The Novel Mitochondrial Gene Arrangement of the Cattle Tick, Boophilus microplus: Fivefold Tandem Repetition of a Coding Region

Transcription

1 The Novel Mitochondrial Gene Arrangement of the Cattle Tick, Boophilus microplus: Fivefold Tandem Repetition of a Coding Region Nick J. H. Campbell* and Stephen C. Barker* *Department of Parasitology and Centre for Molecular and Cellular Biology, University of Queensland, Brisbane, Australia We sequenced across all of the gene boundaries in the mitochondrial genome of the cattle tick, Boophilus microplus, to determine the arrangement of its genes. The mtdna of B. microplus has a coding region, composed of trna Glu and 60 bp of the 3 end of ND1, that is repeated five times. Boophilus microplus is the first coelomate animal known to have more than two copies of a coding sequence. The mitochondrial genome of B. microplus has other unusual features, including (1) reduced T arms in trnas, (2) an AT bias in codon use, (3) two control regions that have evolved in concert, (4) three gene rearrangements, and (5) a stem-loop between trna Gln and trna Phe. The short T arms and small control regions (CRs) of B. microplus and other ticks suggest strong selection for small genomes. Imprecise termination of replication beyond its origin, which can account for the evolution of tandem repeats of coding regions in other mitochondrial genomes, cannot explain the evolution of the fivefold repeated sequence in the mitochondrial genome of B. microplus. Instead, slipped-strand mispairing or recombination are the most plausible explanations for the evolution of these tandem repeats. Introduction The genomes of animal mitochondria typically have genes and are kb long (Wolstenholme 1992). Based on the mitochondrial genomes of a range of vertebrates and arthropods, it is commonly assumed that the arrangement of the major genes, the protein-coding genes and the 2 ribosomal RNA genes, is highly conserved within animal phyla (Wolstenholme 1992; Boore et al. 1995; Macey et al. 1997; Staton, Daehler, and Brown 1997). However, studies of nematodes, mollusks, and now ticks cast doubt on the generality of this assumption (Okimoto et al. 1992; Boore and Brown 1994; Terrett, Miles, and Thomas 1996; Black and Roehrdanz 1998; Campbell and Barker 1998). Rearrangements of mitochondrial genes are powerful phylogenetic markers (Boore et al. 1995; Macey et al. 1997; Black and Roehrdanz 1998; Boore, Lavrov, and Brown 1998; Campbell and Barker 1998). Most emphasis has been placed on ancient phylogenetic relationships. However, gene rearrangements could be found in young taxa too, since each rearrangement of mtdna presumably originates in a single genome in a single individual. Indeed, different gene arrangements have been found in a genus of frogs (Rana; Macey et al. 1997), a genus of sea cucumbers (Cucumaria; Arndt and Smith 1998) and a family of ticks (Ixodidae; Black and Roehrdanz 1998; Campbell and Barker 1998). Tandem duplications followed by deletions of genes probably cause many gene rearrangements (Moritz and Brown 1987; Macey et al. 1997, 1998). The duplication of genes without the deletion of additional copies also leads to new gene arrangements (Moritz and Brown 1987; Stanton et al. 1994; Macey et al. 1998). Tandem repeats of coding sequences with more than two copies are rare in metazoan mtdna. Indeed, they have only been found in one species of nematode (Beck- Key words: Boophilus microplus, mitochondrial genome, Ixodida, gene order, transfer RNA, tandem repeats. Address for correspondence and reprints: Nick J. H. Campbell, Department of Parasitology, The University of Queensland, Brisbane Q. 4072, Australia. n.campbell@mailbox.uq.edu.au. Mol. Biol. Evol. 16(6): by the Society for Molecular Biology and Evolution. ISSN: Azevedo and Hyman 1993) and never in coelomate animals, including vertebrates and arthropods, despite the large number of studies. In contrast, tandem repeats of noncoding sequences with more than two copies are common (see Zhang and Hewitt [1997] for many examples for arthropods). We determined the arrangement of genes in the mitochondrial genome of the cattle tick, Boophilus microplus. This genome has five repeats of a coding sequence: the gene for trna Glu plus 60 bp of the NADH dehydrogenase subunit 1 (ND1) gene. Ours is the first report of a repeated coding region with more than two copies in the mtdna of a coelomate animal. This repeated coding sequence is also the first in a metazoan mtdna that does not include or lie near the end of the CR. Materials and Methods We purified mtdna from eggs of a laboratory strain of B. microplus (N strain) by standard CsCl gradient separation (Dowling et al. 1996). We used long PCR (Cheng et al. 1994) to amplify the entire mitochondrial genome in two large overlapping sections using the primers (1) SR-J (5 -CTGGGCA- TAGTGGGGTATCTAATCC-3 ) with C2-N-3134 (5 - TGATCATGAAAAAAAATTATTTGTTC-3 ) and (2) C1-J-1633 (5 -GGTAATGATCAAATTTATAAT-3 ) with SR-N (5 -AAACTAGGATTAGATACCC- 3 ), named according to where they anneal by gene, strand, and 3 base number relative to the Drosophila yakuba sequence (after Simon et al. 1994). We amplified these fragments with Elongase (GIBCO: BRL), as recommended by the manufacturer, with the following cycling conditions: 94 C for 1 min; 35 cycles of 30 s at 95 C,30sat40 C, and min at 68 C (i.e., incrementing by 14 s per cycle for cycles 1 30); plus an additional 7 min at 68 C. We also amplified smaller (nested) fragments from these large PCR products (conditions as above except that we used 55 C instead of 40 C). We sequenced both strands directly from PCR products generated as described, with the above primers 732

2 Gene Arrangement of the Cattle Tick Mitochondrial Genome 733 FIG. 1. Mitochondrial gene arrangements of the cattle tick Boophilus microplus and other chelicerates. These circular genomes have been linearized for easy comparison. Genes are transcribed from left to right unless otherwise indicated by right to left arrows under their abbreviations. Gene abbreviations: ND NADH dehyrogenase subunit, CO cytochrome c oxidase subunit, A6 ATP synthase 6, A8 ATP synthase 8, Cytb cytochrome b apoenzyme, 12S small-subunit ribosomal RNA, 16S large-subunit ribosomal RNA, CR control region. The trna genes are labeled with the one-character abbreviation of the amino acid their transcripts transfer. Those trna genes labeled with two-character abbreviations are as follows: L1 gene for trna Leu(UUR),L2 gene for trna Leu(CUN),S1 gene for trna Ser(UCN),S2 gene for trna Ser(AGN), E1 putative functional trna Glu gene, E2 E5 putative trna Glu pseudogenes. Double-headed arrows are drawn between genes or groups of genes that vary in position between the genomes shown; the circling arrow indicates an inversion. Dashed lines indicate the tandem repeats of trna Glu in B. microplus compared with the single copy in L. polyphemus and I. hexagonus. The mitochondrial gene arrangement of R. sanguineus is not shown but is identical to B. microplus except for the E1 E5 section. Lines above the mitochondrial genome of B. microplus denote the sections sequenced for this study. plus internal primers. PCR fragments were purified with Wizard columns (Promega), and we used ng in each cycle-sequencing reaction (DyeDeoxy terminator, PE Applied Biosystems). An ABI 373A resolved the sequencing fragments. We identified protein-coding and ribosomal RNA sequences by BLAST searches of GenBank (Altschul et al. 1990) and by alignment with sequences from Limulus polyphemus (a horseshoe crab; Staton, Daehler, and Brown 1997) and D. yakuba (Clary and Wolstenholme 1985) using CLUSTAL W (Thompson, Higgins, and Gibson 1994). We identified trna genes from their secondary structures and anticodons. All boundaries between genes in the mitochondral genome of B. microplus were sequenced (GenBank accession numbers AF AF110622; see also AJ006038, AF067440, and AF from Campbell and Barker 1998). Results The mtdna of B. microplus: Size, Base Composition, Codon Use and Overlaps of Genes We estimate from the sizes of the two major PCR fragments that the mitochondrial genome of B. microplus is about 15 kb long. This genome is similar in size to those of the two other hard ticks studied (family Ixodidae: Ixodes hexagonus 14,539 bp, Rhipicephalus sanguineus 14,710 bp; Black and Roehrdanz 1998) but smaller than those of most arthropods (e.g., D. yakuba 16,019 bp [Clary and Wolstenholme 1985], Artemia franciscana 15,822 bp [Valverde et al. 1994], L. polyphemus 16 kb [Staton, Daehler, and Brown 1997]). We sequenced about half of the mitochondrial genome of B. microplus (7,565 bp). All 37 genes expected were identified: 13 protein-coding genes, 2 ribosomal RNA genes, and 22 trna genes; we sequenced across the boundaries between all of these genes (see fig. 1 for the order and direction of transcription of genes). The base composition of the strand that encodes the majority (23 of the 37) of genes (A: 40.4%, T: 42.0%, C: 9.9%, G: 7.7%) is similar to that of R. sanguineus (A: 37.6%, T: 40.3%, C: 12.1%, G: 9.9%); the overall G C contents are also similar for these species (B. microplus: 17.6%, R. sanguineus: 21%). The mitochondrial genomes of these ticks, which are from the subfamily Rhipicephalinae in the Metastriata (one of the two lineages of hard ticks), are more AT-rich than those of other chelicerates (I. hexagonus [Prostriata] G C 27.4%; L. polyphemus G C 32.2%) but are similar to those of insects (e.g., D. yakuba G C 21.4%; Clary and Wolstenholme 1985). AT-rich codons are much more common in the mtdna of metastriate B. microplus than in those of the

3 734 Campbell and Barker prostriate hard tick I. hexagonus (Black and Roehrdanz 1998) and the horseshoe crab L. polyphemus (Staton, Daehler, and Brown 1997). The ratio of the number of GC-rich codons (Pro, Ala, Arg, and Gly) to AT-rich codons (Phe, Ile, Met, Tyr, Asn, and Lys) for B. microplus (0.15) is about a third that for L. polyphemus (0.43; calculated from Staton, Daehler, and Brown 1997) and half that for I. hexagonus (0.32; calculated from Black and Roehrdanz 1998). This is the lowest ratio reported from an arthropod (cf. Flook, Rowell, and Gellissen 1995a). Translation of mitochondrial protein-coding genes is initiated by three different codons in B. microplus: by ATT in ND1, ND2, ND3, ND5, ND6, and A8; by ATG in ND4L, A6, Cytb, COII, and COIII; and by ATA in ND4 and COI (see fig. 1 legend for key to abbreviations used here and below). Translation is apparently terminated by incomplete codons in seven of the 13 protein-coding genes: by TA in ND4L, Cytb, and A8 and by T in COII, COIII, ND1, and ND5. These termination codons are presumably completed posttranscriptionally by polyadenylation (see Ojala, Montoya, and Attardi 1981). TAA terminates the other six protein-coding genes, ND2, ND3, ND4, ND6, COI, and A6. Two pairs of genes overlap in the mitochondrial genome of B. microplus: (1) trna Lys and trna Asp and (2) trna Arg and trna Asn. In both cases, the genes that overlap are encoded by the same strand. Mitochondrial genes that overlap have also been found in other animals; for example, in some mollusks, annelids, and L. polyphemus (Boore and Brown 1994, 1995; Staton, Daehler, and Brown 1997). In these species and B. microplus, functional transcripts could be produced if the RNA is edited enzymatically (e.g., Yokobori and Pääbo 1995). If we assume translation of ND4L and A8 terminated by complete stop codons (i.e., TAA or TAG), then A8 and A6 overlap, as do ND4L and ND4. Overlap between A8 and A6 is common in metazoan mtdnas. In these cases, functional proteins may be translated from a single mrna with a start site which initiates translation of A6 from within this mrna (Wolstenholme 1992). ND1 and one of the genes for trna Glu also overlap by 2 bp, but, since these genes are encoded by different strands, complete transcripts of one gene could be produced without affecting the transcripts of the other (see below). Transfer RNA Genes (trnas) In B. microplus, we found the 22 trnas that are generally found in the mtdna of Metazoa (fig. 2). However, B. microplus has four extra copies of trna Glu (fig. 2g and below). The mitochondrial genome of B. microplus has 14 trnas with only 2 3 bp in the T C (T) stem (total T-arm length 7 12 bp; fig. 2a c, e, g i, k, o p, r t, and v). Only three trnas have 5-bp T stems; this is the most common size in the mitochondrial trnas of animals (Kumazawa and Nishida 1993; fig. 2f, m, and q). Four other characteristics of the trnas of B. microplus are unusual: (1) trna Arg has a T-T mismatch at the first nucleotide pair (positions 1 and 71; Sprinzl et al. 1996) of the amino acid acceptor stem (AA stem; fig. 2b); (2) trna Gln has a G at position 8 (the most 5 of the two nucleotides between the AA stem and the dihydrourine [DHU or D] stem; fig. 2f), whereas almost all other mitochondrial trnas of Metazoa have a T or an A here (Sprinzl et al. 1996); (3) trna Phe has either a reduced variable loop between the T and AC (anticodon) arms (1 bp, compared to the usual 3 5 bp, as in fig. 2o; Kumazawa and Nishida 1993) or a loop of bases which replaces the T arm and the variable loop between the AC and T arms, i.e., a TV-replacement loop (Wolstenholme et al. 1987; Okimoto et al. 1992); and (4) trna Cys has a secondary structure which is similar to that of trna Phe (fig. 2e). The anticodons of B. microplus trnas are identical to those of the two other ticks studied (R. sanguineus and I. hexagonus; Black and Roehrdanz 1998); all three differ from L. polyphemus in their trna Ser(AGN) anticodon, which is TCT (fig. 2q) instead of GCT (Staton, Daehler, and Brown 1997). Noncoding Regions Like R. sanguineus, but unlike I. hexagonus, (Black and Roehrdanz 1998), there are two large ( 300 bp) noncoding regions in the mitochondrial genome of B. microplus. One is between the 12S rrna and trna Ile genes (CR#1; fig. 1) in a position homologous to that of the CR in other arthropods (CR in L. polyphemus and I. hexagonus; fig. 1). The other is between trna Leu(CUN) and trna Cys (CR#2; fig. 1). CR#1 and CR#2 differ at only 4 nt: all differences are in the first 5 bp at the 12S and trna Leu(CUN) ends of CR#1 and CR#2, respectively. If we define the first and last nucleotides of CR#1 and CR#2 as the nucleotides immediately adjacent to the last and first base pairs of the adjacent genes, respectively, then CR#1 is 3 nt longer at the trna Ile end than is CR#2 (fig. 1). Alternatively, the first three nucleotides of t- RNA Cys might also be the last three nucleotides of CR#1. The next largest noncoding region in the mitochondrial genome of the cattle tick is the 24-nt segment between trna Gln and trna Phe, which appears to form part of a stem-loop (see Discussion). Gene Arrangement There is a large ( 5 kb) rearrangement in the mitochondrial genome of B. microplus (Campbell and Barker 1998) relative to the putative pleisiomorphic arrangement of arthropods (Staton, Daehler, and Brown 1997; L. polyphemus and I. hexagonus in fig. 1). ND5, ND4, ND4L, ND6, Cytb, and five trna genes on the one hand and ND1, 16S, 12S, and four trna genes on the other, have swapped positions relative to other arthropods. Furthermore, we found that in B. microplus, trna Leu(CUN) has moved from between trna Leu(UUR) and 16S rrna to between trna Ser(UCN) and CR#2, and trna Cys has been inverted and moved from between trna Trp and trna Tyr to between CR#2 and trna Met (see fig. 1). These arrangements are also found in R. sanguineus and in representatives of all subfamilies of the Metastriata (Black and Roehrdanz 1998; Campbell and Barker 1998). Fivefold Tandem Repeat Including trna Glu The cattle tick mitochondrial genome has 126 bp of coding region that is repeated five times and arranged

4 Gene Arrangement of the Cattle Tick Mitochondrial Genome 735 FIG. 2. Putative secondary structures of the mitochondrial trna genes of B. microplus. Arms in trnas (clockwise from top): amino acid acceptor (AA) arm, T C (T) arm, anticodon (AC) arm, and dihydrourine (DHU or D) arm. The trna genes are labeled with the three-letter abbreviation of the amino acid their transcripts transfer, their full names, and their abbreviations. All sequences are from 5 to 3, as shown for trna Ala. Lines indicate Watson-Crick bonds, and dots indicate G-T bonds. The possible secondary structures of the four additional copies of of trna Glu (E2 E5) are not shown but are identical to E1 except for a C-T mismatch in the first nucleotide pair of the AA stem. Alternative secondary structures for trna Cys (e) and trna Phe (o) would have single loops instead of T-arms and variable loops, i.e., TV-replacement loops.

5 736 Campbell and Barker FIG. 3. Region of mtdna of B. microplus that has five tandem repeats of the trna Glu gene (E1 E5) and 60 bp of the 3 end of ND1. Sequences are given 5 to 3 from the strand that encodes the majority (23 of the 37) of genes in this mitochondrial genome. The extent of each trna gene is indicated above the nucleotide sequence. The positions of the anticodons in each trna are indicated by ***. The repeated trnas are labeled as follows: E1 putative functional trna Glu, E2 E5 putative trna Glu pseudogenes. The amino acid sequence of ND1 is indicated by single letter abbreviations under the second codon positions. Note that ND1 ends with a T (an incomplete stop codon) on the opposite strand to that presented, i.e., complementary to the A indicated by ˆ. The nucleotides that mismatch in the AA stems of E2 E5 are underlined. Nucleotides that differ in one of the five tandem repeats are indicated by an X above or below the sequence. The 25 bp that is also repeated between the trna Leu(UUR) and 16S rrna genes is in lowercase in the fifth tandem repeat, where it codes for the last eight amino acids in ND1. in tandem between trna Ser(AGN) and ND1 (E1 E5 in fig. 1). The repeat consists of the gene for trna Glu (66 bp) plus a 60-bp section of the 3 end of ND1 (fig. 3). The middle three repeats are identical, whereas the first has an A instead of a C as the most 5 base in the trna Glu. The last repeat differs from the others at three of its final six base positions (fig. 3). The 60 bp of ND1 apparently codes for amino acids in the fifth repeat, whereas in the other four, it is apparently noncoding sequence. Part of the section of ND1 in the tandem repeats (the first 25 bp; fig. 3) is also repeated between the t- RNA Leu(UUR) and 16S rrna genes. Discussion Mitochondrial trnas of B. microplus Metazoa generally have mitochondrial trnas with cloverleaf-like secondary structures (Kumazawa and Nishida 1993). trna Serine (AGN) is the one exception; it lacks the D arm of the other 21 mitochondrial trnas of metazoans (e.g., S2, fig. 2q; see Kumazawa and Nishida 1993). Nematodes have trnas with secondary structures that are different from those of other animals: the T arm and the variable loop between the AC and T arms have apparently been replaced with a single loop of bases in most nematode trnas, i.e., there is a TVreplacement loop (Wolstenholme et al. 1987; Okimoto et al. 1992; Watanabe et al. 1994). Reduced or absent T arms have now also been reported in coelomates, both deuterostomes (snakes; Kumazawa et al. 1996) and protostomes (snails; Yamazaki et al. 1997). The reduction in the lengths of T arms, therefore, may be a relatively common modification of mitochondrial trnas in metazoans. Boophilus microplus has 14 mitochondrial trnas that have T arms that are shorter than usual (2 3 bp in T stems; 7 12 bp total length; fig. 2). In addition, in trna Phe and trna Cys, it seems that TV-replacement loops may have replaced T arms (fig. 2). T arms are also shorter than usual in the other two hard ticks studied so far: I. hexagonus and R. sanguineus have 14 and 16 mitochondrial trnas, respectively, with only 2 3 bp in their T stems (Black and Roehrdanz 1998). In contrast, only three mitochondrial trnas from the horseshoe crab L. polyphemus have 2 3 bp in their T stems (Staton, Daehler, and Brown 1997). Mitochondrial genomes of animals lack introns and substantial intergenic spacers (Wolstenholme 1992). This suggests that there is selection for compact mitochondrial genomes. The maintenance of length heteroplasmy in some species, however, indicates that compact genomes are not always selected for (see Rand 1993). Where the selection pressure for a smaller mitochondrial genome is high, we speculate that selection will reduce those parts of the genome whose function is least impaired by a reduction in size. Given that even trnas without T arms can form tertiary structures (Watanabe et al. 1994), and thus presumably function adequately, T arms of trnas are obvious candidates to be reduced in size by strong selection pressure. Strong selection pressure for compact mitochondrial genomes probably caused the reduction of T arms in the mitochondrial trnas of the three ticks studied so far, B. microplus, R. sanguineus, and I. hexagonus (this study; Black and Roehrdanz 1998). The small sizes of the CRs in ticks relative to those of other arthropods (Valverde et al. 1994; Zhang and Hewitt 1997) is further evidence that selection favors smaller genomes in ticks (but see below). The T-T mismatch in the first nucleotide pair of the AA stem of the trna Arg in B. microplus (fig. 2b), which

6 Gene Arrangement of the Cattle Tick Mitochondrial Genome 737 is also present in R. sanguineus, has not been found in other arthropods (e.g., see Sprinzl et al. 1996). This unusual mismatch coincides with the apparent overlap of this gene with trna Asn. Overlaps of sequences of adjacent trna genes are common in the mitochondrial genomes of some snails; they often accompany AA stems that have apparently been destabilized by mismatches (Terrett, Miles, and Thomas 1996; Yamazaki et al. 1997). In at least one species, RNA-editing corrects some of these mismatches (Euhadra herklotsi; Yokobori and Pääbo 1995). RNA editing might also correct the mismatch we found in B. microplus. The unusual G at position 8 in the trna Gln of B. microplus is also present in R. sanguineus, but not in I. hexagonus, which has a T, like all other arthropod sequences (Sprinzl et al. 1996). Given the extreme conservation of a T at this position in Metazoa, we suggest that there was a T G substitution in a common ancestor of B. microplus and R. sanguineus. If so, the G at this position may be a synapomorphy for a group of metastriate ticks. This would be a strong synapomorphy, because the mitochondrial genomes of Metastriata are ATrich. Three differences in the secondary strucures of the trnas of B. microplus and R. sanguineus may also be synapomorphies for groups of species in the Rhipicephalinae. Whereas R. sanguineus has apparently had the D arm of trna Cys replaced with a loop (the pleisiomorphic condition for the Metastriata; Black and Roehrdanz 1998), based on our putative secondary structure, B. microplus has apparently secondarily regained a D arm, albeit with only two Watson-Crick base-pairings (fig. 2e). The trna Phe and trna Cys of R. sanguineus also seem to lack the unusual features of the putative secondary structures of these trnas in B. microplus (fig. 2e and o). FIG. 4. Putative stem-loops between trna Gln and trna Phe in B. microplus (this study) and R. sanguineus (our interpretation of data from Black and Roehrdanz 1998). Lines indicate Watson-Crick bonds, and dots indicate G-T bonds. The extent of the flanking trna genes is indicated by boxes. Evolution of the Gene Arrangement of Metastriate Ticks Tandem duplications of coding regions are the likely precursor to many rearrangements in the mitochondrial genomes of Metazoa (Stanton et al. 1994; Moritz and Brown 1987; Macey et al. 1997, 1998). Stem-loops occur at the ends of many tandem duplications (Stanton et al. 1994). Structures like these signal the start of replication of mitochondrial genomes (Wong and Clayton 1986). Together, these observations suggest that errors in replication cause at least some tandem duplications. In B. microplus, we found a stem-loop at the boundary between the major segments of the mitochondrial genome that have been transposed in B. microplus and other metastriate ticks (fig. 4). This stem-loop is formed by the 24 nt between trna Gln and trna Phe (Q and F, respectively, in fig. 1) and the adjacent 5 6 nt of these genes (figs. 1 and 4). We also found this stemloop in the mtdna of R. sanguineus (fig. 4). Both of these secondary structures resemble the stem-loops at the origins of replication of the second strand in the other arthropods studied, in that they are also hairpins with AT-rich loops (see Van Raay and Crease 1994 for examples). However, most of the stem-loops at the origins of second-strand replication in arthropods are much longer than those we identified between trna Gln and trna Phe in these Rhipicephaline ticks (fig. 4; cf. Van Raay and Crease 1994). For most insects and crustaceans, the origin of second-strand replication is apparently in the CR (Clary and Wolstenholme 1985; Van Raay and Crease 1994; Zhang and Hewitt 1997; but see Valverde et al. 1994). Black and Roehrdanz (1998) propose that replication in the mtdnas of ticks starts at one of the stem-loops they found in the CRs of meta-

7 738 Campbell and Barker striate and prostriate ticks. There are also stem-loops in the CRs of B. microplus (data not shown). Either a stemloop in the CR or the stem-loop between trna Gln and trna Phe (see above) may be the origin of second-strand replication. Electrophoretic or electron microscopic studies are required to identify the origins of mitochondrial replication for these ticks unambiguously. Whether or not the stem-loop between the genes for trna Gln and trna Phe is the usual origin of replication of the second strand, the presence of this structure between the major segments that have been transposed, along with the presence of two CRs in metastriate ticks, suggests that a tandem duplication led to the large rearrangement of mtdna in this group. Phylogenetic Implications and Timing of Gene Rearrangements Apart from the tandemly repeated region that includes the gene for trna Glu (fig. 3; discussion below), the organization of the mitochondrial genome of B. microplus is identical to that of R. sanguineus and all other metastriate ticks studied. The unique arrangement of genes in the mitochondrial genome of metastriate ticks is thus a seemingly ironclad synapomorphy for a clade supported by morphology (Klompen et al. 1996, 1997) and phylogenies inferred from small subunit ribosomal RNA sequences (Black, Klompen, and Keirans 1997; Dobson and Barker 1999). Interestingly, however, two gene junctions in B. microplus and other metastriate ticks (trna Cys -trna Met and trna Trp -trna Tyr ; C-M and W-Y in fig. 1) are convergent with respect to those of the putative ancestor of protostomes (Boore et al. 1995). Thus, metastriate ticks provide the first example of a reversal and only the second of convergence in arthropod gene arrangements (Flook, Rowell, and Gellissen 1995b). So, as with other phylogenetic characters, sparse taxonomic sampling could lead to errors in the inference of character polarity of gene arrangements. However, it is reassuring that in both known cases of convergent gene arrangements in arthropods, the convergent evolution is obvious once a wider range of taxa are examined. Moreover, in both cases, the alternative phylogenetic relationships suggested by misinterpretation of the convergent arrangements were so novel that further sampling to check for convergence was the obvious next step. Boore et al. (1995) and Macey et al. (1997) are correct when they emphasize that convergent gene arrangements are highly unlikely, but we should add that because they are so rare, they are easy to identify when they do occur. The three gene rearrangements we found in B. microplus are present in both major lineages of the extant Metastriata (the distinctive Australian Aponomma and all other metastriate ticks; see Dobson and Barker 1999). The rearrangements, therefore, apparently occurred before these groups diverged but after the Metastriata and the Prostriata diverged (Black and Roehrdanz 1998; Campbell and Barker 1998). Balashov (1994) and Dobson and Barker (1999) argue that the simplest explanation for the cosmopolitan distribution of both metastriate and prostriate ticks is that these lineages diverged prior to the breakup of Pangaea. Furthermore, that one of the two major lineages in the Metastriata is restricted to Australia (the endemic Aponomma spp.) is consistent with divergence of these lineages before or at the time that Gondwana and Laurasia separated (210 MYA). Thus, the gene rearrangements we found in B. microplus, which diagnose metastriate ticks, occurred a very long time ago possibly more than 210 MYA. If the gene arrangement of metastriate ticks is at least 210 Myr old, then the two CRs present in B. microplus and other metastriate ticks (fig. 1) have persisted and evolved in concert for over 210 Myr (above). This is an extraordinarily long time, given the apparent selection for compact mitochondrial genomes in Metazoa (Wolstenholme 1992). Persistence of two copies of a locus is even more surprising, because the reduced sizes of the trnas and CRs in hard ticks suggests that selection pressure for compact genomes in these animals is even stronger than that in other Metazoa (above). Kumazawa et al. (1996) suggested that duplicate CRs in snakes could be explained if two CRs conferred a selective advantage by enhancing the efficiency of transcription and/or replication. This suggestion could also account for the apparent discrepancy of having two CRs in the otherwise very compact mitochondrial genomes of metastriate ticks. Metastriate ticks represent the third case of duplicate CRs that evolve in concert in the mitochondrial genomes of Metazoa; the other two are from snakes and sea cucumbers (Kumazawa et al. 1996; Arndt and Smith 1998). Since each of these groups is from a different phylum, each duplication was presumably an independent evolutionary event. We predict that duplicate CRs that evolve in concert will be found in other groups of animals too. Fivefold Tandem Repetition of a Coding Region in B. microplus All five trna Glu genes in B. microplus apparently form secondary structures that are stable (fig. 2g). In all but the first trna Glu, however, the first nucleotide pair in the AA stem is a C-T mismatch (figs. 2 and 3); in all other arthropods studied, this is an A-T (Sprinzl et al. 1996; L. polyphemus: Staton, Daehler, and Brown 1997; I. hexagonus and R. sanguineus: Black and Roehrdanz 1998). On this basis, we suggest that the four copies of trna Glu with the C-T mismatch do not produce fully functional transcripts. Provisionally, then, we classify the first trna Glu gene (E1; figs. 1 3) as the functional copy and the other four (E2 E5; figs. 1 3) as nonfunctional pseudogenes. Since R. sanguineus lacks the tandem repeats we found in B. microplus, these repeats presumably evolved after the most recent common ancestor of these two species. The genera Boophilus and Rhipicephalus are very closely related (possibly even paraphyletic; Murrell, Campbell, and Barker 1999). Thus, in contrast to the three gene rearrangements that are in all metastriate ticks (above), the tandem repeats we found in B. microplus must have evolved recently. We are studying other

8 Gene Arrangement of the Cattle Tick Mitochondrial Genome 739 species of Boophilus and Rhipicephalus to see if they have these five trna Glu genes too. Although R. sanguineus does not have the fivefold repeat that we found in B. microplus, it does share with B. microplus an additional copy of the last 25 bp at the 3 end of ND1 (fig. 3; GenBank accession number AF081829). The additional copy of these 25 bp is between the trna Leu(UUR) and 16S rrna genes (L1 and 16S in fig. 1). Since these 25 bp apparently are not part of a secondary structure in the 16S rrna (data not shown) we suggest that they are nonfunctional. However, the other copy encodes the last eight amino acids of ND1 in B. microplus (NCFNICIF in fig. 3) and R. sanguineus (GenBank accession number AF081829); thus, we suggest that this is the original copy. Note that in B. microplus, these 25 bp are also part of the 126-bp repeat which is present five times i.e., the fivefold tandem repeat discussed above. All six copies of the 25 bp repeat in B. microplus and the two copies of this repeat in R. sanguineus are identical. The original duplication apparently preceded the divergence of B. microplus and R. sanguineus, since they both have it. However, since I. hexagonus does not have duplicate sequences that might be homologous with the 25-bp repeats in B. microplus and R. sanguineus (GenBank accession number AF081828), this duplication must have occurred after the most recent common ancestor of B. microplus, R. sanguineus, and I. hexagonus. Slipped-strand mispairing (Levinson and Gutman 1987; Rand and Harrison 1989), transposition (Macey et al. 1997), and replication errors (Fumagalli et al. 1996; Kumazawa et al Macey et al. 1997; 1998) are thought to cause tandem repeats of mtdna. Prior to this study, all known tandemly repeated coding regions included or were near one end of the CR (Moritz and Brown 1987; Stanton et al. 1994), possibly reflecting occasional imprecise termination of replication beyond its origin. If the origins of replication in the mitochondrial genome of B. microplus are in the CR, the distance of the repeats from the CR (fig. 1) suggests that slippedstrand mispairing is a more likely source of the fivefold tandem repeat in B. microplus. Alternatively, given recent evidence that recombination may be more common in mitochondrial evolution than previously thought (Beck-Azevedo and Hyman 1993; Hyman and Azevedo 1996; Kumazawa et al. 1996; Thyagarajan, Padua, and Campbell 1996; Lunt and Hyman 1997; Arndt and Smith 1998), models based on recombination (Rand and Harrison 1989; Hoelzel, Hancock, and Dover 1993) could also explain the evolution of these tandem repeats. Acknowledgments We thank Damien Broderick for help with mitochondrial DNA purification, Peter Green for the gift of the tick eggs, and Chris Schneider for the primer SR-J and encouragement to try long PCR. We received valuable advice from W. Brown, J. Boore, C. Moritz, and Y. Kumazawa. W. Black IV and R. Roehrdanz kindly gave us a copy of their manuscript before it had been published. Funding was from an Australian Research Council grant to S.C.B. and N.J.H.C. LITERATURE CITED ALTSCHUL, S. F., W. GISH, W. MILLER, E. W. MYERS, and D. J. LIPMAN Basic local alignment search tool. J. Mol. Biol. 215: ARNDT, A., and M. J. SMITH Mitochondrial gene rearrangement in the sea cucumber genus Cucumaria. Mol. Biol. Evol. 15: BALASHOV, Y. S Importance of continental drift in the distribution and evolution of ticks. Entomol. Rev. 73: BECK-AZEVEDO, J. L., and B. C. HYMAN Molecular characterisation of lengthy mitochondrial DNA duplications from the parasitic nematode, Romanomermis culicivorax. Genetics 133: BLACK, W. C. IV, J. S. H. KLOMPEN, and J. E. KEIRANS Phylogenetic relationships among tick subfamilies (Ixodida: Ixodidae, Argasidae) based on the 18S nuclear rdna gene. Mol. Phylogenet. Evol. 7: BLACK, W. C. IV, and R. L. ROEHRDANZ Mitochondrial gene order is not conserved in arthropods: prostriate and metastriate tick mitochondrial genomes. Mol. Biol. Evol. 15: BOORE, J. L., and W. M. BROWN Complete DNA sequence of the mitochondrial sequence of the black chiton, Katharina tunicata. Genetics 138: Complete sequence of the mitochondrial DNA of the annelid worm, Lumbricus terrestris. Genetics 141: BOORE, J. L., T. M. COLLINS, D. STANTON, L. L. DAEHLER, and W. M. BROWN Deducing the pattern of arthropod phylogeny from mitochondrial DNA rearrangements. Nature 376: BOORE, J. L., D. V. LAVROV, and W. M. BROWN Gene translocation links insects and crustaceans. Nature 392: CAMPBELL, N. J. H., and S. C. BARKER An unprecedented major rearrangement in an arthropod mitochondrial genome. Mol. Biol. Evol. 15: CHENG, S., C. FOCKLER,W.M.BARNES, and R. HIGUCHI Effective amplification of long targets from cloned inserts and human genomic DNA. Proc. Natl. Acad. Sci. USA 91: CLARY, D. O., and D. R. WOLSTENHOLME The mitochondrial DNA molecule of Drosophila yakuba: nucleotide sequence, gene organisation and genetic code. J. Mol. Evol. 22: DOBSON, S. J., and S. C. BARKER Phylogeny of the hard ticks (Ixodidae) inferred from 18S rrna indicates the genus Aponomma is paraphyletic. Mol. Phylogenet. Evol. (in press). DOWLING, T. E., C. MORITZ, J.D.PALMER, and L. H. RIESE- BERG Nucleic acids III: analysis of fragments and restriction sites. Pp in D. H. HILLIS, C. MORITZ, and B. K. MABLE, eds. Molecular systematics. Sinauer, Sunderland, Mass. FLOOK, P., C. H. F. ROWELL, and G. GELLISSEN. 1995a. The sequence, organisation and evolution of the Locusta migratoria mitochondrial genome. J. Mol. Evol. 41: b. Homoplastic rearrangements of insect mitochondrial trna genes. Naturwissenschaften 82: FUMAGALLI, L., P. TABERLET, L.FAVRE, and J. HAUSSER Origin and evolution of homologous repeated sequences in

9 740 Campbell and Barker the mitochondrial control region of shrews. Mol. Biol. Evol. 13: HOELZEL, A. R., J. M. HANCOCK, and G. A. DOVER Generation of VNTRs and heteroplasmy by sequence turnover in the mitochondrial control region of two elephant seal species. J. Mol. Evol. 37: HYMAN, B. C., and J. L. B. AZEVEDO Similar evolutionary patterning among repeated and single copy nematode mitochondrial genes. Mol. Biol. Evol. 13: KLOMPEN, J. S. H., W. C. BLACK IV,J.E.KEIRANS, and J. H. OLIVER JR Evolution of ticks. Annu. Rev. Entomol. 41: KLOMPEN, J. S. H., J. H. OLIVER JR., J. E. KEIRANS, and P. J. HOMSHER A reevaluation of relationships in the Metastriata (Acari: Parasitiformes: Ixodidae). Syst. Parasitol. 38:1 24. KUMAZAWA, Y., and M. NISHIDA Sequence evolution of mitochondrial trna genes and deep-branch animal phylogenetics. J. Mol. Evol. 37: KUMAZAWA, Y., H. OTA, M. NISHIDA, and T. OZAWA Gene rearrangements in snake mitochondrial genomes: highly concerted evolution of control region-like sequences duplicated and inserted into a trna gene cluster. Mol. Biol. Evol. 13: LEVINSON, G., and G. A. GUTMAN Slipped strand misrepairing: a major mechanism for DNA evolution. Mol. Biol. Evol. 4: LUNT, D. H., and B. C. HYMAN Animal mitochondrial DNA recombination. Nature 387:247. MACEY, J. R., A. LARSON, N. B. ANAJEVA, Z. FANG, and T. J. PAPENFUSS Two novel gene orders and the role of light-strand replication in the rearrangement of the vertebrate mitochondrial genome. Mol. Biol. Evol. 14: MACEY, J. R., J. A. SCHULTE II, A. LARSON, and T. J. PAPEN- FUSS Tandem duplication via light strand synthesis may provide a precursor for mitochondrial genomic rearrangement. Mol. Biol. Evol. 15: MORITZ, C., and W. M. BROWN Tandem duplication in animal mitochondrial DNAs: variation in incidence and gene content among lizards. Proc. Natl. Acad. Sci. USA 84: MURRELL, A., N. J. H. CAMPBELL, and S. C. BARKER Mitochondrial 12S rdna indicates that the Rhipicephalinae (Acari: Ixodida: Ixodidae) is paraphyletic. Mol. Phylogenet. Evol. (in press). OJALA, D. J., J. MONTOYA, and G. ATTARDI trna punctuated model of RNA processing in human mitochondria. Nature 290: OKIMOTO, R., J. L. MACFARLANE, D. O. CLARY, and D. R. WOLSTENHOLME The mitochondrial genome of two nematodes, Caenorhabditis elegans and Ascaris suum. Genetics 130: RAND, D. M Endotherms, ectotherms, and mitochondrial genome-size variation. J. Mol. Evol. 37: RAND, D. M., and R. G. HARRISON Molecular population genetics of mtdna size variation in crickets. Genetics 121: SIMON, C., F. FRATI, A. BECKENBACH, B. CRESPI, H. LIU, and P. FLOOK Evolution, weighting, and phylogenetic utility of mitochondrial gene sequences and a compilation of conserved polymerase chain reaction primers. Ann. Entomol. Soc. Am. 87: SPRINZL, M., C. STEEGBORN, F. HUBEL, and S. STEINBERG Compilation of trna sequences and sequences of trna genes. Nucleic Acids Res. 24: STANTON, D. J., L. L. DAEHLER, C.C.MORITZ, and W. M. BROWN Sequences with the potential to form stemand-loop structures are associated with coding-region duplications in animal mitochondrial DNA. Genetics 137: STATON, J. L., L. L. DAEHLER, and W. M. BROWN Mitochondrial gene arrangement of the horseshoe crab Limulus polyphemus L.: conservation of major features among arthropod classes. Mol. Biol. Evol. 14: TERRETT, J. A., S. MILES, and R. H. THOMAS Complete DNA sequence of the mitochondrial genome of Cepaea nemoralis (Gastropoda: Pulmonata). J. Mol. Evol. 42: THOMPSON, J. D., D. G. HIGGINS, and T. J. GIBSON CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22: THYAGARAJAN, B., R. A. PADUA, and C. CAMPBELL Mammalian mitochondria possess homologous DNA recombination activity. J. Biol. Chem. 271: VALVERDE, J. R., B. BATUECAS, C. MORATILLA, R. MARCO, and R. GARESSE The complete mitochondrial DNA sequence of the crustacean Artemia franciscana. J. Mol. Evol. 39: VAN RAAY, T. J., and T. CREASE Partial mitochondrial DNA sequence of the crustacean Daphnia pulex. Curr. Genet. 25: WATANABE, Y. I., H. TSURUI, T. UEDA, R. FURUSHIMA, S. TA- KAMIYA, K.KITA, K.NISHIKAWA, and K. WATANABE Primary and higher order structures of nematode (Ascaris suum) mt trnas lacking either the T or D stem. J. Biol. Chem. 269: WOLSTENHOLME, D. R Animal mitochondrial DNA: structure and evolution. Int. Rev. Cytol. 141: WOLSTENHOLME, D. R., J. L. MACFARLANE, R. OKIMOTO, D. O. CLARY, and J. A. WAHLEITNER Bizarre trnas inferred from DNA sequences of mitochondrial genomes of nematode worms. Proc. Natl. Acad. Sci. USA 84: WONG, T. W., and D. A. CLAYTON DNA primase of human mitochondria is associated with structural RNA that is essential for enzymatic activity. Cell 45: YAMAZAKI, N., R. UESHIMA, J. A. TERRETT et al. (12 co-authors) Evolution of pulmonate gastropod mitochondrial genomes: comparisons of gene organisations of Euhadra, Cepaea and Albinaria and implications of unusual trna secondary structures. Genetics 145: YOKOBORI, S., and S. PÄÄBO Transfer RNA editing in land snail mitochondria. Proc. Natl. Acad. Sci. USA 92: ZHANG, D.-X., and G. HEWITT Insect mitochondrial control region: a review of its structure, evolution and usefulness in evolutionary studies. Biol. Syst. Ecol. 25: RICHARD THOMAS, reviewing editor Accepted February 15, 1999