* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Rare genomic changes and mitochondrial sequences
Genetic code wikipedia , lookup
Short interspersed nuclear elements (SINEs) wikipedia , lookup
Epitranscriptome wikipedia , lookup
Copy-number variation wikipedia , lookup
Essential gene wikipedia , lookup
Quantitative trait locus wikipedia , lookup
Transposable element wikipedia , lookup
Point mutation wikipedia , lookup
Nutriepigenomics wikipedia , lookup
Non-coding DNA wikipedia , lookup
Oncogenomics wikipedia , lookup
Genomic library wikipedia , lookup
History of genetic engineering wikipedia , lookup
Therapeutic gene modulation wikipedia , lookup
Gene expression programming wikipedia , lookup
Gene desert wikipedia , lookup
Public health genomics wikipedia , lookup
Expanded genetic code wikipedia , lookup
Ridge (biology) wikipedia , lookup
Maximum parsimony (phylogenetics) wikipedia , lookup
Epigenetics of human development wikipedia , lookup
Mitochondrial DNA wikipedia , lookup
Genomic imprinting wikipedia , lookup
Human genome wikipedia , lookup
Site-specific recombinase technology wikipedia , lookup
Biology and consumer behaviour wikipedia , lookup
Genome (book) wikipedia , lookup
Genome editing wikipedia , lookup
Metagenomics wikipedia , lookup
Designer baby wikipedia , lookup
Microevolution wikipedia , lookup
Pathogenomics wikipedia , lookup
Gene expression profiling wikipedia , lookup
Minimal genome wikipedia , lookup
Helitron (biology) wikipedia , lookup
Transfer RNA wikipedia , lookup
Artificial gene synthesis wikipedia , lookup
Molecular Phylogenetics and Evolution 57 (2010) 59–70 Contents lists available at ScienceDirect Molecular Phylogenetics and Evolution journal homepage: www.elsevier.com/locate/ympev Rare genomic changes and mitochondrial sequences provide independent support for congruent relationships among the sea spiders (Arthropoda, Pycnogonida) Susan E. Masta *, Andrew McCall, Stuart J. Longhorn 1 Department of Biology, Portland State University, P.O. Box 751, Portland, OR 97207, USA a r t i c l e i n f o Article history: Received 15 September 2009 Revised 15 June 2010 Accepted 25 June 2010 Available online 30 June 2010 Keywords: Chelicerata Arachnida Mitochondrial genomics tRNA secondary structure Gene rearrangements a b s t r a c t Pycnogonids, or sea spiders, are an enigmatic group of arthropods. Their unique anatomical features have made them difficult to place within the broader group Arthropoda. Most attempts to classify members of Pycnogonida have focused on utilizing these anatomical features to infer relatedness. Using data from mitochondrial genomes, we show that pycnogonids are placed as derived chelicerates, challenging the hypothesis that they diverged early in arthropod history. Our increased taxon sampling of three new mitochondrial genomes also allows us to infer phylogenetic relatedness among major pycnogonid lineages. Phylogenetic analyses based on all 13 mitochondrial protein-coding genes yield well-resolved relationships among the sea spider lineages. Gene order and tRNA secondary structure characters provide independent lines of evidence for these inferred phylogenetic relationships among pycnogonids, and show a minimal amount of homoplasy. Additionally, rare changes in three tRNA genes unite pycnogonids as a clade; these include changes in anticodon identity in tRNALys and tRNASer(AGN) and the shared loss of D-arm sequence in the tRNAAla gene. Using mitochondrial genome changes and tRNA structural changes is especially useful for resolving relationships among the major lineages of sea spiders in light of the fact that there have been multiple independent evolutionary changes in nucleotide strand bias among sea spiders. Such reversed nucleotide biases can mislead phylogeny reconstruction based on sequences, although the use of appropriate methods can overcome these effects. With pycnogonids, we find that applying methods to compensate for strand bias and that using genome-level characters yield congruent phylogenetic signals. Ó 2010 Elsevier Inc. All rights reserved. 1. Introduction Pycnogonida, or sea spiders, are arthropods with anatomical features so unique that their relationships to other members of Arthropoda have been enigmatic. Pycnogonids possess a proboscis for feeding, extreme abdominal reduction, reproductive organs on their legs, and ventral appendages termed ovigers with which males carry eggs. There are two main hypotheses regarding pycnogonid relationships to other arthropods (reviewed in Dunlop and Arango, 2005). One hypothesis is that Pycnogonida is the sister group to all other Arthropoda. Under this hypothesis the arthropods, excluding pycnogonids, have been named Cormogonida (Zrzavý et al., 1998). A second hypothesis is that Pycnogonida are members of Chelicerata, and are the sister group to the Xiphosura (horseshoe crabs) plus Arachnida. Anatomical and developmental * Corresponding author. Fax: +1 503 725 3888. E-mail address: [email protected] (S.E. Masta). 1 Current address: Department of Biology, National University of Ireland, Maynooth. County Kildare, Ireland. 1055-7903/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.ympev.2010.06.020 data have each yielded results that support both of these hypotheses, and therefore, it is still not known which hypothesis, if either, is correct. More data, including data from different sources, is necessary to resolve the placement of Pycnogonida within Arthropoda. Pycnogonids typically possess anterior appendages termed chelifores, and the lack of clear homology of these appendages with other arthropod appendages has fueled debate over the closest relatives of pycnogonids. If sea spiders are chelicerates, then chelifores should be homologous to the chelicerae of other chelicerates, as possession of chelicerae is one of the defining traits for the Chelicerata. However, if chelifores are not directly homologous to chelicerae, then this calls into question whether sea spiders are truly chelicerates. Some analyses based on developmental features have found evidence that these appendages are not direct homologues of chelicerae, and support the hypothesis that sea spiders are sister to all other arthropods (Maxmen et al., 2005). However, more recent data based on developmental expression patterns and neuroanatomy have shown that chelifores are likely chelicerae homologues (Brenneis et al., 2008; Jager et al., 2006; 60 S.E. Masta et al. / Molecular Phylogenetics and Evolution 57 (2010) 59–70 Manuel et al., 2006), consistent with the placement of sea spiders as sister to other chelicerates. Analyses of molecular data are likewise inconsistent in their support for these opposing two hypotheses. Some analyses have found support for pycnogonids being sister to all arthropods (Giribet et al., 2001), or sister to all arthropods except chelicerates (Regier et al., 2008), or sister to the myriapod arthropods (Mallatt and Giribet, 2006; Mallatt et al., 2004). However, most molecular analyses recover phylogenetic trees that are congruent with pycnogonids as a sister group to all other chelicerates, although often with low bootstrap support (Bourlat et al., 2008; Mallatt and Giribet, 2006; Regier et al., 2008). Phylogenomic analyses with expressed sequence tags yielded a tree with more support for pycnogonids as sister to other chelicerates, although taxon sampling was more limited (Dunn et al., 2008). Analyses of mitochondrial genome sequences have suggested a third hypothesis, that sea spiders are derived arachnids. The mitochondrial genomes of two species of sea spiders have previously been sequenced, and a third has been partially sequenced. In phylogenetic analyses of the protein-coding genes from these genomes, along with those of other arthropods, pycnogonids have been recovered as derived arachnids (Hassanin, 2006; Park et al., 2007; Podsiadlowski and Braband, 2006). Given the contradictory results from different sources of data, and the often sparse taxon sampling, it is necessary to further evaluate the phylogenetic placement of sea spiders. Despite the growing number of studies aimed at understanding the relatedness of pycnogonids to other arthropods, very few studies have attempted to infer relationships among the major pycnogonid lineages. Hedgpeth and Fry classified sea spider taxa into families (Fry, 1978; Hedgpeth, 1947), but did not use modern cladistic methods. More recent work by Arango has utilized both morphological and molecular characters to infer phylogenetic relationships within Pycnogonida (Arango and Wheeler, 2007; Arango, 2002, 2003). These phylogenetic studies have lent support to some of the previously defined families of sea spiders, but have also found relationships among pycnogonids that differ from the traditional taxonomy. Additionally, although many relationships among sea spiders were similar using morphological and molecular analyses, some relationships were supported only by molecular data. Arango proposed that more characters were necessary to clearly delineate the pycnogonid lineages (Arango and Wheeler, 2007; Arango, 2002, 2003). Using genomic-level characters to infer relatedness among taxa has become possible with the increased capability of sequencing large contiguous stretches of DNA. The use of data on gene arrangements to help deduce phylogenetic relationships has been particularly useful for inferring relationships among some metazoan taxa. Much of this work was pioneered with comparisons of mitochondrial genome sequences, as the small size of mitochondria made them an early target for studying genome evolution and making phylogenomic inferences (Boore, 2006; Boore and Brown, 1998; Boore et al., 1995; Larget et al., 2005). Moreover, as well as changes in primary sequence and genome arrangements, the structures of molecules can evolve, subject to the selective constraints imposed on them to function within the cell. One of the most highly conserved molecular structures is that of transfer RNA, which maintains a canonical cloverleaf shape across the three domains of life, Eukarya, Bacteria, and Archaea. However, metazoan mitochondrial tRNAs are typically much smaller than their nuclear counterparts, and in some cases may lack either the DHU or TWC arms present in the canonical tRNA structure (see reviews in Soll and RajBhandary, 1995; Wolstenholme, 1992a). Even among metazoans, however, the evolutionary loss of one of the paired stem regions in a mt tRNA gene is rare. Therefore, when organisms share similar non-canonical tRNA structures, it is likely due to shared evolutionary history. An unprecedented number of non-canonical tRNA structures have been found within arachnid mt genomes (Masta and Boore, 2008), but their structures have also proven to be phylogenetically informative within some groups, such as ticks (Murrell et al., 2003) and spiders (Masta and Boore, 2008). Besides potentially providing new types of genome structure characters for making phylogenetic inferences, mitochondrial genomes provide a rich source of sequences for analysis. Typical metazoan mitochondria encode the same 37 genes, consisting of 13 protein-coding genes, 22 tRNA genes, and 2 ribosomal RNA genes. Therefore, it is clear that each of these genes is homologous among metazoans, making full-genome sequence comparisons possible among animal taxa. However, while mt genome data have many useful attributes, there are limits on which genes or gene regions are useful. In part, this depends on the divergence times of the taxa used in the analysis. The rate of evolution in mitochondrial genes tends to be more rapid than in many nuclear loci, and therefore mt gene comparisons tend to show the effects of phylogenetic saturation sooner than other loci. Additionally, arthropod mt genomes tend to possess a pronounced A + T bias, exacerbating the effects of saturation. Because saturation can mislead phylogeny reconstruction, it is necessary to try to correct for the possible misleading effects of saturation. Possible corrective measures for protein-coding genes include omitting saturated third- and first-position nucleotides of codons, or using only translated amino acid data. Alternatively, coding third and first positions of codons as purines or pyrimidines (RY coding) has been found to help reduce the misleading effects of saturation found in mtDNA phylogenetic analyses (Phillips and Penny, 2003). Another potential challenge when using mtDNA sequences in phylogenetic analyses is uneven base composition among taxa included in the study. Genome-wide biases are present in metazoan mitochondrial genomes in that genes encoded on the strand associated with initiating replication tend to have a pronounced A + C bias (Reyes et al., 1998). However, within arthropods, and arachnids in particular, this bias is reversed in some taxa (Hassanin et al., 2005; Masta and Boore, 2004; Masta et al., 2009), so that mitochondrial genes of phylogenetically disjunct taxa can share similar compositional bias. This has the effect of causing the reversed-biased taxa to cluster together in phylogenetic analyses, unless measures are taken to ameliorate the misleading impact of such biases (Hassanin et al., 2005; Jones et al., 2007; Masta et al., 2009). Such measures include recoding some of the nucleotide sites in the codons of the protein-coding genes as purines or pyrimidines (Hassanin et al., 2005; Jones et al., 2007; Masta et al., 2009), or recoding amino acids into categories corresponding to their physiochemical properties (Masta et al., 2009). In this manuscript, we present new data to bear on the relationship of sea spiders to other arthropods, and on their relationships to one another. We have sequenced the complete suite of 13 mt protein-coding genes from three divergent sea spiders. This boosts the taxonomic coverage of pycnogonids with complete mt proteincoding genes to 5 taxa. We conduct phylogenetic analyses using the entire suite of 13 protein-coding genes, and include representative taxa from 8 arachnid orders. Previous analyses that concluded that pycnogonids were derived arachnids had representatives from only three arachnid orders. We compare gene arrangement of our three new pycnogonid mt genomes (one complete, and two missing sequence near the control region) to the two published pycnogonid mt genomes and infer the phylogenetic history of gene arrangements in this group. Gene order provides independent phylogenetic characters that support our findings based on sequences. We also compare secondary structures of the transfer RNA genes in these 5 mt genomes, and use these structures as phylogenetic characters. All three types of data are in agreement. Together, our data provide strong support for relation- S.E. Masta et al. / Molecular Phylogenetics and Evolution 57 (2010) 59–70 ships among the pycnogonids that have not been inferred based on anatomical characters. 2. Materials and methods 2.1. Specimens used, DNA extraction, amplification, and sequencing Specimens of the pycnogonids Tanystylum orbiculare (collected from the coast of Massachusetts; gift from Jeffrey Shultz), Colossendeis sp. (collected from Antarctica; gift from Jeffrey Shultz), and Ammothea hilgendorfi (collected from the coast of California; gift from Bonnie Bain), were used in this study. DNA was extracted from a leg, or part of a leg, from each specimen, following the protocol from the Qiagen™ DNAeasy kit. Initially, short fragments were PCR amplified from both the CO1 and Cytb genes with the primers LCO/HCO and CobF/CobR, respectively (see Masta and Boore, 2008 for further details). Then, taxon specific outward-facing primers (see Appendix) were developed from these resulting fragments, and used for long-PCR amplification. Long-PCR amplification was performed using the Takara™ LA Taq DNA polymerase kit. A 100 ll reaction contained final concentrations of 0.16 mM of each dNTP, 0.4 mM of each primer, 1X Takara™ polymerase buffer, 1 ll of mitochondrial DNA (concentration not determined) and 2.5 Units of Takara™ polymerase. The reaction was cycled at 92 °C for 30 s, 50–58 °C (depending on the primers) for 25 s, and 68 °C for 12 min, for 37 cycles, followed by a final extension at 72 °C for 15 min. The PCR products were electrophoresed in a 0.8% agarose gel to estimate size and concentration, then cleaned by ethanol precipitation, and resuspended in water. The mitochondrial genome for Tanystylum was amplified in one complete 15 kb piece using primers facing out from the Cytb amplicon. The majority of protein-coding genes for Colossendeis and Ammothea (about 9 kb) between CO1 and Cytb were amplified in one fragment using specific primers from the two alternate genes. These long-PCR fragments were sheared into 1.5–2 kB pieces, cloned into plasmid libraries, and subject to automated sequencing at the DOE Joint Genome institute (as in Masta and Boore, 2008). This resulted in approximately 15 coverage depth for Tanystylum and Colossendeis, and 5 coverage depth for Ammothea. We were unable to amplify across the control region and rRNA segments for Colossendeis and Ammothea with long-PCR. Instead, for these two taxa remaining genomic segments of protein-coding genes were amplified by primer-walking. In both cases, a 2 kb region containing ND2 was amplified in a single reaction using complementary primers that targeted tRNAMet and CO1. Additional taxon specific primers were designed within this amplicon for internal sequencing across ND2. To amplify between CytB and ND1, we first amplified a short piece of the 16S rRNA in both taxa, and then designed taxon specific primers amplicon to sequence internally. In addition, we sequenced between 12S rRNA and 16S rRNA in Ammothea using taxon specific primers. 2.2. Sequence annotation, analyses, and tRNA secondary structure determination Protein-coding genes were identified by similarity of inferred amino acid sequences to those of other arthropod mtDNAs. Ribosomal RNA genes were annotated based on alignments with other arachnid rRNA genes. After the protein-coding gene boundaries had been determined, the remaining regions were searched for tRNAs with the use of the program tRNAscan-SE 1.21 (Lowe and Eddy, 1997), and by aligning the candidate sequences to the secondary structure-based alignments of other arachnids (see Masta 61 and Boore, 2008). Inferred secondary structures of tRNAs were made by searching for conserved structural features. Anticodon stems always consist of 5 pairs of nucleotides while acceptor stems always consist of 7 pairs of nucleotides. However, the stems of the TWC or DHU arms (also termed T-arm and D-arms) can vary in length in the mitochondrial genomes of arthropods. These arms were determined to be present if there were potential stems consisting of at least 2 nucleotides that could pair with the nucleotides on the adjacent stem, and if the T- or D-loops consisted of at least 3 nucleotides. If either of these two conditions were not met, the nucleotides present in these regions were determined to form TV or D replacement loops, structures common in other chelicerates (Masta and Boore, 2008). The secondary structures of the tRNAs from the published mt genome sequences of the sea spiders Achelia bituberculata (GenBank accession # NC_009724) and Nymphon gracile (GenBank accession # NC_008572) were also inferred and compared to those from the three newly sequenced sea spider mt genomes. Nucleotide composition was inferred with the aid of the program MacVector version 10.0.2. Strand asymmetry or skew was calculated following Perna and Kocher (1995), where AT skew = (A T)/(A + T) and CG skew = (C G)/(C + G). Skew was calculated for the strand encoding the majority of the protein-coding genes in most arthropods. This strand is termed the alpha strand in this manuscript, while the opposite strand is termed the beta strand. We compared the gene order of the mt genomes of these three sea spiders to those of the published sequences of the sea spiders A. bituberculata and N. gracile, and to the gene arrangement ancestral for chelicerates, as found in horseshoe crabs, mesothele spiders, amblypygids, and some scorpions and ticks (see Fahrein et al., 2009; Masta, 2010). 2.3. Sequence alignment and phylogenetic analyses Sequences from 34 additional complete arthropod mitochondrial genomes that span the diversity of arthropods were selected for phylogenetic analyses, with the onychophoran Epiperipatus used as an outgroup (see Table 1 for a list of taxa). We purposefully excluded candidate mitogenomic data for several mites (e.g., Leptotrombidium, Steganacarus; Metaseiulus), as these taxa show many substitutions that may cause branch attraction artifacts (e.g., Podsiadlowski and Braband, 2006, plus exploratory analyses not shown). All 13 mt protein-coding genes from these 40 taxa were translated with the Drosophila mitochondrial genetic code and the amino acids initially aligned with MacClade (Maddison and Maddison, 2000). Amino acid sequences were then aligned in Clustal X version 1.83 (Thompson et al., 1997) using the PAM350 protein matrix, gap opening equal to 10, and gap extension equal to 0.1. To increase confidence that our alignment accurately reflected homologous amino acids, poorly conserved sites were subsequently excluded with the program Gblocks (Castresana, 2000) version 0.91b. Regions of low sequence similarity were removed by using the half gap setting and conserving alignment regions with at least five consecutive conserved amino acids. The Gblocks-trimmed amino acid sequences from the 13 genes were then concatenated into a single sequence. For analysis of nucleotide sequences, we used only nucleotides that correspond to the amino acids kept by Gblocks for phylogenetic analyses. The output files from Gblocks (that list which amino acids in the alignment were kept) were used to compare to the nucleotide alignments in MacClade, using the ‘‘Nucleotide with Amino Acid Translation” display option. The regions that were excluded from the GBlocks alignment were then manually excluded from the MacClade file. The nucleotides at the first and third positions of each codon were then recoded as either purines or pyrim- 62 S.E. Masta et al. / Molecular Phylogenetics and Evolution 57 (2010) 59–70 Table 1 Taxonomic sampling used in phylogenetic analyses. Taxonomic lineage Chelicerata Species GenBank No. Thelyphonida Opiliones Tanystylum orbiculare Ammothea hilgendorfi Achelia bituberculata Colossendeis sp. Nymphon gracile Limulus polyphemus Ornithodoros moubata Haemaphysalis flava Ixodes hexagonus Rhipicephalus sanguineus Carios capensis Varroa destructor Phrynus sp. Calisoga longitarsisb Ornithoctonus huwena Habronattus oregonensis Nephila clavata Hypochilus thorelli Heptathela hangzhouensis Pseudocellus pearsii Buthus occitanus Centruroides limpidus Uroctonus mordax Eremobates palpisetulosus Nothopuga sp. Mastigoproctus giganteus Phalangium opilio GU370074 GU370075 NC_009724 GU370076 NC_008572 NC_003057 NC_004357 NC_005292 NC_002010 NC_002074 NC_005291 NC_004454 NC_010775 NC_010780 NC_005925 NC_005942 NC_008063 NC_010777 NC_005924 NC_009985 NC_010765 NC_006896 NC_010782 NC_010779 NC_009984 NC_010430 NC_010766 Lithobiidae Scutigeridae Spirobolidae Harpagophoridae Lithobius forficatus Scutigera coleoptrata Narceus annularus Thyropygus sp. DVL-2001 NC_002629 NC_005870 NC_003343 NC_003344 Hexapoda Collembola, Poduridae Insecta, Diptera Drosophilidae Insecta, Diptera Neoptera, Hemiptera Podura aquatica Drosophila yakuba Anopheles gambiae Triatoma dimidiata NC_006075 NC_001322 NC_002084 NC_002609 Crustacea Maxillopoda, Cirripedia, Malacostraca, Paguridae Branchiopoda, Anostraca Diplostraca Pollicipes polymerus Pagurus longicarpus Artemia franciscana Daphnia pulex NC_005936 NC_003058 NC_001620 NC_000844 Peripatidae Epiperipatus biolleya NC_009082 a Pycnogonida Ammotheidae Xiphosura Arachnida Colossendeidae Nymphonidae Limulidae Acari, Ixodida Acari, Mesostigmata Amblypygi Araneae, Mygalomorphae Araneae, Araneomorphae Araneae, Mesothelae Ricinulei Scorpiones, Buthidae Scorpiones, Chactidae Solifugae Myriapoda Chilopoda Diplopoda Pancrustacea Onychophora a Tanystylum was formerly placed in the family Tanystylidae, but found by Arango (2002) to possess morphological characters that placed it within Ammotheidae, and was subsequently considered to be in the family Ammotheidae (Arango and Wheeler, 2007). b This mygalomorph spider was formerly labeled as Aphonopelma sp. in GenBank, but it has now been identified as Calisoga longitarsis (although it is currently placed in the genus Brachythele, until the nomenclature is revised). idines (RY coding), following Phillips and Penny (2003) as this has been found to increase the phylogenetic signal in analyses of mitochondrial genomes (Delsuc et al., 2003; Phillips and Penny, 2003). Phylogenetic analyses of amino acids and nucleotides were performed using maximum likelihood with RaxML, version 7.0.4 (Stamatakis, 2006). For analyses with amino acids, the MtREV model of evolution was used, with gamma plus invariant sites model of rate heterogeneity. The MtREV model of evolution is the best model for mtDNA analyses that is available in RaxML, as determined by the use of ProtTest (Abascal et al., 2005). For analyses of nucleotides, the GTR with gamma plus invariant sites model of evolution was used on the RY recoded dataset. One thousand bootstrap replicates were performed for both of these analyses using the same models, except using the CAT approximation instead of gamma for the amino acid analysis. RaxML using the CAT approximation has been shown to perform well on large datasets (Stamatakis, 2006). Figures of the recovered phylogenetic trees were made with the aid of FigTree v1.2.3. Because some of the pycnogonid mt genomes were found to possess a reversed nucleotide strand bias, as were those of some arachnids (Masta et al., 2009), we also recoded amino acids into six physiochemical groups. We categorized amino acids into the following six groups: sulfhydryl (cysteine); small/neutral (serine, threonine, alanine, proline, and glycine); acidic/amides (aspartate, glutamate, asparagine, and glutamine); basic (histidine, arginine, and lysine); hydrophobic (valine, leucine, isoleucine, and methionine); and aromatic (phenylalanine, tyrosine, and tryptophan). This six-character-state data matrix was then analyzed by parsimony using PAUP* (Swofford, 2000) with a heuristic search with 10 replicates, ACCTRAN, and tree-bisection-reconnection branch-swapping. One thousand bootstrap replicates were performed using these same parameters. 3. Results 3.1. Genome composition We obtained the entire mt genome sequence of T. orbiculare, a circular genome of 15,251 nt (GenBank accession number GU370074). For the mt genome of A. hilgendorfi, we obtained 14,026 nt of sequence (GenBank accession number GU370075), and for Colossendeis we sequenced 12,776 nt (GenBank accession number GU370076). The regions not sequenced in Ammothea and S.E. Masta et al. / Molecular Phylogenetics and Evolution 57 (2010) 59–70 Colossendeis are located near the where the control region is typically found in arthropods. It is not uncommon to have difficulty amplifying and sequencing across this region in arthropods, possibly due to the high A + T content, or the stem-forming stretches of nucleotides that create secondary structures and can inhibit function of polymerase enzymes. As a result, some of the mt tRNA and rRNA genes were not sequenced for Ammothea and Colossendeis; however, we obtained full coverage of all 13 of the mt protein-coding genes for each of the three pycnogonids. The mt genome of Tanystylum contains all 22 tRNA genes and the SSU and LSU rRNA genes that are typically found in metazoan mt genomes. The partial mt genome of Ammothea lacks sequences for the genes for tRNAIle and tRNAGln and has only the 30 ends of the SSU rRNA and tRNAMet genes. The sequence for the putative control region is lacking. Presumably the control region plus the two missing tRNA genes are present in the unsequenced portion of the mt genome. The partial mt genome of Colossendeis possesses 16 tRNA genes, as we did not obtain sequence for tRNAPro, tRNAAla, tRNAAsn, tRNASer(AGN), tRNAGln, and tRNAVal genes. The partial genome of Col- Table 2 Nucleotide composition and skew of the alpha strand of pycnogonid mt genomes. Taxon Tanystylum Ammothea Achelia Nymphon Colossendeis Proportion of nucleotides A C G T 0.389 0.414 0.379 0.454 0.363 0.097 0.117 0.119 0.141 0.116 0.103 0.081 0.110 0.085 0.127 0.410 0.388 0.391 0.319 0.388 AT skew CG skew 0.0263 0.0324 0.0156 0.1746 0.0333 0.03 0.1818 0.0393 0.2478 0.0453 63 ossendeis also lacks sequence for the SSU rRNA gene, and has only the 30 end of the LSU rRNA and tRNAMet genes. Sequence coding for the putative control region is also missing. Surprisingly, there are pronounced differences in the nucleotide skews in these pycnogonid mt genomes. Ammothea and Nymphon have a biased composition with positive AT and CG skews on the alpha strand, while Tanystylum and Colossendeis have the opposite pattern, with negative AT and CG skews (Table 2). Achelia shows a third pattern, with a negative AT skew and a positive CG skew. 3.2. Gene order comparisons The mitochondrial gene arrangements of all 5 pycnogonid taxa with complete protein-coding data from the mt genome are depicted in Fig. 1. Tanystylum has a gene order identical to that of Achelia, which differs from the ancestral chelicerate gene order only in the location of a single tRNA gene. In these mt genomes, the tRNAGln gene is located between the SSU gene and the large noncoding region, instead of between the genes coding for tRNAIle and tRNAMet as it is in the ancestral chelicerate genome. With the exception of the missing region, the gene order for the Ammothea mt genome is identical to that of the ancestral chelicerate, or Tanystylum, and Achelia. The genome of Colossendeis shows a gene order quite different from that of the ammotheid sea spiders, in several ways (Fig. 1). First, four of the six unsequenced tRNA genes (tRNAAla, tRNAAsn, tRNASer(AGN), and tRNAPro) should have been obtained had they been in the same location as in the ancestral chelicerate; thus, these must have been translocated. Second, the relative positions of tRNAGlu and tRNAArg in Colossendeis are opposite to their posi- Fig. 1. Mitochondrial genome arrangement of 5 pycnogonids and ancestral chelicerate arrangement. The circular mt genomes are illustrated as linear by depicting CO1 as the beginning of each linearized genome (except the highly rearranged Nymphon). The putative control regions are designated with ‘‘lnr” (large noncoding region), while tRNA genes are abbreviated with their one-letter amino acid code. The two leucine and two serine tRNA genes are further differentiated by numerals, where L1 = tRNALeu(CUN), L2 = tRNALeu(UUR)), S1 = tRNASer(AGN), and S2 = tRNASer(UCN). Shaded regions indicate that the genes present in that location differ from those present in that location in the ancestral chelicerate. Genes are not shown proportional to size. 64 S.E. Masta et al. / Molecular Phylogenetics and Evolution 57 (2010) 59–70 tions in the ancestral chelicerate and in ammotheid sea spiders. Third, the tRNATyr gene is located next to the tRNAPhe gene on the beta strand of Colossendeis. The gene order between the ND3 and tRNAPhe genes in Colossendeis is therefore the same as in Nymphon, except that the control region of Nymphon is located between the tRNAArg and tRNATyr genes (Fig. 1). Likewise, Colossendeis and Nymphon have identical gene arrangements between ND4L and ND6, as both lack the gene for tRNAPro that is located here in the other pycnogonids. Finally, a unique gene arrangement found in Colossendeis but not other sea spiders, is the location of the tRNAMet gene immediately upstream of the tRNAIle gene. 3.3. tRNA structure comparisons Most of the 22 tRNA genes in the mt genomes of these 5 pycnogonids have inferred secondary structures that are typical of metazoan mt tRNAs. In some cases, we reached different interpretations of secondary structure for Achelia and Nymphon than had been previously published (Park et al., 2007; Podsiadlowski and Braband, 2006). While most of the tRNAs possess cloverleaf structures, some tRNA genes lack the sequence to form the DHU or Darm that is part of the canonical tRNA (see Fig. 2). All pycnogonids lack D-arm sequences in their tRNAAla genes, with the most extreme nucleotide reductions occurring in Tanystylum, Ammothea, and Achelia (Fig. 3), although Colossendeis was unsampled. Nymphon and Colossendeis lack well-paired D-arm stem sequences in their tRNAArg genes. Like other metazoans (Wolstenholme, 1992b), all pycnogonids lack D-arm sequence in their tRNASer(AGN) genes. All 22 of the tRNAs present in the mitochondrial genomes of these pycnogonids possess T-arms (Fig. 3). Another structural difference is present in the tRNAIle gene in Nymphon and Colossendeis. These taxa, but not the ammotheids, possess a variable loop that is only 4 nt long, instead of the 5-ntlong loop found in Tanystylum and Achelia. Differences also exist in the anticodons present in two tRNA genes within sea spiders. The anticodon for tRNALys is UUU in all five pycnogonids, whereas other arthropods possess a CUU anticodon. The anticodon for tRNASer(AGN) is UCU in all four pycnogonids for which this tRNA gene was sequenced. This differs from the typical GCU anticodon found in most other chelicerates. Some mispaired nucleotides are present in the anticodon and acceptor stem regions of 8 tRNA genes within pycnogonids. While most of these are confined to only a single individual, tRNAIle and tRNALeu(CUN) both possess mispaired nucleotides in their acceptor stem regions in Nymphon and Colossendeis, whereas the other pycnogonids have well-paired stems. The anticodon stems of the tRNACys genes have mispaired nucleotides in Nymphon and Colossendeis (Fig. 3). 3.4. Phylogenetic analyses of amino acids There were 3894 amino acids in our initial alignments of the protein-coding genes. Of these, 2313 amino acids from all 13 concatenated protein-coding genes were retained after GBlocks trimming of the dataset. This reduced dataset was used in subsequent phylogenetic analyses. The maximum likelihood analysis of amino acids yielded a phylogenetic tree in which a clade of chelicerates was recovered with 95% bootstrap support (Fig. 4A). Nested within the chelicerates, a clade of pycnogonids was recovered with 100% bootstrap support. While there is not strong support to infer which group of arachnids may be the sister group to pycnogonids, there is ample support to place them within the clade Micrura, which traditionally consists of the orders Acari, Ricinulei, Araneae, Thelyphonida, Schizomida, and Amblypygi (see review in Shultz, 2007). Relationships among the other major lineages of arthropods are similar to those that have been recovered in other analyses of molecular data, with high support for a clade of Pancrustacea. Myriapoda was not recovered as a clade, but instead the centipedes and millipedes were each found to be monophyletic, with very weak support for a sister-group relationship between the Diplopoda (millipedes) and the chelicerates. The interrelationships among the pycnogonids were all strongly supported, with two divergent lineages recovered. One lineage consists of Tanystylum, Ammothea and Achelia, while the other consists of Nymphon and Colossendeis. The parsimony tree that we recovered from analyses of amino acids that had been recoded into 6 functional groups yielded a tree with unresolved relationships among the major lineages of arthropods, and therefore is not shown. However, this tree yielded identical relationships among the pycnogonids, with 100% bootstrap support for a clade consisting of Tanystylum, Ammothea and Achelia, and 100% bootstrap support for a clade consisting of Nymphon and Colossendeis. There was 76% bootstrap support for a sister-group relationship between Tanystylum and Ammothea. 3.5. Phylogenetic analyses of RY-coded nucleotides A B tRNA-Alanine tRNA-Arginine aminoacyl acceptor arm D-arm T T T AG T T A TCA A anticodon arm A T A 5' T A G T A A G T A T T A T T C T-arm CCG AT G GGT G T A T TT T A A C T T C T A TCG A T T G A 3' variable loop 3' A T C T G T T C C G A T A TT A T G C TAT A A A T T T TA T T A T A A T A T A T A T TGC 5' A G G C A A G Fig. 2. Secondary structure drawing of the two types of tRNAs found in pycnogonids. The tRNA structures from Tanystylum orbiculare are illustrated. (A) A canonical tRNA, as illustrated with tRNAArg. Note the 5 regions of the tRNA: acceptor stem, Darm, anticodon arm, variable loop, and T-arm. (B) A tRNA that lacks a D-arm, and instead has only a D-loop, as illustrated with tRNAAla. The maximum likelihood analysis of nucleotides with the first and third positions recoded as purines or pyrimidines yielded a phylogenetic tree that was largely congruent with the aminoacid-based tree. A clade of chelicerates was recovered with 81% bootstrap support, with a clade of pycnogonids nested within Chelicerata (Fig. 4B). Unlike the amino-acid-based tree, however, there was only very weak support for placing Pycnogonida within the micrurans. Additionally, all micrurans were not included in this clade, as Amblypygi fell outside of it. The RY-coded analysis recovered a high level of support (97% bootstrap) for a clade of Pancrustacea, as was also found in the amino acid analysis. While there was generally lower support for most groupings in the recoded nucleotide based tree, there was high support for a monophyletic Myriapoda. Support for a sistergroup relationship between Myriapoda and Chelicerata was weak. Pycnogonida was recovered with high support, as were the same two divergent lineages found in the amino acid analysis (a clade consisting of Tanystylum, Ammothea and Achelia, and a clade consisting of Nymphon and Colossendeis). However, in the RY-recoded nucleotide analysis, Tanystylum and Achelia were recovered as sister taxa with 68% bootstrap support (Fig. 4B). S.E. Masta et al. / Molecular Phylogenetics and Evolution 57 (2010) 59–70 65 Fig. 3. Secondary structure alignments for the 22 tRNAs found in the mt genomes of pycnogonids. The 5 regions of the canonical tRNA structure are given at the top of the figure (see Fig. 2 for an illustration of these regions). Dashes separate these structural features in the alignments. Underlined nucleotides indicate they form base pairs with adjacent underlined nucleotides in the same region of the tRNA. 3.6. Rare genomic changes mapped onto the phylogenetic tree The gene translocations and tRNA structure differences that we inferred were mapped onto the consensus phylogenetic tree of pycnogonids. The pycnogonids are united by three tRNA features that distinguish them from other chelicerates: the loss of a Darm in tRNAAla, and changes in the anticodon identities of tRNALys and tRNASer(AGN) (Fig. 5). Within pycnogonids, the ammotheids Tanystylum and Achelia are united by two rare genomic changes. One is the translocation of tRNAGln between SSU and the large noncoding region, while the other is a reduction in the size of the Dloop in tRNAAla. In contrast, the clade consisting of Nymphon and Colossendeis shows at least 12 changes in genome or tRNA structure that unite it as a monophyletic group. Five of these changes occur within the tRNA genes, while the other seven changes are based upon genome rearrangements (Fig. 5). At least seven mt tRNA translocations must have occurred in the lineage leading to Nymphon and Colossendeis. These are mapped 66 S.E. Masta et al. / Molecular Phylogenetics and Evolution 57 (2010) 59–70 Fig. 3 (continued) onto the phylogenetic tree of pycnogonids (Fig. 5). An eighth translocation, that of tRNAVal moving away from its ancestral location between the LSU and SSU rRNA genes, cannot be placed with confidence onto the phylogeny, since we lack sequence data for this region in Colossendeis. Four unique gene translocations distinguish Nymphon, and therefore must have occurred after its divergence from a common ancestor with Colossendeis (Fig. 5). 4. Discussion 4.1. Variable patterns of nucleotide bias Arthropod mt genomes tend to have an AT nucleotide bias, with a positive AC skew on the major or alpha strand. The strength of the skew varies among taxa, and some arthropod taxa have been found to show the opposite pattern; that is, a negative AC skew, or a positive TG skew (Hassanin, 2006). We found both of these two extreme patterns of skew present within the five pycnogonids we analyzed. Although reversals in skew can impact phylogenetic reconstruction and lead to erroneous conclusions (Hassanin, 2006; Jones et al., 2007; Masta et al., 2009), we used methods de- signed to overcome the effects of saturation and opposite patterns of mutational bias. Within pycnogonids, we found no evidence that altered patterns of nucleotide bias affected phylogeny reconstruction. Biased nucleotide composition may be expected to cause taxa with similar biases to group together. However, Tanystylum and Colossendeis show patterns of skew opposite to that of the other pycnogonids, and indeed to that of most arthropods. However, these taxa were not reconstructed as sister taxa, despite their shared biases. Likewise, other taxa with similar patterns of skew (e.g., opisthothele spiders and scorpions; see Masta et al., 2009) also did not group closely with these pycnogonids. This gives us confidence that our phylogenetic methods have adequately accounted for different mutational biases in this dataset. 4.2. Pycnogonids as derived arachnids Our analyses strongly support the placement of Pycnogonida within Chelicerata, and therefore would reject the hypothesis that sea spiders are the sister group to all other arthropods. Our phylogenetic analyses also consistently place sea spiders as derived chelicerates. These findings are in agreement with other recent 67 S.E. Masta et al. / Molecular Phylogenetics and Evolution 57 (2010) 59–70 Ixodes 100 100 100 100 Rhipicephalus Haemaphysalis Ornithodoros Carios Varroa Tanystylum Ammothea Achelia 100 57 100 80 100 100 54 100 100 52 95 Limulus 56 100 100 55 100 57 93 58 97 Centruroides Buthus Uroctonus Narceus Thyropygus Scutigera Lithobius Pollicipes Pagurus 98 Scorpiones Myriapoda Podura Artemia 98 Araneae Amblypygi Solifugae Opiliones Xiphosura Phrynus Nothopuga Eremobates Phalangium 100 Pycnogonida Nymphon Colossendeis Pseudocellus Ricinulei Mastigoproctus Thelyphonida Heptathela 100 Ornithoctonus Calisoga Habronattus 100 78 Nephila Hypochilus 69 73 Acari Chelicerata A Pancrustacea Daphnia Triatoma 100 Anopheles Drosophila Epiperipatus Onychophora 0.2 B 100 100 Ixodes 100 Rhipicephalus Haemaphysalis Ornithodoros Carios 100 100 Acari Varroa 68 100 Ricinulei Pycnogonida Nymphon Colossendeis 100 Mastigoproctus Heptathela 92 100 81 100 98 100 Phrynus Limulus 84 98 97 77 98 100 Thelyphonida Ornithoctonus Calisoga Habronattus Nephila Hypochilus 100 Nothopuga Eremobates Araneae Solifugae Amblypygi Xiphosura Centruroides Buthus Uroctonus Phalangium 100 Narceus Thyropygus Scutigera Lithobius Pollicipes Pagurus 93 Daphnia Artemia Podura Triatoma Anopheles 100 Drosophila Epiperipatus 100 Chelicerata 100 Pseudocellus Tanystylum Achelia Ammothea Scorpiones Opiliones Myriapoda Pancrustacea Onychophora 0.1 Fig. 4. (A) Maximum likelihood tree based on amino acids from 13 mt protein-coding genes, using the MtREV substitution matrix. Bootstrap values greater than 50% are shown above or below the corresponding branches. Likelihood = 86273.619981; Alpha = 0.857; invariant sites = 0.194572. (B) Maximum likelihood tree based on RY-coded nucleotides from 13 mt protein-coding genes, using the GTR substitution matrix with gamma and invariant sites. Bootstrap values greater than 50% are shown above or below the corresponding branches. For both trees, the bars colored medium gray highlight the three main arthropod groups: Chelicerata, Myriapoda, and Pancrustacea, while the other colored bars are used to highlight the chelicerate groups for which multiple taxa were sequenced. 68 S.E. Masta et al. / Molecular Phylogenetics and Evolution 57 (2010) 59–70 A 9 1 2 3 Tanystylum Ammothea Achelia N I J K L Nymphon 4 5 6 7 8 B C D E F G H M Colossendeis Fig. 5. Phylogeny of pycnogonids with rare genomic changes depicted on the tree. Boxes that are shaded and with letters depict inferred gene translocation events, and unshaded boxes with numbers depict inferred changes in tRNA structure or anticodon identity. Letters for tRNA gene translocations are as follows, using the one-letter amino acid abbreviation: (A) Q translocated between SSU and the large noncoding region (lnr), on same beta strand as original location. (Not assessed in Ammothea); (B) I (or M) translocated between M and ND2, on same alpha strand as original location; (C) Y translocated next to F, on same beta strand as original location; (D) N translocated to near control region; (E) S1 translocated to near control region; (F) A translocated to near control region; (G) either R or E translocated such that E is directly upstream of R on alpha strand; (H) P translocated to near control region; (I) translocation and inversion of K and D to near control region; (J) translocation of region coding for I-ND2-W-C-CO1-CO2 onto opposite strand, upstream of ATP8; (K) M translocated to beta strand, between P and S1; (L) control region translocated to between R and Y; (M) translocation of I to between M and ND2, on same alpha strand and (N) V translocated away from location between LSU and SSU rRNA genes. Numbers for tRNA changes are as follows: (1) loss of the D-arm in tRNAAla; (2) anticodon for tRNALys changes to UUU; (3) anticodon for tRNASer(AGN) changes to UCU; (4) variable loop in tRNAIle decreases to 4 nt in Nymphon and Colossendeis, from 5 nt long in Tanystylum and Achelia; (5) loss of paired D-arm sequence in tRNAArg; (6) mispaired acceptor stems in tRNAIle; (7) Mispaired acceptor stems in tRNALeu(CUN); (8) mispaired anticodon stems of tRNACys and (9) reduced number of nucleotides in D-loop in tRNAAla. analyses of mitochondrial genome data (Jeyaprakash and Hoy, 2009; Park et al., 2007; Podsiadlowski and Braband, 2006). However, unlike those analyses, which had representative taxa from just three of the arachnid orders, our analyses included taxa from 8 arachnid orders that span the diversity of Chelicerata. Our phylogenetic analyses support Pycnogonida as members of micruran arachnids, a group that includes the Acari (mites and ticks) and Araneae (spiders). Pycnogonids have been found to be nested within a clade of micrurans in combined analyses of morphological and nuclear rRNA data (Giribet et al., 2002), but the authors cautioned that they did not find high support for these relationships. The possibility that sea spiders are derived arachnids or chelicerates has not been supported by developmental or anatomical studies. However, most studies that have tried to determine the taxonomic placement of pycnogonids have not included a diversity of micruran arachnids in their analyses (but see Wheeler and Hayashi, 1998). Therefore, it is not clear which, if any, anatomical or developmental features may be shared among pycnogonids and other arachnids. Indeed the idea that sea spiders are derived arachnids may seem unlikely, not only because of sea spiders’ unique and highly derived anatomy, but also due to their strictly marine habitats. Most arachnids are strictly terrestrial, with anatomical features that allow them to live in terrestrial environments. The ecology of sea spiders, however, does not necessarily help us understand their phylogenetic affinities. Although they are restricted to marine environments, pycnogonids are found in a diverse array of ecological niches, from the intertidal zone to deepsea hydrothermal vents to the cold waters of Antarctica. Although it is tempting to think that their marine habitats may imply a close relationship to the (marine) horseshoe crabs, sea spiders are not the sole marine chelicerates. Some mites in the arachnid order Acari inhabit marine environments, whereas others inhabit terrestrial and freshwater environments. Clearly, transitions from sea to land and back again have occurred in the evolutionary history of the chelicerates. Paleontological data are consistent with an aquatic origin of pycnogonids, and suggests they emerged during the Cambrian, before the other extant chelicerate lineages (Dunlop and Selden, 2009). Fossils of horseshoe crabs (Rudkin et al., 2008) exist from the Ordovician, consistent with an early divergence of xiphosurans. Because these horseshoe crab fossils possess derived anatomical features, it was suggested that xiphosurans likely diverged even earlier, in the Cambro-Ordovician (Rudkin et al., 2008). The wellsupported nodes in our phylogenetic trees (e.g., >70% bootstrap support) are not in conflict with this paleontological data. Instead, they suggest a scenario of an ancestral chelicerate diversifying in the Cambrian into at least two lineages. One of these lineages could then have subsequently diversified into pycnogonids and the ancestor of micrurans, while the other lineage diversified into xiphosurans and the ancestor of dromopodan arachnids (i.e., scor- pions, solifuges, harvestmen). Under this scenario, the group Arachnida would be polyphyletic, with two separate origins. Further fossil discoveries and developmental and molecular data with stronger phylogenetic signal will help elucidate if this is a likely scenario. 4.3. Rare genomic changes as phylogenetic markers Although there are no gene arrangements that group sea spiders with any single chelicerate lineage, pycnogonids are supported as chelicerates by their similar gene order. The ammotheid sea spiders share the same basic gene arrangement as other chelicerates, and differ only in the location of a single tRNA gene. Therefore, we can infer that the translocation of this tRNA occurred after the divergence of sea spiders from their common ancestor with other chelicerates. Pycnogonids lack the derived mt gene order found in Pancrustacea, in which the tRNALeu(UUR) gene is located between CO1 and CO2, and not between ND1 and the LSU rRNA genes as in chelicerates (Boore et al., 1995, 1998). Therefore, both our phylogenetic analyses and our genome arrangement data are consistent with pycnogonids being chelicerates. The three tRNA structural characters that unite the pycnogonids (see Fig. 5) show some degree of homoplasy when evaluated across arthropods. A change in the anticodon from GCU to UCU in tRNASer(AGN) has been found in some other arthropod groups. For example, all beetles appear to possess tRNASer(AGN) genes with UCU anticodons, an apparent synapomorphy for this lineage of insects (Sheffield et al., 2008). Likewise, a change in anticodon identity for tRNALys has occurred multiple times in arthropods, and is correlated with changes in tRNASer(AGN) anticodon identity (Abascal et al., 2006). Changes in anticodon identity and remolding of tRNALeu has occurred multiple times in the evolution of metazoan mtDNA (Rawlings et al., 2003), and can lead to incorrect inferences about gene rearrangements if not correctly identified. However, tRNA remolding cannot explain the translocations of whole blocks of tRNA genes to new locations in the genome, as found in Nymphon and Colossendeis. The loss of the D-arm from tRNAAla has also occurred independently in amblypygids (Fahrein et al., 2009). Therefore, while these changes have occurred only rarely in evolution, they do exhibit some degree of homoplasy within arthropods. Searching for other rare evolutionary changes that pycnogonids share with other taxonomic groups may help to more definitively infer their closest relatives. In contrast, many of the genome rearrangements that we inferred to occur in the common ancestor of Nymphon and Colossendeis are unique among arthropods. For example, the gene arrangement of the genes for ND3, tRNAGlu, tRNAArg on the alpha strand, followed by the tRNATyr and tRNAPhe genes on the beta strand, is not found among other arthropods. This gene order required multiple translocations of tRNA genes, and therefore, find- S.E. Masta et al. / Molecular Phylogenetics and Evolution 57 (2010) 59–70 ing evolution convergent with this same arrangement seems highly unlikely. However, the shared changes seen in the tRNA genes among Nymphon and Colossendeis are likely more evolutionarily labile, and may occur on a time scale that allows them to be useful for delimiting pycnogonid taxa, although not at deep phylogenetic levels. Together, though, the inferred gene translocation and tRNA structural characters are synapomorphies that further support our phylogenetic analyses of amino acids and that group Nymphon with Colossendeis. 4.4. Phylogenetic relationships within Pycnogonida A traditional classification of pycnogonids divided them into 8 families (Hedgpeth, 1947), but did not make inferences as to how these may be interrelated. The taxa in our study included representatives from what have been considered to be quite divergent lineages: Ammotheidae, Nymphonidae, and Colossendeidae. A cladistic analysis of morphological characters (Arango, 2002) found that Ammotheidae was a diverse and paraphyletic group, with Colossendeidae derived within the ammotheids. Taxa from the genera Achelia and Tanystylum were found to be more closely related to one another than they are to Ammothea. This paraphyletic group of ammotheids plus colossendeids was found to be sister to a clade that contained Nymphonidae. Our analyses provide support for a group of ammotheids, but only phylogenetic analyses of RY-coded nucleotide data give support for the closest relationship being between Achelia and Tanystylum. Analyses of amino acids yield trees with high support for a closer relationship of Tanystylum with Ammothea. We do not find rare genomic changes that support or refute either of these scenarios. However, we do find very strong support for a close relationship between Colossendeis and Nymphon, in contrast to inferences made based upon morphological characters (Arango, 2002) and morphology plus six loci (Arango and Wheeler, 2007). We also find Nymphon to possess a highly derived gene arrangement that clearly shares its origins with those of Colossendeis. Our findings are consistent with those of Arango (2003), who found that 18S and 28S rRNA sequences yielded a tree with the ammotheids sister to a clade that contains Nymphon and Colossendeis (which were reconstructed as sister taxa). However, our findings differ from those of Nakamura et al. (2007), who found that 18S rRNA sequences resolved the clade containing Colossendeidae as sister to a clade containing Ammotheidae, with Nymphonidae sister to that large clade. Moreover, our findings differ from those of Arango and Wheeler (2007), who conducted a comprehensive analysis built on the earlier nuclear rRNA study of Arango (2003) by adding many more taxa and using four additional loci, including 3 mitochondrial genes (CO1, SSU rRNA, and LSU rRNA) and histone 3. Arango and Wheeler (2007) recovered Nymphonidae in a clade with Ammotheidae, with Colossendeidae sister to that larger clade containing most families of sea spiders. Why our results should differ from those of two of the three previous molecular analyses is not clear. Arango and Wheeler (2007) analyzed six loci in conjunction with morphological characters, so it is not possible to determine which of these other loci, if any, conflicted with our complete mt protein-coding gene analyses and the prior nuclear rRNA analyses. However, given the high support we found for a clade of Nymphon plus Colossendeis in all of our analyses – with both sequence data and rare genomic changes – our data strongly indicates that these two specimens are much more closely related to one another than either are to the ammotheids Tanystylum, Achelia, and Ammothea. One possibility is that the Colossendeis specimen could have been originally misidentified, and in fact is closely related to nymphonid sea spiders. This specimen has been used extensively for analyses of nuclear loci (Regier et al., 2008, 2010), and little remains of it to aid in re-identification. 69 Both the analyses of Arango and Wheeler (2007) and Nakamura et al. (2007) consistently recovered Callipallenidae as sister to Nymphonidae. Our findings would be consistent with these previous analyses if our specimen were actually a member of the family Callipallenidae, and not Colossendidae. However, both our sequence data and our rare genomic change data indicate that the specimen is quite divergent from the N. gracile mt genome, and that rare genomic changes are useful for resolving sea spider relationships. Our analyses provide strong support for the placement of sea spiders as chelicerates, and moderate support for sea spiders as derived chelicerates. There are deep divergences within sea spiders, resulting in phylogenetic trees with lineages separated by long branches. Additionally, we find multiple cases of changes in mitochondrial strand bias among these lineages. Our analyses attempted to overcome the misleading effects of changes in strand bias by recoding data as amino acids or as purines and pyrimidines. However, it is more difficult to overcome potential problems associated with long branches causing the recovery of artifactual relationships. Future sampling of additional sea spider mitochondrial genomes may help break up the branch leading to sea spiders, as well as branches among the major lineages of sea spiders. Rare changes in the mitochondrial tRNA genes and in mitochondrial gene arrangements hold much promise for helping to resolve relationships among this enigmatic group of animals. Acknowledgments This work was funded by NSF award no. DEB-0416628 to S.E.M. The manuscript was improved by the comments of Claudia Arango, Alfried Vogler, and an anonymous reviewer. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.ympev.2010.06.020. References Abascal, F., Zardoya, R., Posada, D., 2005. ProtTest: selection of best-fit models of protein evolution. Bioinformatics 21, 2104–2105. Abascal, F., Posada, D., Knight, R.D., Zardoya, R., 2006. Parallel evolution of the genetic code in arthropod mitochondrial genomes. PLoS Biol. 4, e127. Arango, C.P., 2002. Morphological phylogenetics of the sea spiders (Arthropoda: Pycnogonida). Organ. Diversity Evol. 2, 107–125. Arango, C.P., 2003. Molecular approach to the phylogenetics of sea spiders (Arthropoda: Pycnogonida) using partial sequences of nuclear ribosomal DNA. Mol. Phylogen. Evol. 28, 588–600. Arango, C.P., Wheeler, W.C., 2007. Phylogeny of the sea spiders (Arthropoda, Pycnogonida) based on direct optimization of six loci and morphology. Cladistics 23, 255–293. Boore, J.L., 2006. The use of genome-level characters for phylogenetic reconstruction. Trends Ecol. Evol. 21, 439–446. Boore, J.L., Brown, W.M., 1998. Big trees from little genomes: mitochondrial gene order as a phylogenetic tool. Curr. Opin. Genet. Dev. 8, 668–674. Boore, J.L., Collins, T.M., Stanton, D., Daehler, L.L., Brown, W.M., 1995. Deducing the pattern of arthropod phylogeny from mitochondrial DNA rearrangements. Nature 376, 163–167. Boore, J.L., Lavrov, D.V., Brown, W.M., 1998. Gene translocation links insects and crustaceans. Nature 392, 667–668. Bourlat, S.J., Nielsen, C., Economou, A.D., Telford, M.J., 2008. Testing the new animal phylogeny: a phylum level molecular analysis of the animal kingdom. Mol. Phylogen. Evol. 49, 23–31. Brenneis, G., Ungerer, P., Scholtz, G., 2008. The chelifores of sea spiders (Arthropoda, Pycnogonida) are the appendages of the deutocerebral segment. Evol. Develop. 10, 717–724. Castresana, J., 2000. Selection of conserved blocks from multiple alignments for their use in phylogenetic analysis. Mol. Biol. Evol. 17, 540–552. Delsuc, F., Phillips, M.J., Penny, D., 2003. Comment on Hexapod origins: monophyletic or paraphyletic? Science 301. Dunlop, J.A., Arango, C.P., 2005. Pycnogonid affinities: a review. J. Zool. Syst. Evol. Res. 43, 8–21. Dunlop, J.A., Selden, P.A., 2009. Calibrating the chelicerate clock: a paleontological reply to Jeyaprakash and Hoy. Exp. Appl. Acarol. 48, 183–197. 70 S.E. Masta et al. / Molecular Phylogenetics and Evolution 57 (2010) 59–70 Dunn, C.W., Hejnol, A., Matus, D.Q., Pang, K., Browne, W.E., Smith, S.A., Seaver, E., Rouse, G.W., Obst, M., Edgecombe, G.D., Sorensen, M.V., Haddock, S.H.D., Schmidt-Rhaesa, A., Okusu, A., Kristensen, R.M., Wheeler, W.C., Martindale, M.Q., Giribet, G., 2008. Broad phylogenomic sampling improves resolution of the animal tree of life. Nature 452, 745–749. Fahrein, K., Masta, S.E., Podsiadlowski, L., 2009. The first complete mitochondrial genome sequences of Amblypygi (Chelicerata: Arachnida) reveal a conservation of the ancestral arthropod gene order. Genome 52, 456–466. Fry, W.G., 1978. A classification within the pycnogonids. Zool. J. Linn. Soc. 63, 35–38. Giribet, G., Edgecombe, G.D., Wheeler, W.C., 2001. Arthropod phylogeny based on eight molecular loci and morphology. Nature 413, 157–161. Giribet, G., Edgecombe, G.D., Wheeler, W.C., Babbitt, C., 2002. Phylogeny and systematic position of Opiliones: a combined analysis of chelicerate relationships using morphological and molecular data. Cladistics 18, 5–70. Hassanin, A., 2006. Phylogeny of Arthropoda inferred from mitochondrial sequences: strategies for limiting the misleading effects of multiple changes in pattern and rates of substitution. Mol. Phylogen. Evol. 38, 100–116. Hassanin, A., Leger, N., Deutsch, J., 2005. Evidence for multiple reversals of asymmetric mutational constraints during the evolution of the mitochondrial genome of Metazoa, and consequences for phylogenetic inferences. Syst. Biol. 54, 277–298. Hedgpeth, J.W., 1947. On the evolutionary significance of the Pycnogonida. Smiths. Misc. Coll., 1–53. Jager, M., Murienne, J., Clabaut, C., Deutsch, J., Le Guyader, H., Manuel, M., 2006. Homology of arthropod anterior appendages revealed by Hox gene expression in a sea spider. Nature 441, 506–508. Jeyaprakash, A., Hoy, M.A., 2009. First divergence time estimate of spiders, scorpions, mites and ticks (subphylum: Chelicerata) inferred from mitochondrial phylogeny. Exp. Appl. Acarol. 47, 1–18. Jones, M., Gantenbein, B., Fet, V., Blaxter, M., 2007. The effect of model choice on phylogenetic inference using mitochondrial sequence data: lessons from the scorpions. Mol. Phylogen. Evol. 43, 583–595. Larget, B., Simon, D.L., Kadane, J.B., Sweet, D., 2005. A Bayesian analysis of metazoan mitochondrial genome arrangements. Mol. Biol. Evol. 22, 486–495. Lowe, T.M., Eddy, S.R., 1997. TRNAscan-SE: a program for improved detection of transfer RNA genes in genomic sequence. Nucleic Acids Res. 25, 955–964. Maddison, D.R., Maddison, W.P., 2000. MacClade 4: Analysis of Phylogeny and Character Evolution. Sinauer Associates, Inc., Sunderland, Massachusetts, USA. Mallatt, J., Giribet, G., 2006. Further use of nearly complete, 28S and 18S rRNA genes to classify Ecdysozoa: 37 more arthropods and a kinorhynch. Mol. Phylogen. Evol. 40, 772–794. Mallatt, J.M., Garey, J.R., Shultz, J.W., 2004. Ecdysozoan phylogeny and Bayesian inference. First use of nearly complete 28S and 18S rRNA gene sequences to classify the arthropods and their kin. Mol. Phylogen. Evol. 31, 178–191. Manuel, M., Jager, M., Murienne, J., Clabaut, C., Le Guyader, H., 2006. Hox genes in sea spiders (Pycnogonida) and the homology of arthropod head segments. Dev. Genes. Evol. 216, 481–491. Masta, S.E., 2010. Mitochondrial rRNA secondary structures and genome arrangements distinguish chelicerates: comparisons with a harvestman (Arachnida: Opiliones: Phalangium opilio). Gene 449, 9–21. Masta, S.E., Boore, J.L., 2004. The complete mitochondrial genome sequence of the spider Habronattus oregonensis reveals rearranged and extremely truncated tRNAs. Mol. Biol. Evol. 21, 893–902. Masta, S.E., Boore, J.L., 2008. Parallel evolution of truncated transfer RNA genes in arachnid mitochondrial genomes. Mol. Biol. Evol. 25, 949–959. Masta, S.E., Longhorn, S.J., Boore, J.L., 2009. Arachnid relationships based on mitochondrial genomes: asymmetric nucleotide and amino acid bias affects phylogenetic analyses. Mol. Phylogen. Evol. 50, 117–128. Maxmen, A., Brown, W.E., Martindale, M.Q., Giribet, G., 2005. Neuroanatomy of sea spiders implies an appendicular origin of the protocerebral segment. Nature 437, 1144–1148. Murrell, A., Campbell, N.J.H., Barker, S.C., 2003. The value of idiosyncratic markers and changes to conserved tRNA sequences from the mitochondrial genome of hard ticks (Acari: Ixodida: Ixodidae) for phylogenetic inference. Syst. Biol. 52, 296–310. Nakamura, K., Kano, Y., Suzuki, N., Namatame, T., Kosaku, A., 2007. 18S rRNA phylogeny of sea spiders with emphasis on the position of Rhynchothoracidae. Mar. Biol. 153, 213–223. Park, S.-J., Lee, Y.-S., Hwang, U., 2007. The complete mitochondrial genome of the sea spider Achelia bituberculata (Pycnogonida, Ammotheidae): arthropod ground pattern of gene arrangement. BMC Genomics 8, 343. Perna, N.T., Kocher, T.D., 1995. Patterns of nucleotide composition at fourfold degenerate sites of animal mitochondrial genomes. J. Mol. Evol. 41, 353–358. Phillips, M.J., Penny, D., 2003. The root of the mammalian tree inferred from whole mitochondrial genomes. Mol. Phylogen. Evol. 28, 171–185. Podsiadlowski, L., Braband, A., 2006. The complete mitochondrial genome of the sea spider Nymphon gracile (Arthropoda: Pycnogonida). BMC Genomics 7, 284. Rawlings, T.A., Collins, T.M., Bieler, R., 2003. Changing identities: tRNA duplication and remolding within animal mitochondrial genomes. Proc. Natl Acad. Sci. USA 100, 15700–15705. Regier, J., Shultz, J., Ganley, A., Hussey, A., Shi, D., Ball, B., Zwick, A., Stajich, J., Cummings, M., Martin, J., Cunningham, C., 2008. Resolving arthropod phylogeny: exploring phylogenetic signal within 41 kb of protein-coding nuclear gene sequence. Syst. Biol. 57, 920–938. Regier, J.C., Shultz, J.W., Zwick, A., Hussey, A., Ball, B., Wetzer, R., Martin, J.W., Cunningham, C.W., 2010. Arthropod relationships revealed by phylogenomic analysis of nuclear protein-coding sequences. Nature 463, 1079–1083. Reyes, A., Gissi, C., Pesole, G., Saccone, C., 1998. Asymmetrical directional mutation pressure in the mitochondrial genome of mammals. Mol. Biol. Evol. 15, 957– 966. Rudkin, D.M., Young, G.A., Nowlan, G.S., 2008. The oldest horseshoe crab: a new xiphosurid from late Ordovician Konservatt-Lagerstätten deposits, Manitoba, Canada. Palaeont 51, 1–9. Sheffield, N.C., Song, H., Cameron, L., Whiting, M.F., 2008. A comparative analysis of mitochondrial genomes in Coleoptera (Arthropoda: Insecta) and genome descriptions of six new beetles. Mol. Biol. Evol. 25, 2499–2509. Shultz, J.W., 2007. A phylogenetic analysis of the arachnid orders based on morphological characters. Zool. J. Linn. Soc. 150, 221–265. Soll, D., RajBhandary, U.L., 1995. TRNA: Structure, Biosynthesis, and Function. ASM Press, Washington, DC. Stamatakis, A., 2006. RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics 22, 2688– 2690. Swofford, D.L., 2000. PAUP*. Phylogenetic Analysis Using Parsimony (*and Other Methods). Version 4. Sinauer Associates, Sunderland, Massachusetts. Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F., Higgins, D.G., 1997. The CLUSTAL_X windows interface. Flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res. 25, 4876–4882. Wheeler, W.C., Hayashi, C.Y., 1998. The phylogeny of the extant chelicerate orders. Cladistics 14, 173–192. Wolstenholme, D.R., 1992a. Animal mitochondrial DNA: structure and evolution. In: Wolstenholme, D.R., Jeon, K.W. (Eds.), Mitochondrial Genomes. Academic Press, New York, pp. 173–216. Wolstenholme, D.R., 1992b. Genetic novelties in mitochondrial genomes of multicellular animals. Curr. Opin. Genet. Develop. 2, 918–925. Zrzavý, J., Hypsa, V., Vlásková, M., 1998. Arthropod phylogeny: taxonomic congruence, total evidence and conditional combination approaches to morphological and molecular data sets. In: Fortey, R.A., Thomas, R.H. (Eds.), Arthropod Relationships. Chapman & Hall, London, pp. 97–107.