Complete Sequence of the Mitochondrial DNA of the Annelid Worm Lumbricus terrestrk Jeffrey L . Boore ' and Wesley

We have determined the complete nucleotide (nt) sequence of the mitochondrial genome of an oligochaete annelid, the earthworm Lumbricus tarestris. This genome contains the 37 genes typical of metazoan mitochondrial DNA (mtDNA), including ATPase8, which is missing from some invertebrate mtDNAs. ATPase8 is not immediately upstream of ATPase6, a condition found previously only in the mtDNA of snails. All genes are transcribed from the same DNA strand. The largest noncoding region is 384 nt and is characterized by several homopolymer runs, a tract of alternating TA pairs, and potential secondary structures. All proteinencoding genes either overlap the adjacent downstream gene or end at an abbreviated stop codon. In Lumbricus mitochondria, the variation of the genetic code that is typical of most invertebrate mitochondrial genomes is used. Only the codon ATG is used for translation initiation. Lumbricus mtDNA is A + T rich, which appears to affect the codon usage pattern. The DHU arm appears to be unpaired not only in tRNAse'(AGN), as is typical for metazoans, but perhaps also in , a condition found previously only in a chiton and among nematodes. Relating the Lumbricus gene organization to those of other major protostome groups requires numerous rearrangements. t m ~ ( U C N ) T HE mitochondrial genome of multicellular animals consists of a closed circular DNA molecule except among some cnidarians, where it consists of one or two linear molecules (WARRIOR and GALL 1985; BRIDGE et al. 1992). Its usual size range is from 14 to 17 kb, but variations as large as 40 kb are known (MORITZ et al. 1987; SNYDER et al. 1987; WOLSTENHOLME 1992). Despite this size variation, there is little variation in gene content, since in all cases analyzed the larger size is due either to variation in the length of a noncoding region (BROWN 1985; HARRISON 1989; ~ O C H E et al. 1990) or to iteration of some portion of the mitochondrial DNA (mtDNA) (MORITZ and BROWN 1986, 1987; WALLIS 1987; ZEVERING et al. 1991; AZEVEDO and HYMAN 1993; FULLER and Z o u R o s 1993; STANTON et al. 1994). Metazoan mtDNAs ordinarily contain 36 or 37 genes: 2 for ribosomal RNAs (+rRNA and 1-rRNA), 22 for tRNAs and 13 for subunits of multimeric proteins of the inner mitochondrial membrane [cytochrome oxidase subunits 1-111 (COI-III), cytochrome b apoenzyme (Cytb), ATP synthase subunits 6 and (usually) 8 (ATPase6 and ATPases) and NADH dehydrogenase subunits 1-6 and 4L (ND1-6, ND4L)l. In addition, there is usually at least one sequence of variable length that does not encode any gene. In vertebrates (MONTOYA et Corresponding author: Wesley M. Brown, Department of Biology, University of Michigan, 830 N. University Ave., Ann Arbor, MI 481091048. E-mail: [email protected] ' Present address: Department of Cell Biology and Neuroanatomy, University of Minnesota, 4135 Jackson Hall, 321 Church St. SE, Minneapolis, MN 55455. Genetics 141: 305-319 (September, 1995) al. 1982; BOGENHAGEN et al. 1985; KING and Low 1987; FORAN et al. 1988; CLAY~ON 1991, 1992) and insects (CLARY and WOLSTENHOLME 1985a) these noncoding sequences are known to include elements that regulate and initiate mtDNA replication and transcription. Complete mitochondrial gene arrangements have been published for 28 metazoans: 14 vertebrates (human, ANDERSON et al. 1981; domestic mouse, BIBB et al. 1981; cow, ANDERSON et al. 1982; rat, GADALETA et al. 1989; opossum, JANKE et al. 1994; frog, ROE et al. 1985; carp, CHANG et al. 1994; cod, JOHANSEN et al. 1990; loach, TZENG et al. 1992; chicken, DESJARDINS and MORAIS 1990; harbor seal, ARNASON and JOHNSSON 1992; blue whale, ARNASON and GULLBERG 1993; fin whale, ARNASON et ai. 1991; sea lamprey, LEE and KOCHER 1995), 3 echinoderms [sea urchin (Stronglyocentrotus), JACOBS et al. 1988a; sea urchin (Paracentrotus), CANTATORE et al. 1989; sea star, ASAKAWA et al. 19951, 5 arthropods (Drosophila yakuba, CLARY and WOLSTENHOLME 1985a; D. mehnogaster, GARESSE 1988; honeybee, CROZIER and CROZIER 1993; mosquito, MITCHELL et al. 1993; ktemia, VALVERDE et al. 1994), 2 mollusks (blue mussel, HOFEMANN et al. 1992; black chiton, BOORE and BROWN 1994a), 3 nematodes (Ascaris, WOLSTENHOLME et al. 1987; Caenorhabditis, OKIMOTO et al. 1992; Meloidogyne, OKIMOTO et al. 1991) and 1 cnidarian (Metridium, WOLSTENHOLME 1992). Comparisons among these and among many other partially determined metazoan mitochondrial gene arrangements suggest several generalities (WOLSTENHOLME 1992; BOORE and BROWN 199413): (1) gene content is nearly unvarying, (2) gene 306 J. L. Boore and W. M. Brown arrangements generally remain unchanged over long periods of evolutionary time, (3) stability of gene arrangements is best explained by a lack of genetic recombination and by the infrequency of rearrangements that maintain functional genomes, rather than by invoking selection for or against any particular gene order, (4) the great number of potential gene arrangements makes it very unlikely that different axa would independently arrive at identical arrangements, and (5) with a few possible exceptions, gene arrangements seem relatively stable within major groups but variable between them. These characteristics make the comparison of mitochondrial gene arrangements a potentially powerful means for inferring phylogenetic relationships among major metazoan groups (BROWN 1985; MORITZ et al. 1987;JACOBS et al. 1988b; SMITH et al. 1993; BOORE and BROWN 1994b; BOORE et aZ. 1995). In addition, many features of the molecular biology of mtDNA have been shown to vary and can be compared among metazoans (alternative translation start codons, unusual tRNA and rRNA structures, genetic code variations, features of mtDNA replication and transcription, etc.) (see BROWN 1985; WOLSTENHOLME 1992; BOORE and BROWN 1994b). While comparisons of these and other aspects of the nuclear genome are possible, they are currently more accessible from the much smaller and simpler mitochondrial genome. As the data set expands, we anticipate that mitochondrial gene arrangements, as well as many other aspects of the molecular biology of mtDNA, will contribute importantly to our knowledge of evolutionary relationships among many major metazoan groups. To this end, we have determined the gene arrangement and DNA sequence of the mitochondrial genome of Lumbricus terrestris, the common earthworm, the first annelid to be so characterized. We outline the notable features of this genome and compare them in detail to those determined for representatives of the other two major protostome phyla, Arthropoda and Mollusca. MATERIALS AND METHODS Mitochondrial DNA was isolated from the common earthworm, L. tmestris, and purified by cesium chloride-ethidium bromide centrifugation. A detailed restriction endonuclease cleavage map was constructed for this mtDNA using eight enzymes. DNA fragments from the enzymatic digests were resolved electrophoretically in both 1% agarose and 3.5% acrylamide gels, which allows visualization and accurate size estimation of fragments from a few bp to -15 kb. The methods we used to clone and sequence mtDNA and to identify individual mitochondrial genes are described in Boom and BROWN (1994a). The entire Lumbricus mtDNA was cloned by linearizing it at its unique BclI site at position 5900 (see Figure 1 and APPENDIX) and inserting it as a unit into the BamHI site of bacteriophage lambda EMBL3 DNA (BamHI cleavage leaves a BclI-compatible overhang). BamHI and SulI codigestion produces five DNA fragments from the mtDNA portion of this recombinant phage, of -8.0, 4.3, 1.0, 0.93 and 0.76 kb. Since ligation of the compatible BclI and BamHI ends of the phage fails to regenerate either site, the terminal portions of the mtDNA insert (0.93 and 1.0 kb) were liberated by cleaving the SalI sites that flank the BamHI site of the vector. Each of these fragments was cloned into pBluescript as in BOORE and BROWN (1994a). Cleavage maps of the recombinant plasmids, designated L8.0, L4.3, L1.O, L0.93 and L0.76, respectively, were constructed for 20 other restriction enzymes. Further subcloning yielded a total of 14 additional clones (Figure 1). To verify further that no portion of the mtDNA was lost in cloning, a fragment of -500 bp spanning the BclI cloning site was amplified from Lumbricus mtDNA by PCR, using primers designed on the basis of the flanking sequences. The sequence of the PCR fragment was determined and verified as identical to that obtained from fusing the sequence of the two flanking plasmid clones, For PCR amplification, we employed 30 cycles of (30 sec at 94", 1 min at 55" and 2 min at 72"). Single-stranded templates for sequencing were generated by using unequal ratios of the amplification primers. Open reading frames (ORFs) were determined using the program MacVectoF (IBI, version 3.0) and the genetic code for invertebrate mtDNA. ORFs were identified based on the similarity of their inferred amino acid sequences to those of the corresponding mitochondrial genes of D. yakuba (CLARY and WOLSTENHOLME 1985a) and Kathan'na tunicata (BOORE and BROWN 1994a). In the case of h!D4L, ND6 and ATPase& identity was further confirmed by a comparison of hydrophilicity profiles (KITE and DOOLITTLE 1982) (see Figure 3). Transfer RNA genes were identified generically by their potential to be folded into characteristic mitochondrial tRNA structures and specifically by their anticodon sequences. Ribosomal RNAs were identified by their similarity in sequence and potential secondary structure to the mitochondrial rRNAs of D. yakuba (CLARY and WOLSTENHOLME 1985b). Sequence of the L. tmestn's mitochondrial genome has been deposited in GenBank under accession number U24570. RESULTS AND DISCUSSION Genome composition: The mtDNA of the annelid worm L. terrestris is 14,998 bp in size. It contains no introns and, in keeping with its small size, intergenic regions are few and short; most of its genes abut directly or overlap. Lumbricus mtDNA is 61.6% A + T, at the low end of the range for invertebrates. By contrast, mosquito (MITCHELL et al. 1993) and Drosophila (CLARY and WOLSTENHOLME 1985a) mtDNAs are -77% A + T, Apis mtDNA is 84.9% A + T (CROZIER and CROZIER 1993) and Katharina mtDNA is 69% A + T (BOORE and BROWN 1994a). A + T richness is similar to the sequenced portions of Mytilus mtDNA (62%) (HOFFMANN et aZ. 1992). Both the coding and noncoding DNA strands have approximately equal proportions of purines and pyrimidines (0.087 on a scale where 0.0 is an equal ratio of pyrimidines and purines and 1.0 represents all purines or all pyrimidines on one strand). Gene content and organization: Lumbricus mtDNA contains genes for 13 proteins (ND1-6, ND4L, COI111, ATPase6, ATPase8 and Cytb) for 2 rRNAs (sand Annelid Mitochondrial Genome Lumbricus terrestris 307 Y 2 ND4 ND2 ND3 ND1 I r m E L COI ND5 J A6 Cytb ND6 Call A8 COll z ih