Transcript
Phylogenetic Relationships of the Acanthocephala Inferred from 18SRibosomal DNA Sequences ThomasJ. Near, * ,1 JamesR. Garey, † and Steven A. Nadler ‡ * Center for Biodiversity, Illinois Natural History Survey, Champaign, Illinois 61820; † Department of Biological Sciences, University of South Florida, Tampa, Florida 33620; and ‡ Department of Nematology, University of California at Davis, Davis, California 95616 Received July 18, 1997; revised June 25, 1998 Phylogenetic relationships within the Acantho-cephala have remained unresolved. Past systematicefforts have focused on creating classifications withlittle consideration of phylogenetic methods. The Acan-thocephala are currently divided into three majortaxonomic groups: Archiacanthocephala, Palaeacan-thocephala, and Eoacanthocephala. These groups arecharacterized by structural features in addition to thetaxonomy and habitat of hosts parasitized. In thisstudy the phylogenetic relationships of 11 acantho-cephalan species are examined with 18S rDNA se-quences. Maximum parsimony, minimum evolution,and maximum likelihood methods are used to estimatephylogenetic relationships. Within the context of sampled taxa, all phylogenetic analyses are consistentwith monophyly of the major taxonomic groups of theAcanthocephala, suggesting that the current higherorder classification is natural. The molecular phylog-eny is used to examine patterns of character evolutionfor various structural and ecological characteristics of the Acanthocephala. Arthropod intermediate host dis-tributions, when mapped on the phylogeny, are consis-tent with monophyletic groups of acanthocephalans.Vertebrate definitive host distributions among theAcanthocephala display independent radiations intosimilar hosts. Levels of uncorrected sequence diver-gence among acanthocephalans are high; however,relative-rate tests indicate significant departure fromrate uniformity among acanthocephalans, arthropods,and vertebrates. This precludes comparison of 18Sdivergence levels to assess the relative age of theAcanthocephala. However, other evidence suggests anancient srcin of the acanthocephalan–arthropod para-sitic association. 1998 Academic Press INTRODUCTION Acanthocephala are helminth parasites that usearthropods and vertebrates to complete their life cycles.These helminths lack an alimentary tract and arecharacterized by the presence of a proboscis armed withrecurved hooks, a syncytial epidermis, and a lacunarsystem with circulatory channels that promotes directabsorption of nutrients through the body wall. Approxi-mately 820 species representing 125 genera have beendescribed from the three major classes (Amin, 1985).Amphipods, isopods, ostracods, copepods, insects, andmyriapods serve as intermediate hosts ofacanthocepha-lans. Teleosts, amphibians, turtles, snakes, lizards,birds, and mammals serve as definitive hosts. Someacanthocephalan life cycles involve paratenic (trans-port) hosts. Paratenic hosts are usually vertebratesthat ingest infected intermediate hosts and subse-quently are preyed upon by the definitive host (Nickol,1985). The diversity exhibited by the Acanthocephalain host distribution, host habitat, morphology, and lifehistory provides a wealth of material to examine inassociation with a phylogenetic hypothesis.The relationship of the Acanthocephala to otherinvertebrate phyla has been estimated recently byanalysis of structural and molecular data. The hypoth-esis that the Acanthocephala and Rotifera are sistertaxa has been supported in several phylogenetic stud-ies (Backeljau et al., 1993; Raff et al., 1994; Schram,1991; Winnepenninckx et al., 1995). In a cladisticanalysis of structural characters, Lorenzen (1985) hy-pothesized that the Acanthocephala share most recentcommon ancestry with rotifers of the class Bdelloidea.This hypothesized sister-group relationship betweenbdelloid rotifers and acanthocephalans was stronglysupported by phylogenetic analysis of complete 18SrDNA sequences (Garey et al., 1996). As a consequenceof this tree topology, recognition of Rotifera in phyloge-netic taxonomy necessitates the inclusion of the acan-thocephalans, as was recommended by Garey et al. (1996).Most investigations of acanthocephalan relation-ships predated the development ofphylogeneticsystem-atic methods (Hennig, 1966). Structural charactersprovided the basis for interpreting systematic group-ings of acanthocephalans; however, using these data to 1 To whom correspondence and reprint requests should be ad-dressed. Fax: (217) 333-4949. E-mail:
[email protected]. MOLECULAR PHYLOGENETICS AND EVOLUTION Vol. 10, No. 3, December, pp. 287–298, 1998 ARTICLE NO. FY980569287 1055-7903/98 $25.00Copyright 1998 by Academic PressAll rights of reproduction in any form reserved. develop phylogenetic hypotheses has been hamperedby a paucity of informative characters, morphologicaland ecological divergence among extant acanthocepha-lan groups, and an inability to polarize character states(Bullock, 1969; Conway Morris and Crompton, 1982).With no known fossil record and conflicting hypothesesconcerning free-living sister taxa (outgroups), determi-nation of shared derived character states among majoracanthocephalan groups has been complicated (Con-way Morris and Crompton, 1982). To date, relation-ships among the major lineages have not been suffi-ciently resolved.The currently accepted classification of the Acantho-cephala is an amalgamation of certain taxonomic hy-potheses ofMeyer and Van Cleave as modified by others(Meyer, 1932, 1933; Van Cleave, 1948, 1952; Bullock,1969;Amin, 1985). Three major taxonomic groups wererecognized as classes in the phylum Acanthocephala:Archiacanthocephala, Palaeacanthocephala, and Eoac-anthocephala (Golvan, 1959a, 1960, 1961, 1962). Thethree groups are distinguished by location of lacunarcanals, the persistence of ligament sacs in females,number and type of cement glands in males, numberand arrangement of proboscis hooks, intermediate anddefinitive hosts, and host ecology (Bullock, 1969; Duna-gan and Miller, 1991).In this study we examined the phylogenetic relation-ships of 11 acanthocephalan species as inferred fromcomplete 18S rDNAsequences. The monophyly of tradi-tionally recognized major taxonomic groupings wasassessed, and alternative phylogenetic and taxonomichypotheses of relationships were compared statisticallyusing the rDNA data. Hypotheses for the evolution of various structural and life history features used tradi-tionally in acanthocephalan taxonomy were developedby parsimony mapping on the molecular tree. MATERIALS AND METHODS Collection of Specimens Acanthocephalans were collected from vertebratedefinitive hosts or arthropod intermediate hosts. Speci-mens were stored at ultracold temperatures ( Ϫ 70°C) orin 95% ethanol until nucleic acids were extracted. Thespecies used in this analysis, GenBank accession num-bers for the acanthocephalan sequences, and theirclassification ( sensu Amin, 1985), with source hosts (h)in parentheses, are as follows: Archiacanthocephala- Gigantorhynchida -Gigantorhynchidae, Mediorhynchusgrandis AF001843 (h ϭ Sturnella magna, westernmeadowlark); Archiacanthocephala- Moniliformida -Moniliformidae, Moniliformis moniliformis Z19562(Telford and Holland, 1993) (h ϭ Rattus rattus, rat);Archiacanthocephala- Oligacanthorhynchida -Oligacan-thorhynchidae, Macracanthorhynchus ingens AF001844(h ϭ Procyon lotor, raccoon); Palaeacanthocephala- Echinorhynchida -Pomphorhynchidae, Pomphorhyn-chus bulbocolli AF001841 (h ϭ Onchorhynchus mykiss, rainbow trout);Palaeacanthocephala- Echinorhynchida -Rhadinorhynchidae, Leptorhynchoides thecatus AF001840(h ϭ Lepomis cyanellus, green sunfish); Palaeacantho-cephala- Polymorphida -Centrorhynchidae, Centrorhyn-chus conspectus U41399 (h ϭ Strix varia, barred owl);Palaeacanthocephala- Polymorphida -Plagiorhynchi-dae, Plagiorhynchus cylindraceus AF001839 (h ϭ Armadillidum vulgare, pillbug); Palaeacanthocephala- Polymorphida -Polymorphidae, Corynosoma enhydri AF001837 (h ϭ Enhydra lutris, sea otter) and Polymor- phus altmani AF001838 (h ϭ Enhydra lutris, sea ot-ter);Eoacanthocephala- Neoechinorhynchida -Neochino-rhynchidae, Neoechinorhynchus crassus AF001842 (h ϭ Catostomus commersoni, white sucker) and Neoechino-rhynchus pseudemydis U41400 (h ϭ Trachemys scriptaelegans, red-eared slider). Nucleic Acid Isolation, Polymerase Chain Reaction,and Sequencing Total nucleic acids were extracted from individualacanthocephalan specimens. Tissues were homog-enized on ice in STE buffer (10 mM Tris–HCl, pH 7.5,10 mM NaCl, 1 mM EDTA) and digested by adding 50µl of 20% SDS and 20 µl of proteinase K (10 mg/mL) andincubating at 50°C. The supernatant was extractedtwice with buffered phenol (pH 8.0) and once withchloroform/isoamyl alcohol (24:1). The nucleic acidswere precipitated overnight ( Ϫ 20°C) in a solution con-taining 50 µl of 3 M sodium acetate (pH 5.2) and 1.0 mlof absolute ethanol. The pellet was washed twice with70% ethanol, dried, resuspended in 100 µl of TE (10mM Tris–HCl, pH 7.5, 1 mM EDTA, pH 8.0) and storedat Ϫ 20°C. The concentration of nucleic acids wasestimated by spectrophotometry.The polymerase chain reaction (PCR) was used toamplify a homologous region of the 18S ribosomal DNA(rDNA) that ranged from 1745 to 1773 bp in 10acanthocephalan species. PCR was performed in 50-µlreactions containing 2.5 mM MgCl 2 , 0.25 mM eachdeoxynucleotide, 0.5 mM each primer (forward primer5 Ј -AGATTAAGCCATGCATGCGTAAG-3 Ј , reverse primer5 Ј -TGATCCTTCTGCAGGTTCACCTAC-3 Ј ), and 2.5units of Thermus aquaticus DNA polymerase in areaction buffer of 50 mM KCl, 10 mM Tris–HCl (pH8.3), and 0.1% Triton X-100. Template DNA used inPCR ranged from 100 to 300 ng. Water used in the PCRwas double distilled, autoclaved, and irradiated with400 mJ/cm 2 of254-nm light in an ultraviolet crosslinkerto inactivate potential contaminating nucleic acids(Sarkar and Sommer, 1990). Thermal cycling was per-formed using an initial denaturation of 94°C for 4.0 minfollowed by 25 cycles of 94°C (30 s), 60°C (30 s.), and72°C (1.5 min). A final incubation of 5 min (72°C) wasperformed to completely extend the amplified product.PCR product size was verified by electrophoresis in a1% agarose gel using DNAsize standards.288 NEAR, GAREY, AND NADLER The amplified 18S product was separated from PCRreactants by ultrafiltration using a 30,000 MW cutoff polysurfone tube (Millipore Corporation), ligated intopGEM-T vector plasmid using T 4 ligase (Promega), andused to transform DH5 ␣ - Escherichia coli. Colonies thatwere positive by blue/white selection were screenedusing internal 18S primers and PCR to verify theidentity of the insert. Plasmid DNA was isolated fromindividual clones and used as template for the sequenc-ing reactions. At least two clones from each individualwere sequenced.Chain-termination cycle sequencing was performedusing the ⌬ Taq Cycle Sequencing kit (Amersham UnitedStates Biochemical, Cleveland, OH) with [ 32 P] dATP asthe radionuclide. Each species was sequenced for bothstrands using a total of 13 internal primers and 2 vectorprimers. Sequences of the internal forward primerswith their 5 Ј annealing positions numbered accordingtothe 18S rDNAof Moniliformis moniliformis (Acantho-cephala: Archiacanthocephala, Genbank Z19562) areAACCGCGAATGGCTCATT (46), CGGAGAGGGAGCC-TGAGAAACGGC (346), GCCGCGGTAATTCCAGCTC(537), CGGAAGCTGAGGTAATGATT (812), CGGGGG-GAGTATGGTTGC (1073), CTTAAGCACACGAAGAG-GAGC (1371), and ACACCGCCCGTCGCTACT (1600).Reverse primers used were CTCATGCTCTCTCTC-CGG (363), GAATTACCGCGGCTGCTGG (549), GTT-GTTCGTCTGGCGGTGATC (904), CTGGTGTGCCCC-TCCGTC (1133), CCATTGTAGCGCGCGTG (1446), andTGATCCTTCTGCAGGTTCACCTAC (1766). Sequenc-ing products were separated by electrophoresis in 6%polyacrylamide/8.3 M urea gels and visualized by auto-radiography. Complete 18S sequences were assembledby overlapping individual sequence files in the ASSEM-GEL program of the PC Gene package (IntelliGeneticsMountain View, CA). Ambiguities between overlappingsequence files were rechecked and resolved by sequenc-ing additional clones as required. Phylogenetic Analysis of Sequence Data Sequences of 11 acanthocephalan species and 2 ro-tifer species (Garey et al., 1996) were aligned accordingto a secondary structure model (Van de Peer et al., 1994) using the DCSE editor (De Rijk and De Wachter,1993). Brachionus plicatilis (GenBank U29235) and Philodina acuticornis (GenBank U41281) were used toroot trees in the acanthocephalan analyses. These taxawere selected as outgroups because a more completephylogenetic analysis of invertebrate diversity (Garey et al., 1996) based on 18S sequences represented thesetworotifers as most closely related tothe acanthocepha-lans.Maximum parsimony analyses were performed usinga test version of PAUP* (4.0d54) (Swofford, 1997). In allmaximum parsimony analyses, character-states in-ferred as gaps were treated as missing data, onlyminimal-length trees were retained, and zero-lengthbranches collapsed. The branch-and-bound algorithmwas utilized and bootstrap analysis (1000 replications,branch-and-bound algorithm) was used to examine therelative robustness of inferred monophyletic groups.Levels of support for groups recovered in parsimonyanalyses was also evaluated by decay analysis (Bremer,1988), wherein strict consensus trees were constructedfor all trees at successive steps longer than the shortesttree until the consensus tree collapsed to an unresolvedbush. The decay index shows the number of substitu-tions that must be added to the most parsimonioushypothesis before each clade is no longer supported.The data set was assessed for phylogenetic signal byexamination of the g 1 value of the tree length distribu-tion for 10 5 randomly generated trees (Hillis andHuelsenbeck, 1992).Phylogenetic relationships were also estimated usingthe minimum evolution method (Rzhetsky and Nei,1992) as executed in PAUP* (4.0d54) with LogDet/ paralinear distances (Lockhart et al., 1994;Lake, 1994).Minimum evolution trees were recovered using heuris-tic searches with random addition of taxa (10 repli-cates), tree-bisection/reconnection branch-swapping,and the steepest descent option. Minimum evolutionbootstrap analysis involved 1000 replications, withheuristic search options as described previously, exceptthat random additions of taxa were not replicated.Maximum likelihood analysis was executed usingPAUP* (4.0d54) and included all sites in the alignedsequence data set. Options invoked in maximum likeli-hood analysis included nucleotide frequencies esti-mated from the data, number of substitution types setat 2, and rates assumed to follow a ␥ distribution withthe ␣ -shape parameter and proportion of invariablesites estimated via maximum likelihood. The ␥ approxi-mation was set to four rate categories and the averagerate for each category was represented by the mean. Inaddition, the HKY85 model (Hasegawa et al., 1985)with rate heterogeneity was used, with the transition/ transversion ratio estimated via maximum likelihoodand starting branch lengths obtained using the Rogers–Swofford approximation. Bootstrap analysis employed100 replications using model parameters estimated forthe original dataset.Alternative phylogenetic hypotheses (Fig. 1) weretested statistically by two different methods using theTree Scores option in PAUP* (4.0d54). Topologies wereassessed using a pairwise parsimony method proposedby Templeton (1983)and modified by Felsenstein (1993).This method uses the mean and variance of stepdifferences between alternative topologies and is re-lated to the test using log-likelihood differences of Kishino and Hasegawa (1989). For maximum likeli-hood analyses, alternative topologies were assessedusing the test proposed by Kishino and Hasegawa(1989), where the mean and variance of log-likelihooddifferences are compared between trees. In both parsi-289 ACANTHOCEPHALAN PHYLOGENY mony and maximum likelihood assessments, alterna-tive topologies were considered significantly different if the mean exceeded 1.96 standard deviations. Alterna-tive hypotheses assessed included the inferred topolo-gies from maximum parsimony, maximum likelihood,and minimum evolution analyses. Other topologiesexamined by these methods were representative of previous systematic hypotheses of acanthocephalanrelationships and included the phylogenetic hypothesisof Van Cleave (1952), which depicts the Archiacantho-cephala and Palaeacanthocephala as sister taxa (Meta-canthocephala) (Fig. 1a), a sister-group relationshipbetween the Archiacanthocephala and Eoacantho-cephala (Fig. 1b), a hypothesis proposed by Brooks andMcLennan (1993, pp. 369–373) (Fig. 1c), and the topol-ogy inferred from Petrochenko’s (1956, pp. 159–162)classification of the Acanthocephala (Fig. 1d). Analysis of Morphological and Ecological Characters Eight morphological characters and four ecologicalcharacters were coded for analysis and treated asunordered (Table 1).All morphological characters exam-ined were binary and three of four ecological characterswere coded as multistate. Character states for indi- FIG. 1. Alternative topologies tested by statistical methods (see Table 3). Bpl, Brachionus plicatilis; Cco, Centrorhynchus conspectus; Cen, Corynosoma enhydri; Lth, Leptorhynchoides thecatus; Mgr, Mediorhynchus grandis; Min, Macracanthorhynchus ingens; Mmo, Moniliformismoniliformis; Ncr, Neoechinorhynchus crassus; Nps, Neoechinorhynchus pseudemydis; Pal, Polymorphus altmani; Pac, Philodina acuticornis; Pbu, Pomphorhynchus bulbocolli; Pcy, Plagiorhynchus cylindraceus. 290 NEAR, GAREY, AND NADLER vidual acanthocephalan species were taken from theliterature (Bullock, 1969; Meyer, 1932, 1933; VanCleave, 1952). Characters examined and coding of character states were as follows: (A) Proboscis recep-tacle: 0 ϭ single-walled; 1 ϭ double-walled. (B) Cementgland i: 0 ϭ giant nuclei; 1 ϭ fragmented nuclei. (C)Cement gland ii: 0 ϭ multiple; 1 ϭ single. (D) Giantsubcuticular nuclei: 0 ϭ absent; 1 ϭ present. (E)Ligament sac i: 0 ϭ persistent; 1 ϭ nonpersistent. (F)Ligament sac ii: 0 ϭ double; 1 ϭ single. (G) Lacunarsystem: 0 ϭ dorsal or dorsal and ventral; 1 ϭ lateral.(H) Protonephridia: 0 ϭ absent; 1 ϭ present. (I) Inter-mediate host: 0 ϭ amphipod or isopod; 1 ϭ insect ordiplopod;2 ϭ ostracod or copepod. (J)Intermediate hosthabitat:0 ϭ aquatic;1 ϭ terrestrial. (K)Definitive host:0 ϭ fish; 1 ϭ turtle; 2 ϭ bird; 3 ϭ mammal. (L)Definitive host habitat: 0 ϭ aquatic; 1 ϭ terrestrial.MacClade 3.0 (Maddison and Maddison, 1992) wasused to map these characters by parsimony and tocalculate consistency indices for these characters onthe 18S gene tree and alternative topologies (Fig. 1). 18S Ribosomal DNA Sequence Divergence Sequence divergence among acanthocephalans wascompared to 18S rDNA sequence divergence amongarthropods and vertebrates. Given the absence of aknown fossil record for the Acanthocephala, compara-tive analysis of molecular data are among the onlymethods with the potential to provide estimates of therelative times of srcin and divergence among acantho-cephalan groups.Nineteen arthropods, seven vertebrates, and a singlehemichordate (Table 2) were aligned according to asecondary structure model (Van de Peer et al., 1994)using the DCSE editor (De Rijk and De Wachter, 1993)and are deposited in TreeBASE (Sanderson et al., 1994). Reported values for pairwise sequence diver-gence are based on this global alignment. Tree topolo-gies (deposited in TreeBASE, not shown) were inferredusing maximum parsimony as implemented in PAUP*4.0. Three most-parsimonious trees were recovered andthese trees were used to evaluate relative-rate varia-tion among OTUs using the method of Wu and Li (1985)as implemented in the r8s program (Sanderson, 1997,version 0.10). RESULTS Phylogenetic Analysis of 18S rDNA Sequences The aligned 18S rDNA sequence data consisted of 1848 nucleotide sites for the 11 acanthocephalan and 2rotifer outgroup species. Eight hundred forty-one of thealigned nucleotide sites were variable and when allsites with gaps were excluded 713 sites were variable.Five hundred seventy-five of the 841 sites that variedwere phylogenetically informative in maximum parsi-mony analysis. Unweighted maximum parsimonyanalysis of the aligned data set yielded a single mostparsimonious tree of 1673 steps, with a consistencyindex (excluding uninformative characters) of 0.674(Fig. 2). The minimum possible branch lengths support-ing internal nodes within the Acanthocephala rangedfrom 4 to 150 apomorphies. The g 1 statistic for distribu-tion of tree lengths from 10 5 randomly generated trees( Ϫ 0.855) was significant ( P Ͻ 0.01), indicating that thedata set is more structured than are random data(Hillis and Huelsenbeck, 1992).Minimum evolution analysis resulted in a single treewith a score of 0.9387 (Fig. 3a). This tree differed fromthe maximum parsimony tree (Fig. 2) with respect torelationships within the Archiacanthocephala, and the TABLE 1Matrix of Morphological and Ecological Characters for Species of Acanthocephala Used for ParsimonyMapping on Phylogenetic Hypotheses A B C D E F G H I J K L Moniliformis moniliformis 0 0 1 0 0 0 0 0 1 1 3 1 Macracanthorhynchus ingens 0 0 1 0 0 0 0 1 1 1 3 1 Mediorhynchus grandis 0 0 1 0 0 0 0 0 1 1 2 1 Neoechinorhynchus pseudemydis 0 0 0 1 0 1 0 0 2 0 1 0 Neoechinorhynchus crassus 0 0 0 1 0 1 0 0 2 0 0 0 Pomphorhynchus bulbocolli 1 1 1 0 1 1 1 0 0 0 0 0 Leptorhynchoides thecatus 1 1 1 0 1 1 1 0 0 0 0 0 Plagiorhynchus cylindraceus 1 1 1 0 1 1 1 0 0 1 2 1 Centrorhynchus conspectus 1 1 1 0 1 1 1 0 0 1 2 1 Polymorphus altmani 1 1 1 0 1 1 1 0 0 0 3 0 Corynosoma enhydri 1 1 1 0 1 1 1 0 0 0 3 0Steps 1 1 1 1 1 1 1 1 2 3 5 3Consistency index 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 0.33 0.60 0.33 Note. See text for description of characters and coding of character states. Number of steps and consistency index for each character, whenoptimized on the phylogenetic hypothesis, are indicated in the last two rows. 291 ACANTHOCEPHALAN PHYLOGENY
