© 2000 by The Society for Integrative and Comparative Biology
Controversy and Consensus in Asteroid Systematics: New Insights to Ordinal and Familial Relationships1
1 Department of Ecology and Evolution, State University of New York, Stony Brook, New York 11794-5245
 |
SYNOPSIS |
|---|
TOP
 SYNOPSIS INTRODUCTIONBACKGROUNDMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSReferences |
|---|
Phylogenetic approaches have sparked controversy in asteroid systematics since 1987. Despite recent attempts at resolving these differences and evidence of some consensus, our understanding of relationships among asteroid taxa remains unsatisfactory. This paper presents results of an investigation into asteroid evolutionary history using DNA sequence data from mitochondrial transfer RNA and the cytochrome oxidase c subunit I genes analyzed with and without previously published ribosomal gene sequences. Analysis of these genes provides an assessment of familial relationships but does little to elucidate ordinal relationships. A basal position for the Paxillosida is not supported. However, close relationships of some velatid and valvatid taxa are upheld. The resulting phylogenies are not a definitive answer to controversies in asteroid systematics. However, with new insights to some asteroid relationships, they highlight the need for a redirection of future systematic studies so a consensus can be made.
|
INTRODUCTION |
|---|
|
TOP
SYNOPSISINTRODUCTIONBACKGROUNDMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSReferences |
|---|
The Asteroidea (sea stars or starfishes) is one of the most diverse classes of the Phylum Echinodermata. Seven orders of approximately 35 families, 300 genera and 1500 species are recognized (Hyman, 1955
; Clark and Downey, 1992). The diversity of asteroids is exemplified by their varied morphological forms, their occupation of a wide range of habitats and geographic localities, and their complex developmental modes and life histories. Although many asteroid groups have been studied in detail, a consensus on most aspects of their evolutionary relationships has not been achieved. The need for an asteroid phylogeny is becoming increasingly apparent as biologists frame their studies in a comparative context (Emson and Young, 1994; Byrne, 1995; Janies, 1995; Hart et al., 1997; McEdward and Janies, 1997; Smith, 1997). A robust phylogeny will provide a basis for such comparative studies and an evolutionary scenario for testing generalizations about asteroid morphology, biogeography, ecology and development. |
BACKGROUND |
|---|
|
TOP
SYNOPSISINTRODUCTIONBACKGROUNDMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSReferences |
|---|
Two phylogenetic hypotheses based on analyses of morphological data (Blake, 1987
; Gale, 1987) have been proposed for the Asteroidea (summarized in Fig. 1). These conflicting proposals initiated some of the controversies in asteroid systematics and are still unresolved. Their conflicts are primarily due to analysis of different morphological characters, using different levels of taxon sampling and different assignments of character state polarity. Gale follows earlier arguments and classification (i.e., Spencer and Wright [1966] and McKnight [1975]) by including many different families in the large order Valvatida. Blake separates these families into three different orders (Spinulosida, Velatida and Valvatida). Gale's Valvatida is interpreted to be the sister group to the Forcipulatida (includes Brisingida); however, Blake's smaller Valvatida is interpreted to be the sister group of the Notomyotida. Gale considers characters uniting the Paxillosida as ancestral, whereas Blake considers them to be derived.
|
Recent studies have used molecular data in an effort to resolve the discrepancies of the Blake and Gale phylogenies. Lafay and colleagues (1995)
present an analysis of molecular (28S rRNA gene sequences) and morphological data (select characters from Blake [1987] and Gale [1987]). Their analysis of the molecular data alone results in several conflicting topologies due to weak phylogenetic signal. This signal is masked by the signal from morphological data when the two data sets are combined. Nonetheless, Lafay and colleagues (1995) follow the results from molecular data, concluding that the Paxillosida may not be monophyletic and that the paxillosid family Astropectinidae may be the sister group to the remaining asteroids (Fig. 2A). Wada and colleagues (1996) present an analysis of a different molecular data set (12S and 16S rDNA gene sequences). When using multiple optimality criteria for tree reconstruction, they find that the Paxillosida are paraphyletic (with Luidiidae sister to all other asteroids) and that the Spinulosida may have a very different relationship to other groups than has ever been proposed (Fig. 2B). In their phylogenies, the Spinulosida are either basal to all asteroids except the paxillosid Luidiidae or basal to a large Forcipulatida–Valvatida–Velatida clade.
|
Although Lafay and colleagues (1995)
and Wada and colleagues (1996) appear to resolve the position of the Paxillosida, they are not in agreement and do not address other questions of asteroid relationships. In fact, the results of these studies raise questions of ordinal and familial monophyly that were not realized in analyses of morphological data. However, due to their limited taxon sampling, neither study can address questions of ordinal or familial monophyly. Their discrepancies indicate that reconciliation of controversy in asteroid systematics will require more than choosing between the phylogenies proposed by Blake (1987) and Gale (1987). Here we present results from a study of multiple gene sequences from many asteroid species. Analysis of several genes with different evolutionary rates of change provides a means to evaluate both basal and more recent relationships. Analysis of a large taxon sample allows for assessment of familial as well as ordinal monophyly. With this approach we hope to redirect the focus of asteroid systematics back to the relationships themselves rather than the controversies of previous studies.
|
MATERIALS AND METHODS |
|---|
|
TOP
SYNOPSISINTRODUCTIONBACKGROUNDMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSReferences |
|---|
Thirty-two species from 17 families of six asteroid orders were collected, sequenced and included in one or more of our phylogenetic analyses (Table 1). Previously published mitochondrial gene sequences of the asteroids Patiriella vivipara and Oreaster reticulatus were also analyzed. Published sequences from an ophiuroid (Ophiopholis aculeata) and an echinoid (Strongylocentrotus purpuratus) were combined and included in the analysis with sequences from a more distantly related holothuroid (Cucumaria miniata) for outgroup comparison (see Table 1 for GenBank accession numbers).
|
DNA preparation and sequencing
All sea stars were preserved in 95% EtOH. Genomic DNA was extracted from tube foot tissue using phenol–chloroform techniques (Ausubel et al., 1997
). A 1826 bp segment of mitochondrial DNA comprising five transfer RNA (tRNA) gene sequences and a large portion of the cytochrome oxidase c subunit I (COI) gene was amplified and sequenced in three sections (A, B, and C) from genomic DNA using published primers and primers of our own design (Table 2). Amplification reactions were carried out in 25 µl volumes of a standard reaction mix with Taq DNA polymerase (GIBCO Life Technologies) using a MJ Research PTC-200 thermocycler (specific protocol available from the authors). All samples were purified in 2% NuSieve agarose (FMC BioProducts) and gel extracted (QIAGEN or GIBCO Life Technologies kits). Purified samples were transformed (Brown, 1991) into XLI Blue (Stratagene) competent cells using pGEM-T vector (Promega). Cloned samples were purified with a Wizard Plus Miniprep purification kit (Promega) and sequenced in forward and reverse directions using vector sequence primers [M13 (-20) M13 (rev)] on an ABI 373 automated sequencer with dye-terminator sequencing reaction mix (PE Applied Biosystems).
|
Analysis and phylogeny reconstruction
Sequence data from both strands were combined with Sequencher 3.0 (Gene Codes Corporation, 1995
), aligned using default parameters of the Wisconsin Sequence Analysis Program (GCG; Genetics Computer Group, 1997) and edited by eye (available from the authors). Some regions of the tRNA data could not be aligned. Since an accurate homology assessment in these regions (109 bp) is impossible, they were omitted from phylogenetic analyses. Alignment of the COI data, on the other hand, was unambiguous. Gaps in the sequence alignment were coded as missing. Small amounts of missing data in some taxa did not appear to affect taxon placement. These characters were included in all analyses. However, section B (658 bp) could not be amplified in ten of the 32 taxa. This large span of missing data influenced placement of taxa more than the sequence data that was available. Therefore, we performed analyses on two different taxon sets (Table 1). Set I included taxa (22 species) with all data present (sections A, B and C minus regions of ambiguous alignment, 1,717 bp). Set II included taxa (32 species) for which data from only sections A and C were available (minus regions of ambiguous alignment, 1,059 bp). Previously published sequence data from Patiriella vivipara and Oreaster reticulatus were included in both taxon sets.
The sequence data for each taxon set described above were analyzed with multiple optimality criteria using PAUP* v.4.0b1 (Swofford, 1998) to assess phylogenetic relationships. Tree reconstruction parameters were modified for each as follows: 1) Maximum parsimony (MP)—all data were unweighted. Analyses used heuristic search methods with 100 replicates of random stepwise addition and tree-bisection-reconnection (TBR) branch swapping algorithms. 2) Maximum likelihood (ML)—analyses were first performed with an assumption of the Hasegawa–Kishino–Yano (1985; HKY85) model of evolutionary change using empirical base frequencies, equal rates of variability at all sites and a transition/transversion ratio of two. Heuristic search methods were used as described for MP analyses but with only two replicates of random stepwise addition. Analyses were then performed with variability of sites, rate heterogeneity and transition/transversion ratio estimated by maximum likelihood (ML-estimated). Heuristic search methods were used with a single as-is stepwise addition and TBR algorithms. 3) Neighbor-Joining (NJ)—all analyses used distances calculated with HKY85 models and neighbor-joining tree reconstruction methods. Bootstrapping (Felsenstein, 1985) was employed in all analyses except ML-estimated which was computationally intensive.
A portion of the mitochondrial data (Section A) was combined with previously published 28S rRNA data (Lafay et al., 1995) and 12S and 16S rDNA data (Wada et al., 1996) for further analysis (set III, Table 1). Alignment of the 28S rRNA data set was identical to that published by Lafay and colleagues (1995). Original alignment of the 12S and 16S rDNA data could not be obtained and so these sequences were aligned using default parameters of GCG (Genetics Computer Group, 1997). Ambiguously aligned sequences from each data set were omitted, leaving a data set of 1581 characters. Only taxa that were in common (by genus or by family) between the three data sets were included (Table 1). Set III was analyzed with multiple optimality criteria as described above except that branch and bound search methods with a simple stepwise addition were used in MP analyses and heuristic search methods with ten replicates of random sequence addition were used in ML analysis. Bootstrapping (Felsenstein, 1985) was employed in all analyses.
|
RESULTS |
|---|
|
TOP
SYNOPSISINTRODUCTIONBACKGROUNDMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSReferences |
|---|
Analysis of taxon sets I and II (see Materials and Methods, Table 1) produced phylogenies that differed only slightly. Therefore, only results from analysis of the larger taxon set (set II) are shown. Different results seen in analysis of set I are discussed here. Unless specifically noted, reference to ordinal names indicates those groups designated by Blake (1987)
.
Regardless of optimality criterion used in tree reconstruction, a Brisingida–Forcipulatida–Spinulosida–Pterasteridae clade is recovered (Fig. 3). This clade has moderate bootstrap support in ML and NJ analyses. Smaller groups within this assemblage (e.g., Spinulosida and Pterasteridae) have much higher bootstrap support. The Brisingida group with the Pterasteridae in all analyses except NJ. These taxa and the Spinulosida group among forcipulate taxa making the Forcipulatida paraphyletic. In contrast, analysis of taxon set I indicates that the Forcipulatida and Brisingida are a monophyletic group. Close relationships between the Forcipulatida and Brisingida have been suggested before (Spencer and Wright, 1966; Gale, 1987), but the Forcipulatida have been considered unique because they share robust morphological characteristics that distinguish them from other asteroids (Blake, 1987). The Pterasteridae are assigned to the order Velatida but do not group with other velatid taxa, instead grouping with the brisingid or as basal to all other taxa in this Brisingida–Forcipulatida–Spinulosida–Pterasteridae clade.
|
The remaining taxa (Paxillosida–Valvatida–Solasteridae) show varied relationships depending on the different optimality criteria used in tree reconstruction (see Fig. 3). The Paxillosida is monophyletic except in MP analyses. As in the phylogeny proposed by Blake (1987)
, the Paxillosida is not basal in our trees. Constraining analyses to produce a basal Paxillosida results in phylogenies that are longer and have lower likelihood values. However, these phylogenies do not differ significantly from phylogenies of unconstrained analyses when tested with the Kishino–Hasegawa pairwise test (Kishino and Hasegawa, 1989; data not shown). The Valvatida and Solasteridae (assigned to the order Velatida) are variable in their placement with other asteroid taxa. Relationships among valvatid taxa are ambiguous and, except for close relatives, are not strongly supported by bootstrap analysis.
Analysis of set III ([Table 1], including sequence data from our section A, 28S rRNA [Lafay et al., 1995] and 12S and 16S rDNA [Wada et al., 1996]) did not result in a monophyletic Asteroidea when using ML and NJ optimality criteria (Fig. 4). The Paxillosida is monophyletic only in the NJ analysis. The Paxillosida does not have a basal position in these analyses, but its position is inconsistently resolved and is not supported by bootstrap analysis. Constraining analyses to produce a basal Paxillosida results in phylogenies that are longer and have lower likelihood values. However, these phylogenies do not differ significantly from phylogenies of unconstrained analyses when tested with the Kishino–Hasegawa pairwise test (Kishino and Hasegawa, 1989; data not shown). A clade containing members of the Forcipulatida (Asterias + Coscinasterias) and one of the representative spinulosid taxa (Echinaster) is consistently recovered with high bootstrap support in both MP and ML analyses. The spinulosid taxa (Echinaster and Henricia) never group together or with the representative velatid taxon (Crossaster), a presumed close relative.
|
|
DISCUSSION |
|---|
|
TOP
SYNOPSISINTRODUCTIONBACKGROUNDMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSReferences |
|---|
Analysis of our data alone and in combination with other data sets indicates some strong phylogenetic relationships that can be consistently recovered from DNA sequence of genes with rapid evolutionary rates of change (e.g., Pterasteridae, Solasteridae, Asterias + Coscinasterias). Ambiguous and weakly supported relationships among taxa seen in these analyses indicate problems that need to be addressed with further study. For example, the Spinulosida groups with taxa as varied as valvatids, paxillosids and forcipulates. Such groupings may not be indicative of actual species relationships, but may be due to greater divergence of these genes within the Spinulosida relative to other groups or greater divergence of the Spinulosida irrespective of the genes being studied. Such inconsistencies in molecular data could produce a phylogenetic bias. Unique divergence is most likely the case for the spinulosid genus Echinaster. Results from NJ analysis of set III (Fig. 4D) show Echinaster to have a much longer branch length than the other asteroid taxa, causing it to group with the holothuroid outgroup.
In addition to effects of gene choice on phylogeny reconstruction, effects of choosing representative asteroid taxa for inclusion in phylogenetic analyses must be considered. For example, the Valvatida historically has been a very large order (Spencer and Wright, 1966; Gale, 1987) that is still large despite Blake's (1987) division of it into three smaller orders (Valvatida, Velatida, and Spinulosida). A more detailed understanding of the members of this order and their relationships will be necessary before we can feel confident using only a few valvatid taxa as representatives of this diverse taxonomic group. Our results never show a monophyletic Valvatida. Further studies of the Valvatida are needed to test this result and to eliminate the possibility that taxon choice is confounding our analyses.
Relationships within and between asteroid families also need to be re-evaluated. The Forcipulatida (possibly including the Brisingida) is likely to be a monophyletic group. However, traditional definitions of families within this order may not be correct. Our results consistently show a Pycnopodia helianthoides–Rathbunaster californicus clade rather than grouping Pycnopodia with other members of the Asteriidae. Rathbunaster is a member of the deep-sea dwelling family Labidiasteridae that is distinguished from other forcipulate groups by some morphological characters. Families within the Velatida (Pterasteridae and Solasteridae) do not group together in our analyses. Close relationship between these taxa was expected because of their shared morphological characteristics. If the relationships recovered here are true, our interpretation of characters defining these families and orders must change. In this respect, molecular data may shed some light on our understanding of morphological evolution in the Asteroidea. Relationships within and between asteroid families may have to be discerned using molecular characters if morphological characters are more variable than previously thought.
|
CONCLUSIONS |
|---|
|
TOP
SYNOPSISINTRODUCTIONBACKGROUNDMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSReferences |
|---|
These analyses provide some new insights to asteroid relationships by incorporating multiple gene sequences and many representative taxa. Although inconclusive in some respects, our results have shed light on problem areas of asteroid systematics and identified routes through which phylogenetic analyses of the Asteroidea may be biased. Previous molecular-based phylogenetic studies of this group have focused on resolving controversies of morphological-based phylogenies and, in particular, the questionable position of the Paxillosida. Although previous molecular-based studies are in consensus that the Paxillosida are basal, our results show that the Paxillosida may be derived, providing an argument for a more careful investigation of asteroid relationships in the future. Such investigations should include an emphasis on testing familial and ordinal monophyly with consideration of taxon sampling and gene choice.
|
|
|
ACKNOWLEDGMENTS |
|---|
Many thanks go to Elizabeth Balser, Richard Emlet, Will Jaeckle, Dan Janies, Sue Lisin, Christopher Lowe, Chris Mah and Richard Turner for providing the samples used in this study. Ehab Abouheif, Alexa Bely, Dan Blake and two anonymous reviewers provided comments that helped improve the manuscript.
|
FOOTNOTES |
|---|
1 From the Symposium Evolution of Starfishes: Morphology, Molecules, Development, and Paleobiology presented at the Annual Meeting of the Society for Integrative and Comparative Biology, 6–10 January 1999, at Denver, Colorado.
2 E-mail: keknott{at}life.bio.sunysb.edu
3 Present address of Gregory A. Wray is Department of Zoology, Duke University, Box 90325, Durham, North Carolina 27708-0325.
|
References |
|---|
|
TOP
SYNOPSISINTRODUCTIONBACKGROUNDMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSReferences |
|---|
Ausubel, F. M., R. Brent, R. E. Kingston, D. D. Moore, J. G. Seidman, J. A. Smith, and K. Struhl.(eds.) 1997. Current protocols in molecular biology, 3 vols. John Wiley and Sons, New York.
Blake, D. B. 1987. A classification and phylogeny of post-Paleozoic sea stars (Asteroidea: Echinodermata). J. Nat. Hist, 21:481-528.[CrossRef]
Brown, T. A. 1991. Essential molecular biology: A practical approach.. Oxford Press, New York.
Byrne, M. 1995. Changes in larval morphology in the evolution of benthic development by Patiriella exigua (Asteroidea: Asterinidae). A comparison with the larvae of Patiriella species with planktonic development. Biol. Bull, 188:293-305.[Abstract]
Clark, A. M., and M. E. Downey. 1992. Starfishes of the Atlantic.. Chapman and Hall, London.
Emson, R. H., and C. M. Young. 1994. Comparative studies of the structure of the eye spots of Marthasterias glacialis and the bathyal sea-star Novodinia antillensis.. In B. David, A. Guille, J.-P. Feral, and M. Roux (eds.), Echinoderms: Dijon. Proceedings of the 8th International Echinoderm Conference, p. 284. A. A. Balkema, Rotterdam.
Felsenstein, J. 1985. Confidence limits on phylogenies: An approach using the bootstrap. Evolution, 39:783-791.[CrossRef][ISI]
Folmer, O., M. Black, W. Hoeh, R. Lutz, and R. Vrijenhoek. 1994. DNA primers for amplification of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates. Mol. Mar. Biol, 3:294-299.
Gale, A. S. 1987. Phylogeny and classification of the Asteroidea (Echinodermata). Zool. Soc. Linnean Soc, 89:107-132.
Gene Codes Corporation., 1995. Sequencher 3.0.. Gene Codes Corporation, Ann Arbor, MI.
Genetics Computer Group., 1997. Wisconsin Sequence Analysis Package, version 9.. Genetics Computer Group, Madison, WI.
Hart, M. W., M. Byrne, and M. J. Smith. 1997. Molecular phylogenetic analysis of life-history evolution in asterinid starfish. Evolution, 51:1848-1861.[CrossRef][ISI]
Hasegawa, M., H. Kishino, and T. Yano. 1985. Dating of the human-ape splitting by a molecular clock of mitochondrial DNA. J. Mol. Evol, 21:160-174.
Hyman, L. H. 1955. The invertebrates: Echinodermata, Vol. IV. McGraw-Hill, New York.
Janies, D. A. 1995. Reconstructing the evolution of morphogenesis and dispersal among velatid asteroids. Doctoral Ph.D. Diss., University of Florida, Gainesville, FL.
Kishino, H., and M. Hasegawa. 1989. Evaluation of the maximum likelihood estimate of the evolutionary tree topologies from DNA-sequence data and the branching order in Hominoidea. J. Mol. Evol, 29:170-179.[CrossRef][ISI][Medline]
Lafay, B., A. B. Smith, and R. Christen. 1995. A combined morphological and molecular approach to the phylogeny of asteroids (Asteroidea: Echinodermata). Syst. Biol, 44:190-208.[CrossRef]
McEdward, L. R., and D. A. Janies. 1997. Relationships among development, ecology and morphology in the evolution of echinoderm larvae and life cycles. Biol. J. Linn. Soc, 60:381-400.[CrossRef]
McKnight, D. G. 1975. Classification of somasteroids and asteroids (Asterozoa: Echinodermata). J. Royal Soc. N.Z, 5:13-19.
Smith, A. B. 1997. Echinoderm larvae and phylogeny. Ann. Rev. Ecol. Syst, 28:219-241.[CrossRef][ISI]
Smith, M. J., A. Arndt, S. Gorski, and E. Fajber. 1993. The phylogeny of echinoderm classes based on mitochondrial gene arrangements. J. Mol. Evol, 36:545-554.[CrossRef][ISI][Medline]
Spencer, W. K., and C. W. Wright. 1966. Asterozoans. In R.C. Moore. (ed.), Treastise on Invertebrate Paleontology, Part U, Echinodermata, pp. U4-U107. Geological Society of America and University of Kansas Press, Lawrence, KS.
Swofford, D. L. 1998. PAUP*: Phylogenetic Analysis Using Parsimony (and Other Methods), version 4.0.. Sinauer Associates, Sunderland, MA.
Wada, H., M. Komatsu, and N. Satoh. 1996. Mitochondrial rDNA phylogeny of the Asteroidea suggests the primitiveness of the Paxillosida. Mol. Phyl. Evol, 6:97-106.[Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||