Molecular Phylogeny of Early Vertebrates: Monophyly of the Agnathans as Revealed by Sequences of 35 Genes
Max-Planck-Institut für Biologie, Abteilung Immungenetik, Tübingen, Germany4ti, http://www.100md.com
Abstract4ti, http://www.100md.com
Extant vertebrates are divided into three major groups: hagfishes (Hyperotreti, myxinoids), lampreys (Hyperoartia, petromyzontids), and jawed vertebrates (Gnathostomata). The phylogenetic relationships among the groups and within the jawed vertebrates are controversial, for both morphological and molecular studies have rendered themselves to conflicting interpretations. Here, we use the sequences of 35 nuclear protein-encoding genes to provide definitive evidence for the monophyly of the Agnatha (jawless vertebrates, a group encompassing the hagfishes and lampreys). Our analyses also give a strong support for the separation of Chondrichthyes (cartilaginous fishes) before the divergence of Osteichthyes (bony fishes) from the other gnathostomes.4ti, http://www.100md.com
Key Words: molecular phylogeny • lamprey • hagfish • cartilaginous fish • bony fish
Introductionff](g, 百拇医药
Hagfishes and lampreys are two major extant lineages of jawless vertebrates, which represent the most basal group of vertebrates. To clarify their phylogenetic relationship is of great importance in understanding many aspects of early vertebrate evolution including morphology, palaeontology, and development of body plans, as well as the emergence of complex biological systems such as adaptive immunity in jawed vertebrates. Hagfishes and lampreys share several derived characters (synapomorphies) that set them apart from jawed vertebrates . In addition to the absence of a biting apparatus derived from gill arches (jaws), they are distinguished by the following characteristics: absence of the third (horizontal) semicircular canal in the inner ear; large notochord; horny teeth; absence of paired appendages; gill passages expanded into pouches (internal gills) connected through pores (rather than slits) with the exterior; more than five external openings of the gills; and branchial skeleton in the form of lattice-work about the pouches. These and other shared characters were originally used to assign hagfishes and lampreys to Cyclostomata or Agnatha (when extinct jawless vertebrates were included), separate from Gnathostomata . Later, however, several authors pointed out that lampreys share various other characters with jawed fishes, all of which are absent in hagfishes and the cephalochordate amphioxus, Branchiostoma. These lamprey-gnathostome synapomorphies include control of the heartbeat via a branch of the vagus nerve (neurogenic as opposed to myogenic control); ectodermal (rather than endodermal) origin and complex histological structure of the adenohypophysis; presence of arcualia—small irregular cartilages on either side of the dorsal nerve cord, representing elements of the vertebral column; absence of a functional pronephros; capability to control osmotic pressure of the blood by means of descending and ascending tubules of the kidney; eyes with a lens and extrinsic muscles enabling their motility; and pterygial muscles attached to and capable of changing the shape of the fins. Under the influence of these studies, the view prevailed that lampreys were more closely related to jawed fishes than to hagfishes, and the last of the above was used to erect a separate group of "myopterygians" comprising lampreys and gnathostomes . This interpretation was also supported by some molecular data, such as the sequences of mtDNA , 18S rRNA , the 7S K RNA gene , and the vasotocin gene . However, recent studies of mtDNA , 18S, 28S, and 5.8S rRNA (e.g.), globin (e.g.), and other genes (e.g.) support the monophyly of the agnathans.
The published molecular studies are inconclusive primarily because analysis of a single locus or small numbers of loci suffers from a large sampling error and from lack of high statistical support. Further, different tree-drawing methods often generate different phylogenies (Delarbre et al. 2000). Even in cases in which some methods have given high statistical support for a particular phylogeny, other methods have supported this phylogeny poorly (e.g). The choice of a too distant outgroup and the use of only a few vertebrate lineages may have distorted the phylogenetic relationships among the lineages (e.g.,;). A number of authors have pointed out that mitochondrial genes are unsuitable for resolving phylogenetic relationships among major vertebrate groups . Similarly, the use of rRNA is prone to potential problems such as the choice of alignment or difference in base frequencies in different lineages . To overcome these problems, we obtained nucleotide sequences of the coding regions of 35 nuclear genes from each of the following groups: cephalochordates (used as an outgroup), hagfishes, lampreys, cartilaginous fishes (in this group only 31/35 genes were available), bony fishes (represented by teleost fishes), and tetrapods (represented by human).
Materials and Methodsgd5r, 百拇医药
Acquisition of Sequence Datagd5r, 百拇医药
Total RNA was isolated from whole bodies of Branchiostoma lanceolatum and Danio rerio or liver of Myxine glutinosa, Petromyzon marinus, and Scyliorhinus canicula and used for the cDNA synthesis with the help of the Smart cDNA library construction kit (Clontech). The cDNA was amplified by the polymerase chain reaction (PCR), and the amplification products were cloned and sequenced. The PCR amplification was carried out in the PTC-200 Programmable Thermal Controller (Biozym) with the aid of the Advantage2PCR kit (Clontech) with 1 µl of the cDNA and 1 mM of each of the sense and antisense primers. Selected PCR products were isolated from low-melting-point agarose gels (Gibco BRL), purified with the GFX kit (Amersham Biosciences) and cloned into pcr 2.1 Topo vector with the aid of the TOPO TA kit (Invitrogen). Sequencing reactions were carried out using the Thermo sequenase Primer Cycle Sequencing Kit (Amersham Biosciences) and processed by the LI-COR Long ReadIR 4200 DNA Sequencer (MWG Biotech). Other sequences used in this study were obtained by a Blast search conducted against the GenBank non-redundant protein database for the orthologs of the selected proteins. Multiple alignments of amino acid sequences were produced by the CLUSTALW 1.82 and checked by eye. Orthologies of the sequences were examined with the help of Neighbor-Joining (NJ) trees (Saitou and Nei 1987). Species names, accession numbers, and alignments of the sequences used in this study are provided on Supplementary Material at the MBE Web site.
Phylogenetic Analysis*/, 百拇医药
Phylogenetic trees were drawn by the maximum parsimony (MP) method with PAUP 4.0b10 , the NJ method, the minimum evolution (ME) method , and the maximum likelihood (ML) method with PAML 3.0c and MOLPHY 2.3b3 . In all the analyses, amino acid sequences were used after excluding positions with indels. Phylogenetic trees were drawn for each protein and for concatenated sequences of all the proteins. The NJ and ME trees were drawn using Poisson-correction and gamma distances . The gamma parameter was estimated by the ML method by using the tree topologies in and and the JTT model . The ML trees were drawn using JTT, Dayhoff, and Poisson models with or without assuming gamma distribution for rate variation across amino acid positions (G option) and with or without using observed amino acid frequencies in the data (F option). The MP trees were obtained by branch and bound search. In the ME and ML methods all the possible tree topologies (15 for five taxa and 105 for six taxa) were examined. Two thousand bootstrap iterations were conducted for the MP, NJ, and ME trees; 10,000 bootstrap replications were generated for the ML trees by the RELL method for all the possible tree topologies. The bootstrap probabilities for each branch on the ML trees and were calculated by summing up the RELL bootstrap values for tree topologies that contain the same branch. This way of calculating bootstrap values for branches is essentially the same procedure taken by the standard bootstrap test, except that the likelihood values for different tree topologies in bootstrap replications are computed by the RELL method rather than by carrying out the optimization process in the ML method each time. In the ML method, the results for individual proteins were combined by the TOTALML approach . The bootstrap test by this approach was conducted by the TOTALML program in MOLPHY . In TOTALML, the G option was not available.
fig.ommittedk, 百拇医药
FIG. 1. Phylogenetic trees based on concatenated amino acid sequences of 35 nuclear protein-encoding genes (cartilaginous fishes excluded). A, Maximum parsimony (MP) tree. B, Neighbor-joining (NJ) tree. C, Maximum likelihood (ML) tree. The numbers shown on the interior nodes are bootstrap probabilities (BP) in percent. The scale bars show the number of amino acid replacements per sequence for the MP tree and per amino acid position for the NJ and ML trees. For the MP tree, the tree length TL = 8,039, the consistency index CI = 0.95 and the retention index RI = 0.53. Branch lengths of the MP tree were computed by MINF option in PAUP. The NJ tree shown was drawn with Poisson-correction distance. The ML tree shown was obtained using the JTT modelk, 百拇医药
fig.ommittedk, 百拇医药
FIG. 2. Phylogenetic trees based on concatenated amino acid sequences of 31 nuclear protein-encoding genes (cartilaginous fishes included). A, Maximum parsimony (MP) tree. B, Neighbor-Joining (NJ) tree. C, Maximum-likelihood (ML) tree. Numbers on the interior branches are bootstrap probabilities in percent. For the MP tree, the tree length TL = 4,889, the consistency index CI = 0.89 and the retention index RI = 0.49. The NJ tree shown was drawn using Poisson-correction distance. The ML tree was drawn using the JTT model
Topologies of the best trees obtained from concatenated sequences were compared to the topologies of all the other trees by the non-parametric and tests for MP trees, the Kishino-Hasegawa test () for ML trees, and the test for ME trees. In the TOTALML approach, the significance of the difference between the likelihood values of the best and second best trees was examined by Kishino-Hasegawa test. In TOTALML approach log-likelihood of tree T is the sum (lTi) of log-likelihood values (lTi) of tree T for each protein. The difference of log-likelihood values of trees T and U is the sum of the log-likelihood difference for each protein82!.g$1, 百拇医药
The variance [V(l)] of the difference of the two trees is calculated as the sum of the variance of the log-likelihood difference for the two trees of each protein82!.g$1, 百拇医药
Significance of the difference of log-likelihood values of the two trees is tested by the statistic
The Kishino-Hasegawa test examines the null hypothesis H0: l = 0. If the likelihood values are computed for a subset of all the possible tree topologies, the tree topology with the highest likelihood may not be the best tree of all the possible tree topologies. Even in such a case it is possible that the best tree in the subset is significantly supported by the Kishino-Hasegawa test in comparison to other trees. However, in this study we computed likelihood values for all the possible tree topologies in all cases. Therefore, if the best tree has a significantly higher likelihood value than the second best tree, the likelihood value of the best tree is expected to be significantly higher than the rest of the possible tree topologies.uus, 百拇医药
All trees drawn, including those using LogDet distance , gave the same topologies as those in and . The NJ and ME trees with LogDet distance are not discussed because there was no sign of amino acid frequency changes in different lineages.
Results and Discussionh;, http://www.100md.com
Collection of Protein Sequencesh;, http://www.100md.com
The list of the genes used in this study appears in . The dataset consists of amino acid sequences of 35 nuclear genes from each of the following groups: cephalochordates (Branchiostoma lanceolatum, B. belcheri, B. floridae, used as an outgroup), hagfishes (Myxine glutinosa, Eptatretus burgeri, E. stoutii), lampreys (Petromyzon marinus, Lampetra fluviatilis, Lethenteron japonicum, L. reissneri), cartilaginous fishes (Scyliorhinus canicula, S. stellaris, Ginglystoma cirratum, Chilscyllium punctatum; in this group only 31/35 genes were available), bony fishes (Danio rerio, Cyprinus carpio, Carassius auratus, Ictalurus punctatus, Takifugu rubripes, Xiphophorus maculatus, Opsanus tau, Paralichthys olivaceus, Oncorhynchus mykiss), and tetrapods (Homo sapiens). Over 60% of the sequences were obtained in our laboratory; the remaining sequences were from databases. In the final analysis, tetrapods were represented by human sequences, but in earlier examinations, the comparisons also included amphibians (mostly Xenopus laevis) and birds (domestic fowl, Gallus gallus) with results that were congruent with those based on the human dataset. Similarly, the inclusion of sarcopterygian fishes (the African lungfish Protopterus dolloi and the coelacanth Latimeria chalumnae) in the analysis did not change the conclusions reached without these taxa.
fig.ommittedtwn, 百拇医药
Table 1 Genes Used in Phylogenetic Analyses.twn, 百拇医药
The final analysis was limited to a single species for each of the six lineages examined in each gene. We believe this approach to be justified by observations made in the process of selecting suitable genes for inclusion in the final dataset, a process which involved phylogenetic analyses of sequences derived from different species of each of the six lineages. The analyses revealed the interspecies differences within the amphioxus, hagfish, and lamprey lineages to be small and hence without influence on the overall tree topologies (see also ). By contrast, interspecies differences within the cartilaginous and bony fish lineages were often large, and here some intermingling of species from different lineages occurred in the trees. In the majority of cases, however, the separation of the lineages was clear.twn, 百拇医药
For each of the 35 loci, orthology of the genes in the different groups was determined by phylogenetic analysis of as many related sequences as were available. In cases of doubt, an effort was made to identify the cDNAs of the true orthologs, and if that search failed, the locus was excluded from the final analysis. The assembled loci represent a diverse group of nuclear protein-encoding genes which includes housekeeping as well as regulatory and large as well as small genes of different functions . About half of the genes encode ribosomal proteins which are known to evolve slowly and to be relatively short. In our analyses, they yielded trees differing in their topologies from those produced by the combined analyses of the whole set somewhat more often than the other sequences, probably because of the relative paucity of phylogenetic information. The loci included in our data set were relatively conserved. The number of amino acid substitutions per position between tetrapod and amphioxus sequences was 0.4 on average for the different proteins; the largest number was 1.3 .
Monophyly of the Agnathans&, 百拇医药
shows the phylogenetic trees drawn for the concatenated sequences of 35 loci by the MP, NJ, and ML methods. The set consisted of a total of 9,098 shared amino acid positions (6,786 positions when cartilaginous fishes were included; ). All the methods applied to the concatenated sequences using different substitution models and distance measures and the TOTALML approach generated the same tree topology shown in , in which the agnathans form a monophyletic cluster. The monophyly of the agnathans was supported by all the methods used with a high statistical confidence. The bootstrap probabilities (BP) for the two interior branches were 99%–100% for different substitution models by the ML method including the TOTALML approach (not shown). In all the tree-drawing methods the second best tree was ((tetrapod, bony fish), lamprey), hagfish, amphioxus) and the best tree was significantly supported by the tests of tree comparison (for the MP tree P = 0.002 by Kishino-Hasegawa's and Templeton's nonparametric tests; for the ME trees P = 0.0001 using Poisson correction distance and P = 0.02 using the gamma distance; for the ML trees P = 0.0001 by the Kishino-Hasegawa test using all the substitution models examined; and for the TOTALML approach P < 0.0016 by the Kishino-Hasegawa test using all the substitution models.) In the MP method the tree length of the second best tree was larger than that of the best tree by 50. The gamma parameter, inversely related to the extent of rate variation across amino acid positions, estimated for the concatenated sequences by the ML method was 0.38.
The same tree topology as that generated from the concatenated sequences and the TOTALML approach was obtained for 19 genes (54%) by the NJ method and for 14 genes (45%) with the MP and ML methods (see Supplementary Material). By considering only whether or not the cluster of hagfish and lamprey appeared in the trees drawn, the monophyly of the agnathans was supported by the majority of the trees (57% to 62% depending on the methods) of the individual proteins. The alternative agnathan phylogeny in which lampreys are more closely related to gnathostomes than to hagfishes was supported by only 10% to 27% of the trees.lq&?ix, http://www.100md.com
Phylogenetic Relationship of Cartilaginous Fish and Bony Fish and the Impact of the Inclusion of the Former on Agnathan Monophylylq&?ix, http://www.100md.com
shows the phylogenetic trees drawn from concatenated sequences of 31 nuclear protein-encoding genes for which sequences of cartilaginous fishes are available. All the tree-drawing methods used for the concatenated sequences and the TOTALML approach produced the tree topology shown in, in which the agnathans form a monophyletic cluster and cartilaginous fishes are positioned outside the cluster of tetrapod and bony fish. With this data set the monophyly of the agnathans was also strongly supported. Bootstrap probabilities for the ancestral branches of hagfish and lamprey were 99%–100% by all the methods. The position of cartilaginous fishes outside the cluster of bony fishes and tetrapods is significantly supported by most of the methods applied, but with slightly lower confidence than the agnathan monophyly. Bootstrap probabilities for the ancestral branch of tetrapods and bony fishes were 97%–100% by the NJ method using all the distance measures, and by the ML methods using different substitution models including the TOTALML approach. The BP (94%) of the MP method for the clustering of tetrapods and bony fishes was slightly lower than the 5% significance level. The second best tree was (((tetrapod, bony fish), cartilaginous fish), lamprey), hagfish, amphioxus) by the MP method and by the TOTALML approach using the Poisson model with the F option regardless of whether the G option was specified; and (((tetrapod, (bony fish, cartilaginous fish)), (lamprey, hagfish), amphioxus) by the ME method and by the ML method except for the TOTALML approach mentioned earlier. Tests of tree comparisons were significant by the ME method using the Poisson-correction distance (P = 0.001), by the ML method (P = 0.02–0.004), and by the TOTALML approach (P = 0.017–0.0001), but not significant by the MP method (P = 0.12) and by the ME method using the gamma distance (P = 0.26). In the MP method, the difference in the tree lengths between the best and the second best trees was 18. The gamma parameter estimated for the concatenated sequences by the ML method was 0.38.
The same tree topology as that generated from the concatenated sequences and the TOTALML approach () was obtained for eight, six, and five genes by the NJ, MP, and ML methods, respectively. The hagfish-lamprey cluster was observed in the majority (50%–65%) of genes, as in the case of the dataset without cartilaginous fishes mentioned in the previous section. By contrast, the branching pattern ((tetrapod, bony fish), cartilaginous fish) shown in appeared in the trees of only about 30% of genes. The alternative branching patterns for the relationship of the cartilaginous and bony fish were observed in even lower frequencies: the pattern (tetrapod, (bony fish, cartilaginous fish)) was observed in about 20% of the genes by all the methods, and the pattern ( (tetrapod, cartilaginous fish), bony fish) appeared in 19% of the NJ trees and 6% of the MP and ML trees.0[v9, 百拇医药
Is the Agnathan Monophyly a Result of Long Branch Attraction?0[v9, 百拇医药
Because hagfishes and lampreys are represented by long branches on the phylogenetic trees ( and ), it could be argued that their monophyly is the result of a long branch attraction . Distortions of phylogenetic reconstructions by long branch attraction occur especially when grossly simplifying models of the evolutionary process and the parsimony method are used . To avoid this problem, we used other methods of phylogenetic reconstruction in addition to those based on the maximum parsimony principle, and in the analyses we took into account different substitution rates across sites, genes, and lineages. We therefore do not think that long branch attraction is the cause of the observed agnathan monophyly. It has been argued that potential artifactual attractions between branches can be avoided by including more taxa in the data set (e.g). Others have pointed out, however, that adding taxa to break long branches increases the probability of distorting phylogenetic relationships in certain cases and that a better way of dealing with this problem is to increase the number of amino acid positions in the analyses . This is the way we have chosen in the present study.
Conclusion|jv, 百拇医药
The phylogenetic analysis leads to two important conclusions. First, the hagfishes and lampreys form a monophyletic group which constitutes a sister group of the gnathostomes, with cephalochordates as an outgroup. And second, the cartilaginous fishes diverged from the common ancestor of bony fishes and tetrapods before the latter two lineages diverged from each other. The first conclusion is well supported statistically by all tree-drawing methods applied to the dataset of the 35 proteins. It is also supported by the majority (57% to 62%, depending on the method) of the trees of the individual proteins. The second conclusion is supported by a high statistical confidence except by the maximum parsimony method and the minimum evolution method with gamma distance . The lower statistical support by the latter methods is probably due to a smaller number of amino acid positions in the data set that included cartilaginous fishes and to the relative shortness of the time interval during which the cartilaginous fishes and the bony fishes separated from the tetrapod lineage. The placement of cartilaginous fishes outside the cluster formed by bony fishes and tetrapods is in agreement with the traditional interpretation of this part of vertebrate phylogeny and in contradiction to the recent claim by placing cartilaginous fishes inside the bony fish clade. This latter phylogeny is based on mtDNA sequences and supported by high statistical significance. As mentioned earlier, however, sequence data of mitochondrial protein-encoding genes are not suitable for deep phylogenetic reconstructions .
The origin and nature of the morphological characters previously claimed to support the close relationship of lampreys to jawed fishes will have to be re-examined in view of the strong support of the molecular data for the agnathan monophyly. Many of these characters may not be homologous . Some may have arisen independently by convergent evolution in the lamprey and gnathostome lineages, while others still may have been present in the agnathan ancestors and lost in the hagfish . A good example of the uncertainties surrounding the nature of some of these characters is the agnathan "tongue." Originally, the resemblance between the "rasping tongue" of the lampreys and the "laterally biting jaws" of the hagfishes was one of the reasons for grouping the two lineages into Cyclostomata. Later, the resemblance was proclaimed to be superficial and the structures were interpreted as having been acquired independently by the two lineages (e.g.). Later still, however, pointed out at least 11 synapomorphous anatomical features of the feeding apparatus were shared by lampreys and hagfishes, but not by gnathostomes. Now the molecular data clearly side with the view that the "tongue" of hagfishes and lampreys is indeed a homologous organ.
Acknowledgements3/]rl, 百拇医药
We thank Jane Kraushaar for editorial assistance and Ryszard Lorenz for technical assistance. We are grateful to Dr. Hirohisa Kishino for the advice on the Kishino-Hasegawa test for TOTALML approach, and Drs. J.-O. Strömberg and B. Bergström for providing us M. glutinosa tissues.3/]rl, 百拇医药
Literature Cited3/]rl, 百拇医药
Adachi, J., and M. Hasegawa. 1996. MOLPHY2.3b3: programs for molecular phylogenetics based on maximum likelihood. Institute of Statistical Mathematics, Tokyo.3/]rl, 百拇医药
Carroll, R. L. 1987. Vertebrate paleontology and evolution. W. H. Freeman, New York.3/]rl, 百拇医药
Delarbre C., H. Escriva, C. Gallut, V. Barriel, P. Kourilsky, P. Janvier, V. Laudet, and G. Gachelin. 200l. The complete nucleotide sequence of the mitochondrial DNA of the agnathan Lampetra fluviatilis: bearings on the phylogeny of cyclostomes. Mol. Biol. Evol. 17:519-529.3/]rl, 百拇医药
Delarbre, C., C. Gallut, V. Barriel, P. Janvier, and G. Gachelin. 2002. Complete mitochondrial DNA of the hagfish, Eptatretus burgeri: the comparative analysis of mitochondrial DNA sequences strongly supports the cyclostome monophyly. Mol. Phylogenet. Evol. 22:184-192.
Goodman, M., M. M. Miyamoto, and J. Czelusniak. 1987. Pattern and process in vertebrate phylogeny revealed by coevolution of molecules and morphologies. Pp. 141–176 in C. Patterson, ed. Molecules and morphology in evolution: conflict or compromise? Cambridge University Press, New York.[4e1#, 百拇医药
Graybeal, A. 1998. Is it better to add taxa or characters to a difficult phylogenetic problem? Syst. Biol. 47:9-17.[4e1#, 百拇医药
Gu, X., and W.-H. Li. 1996. Bias-corrected paralinear and LogDet distances and tests of molecular clocks and phylogenies under nonstationary nucleotide frequencies. Mol. Biol. Evol. 13:1375-1383.[4e1#, 百拇医药
Gürsoy, H. C., D. Koper, and B. J. Benecke. 2000. The vertebrate 7S K RNA separates hagfish (Myxine glutinosa) and lamprey (Lampetra fluviatilis). J. Mol. Evol. 50:456-464.[4e1#, 百拇医药
Hardisty, M. W. 1982. Lampreys and hagfishes: analysis of cyclostome relationships. Pp. 165–259 in M. W. Hardisty and I. C. Potter, eds. The biology of lampreys. Academic Press, London.[4e1#, 百拇医药
Hedges, S. B. 2001. Molecular evidence for the early history of living vertebrates. Pp. 119–134 in P. E. Ahlberg, ed. Major events in early vertebrate evolution: paleontology, phylogeny, genetics, and development. Taylor & Francis, London.
Janvier, P. 1981. The phylogeny of the Craniata, with particular reference to the significance of fossil "agnathans." J. Vert. Paleontol. 1:121-159.?ks!aq, 百拇医药
Janvier, P. 1996. The dawn of the vertebrates: characters versus common ascent in the rise of current vertebrate phylogenies. Palaeontology 39:259-287.?ks!aq, 百拇医药
Jefferies, R. P. S. 1986. The ancestry of the vertebrates. British Museum, London.?ks!aq, 百拇医药
Jones, D. T., W. R. Taylor, and J. M. Thornton. 1992. The rapid generation of mutation data matrices from protein sequences. Comput. Appl. Biosci. 8:275-282.?ks!aq, 百拇医药
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.?ks!aq, 百拇医药
Kuraku, S., D. Hoshiyama, K. Katoh, H. Suga, and T. Miyata. 1999. Monophyly of lampreys and hagfishes supported by nuclear DNA-coded genes. J. Mol. Evol. 49:729-735.?ks!aq, 百拇医药
Lanfranchi, G., A. Pallavicini, P. Laveder, and G. Valle. 1994. Ancestral hemoglobin switching in lampreys. Dev. Biol. 164:402-408.
Lipscomb, D. L., J. S. Farris, M. Källersjö, and A. Tehler. 1998. Support, ribosomal sequences, and the phylogeny of the eukaryotes. Cladistics 14:303-338.:^asj0, 百拇医药
Løvtrup, S. 1977. The phylogeny of vertebrata. Wiley, New York.:^asj0, 百拇医药
Mallatt, J., and J. Sullivan. 1998. 28S and 18S rDNA sequences support the monophyly of lampreys and hagfishes. Mol. Biol. Evol. 15:1706-1718.:^asj0, 百拇医药
Mallatt, J., J. Sullivan, and C. J. Winchell. 2001. The relationship of lampreys to hagfishes: a spectral analysis of ribosomal DNA sequences. Pp. 106–118 in P. E. Ahlberg, ed. Major events in early vertebrate evolution: paleontology, phylogeny, genetics and development. Taylor & Francis, London.:^asj0, 百拇医药
Nei, M., and S. Kumar. 2000. Molecular evolution and phylogenetics. Oxford University Press, New York.:^asj0, 百拇医药
Philippe, H., A. Chenuil, and A. Adoutte. 1994. Can the Cambrian explosion be inferred through molecular phylogeny? Development Suppl:15–25.:^asj0, 百拇医药
Poe, S., and D. L. Swofford. 1999. Taxon sampling revisited. Nature 398:299-300.
Rasmussen, A. S., and U. Arnason. 1999. Molecular studies suggest that cartilaginous fishes have a terminal position in the piscine tree. Proc. Natl. Acad. Sci. USA 96:2177-2182.$, 百拇医药
Rasmussen, A. S., A. Janke, and U. Arnason. 1998. The mitochondrial DNA molecule of the hagfish (Myxine glutinosa) and vertebrate phylogeny. J. Mol. Evol. 46:382-388.$, 百拇医药
Romer, A. S. 1966. Vertebrate paleontology. University of Chicago Press, Chicago.$, 百拇医药
Rosenberg, M. S., and S. Kumar. 2001. Incomplete sampling is not a problem for phylogenetic inference. Proc. Natl. Acad. Sci. USA 98:10751-10756.$, 百拇医药
Rzhetsky, A., and M. Nei. 1992. A simple method for estimating and testing minimum-evolution trees. Mol. Biol. Evol. 9:954-967.$, 百拇医药
Stock, D. W., and G. S. Whitt. 1992. Evidence from 18S ribosomal RNA sequences that lampreys and hagfishes form a natural group. Science 57:787-789.$, 百拇医药
Suga, H., D. Hoshiyama, S. Kuraku, K. Katoh, K. Kubokawa, and T. Miyata. 1999. Protein tyrosine kinase cDNAs from amphioxus, hagfish, and lamprey: isoform duplications around the divergence of cyclostomes and gnathostomes. J. Mol. Evol. 49:601-608.
Suzuki, M., K. Kubokawa, H. Nagasawa, and A. Urano. 1995. Sequence analysis of vasotocin cDNAs of the lamprey, Lampetra japonica, and the hagfish, Eptatretus burgeri: evolution of cyclostome vasotocin precursors. J. Mol. Endocrinol. 14:67-77.@;mm?3i, 百拇医药
Swofford, D. L. 2002. PAUP*: phylogenetic analysis using parsimony (*and other methods). Sinauer Associates, Sunderland, Mass.@;mm?3i, 百拇医药
Takezaki, N., and T. Gojobori. 1999. Correct and incorrect vertebrate phylogenies obtained by the entire mitochondrial DNA sequences. Mol. Biol. Evol. 16:590-601.@;mm?3i, 百拇医药
Templeton, A. 1983. Phylogenetic inference from restriction endonuclease cleavage site maps with particular reference to the evolution of humans and the apes. Evolution 37:221-244.@;mm?3i, 百拇医药
Thompson, J. D., D. G. Higgins, and T. J. Gibson. 1994. ClustalW: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22:4673-4680.@;mm?3i, 百拇医药
Turbeville, J. M., J. R. Schulz, and R. A. Raff. 1994. Deuterostome phylogeny and the sister group of the chordates: evidence from molecules and morphology. Mol. Biol. Evol. 11:648-655.@;mm?3i, 百拇医药
Yalden, D. W. 1985. Feeding mechanisms as evidence for cyclostome monophyly. Zool. J. Linn. Soc. 84:291-300.@;mm?3i, 百拇医药
Yang, Z. 1996. Phylogenetic analysis using parsimony and likelihood methods. J. Mol. Evol. 42:294-307.@;mm?3i, 百拇医药
Yang, Z. 1999. PAML: a program package for phylogenetic analysis by maximum likelihood. Version 3.0c. University College London, London.@;mm?3i, 百拇医药
Accepted for publication October 29, 2002.(Naoko Takezaki Felipe Figueroa Zofia Zaleska-Rutczynska and Jan Klein)
Abstract4ti, http://www.100md.com
Extant vertebrates are divided into three major groups: hagfishes (Hyperotreti, myxinoids), lampreys (Hyperoartia, petromyzontids), and jawed vertebrates (Gnathostomata). The phylogenetic relationships among the groups and within the jawed vertebrates are controversial, for both morphological and molecular studies have rendered themselves to conflicting interpretations. Here, we use the sequences of 35 nuclear protein-encoding genes to provide definitive evidence for the monophyly of the Agnatha (jawless vertebrates, a group encompassing the hagfishes and lampreys). Our analyses also give a strong support for the separation of Chondrichthyes (cartilaginous fishes) before the divergence of Osteichthyes (bony fishes) from the other gnathostomes.4ti, http://www.100md.com
Key Words: molecular phylogeny • lamprey • hagfish • cartilaginous fish • bony fish
Introductionff](g, 百拇医药
Hagfishes and lampreys are two major extant lineages of jawless vertebrates, which represent the most basal group of vertebrates. To clarify their phylogenetic relationship is of great importance in understanding many aspects of early vertebrate evolution including morphology, palaeontology, and development of body plans, as well as the emergence of complex biological systems such as adaptive immunity in jawed vertebrates. Hagfishes and lampreys share several derived characters (synapomorphies) that set them apart from jawed vertebrates . In addition to the absence of a biting apparatus derived from gill arches (jaws), they are distinguished by the following characteristics: absence of the third (horizontal) semicircular canal in the inner ear; large notochord; horny teeth; absence of paired appendages; gill passages expanded into pouches (internal gills) connected through pores (rather than slits) with the exterior; more than five external openings of the gills; and branchial skeleton in the form of lattice-work about the pouches. These and other shared characters were originally used to assign hagfishes and lampreys to Cyclostomata or Agnatha (when extinct jawless vertebrates were included), separate from Gnathostomata . Later, however, several authors pointed out that lampreys share various other characters with jawed fishes, all of which are absent in hagfishes and the cephalochordate amphioxus, Branchiostoma. These lamprey-gnathostome synapomorphies include control of the heartbeat via a branch of the vagus nerve (neurogenic as opposed to myogenic control); ectodermal (rather than endodermal) origin and complex histological structure of the adenohypophysis; presence of arcualia—small irregular cartilages on either side of the dorsal nerve cord, representing elements of the vertebral column; absence of a functional pronephros; capability to control osmotic pressure of the blood by means of descending and ascending tubules of the kidney; eyes with a lens and extrinsic muscles enabling their motility; and pterygial muscles attached to and capable of changing the shape of the fins. Under the influence of these studies, the view prevailed that lampreys were more closely related to jawed fishes than to hagfishes, and the last of the above was used to erect a separate group of "myopterygians" comprising lampreys and gnathostomes . This interpretation was also supported by some molecular data, such as the sequences of mtDNA , 18S rRNA , the 7S K RNA gene , and the vasotocin gene . However, recent studies of mtDNA , 18S, 28S, and 5.8S rRNA (e.g.), globin (e.g.), and other genes (e.g.) support the monophyly of the agnathans.
The published molecular studies are inconclusive primarily because analysis of a single locus or small numbers of loci suffers from a large sampling error and from lack of high statistical support. Further, different tree-drawing methods often generate different phylogenies (Delarbre et al. 2000). Even in cases in which some methods have given high statistical support for a particular phylogeny, other methods have supported this phylogeny poorly (e.g). The choice of a too distant outgroup and the use of only a few vertebrate lineages may have distorted the phylogenetic relationships among the lineages (e.g.,;). A number of authors have pointed out that mitochondrial genes are unsuitable for resolving phylogenetic relationships among major vertebrate groups . Similarly, the use of rRNA is prone to potential problems such as the choice of alignment or difference in base frequencies in different lineages . To overcome these problems, we obtained nucleotide sequences of the coding regions of 35 nuclear genes from each of the following groups: cephalochordates (used as an outgroup), hagfishes, lampreys, cartilaginous fishes (in this group only 31/35 genes were available), bony fishes (represented by teleost fishes), and tetrapods (represented by human).
Materials and Methodsgd5r, 百拇医药
Acquisition of Sequence Datagd5r, 百拇医药
Total RNA was isolated from whole bodies of Branchiostoma lanceolatum and Danio rerio or liver of Myxine glutinosa, Petromyzon marinus, and Scyliorhinus canicula and used for the cDNA synthesis with the help of the Smart cDNA library construction kit (Clontech). The cDNA was amplified by the polymerase chain reaction (PCR), and the amplification products were cloned and sequenced. The PCR amplification was carried out in the PTC-200 Programmable Thermal Controller (Biozym) with the aid of the Advantage2PCR kit (Clontech) with 1 µl of the cDNA and 1 mM of each of the sense and antisense primers. Selected PCR products were isolated from low-melting-point agarose gels (Gibco BRL), purified with the GFX kit (Amersham Biosciences) and cloned into pcr 2.1 Topo vector with the aid of the TOPO TA kit (Invitrogen). Sequencing reactions were carried out using the Thermo sequenase Primer Cycle Sequencing Kit (Amersham Biosciences) and processed by the LI-COR Long ReadIR 4200 DNA Sequencer (MWG Biotech). Other sequences used in this study were obtained by a Blast search conducted against the GenBank non-redundant protein database for the orthologs of the selected proteins. Multiple alignments of amino acid sequences were produced by the CLUSTALW 1.82 and checked by eye. Orthologies of the sequences were examined with the help of Neighbor-Joining (NJ) trees (Saitou and Nei 1987). Species names, accession numbers, and alignments of the sequences used in this study are provided on Supplementary Material at the MBE Web site.
Phylogenetic Analysis*/, 百拇医药
Phylogenetic trees were drawn by the maximum parsimony (MP) method with PAUP 4.0b10 , the NJ method, the minimum evolution (ME) method , and the maximum likelihood (ML) method with PAML 3.0c and MOLPHY 2.3b3 . In all the analyses, amino acid sequences were used after excluding positions with indels. Phylogenetic trees were drawn for each protein and for concatenated sequences of all the proteins. The NJ and ME trees were drawn using Poisson-correction and gamma distances . The gamma parameter was estimated by the ML method by using the tree topologies in and and the JTT model . The ML trees were drawn using JTT, Dayhoff, and Poisson models with or without assuming gamma distribution for rate variation across amino acid positions (G option) and with or without using observed amino acid frequencies in the data (F option). The MP trees were obtained by branch and bound search. In the ME and ML methods all the possible tree topologies (15 for five taxa and 105 for six taxa) were examined. Two thousand bootstrap iterations were conducted for the MP, NJ, and ME trees; 10,000 bootstrap replications were generated for the ML trees by the RELL method for all the possible tree topologies. The bootstrap probabilities for each branch on the ML trees and were calculated by summing up the RELL bootstrap values for tree topologies that contain the same branch. This way of calculating bootstrap values for branches is essentially the same procedure taken by the standard bootstrap test, except that the likelihood values for different tree topologies in bootstrap replications are computed by the RELL method rather than by carrying out the optimization process in the ML method each time. In the ML method, the results for individual proteins were combined by the TOTALML approach . The bootstrap test by this approach was conducted by the TOTALML program in MOLPHY . In TOTALML, the G option was not available.
fig.ommittedk, 百拇医药
FIG. 1. Phylogenetic trees based on concatenated amino acid sequences of 35 nuclear protein-encoding genes (cartilaginous fishes excluded). A, Maximum parsimony (MP) tree. B, Neighbor-joining (NJ) tree. C, Maximum likelihood (ML) tree. The numbers shown on the interior nodes are bootstrap probabilities (BP) in percent. The scale bars show the number of amino acid replacements per sequence for the MP tree and per amino acid position for the NJ and ML trees. For the MP tree, the tree length TL = 8,039, the consistency index CI = 0.95 and the retention index RI = 0.53. Branch lengths of the MP tree were computed by MINF option in PAUP. The NJ tree shown was drawn with Poisson-correction distance. The ML tree shown was obtained using the JTT modelk, 百拇医药
fig.ommittedk, 百拇医药
FIG. 2. Phylogenetic trees based on concatenated amino acid sequences of 31 nuclear protein-encoding genes (cartilaginous fishes included). A, Maximum parsimony (MP) tree. B, Neighbor-Joining (NJ) tree. C, Maximum-likelihood (ML) tree. Numbers on the interior branches are bootstrap probabilities in percent. For the MP tree, the tree length TL = 4,889, the consistency index CI = 0.89 and the retention index RI = 0.49. The NJ tree shown was drawn using Poisson-correction distance. The ML tree was drawn using the JTT model
Topologies of the best trees obtained from concatenated sequences were compared to the topologies of all the other trees by the non-parametric and tests for MP trees, the Kishino-Hasegawa test () for ML trees, and the test for ME trees. In the TOTALML approach, the significance of the difference between the likelihood values of the best and second best trees was examined by Kishino-Hasegawa test. In TOTALML approach log-likelihood of tree T is the sum (lTi) of log-likelihood values (lTi) of tree T for each protein. The difference of log-likelihood values of trees T and U is the sum of the log-likelihood difference for each protein82!.g$1, 百拇医药
The variance [V(l)] of the difference of the two trees is calculated as the sum of the variance of the log-likelihood difference for the two trees of each protein82!.g$1, 百拇医药
Significance of the difference of log-likelihood values of the two trees is tested by the statistic
The Kishino-Hasegawa test examines the null hypothesis H0: l = 0. If the likelihood values are computed for a subset of all the possible tree topologies, the tree topology with the highest likelihood may not be the best tree of all the possible tree topologies. Even in such a case it is possible that the best tree in the subset is significantly supported by the Kishino-Hasegawa test in comparison to other trees. However, in this study we computed likelihood values for all the possible tree topologies in all cases. Therefore, if the best tree has a significantly higher likelihood value than the second best tree, the likelihood value of the best tree is expected to be significantly higher than the rest of the possible tree topologies.uus, 百拇医药
All trees drawn, including those using LogDet distance , gave the same topologies as those in and . The NJ and ME trees with LogDet distance are not discussed because there was no sign of amino acid frequency changes in different lineages.
Results and Discussionh;, http://www.100md.com
Collection of Protein Sequencesh;, http://www.100md.com
The list of the genes used in this study appears in . The dataset consists of amino acid sequences of 35 nuclear genes from each of the following groups: cephalochordates (Branchiostoma lanceolatum, B. belcheri, B. floridae, used as an outgroup), hagfishes (Myxine glutinosa, Eptatretus burgeri, E. stoutii), lampreys (Petromyzon marinus, Lampetra fluviatilis, Lethenteron japonicum, L. reissneri), cartilaginous fishes (Scyliorhinus canicula, S. stellaris, Ginglystoma cirratum, Chilscyllium punctatum; in this group only 31/35 genes were available), bony fishes (Danio rerio, Cyprinus carpio, Carassius auratus, Ictalurus punctatus, Takifugu rubripes, Xiphophorus maculatus, Opsanus tau, Paralichthys olivaceus, Oncorhynchus mykiss), and tetrapods (Homo sapiens). Over 60% of the sequences were obtained in our laboratory; the remaining sequences were from databases. In the final analysis, tetrapods were represented by human sequences, but in earlier examinations, the comparisons also included amphibians (mostly Xenopus laevis) and birds (domestic fowl, Gallus gallus) with results that were congruent with those based on the human dataset. Similarly, the inclusion of sarcopterygian fishes (the African lungfish Protopterus dolloi and the coelacanth Latimeria chalumnae) in the analysis did not change the conclusions reached without these taxa.
fig.ommittedtwn, 百拇医药
Table 1 Genes Used in Phylogenetic Analyses.twn, 百拇医药
The final analysis was limited to a single species for each of the six lineages examined in each gene. We believe this approach to be justified by observations made in the process of selecting suitable genes for inclusion in the final dataset, a process which involved phylogenetic analyses of sequences derived from different species of each of the six lineages. The analyses revealed the interspecies differences within the amphioxus, hagfish, and lamprey lineages to be small and hence without influence on the overall tree topologies (see also ). By contrast, interspecies differences within the cartilaginous and bony fish lineages were often large, and here some intermingling of species from different lineages occurred in the trees. In the majority of cases, however, the separation of the lineages was clear.twn, 百拇医药
For each of the 35 loci, orthology of the genes in the different groups was determined by phylogenetic analysis of as many related sequences as were available. In cases of doubt, an effort was made to identify the cDNAs of the true orthologs, and if that search failed, the locus was excluded from the final analysis. The assembled loci represent a diverse group of nuclear protein-encoding genes which includes housekeeping as well as regulatory and large as well as small genes of different functions . About half of the genes encode ribosomal proteins which are known to evolve slowly and to be relatively short. In our analyses, they yielded trees differing in their topologies from those produced by the combined analyses of the whole set somewhat more often than the other sequences, probably because of the relative paucity of phylogenetic information. The loci included in our data set were relatively conserved. The number of amino acid substitutions per position between tetrapod and amphioxus sequences was 0.4 on average for the different proteins; the largest number was 1.3 .
Monophyly of the Agnathans&, 百拇医药
shows the phylogenetic trees drawn for the concatenated sequences of 35 loci by the MP, NJ, and ML methods. The set consisted of a total of 9,098 shared amino acid positions (6,786 positions when cartilaginous fishes were included; ). All the methods applied to the concatenated sequences using different substitution models and distance measures and the TOTALML approach generated the same tree topology shown in , in which the agnathans form a monophyletic cluster. The monophyly of the agnathans was supported by all the methods used with a high statistical confidence. The bootstrap probabilities (BP) for the two interior branches were 99%–100% for different substitution models by the ML method including the TOTALML approach (not shown). In all the tree-drawing methods the second best tree was ((tetrapod, bony fish), lamprey), hagfish, amphioxus) and the best tree was significantly supported by the tests of tree comparison (for the MP tree P = 0.002 by Kishino-Hasegawa's and Templeton's nonparametric tests; for the ME trees P = 0.0001 using Poisson correction distance and P = 0.02 using the gamma distance; for the ML trees P = 0.0001 by the Kishino-Hasegawa test using all the substitution models examined; and for the TOTALML approach P < 0.0016 by the Kishino-Hasegawa test using all the substitution models.) In the MP method the tree length of the second best tree was larger than that of the best tree by 50. The gamma parameter, inversely related to the extent of rate variation across amino acid positions, estimated for the concatenated sequences by the ML method was 0.38.
The same tree topology as that generated from the concatenated sequences and the TOTALML approach was obtained for 19 genes (54%) by the NJ method and for 14 genes (45%) with the MP and ML methods (see Supplementary Material). By considering only whether or not the cluster of hagfish and lamprey appeared in the trees drawn, the monophyly of the agnathans was supported by the majority of the trees (57% to 62% depending on the methods) of the individual proteins. The alternative agnathan phylogeny in which lampreys are more closely related to gnathostomes than to hagfishes was supported by only 10% to 27% of the trees.lq&?ix, http://www.100md.com
Phylogenetic Relationship of Cartilaginous Fish and Bony Fish and the Impact of the Inclusion of the Former on Agnathan Monophylylq&?ix, http://www.100md.com
shows the phylogenetic trees drawn from concatenated sequences of 31 nuclear protein-encoding genes for which sequences of cartilaginous fishes are available. All the tree-drawing methods used for the concatenated sequences and the TOTALML approach produced the tree topology shown in, in which the agnathans form a monophyletic cluster and cartilaginous fishes are positioned outside the cluster of tetrapod and bony fish. With this data set the monophyly of the agnathans was also strongly supported. Bootstrap probabilities for the ancestral branches of hagfish and lamprey were 99%–100% by all the methods. The position of cartilaginous fishes outside the cluster of bony fishes and tetrapods is significantly supported by most of the methods applied, but with slightly lower confidence than the agnathan monophyly. Bootstrap probabilities for the ancestral branch of tetrapods and bony fishes were 97%–100% by the NJ method using all the distance measures, and by the ML methods using different substitution models including the TOTALML approach. The BP (94%) of the MP method for the clustering of tetrapods and bony fishes was slightly lower than the 5% significance level. The second best tree was (((tetrapod, bony fish), cartilaginous fish), lamprey), hagfish, amphioxus) by the MP method and by the TOTALML approach using the Poisson model with the F option regardless of whether the G option was specified; and (((tetrapod, (bony fish, cartilaginous fish)), (lamprey, hagfish), amphioxus) by the ME method and by the ML method except for the TOTALML approach mentioned earlier. Tests of tree comparisons were significant by the ME method using the Poisson-correction distance (P = 0.001), by the ML method (P = 0.02–0.004), and by the TOTALML approach (P = 0.017–0.0001), but not significant by the MP method (P = 0.12) and by the ME method using the gamma distance (P = 0.26). In the MP method, the difference in the tree lengths between the best and the second best trees was 18. The gamma parameter estimated for the concatenated sequences by the ML method was 0.38.
The same tree topology as that generated from the concatenated sequences and the TOTALML approach () was obtained for eight, six, and five genes by the NJ, MP, and ML methods, respectively. The hagfish-lamprey cluster was observed in the majority (50%–65%) of genes, as in the case of the dataset without cartilaginous fishes mentioned in the previous section. By contrast, the branching pattern ((tetrapod, bony fish), cartilaginous fish) shown in appeared in the trees of only about 30% of genes. The alternative branching patterns for the relationship of the cartilaginous and bony fish were observed in even lower frequencies: the pattern (tetrapod, (bony fish, cartilaginous fish)) was observed in about 20% of the genes by all the methods, and the pattern ( (tetrapod, cartilaginous fish), bony fish) appeared in 19% of the NJ trees and 6% of the MP and ML trees.0[v9, 百拇医药
Is the Agnathan Monophyly a Result of Long Branch Attraction?0[v9, 百拇医药
Because hagfishes and lampreys are represented by long branches on the phylogenetic trees ( and ), it could be argued that their monophyly is the result of a long branch attraction . Distortions of phylogenetic reconstructions by long branch attraction occur especially when grossly simplifying models of the evolutionary process and the parsimony method are used . To avoid this problem, we used other methods of phylogenetic reconstruction in addition to those based on the maximum parsimony principle, and in the analyses we took into account different substitution rates across sites, genes, and lineages. We therefore do not think that long branch attraction is the cause of the observed agnathan monophyly. It has been argued that potential artifactual attractions between branches can be avoided by including more taxa in the data set (e.g). Others have pointed out, however, that adding taxa to break long branches increases the probability of distorting phylogenetic relationships in certain cases and that a better way of dealing with this problem is to increase the number of amino acid positions in the analyses . This is the way we have chosen in the present study.
Conclusion|jv, 百拇医药
The phylogenetic analysis leads to two important conclusions. First, the hagfishes and lampreys form a monophyletic group which constitutes a sister group of the gnathostomes, with cephalochordates as an outgroup. And second, the cartilaginous fishes diverged from the common ancestor of bony fishes and tetrapods before the latter two lineages diverged from each other. The first conclusion is well supported statistically by all tree-drawing methods applied to the dataset of the 35 proteins. It is also supported by the majority (57% to 62%, depending on the method) of the trees of the individual proteins. The second conclusion is supported by a high statistical confidence except by the maximum parsimony method and the minimum evolution method with gamma distance . The lower statistical support by the latter methods is probably due to a smaller number of amino acid positions in the data set that included cartilaginous fishes and to the relative shortness of the time interval during which the cartilaginous fishes and the bony fishes separated from the tetrapod lineage. The placement of cartilaginous fishes outside the cluster formed by bony fishes and tetrapods is in agreement with the traditional interpretation of this part of vertebrate phylogeny and in contradiction to the recent claim by placing cartilaginous fishes inside the bony fish clade. This latter phylogeny is based on mtDNA sequences and supported by high statistical significance. As mentioned earlier, however, sequence data of mitochondrial protein-encoding genes are not suitable for deep phylogenetic reconstructions .
The origin and nature of the morphological characters previously claimed to support the close relationship of lampreys to jawed fishes will have to be re-examined in view of the strong support of the molecular data for the agnathan monophyly. Many of these characters may not be homologous . Some may have arisen independently by convergent evolution in the lamprey and gnathostome lineages, while others still may have been present in the agnathan ancestors and lost in the hagfish . A good example of the uncertainties surrounding the nature of some of these characters is the agnathan "tongue." Originally, the resemblance between the "rasping tongue" of the lampreys and the "laterally biting jaws" of the hagfishes was one of the reasons for grouping the two lineages into Cyclostomata. Later, the resemblance was proclaimed to be superficial and the structures were interpreted as having been acquired independently by the two lineages (e.g.). Later still, however, pointed out at least 11 synapomorphous anatomical features of the feeding apparatus were shared by lampreys and hagfishes, but not by gnathostomes. Now the molecular data clearly side with the view that the "tongue" of hagfishes and lampreys is indeed a homologous organ.
Acknowledgements3/]rl, 百拇医药
We thank Jane Kraushaar for editorial assistance and Ryszard Lorenz for technical assistance. We are grateful to Dr. Hirohisa Kishino for the advice on the Kishino-Hasegawa test for TOTALML approach, and Drs. J.-O. Strömberg and B. Bergström for providing us M. glutinosa tissues.3/]rl, 百拇医药
Literature Cited3/]rl, 百拇医药
Adachi, J., and M. Hasegawa. 1996. MOLPHY2.3b3: programs for molecular phylogenetics based on maximum likelihood. Institute of Statistical Mathematics, Tokyo.3/]rl, 百拇医药
Carroll, R. L. 1987. Vertebrate paleontology and evolution. W. H. Freeman, New York.3/]rl, 百拇医药
Delarbre C., H. Escriva, C. Gallut, V. Barriel, P. Kourilsky, P. Janvier, V. Laudet, and G. Gachelin. 200l. The complete nucleotide sequence of the mitochondrial DNA of the agnathan Lampetra fluviatilis: bearings on the phylogeny of cyclostomes. Mol. Biol. Evol. 17:519-529.3/]rl, 百拇医药
Delarbre, C., C. Gallut, V. Barriel, P. Janvier, and G. Gachelin. 2002. Complete mitochondrial DNA of the hagfish, Eptatretus burgeri: the comparative analysis of mitochondrial DNA sequences strongly supports the cyclostome monophyly. Mol. Phylogenet. Evol. 22:184-192.
Goodman, M., M. M. Miyamoto, and J. Czelusniak. 1987. Pattern and process in vertebrate phylogeny revealed by coevolution of molecules and morphologies. Pp. 141–176 in C. Patterson, ed. Molecules and morphology in evolution: conflict or compromise? Cambridge University Press, New York.[4e1#, 百拇医药
Graybeal, A. 1998. Is it better to add taxa or characters to a difficult phylogenetic problem? Syst. Biol. 47:9-17.[4e1#, 百拇医药
Gu, X., and W.-H. Li. 1996. Bias-corrected paralinear and LogDet distances and tests of molecular clocks and phylogenies under nonstationary nucleotide frequencies. Mol. Biol. Evol. 13:1375-1383.[4e1#, 百拇医药
Gürsoy, H. C., D. Koper, and B. J. Benecke. 2000. The vertebrate 7S K RNA separates hagfish (Myxine glutinosa) and lamprey (Lampetra fluviatilis). J. Mol. Evol. 50:456-464.[4e1#, 百拇医药
Hardisty, M. W. 1982. Lampreys and hagfishes: analysis of cyclostome relationships. Pp. 165–259 in M. W. Hardisty and I. C. Potter, eds. The biology of lampreys. Academic Press, London.[4e1#, 百拇医药
Hedges, S. B. 2001. Molecular evidence for the early history of living vertebrates. Pp. 119–134 in P. E. Ahlberg, ed. Major events in early vertebrate evolution: paleontology, phylogeny, genetics, and development. Taylor & Francis, London.
Janvier, P. 1981. The phylogeny of the Craniata, with particular reference to the significance of fossil "agnathans." J. Vert. Paleontol. 1:121-159.?ks!aq, 百拇医药
Janvier, P. 1996. The dawn of the vertebrates: characters versus common ascent in the rise of current vertebrate phylogenies. Palaeontology 39:259-287.?ks!aq, 百拇医药
Jefferies, R. P. S. 1986. The ancestry of the vertebrates. British Museum, London.?ks!aq, 百拇医药
Jones, D. T., W. R. Taylor, and J. M. Thornton. 1992. The rapid generation of mutation data matrices from protein sequences. Comput. Appl. Biosci. 8:275-282.?ks!aq, 百拇医药
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.?ks!aq, 百拇医药
Kuraku, S., D. Hoshiyama, K. Katoh, H. Suga, and T. Miyata. 1999. Monophyly of lampreys and hagfishes supported by nuclear DNA-coded genes. J. Mol. Evol. 49:729-735.?ks!aq, 百拇医药
Lanfranchi, G., A. Pallavicini, P. Laveder, and G. Valle. 1994. Ancestral hemoglobin switching in lampreys. Dev. Biol. 164:402-408.
Lipscomb, D. L., J. S. Farris, M. Källersjö, and A. Tehler. 1998. Support, ribosomal sequences, and the phylogeny of the eukaryotes. Cladistics 14:303-338.:^asj0, 百拇医药
Løvtrup, S. 1977. The phylogeny of vertebrata. Wiley, New York.:^asj0, 百拇医药
Mallatt, J., and J. Sullivan. 1998. 28S and 18S rDNA sequences support the monophyly of lampreys and hagfishes. Mol. Biol. Evol. 15:1706-1718.:^asj0, 百拇医药
Mallatt, J., J. Sullivan, and C. J. Winchell. 2001. The relationship of lampreys to hagfishes: a spectral analysis of ribosomal DNA sequences. Pp. 106–118 in P. E. Ahlberg, ed. Major events in early vertebrate evolution: paleontology, phylogeny, genetics and development. Taylor & Francis, London.:^asj0, 百拇医药
Nei, M., and S. Kumar. 2000. Molecular evolution and phylogenetics. Oxford University Press, New York.:^asj0, 百拇医药
Philippe, H., A. Chenuil, and A. Adoutte. 1994. Can the Cambrian explosion be inferred through molecular phylogeny? Development Suppl:15–25.:^asj0, 百拇医药
Poe, S., and D. L. Swofford. 1999. Taxon sampling revisited. Nature 398:299-300.
Rasmussen, A. S., and U. Arnason. 1999. Molecular studies suggest that cartilaginous fishes have a terminal position in the piscine tree. Proc. Natl. Acad. Sci. USA 96:2177-2182.$, 百拇医药
Rasmussen, A. S., A. Janke, and U. Arnason. 1998. The mitochondrial DNA molecule of the hagfish (Myxine glutinosa) and vertebrate phylogeny. J. Mol. Evol. 46:382-388.$, 百拇医药
Romer, A. S. 1966. Vertebrate paleontology. University of Chicago Press, Chicago.$, 百拇医药
Rosenberg, M. S., and S. Kumar. 2001. Incomplete sampling is not a problem for phylogenetic inference. Proc. Natl. Acad. Sci. USA 98:10751-10756.$, 百拇医药
Rzhetsky, A., and M. Nei. 1992. A simple method for estimating and testing minimum-evolution trees. Mol. Biol. Evol. 9:954-967.$, 百拇医药
Stock, D. W., and G. S. Whitt. 1992. Evidence from 18S ribosomal RNA sequences that lampreys and hagfishes form a natural group. Science 57:787-789.$, 百拇医药
Suga, H., D. Hoshiyama, S. Kuraku, K. Katoh, K. Kubokawa, and T. Miyata. 1999. Protein tyrosine kinase cDNAs from amphioxus, hagfish, and lamprey: isoform duplications around the divergence of cyclostomes and gnathostomes. J. Mol. Evol. 49:601-608.
Suzuki, M., K. Kubokawa, H. Nagasawa, and A. Urano. 1995. Sequence analysis of vasotocin cDNAs of the lamprey, Lampetra japonica, and the hagfish, Eptatretus burgeri: evolution of cyclostome vasotocin precursors. J. Mol. Endocrinol. 14:67-77.@;mm?3i, 百拇医药
Swofford, D. L. 2002. PAUP*: phylogenetic analysis using parsimony (*and other methods). Sinauer Associates, Sunderland, Mass.@;mm?3i, 百拇医药
Takezaki, N., and T. Gojobori. 1999. Correct and incorrect vertebrate phylogenies obtained by the entire mitochondrial DNA sequences. Mol. Biol. Evol. 16:590-601.@;mm?3i, 百拇医药
Templeton, A. 1983. Phylogenetic inference from restriction endonuclease cleavage site maps with particular reference to the evolution of humans and the apes. Evolution 37:221-244.@;mm?3i, 百拇医药
Thompson, J. D., D. G. Higgins, and T. J. Gibson. 1994. ClustalW: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22:4673-4680.@;mm?3i, 百拇医药
Turbeville, J. M., J. R. Schulz, and R. A. Raff. 1994. Deuterostome phylogeny and the sister group of the chordates: evidence from molecules and morphology. Mol. Biol. Evol. 11:648-655.@;mm?3i, 百拇医药
Yalden, D. W. 1985. Feeding mechanisms as evidence for cyclostome monophyly. Zool. J. Linn. Soc. 84:291-300.@;mm?3i, 百拇医药
Yang, Z. 1996. Phylogenetic analysis using parsimony and likelihood methods. J. Mol. Evol. 42:294-307.@;mm?3i, 百拇医药
Yang, Z. 1999. PAML: a program package for phylogenetic analysis by maximum likelihood. Version 3.0c. University College London, London.@;mm?3i, 百拇医药
Accepted for publication October 29, 2002.(Naoko Takezaki Felipe Figueroa Zofia Zaleska-Rutczynska and Jan Klein)