The Mitochondrial Genome of the Sea Anemone Metridium senile (Cnidaria): Introns, a Paucity of tRNA Genes, and a Near-Standard Genetic Code

Corresponding author: David R. Wolstenholme, Department of Biology, University of Utah, Salt Lake City, Utah 84112, wolstenholme{at}biology.utah.edu (E-mail).

Communicating editor: G. B. GOLDING

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUDING REMARKS LITERATURE CITED

The circular, 17,443 nucleotide-pair mitochondrial (mt) DNA molecule of the sea anemone, Metridium senile (class Anthozoa, phylum Cnidaria) is presented. This molecule contains genes for 13 energy pathway proteins and two ribosomal (r) RNAs but, relative to other metazoan mtDNAs, has two unique features: only two transfer RNAs (tRNA^f-Met and tRNA^Trp) are encoded, and the cytochrome c oxidase subunit I (COI) and NADH dehydrogenase subunit 5 (ND5) genes each include a group I intron. The COI intron encodes a putative homing endonuclease, and the ND5 intron contains the molecule's ND1 and ND3 genes. Most of the unusual characteristics of other metazoan mtDNAs are not found in M. senile mtDNA: unorthodox translation initiation codons and partial translation termination codons are absent, the use of TGA to specify tryptophan is the only genetic code modification, and both encoded tRNAs have primary and secondary structures closely resembling those of standard tRNAs. Also, with regard to size and secondary structure potential, the mt-s-rRNA and mt-l-rRNA have the least deviation from Escherichia coli 16S and 23S rRNAs of all known metazoan mt-rRNAs. These observations indicate that most of the genetic variations previously reported in metazoan mtDNAs developed after Cnidaria diverged from the common ancestral line of all other Metazoa.

COMPLETE nucleotide sequences have been reported for mitochondrial (mt) genomes of more than 40 Metazoa (multicellular animals). About half of these mt genomes are from chordates, but the remainder are from organisms representing at least six invertebrate phyla (for references see WOLSTENHOLME 1992A ; OKIMOTO et al. 1992 ; BOORE and BROWN 1994 ; BOORE and BROWN 1995 ; KRETTEK et al. 1995 ; ASAKAWA et al. 1995 ; FLOOK et al. 1995 ; HATZOGLOU et al. 1995 ; BEAGLEY et al. 1995 , BEAGLEY et al. 1996 ; ARNASON et al. 1996 ). Most metazoan mt genomes comprise a single circular, double-stranded DNA molecule between 14 and 18 kb that contains a uniform set of 37 genes. There are genes for 13 energy pathway proteins: Cytochrome b (Cyt b), subunits I–III of cytochrome c oxidase (COI–COIII), subunits 6 and 8 of the F₀ ATP synthetase complex (ATPase6 and 8), and subunits 1–6 and 4L of the respiratory chain NADH dehydrogenase (ND1–ND6 and ND4L), for the two RNA components (s-rRNA and l-rRNA) of mt ribosomes, and for 22 tRNAs. These genes are arranged with very few intervening nucleotide pairs (ntp), except that in each metazoan mtDNA, there is a sequence of between 125 and 8000 ntp that lacks genes but includes the molecules' major transcription promoters and origin of replication and is therefore designated the control region (see CLAYTON 1991 , CLAYTON 1992 ).

Exceptions to the constant gene content of metazoan mtDNAs include (1) the lack of an ATPase8 gene in mtDNAs of nematodes and of the mollusc Mytilus edulis (WOLSTENHOLME et al. 1987 ; OKIMOTO et al. 1991 , OKIMOTO et al. 1992 ; HOFFMAN et al. 1992), (2) the occurrence of an extra gene for a tRNA expected to recognize AUA (methionine) codons in mtDNAs of M. edulis and Mytilus californianus (HOFFMAN et al. 1992; C. T. BEAGLEY, R. OKIMOTO and D. R. WOLSTENHOLME, unpublished data), (3) the occurrence of a gene for a bacterial MutS homologue in mtDNAs of octocorals (PONT-KINGDON et al. 1995 , PONT-KINGDON et al. 1997), and (4) the occurrence of two group I introns in mtDNAs of sea anemones (BEAGLEY et al. 1995 , BEAGLEY et al. 1996 ).

Metazoan mtDNAs exhibit a number of unusual features. Modifications relative to the standard genetic code are found in each of the metazoan mt genetic codes examined to date. In all metazoan mtDNAs, TGA serves to specify tryptophan rather than termination. ATA has the standard code specification of isoleucine only in mtDNAs of echinoderms ( JACOBS et al. 1988 ; CANTATORE et al. 1989 ) and some Platyhelminthes (BESSHO et al. 1992 ) and Cnidaria (PONT-KINGDON et al. 1994 ; BEAGLEY et al. 1995 , BEAGLEY et al. 1996 ); in all other metazoan mt genetic codes, ATA acts as a second methionine codon. AGA and AGG have the standard code specification of arginine only in Cnidaria (PONT-KINGDON et al. 1994 ; BEAGLEY et al. 1995 , BEAGLEY et al. 1996 ). Among other invertebrates, these codons specify serine (except that AGG codons are absent from Drosophila mtDNAs); in ascidians, they specify glycine, but in vertebrates, neither specifies an amino acid. In the latter case, they may act as rare stop codons (reviewed in WOLSTENHOLME and FAURON 1995 ).

In most metazoan mtDNAs, some protein genes begin with unorthodox translation initiation codons that include ATA, ATT, ATC, GTG, GTT, and TTG (OKIMOTO and WOLSTENHOLME 1990 ; WOLSTENHOLME 1992A ). In organisms from different metazoan phyla, some mt protein genes end in T or TA, and in mammals, it has been shown that precise cleavage of transcripts 3' to these nucleotides, followed by polyadenylation, creates a complete UAA termination codon (OJALA et al. 1981 ).

Unusual wobble rules that allow the anticodon of some tRNAs to recognize all codons of a four-codon family have been suggested as the explanation for why the number of tRNAs encoded in metazoan mtDNAs is only 22, the number that under these conditions would be sufficient to translate mt protein gene transcripts (BARRELL et al. 1979 , BARRELL et al. 1980 ). Numerous structural modifications relative to standard tRNAs are found among metazoan mtDNA–encoded tRNAs. These include some in which one or other (but never both) of the side arms are replaced with a single loop of nucleotides lacking the usual secondary structure potential (DE BRUIJN et al. 1980 ; WOLSTENHOLME et al. 1987 , WOLSTENHOLME et al. 1994 ; OKIMOTO and WOLSTENHOLME 1990 ).

The two rRNAs encoded by metazoan mtDNAs (s-rRNA and l-rRNA) are clearly homologous to Escherichia coli 16S and 23S rRNAs (GUTTEL and FOX 1988; GUTTEL et al. 1993). The metazoan mt-s-rRNAs and mt-l-rRNAs, however, are all smaller by factors of as much as 2 and 3, respectively, than their E. coli counterparts. Nevertheless, all metazoan mt-rRNAs have the potential to fold into structures that resemble at least the core structures of the corresponding secondary structures of E. coli rRNAs (GRAY et al. 1984 ; GUTTEL et al. 1993; OKIMOTO et al. 1994 ; PONT-KINGDON et al. 1994 ).

We presented previously the sequence of a 6.1-kb segment of the circular mtDNA molecule of the sea anemone Metridium senile (PONT-KINGDON et al. 1994 ), and we have reported separately on the structure and processing of two group I introns in the M. senile COI and ND5 genes (BEAGLEY et al. 1996 ). In this paper, we present the entire sequence and analysis of the M. senile 17.4 -kb mtDNA molecule. The data confirm our initial conclusions regarding the M. senile mt genetic code, and they also show that with regard to primary sequence and secondary structure potential, the M. senile mt-l-rRNA bears the closest resemblance to the E. coli 23S rRNA than any presently known mt-l-rRNA. A further noteworthy finding is the occurrence in M. senile mt-DNA of only two tRNA genes. This is the first recorded case of a metazoan mtDNA that does not encode sufficient tRNAs to carry out protein synthesis in mitochondria. We discuss the evolutionary implications of this finding with regard to similar deficiencies of tRNAs in some protozoan and plant mtDNAs. Finally, in this paper, we also discuss factors that might be expected to contribute to the maintenance of the two introns in M. senile mtDNA.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUDING REMARKS LITERATURE CITED

Animals and mt nucleic acid isolations:
Specimens of the white morph of M. senile with an average size <5 cm were obtained from Dr. RIMOND C. FAY. Mitochondria were isolated by methods described previously (WOLSTENHOLME and FAURON 1976 ), except that sucrose was replaced with mannitol and 0.1–0.2% bovine serum albumin was present in all solutions. Mitochondria were lysed with 2% Sarkosyl, and mtDNA was isolated by cesium chloride-ethidium bromide centrifugation (WOLSTENHOLME and FAURON 1976 ). RNA was isolated from M. senile mitochondria using the RNA-Gents Total RNA Isolation Kit (Promega, Madison, WI) and was treated with RNase-free DNase I (Stratagene, La Jolla, CA) using the supplier's recommended reaction conditions.

Restriction enzyme digestions and cloning:
M. senile mtDNA was cleaved with a combination of BamHI and Bgl II to yield five fragments of ~5.0, 4.5, 4.2, 2.8, and 1.0 kb, and each was cloned into BamHI-cleaved bacteriophage M13mp19 DNA. From the cloned fragments, nested sets of deletion clones were produced (DALE et al. 1985 ). Clones of fragments that cross each of the M. senile mtDNA BamHI or Bgl II sites were also obtained using various other restriction enzymes.

DNA sequencing and sequence analysis:
DNA sequences of overlapping deletion clones were obtained by the method of SANGER et al. 1977 , using Sequenase (Amersham, Arlington Heights, IL), and assembled (STADEN 1982 ). Each of the five BamHI-Bgl II restriction fragments was completely sequenced in both directions, and all BamHI or Bgl II sites were sequenced across in one direction. Sequences were analyzed using Wisconsin Genetics Computer Group programs and Lasergene software from DNASTAR, Inc. (Madison, WI). The nucleotide sequence of the M. senile mtDNA molecule has been submitted to GenBank under the accession number BankIt 108562 AF000023. Secondary structure potentials of intergenic regions were analyzed using the program of ZUCKER (1989).

Determination of the 5' and 3' ends of the l-rRNA gene:
5' and 3' RACE (rapid amplification of cDNA ends) analyses (FROHMAN 1990 ) were carried out to locate the ends of the M. senile mt-l-rRNA by following similar procedures and by using the same 5' and 3' RACE pools that were previously used to determine the ends of the M. senile s-rRNA (PONT-KINGDON et al. 1994 ). From the 5' RACE pool, 5' end regions were PCR amplified using primer TB17-2 (5' CCAGATCTGGATCCTGCAG) and the antisense primer TBLR-1 (5' CGCTCGCCGCTACTTACGG, nt 269–251, Figure 1). The double-stranded PCR product was sequenced directly using primer TBLR-1 and the Taq Cycle Sequencing kit (Amersham). A second 5' RACE determination was made using a new 5' RACE pool prepared from cDNAs (generated as before with reverse transcriptase and random hexamers) that had been 3' polyguanylated and then polyadenylated using terminal deoxynucleotidyl transferase (Amersham). For determination of the mt-l-rRNA 3' end, 3' end regions were PCR amplified from the 3' RACE pool using primer TB17-2 and the sense oligonucleotide TBLR-2 (5' TTAAATAAAAAGTTTAGATAATGTG, nt 1614 –1638, Figure 1). The double-stranded amplification product was digested with BamHI, which recognizes a site at nt 1639 (Figure 1) and in the TB17-2 sequence, and was cloned into BamHI-cleaved M13mp19. The resulting clones were screened by dot blot analysis (DAVIS et al. 1986 ) using oligonucleotide TBLR-3 (5' GGTTTCTATCTACAATATG, nt 2059–2077, Figure 1), and eight positive clones were sequenced using TBLR-3 as primer.

View larger version (349K): [in this window] [in a new window] [Download PPT slide]   Figure 1. —Nucleotide sequence of the circular 17,443-ntp mtDNA molecule of the sea anemone M. senile. The sequence contains the genes for 13 energy pathway proteins: Cyt b, cytochrome b; COI–COIII, subunits I–III of cytochrome c oxidase; ATPase6 and ATPase8, subunits 6 and 8 of the F0 ATP synthetase; ND1–6 and 4L, respiratory chain NADH dehydrogenase subunits 1–6 and 4L; and for two ribosomal RNAs (s-rRNA and l-rRNA) and two transfer RNAs (tRNAf-Met and tRNATrp). The locations of the single group I intron within each of the COI and ND5 genes are indicated. ORF is a gene for a putative homing endonuclease within the COI intron. Note that the ND1 and ND3 genes are within the ND5 intron. The arrow at the beginning of the sequence indicates the single direction of transcription for all genes: left to right, and therefore the strand of each gene shown is the sense strand. The predicted amino acid sequences of each of the 14 protein genes is given below the sequence. TGA codons are shown to specify tryptophan rather than termination (ter, termination codons). Intron splice sites were determined by RNA sequence analyses (BEAGLEY et al. 1996 ). The 5' and 3' ends of the s-rRNA and l-rRNA genes were determined by 5' and 3' RACE analyses (PONT-KINGDON et al. 1994 ; this study). Within an overline of each tRNA gene, brackets identify the anticodon. Nucleotides 1–477 and 11786–17443 are from PONT-KINGDON et al. 1994 .

	RESULTS AND DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUDING REMARKS LITERATURE CITED

The complete nucleotide sequence of the circular mtDNA molecule of M. senile is given in Figure 1, and gene content and organization are summarized in the map shown in Figure 2. The size of this molecule, 17,443 ntp, is within the range of sizes found for mtDNAs of other Metazoa (WOLSTENHOLME 1992A , WOLSTENHOLME 1992B ). The circularity of the M. senile mtDNA molecule is in keeping with all other noncnidarian metazoan mtDNAs examined to date. It has been reported, however, that mtDNAs of members of three other classes of Cnidaria are noncircular: the mtDNAs of Hydra attenuata and of other Hydra species (class Hydrozoa) are in the form of two 8-kb linear molecules (WARRIOR and GALL 1985 ; BRIDGE et al. 1992 ), and mtDNAs of other Hydrozoa, Scyphozoa and Cubozoa, are in the form of a single linear molecule of 14 –18 kb (BRIDGE et al. 1992 ).

View larger version (34K): [in this window] [in a new window] [Download PPT slide]   Figure 2. —Gene map of the M. senile mtDNA molecule. The molecule contains the genes for 13 energy pathway proteins, two rRNAs and two tRNAs. All genes are transcribed in the direction shown by the arrow. Numbers at gene boundaries on the inside of the map indicate apparently noncoding nucleotides between genes. A single group I intron (stippled) occurs in each of the COI and ND5 genes. The ORF within the COI intron encodes a putative homing endonuclease. The ND1 and ND3 genes occur within the ND5 intron.

The M. senile mtDNA molecule contains the same 13 protein genes and two rRNA genes found in other metazoan mtDNAs; however, there are genes for only two tRNAs, tRNA^f-Met and tRNA^Trp, rather than for 22 tRNAs reported for all other fully sequenced metazoan mtDNAs. In the M. senile mtDNA molecule, all genes are transcribed from the same strand of the molecule. This is also the case in nematode, bivalve mollusc, and annelid mtDNAs. In insects, echinoderms, gasteropods, and vertebrates, however, both strands are involved as gene templates.

Two of the M. senile mt protein genes, COI and ND5, each contain a group I intron. The COI intron contains a free-standing open reading frame (ORF) that appears to encode a protein homologous to homing endonucleases encoded in group I introns from a variety of organisms (LAMBOWITZ and BELFORT 1993 ; DOOLITTLE 1993 ). The ND5 intron contains the molecule's only copies of the ND1 and ND3 genes. We have described the structure and processing of these introns elsewhere (BEAGLEY et al. 1996 ).

Protein genes:
Four of the 13 M. senile energy pathway mt protein genes have been described previously (PONT-KINGDON et al. 1994 ). Six of the nine other mt protein genes (COI, COIII, ND1, ND2, ND3, and ND5) were identified by comparisons of amino acid sequences to amino acid sequences of previously identified mt protein genes of mouse and Drosophila yakuba (Figure 3). The remaining three mt protein genes were identified as ATPase6, ATPase8, and ND4L from minimal predicted amino acid sequence and size similarities (Figure 3) and hydropathic profile similarities (data not shown). All the M. senile genes, with the exception of ND5, are larger than the homologous genes of mouse by an average of 5.8% [range 0.38% (COIII) to 11.74% (ND6)], and with the exception of COIII, they are larger than the homologous genes of D. yakuba by 8.5% [range 2.7% (ATPase6) to 35.9% (ATPase8)].

View larger version (146K): [in this window] [in a new window] [Download PPT slide]   Figure 3. —Comparisons of the predicted amino acid sequences of nine M. senile (M.s) mt proteins with the corresponding proteins of Mus musculus (M.m, BIBB et al. 1981 ) and Drosophila yakuba (D.y, CLARY and WOLSTENHOLME 1985 ). Similar comparisons of the remaining four M. senile mtDNA-encoded energy pathway proteins are given in Pont-Kingdon et al. 1994. In the mouse and D. yakuba sequences, an asterisk indicates an amino acid that is identical in the M. senile sequence. A dash indicates the absence in one sequence of an amino acid that occurs in one or more of the other corresponding sequences. For translation of mtDNA nucleotide sequences, the following nonstandard genetic code specifications were used: TGA specifies tryptophan in all three species, ATA specifies methionine in the mouse and D. yakuba sequences, and AGA specifies serine in the D. yakuba sequences [AGG codons do not specify an amino acid in D. yakuba mtDNA, and neither AGA nor AGG specify an amino acid in mouse mtDNA (BIBB et al. 1981 ; CLARY and WOLSTENHOLME 1985 )]. Above some of the M. senile sequences, dots identify amino acids interpreted as isoleucine, specified by ATA codons; ± and signs identify amino acids interpreted as arginine specified by AGA and AGG, respectively. Amino acids in all the sequences shown that are specified by TGA codons and interpreted as tryptophan are identified by a # sign. Group I intron insertion sites in corresponding M. senile COI and ND5 gene sequences are indicated. Percentage amino acid sequence identities of M. senile and mouse mt proteins are in the range 13.4% (ND6) to 66% (COI; mean = 39.0%), and those of M. senile and D. yakuba mt proteins are in the range 9.9% (ND6) to 67.7% (COI; mean = 37.7%).

Similarity in gene arrangement between the M. senile mtDNA molecule and mtDNA molecules of other metazoans is severely limited. The gene arrangement ATPase8–ATPase6 in M. senile mtDNA, where both genes are transcribed in the same direction, is conserved in mtDNAs of vertebrates, echinoderms, arthropods, and the polyplacophoran mollusc Katharina tunicata, but in vertebrate, echinoderm, and arthropod mtDNAs, the 3' end of the ATPase8 gene overlaps the 5' end of the ATPase6 gene by as many as 43 ntp (BIBB et al. 1981 ; JACOBS et al. 1988 ; CANTATORE et al. 1989 ; ASAKAWA et al. 1995 ; CLARY and WOLSTENHOLME 1985 ; CROZIER and CROZIER 1993 ; VALVERDE et al. 1994 ; BOORE and BROWN 1994 ). The gene arrangement ND6-Cyt b, where both genes are transcribed from left to right, is also found in arthropods, in the mollusc K. tunicata, and in the annelid Lumbricus terrestris (BOORE and BROWN 1994 , BOORE and BROWN 1995 ).

Codon usage and the genetic code:
Codon usage among the 13 energy pathway protein genes and, separately, in the COI intron ORF are given in Table 1. All codons are used in the 13 energy pathway protein genes. The relative frequencies of nucleotides in codon third positions (T, 40.2%; A, 28.1%; G, 17.1%; C, 14.7%) are similar to the relative frequencies of all nucleotides in the sense strand of these genes (T, 38.0%; A, 24.5%; G, 20.5%; C, 17.0%). The overall A plus T content of protein gene sense strands (62.5%) is within the range of values found for other metazoan mt protein genes: 83.3% (Apis mellifera) to 54.9% (human). In the COI intron ORF, 10 codons, nine of which end in C (five) or G (four), do not occur. This is correlated with the lower overall usage of G (13.5%) and C (10.8%) in codon third positions in the COI intron ORF than in the energy pathway protein genes (G, 17.1%; C, 14.7%). As expected for a protein predicted to interact with a nucleic acid, there are much higher frequencies of lysine- and arginine-specifying codons (11.7 and 6.7%, respectively) in the COI intron ORF than in the energy pathway protein genes (2.6 and 2.5%, respectively; Table 1). In contrast to the finding that in most invertebrate and vertebrate mtDNAs either a fraction or all of the protein genes begin with various unorthodox translation initiation codons, all of the M. senile mt protein genes begin with ATG. Such a situation has been recorded previously only for the mtDNAs of the amphibian Xenopus laevis and the annelid L. terrestris (ROE et al. 1985 ; BOORE and BROWN 1995 ). All of the M. senile mt protein genes end with either TAA (nine) or TAG (four).

Data obtained from analysis of the four complete mt protein genes (Cyt b, COII, ND4, and ND6) and one partial mt protein gene (ND2) of M. senile (PONT-KINGDON et al. 1994 ) were consistent with the interpretation that TGA specifies an amino acid rather than termination in the M. senile mt genetic code. These data, however, were insufficient to permit the conclusion that the amino acid specified was tryptophan as in all other known invertebrate mt genetic codes. Among the 13 M. senile energy pathway mt protein genes, there are a total of 12 internal TGA codons: between one and three TGAs occur in eight genes, and the remaining four genes (Cyt b, ATPase6, ATPase8, and ND6) lack a TGA (Figure 3). Three of the M. senile TGA codons are located at positions that are occupied by tryptophan-specifying codons in the corresponding genes of both mouse and D. yakuba. Four other M. senile TGA codons correspond in location to a tryptophan-specifying codon in either D. yakuba (three) or mouse (one) mt protein genes. The remaining M. senile TGA codons correspond in location to codons for five other amino acids in both the mouse and the D. yakuba genes, but in neither species do any of these amino acids have a higher frequency than tryptophan. These data, together with the occurrence in M. senile mtDNA of a gene for a predicted tRNA with a 5' UCA anticodon expected to recognize 5' UGA and 5' UGG codons, support the view that as in all other metazoan, protozoan, and some fungal mt genetic codes, TGA specifies tryptophan. As noted from analysis of part of the M. senile mt genome (PONT-KINGDON et al. 1994 ), in the entire M. senile mt genome, the use of TGA codons (12 plus two in the COI intron ORF) is much less than that of TGG codons (86 plus two in the COI intron ORF). The ratio of these codons (1:6.3) is inversely correlated to the overall frequency of other codons ending in A and G (1.7:1) in the 14 M. senile mt protein genes (Table 1). In mouse and D. yakuba mt protein genes, the TGA:TGG codon ratios (13.9:1 and 16.0:1, respectively) are closely correlated to the overall relative use of codons ending in A and G (14.6:1 and 15.7:1) (PONT-KINGDON et al. 1994 ). As the codon TGA is used to specify tryptophan in mt protein genes of protozoa, most fungi, and all other Metazoa so far examined, it seems more likely that the low number of TGA codons observed in M. senile mt protein genes represents an intermediate state in elimination of this codon rather than an increase in the use of this codon to specify tryptophan.

Analysis of AGA and AGG codon usage in the full complement of M. senile mt energy pathway protein genes (Figure 3) confirmed the conclusion drawn from analysis of the M. senile COII and Cyt b genes (PONT-KINGDON et al. 1994 ), that these codons specify arginine as in the standard genetic code rather than serine as in all other known invertebrate mt genetic codes [AGA and AGG codons specify glycine in an ascidian mt genetic code but do not specify an amino acid in vertebrate mtDNAs (see WOLSTENHOLME and FAURON 1995 )].

Analysis of ATA codon usage in the four most conserved M. senile mt protein genes (COI, COII, COIII, and Cyt b; Figure 3) also confirmed our earlier conclusion that these codons have the standard genetic code specification of isoleucine (PONT-KINGDON et al. 1994 ). This contrasts with the use of ATA codons to specify methionine in all other known metazoan mt genetic codes, except those of echinoderms and Turbellaria ( JACOBS et al. 1988 ; CANTATORI et al. 1989; BESSHO et al. 1992 ).

Segments of unassigned nucleotides:
Although the tRNA^f-Met and l-rRNA genes are immediately adjacent to each other (see below), segments of noncoding nucleotides occur between all other adjacent genes in the M. senile mtDNA molecule (Figure 1 and Figure 2). The largest such segment is 324 ntp between the tRNA^Trp and ND2 genes and, therefore, is the most likely candidate for the control region. The remaining 14 noncoding segments range from 4 to 143 ntp, and they total 614 ntp, the largest fraction (3.5%) of any known metazoan mt genome. Both the 143 ntp and 324 ntp noncoding segments have a higher A plus T content (72.0 and 70.4%, respectively) than the mt protein gene sense strands (62.5%), and RNAs transcribed from these regions in the same direction as the mt genes would have extensive, stable secondary structure potential (data not shown). This latter observation is of interest because it has been suggested that secondary structures in the mtDNA control region may play some part in the initiation of replication and transcription (see CHANG et al. 1985 ; CLARY and WOLSTENHOLME 1987 ; MONNEROT et al. 1990 ; HOFFMANN et al. 1992 ). Also, the possibility that secondary structure involving intergenic regions and protein gene termini in mtDNA transcripts may be important to transcript processing has been discussed extensively (BIBB et al. 1981 ; CLARY and WOLSTENHOLME 1985 ; OKIMOTO et al. 1992 ).

The M. senile mt-tRNA^Trp gene:
The 70 ntp sequence that begins 90 ntp downstream from the 3' end of the ND5 gene has primary structure and secondary structure potential of a gene for tRNA^Trp (Figure 1 and Figure 4). Almost all the nucleotides that are highly conserved among standard tRNAs (prokaryotic tRNAs and tRNAs encoded in eukaryotic nuclear and chloroplast DNAs) are found at specific positions within this structure: T₈, A₁₄, G₁₈, G₁₉, Pu₂₁, Py₃₂, T₃₃, Pu₃₇, Py₄₈, G₅₈, T₅₄, T₅₅, C₅₆, A₅₈, Py₆₀, and C₆₁ (RICH and RAJBHANDARY 1976 ; DIRHEIMER et al. 1979 ; SINGHAL and FALLIS 1979 ; SPRINZL et al. 1989 ). No other metazoan mt-tRNA^Trp gene so far reported has all of the three major features that are characteristic of standard tRNAs: G₁₈ and G₁₉ in the DHU loop, the G₅₃-C₆₁ pair at the distal end of the TC stem, and the complete 5' TTCRANY TC loop. Also, in the M. senile mt-tRNA^Trp gene sequence, there are a number of nucleotides that are considered semi-invariant in standard tRNAs: Pu₁₀, Pu₁₃, Pu₁₅, Pu₂₂, Py₂₅, Py₂₇, and Pu₄₃ (SINGHAL and FALLIS 1979 ). Therefore, the M. senile mt-tRNA^Trp gene more closely resembles the tRNA^Trp genes of prokaryotes and eukaryotic nuclear and chloroplast DNAs than it does the mt-tRNA genes of other metazoa. Interestingly, however, rather than having a Py₁₁-Pu₂₄, which is an invariant feature of prokaryotic and eukaryotic nuclear and chloroplast tRNA^Trp genes (and mt-tRNA^Trp genes of plants and fungi), the M. senile mt-tRNA^Trp gene has a Pu₁₁-Py₂₄ (G₁₁-C₂₄) pair that occurs in all known metazoan mt-tRNA^Trp genes (WOLSTENHOLME 1992A ).

View larger version (23K): [in this window] [in a new window] [Download PPT slide]   Figure 4. —The gene for mt-tRNATrp of M. senile is shown in the presumed secondary structure of the corresponding tRNA. The numbers shown follow the yeast tRNAPhe numbering system (SPRINZL et al. 1989 ) and, relative to the latter, they imply two nucleotides less in the dihydrouridine (DHU) arm and one nucleotide less in the variable loop.

The finding that the M. senile mt-tRNA^Trp gene as well as the mt-tRNA^f-Met gene have near-standard primary and secondary structures strengthens the view that almost all the various tRNA gene modifications found in metazoan mtDNAs postdated divergence of the cnidarian ancestral line.

The M. senile mt-l-rRNA gene:
The 5' terminus of the M. senile mt-l-rRNA gene was located using 5' RACE analysis. Comparisons of the nucleotide sequence of the RACE PCR amplification product with the mt-l-rRNA gene sequence (Figure 1) indicated that the 5' nucleotide could be any of nucleotides 71–73. This was because nucleotides 71 and 72 are Ts, which cannot be distinguished from the reiterated 3' poly-T sequence (the complement of the added 3' poly-A sequence). To precisely determine the mt-l-rRNA 5' end, a further 5' RACE analysis was carried out using a second 5' RACE pool comprising double-stranded cDNAs that were first 3' polyguanylated and then 3' polyadenylated. The result (Figure 5A) indicates that the 5' nucleotide is 71, a T (Figure 1); therefore, the 5' end of the mt-l-rRNA gene is immediately adjacent to the 3' end of the tRNA^f-Met gene.

View larger version (107K): [in this window] [in a new window] [Download PPT slide]   Figure 5. —Identification of the 5' and 3' ends of M. senile mt-l-rRNA by 5' and 3' RACE analyses. (A) Autoradiograph showing the nucleotide sequence of the amplified product of a 5' polyguanylated (by terminal deoxynucleotidyl transferase) cDNA of the mt-l-rRNA 5' end-proximal region. This sequence was generated by direct cycle sequencing of the double-stranded PCR product. Since the sequence is the transcript equivalent, the terminal reiterated nucleotide is C rather than G. The arrowhead indicates the 5' terminal nucleotide of the mt-l-rRNA gene. The sequence downstream from the poly-C sequence is written to the left. (B) An autoradiograph showing one of eight cloned sequences of the amplification product of cDNAs of the 3' end-proximal region of mt-l-rRNAs that have been 3' polyadenylated by yeast poly-A polymerase. The sequence shown is equivalent to the predicted mt-l-rRNA sense strand. The arrowheads indicate the 3'-terminal nucleotide of the mt-l-rRNA gene (Figure 2). The sequence upstream from the poly-A sequences is written to the right.

The 3' terminus of the M. senile mt-l-rRNA was located using 3' RACE analysis. In all of the resulting eight cloned sequences of the 3' end–proximal region of the mt-l-rRNA gene (Figure 5B), the position of the added poly-A sequence indicated that the 3'-terminal nucleotide of the mt-l-rRNA is nt 2260 (Figure 1). Therefore, 70 (apparently noncoding) nucleotides occur between the M. senile mt-l-r-RNA and COIII genes (Figure 1).

The aforementioned data define the size of the M. senile mt-l-rRNA as 2189 nt. This is smaller than the E. coli 23S rRNA by 715 nt but larger than any other presently known metazoan mt-l-rRNA. Examples of other predicted metazoan mt-l-rRNA sizes are Caenorhabditis elegans, 953 nt; D. yakuba, 1326 nt; X. laevis, 1640 nt; mouse, 1582 nt (WOLSTENHOLME 1992A ).

A secondary structure model for the M. senile l-rRNA based on the secondary structure model of E. coli 23S rRNA (BRIMACOMBE et al. 1990 ; GUTELL et al. 1993 ) is shown in Figure 6. As was found for the M. senile s-rRNA, with regard to secondary structure elements and size, the M. senile l-rRNA is closest to that of E. coli 23S rRNA than is any other metazoan mt-l-rRNA. In accordance with modeling of the E. coli 23S rRNA, the mt-l-rRNA model (Figure 6) is divided into 5' and 3' regions that collectively comprise six domains (I–VI), which in turn comprise helical structures numbered according to BRIMACOMBE et al. 1990 . From comparisons of the E. coli and M. senile models, it appears that there are a total of 16 specific segments of the E. coli sequence (shown as negative numbers of nucleotides) that are not present in the M. senile sequence (Figure 7). Five of these segments, comprising between 42 and 172 nt, each account for loss of most or all of multiple helices: helices 9 and 10, 15–21, 53–58, and 62 and 63. Two other segments each account for loss of a single helix and shortening of a second helix: helices 6 and 7, and 84 and 85. A further nine segments result in either shortening, reduction to a short loop, or complete loss of a single helix: helices 24, 34, 38, 39, 50, 52, 68, 78, and 79. Furthermore, relative to the E. coli sequence, the M. senile mt-l-rRNA sequence appears to have gained two segments, each of 9 nt, that result in addition of a helix (11A) in domain I and extension of a helix (87) in domain V. Also, it seems likely that secondary structure potential in domain VI favors the shortening of helices 96 and 97 and the inclusion of a helix (96A) not present in the E. coli 23S rRNA model. The sum of the nucleotides predicted to be lost (691 in a total of 16 locations) and gained (18 in two locations) is 673, which is very close to the difference in size of 687 nt between M. senile mt-l-rRNA and E. coli 23S rRNA. As noted for other metazoans (OKIMOTO et al. 1994 ), the losses in nucleotides in the M. senile mt-l-rRNA relative to E. coli 23S rRNA are considerably greater in the 5' region (425 nt) than in the 3' region (248 nt).

View larger version (69K): [in this window] [in a new window] [Download PPT slide]   Figure 6. —Secondary structure model of the M. senile mt-l-rRNA. The 5' and 3' regions are contained in A and B, respectively. The 5' and 3' ends were defined by RACE analyses (Figure 5). The sequence is numbered every 25 nt from the 5' end. Watson-Crick basepairings are indicated by a dash between nucleotides. Base pairing of G and U is indicated by G U, and presumed pairing of G and A is indicated by G A. The solid outlines identify either a sequence of 9 nt or more or shorter sequences in interhairpin regions that are 100% identical and in similar locations to sequences in the E. coli 23S rRNA secondary structure model (NOLLER et al. 1986 ; GUTELL et al. 1993 ). Roman numerals identify six major domains corresponding to the six domains of the E. coli 23S rRNA model (GUTELL et al. 1993 ). Bold numbers identify helixes that appear to have been conserved relative to those of the E. coli model (1–101, as defined by BRIMACOMBE et al. 1990 ). Bold numbers in parentheses indicate the equivalent locations of E. coli 23S rRNA helixes that clearly cannot be formed from the M. senile sequence delimited in each case by pairs of open-headed, bent-tailed arrows. The [59] indicates uncertainty as to whether the helix shown corresponds to helix 59 in the E. coli 23S rRNA model. The larger, bold negative numbers shown indicate differences in nucleotides to either the corresponding sequences delimited by the pairs of open-headed, bent-tailed arrows or (together with two positive numbers) to a single simple or complex hairpin structure. The two asterisks indicate two single hairpin structures (11A and 96A) not predicted in the E. coli 23S rRNA model.

	CONCLUDING REMARKS

TOP ABSTRACT MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUDING REMARKS LITERATURE CITED

The M. senile mt genome has two major unusual features: (1) the absence of all but two tRNA genes and (2) the presence of two group I introns, one in the COI gene and the other in the ND5 gene.

It seems likely that retention of genes for tRNA^f-Met and tRNA^Trp in the M. senile mt genome is related to the specific mt functions of these tRNAs. Initiation of protein synthesis in mammalian mitochondria has been shown to have the bacterial characteristic of using formyl-methionine rather than methione (CHOMYN et al. 1981 ). If this is a common feature of Metazoa, therefore, then there is a rationale for selection favoring the retention of the tRNA^f-Met gene in M. senile mtDNA. The presence of a small number of TGA codons in most of the M. senile mt protein genes provides a plausible explanation for the retention of a gene for a tRNA^Trp with a 5' UCA anticodon that is expected to recognize UGA and UGG codons.

A deficiency of genes in mtDNA for tRNAs necessary for protein synthesis in mitochondria has also been reported for the Protozoa Leishmania tarentollae, Trypanosoma brucei, Paramecium tetraurelia, Tetrahymena pyriformis, Acanthamoeba castellanii, and Chlamydomonas reinhardtii, as well as for some angiosperm plants, including maize (Zea mays), wheat (Triticum aestivum), and common bean (Phaseolus vulgaris). No tRNAs are encoded in mtDNAs of L. tarentollae and T. brucei (SIMPSON et al. 1989 ; HANCOCK and HAJDUK 1990 ; HANCOCK et al. 1992 ; SCHNEIDER et al. 1994A , SCHNEIDER et al. 1994B ). The respective numbers of mtDNA-encoded tRNAs in the other organisms are as follows: P. tetraurelia, three; T. pyriformis, eight; A. castellanii, 16; C. reinhardtii, three; and angiosperm plants, 16–20 (SUYAMA 1986 ; BOER and GRAY 1988 ; PRITCHARD et al. 1990 ; DIETRICH et al. 1992 ; MARECHAL-DROUARD et al. 1993 ; BURGER et al. 1995 ). Therefore, for protein synthesis to occur in the mitochondria of these organisms, tRNAs must be imported. It has been concluded from studies using two-dimensional electrophoresis and hybridization methods that there are 35–40 tRNAs present in L. tarentollae and T. brucei mitochondria, and that these tRNAs are nuclear DNA–encoded (SIMPSON et al. 1989 ; HANCOCK and HAJDUK 1990 ). Some tRNAs present in angiosperm plant mitochondria also appear to be transcribed from nuclear genes (DIETRICH et al. 1992 ; MARECHAL-DROUARD et al. 1993 ). Mitochondria of the yeast Saccharomyces cerevisiae contain a single tRNA for lysine that is encoded in nuclear DNA. Importation into mitochondria of this tRNA^Lys, of a nuclear DNA-encoded plant tRNA, and of a nuclear DNA–encoded T. brucei tRNA have been demonstrated directly (SMALL et al. 1992 ; TARASSOV and ENTELLIS 1992; SCHNEIDER et al. 1994B ).

It seems likely that in plants and T. brucei at least, some specific nuclear DNA–encoded tRNAs are used in both cytoplasmic and mt protein synthesis. In P. vulgaris, three different tRNAs in mitochondria, each specifying leucine, are identical in sequence to three leucine-specifying tRNAs in the cytoplasm, except that the G₁₈ of each mt-tRNA^Leu is methylated. Similarly, three tRNAs with corresponding specificity (lycine, leucine, and tyrosine) in the mitochondria and cytoplasm of T. brucei are identical, except that C₃₂ is modified in each mt-tRNA (SCHNEIDER et al. 1994A ). However, because modification occurs after entry of the tRNAs into mitochondria in each organism, modification is clearly not a mediator of importation. In S. cerevisiae, entry of the nuclear DNA–encoded tRNA^Lys into mitochondria requires a functional mt protein importation apparatus (TARASSOV et al. 1995A ), as well as the cooperation of the cytosolic and mt-lysyl-tRNA synthetases (TARASSOV et al. 1995B ).

The evolutionary origin of nuclear DNA–encoded genes for tRNAs postulated to be imported into mitochondria in cnidaria, protozoa, and plants is undetermined. In the past, such genes might have been transferred from mtDNA to nuclear DNA either by direct gene transfer or, as seems to have occurred for some plant mt protein genes, through an RNA intermediate (NUGENT and PALMER 1991 ). In any event, transfer must have been either accompanied or followed by loss of all or some mt-tRNA genes. An alternative explanation, supported in part by observations of sequence-identical tRNAs in mitochondria and cytoplasm of plants and T. brucei, mentioned above, is that all or most mt-tRNA genes are lost and their function has been taken over by imported cytoplasmic tRNAs. Our present knowledge of the number and specificity of different tRNAs found in plants and in T. brucei mitochondria indicates that in both cases, codon–anticodon interactions involved in protein translation are governed by standard wobble rules (G ·U pairing) rather than the extended wobble rules (superwobble U·N pairing) that seems to operate in noncnidarian metazoan mitochondria. Because of this, and the lack of necessity to postulate importation of tRNAs into mitochondria of any noncnidarian metazoan for which the entire mtDNA sequence is known, it seems likely that M. senile mt protein synthesis involves standard anticodon–codon interactions and, therefore, requires importation of at least 30 tRNAs. If this is the case, then in M. senile there would likely be the correlation of the largest known metazoan mt-rRNAs with a full complement of tRNAs in mitochondria that have standard primary and secondary structures.

Because the tRNAs encoded in all noncnidarian metazoan mtDNAs examined so far are sufficient to carry out mt translation, it seems most likely that the apparent replacement of mtDNA-encoded tRNAs with nuclear DNA-encoded tRNAs in M. senile mitochondria developed in an ancestral organism after divergence of the cnidarian line from the line leading to all other Metazoa. Therefore, because all extant Metazoa are considered to be monophyletic, it seems reasonable to argue that use of nuclear DNA–encoded tRNAs to carry out mt protein synthesis in cnidarian mitochondria was developed independently from the corresponding events in Protozoa and plants. The reported involvement of tRNA synthetases in tRNA^Lys importation into mitochondria in S. cerevisiae (TARASSOV et al. 1995B ) is in line with the expectation that a mechanism that has arisen multiple times in evolution makes use of a cellular system common to the different organisms.

The two group I introns present in the M. senile COI and ND5 genes are of particular interest for the following reasons. These are the only introns so far recorded in a metazoan mtDNA, and they are the only introns of the potentially self-splicing kinds (groups I, II, and III) to be found in Metazoa. Furthermore, we have shown that these introns appear to be limited in occurrence to members of the anthozoan order Actiniaria (BEAGLEY et al. 1996 ). Specifically, we were unable to detect either intron in the COI and ND5 genes of a total of five species representing four other orders of the anthozoan subclass Hexacorallia, of one species of each of two orders of the anthozoan subclass Octocorallia, and of two species of the class Hydrozoa. Because of this limited distribution of the COI and ND5 introns, we have given serious consideration to the possibility that, rather than being inherited from an ancestral cnidarian, these introns were acquired by an early actinarian from organisms ancestral to Zooxanthellae or Zoochlorellae, endosymbionts commonly found in present-day sea anemones and other cnidaria (BEAGLEY et al. 1996 ).

The amino acid sequence of the protein encoded by the COI intron ORF includes a consensus signature sequence (RLAGLVDGEGVF) of the LAGLI-DADG family of group I intron-encoded, site-specific endonucleases designated homing endonucleases (LAMBOWITZ and BELFORT 1993 ; DOOLITTLE 1993 ). In S. cerevisiae, it has been demonstrated that mating of a cell possessing a gene harboring the homing endonuclease–containing intron with a cell possessing an intronless allele results in a copy of the entire intron being transferred to the intronless allele, and that the process is mediated by the homing endonuclease (COLLEAUX et al. 1988 ; SARGUEIL et al. 1991 ). Therefore, the endonuclease-containing intron ensures its own propagation. In most other Metazoa examined, mtDNA inheritance is strictly maternal (BUZZO et al. 1978 ; FAURON and WOLSTENHOLME 1980 ). If mtDNA inheritance in cnidaria is also maternal, then transfer of the intron between mitochondria originating from different organisms would not be expected, and the most likely function of the COI intron ORF protein would be to reintroduce the intron into COI alleles from which the intron has been lost. If, as must occur in S. cerevisiae and has been demonstrated in mammals (SWIFT and WOLSTENHOLME 1969 ; WALLACE 1987 ; HAYASHI et al. 1994 ), there is fusion of mitochondria in cnidarian cells, then intron reintroduction could occur between intron-containing and intron-deficient alleles originating from different mitochondria within one cell. This proposed function for the M. senile COI intron ORF protein, however, does not address the question of what selective advantage is conveyed by this protein (or the catalytic core of the COI intron) to the COI genes that contain it.

In contrast, maintenance of the intron in the M. senile ND5 gene, which does not encode a homing endonuclease, is ensured by the inclusion of ND1 and ND3 genes. Because these essential genes do not appear to be duplicated in the cell, the loss of the entire ND5 intron would be lethal. Therefore, elimination of the ND5 intron could only be accomplished by a multistep event, the most likely of which is transposition (directly or after duplication) of the ND1 and ND3 genes (as a unit or separately) to another nonintron location in the M. senile genome, followed by intron loss.

The above considerations of M. senile COI and ND5 introns provide a clear-cut case of two introns that occur in the same small genome, whose continued presence seems likely to be ensured by different selective forces.

	FOOTNOTES

¹ Present address: Ronald Okimoto, Department of Poultry Science, University of Arkansas, Fayetteville, AR 72701.

	ACKNOWLEDGMENTS

We thank JANE L. MACFARLANE for assistance at various times. This work was supported by National Institutes of Health grant GM-18375

Manuscript received July 3, 1997; Accepted for publication November 14, 1997.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUDING REMARKS LITERATURE CITED

ARNASON, U., X. XU, and A. GULLBERG, 1996  Comparison between the complete mitochondrial DNA sequences of Homo and the common chimpanzee based on nonchimeric sequences. J. Mol. Evol. 42:145-152[Medline].

ASAKAWA, S., H. HIMENO, K.-I. MIURA, and K. WATANABE, 1995  Nucleotide sequence and gene organization of the starfish Asterina pectinifera mitochondrial genome. Genetics 140:1047-1060[Abstract].

BARRELL, B. G., A. T. BANKIER, and J. DROUIN, 1979  A different genetic code in human mitochondria. Nature 282:189-194[Medline].

BARRELL, B. G., S. ANDERSON, A. T. BANKIER, M. H. L. DE BRUIJN, and E. CHEN et al., 1980  Different pattern of codon recognition by mammalian mitochondrial tRNAs. Proc. Natl. Acad. Sci. USA 77:3164-3166[Abstract/Free Full Text].

BEAGLEY, C. T., J. L. MACFARLANE, G. A. PONT-KINGDON, R. OKIMOTO, N. A. OKADA et al., 1995 Mitochondrial genomes of Anthozoa (Cnidaria), pp. 149–153 in Progress in Cell Research—Symposium on Thirty Years of Progress in Mitochondrial Bioenergetics and Molecular Biology, edited by F. PALMIERI, S. PAPA, C. SACCONE and N. GADALETA. Elsevier, Amsterdam.

BEAGLEY, C. T., N. A. OKADA, and D. R. WOLSTENHOLME, 1996  Two mitochondrial group I introns in a metazoan, the sea anemone Metridium senile : one intron contains genes for subunits 1 and 3 of NADH dehydrogenase. Proc. Natl. Acad. Sci. USA 93:5619-5623[Abstract/Free Full Text].

BESSHO, Y., T. OHAMA, and S. OSAWA, 1992  Planarian mitochondria. II. The unique genetic code as deduced from cytochrome c oxidase subunit gene sequences. J. Mol. Evol. 34:331-335[Medline].

BIBB, M. H., R. A. VAN ETTEN, C. T. WRIGHT, M. W. WALBERG, and D. A. CLAYTON, 1981  Sequence and gene organization of mouse mitochondrial DNA. Cell 26:167-180[Medline].

BOER, P. H. and M. W. GRAY, 1988  Transfer RNA genes and the genetic code in Chlamydomonas reinhardii.. Curr. Genet. 14:583-590[Medline].

BOORE, J. L. and W. M. BROWN, 1994  Complete DNA sequence of the mitochondrial genome of the black chiton Katharina tunicata.. Genetics 138:423-443[Abstract].

BOORE, J. L. and W. M. BROWN, 1995  Complete sequence of the mitochondrial DNA of the annelid worm Lumbricus terrestris.. Genetics 141:305-319[Abstract].

BRIDGE, D., C. W. CUNNINGHAM, B. SCHIERWATER, R. DESALLE, and L. W. BUSS, 1992  Class level relationships in the phylum Cnidaria: evidence from mitochondrial genome structure. Proc. Natl. Acad. Sci. USA 89:8750-8753[Abstract/Free Full Text].

BRIMACOMBE, R., B. GREUER, P. MITCHELL, M. OBWALD, J. RINKE-APPEL et al., 1990 Three dimensional structure and function of Escherchia coli 16S and 23S rRNA as studied by crosslinking techniques, pp. 93–106 in The Ribosome: Structure, Function and Evolution, edited by W. E. HILL, A. DAHLBERG, R. A. GARRETT, P. B. MOORE, D. SCHLESSINGER and J. R. WARNER. ASM Press, Washington, DC.

BURGER, G., I. PLANTE, K. M. LONERGAN, and M. W. GRAY, 1995  The mitochondrial DNA of the amoeboid protozoon Acanthamoeba castellanii: Complete sequence, gene content and genome organization. J. Mol. Biol. 245:522-537[Medline].

BUZZO, K., D. L. FOUTS, and D. R. WOLSTENHOLME, 1978  EcoRI cleavage site variants of mitochondrial DNA molecules from rats. Proc. Natl. Acad. Sci. USA 75:909-913[Abstract/Free Full Text].

CANTATORE, P., M. ROBERTI, G. RAINALDI, M. N. GADALETA, and C. SACCONE, 1989  The complete nucleotide sequence, gene organization, and genetic code of the mitochondrial genome of Paracentrotus lividus.. J. Biol. Chem. 264:10965-10975[Abstract/Free Full Text].

CHANG, D. D., T. W. WONG, J. E. HIXON and D. A. CLAYTON, 1985 Regulatory sequences for mammalian mitochondrial transcription and replication, pp. 135–144 in Achievements and Perspectives of Mitochondrial Research, Vol. II. Biogenesis edited by E. QUAGLIARIELLO, E. C. SLATER, F. PALMIERI, C. SACCONE and A. M. KROON. Elsevier, Amsterdam.

CHOMYN, A., M. W. HUNKAPILLAR, and G. ATTARDI, 1981  Alignment of the amino terminal amino acid sequence of human cytochrome c oxidase subunits I and II with the sequence of their putative mRNAs. Nucleic Acids Res. 9:867-877[Abstract/Free Full Text].

CLARY, D. O. and D. R. WOLSTENHOLME, 1985  The mitochondrial DNA molecule of Drosophila yakuba: nucleotide sequence, gene organization, and genetic code. J. Mol. Evol. 22:252-271[Medline].

CLARY, D. O. and D. R. WOLSTENHOLME, 1987  Drosophila mitochondrial DNA: conserved sequences in the A+T-rich region and supporting evidence for a secondary structure model of the small ribosomal RNA. J. Mol. Evol. 25:116-125[Medline].

CLAYTON, D. A., 1991  Replication and transcription of vertebrate mitochondrial DNA. Ann. Rev. Cell Biol. 7:453-478.

CLAYTON, D. A., 1992 Transcription and replication of animal mitochondrial DNAs, pp. 217–232 in Mitochondrial Genomes: International Review of Cytology, Vol. 141, edited by D. R. WOLSTENHOLME and K. W. JEON. Academic Press, New York.

COLLEAUX, L., L. D'AURIOL, F. GALIBERT, and B. DUJON, 1988  Recognition and cleavage site of the intron-encoded omega transposase. Proc. Natl. Acad. Sci. USA 85:6022-6026[Abstract/Free Full Text].

CROZIER, R. H. and Y. C. CROZIER, 1993  The mitochondrial genome of the honey bee Apis melifera: complete sequence and genome organization. Genetics 133:97-117[Abstract].

DALE, R. M. K., B. MCCLURE, and J. P. HOUCHINS, 1985  A rapid single-stranded cloning strategy for producing overlapping clones for use in DNA sequencing: application to sequencing the corn mitochondrial 18S rDNA. Plasmid 13:31-40[Medline].

DAVIS, L. G., M. D. DIBNER and J. F. BALTEY, 1986 Basic Methods in Molecular Biology. Elsevier, New York.

DE BRUIJN, M. H. L., P. H. SCHREIER, I. C. EPERON, B. G. BARRELL, and E. Y. CHEN et al., 1980  A mammalian mitochondrial serine transfer RNA lacking the "dihydrouridine" loop and stem. Nucleic Acids Res. 8:5213-5222[Abstract/Free Full Text].

DIETRICH, A., J. H. WEIL, and L. MARÉCHAL-DROUARD, 1992  Nuclear-encoded transfer RNAs in plant mitochondria. Annu. Rev. Cell Biol. 8:115-131.

DIRHEIMER, G., G. KEITH, A.-P. SIBLER and R. P. MARTIN, 1979 The primary structure of tRNAs and their rare nucleosides, pp. 19–41 in Transfer RNA: Structure, Properties and Recognition, edited by P. R. SCHIMMEL, D. SOLL and J. N. ABELSON. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

DOOLITTLE, R. F., 1993  The comings and goings of homing endonucleases and mobile introns. Proc. Natl. Acad. Sci. USA 90:5379-5381[Free Full Text].

FAURON, C. M.-R. and D. R. WOLSTENHOLME, 1980  Extensive diversity among Drosophila species with respect to nucleotide sequences within the adenine + thymine-rich region of mitochondrial DNA molecules. Nucleic Acids Res. 8:2439-2452[Abstract/Free Full Text].

FLOOK, P. K., C. H. F. ROWELL, and G. GELLISSEN, 1995  The sequence, organization, and evolution of the Locusta migratoria mitochondrial genome. J. Mol. Evol. 41:928-941[Medline].

FROHMAN, M. A., 1990 RACE: rapid amplification of cDNA ends, pp. 28–38 in PCR Protocols, A Guide to Methods and Applications, edited by M. A. INNIS, D. H. GELFAND, J. J. SNINSKY and T. J. WHITE. Academic Press, San Diego.

GRAY, M. W., D. SANKOFF, and R. J. CEDERGREN, 1984  On the evolutionary descent of organisms and organelles: A global phylogeny based on a highly conserved structural core in small subunit ribosomal RNA. Nucleic Acids Res. 12:5837-5852[Abstract/Free Full Text].

GUTELL, R. and G. E. FOX, 1988  A compilation of large subunit RNA sequences presented in a structural format. Nucleic Acids Res. 16:r175-r269.

GUTELL, R. R., M. W. GRAY, and M. N. SCHNARE, 1993  A compilation of large subunit (23S and 23S-like) ribosomal RNA structures. Nucleic Acids Res. 21:3055-3074[Free Full Text].

HANCOCK, K. and S. L. HAJDUK, 1990  The mitochondrial tRNAs of Trypanosoma brucei are nuclear encoded. J. Biol. Chem. 265:19208-19215[Abstract/Free Full Text].

HANCOCK, K., A. J. LEBLANC, D. DONZE, and S. L. HAJDUK, 1992  Identification of nuclear encoded precursor tRNAs within the mitochondrion of Trypanosoma brucei.. J. Biol. Chem. 267:23963-23971[Abstract/Free Full Text].

HATZOGLOU, E., G. C. RODAKIS, and R. LECANIDOU, 1995  Complete sequence and gene organization of the mitochondrial genome of the land snail Albinaria coerulea.. Genetics 140:1353-1366[Abstract].

HAYASHI, J.-I., M. TAKEMITSU, Y.-I. GOTO, and I. NONAKA, 1994  Human mitochondria and mitochondrial genome function as a single dynamic cellular unit. J. Cell Biol. 125:43-50[Abstract/Free Full Text].

HOFFMANN, R. J., J. L. BOORE, and W. M. BROWN, 1992  A novel mitochondrial genome organization for the blue mussel, Mytilus edulis.. Genetics 131:397-412[Abstract].

JACOBS, H. T., D. J. ELLIOT, V. B. MATH, and A. FARQUHARSON, 1988  Nucleotide sequence and gene organization of sea urchin mitochondrial DNA. J. Mol. Biol. 202:185-217[Medline].

KRETTEK, A., A. GULLBERG, and U. ARNASON, 1995  Sequence analysis of the complete mitochondrial DNA molecule of the hedgehog, Erinaceus europaeus, and the phylogenetic position of the Lipotyphla. J. Mol. Evol. 441:952-957.

LAMBOWITZ, A. M. and M. BELFORT, 1993  Introns as mobile genetic elements. Annu. Rev. Biochem. 62:587-622[Medline].

MARÉCHAL-DROUARD, L., J. H. WEIL, and A. DIETRICH, 1993  Transfer RNAs and transfer RNA genes in plants. Annu. Rev. Plant Physiol. Plant Mol. Biol. 44:13-32.

MONNEROT, M., M. SOLIGNAC, and D. R. WOLSTENHOLME, 1990  Discrepancy in divergence of the mitochondrial and nuclear genes of Drosophila teissieri and Drosophila yakuba.. J. Mol. Evol. 30:500-508[Medline].

NOLLER, H. F., M. ASIRE, A. BARTA, S. DOUTHWAITE, T. GOLDSTEIN et al., 1986 Studies on the structure and function of ribosomal RNAs, pp. 141–163 in Structure, Function and Genetics of Ribosomes, edited by B. HARDESTY and G. KRAMER (Springer Series in Molecular Biology, edited by A. RICH). Springer-Verlag, New York.

NUGENT, J. M. and J. D. PALMER, 1991  RNA-mediated transfer of the gene coxII from the mitochondrion to the nucleus during flowering plant evolution. Cell 66:473-481[Medline].

OJALA, D., J. MONTOYA, and G. ATTARDI, 1981  RNA punctuation model of RNA processing in human mitochondria. Nature 290:470-474[Medline].

OKIMOTO, R. and D.R. WOLSTENHOLME, 1990  A set of tRNAs that lack either the TC arm or the dihydrouridine arm: towards a minimal tRNA adaptor. EMBO J. 9:3405-3411[Medline].

OKIMOTO, R., H. M. C