Evolutionary History of Free-Swimming and Sessile Lifestyles in Urochordates As Deduced from 18S Rdna Molecular Phylogeny
Total Page：16
File Type：pdf, Size：1020Kb
Evolutionary History of Free-Swimming and Sessile Lifestyles in Urochordates as Deduced from 18S rDNA Molecular Phylogeny Hiroshi Wada1 School of Animal and Microbial Sciences, The University of Reading Whiteknights, Reading, England Whether the ancestral chordates were free-swimming or sessile is a longstanding question that remains to be settled. Vertebrates and amphioxi are free-swimming, but the most basal chordate subphylum (the urochordates) includes both sessile and free-swimming species. Here, I report molecular phylogenetic analyses of 18S rDNA of urochor- dates to deduce which lifestyle is ancestral. This revealed a close relationship between salps and doliolids and paraphyly of the ascidians. An early divergence of larvaceans, which show a tadpole-like body plan throughout life, is also supported by the analyses. Based on this phylogeny, a free-swimming ancestor for chordates is more parsimonious than a sessile ancestor. The evolutionary history of various lifestyles of chordates from this ancestral form is proposed. Downloaded from https://academic.oup.com/mbe/article/15/9/1189/1412733 by guest on 27 September 2021 Introduction One of the main controversies concerning the ori- as for locomotion. The points at issue are which of these gin of vertebrates is about the nature of the chordate lifestyles represents the primitive condition for urochor- ancestors. In particular, did vertebrates evolve from free- dates, and which lifestyle was possessed by the ancestor living ancestors, or did they derive from sessile ances- of all chordates. Haeckel (1868) extended his recapitu- tors, similar to ascidians, via paedomorphosis? Verte- lation theory to this case and proposed that the ®rst brates are a member of phylum Chordata, which also chordates were free-living, retaining a tadpole-like body includes cephalochordates (represented by amphioxi) plan throughout life, and that the sessile lifestyle of the and urochordates. Among them, urochordates are the ascidian has been acquired secondarily as a terminal ad- most basal group, with cephalochordates and vertebrates dition to development. This idea has been supported by being sister groups (Maisey 1986; Schaeffer 1987; Brus- several authors, including Darwin (1871) and, more re- ca and Brusca 1990; Wada and Satoh 1994; Turbeville cently, Tokioka (1971), Jollie (1973), and Jefferies et al. 1994; Nielsen 1995). Cephalochordates and ver- (1986). In contrast, Garstang (1928) proposed that the tebrates show a highly motile body plan throughout their ancestral chordates had sessile adults and that the tad- lives which is referred to here as the tadpole-like body pole-like body plan evolved in the larval stage of these plan. In both taxa, the motility is driven by lateral mus- ancestors. Cephalochordates and vertebrates then cles using a notochord or vertebrae as a support, with a evolved their fully motile lifestyle by paedomorphosis neural tube and a gut lying dorsally and ventrally, re- of the sessile ancestors. Garstang (1928) also insisted spectively. There is little doubt that the common ances- that the larvaceans evolved by paedomorphosis from the tor of the cephalochordates and the vertebrates showed doliolids. Although Garstang's views have been accept- this tadpole-like body plan throughout life. The contro- ed and developed by several modern authors (Berrill versy centers on the more basal ancestors of all chor- 1955; Romer 1967), the controversy still remains to be dates; a problem that has been dif®cult to resolve due settled. A reliable phylogeny of the urochordates should to the existence of a variety of lifestyles in the uro- reveal the polarity of changes in the evolution of life- chordates. Urochordates are classi®ed into ®ve groups: styles of the urochordates, which, in turn, should allow ascidians, salps, doliolids, pyrosomes, and larvaceans. implications to be drawn concerning the origin of the Three types of lifestyles are found for them. The ®rst is chordates. observed for ascidians, which have a sessile adult; the Wada and Satoh (1994) have reported analyses of second includes the pelagic adult of salps, doliolids, and urochordate relationships using 18S rDNA sequences pyrosomes. In these two types, a tadpole-like body plan where only larvaceans, salps, and ascidians were stud- is seen only in the larval stage, although some species ied. Although the early divergence of larvaceans is sup- have lost the tadpole larvae secondarily. In the third type ported by those analyses, further studies including py- of lifestyle, observed in larvaceans, the tadpole-like rosomes and doliolids have been weighted in order to body plan is retained throughout life, although the mo- deduce the evolutionary history of urochordates and to tile tail in the adult is used for collecting foods as well draw implications for the evolutionary origin of chor- dates. The relationship between larvaceans and doliolids is especially crucial in this regard, because several au- 1Present address: Seto Marine Biological Laboratory, Kyoto Uni- versity, Japan. thors have proposed that the larvaceans may have evolved by paedomorphosis from doliolids (Garstang Key words: Urochordata, 18S rDNA, molecular phylogeny, chor- date evolution, ascidian, larvacean. 1928; Bone 1960; Nielsen 1995). Here, I report molec- ular phylogenetic analyses of 18S rDNAs from all ®ve Address for correspondence and reprints: Hiroshi Wada, Seto Ma- rine Biological Laboratory, Kyoto University, 459 Shirahama-cho, representative groups of urochordates (larvaceans, as- Nishimuro-gun, Wakayama 649-2211, Japan. E-mail: cidians, salps, pyrosomes, and doliolids), and, based on [email protected]. the phylogeny concluded here, I propose an evolution- Mol. Biol. Evol. 15(9):1189±1194. 1998 ary history of urochordates, with special emphasis on q 1998 by the Society for Molecular Biology and Evolution. ISSN: 0737-4038 lifestyle evolution. 1189 1190 Wada Downloaded from https://academic.oup.com/mbe/article/15/9/1189/1412733 by guest on 27 September 2021 FIG. 1.ÐMolecular phylogenetic trees constructed from urochordate 18S rDNA sequences. The tree topology and branch lengths come from the NJ analysis. Bootstrap values for internal branches are shown in each node; the upper numbers are bootstrap values by NJ and the italic numbers under them are those by MP. Materials and Methods Phylogenetic Analyses DNA Isolation and PCR Ampli®cation An alignment of the sequences was constructed by eye using SeqApp manual aligner. The alignment is Genomic DNA extraction and PCR ampli®cation available upon request. Clustal V (Higgins, Bleasby, and were performed as described in Wada and Satoh (1994), Fuchs 1992) was used for the neighbor-joining (NJ) except that for ampli®cation of Ciona intestinalis and method (Saitou and Nei 1987). Evolutionary distances Halocynthia roretzi rDNA, Pfu DNA polymerase (Stra- were calculated according to Kimura's (1980) two-pa- tagene) was used. For ampli®cation of Oikopleura dioi- rameter method. Gaps and insertions were excluded in ca, DNA from a genomic DNA library was used as the the NJ analyses. The con®dence of the tree topology was template. assessed by 1,000 bootstrap resamplings (Felsenstein 1985). For maximum-parsimony (MP) analyses, PAUP Sequences 3.1.1 branch-and-bound options (Swofford 1993) were Sequences were determined after subcloning am- used. The con®dence was assessed by 100 bootstrap re- pli®ed DNAs into pUC 18 plasmid vector (Pharmacia). samplings. fastDNAml 1.0 (Olsen et al. 1993) was used Primers used for sequencing reactions are described in for the maximum-likelihood (ML) analyses (Felsenstein Wada and Satoh (1994). In order to exclude any se- 1981). Jumble options were used to ®nd a true ML tree. quence errors arising from misampli®cation by Taq DNA polymerase (Wada et al., 1992), I determined se- Results quences of three clones from independent PCR ampli- I determined almost-full-length 18S rDNA se- ®cations for each species; identical sequence in at least quences from a larvacean, O. dioica; a pyrosome, Py- two clones was taken as representative. In the course of rosoma atlanticum; a doliolid, D. nationalis; and two sequence determination, I found major and minor ver- species of ascidian, C. intestinalis and H. roretzi. Pre- sions of 18S rDNA in the O. dioica library; the minor viously reported 18S rDNA sequences of the ascidian copy (one clone from eight) is very similar to a previ- Styela plicata and the salp Thalia democratica are also ously reported sequence from Oikopleura sp. in Wada included for molecular phylogenetic analyses, together and Satoh (1994) (Oikopleura sp. 1 in ®g. 1). It is likely, with those of Balanoglossus carnosus (hemichordate therefore, that the minor version of 18S rDNA is from acornworm) and Asterias amurensis (echinoderm star- a different species of Oikopleura (Oikopleura sp. 2: ac- ®sh) as outgroups. These outgroup taxa were chosen be- cession number AB013015) contaminated during the cause of their slower substitution rate of 18S rDNAs, preparation of the genomic library. I also found a minor which is re¯ected in the shorter branch lengths of the copy (1 clone from 11) from Doliolum nationalis which phylogenetic tree in Wada and Satoh (1994). Accession is very divergent but shows a clear af®nity to the se- numbers of the sequences studied here are listed in table quence of the major D. nationalis 18S rDNA. The un- 1. The full sequence of 18S rDNA from the ascidian usual divergence suggests it is probably a pseudogene, Herdmania momus has been reported