Complete Sequences of the Highly Rearranged Molluscan Mitochondrial Genomes of the Scaphopod Graptacme eborea and the Bivalve Mytilus edulis
http://www.100md.com
分子生物学进展 2004年第8期
* Department of Biology, University of Michigan, Ann Arbor
DOE Joint Genome Institute and Lawrence Berkeley National Laboratory, Walnut Creek, California
E-mail: jlboore@lbl.gov.
Abstract
We have determined the complete sequence of the mitochondrial genome of the scaphopod mollusk Graptacme eborea (14,492 nts) and completed the sequence of the mitochondrial genome of the bivalve mollusk Mytilus edulis (16,740 nts). (The name Graptacme eborea is a revision of the species formerly known as Dentalium eboreum.) G. eborea mtDNA contains the 37 genes that are typically found and has the genes divided about evenly between the two strands, but M. edulis contains an extra trnM and is missing atp8, and it has all genes on the same strand. Each has a highly rearranged gene order relative to each other and to all other studied mtDNAs. G. eborea mtDNA has almost no strand skew, but the coding strand of M. edulis mtDNA is very rich in G and T. This is reflected in differential codon usage patterns and even in amino acid compositions. G. eborea mtDNA has fewer noncoding nucleotides than any other mtDNA studied to date, with the largest noncoding region only 24 nt long. Phylogenetic analysis using 2,420 aligned amino acid positions of concatenated proteins weakly supports an association of the scaphopod with gastropods to the exclusion of Bivalvia, Cephalopoda, and Polyplacophora, but it is generally unable to convincingly resolve the relationships among major groups of the Lophotrochozoa, in contrast to the good resolution seen for several other major metazoan groups.
Key Words: scaphopod ? bivalve ? mollusk ? mitochondria ? evolution ? genome
Introduction
Complete mitochondrial (mt) DNA sequences have been reported for nine mollusks representing four classes: the polyplacophoran Katharina tunicata (Boore and Brown 1994a, 1994b); the cephalopod Loligo bleekeri (Tomita et al. 2002); the bivalves Crassostrea gigas (Kim, Je, and Park 1999), Venerupis (Ruditapes) philippinarum (Okazaki and Ueshima 2002), and Inversidens japanensis (incomplete by only a small fragment of apparently noncoding sequence; Okazaki and Ueshima 2001); and the gastropods Cepaea nemoralis (Terrett, Miles, and Thomas 1996), Albinaria coerulea (Hatzoglou, Rodakis, and Lecanidou 1995), Roboastra europaea (Grande et al. 2002), and Pupa strigosa (Kurabayashi and Ueshima 2000a). Additionally, enough of the mtDNA sequence of another gastropod, Euhadra herkotsi (Yamazaki et al. 1997), and of the bivalve Mytilus edulis (F-type, see below) (Hoffmann, Boore, and Brown 1992) have been determined to assess the arrangement of all encoded genes. Like the mitochondrial genomes of nearly all metazoan animals (see Boore 1999), these are small, circular DNA molecules containing almost universally the same 37 genes: 13 for protein subunits of oxidative phosphorylation enzymes (cox1-3, nad1-6, nad4L, cob, atp6, atp8), two for mitochondrial ribosomal RNAs (rrnS and rrnL), and 22 for the tRNAs necessary to translate these 13 proteins (designated by trnX, with X being the one-letter code for the corresponding amino acid, and the two for each of serine and for leucine differentiated by numeral). M. edulis varies from this by the loss of atp8 and the gain of a second trnM. M. edulis is unusual also in maintaining two different mtDNAs, one called F-type and one M-type (Hoeh, Blakley, and Brown 1991), transmitted by an unusual system dubbed "doubly-uniparental inheritance" (Stewart et al. 1995), which now appears to be widespread in bivalves (Passamonti and Scali 2001; Hoeh, Stewart, and Guttman 2002). Partial mtDNA sequences are also available for the bivalve M. californianus (also F-type; Beagley, Okimoto, and Wolstenholme 1999) and the gastropods Albinaria turrita (Lecanidou, Douris, and Rodakis 1994), Omalogyra atomus (Kurabayashi and Ueshima 2000b), Littorina saxatilis (Wilding, Mill, and Grahame 1999) and several vermetids (Rawlings, Collins, and Bieler 2001).
For some phyla of animals, mitochondrial gene arrangements seem seldom to have changed. For example, with few notable exceptions, those vertebrates studied have identical gene arrangements, as do most studied arthropods. (In the latter case, exceptions include the highly rearranged mtDNAs found for wallaby louse [Shao, Campbell, and Barker 2001], hermit crab [Hickerson and Cunningham 2000], and metastriate ticks [Black and Roehrdanz 1998; Campbell and Barker 1998].) Mollusks differ, with many gene rearrangements noted for the molluscan taxa listed above. High levels of rearrangement have also been noted for nematodes and brachiopods (see Boore 1999, 2002). Gene rearrangements have been shown to be very powerful characters for reconstructing evolutionary relationships (see Boore and Brown 1998), and the rapidity of rearrangement within a lineage determines the level at which rearrangements are likely to be phylogenetically informative.
The phylogenetic relationships among the different extant molluscan classes are not well established, and anatomical studies have proposed multiple alternatives to this issue. A common proposal is a gradist scenario where chitons (Polyplacophora) and solenogasters and caudofoveates (Aplacophora) are the basal lineages to a grade of valve-bearing taxa (Gastropoda, Bivalvia, Cephalopoda, Monoplacophora, and Scaphopoda), collectively known as the Conchifera (Salvini-Plawen 1985; Salvini-Plawen and Steiner 1996; Haszprunar 2000) (fig. 1A). Some propose that chitons and aplacophorans form a monophyletic clade rather than a grade (Scheltema 1993, 1996), and some view Conchifera as split into a cephalopod/gastropod clade and a scaphopod/bivalve clade with monoplacophorans as the basal conchiferan lineage (Runnegar and Pojeta 1974) (each as in fig. 1B). Recent evaluations of morphological and paleontological data (Waller 1998) as well as 18S rRNA sequences (Steiner and Dreyer 2002) alternatively conclude that Scaphopoda is the sister group to Cephalopoda (fig. 1C). The many ribosomal RNA sequences have so far only poorly resolved molluscan phylogeny, rendering some taxa paraphyletic (e.g., bivalves; Steiner and Müller 1996), making it difficult to assess whether the proposed anatomical interpretations are identifying true synapomorphies (Steiner and Müller 1996; Steiner and Hammer 2000). Consequently, we are in need of additional characters that can help address phylogenetic relationships among major molluscan lineages. As the mitochondrial genome database continues to grow, we will be able to incorporate both gene order and sequence data into this analysis. Here we present the first complete mitochondrial genome sequence for a member of the Scaphopoda (Graptacme eborea), a previously unsampled class, and the completed sequence of the F-type mitochondrial genome of the bivalve Mytilus edulis.
FIG. 1. A comparison of the most commonly held views on the evolutionary relationships of molluscan classes
Materials and Methods
Determining the mtDNA Sequence of Graptacme eborea
A preparation of total DNA prepared from G. eborea was the gift of K. Fraser and T. Collins. PCR amplification of a portion of cox1 with the LCO1490 and HCO2198 oligonucleotides (Folmer et al. 1994) employed an initial heating at 97°C for 5 min, a hot start by adding Taq polymerase (Fisher) at 72°C, then 36 cycles of: (94°C, 1 min; 45°C, 1 min; 72°C, 2 min), followed by incubation at 72°C for 15 min. Reaction volume was 25 μl, which then yielded a single band of approximately 710 nts on a 1% agarose gel stained with ethidium bromide. This amplification product was gel purified using Gene Clean (Bio 101) and ligated into pBluescript (Stratagene), which had been prepared by: digestion with EcoRV, tailing on each end with a single T using Taq polymerase and dTTP, then gel purification using Gene Clean (Bio 101). DNA was prepared from recombinant plasmids using alkaline lysis followed by organic extraction and ethanol precipitation (Sambrook, Fritsch, and Maniatis 1989). DNA sequence was determined using AmplitaqFS, dye-terminator cycle sequencing (Perkin-Elmer) according to supplier's instructions and an ABI 377 Automated DNA Sequencer. This sequence was verified as the expected portion of the cox1 gene by comparison to the homologous regions of Katharina tunicata mtDNA (Boore and Brown 1994a).
A set of two oligonucleotide primers was designed facing "out" from this fragment, matching to positions that are separated by only 27 nts of the G. eborea cox1 sequence. These were used to amplify 14,465 nt, nearly the entire mtDNA, in a single reaction. This PCR used rTth-XL polymerase (Perkin-Elmer) with 1.3 mM MgOAc, and was otherwise performed according to supplier's instructions. Reaction volume was 100 μl and conditions were 94°C for 45 sec, followed by 37 cycles of: (94°C, 10 sec; 55°C, 20 sec; 65°C, 12 min, with an additional 15 sec per cycle after the 16th), then an incubation at 72°C for 12 min. An aliquot yielded a single band on a 1% agarose gel stained with ethidium bromide.
Approximately 2 μg of this product was digested separately with the restriction enzymes MboI and TaqI, each of which recognizes 4-nt sites. Several fragments were selected from each digest and gel purified as above, then they were ligated into the compatible BamHI and ClaI sites, respectively, of pBluescript plasmid (Stratagene), followed by DNA preparation and sequence determination as above. Additional oligonucleotide primers were designed for determining the sequence "out" from each of these cloned fragments. The 14,465 nt PCR product was passed three times through an Ultrafree Spin Column (30,000 NMWL; Millipore) to eliminate amplification primers and PCR reagents and then used directly as a template for sequencing reactions as above. Using a combination of oligonucleotides matching the ends of the amplified fragments with those matching internal sequences obtained from the cloned MboI and TaqI fragments greatly reduced the time required to "primer walk" through this fragment. All sequence was determined in both directions.
Completing the mtDNA Sequence of Mytilus edulis
Most (13.9 kb) of the F-type mtDNA sequence of M. edulis has been previously reported (Hoffmann, Boore, and Brown 1992). Although this was sufficient to determine the gene content and arrangement, it omitted the sequences of the central portions of many genes. To complete this, we designed oligonucleotide primers for PCR that match the ends of the previously reported sequences and used these to amplify the undetermined portions using DNA preparations of the appropriate M. edulis clones (Hoffmann, Boore, and Brown 1992) as templates. Each PCR reaction yielded a single band on a 1% agarose gel when visualized by ethidium bromide staining and UV irradiation. DNA was purified and the DNA sequence was determined as for G. eborea, using the amplifying or internal primers as necessary. All sequence was determined in both directions and was assembled with that previously reported into a complete mtDNA sequence.
Gene Annotation and Gene Order Comparison
Protein-encoding genes of each mtDNA were identified by sequence similarity of open reading frames to mitochondrial gene sequences of Katharina tunicata (Boore and Brown 1994a). Ribosomal RNA genes were identified by sequence similarity and potential secondary structures. As a class, tRNA genes were identified by their potential to form tRNA-like secondary structures; specific identifications were made according to anticodon sequence.
A search for shared gene arrangements was conducted against all mitochondrial sequence data available in GenBank that included sequence from three or more genes (3,376 entries). This search employed a PERL script that decomposed the query genome into all binary gene arrangements, searched for shared gene orders, and then reassembled any overlapping pairs for each comparison.
Phylogenetic Analysis of Protein Data
We included 27 taxa in the phylogenetic analysis (table 1), 15 of which are mollusks or other lophotrochozoans and 12 of which are metazoan outgroups (five non-lophotrochozoan protostomes, six deuterostomes, and one cnidarian). (One taxon, Inversidens japanensis, is represented by two sets of sequences, one F-type and one M-type.) We performed multiple sequence alignments for each protein using the pileup program in the GCG package. Each alignment was then refined by eye and subsequently combined into a concatenated data set. Because atp8 is missing in several of the taxa, it was excluded from all analyses. Regions of ambiguous alignment were also excluded; table 2 shows the regions corresponding to each gene in the concatenated alignment, the total number of positions per protein, and the number of amino acid sites included in the final analysis. Maximum parsimony (MP) reconstructions were conducted with PAUP*4.0b (Swofford 2001), with branch support estimated from 1,000 bootstrap pseudoreplicates. Quartet-puzzling (QP) was performed with Tree-Puzzle (Strimmer and Haeseler 1997) using both the mtREV24 and Blosum62 models with 100,000 quartet-puzzling steps, with a gamma correction and eight rate categories, and estimating amino acid frequencies from the data set. Bayesian reconstructions (MB) used MrBayes 3.0 (Huelsenbeck and Ronquist 2001). Exploratory Markov Chain Monte Carlo runs were performed starting with different amino acid substitution priors (i.e., mixed models, Poisson). Subsequently, we ran the heated MCMC chain for 1,000,000 generations, which was sampled every 100 updates using the models with higher posterior probabilities (mtREV and Blossum) from the mixed model prior. We discarded 1,000 cycles as burn-in before estimating joint posterior probabilities. We also analyzed each gene individually using MP and QP as above, and MB with mixed and mtRev amino acid models. The Nexus-formatted file of the alignment is available as online Supplementary Material.
Table 1 Species Used in the Phylogenetic Analysis with Current Taxonomic Classification and Mitochondrial Genome Accession Numbers.
Table 2 Number of Amino Acid Positions Used in the Final Analysis.
Results and Discussion
Gene Content and Organization
The mitochondrial genome of G. eborea (GenBank accession number AY484748) contains the 37 genes most commonly found in animal mtDNAs (Boore 1999), including atp8, which is missing in M. edulis (complete sequence deposited in GenBank as accession number AY484747). (atp8 is also missing from the mtDNAs of other bivalves [Kim, Je, and Park 1999; Okazaki and Ueshima 2001], secernentean nematodes [Okimoto et al. 1992], platyhelminths [Le et al. 2000], and chaetognaths [unpublished data].) Genes are divided between the two strands about evenly, with one having 12 tRNA- and five protein-encoding genes and the other having 10 tRNA-, eight protein-, and two rRNA-encoding genes (fig. 2). (All genes are on the same strand for M. edulis mtDNA.) On the strand reading clockwise in figure 2, there are two regions of co-oriented gene clusters: (1) from atp6 to trnH (eight genes) and (2) from trnA to trnW (six genes). In addition, this strand contains three tRNA genes (trnG, trnR, and trnS2) that punctuate the gene arrangement on the opposite strand. On the strand reading counter-clockwise in figure 2, there are two regions of gene clusters: (1) from nad4L to trnF (six genes) and (2) from trnV to cox2 (14 genes), except as interrupted by trnG, trnR, and trnS2.
FIG. 2. Gene map of the mitochondrial genome of the scaphopod mollusk Graptacme eborea. Genes have standard abbreviations except for tRNAs, which are designated by the one-letter code for the corresponding amino acid. S1, S2, L1, and L2 designate genes for those tRNAs recognizing the codons AGN, UCN, CUN, and UUR, respectively. Transfer RNA genes shown outside of the circle are transcribed clockwise while those inside are transcribed counter-clockwise. Transcriptional orientation for each of the other genes is indicated with an arrow. Gene scaling is only approximate
As is typical of animal mtDNAs, genes start with nonstandard initiation codons, but with no consistency of usage among the genes (table 3). Similarly, there is no pattern evident in the use of complete vs. abbreviated stop codons.
Table 3 Comparisons of the Mitochondrial Protein Coding Genes of Four Mollusks, the Scaphopod Graptacme eborea (Geb), Bivalve Mytilus edulis (Med), Polyplacophoran Katharina tunicata (Ktu), and the Gastropod Albinaria coerulea (Aco).
In the few cases where it has been studied, mitochondrial genes are expressed as a polycistron, and then the tRNAs are enzymatically removed to liberate the flanking gene specific mRNAs. G. eborea mtDNA contains only four cases where non-tRNA genes abut: nad2-cob, atp8-nad1, nad4L-nad4, and nad4L-atp6. For this last case, the genes are on opposite strands. We infer that nad2 and nad4L end on abbreviated stop codons (T and TA, respectively; see fig. 3) that would be completed by polyadenylation, but it is not clear how cleavage would occur to end these mRNAs at these nucleotides. The genes for atp8 and nad1 are separated by 14 nt, raising the same issue. It has been speculated that other secondary structures can substitute for tRNAs for message cleavage (e.g., Boore and Brown 1994a), but no potential secondary structures are obvious at these gene boundaries. Further, it is not clear by what alternative mechanism transcript cleavage could occur at the 3' end of cox2, cox3, or nad6 since these are flanked by tRNA genes on the opposite strand.
FIG. 3. A greatly-abbreviated, schematic representation of the mtDNA sequence of Graptacme eborea. Numbers within the slash marks indicate omitted nucleotides. Translation for each gene is assumed to initiate with formyl-methionine, so for those genes starting with other than ATR (cox1, nad1, nad4, and nad4L), the initiator M is shown in parenthesis to indicate presumed nonconformity with the generally employed genetic code. Stop codons, including those inferred to be "abbreviated," are marked by an asterisk. A dart (>) marks the last nucleotide of each gene and indicates the direction of transcription
There are several cases where genes seem to overlap on the same strand. The cox1 gene is interpreted to overlap the downstream trnS1 by six nts. The only reasonable alternative is that cox1 ends on an abbreviated stop codon; the minimal truncation to allow this would be by 24 nucleotides. The seven amino acids that would then not be coded at the carboxyl end are not identical to those of any other animal, so this is possible, but they are chemically similar to the amino acids commonly found at the carboxyl end of animal Cox1 proteins. Secondly, trnY overlaps trnR by three nts, GGA, which match well in each of these two tRNAs (fig. 4). The third case is for the pair trnR-trnS2, which overlaps by a single T; perhaps the tRNA(R) discriminator nucleotide is not encoded in the genomic DNA. Overlapping genes could not be resolved into separate, complete messages from a single polycistron, so this would imply that (1) these genes are transcribed from multiple promoters, (2) there is differential cleavage to generate sometime one or sometimes the other complete RNA, or (3) complete RNAs are restored by some type of post transcriptional mechanism.
FIG. 4. The potential secondary structures of the 22 inferred tRNAs of Graptacme eborea mtDNA. Nomenclature for portions of tRNA structures is shown for tRNA(V). Codons recognized are shown for the pairs of leucine and serine tRNAs. The arrow on tRNA(L1) marks the beginning of overlap with the downstream gene trnL2.
The nucleotides CTAA depicted as being part of trnL1 in figure 4 do not pair well with the 5' end of the tRNA and overlap with the downstream gene, trnL2. It is possible, alternatively, that these genomic nucleotides are not part of the trnL1 gene, but that the necessary nucleotides to complete the tRNA are added posttranscriptionally after tRNA(L2) is cleaved from the transcript. This could, potentially, be done by an RNA-dependent RNA polymerase, using the 5' end of the tRNA as a template, as seems likely for centipede mt-tRNAs (Lavrov, Brown, and Boore 2000). However, because the potentially edited nucleotides would all be A's (to match the T's at the 5' end), this might be completed by a simpler mechanism, tRNA polyadenylation, as is the case for some mt-tRNAs (Yokobori and P??bo 1995). Otherwise, there is little reason to suppose this as a common process, since the amino-acyl acceptor stems are well-matched for the most 3' sequences for every other tRNA except tRNA(S2), which has a single mismatched TT terminal pair. This mismatch could be corrected by tRNA polyadenylation, although this gene (as inferred with the mismatch) does not overlap any other.
There have been extensive, unique rearrangements involving nearly every gene (fig. 5) of both of these mtDNAs. G. eborea and M. edulis mtDNAs have only a few gene boundaries in common with any other animal studied to date. An examination of all 3,376 entries in GenBank of sequences having three or more mitochondrial genes reveals that G. eborea shares the arrangement rrnL, trnM, rrnS with the Yesso scallop, Mizuhopecten yessoensis (GenBank accession AB052599), and nad1, P, nad6 with the squid, Loligo bleekeri. The first is an interesting potential synapomorphy that would exclude, among the sampled mollusks, only the polyplacophoran Katharina tunicata (Boore and Brown 1994a) and the cephalopod Loligo bleekeri (Sasuga et al. 1999); others have these genes in autapomorphic arrangements. It is interesting that the sampled gastropods have the arrangement trnM, rrnS, although rrnL is elsewhere. The inferred basal group, Polyplacophora, is represented by K. tunicata, which has an arrangement similar to the second case, nad1, -P, nad6; the other studied mollusks have further rearrangements of these genes. The same analysis of M. edulis mtDNA reveals that it shares the arrangement trnL1, trnL2, nad1 with the hemichordate Balanoglossus carnosus (Castresana et al. 1998) and with K. tunicata. Other than these arrangements, neither mtDNA shares more than a single gene boundary with any other animal.
FIG. 5. The gene map of Graptacme eborea mtDNA compared with those of all other mollusks whose complete mt gene arrangement has been published and with that of an outgroup, the annelid Lumbricus terrestris (Boore and Brown 1995). Genes are abbreviated as in figure 2. All genes are transcribed from left-to-right except for those with a left-facing arrow
Transfer RNAs
G. eborea mtDNA contains the typical complement of 22 tRNAs. Most have potential to fold into a normal cloverleaf structure, although tRNA(R) and tRNA(S1) do not have paired DHU arms, and a few others have a mismatched nt pair. All have T immediately preceding and R following the anticodon. tRNAs have been described previously for M. edulis (Hoffmann, Boore, and Brown 1992).
Mutational Bias Is Reflected in Codon Usage Patterns and Protein Amino Acid Composition
G. eborea mtDNA is 14,492 bp in length and is 74.1% A+T, very high even for a mitochondrial genome. Strand skew measures (Perna and Kocher 1995) for the distribution of GC pairs [(G–C)/(G+C)] and TA pairs [(T–A)/(T+A)] between the two strands are nearly zero (–0.02 and +0.002, respectively).
M. edulis mtDNA is 16,740 bp in length and is 61.8% A+T. GC skew is +0.246 and TA skew is +0.110, indicating that the strand containing the genes is quite rich in G and T relative to the other strand. This bias is very evident in comparisons of synonymous codon usage pattern between the two genomes (table 4); for every case where an amino acid can be specified by any NNR codon, M. edulis has a much greater proportion of NNG:NNA relative to G. eborea.
Table 4 Codon Usage for Graptacme eborea (Geb; 3,649 Codons) and Mytilus edulis (Med; 3,681 Codons) mtDNAs.
As has been found for many mtDNAs (Cardon et al. 1994), CpG is the least frequent dinucleotide for these two mtDNAs, both in absolute number and relative to expected frequency calculated from the percentage of C and G in the genomes. For both mtDNAs, all four homodimers are significantly more common than would occur by chance; these values are the four highest except for GC in both mtDNAs and CT for M. edulis alone.
The amino acid leucine can be specified by six different codons (TTR and CTN) and the proteins of these two mtDNAs have a very similar number of leucines. As reflects the mutation bias, M. edulis has a much greater proportion of leucines specified by TTG and CTG codons, mainly at the expense of TTA codons. The amino acid serine can be specified by eight different codons (TCN and AGN); the proteins of the two mtDNAs also contain similar numbers of this amino acid and, again, the distribution reflects the mutation bias of M. edulis toward higher G and T. All AGN codons are used much more frequently in M. edulis, especially AGG; TCG usage is also elevated, all at the expense of TCT, TCC, and TCA codons.
This bias is also reflected in patterns of amino acid substitutions between these two mtDNAs. G. eborea and M. edulis mitochondrial proteins contain nearly identical numbers of nonpolar (A, V, L, I, P, M, F, W) (2,082 and 2,020, respectively) and polar (G, S, T, C, Y, N, Q) (1,170 and 1,219, respectively) amino acids. For nonpolar amino acids, M. edulis proteins use many more alanines (GCN) and valines (GTN) at the expense of isoleucine (ATY), methionine (ATR), and phenylalanine (TTY). For polar amino acids, M. edulis proteins contain more glycine (GGN) at the expense of asparagine (AAY). Presumably, the bias toward G and T in the gene-containing strand of M. edulis has resulted in amino acid replacements within the tolerance of physio-chemical similarity.
Unassigned DNA
G. eborea mtDNA is very uncommon for lacking any large noncoding regions, as are usually inferred to contain the origin(s) of replication and transcription control signals. The largest noncoding region is only 24 nt between trnK and trnF. Next in size are the 19 nt gap between cox3 and trnG and the 18 nt between trnG and trnQ. Noncoding DNA of M. edulis mtDNA has been analyzed and described earlier (Hoffmann, Boore, and Brown 1992). There is no obvious conservation of either nucleotide identities or potential secondary structures between the mollusks' noncoding regions. Whatever regulatory elements may be present are apparently short, dispersed, and/or rapidly changing.
Phylogenetic Analysis
Figure 6A presents a 70% majority rule consensus tree of MP bootstrap analysis for the taxa outlined in table 1, which has a topology congruent with those from quartet puzzling (QP) and some of the Bayesian analysis. These analyses support the monophyly of the lophophorates, annelids, and brachiopods. Relationships among the major molluscan lineages, however, remain unresolved, as are those among mollusks, brachiopods, and annelids, despite using this relatively large data set of 2,420 confidently aligned amino acid positions. In contrast, relationships among the major groups of deuterostomes and of arthropods are well resolved and conforming to expectation from other analyses, bolstering the view that the relationships among the lophotrochozoan groups are especially difficult to resolve.
FIG. 6. Comparisons of various phylogenetic analyses of 2,420 aligned amino acid positions of 12 concatenated mitochondrial proteins. (A) 70% majority rule consensus tree of maximum parsimony bootstrap analysis. This tree is congruent with the quartet puzzling (QP) analyses, and the numerals at the nodes are MP bootstrap values followed by QP support from the mtREV24 model analysis (which were very similar to those using Blosum62). Results of Bayesian analyses (50% majority rule consensus) are shown using the Blosum62 (B), mtREV24 (C), mixed (D), and Poisson (E) models. Any node unmarked by a numeral has support of 100%
Results of MB analyses with different sets of prior probabilities for the amino acid model are depicted in figure 6, panels B–E. Concerns over the inability to resolve many of these phylogenetic relationships are exacerbated by the observation that the Bayesian analyses (unlike the MP and QP analyses) return substantially different topologies with high posterior probability values when the prior amino acid substitution models vary.
However, this data set of concatenated protein sequences does give consistent results for support of many metazoan clades regardless of the type of analysis performed (table 5). In contrast, analyses using each individual gene recovered only the deuterostome and protostome nodes with high levels of confidence (>70%) in six cases and not all methods were consistent. The arthropod clade was recovered by only three individual genes and by only one or two methods in the best of cases. The Lophotrochozoa clade was recovered only by Nad2 and only in the case of using the mixed model prior analysis.
Table 5 Support of Particular Metazoan Clades in Individual Gene (Excluding Atp8) Analyses Using Maximum Parsimony (MP), Quartet Puzzling (QP) with mtREV24, and MrBayes (MB) with a Mixed Model Prior Probability Assignment.
Although there are key nodes within the Lophotrochozoa that are still unresolved, one hopes that they will yield to further taxon sampling, and these results illustrate the importance of sequencing complete mitochondrial genomes, as opposed to using only single gene sequences, when analyzing deep divergences.
Acknowledgements
Thanks to K. Fraser and T. Collins for the gift of G. eborea DNA and to B. Dayrat for helpful comments. This work was supported by DEB-9807100 from the National Science Foundation. Part of this work was performed under the auspices of the U.S. Department of Energy, Office of Biological and Environmental Research, by the University of California, Lawrence Berkeley National Laboratory, under contract No. DE-AC03-76SF00098.
Literature Cited
Beagley, C. T., R. Okimoto, and D. R. Wolstenholme. 1999. Mytilus mitochondrial DNA contains a functional gene for a tRNASer(UCN) with a dihydrouridine arm-replacement loop and a pseudo-tRNASer(UCN) gene. Genetics 152:641-652.
Black, W. C., and R. L. Roehrdanz. 1998. Mitochondrial gene order is not conserved in arthropods: Prostriate and metastriate tick mitochondrial genomes. Mol. Biol. Evol. 15:1772-1785.
Boore, J. L. 1999. Animal mitochondrial genomes. Nucleic Acids Res. 27:1767-1780.
Boore, J. L. 2002. "Mitochondrial Gene Arrangement Source Guide.". (http://www.jgi.doe.gov/programs/comparative/Mito_top_level.html).
Boore, J. L., and W. M. Brown. 1994a. Complete DNA sequence of the mitochondrial genome of the black chiton, Katharina tunicata. Genetics 138:423-443.
Boore, J. L., and W. M. Brown. 1994b. Mitochondrial genomes and the phylogeny of mollusks. Nautilus 108:(Suppl. 2): 61-78.
Boore, J. L., and W. M. Brown. 1995. The complete DNA sequence of the mitochondrial genome of the annelid worm Lumbricus terrestris. Genetics 141:305-319.
Boore, J. L., and W. M. Brown. 1998. Big trees from little genomes: Mitochondrial gene order as a phylogenetic tool. Curr. Opin. Genet. Dev. 8:668-674.
Campbell, N. J. H., and S. C. Barker. 1998. An unprecedented major rearrangement in an arthropod mitochondrial genome. Mol. Biol. Evol. 15:1786-1787.
Cardon, L. R., C. Burge, D. A. Clayton, and S. Karlin. 1994. Pervasive CpG suppression in animal mitochondrial genomes. Proc. Natl. Acad. Sci. USA 91:3799-3803.
Castresana, J., G. Feldmaier-Fuchs, S.-I. Yokobori, N. Satoh, and S. P??bo. 1998. The mitochodrial genome of the hemichordate Balanoglossus carnosus and the evolution of deuterostome mitochondria. Genetics 150:1115-1123.
Folmer, O., M. Black, W. Hoeh, R. Lutz, and R. Vrijenhoek. 1994. DNA primers for amplification of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates. Mol. Mar. Biol. Biotech. 3:294-299.
Grande, C., J. Templado, J. L. Cervera, and R. Zardoya. 2002. The complete mitochondrial genome of the nudibranch Roboastra europaea (Mollusca: Gastropoda) supports the monophyly of opisthobranchs. Mol. Biol. Evol. 19:1672-1685.
Haszprunar, G. 2000. Is the Aplacophora monophyletic? A cladistic point of view. American Malacological Bulletin 15:115-130.
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.
Hickerson, M. J., and C. W. Cunningham. 2000. Dramatic mitochondrial gene rearrangements in the Hermit Crab Pagurus longicarpus (Crustacea, Anomura). Mol. Biol. Evol. 17:639-644.
Hoeh, W. R., K. H. Blakley, and W. M. Brown. 1991. Heteroplasmy suggests limited biparental inheritance of Mytilus mitochondrial DNA. Science 251:1488-1490.
Hoeh, W. R., D. T. Stewart, and S. I. Guttman. 2002. High fidelity of mitochondrial genome transmission under the doubly uniparental mode of inheritance in freshwater mussels (Bivalvia: Unionoidea). Evolution 56:2252-2261.
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.
Huelsenbeck, J. P., and F. Ronquist. 2001. MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17:754-755.
Kim, S.-H., E.-Y. Je, and D.-W. Park. 1999. Crassostrea gigas mitochondrial DNA. GenBank accession number AF177226.
Kurabayashi, A., and R. Ueshima. 2000a. Complete sequence of the mitochondrial DNA of the primitive opisthobranch Pupa strigosa: Systematic implications of the genome organization. Mol. Biol. Evol. 17:266-277.
Kurabayashi, A., and R. Ueshima. 2000b. Partial mitochondrial genome organization of the heterostrophan gastropod Omalogyra atomus and its systematic significance. Venus (Jap. J. Malac.) 59:7-18.
Lavrov, D. V., W. M. Brown, and J. L. Boore. 2000. A novel type of RNA editing occurs in the mitochondrial tRNAs of the centipede Lithobius forticatus. Proc. Natl. Acad. Sci. USA 97:13738-13742.
Le, T. H., D. Blair, and T. Agatsuma, et al. (14 co-authors). 2000. Phylogenies inferred from mitochondrial gene orders-a cautionary tale from the parasitic flatworms. Mol. Biol. Evol. 17:1123-1125.
Lecanidou, R., V. Douris, and G. Rodakis. 1994. Novel features of metazoan mtDNA revealed from sequence analysis of three mitochondrial DNA segments of the land snail Albinaria turrita (Gastropoda: Clausiliidae). J. Mol. Evol. 38:369-382.
Okazaki, M., and R. Ueshima. 2001. Evolutionary diversity between the gender-associate mitochondrial DNA genomes of freshwater mussels. Genbank accession numbers AB055624 (male haplotype) and AB055625 (female haplotype).
Okazaki, M., and R. Ueshima. 2002. Gender-associated mtDNA of Tapes philippinarum. Genbank accession number NC_003354.
Okimoto, R., J. L. Macfarlane, D. O. Clary, and D. R. Wolstenholme. 1992. The mitochondrial genomes of two nematodes, Caenorhabditis elegans and Ascaris suum. Genetics 130:471-498.
Passamonti, M., and V. Scali. 2001. Gender-associated mitochondrial DNA heteroplasmy in the venerid clam Tapes philippinarum (Mollusca Bivalvia). Curr. Genet. 39:117-124.
Perna, N. T., and T. D. Kocher. 1995. Patterns of nucleotide composition at fourfold degenerate sites of animal mitochondrial genomes. J. Mol. Evol. 41:353-358.
Rawlings, T., T. Collins, and R. Bieler. 2001. A major mitochondrial gene rearrangement among closely related species. Mol. Biol. Evol. 18:1604-1609.
Runnegar, B., and J. Pojeta. 1974. Molluscan phylogeny: The paleontological viewpoint. Science 186:311-317.
Salvini-Plawen, L. V. 1985. Early evolution and the primitive groups. Pp. 59–150 in E. R. Trueman and M. R. Clark, eds. Evolution. Academic Press, Orlando, Fla.
Salvini-Plawen, L. V., and G. Steiner. 1996. Synapomorphies and plesiomorphies in higher classification of Mollusca. Pp. 29–51 in J. Taylor, ed. Origin and evolutionary radiation of the Mollusca. Oxford University Press, Oxford.
Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning, a laboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York.
Sasuga, J., S.-I. Yokobori, M. Kaifu, T. Ueda, K. Nishikawa, and K. Watanabe. 1999. Gene contents and organization of a mitochondrial DNA segment of the squid Loligo bleekeri. J. Mol. Evol. 48:692-702.
Scheltema, A. H. 1993. Aplacophora as progenetic aculiferans and the coelomatic origin of mollusks as the sister taxon of Sipuncula. Biol. Bull. 184:57-78.
Scheltema, A. H. 1996. Phylogenetic position of Sipuncula, Mollusca and the progenetic Aplacophora. Pp. 53–58 in J. Taylor, ed. Origin and evolutionary radiation of the Mollusca. Oxford University Press, Oxford.
Shao, R., N. J. H. Campbell, and S. C. Barker. 2001. Numerous gene rearrangements in the mitochondrial genome of the wallaby louse, Heterodoxus macropus (Phthiraptera). Mol. Biol. Evol. 18:858-865.
Steiner, G., and H. Dreyer. 2002. Scaphopoda and Cephalopoda are sister taxa—an evolutionary scenario. Abstract for the American Malacological Society Annual Meeting, Charleston, SC.
Steiner, G., and S. Hammer. 2000. Molecular phylogeny of the Bivalvia inferred from 18S rDNA sequences with particular reference to the Pteriomorpha. Pp. 11–29 in E. M. Harper, J. D. Taylor, and J. A. Crame, eds. The evolutionary biology of the Bivalvia. The Geological Society of London Special Publications, London.
Steiner, G., and M. Müller. 1996. What can 18S rDNA do for bivalve phylogeny? J. Mol. Evol. 43:58-70.
Stewart, D. T., C. Saavedra, R. R. Stanwood, A. O. Ball, and E. Zouros. 1995. Male and female mitochondrial DNA lineages in the blue mussel (Mytilus edulis) species group. Mol. Biol. Evol. 12:735-747.
Strimmer, K., and A. von Haeseler. 1997. Puzzle. Maximum likelihood analysis for nucleotide and amino acid alignments. Zoologisches Institut, Munchen, Germany.
Swofford, D. L. 2001. PAUP*: phylogenetic analysis using parsimony (*and other methods). Beta Version 4.0b8. Sinauer Associates, Sunderland, Mass.
Terrett, J. A., S. Miles, and R. H. Thomas. 1996. Complete DNA sequence of the mitochondrial genome of Cepaea nemoralis (Gastropoda: Pulmonata). J. Mol. Evol. 42:160-168.
Tomita, K., S. Yokobori, T. Oshima, T. Ueda, and K. Watanabe. 2002. The cephalopod Loligo bleekeri mitochondrial genome: Multiplied noncoding regions and transposition of tRNA genes. J. Mol. Evol. 54:486-500.
Waller, T. R. 1998. Origin of the molluscan class Bivalvia and a phylogeny of major groups. Pp. 1–47 in P. A. Johnston and J. W. Haggart, eds. Bivalves: An eon of evolution—plaeobiological studies honoring Norman D. Newell. University of Calgary Press, Calgary, Canada.
Wilding, C. S., P. J. Mill, and J. Grahame. 1999. Partial sequence of the mitochondrial genome of Littorina saxatilis: Relevance to gastropod phylogenetics. J. Mol. Evol. 48:348-359.
Yamazaki, N., R. Ueshima, and J. Terrett, et al. (12 co-authors). 1997. Evolution of pulmonate gastropod mitochondrial genomes: comparisons of gene organizations of Euhadra, Cepaea and Albinaria and implications of unusual tRNA secondary structures. Genetics 145:749-758.
Yokobori, S., and S. P??bo. 1995. Transfer RNA editing in land snail mitochondria. Proc. Natl. Acad. Sci. USA 92:10432-10435.(Jeffrey L. Boore*,, Monic)
DOE Joint Genome Institute and Lawrence Berkeley National Laboratory, Walnut Creek, California
E-mail: jlboore@lbl.gov.
Abstract
We have determined the complete sequence of the mitochondrial genome of the scaphopod mollusk Graptacme eborea (14,492 nts) and completed the sequence of the mitochondrial genome of the bivalve mollusk Mytilus edulis (16,740 nts). (The name Graptacme eborea is a revision of the species formerly known as Dentalium eboreum.) G. eborea mtDNA contains the 37 genes that are typically found and has the genes divided about evenly between the two strands, but M. edulis contains an extra trnM and is missing atp8, and it has all genes on the same strand. Each has a highly rearranged gene order relative to each other and to all other studied mtDNAs. G. eborea mtDNA has almost no strand skew, but the coding strand of M. edulis mtDNA is very rich in G and T. This is reflected in differential codon usage patterns and even in amino acid compositions. G. eborea mtDNA has fewer noncoding nucleotides than any other mtDNA studied to date, with the largest noncoding region only 24 nt long. Phylogenetic analysis using 2,420 aligned amino acid positions of concatenated proteins weakly supports an association of the scaphopod with gastropods to the exclusion of Bivalvia, Cephalopoda, and Polyplacophora, but it is generally unable to convincingly resolve the relationships among major groups of the Lophotrochozoa, in contrast to the good resolution seen for several other major metazoan groups.
Key Words: scaphopod ? bivalve ? mollusk ? mitochondria ? evolution ? genome
Introduction
Complete mitochondrial (mt) DNA sequences have been reported for nine mollusks representing four classes: the polyplacophoran Katharina tunicata (Boore and Brown 1994a, 1994b); the cephalopod Loligo bleekeri (Tomita et al. 2002); the bivalves Crassostrea gigas (Kim, Je, and Park 1999), Venerupis (Ruditapes) philippinarum (Okazaki and Ueshima 2002), and Inversidens japanensis (incomplete by only a small fragment of apparently noncoding sequence; Okazaki and Ueshima 2001); and the gastropods Cepaea nemoralis (Terrett, Miles, and Thomas 1996), Albinaria coerulea (Hatzoglou, Rodakis, and Lecanidou 1995), Roboastra europaea (Grande et al. 2002), and Pupa strigosa (Kurabayashi and Ueshima 2000a). Additionally, enough of the mtDNA sequence of another gastropod, Euhadra herkotsi (Yamazaki et al. 1997), and of the bivalve Mytilus edulis (F-type, see below) (Hoffmann, Boore, and Brown 1992) have been determined to assess the arrangement of all encoded genes. Like the mitochondrial genomes of nearly all metazoan animals (see Boore 1999), these are small, circular DNA molecules containing almost universally the same 37 genes: 13 for protein subunits of oxidative phosphorylation enzymes (cox1-3, nad1-6, nad4L, cob, atp6, atp8), two for mitochondrial ribosomal RNAs (rrnS and rrnL), and 22 for the tRNAs necessary to translate these 13 proteins (designated by trnX, with X being the one-letter code for the corresponding amino acid, and the two for each of serine and for leucine differentiated by numeral). M. edulis varies from this by the loss of atp8 and the gain of a second trnM. M. edulis is unusual also in maintaining two different mtDNAs, one called F-type and one M-type (Hoeh, Blakley, and Brown 1991), transmitted by an unusual system dubbed "doubly-uniparental inheritance" (Stewart et al. 1995), which now appears to be widespread in bivalves (Passamonti and Scali 2001; Hoeh, Stewart, and Guttman 2002). Partial mtDNA sequences are also available for the bivalve M. californianus (also F-type; Beagley, Okimoto, and Wolstenholme 1999) and the gastropods Albinaria turrita (Lecanidou, Douris, and Rodakis 1994), Omalogyra atomus (Kurabayashi and Ueshima 2000b), Littorina saxatilis (Wilding, Mill, and Grahame 1999) and several vermetids (Rawlings, Collins, and Bieler 2001).
For some phyla of animals, mitochondrial gene arrangements seem seldom to have changed. For example, with few notable exceptions, those vertebrates studied have identical gene arrangements, as do most studied arthropods. (In the latter case, exceptions include the highly rearranged mtDNAs found for wallaby louse [Shao, Campbell, and Barker 2001], hermit crab [Hickerson and Cunningham 2000], and metastriate ticks [Black and Roehrdanz 1998; Campbell and Barker 1998].) Mollusks differ, with many gene rearrangements noted for the molluscan taxa listed above. High levels of rearrangement have also been noted for nematodes and brachiopods (see Boore 1999, 2002). Gene rearrangements have been shown to be very powerful characters for reconstructing evolutionary relationships (see Boore and Brown 1998), and the rapidity of rearrangement within a lineage determines the level at which rearrangements are likely to be phylogenetically informative.
The phylogenetic relationships among the different extant molluscan classes are not well established, and anatomical studies have proposed multiple alternatives to this issue. A common proposal is a gradist scenario where chitons (Polyplacophora) and solenogasters and caudofoveates (Aplacophora) are the basal lineages to a grade of valve-bearing taxa (Gastropoda, Bivalvia, Cephalopoda, Monoplacophora, and Scaphopoda), collectively known as the Conchifera (Salvini-Plawen 1985; Salvini-Plawen and Steiner 1996; Haszprunar 2000) (fig. 1A). Some propose that chitons and aplacophorans form a monophyletic clade rather than a grade (Scheltema 1993, 1996), and some view Conchifera as split into a cephalopod/gastropod clade and a scaphopod/bivalve clade with monoplacophorans as the basal conchiferan lineage (Runnegar and Pojeta 1974) (each as in fig. 1B). Recent evaluations of morphological and paleontological data (Waller 1998) as well as 18S rRNA sequences (Steiner and Dreyer 2002) alternatively conclude that Scaphopoda is the sister group to Cephalopoda (fig. 1C). The many ribosomal RNA sequences have so far only poorly resolved molluscan phylogeny, rendering some taxa paraphyletic (e.g., bivalves; Steiner and Müller 1996), making it difficult to assess whether the proposed anatomical interpretations are identifying true synapomorphies (Steiner and Müller 1996; Steiner and Hammer 2000). Consequently, we are in need of additional characters that can help address phylogenetic relationships among major molluscan lineages. As the mitochondrial genome database continues to grow, we will be able to incorporate both gene order and sequence data into this analysis. Here we present the first complete mitochondrial genome sequence for a member of the Scaphopoda (Graptacme eborea), a previously unsampled class, and the completed sequence of the F-type mitochondrial genome of the bivalve Mytilus edulis.
FIG. 1. A comparison of the most commonly held views on the evolutionary relationships of molluscan classes
Materials and Methods
Determining the mtDNA Sequence of Graptacme eborea
A preparation of total DNA prepared from G. eborea was the gift of K. Fraser and T. Collins. PCR amplification of a portion of cox1 with the LCO1490 and HCO2198 oligonucleotides (Folmer et al. 1994) employed an initial heating at 97°C for 5 min, a hot start by adding Taq polymerase (Fisher) at 72°C, then 36 cycles of: (94°C, 1 min; 45°C, 1 min; 72°C, 2 min), followed by incubation at 72°C for 15 min. Reaction volume was 25 μl, which then yielded a single band of approximately 710 nts on a 1% agarose gel stained with ethidium bromide. This amplification product was gel purified using Gene Clean (Bio 101) and ligated into pBluescript (Stratagene), which had been prepared by: digestion with EcoRV, tailing on each end with a single T using Taq polymerase and dTTP, then gel purification using Gene Clean (Bio 101). DNA was prepared from recombinant plasmids using alkaline lysis followed by organic extraction and ethanol precipitation (Sambrook, Fritsch, and Maniatis 1989). DNA sequence was determined using AmplitaqFS, dye-terminator cycle sequencing (Perkin-Elmer) according to supplier's instructions and an ABI 377 Automated DNA Sequencer. This sequence was verified as the expected portion of the cox1 gene by comparison to the homologous regions of Katharina tunicata mtDNA (Boore and Brown 1994a).
A set of two oligonucleotide primers was designed facing "out" from this fragment, matching to positions that are separated by only 27 nts of the G. eborea cox1 sequence. These were used to amplify 14,465 nt, nearly the entire mtDNA, in a single reaction. This PCR used rTth-XL polymerase (Perkin-Elmer) with 1.3 mM MgOAc, and was otherwise performed according to supplier's instructions. Reaction volume was 100 μl and conditions were 94°C for 45 sec, followed by 37 cycles of: (94°C, 10 sec; 55°C, 20 sec; 65°C, 12 min, with an additional 15 sec per cycle after the 16th), then an incubation at 72°C for 12 min. An aliquot yielded a single band on a 1% agarose gel stained with ethidium bromide.
Approximately 2 μg of this product was digested separately with the restriction enzymes MboI and TaqI, each of which recognizes 4-nt sites. Several fragments were selected from each digest and gel purified as above, then they were ligated into the compatible BamHI and ClaI sites, respectively, of pBluescript plasmid (Stratagene), followed by DNA preparation and sequence determination as above. Additional oligonucleotide primers were designed for determining the sequence "out" from each of these cloned fragments. The 14,465 nt PCR product was passed three times through an Ultrafree Spin Column (30,000 NMWL; Millipore) to eliminate amplification primers and PCR reagents and then used directly as a template for sequencing reactions as above. Using a combination of oligonucleotides matching the ends of the amplified fragments with those matching internal sequences obtained from the cloned MboI and TaqI fragments greatly reduced the time required to "primer walk" through this fragment. All sequence was determined in both directions.
Completing the mtDNA Sequence of Mytilus edulis
Most (13.9 kb) of the F-type mtDNA sequence of M. edulis has been previously reported (Hoffmann, Boore, and Brown 1992). Although this was sufficient to determine the gene content and arrangement, it omitted the sequences of the central portions of many genes. To complete this, we designed oligonucleotide primers for PCR that match the ends of the previously reported sequences and used these to amplify the undetermined portions using DNA preparations of the appropriate M. edulis clones (Hoffmann, Boore, and Brown 1992) as templates. Each PCR reaction yielded a single band on a 1% agarose gel when visualized by ethidium bromide staining and UV irradiation. DNA was purified and the DNA sequence was determined as for G. eborea, using the amplifying or internal primers as necessary. All sequence was determined in both directions and was assembled with that previously reported into a complete mtDNA sequence.
Gene Annotation and Gene Order Comparison
Protein-encoding genes of each mtDNA were identified by sequence similarity of open reading frames to mitochondrial gene sequences of Katharina tunicata (Boore and Brown 1994a). Ribosomal RNA genes were identified by sequence similarity and potential secondary structures. As a class, tRNA genes were identified by their potential to form tRNA-like secondary structures; specific identifications were made according to anticodon sequence.
A search for shared gene arrangements was conducted against all mitochondrial sequence data available in GenBank that included sequence from three or more genes (3,376 entries). This search employed a PERL script that decomposed the query genome into all binary gene arrangements, searched for shared gene orders, and then reassembled any overlapping pairs for each comparison.
Phylogenetic Analysis of Protein Data
We included 27 taxa in the phylogenetic analysis (table 1), 15 of which are mollusks or other lophotrochozoans and 12 of which are metazoan outgroups (five non-lophotrochozoan protostomes, six deuterostomes, and one cnidarian). (One taxon, Inversidens japanensis, is represented by two sets of sequences, one F-type and one M-type.) We performed multiple sequence alignments for each protein using the pileup program in the GCG package. Each alignment was then refined by eye and subsequently combined into a concatenated data set. Because atp8 is missing in several of the taxa, it was excluded from all analyses. Regions of ambiguous alignment were also excluded; table 2 shows the regions corresponding to each gene in the concatenated alignment, the total number of positions per protein, and the number of amino acid sites included in the final analysis. Maximum parsimony (MP) reconstructions were conducted with PAUP*4.0b (Swofford 2001), with branch support estimated from 1,000 bootstrap pseudoreplicates. Quartet-puzzling (QP) was performed with Tree-Puzzle (Strimmer and Haeseler 1997) using both the mtREV24 and Blosum62 models with 100,000 quartet-puzzling steps, with a gamma correction and eight rate categories, and estimating amino acid frequencies from the data set. Bayesian reconstructions (MB) used MrBayes 3.0 (Huelsenbeck and Ronquist 2001). Exploratory Markov Chain Monte Carlo runs were performed starting with different amino acid substitution priors (i.e., mixed models, Poisson). Subsequently, we ran the heated MCMC chain for 1,000,000 generations, which was sampled every 100 updates using the models with higher posterior probabilities (mtREV and Blossum) from the mixed model prior. We discarded 1,000 cycles as burn-in before estimating joint posterior probabilities. We also analyzed each gene individually using MP and QP as above, and MB with mixed and mtRev amino acid models. The Nexus-formatted file of the alignment is available as online Supplementary Material.
Table 1 Species Used in the Phylogenetic Analysis with Current Taxonomic Classification and Mitochondrial Genome Accession Numbers.
Table 2 Number of Amino Acid Positions Used in the Final Analysis.
Results and Discussion
Gene Content and Organization
The mitochondrial genome of G. eborea (GenBank accession number AY484748) contains the 37 genes most commonly found in animal mtDNAs (Boore 1999), including atp8, which is missing in M. edulis (complete sequence deposited in GenBank as accession number AY484747). (atp8 is also missing from the mtDNAs of other bivalves [Kim, Je, and Park 1999; Okazaki and Ueshima 2001], secernentean nematodes [Okimoto et al. 1992], platyhelminths [Le et al. 2000], and chaetognaths [unpublished data].) Genes are divided between the two strands about evenly, with one having 12 tRNA- and five protein-encoding genes and the other having 10 tRNA-, eight protein-, and two rRNA-encoding genes (fig. 2). (All genes are on the same strand for M. edulis mtDNA.) On the strand reading clockwise in figure 2, there are two regions of co-oriented gene clusters: (1) from atp6 to trnH (eight genes) and (2) from trnA to trnW (six genes). In addition, this strand contains three tRNA genes (trnG, trnR, and trnS2) that punctuate the gene arrangement on the opposite strand. On the strand reading counter-clockwise in figure 2, there are two regions of gene clusters: (1) from nad4L to trnF (six genes) and (2) from trnV to cox2 (14 genes), except as interrupted by trnG, trnR, and trnS2.
FIG. 2. Gene map of the mitochondrial genome of the scaphopod mollusk Graptacme eborea. Genes have standard abbreviations except for tRNAs, which are designated by the one-letter code for the corresponding amino acid. S1, S2, L1, and L2 designate genes for those tRNAs recognizing the codons AGN, UCN, CUN, and UUR, respectively. Transfer RNA genes shown outside of the circle are transcribed clockwise while those inside are transcribed counter-clockwise. Transcriptional orientation for each of the other genes is indicated with an arrow. Gene scaling is only approximate
As is typical of animal mtDNAs, genes start with nonstandard initiation codons, but with no consistency of usage among the genes (table 3). Similarly, there is no pattern evident in the use of complete vs. abbreviated stop codons.
Table 3 Comparisons of the Mitochondrial Protein Coding Genes of Four Mollusks, the Scaphopod Graptacme eborea (Geb), Bivalve Mytilus edulis (Med), Polyplacophoran Katharina tunicata (Ktu), and the Gastropod Albinaria coerulea (Aco).
In the few cases where it has been studied, mitochondrial genes are expressed as a polycistron, and then the tRNAs are enzymatically removed to liberate the flanking gene specific mRNAs. G. eborea mtDNA contains only four cases where non-tRNA genes abut: nad2-cob, atp8-nad1, nad4L-nad4, and nad4L-atp6. For this last case, the genes are on opposite strands. We infer that nad2 and nad4L end on abbreviated stop codons (T and TA, respectively; see fig. 3) that would be completed by polyadenylation, but it is not clear how cleavage would occur to end these mRNAs at these nucleotides. The genes for atp8 and nad1 are separated by 14 nt, raising the same issue. It has been speculated that other secondary structures can substitute for tRNAs for message cleavage (e.g., Boore and Brown 1994a), but no potential secondary structures are obvious at these gene boundaries. Further, it is not clear by what alternative mechanism transcript cleavage could occur at the 3' end of cox2, cox3, or nad6 since these are flanked by tRNA genes on the opposite strand.
FIG. 3. A greatly-abbreviated, schematic representation of the mtDNA sequence of Graptacme eborea. Numbers within the slash marks indicate omitted nucleotides. Translation for each gene is assumed to initiate with formyl-methionine, so for those genes starting with other than ATR (cox1, nad1, nad4, and nad4L), the initiator M is shown in parenthesis to indicate presumed nonconformity with the generally employed genetic code. Stop codons, including those inferred to be "abbreviated," are marked by an asterisk. A dart (>) marks the last nucleotide of each gene and indicates the direction of transcription
There are several cases where genes seem to overlap on the same strand. The cox1 gene is interpreted to overlap the downstream trnS1 by six nts. The only reasonable alternative is that cox1 ends on an abbreviated stop codon; the minimal truncation to allow this would be by 24 nucleotides. The seven amino acids that would then not be coded at the carboxyl end are not identical to those of any other animal, so this is possible, but they are chemically similar to the amino acids commonly found at the carboxyl end of animal Cox1 proteins. Secondly, trnY overlaps trnR by three nts, GGA, which match well in each of these two tRNAs (fig. 4). The third case is for the pair trnR-trnS2, which overlaps by a single T; perhaps the tRNA(R) discriminator nucleotide is not encoded in the genomic DNA. Overlapping genes could not be resolved into separate, complete messages from a single polycistron, so this would imply that (1) these genes are transcribed from multiple promoters, (2) there is differential cleavage to generate sometime one or sometimes the other complete RNA, or (3) complete RNAs are restored by some type of post transcriptional mechanism.
FIG. 4. The potential secondary structures of the 22 inferred tRNAs of Graptacme eborea mtDNA. Nomenclature for portions of tRNA structures is shown for tRNA(V). Codons recognized are shown for the pairs of leucine and serine tRNAs. The arrow on tRNA(L1) marks the beginning of overlap with the downstream gene trnL2.
The nucleotides CTAA depicted as being part of trnL1 in figure 4 do not pair well with the 5' end of the tRNA and overlap with the downstream gene, trnL2. It is possible, alternatively, that these genomic nucleotides are not part of the trnL1 gene, but that the necessary nucleotides to complete the tRNA are added posttranscriptionally after tRNA(L2) is cleaved from the transcript. This could, potentially, be done by an RNA-dependent RNA polymerase, using the 5' end of the tRNA as a template, as seems likely for centipede mt-tRNAs (Lavrov, Brown, and Boore 2000). However, because the potentially edited nucleotides would all be A's (to match the T's at the 5' end), this might be completed by a simpler mechanism, tRNA polyadenylation, as is the case for some mt-tRNAs (Yokobori and P??bo 1995). Otherwise, there is little reason to suppose this as a common process, since the amino-acyl acceptor stems are well-matched for the most 3' sequences for every other tRNA except tRNA(S2), which has a single mismatched TT terminal pair. This mismatch could be corrected by tRNA polyadenylation, although this gene (as inferred with the mismatch) does not overlap any other.
There have been extensive, unique rearrangements involving nearly every gene (fig. 5) of both of these mtDNAs. G. eborea and M. edulis mtDNAs have only a few gene boundaries in common with any other animal studied to date. An examination of all 3,376 entries in GenBank of sequences having three or more mitochondrial genes reveals that G. eborea shares the arrangement rrnL, trnM, rrnS with the Yesso scallop, Mizuhopecten yessoensis (GenBank accession AB052599), and nad1, P, nad6 with the squid, Loligo bleekeri. The first is an interesting potential synapomorphy that would exclude, among the sampled mollusks, only the polyplacophoran Katharina tunicata (Boore and Brown 1994a) and the cephalopod Loligo bleekeri (Sasuga et al. 1999); others have these genes in autapomorphic arrangements. It is interesting that the sampled gastropods have the arrangement trnM, rrnS, although rrnL is elsewhere. The inferred basal group, Polyplacophora, is represented by K. tunicata, which has an arrangement similar to the second case, nad1, -P, nad6; the other studied mollusks have further rearrangements of these genes. The same analysis of M. edulis mtDNA reveals that it shares the arrangement trnL1, trnL2, nad1 with the hemichordate Balanoglossus carnosus (Castresana et al. 1998) and with K. tunicata. Other than these arrangements, neither mtDNA shares more than a single gene boundary with any other animal.
FIG. 5. The gene map of Graptacme eborea mtDNA compared with those of all other mollusks whose complete mt gene arrangement has been published and with that of an outgroup, the annelid Lumbricus terrestris (Boore and Brown 1995). Genes are abbreviated as in figure 2. All genes are transcribed from left-to-right except for those with a left-facing arrow
Transfer RNAs
G. eborea mtDNA contains the typical complement of 22 tRNAs. Most have potential to fold into a normal cloverleaf structure, although tRNA(R) and tRNA(S1) do not have paired DHU arms, and a few others have a mismatched nt pair. All have T immediately preceding and R following the anticodon. tRNAs have been described previously for M. edulis (Hoffmann, Boore, and Brown 1992).
Mutational Bias Is Reflected in Codon Usage Patterns and Protein Amino Acid Composition
G. eborea mtDNA is 14,492 bp in length and is 74.1% A+T, very high even for a mitochondrial genome. Strand skew measures (Perna and Kocher 1995) for the distribution of GC pairs [(G–C)/(G+C)] and TA pairs [(T–A)/(T+A)] between the two strands are nearly zero (–0.02 and +0.002, respectively).
M. edulis mtDNA is 16,740 bp in length and is 61.8% A+T. GC skew is +0.246 and TA skew is +0.110, indicating that the strand containing the genes is quite rich in G and T relative to the other strand. This bias is very evident in comparisons of synonymous codon usage pattern between the two genomes (table 4); for every case where an amino acid can be specified by any NNR codon, M. edulis has a much greater proportion of NNG:NNA relative to G. eborea.
Table 4 Codon Usage for Graptacme eborea (Geb; 3,649 Codons) and Mytilus edulis (Med; 3,681 Codons) mtDNAs.
As has been found for many mtDNAs (Cardon et al. 1994), CpG is the least frequent dinucleotide for these two mtDNAs, both in absolute number and relative to expected frequency calculated from the percentage of C and G in the genomes. For both mtDNAs, all four homodimers are significantly more common than would occur by chance; these values are the four highest except for GC in both mtDNAs and CT for M. edulis alone.
The amino acid leucine can be specified by six different codons (TTR and CTN) and the proteins of these two mtDNAs have a very similar number of leucines. As reflects the mutation bias, M. edulis has a much greater proportion of leucines specified by TTG and CTG codons, mainly at the expense of TTA codons. The amino acid serine can be specified by eight different codons (TCN and AGN); the proteins of the two mtDNAs also contain similar numbers of this amino acid and, again, the distribution reflects the mutation bias of M. edulis toward higher G and T. All AGN codons are used much more frequently in M. edulis, especially AGG; TCG usage is also elevated, all at the expense of TCT, TCC, and TCA codons.
This bias is also reflected in patterns of amino acid substitutions between these two mtDNAs. G. eborea and M. edulis mitochondrial proteins contain nearly identical numbers of nonpolar (A, V, L, I, P, M, F, W) (2,082 and 2,020, respectively) and polar (G, S, T, C, Y, N, Q) (1,170 and 1,219, respectively) amino acids. For nonpolar amino acids, M. edulis proteins use many more alanines (GCN) and valines (GTN) at the expense of isoleucine (ATY), methionine (ATR), and phenylalanine (TTY). For polar amino acids, M. edulis proteins contain more glycine (GGN) at the expense of asparagine (AAY). Presumably, the bias toward G and T in the gene-containing strand of M. edulis has resulted in amino acid replacements within the tolerance of physio-chemical similarity.
Unassigned DNA
G. eborea mtDNA is very uncommon for lacking any large noncoding regions, as are usually inferred to contain the origin(s) of replication and transcription control signals. The largest noncoding region is only 24 nt between trnK and trnF. Next in size are the 19 nt gap between cox3 and trnG and the 18 nt between trnG and trnQ. Noncoding DNA of M. edulis mtDNA has been analyzed and described earlier (Hoffmann, Boore, and Brown 1992). There is no obvious conservation of either nucleotide identities or potential secondary structures between the mollusks' noncoding regions. Whatever regulatory elements may be present are apparently short, dispersed, and/or rapidly changing.
Phylogenetic Analysis
Figure 6A presents a 70% majority rule consensus tree of MP bootstrap analysis for the taxa outlined in table 1, which has a topology congruent with those from quartet puzzling (QP) and some of the Bayesian analysis. These analyses support the monophyly of the lophophorates, annelids, and brachiopods. Relationships among the major molluscan lineages, however, remain unresolved, as are those among mollusks, brachiopods, and annelids, despite using this relatively large data set of 2,420 confidently aligned amino acid positions. In contrast, relationships among the major groups of deuterostomes and of arthropods are well resolved and conforming to expectation from other analyses, bolstering the view that the relationships among the lophotrochozoan groups are especially difficult to resolve.
FIG. 6. Comparisons of various phylogenetic analyses of 2,420 aligned amino acid positions of 12 concatenated mitochondrial proteins. (A) 70% majority rule consensus tree of maximum parsimony bootstrap analysis. This tree is congruent with the quartet puzzling (QP) analyses, and the numerals at the nodes are MP bootstrap values followed by QP support from the mtREV24 model analysis (which were very similar to those using Blosum62). Results of Bayesian analyses (50% majority rule consensus) are shown using the Blosum62 (B), mtREV24 (C), mixed (D), and Poisson (E) models. Any node unmarked by a numeral has support of 100%
Results of MB analyses with different sets of prior probabilities for the amino acid model are depicted in figure 6, panels B–E. Concerns over the inability to resolve many of these phylogenetic relationships are exacerbated by the observation that the Bayesian analyses (unlike the MP and QP analyses) return substantially different topologies with high posterior probability values when the prior amino acid substitution models vary.
However, this data set of concatenated protein sequences does give consistent results for support of many metazoan clades regardless of the type of analysis performed (table 5). In contrast, analyses using each individual gene recovered only the deuterostome and protostome nodes with high levels of confidence (>70%) in six cases and not all methods were consistent. The arthropod clade was recovered by only three individual genes and by only one or two methods in the best of cases. The Lophotrochozoa clade was recovered only by Nad2 and only in the case of using the mixed model prior analysis.
Table 5 Support of Particular Metazoan Clades in Individual Gene (Excluding Atp8) Analyses Using Maximum Parsimony (MP), Quartet Puzzling (QP) with mtREV24, and MrBayes (MB) with a Mixed Model Prior Probability Assignment.
Although there are key nodes within the Lophotrochozoa that are still unresolved, one hopes that they will yield to further taxon sampling, and these results illustrate the importance of sequencing complete mitochondrial genomes, as opposed to using only single gene sequences, when analyzing deep divergences.
Acknowledgements
Thanks to K. Fraser and T. Collins for the gift of G. eborea DNA and to B. Dayrat for helpful comments. This work was supported by DEB-9807100 from the National Science Foundation. Part of this work was performed under the auspices of the U.S. Department of Energy, Office of Biological and Environmental Research, by the University of California, Lawrence Berkeley National Laboratory, under contract No. DE-AC03-76SF00098.
Literature Cited
Beagley, C. T., R. Okimoto, and D. R. Wolstenholme. 1999. Mytilus mitochondrial DNA contains a functional gene for a tRNASer(UCN) with a dihydrouridine arm-replacement loop and a pseudo-tRNASer(UCN) gene. Genetics 152:641-652.
Black, W. C., and R. L. Roehrdanz. 1998. Mitochondrial gene order is not conserved in arthropods: Prostriate and metastriate tick mitochondrial genomes. Mol. Biol. Evol. 15:1772-1785.
Boore, J. L. 1999. Animal mitochondrial genomes. Nucleic Acids Res. 27:1767-1780.
Boore, J. L. 2002. "Mitochondrial Gene Arrangement Source Guide.". (http://www.jgi.doe.gov/programs/comparative/Mito_top_level.html).
Boore, J. L., and W. M. Brown. 1994a. Complete DNA sequence of the mitochondrial genome of the black chiton, Katharina tunicata. Genetics 138:423-443.
Boore, J. L., and W. M. Brown. 1994b. Mitochondrial genomes and the phylogeny of mollusks. Nautilus 108:(Suppl. 2): 61-78.
Boore, J. L., and W. M. Brown. 1995. The complete DNA sequence of the mitochondrial genome of the annelid worm Lumbricus terrestris. Genetics 141:305-319.
Boore, J. L., and W. M. Brown. 1998. Big trees from little genomes: Mitochondrial gene order as a phylogenetic tool. Curr. Opin. Genet. Dev. 8:668-674.
Campbell, N. J. H., and S. C. Barker. 1998. An unprecedented major rearrangement in an arthropod mitochondrial genome. Mol. Biol. Evol. 15:1786-1787.
Cardon, L. R., C. Burge, D. A. Clayton, and S. Karlin. 1994. Pervasive CpG suppression in animal mitochondrial genomes. Proc. Natl. Acad. Sci. USA 91:3799-3803.
Castresana, J., G. Feldmaier-Fuchs, S.-I. Yokobori, N. Satoh, and S. P??bo. 1998. The mitochodrial genome of the hemichordate Balanoglossus carnosus and the evolution of deuterostome mitochondria. Genetics 150:1115-1123.
Folmer, O., M. Black, W. Hoeh, R. Lutz, and R. Vrijenhoek. 1994. DNA primers for amplification of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates. Mol. Mar. Biol. Biotech. 3:294-299.
Grande, C., J. Templado, J. L. Cervera, and R. Zardoya. 2002. The complete mitochondrial genome of the nudibranch Roboastra europaea (Mollusca: Gastropoda) supports the monophyly of opisthobranchs. Mol. Biol. Evol. 19:1672-1685.
Haszprunar, G. 2000. Is the Aplacophora monophyletic? A cladistic point of view. American Malacological Bulletin 15:115-130.
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.
Hickerson, M. J., and C. W. Cunningham. 2000. Dramatic mitochondrial gene rearrangements in the Hermit Crab Pagurus longicarpus (Crustacea, Anomura). Mol. Biol. Evol. 17:639-644.
Hoeh, W. R., K. H. Blakley, and W. M. Brown. 1991. Heteroplasmy suggests limited biparental inheritance of Mytilus mitochondrial DNA. Science 251:1488-1490.
Hoeh, W. R., D. T. Stewart, and S. I. Guttman. 2002. High fidelity of mitochondrial genome transmission under the doubly uniparental mode of inheritance in freshwater mussels (Bivalvia: Unionoidea). Evolution 56:2252-2261.
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.
Huelsenbeck, J. P., and F. Ronquist. 2001. MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17:754-755.
Kim, S.-H., E.-Y. Je, and D.-W. Park. 1999. Crassostrea gigas mitochondrial DNA. GenBank accession number AF177226.
Kurabayashi, A., and R. Ueshima. 2000a. Complete sequence of the mitochondrial DNA of the primitive opisthobranch Pupa strigosa: Systematic implications of the genome organization. Mol. Biol. Evol. 17:266-277.
Kurabayashi, A., and R. Ueshima. 2000b. Partial mitochondrial genome organization of the heterostrophan gastropod Omalogyra atomus and its systematic significance. Venus (Jap. J. Malac.) 59:7-18.
Lavrov, D. V., W. M. Brown, and J. L. Boore. 2000. A novel type of RNA editing occurs in the mitochondrial tRNAs of the centipede Lithobius forticatus. Proc. Natl. Acad. Sci. USA 97:13738-13742.
Le, T. H., D. Blair, and T. Agatsuma, et al. (14 co-authors). 2000. Phylogenies inferred from mitochondrial gene orders-a cautionary tale from the parasitic flatworms. Mol. Biol. Evol. 17:1123-1125.
Lecanidou, R., V. Douris, and G. Rodakis. 1994. Novel features of metazoan mtDNA revealed from sequence analysis of three mitochondrial DNA segments of the land snail Albinaria turrita (Gastropoda: Clausiliidae). J. Mol. Evol. 38:369-382.
Okazaki, M., and R. Ueshima. 2001. Evolutionary diversity between the gender-associate mitochondrial DNA genomes of freshwater mussels. Genbank accession numbers AB055624 (male haplotype) and AB055625 (female haplotype).
Okazaki, M., and R. Ueshima. 2002. Gender-associated mtDNA of Tapes philippinarum. Genbank accession number NC_003354.
Okimoto, R., J. L. Macfarlane, D. O. Clary, and D. R. Wolstenholme. 1992. The mitochondrial genomes of two nematodes, Caenorhabditis elegans and Ascaris suum. Genetics 130:471-498.
Passamonti, M., and V. Scali. 2001. Gender-associated mitochondrial DNA heteroplasmy in the venerid clam Tapes philippinarum (Mollusca Bivalvia). Curr. Genet. 39:117-124.
Perna, N. T., and T. D. Kocher. 1995. Patterns of nucleotide composition at fourfold degenerate sites of animal mitochondrial genomes. J. Mol. Evol. 41:353-358.
Rawlings, T., T. Collins, and R. Bieler. 2001. A major mitochondrial gene rearrangement among closely related species. Mol. Biol. Evol. 18:1604-1609.
Runnegar, B., and J. Pojeta. 1974. Molluscan phylogeny: The paleontological viewpoint. Science 186:311-317.
Salvini-Plawen, L. V. 1985. Early evolution and the primitive groups. Pp. 59–150 in E. R. Trueman and M. R. Clark, eds. Evolution. Academic Press, Orlando, Fla.
Salvini-Plawen, L. V., and G. Steiner. 1996. Synapomorphies and plesiomorphies in higher classification of Mollusca. Pp. 29–51 in J. Taylor, ed. Origin and evolutionary radiation of the Mollusca. Oxford University Press, Oxford.
Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning, a laboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York.
Sasuga, J., S.-I. Yokobori, M. Kaifu, T. Ueda, K. Nishikawa, and K. Watanabe. 1999. Gene contents and organization of a mitochondrial DNA segment of the squid Loligo bleekeri. J. Mol. Evol. 48:692-702.
Scheltema, A. H. 1993. Aplacophora as progenetic aculiferans and the coelomatic origin of mollusks as the sister taxon of Sipuncula. Biol. Bull. 184:57-78.
Scheltema, A. H. 1996. Phylogenetic position of Sipuncula, Mollusca and the progenetic Aplacophora. Pp. 53–58 in J. Taylor, ed. Origin and evolutionary radiation of the Mollusca. Oxford University Press, Oxford.
Shao, R., N. J. H. Campbell, and S. C. Barker. 2001. Numerous gene rearrangements in the mitochondrial genome of the wallaby louse, Heterodoxus macropus (Phthiraptera). Mol. Biol. Evol. 18:858-865.
Steiner, G., and H. Dreyer. 2002. Scaphopoda and Cephalopoda are sister taxa—an evolutionary scenario. Abstract for the American Malacological Society Annual Meeting, Charleston, SC.
Steiner, G., and S. Hammer. 2000. Molecular phylogeny of the Bivalvia inferred from 18S rDNA sequences with particular reference to the Pteriomorpha. Pp. 11–29 in E. M. Harper, J. D. Taylor, and J. A. Crame, eds. The evolutionary biology of the Bivalvia. The Geological Society of London Special Publications, London.
Steiner, G., and M. Müller. 1996. What can 18S rDNA do for bivalve phylogeny? J. Mol. Evol. 43:58-70.
Stewart, D. T., C. Saavedra, R. R. Stanwood, A. O. Ball, and E. Zouros. 1995. Male and female mitochondrial DNA lineages in the blue mussel (Mytilus edulis) species group. Mol. Biol. Evol. 12:735-747.
Strimmer, K., and A. von Haeseler. 1997. Puzzle. Maximum likelihood analysis for nucleotide and amino acid alignments. Zoologisches Institut, Munchen, Germany.
Swofford, D. L. 2001. PAUP*: phylogenetic analysis using parsimony (*and other methods). Beta Version 4.0b8. Sinauer Associates, Sunderland, Mass.
Terrett, J. A., S. Miles, and R. H. Thomas. 1996. Complete DNA sequence of the mitochondrial genome of Cepaea nemoralis (Gastropoda: Pulmonata). J. Mol. Evol. 42:160-168.
Tomita, K., S. Yokobori, T. Oshima, T. Ueda, and K. Watanabe. 2002. The cephalopod Loligo bleekeri mitochondrial genome: Multiplied noncoding regions and transposition of tRNA genes. J. Mol. Evol. 54:486-500.
Waller, T. R. 1998. Origin of the molluscan class Bivalvia and a phylogeny of major groups. Pp. 1–47 in P. A. Johnston and J. W. Haggart, eds. Bivalves: An eon of evolution—plaeobiological studies honoring Norman D. Newell. University of Calgary Press, Calgary, Canada.
Wilding, C. S., P. J. Mill, and J. Grahame. 1999. Partial sequence of the mitochondrial genome of Littorina saxatilis: Relevance to gastropod phylogenetics. J. Mol. Evol. 48:348-359.
Yamazaki, N., R. Ueshima, and J. Terrett, et al. (12 co-authors). 1997. Evolution of pulmonate gastropod mitochondrial genomes: comparisons of gene organizations of Euhadra, Cepaea and Albinaria and implications of unusual tRNA secondary structures. Genetics 145:749-758.
Yokobori, S., and S. P??bo. 1995. Transfer RNA editing in land snail mitochondria. Proc. Natl. Acad. Sci. USA 92:10432-10435.(Jeffrey L. Boore*,, Monic)