^Glu(UUC), tRNA^His(GUG), and tRNA^Arg(UCU) by their anticodons and by their similarity to other previously identified tRNA genes from the chloroplast DNAs of higher plants or from E. gracilis. In addition,two other tRNA genes, tRNA^Gly (UCC) and tRNA^Ile(GAU), have been partially sequenced. The tRNA genes are compared to other known chloroplast tRNA genes from higher plants and are found to be 90–100% homologous. In addition there are similarities in the overall arrangement of the individual genes between different plants. The 5 flanking regions and the internal sequences of tRNA genes have been studied for conserved regions and consensus sequences. Two unusual features have been found: there is an apparent intron in the D-loop of the tRNA^Gly(UCC), and the tRNA^Glu(UUC) contains GATTC in its T-loop.     

Organization of the Mitochondrial Genome of a Deep-Sea Fish, Gonostoma gracile (Teleostei: Stomiiformes): First Example of Transfer RNA Gene Rearrangements in Bony Fishes

We determined the complete nucleotide sequence of the mitochondrial genome (except for a portion of the putative control region) for a deep-sea fish, Gonostoma gracile. The entire mitochondrial genome was purified by gene amplification using long polymerase chain reaction (long PCR), and the products were subsequently used as templates for PCR with 30 sets of newly designed, fish-universal primers that amplify contiguous, overlapping segments of the entire genome. Direct sequencing of the PCR products showed that the genome contained the same 37 mitochondrial structural genes as found in other vertebrates (two ribosomal RNA, 22 transfer RNA, and 13 protein-coding genes), with the order of all rRNA and protein-coding genes, and 19 tRNA genes being identical to that in typical vertebrates. The gene order of the three tRNAs (tRNA^Glu, tRNA^Thr, and tRNA^Pro) relative to cytochrome b, however, differed from that determined in other vertebrates. Two steps of tandem duplication of gene regions, each followed by deletions of genes, can be invoked as mechanisms generating such rearrangements of tRNAs. This is the first example of tRNA gene rearrangements in a bony fish mitochondrial genome. Received August 5, 1998; accepted February 19, 1999.