Evolution and Development of the Chordates: Collagen and
Pharyngeal Cartilage
Amanda L. Rychel,*! Shannon E. Smith,* Heather T. Shimamoto,* and Billie J. Swalla*!
*Center for Developmental Biology and Biology Department, University of Washington, Seattle; and
!Friday Harbor Laboratories, University of Washington, Friday Harbor
Chordates evolved a unique body plan within deuterostomes and are considered to share five morphological characters,
a muscular postanal tail, a notochord, a dorsal neural tube, an endostyle, and pharyngeal gill slits. The phylum Chordata
typically includes three subphyla, Cephalochordata, Vertebrata, and Tunicata, the last showing a chordate body plan only
as a larva. Hemichordates, in contrast, have pharyngeal gill slits, an endostyle, and a postanal tail but appear to lack a notochord and dorsal neural tube. Because hemichordates are the sister group of echinoderms, the morphological features
shared with the chordates must have been present in the deuterostome ancestor. No extant echinoderms share any of the
chordate features, so presumably they have lost these structures evolutionarily. We review the development of chordate
characters in hemichordates and present new data characterizing the pharyngeal gill slits and their cartilaginous gill bars.
We show that hemichordate gill bars contain collagen and proteoglycans but are acellular. Hemichordates and cephalochordates, or lancelets, show strong similarities in their gill bars, suggesting that an acellular cartilage may have preceded
cellular cartilage in deuterostomes. Our evidence suggests that the deuterostome ancestor was a benthic worm with gill slits
and acellular gill cartilages.
Introduction
The unique chordate body plan evolved within the
deuterostome animals sometime before the Cambrian
(Valentine, Jablonski, and Erwin 1999; Blair and Hedges
2005). Chordates traditionally include vertebrates, lancelets
(cephalochordates), and tunicates, but tunicates do not exhibit a chordate body plan as adults (Zeng and Swalla 2005)
(fig. 1). Hemichordates are sister group to echinoderms and
both phyla are an outgroup to the rest of the chordates
(Cameron, Garey, and Swalla 2000; Peterson 2004; Smith
et al. 2004; Blair and Hedges 2005; Zeng and Swalla 2005)
(fig. 1). Xenoturbella has been recently included in the deuterostomes as molecular evidence unites them with hemichordates and echinoderms, but their exact position
within the deuterostomes is not yet clear (Bourlat et al.
2003) (fig. 1). All five chordate characteristics (postanal
tail, dorsal nerve cord, notochord, endostyle, and pharyngeal gill slits) have at one time or another been suggested
to have homologous structures present in hemichordates,
but all these features are lacking in echinoderms and Xenoturbella, the closest relatives to hemichordates, suggesting
that they were lost during their evolution (Smith et al. 2004,
Zeng and Swalla 2005).
Recent evidence suggests that hemichordates lack a
notochord and a dorsal central nervous system (CNS), which
are found in chordates (Lowe et al. 2003; Gerhart, Lowe, and
Kirschner 2005). These two structures are linked in chordates as the notochord induces the CNS during development
(Von Straaten et al. 1988; Smith and Schoenwolf 1989;
Tanabe and Jessell 1996). The notochord is a stiff rod-like
structure present in chordates that provides support against
the hydrostatic skeleton for propelling the animal forward
when it swims (Kardong 2002). In hemichordates, an anterior diverticulum of the gut has been described as histologically similar to the chordate notochord, so has been called
Key words: evolution and development, chordate evolution,
tunicates, cephalochordates, hemichordates, cartilage, collagen.
E-mail: [email protected].
Mol. Biol. Evol. 23(3):541–549. 2006
doi:10.1093/molbev/msj055
Advance Access publication November 9, 2005
! The Author 2005. Published by Oxford University Press on behalf of
the Society for Molecular Biology and Evolution. All rights reserved.
For permissions, please e-mail: [email protected]
a stomochord (Willey 1899; Balser and Ruppert 1990).
Brachyury is a T-box transcription factor that specifies notochord and is expressed in tunicate, cephalochordate, and
vertebrate notochords (mesoderm) during development
(Wilkinson, Bhatt, and Herrmann 1990; Yasuo and Satoh
1993, 1994; Holland et al. 1995). However, it has been shown
that this gene is not expressed in the hemichordate stomochord, though it is expressed in other tissues (Peterson
et al. 1999; Gerhart, Lowe, and Kirschner 2005). In sea
urchins, brachyury is expressed in the secondary mesenchyme cells (mesoderm) (Harada, Yasuo, and Satoh
1995), and in Drosophila, it is expressed in the posterior
gut (endoderm) (Singer et al. 1996). Thus, developmental
gene expression data do not confirm the homology between
the notochord and the stomochord, but it does not completely
rule it out either (see Gerhart, Lowe, and Kirschner 2005).
The dorsal neural tube is a chordate characteristic at
one time thought to be present in hemichordates, albeit in
a more limited form, namely the collar nerve cord (Bateson
1886). Hemichordates have a diffuse nervous system composed of a middorsal nerve cord, a midventral nerve cord in
the collar and trunk, as well as an epidermal nerve net
throughout all body regions (Knight-Jones 1952). The dorsal nerve cord in the collar region has been hypothesized to
have homology with the chordate neural tube because it is
hollow and forms by rolling up into a tube during development (Morgan 1891). However, Lowe et al. (2003) recently
showed that vertebrate neural-patterning genes are expressed circumferentially in the hemichordate ectoderm
and, strikingly, that vertebrate forebrain genes are expressed
in the proboscis ectoderm, midbrain genes in the collar ectoderm, and hindbrain genes in the trunk ectoderm. These
results suggest that the deuterostome ancestor may have
had an epidermal nerve net rather than a CNS, yet the
anterior-posterior patterning of the ectoderm was already
present (Holland 2003).
In contrast, three of the chordate characteristics are
present in hemichordates, a ventral postanal tail, an endostyle, and pharyngeal gill slits. A ventral postanal tail is
present in juvenile harrimaniid hemichordate worms
(Bateson 1885; Burdon-Jones 1952; Cameron 2002), and
542 Rychel et al.
FIG. 1.—Current understanding of deuterostome phylogeny, with synapomorphies for each node indicated on the phylogeny. Echinoderms and
hemichordates are sister groups, while tunicate relationships are still not well resolved. The placement of Xenoturbella is still not exactly certain. Chordate
characters indicated on the phylogeny that unite tunicates with cephalochordates and vertebrates are present only in the tunicate tadpole larva. (Redrawn
from Swalla 2001 and Zeng and Swalla 2005)
it has been proposed to be homologous to the chordate
postanal tail due to posterior Hox gene expression data
(Lowe et al. 2003). Hemichordates have three posterior
Hox genes, Hox 11/13a, 11/13b, and 11/13c, which are also
found in echinoderms (Peterson 2004). Furthermore, they
share specific amino acid motifs with the echinoderms
(Peterson 2004), strongly supporting their sister-group relationship (Zeng and Swalla 2005). Both hemichordate
Hox 11/13 and the orthologous vertebrate Hox 10-13 are
expressed in the postanal tail (Lowe et al. 2003). It is interesting that a postanal tail has only been reported in
species of the direct-developing Harrimaniidae (Bateson
1885; Burdon-Jones 1952; Cameron 2002), not in the
Ptychoderidae (Urata and Yamaguchi 2004). These two
families are paraphyletic with 18S rDNA analyses, with
harrimaniids a sister group to the colonial pterobranchs
(Halanych 1995; Cameron, Garey, and Swalla 2000); however, theses two families are monophyletic when a 28S
rDNA (Winchell et al. 2002) or morphological (Cameron
2005) analysis is performed. A postanal tail not being present in the indirect-developing ptychoderid acorn worms
means that either it was lost in ptychoderids or the directdeveloping harrimaniid worms more closely resemble the
chordate ancestor.
The endostyle is an iodine-binding organ present in the
tunicate and cephalochordate pharynx and is considered by
many researchers to be a homologous organ to the vertebrate thyroid (Sasaki et al. 2003). The hemichordate epibranchial ridge, like the chordate endostyle, is composed
of specialized secretory cells (Ruppert, Cameron, and Frick
1999), and these cells bind iodine. However, iodine binding
is not restricted to this region but instead occurs all throughout the pharynx (Ruppert 2005). Likewise, the hemichordate homolog of the gene NK2.1, which in chordates is
expressed in the endostyle and the dorsal nervous system,
was seen to be expressed in hemichordate larval neuronal
structures as well as broadly in the pharyngeal endoderm,
stomochord, and hindgut (Takacs, Moy, and Peterson
2002). Based on these data, it appears that the endostyle
function in chordates is accomplished broadly by the
pharynx in hemichordates.
The most convincing homology between hemichordates and chordates is the pharyngeal gills. The transcription factor Pax1/9 is expressed in the pharyngeal endoderm
in hemichordates (Ogasawara et al. 1999; Lowe et al. 2003),
tunicates (Ogasawara et al. 1999), cephalochordates (N. D.
Holland and L. Z. Holland 1995), lamprey (Ogasawara
et al. 2000), and gnathostomes (Neubüser, Koseki, and
Balling 1995; Wallin et al. 1996; Peters et al. 1998) (fig.
2). The expression of the most anterior Hox gene, Hox1,
is first seen in vertebrates in the second pharyngeal slit
and the expression of Hox1 in hemichordates is seen at
the level between the first and second gill slits, suggesting
that the location of the gill slits along the anterior-posterior
axis is also homologous (Lowe et al. 2003). Additionally,
the pharyngeal skeletal elements of hemichordates and
cephalochordates are strikingly similar in appearance
(Hyman 1959; Schaeffer 1987). Both have been reported
to contain fibrillar collagen, as seen by transmission electron microscopy (TEM) studies (Rahr 1982; Pardos and
Benito 1988). Pharyngeal gills appear to function primarily
for feeding and possibly for oxygen absorption, and thus in
burrowing animals, cartilaginous elements may play an important structural role. All deuterostomes with gills, except
for tunicates, have structural elements associated with
pharyngeal gills. Tunicates lack any inner structural support
but instead have a cellulose-based tunic (Rånby 1952;
Matthysse et al. 2004). Phylogeny results have shown that
it is likely that the ancestral deuterostome was a worm that
had pharyngeal gills supported by a collagenous acellular
skeleton (Cameron, Garey, and Swalla 2000; Bourlat et al.
2003; Smith et al. 2004; Zeng and Swalla 2005).
Chordate Evolution 543
FIG. 2.—Expression of Pax1/9 in the deuterostomes. Pax1/9 is expressed in the developing and adult pharyngeal gill slits in all the gnathostomes,
lamprey (Lampetra), cephalochordates, tunicates, and hemichordates. Neither Xenoturbella nor Echinodermata has been reported to express this transcription factor. In lamprey and gnathostomes, Pax1/9 is also expressed in the nasal placode and facial ectomesenchyme, and in gnathostomes, it is
expressed in the somites as well.
The most abundant and studied members of the collagen family are the fibrillar collagens. Each fibrillar collagen
possesses the following characteristics: a long uninterrupted
triple helical collagenous domain of approximately 340
Gly-X-Y repeats, a conserved noncollagenous C-terminal
domain, and an N-terminal domain with either a von Willebrand C (VWC) or thrompospondin (TSPN) domain within it
(Kielty and Grant 2002). In this article, we construct gene
trees with invertebrate and vertebrate triple helical domain
sequences to show that hemichordates and lancelets have
at least one fibrillar collagen and that it is similar to type I
and type II vertebrate collagens.
We have been testing the hypothesis that hemichordates have a pharyngeal cartilage similar to vertebrates,
containing collagen and proteoglycans in the matrix. Here,
we show that the matrix contains at least one fibrillar collagen and probably contains proteoglycans as well, even if
it is not cellular. The extracellular matrix (ECM) that com-
prises the gill bar skeleton in both hemichordates and cephalochordates is probably secreted by the endoderm, unlike
the vertebrate pharyngeal cartilage which is neural crest derived. We hypothesize that the earliest gill cartilages were
acellular, and these were gradually replaced by neural crest
cells in vertebrates. Therefore, the deuterostome ancestor
would have been a benthic worm with gill cartilages,
a mouth, and the ability to filter feed. Cephalochordates
may have lost their mouth and evolved a velum, while vertebrates lost the ability to filter feed. We next show evidence
that invertebrate deuterostome cartilage is acellular, with an
ECM of collagen and proteoglycans.
Materials and Methods
Animals
Saccoglossus bromophenolosus were collected at Bay
View State Park in Padilla Bay, Washington, during low
544 Rychel et al.
tide. Worms were dug with shovels and transported back to
the laboratory in 15-ml falcon tubes filled with seawater.
Sites in which to dig for the worms were recognized by
small circular fecal casts that the animals deposit on the
surface of the mud. Branchiostoma floridae adults were
purchased from Gulf Specimen Marine Lab (Panacea,
Fla; http://www.gulfspecimen.org/).
Histology and Immunocytochemistry
Animals were fixed for 1 h in 4% paraformaldehyde in
phosphate-buffered saline. They were then dehydrated
through a 30%, 50%, 80%, and 100% ethanol series and embedded in polyester wax. Seven-micrometer sections were
made on a Spencer 829 microtome and mounted on gelatinsubbed slides. The wax was removed with 100% ethanol,
and slides with sections were stained with Milligan’s trichrome (Presnell and Schreibman 1997) or hydrated and
stained with antibodies against type II collagen (dilution
1:500) and DAPI (Sigma, St. Louis, Mo.). Photographs were
taken using Cool Snap Version 1.2.0 for Macintosh and
a Nikon E600 microscope. Type II collagen monoclonal
antibody (ascites II-II6b3) was purchased from the Developmental Studies Hybridoma Bank (University of Iowa,
Iowa City, Iowa; http://www.uiowa.edu/;dshbwww/
iiii6b3.html). This monoclonal antibody is specific for vertebrate type II collagen, has a broad species specificity, and
does not cross-react with vertebrate nonfibrillar collagens
(Linsenmayer and Hendrix 1980).
Collagen Phylogenetic analysis
All known Homo sapiens fibrillar collagen amino
acid sequences were downloaded from GenBank, and they
have the following accession numbers: NP_000084.2,
NP_000384.1, NP_056534.1, NP_000079.1, NP_000080.2,
CAA34488.1, AAA51891.1, AAC50214.1, AAL13167.1,
NP_116277.2, NP_690850.1. Invertebrate accession
numbers are as follows for all known fibrillar collagens in
GenBank—AAC35289.2: polychaete Alvinella pompejana;
NP_999674.1, NP_999675.1: sea urchin Strongylocentrotus purpuratus; CAE53096.1: sea urchin Paracentrotus lividus; BAA75669.1, BAA75668.1: abalone Haliotis discus;
and AAM77398.1: Hydra vulgaris. Tunicate fibrillar
collagen sequences were downloaded from the Ciona
genome Web site (http://genome.jgi-psf.org/ciona4/ciona4.
home.html) and have the following gene model
numbers: ci0100154301, ci0100150759, ci0100131606,
ci0100144916. The mosquito Anopheles gambiae and bee
Apis mellifera fibrillar collagen sequences were found via
Blast on the Ensembl genome Web site (http://www.
ensembl.org/Multi/blastview) and are identified by the
following Ensembl novel transcript numbers: ENSANGP00000021001, ENSAPMT00000018541, ENSAPMT00000025599. The lancelet B. floridae sequence
was found by Blasting the Max Planck Amphioxus gene
catalogue
(http://goblet.molgen.mpg.de/cgi-bin/Blastamphioxus.cgi) with a vertebrate C-terminal fibrillar collagen domain, and it has the identifier of cluster00172.2.
For all collagen sequences found via Blast, only those having
a triple helical sequence in the same gene or genomic vicinity
of a highly conserved noncollagenous domain (known to be
present in all fibrillar collagens) were used. The hemichordate Saccoglossus kowalevskii sequence is from an expressed sequence tag (EST) clone kindly provided by
John Gerhart and Marc Kirschner (Lowe at al. 2003), and
its accession number is DQ233249. Amino acid sequences
were trimmed to include only the major triple helix, which in
complete sequences varied from 657 to 1,017 amino acids
with the majority being in the 1,011–1,017 amino acid range.
Each sequence in the data set is represented by a complete
major helical domain except for the polychaete, hemichordate, and lancelet sequences. Because only the C-terminal
ends of these sequences were available, an unknown
sequence of the appropriate length was added to the Nterminus of these sequences to aid in proper alignment.
These amino acid sequences were then aligned using the
program ClustalX (Jeanmougin et al. 1998) and produced
an alignment of 1,060 total characters.
The data set was subjected to phylogenetic analysis
using PAUP* v4b10 (Swofford 1999) under the optimality
criterion parsimony. To assess levels of clade support in this
analysis, bootstrap support was calculated using 1,000
pseudoreplicates. Clades with values of 70 and greater were
considered well supported (Hillis and Bull 1993). The data
set was further analyzed using Bayesian methods (MrBayes
v3.0b4) (Huelsenbeck and Ronquist 2001; Ronquist and
Huelsenbeck 2003). Bayesian methods were used to compute tree topologies based on a mixed model of amino acid
substitution and allowing for invariant sites and rate heterogeneity among sites (gamma parameter). The analysis was
run for 1,000,000 generations, and the first 15,000 trees
were discarded as burn-in after the distribution was graphed
to ensure that only the trees on the plateau were included.
Clades with posterior probabilities 95 and greater were considered strongly supported because they are true probabilities of clades under the assumed models (Rannala and
Yang 1996).
Results and Discussion
Hemichordates are tricoelomic, with an anterior proboscis, a middle neck region, and a posterior abdomen
(fig. 3). The abdomen contains a number of gill slits, which
are shown in figure 3A and C. In the harrimaniids, the reported range of gill slits is between 30 and 150, although the
exact number will vary even within species (Smith et al.
2003). If the worm is dissected, it is clear that there is
an anterior cartilaginous proboscis skeleton and cartilaginous gill bars that support the gill pores (fig. 3B). These
gill bars have been shown to stain with alcian blue (Smith
et al. 2003), which stains sulfated proteoglycans. TEM
studies have shown that the matrix of these gill bars appears
to contain collagen fibrils (Pardos and Benito 1988), and we
show here that the matrix also stains with type II collagen
antibodies (fig. 4C and D).
Milligan trichrome histology on posterior gill bars has
revealed that pharyngeal endoderm cells adjacent to the gill
bars appear to be secreting the matrix that makes up the gill
bar matrix (fig. 4B). Hyman (1959) also described the
gill bars as a ‘‘thickening of the basement membrane of the
pharyngeal epithelium.’’ However, we cannot rule out that
another population of cells could have migrated to the
Chordate Evolution 545
FIG. 3.—(A) Diagram of a hemichordate, showing the tripartite body
plan. (B) Dissection of a Saccoglossus bromophenolosus showing the
wishbone-shaped proboscis skeleton of the neck region above (white arrow, top) and the hairpin-shaped cartilaginous gill bars below (white arrows, bottom). (C) Live view of the external gill pores of the branchial
region of the hemichordate S. bromophenolosus.
pharyngeal area during development and are responsible for
secreting the matrix. However, developmental cell migrations have not yet been described in hemichordates or cephalochordates at these stages. In no case have any nuclei
been seen within the matrix (fig. 4B–D), so it appears that
the gill bars form in an acellular manner, even though
Benito and Pardos (1997) show that a ‘‘wandering’’ nucleus
from connective tissue cell can be seen on occasion within
the gill bar matrix. The way that gill bars form could be
similar to some types of acellular vertebrate bone formation, such as the formation of acellular vertebrae in teleost
fish (Ekanayake and Hall 1987, 1988; Fleming, Keynes,
and Tannahill 2004). Milligan trichrome staining on lancelet gill bars shows similar qualities to that of the hemichordate gill bars. They are acellular and adjacent to
endodermally derived epithelial cells (fig. 4E and F). Also,
the outer layer of these gill bars is positive for type II
collagen when stained with the same monoclonal antibody
(II-II6b3) as the hemichordates (fig. 4G and H).
A number of studies have attempted to determine the
affinities between invertebrate fibrillar collagens and vertebrate fibrillar collagens. Some have used positive staining
with vertebrate monoclonal and polyclonal antibodies
(Corbetta, Bairati, and Vitellaro Zuccarello 2002; Rigo,
Hartmann, and Bairati 2002; Cole and Hall 2004a,
2004b) and others phylogenetic approaches (Sicot et al.
1997; Exposito et al. 2002; Aouacheria et al. 2004). The
study of Aouacheria et al. (2004) is the first to show any
resolution in the evolution of invertebrate and vertebrate
collagens. They accomplished this by using the major triple
helical collagen domain rather than the C-terminal domain
for their analysis. Their published trees appeared to have an
arbitrary root placement as well as a smaller taxonomic
sample, so it was not clear to what major clade of collagens
the invertebrate sequences belonged. In our analysis, both
parsimony and Bayesian trees show a well-supported division between the A clade and the B/C clades (fig. 5). This
division is furthermore well supported in both analyses with
a bootstrap value of 87 and posterior probability of 99. This
division is also supported by known N-terminal domains
for these two groups. All human Clade A collagen genes
except for Hsa col1a2 have a VWC domain at the N-terminus, while all human clade B and C collagen genes have
a TSPN domain at the N-terminus (Aouacheria et al. 2004;
fig. 5).
The majority of invertebrate sequences found to date
belong to clade A based on major triple helix sequence, including abalone H. discus, polychaete A. pompejana,
urchins P. lividus and S. purpuratus, lancelet B. floridae,
hemichordate S. kowalevskii, and some of the sequences
from bee A. mellifera and tunicate C. intestinalis. These
findings are additionally corroborated by the fact that Ciona759, Urchin2, and Urchin5 have N-terminal VWC
domains (Aouacheria et al. 2004). Also, Ciona301, Bee1,
and Mosquito1 sequences were found within the vertebrate
B clade, and like the vertebrate collagen genes of this clade,
these invertebrate collagen genes have N-terminal TSPN
domains (Aouacheria et al. 2004). Furthermore, the presence of insect and tunicate collagen genes among B/C vertebrate clade sequences points to the possibility that there
were at least two, if not three, ancestral collagen types in the
bilaterian ancestor. Both methods of tree reconstruction
agreed well in overall topology and level of support with
the exception of H. vulgaris. Parsimony places it with
a bootstrap value of 87 in clade B/C, while in the Bayesian
tree it is part of the A clade with a posterior probability of 99
(fig. 5). A comparison using the Templeton’s Wilcoxon
signed-rank test (Templeton 1983) of the placement of
the Hydra collagen also shows incongruence in its position
between the parsimony tree and the Bayesian tree. Comparing the parsimony consensus tree to the same topology,
with only the position of Hydra changed, results in a statistically significant difference between them (P , 0.0001).
Hydra’s tendency to group with Ciona606 and Ciona916
in the parsimony tree is potentially due to long-branch attraction, a problem known to mislead parsimony analyses
(Felsenstein 1978).
Collagen types I and II, which cluster in the A clade,
are the most abundant collagen types in vertebrates (fig. 5).
The hemichordate and lancelet collagen sequences in the A
clade were found due to their transcript abundance in EST
studies (Lowe et al. 2003; Panopoulou et al. 2003), so it is
likely that these sequences represent the major fibrillar collagens found in these animals. It is also likely that these are
the collagens identified in gill bars stained with chicken
type II collagen monoclonal antibodies. It still remains to
be seen whether or not more fibrillar collagens are present
in the hemichordates, echinoderms, and lancelets. This will
be answered in the near future as more of these organisms’
genomes are sequenced and more ESTs become available.
546 Rychel et al.
FIG. 4.—(A–D) Saccoglossus bromophenolosus and (E–H) Branchiostoma floridae. (A, E) Transverse section, with dorsal facing up, through the
trunk region of the hemichordate S. bromophenolosus (A) and through the pharynx of the lancelet B. floridae (E), stained with Milligan’s trichrome (nuclei
and muscle stain red, collagen stains blue or green). (B, F) Close up of gill bar cross section stains blue-green with Milligan’s trichrome. (C, G) Similar gill
bar cross section stained in orange with a collagen type II monoclonal antibody (C also shows purple DAPI-stained nuclei). (D, H) Control section with the
type II collagen antibody omitted (D also stained with DAPI). D, dorsal nerve cord; G, gill bars; N, notochord; M, trunk musculature; and O, ovaries. Scale
bars (A, E): 0.2 mm; (B–D), (F–H): 50 lm.
Cartilaginous ECM containing collagen is typically
thought of as a vertebrate-specific tissue type, although similar tissues have been described for many invertebrates
(Cole and Hall 2004a, 2004b). Hemichordates and lancelets
have pharyngeal cartilages that are structurally and chemically similar to vertebrate pharyngeal cartilages, so far as
they are composed of collagen and proteoglycans. Despite
the fact that different cell types secrete the cartilage matrix
in hemichordates and lancelets versus vertebrates, it will be
interesting to discover if similar gene pathways control the
secretion of these ECM proteins.
Cartilaginous ECM appears to be endodermally derived in both hemichordates and lancelets. An invertebrate
collagen similar to vertebrate type II collagen is an important
component of the cartilage of hemichordates and lancelets.
In addition, they secrete the matrix in such a way that the
Chordate Evolution 547
FIG. 5.—Fibrillar collagen gene relationships based on the major triple helix amino acid sequence. (A) Parsimony tree. A total of 1,000 bootstrap
replicates were performed, and their values .50 are reported on the nodes, and well-supported values are in bold. (B) Bayesian tree with branch lengths
found using mixed amino acid model with invariant sites and gamma parameter. Posterior probabilities are reported on the nodes, and well-supported
values are in bold. Bee1, Bee2: Apis mellifera; Mosquito1: Anopheles gambiae; polychaete: Alvinella pompejana; Abalone1, Abalone2: Haliotis discus;
lancelet: Branchiostoma floridae; hemichordate: Saccoglossus kowalevskii; Urchin1, Urchin2: Strongylocentrotus purpuratus; Urchin5: Paracentrotus
lividus; Ciona301, Ciona759, Ciona606, and Ciona916 are the Ciona intestinalis gene model numbers ci0100154301, ci0100150759, ci0100131606,
ci0100144916, respectively; Hsa: Homo sapiens; and Hydra: Hydra vulgaris.
matrix-secreting cells themselves remain adjacent to, but are
never inside, the matrix, which is different from the mode of
cartilage formation found in vertebrates where chondrocytes
are trapped within the matrix they secrete. Neural crest cartilages are found in the vertebrate pharyngeal gills and head,
while the rest of the vertebrate cartilages are mesodermally
derived from the sclerotome of somites. The types of proteoglycans present in hemichordate and lancelet gill bars
remains to be determined. Hemichordates and lancelets
share a similar benthic lifestyle, and it is possible that pharyngeal endodermal cartilage secretion is the ancestral mode
of making cartilage in deuterostomes. In vertebrates, neural
crest cells take on this role of pharyngeal cartilage secretion,
but the developmental timing, place, and composition of the
matrix may have stayed the same.
In summary, we have reviewed the chordate characteristics and have emphasized research on characteristics that
are shared by chordates and hemichordates. The homologous pharyngeal gills of hemichordates, cephalochordates,
and vertebrates are supported by collagenous cartilage skeletons. The similarity in formation and composition between
hemichordate and cephalochordate gill bar matrix is striking, and it is likely that endodermally secreted acellular cartilage is the basal chordate pharyngeal cartilage. Further
developmental work is critical to understanding the origin
of the hemichordate stomochord and the possible patterning
in the collar nerve chord.
Supplementary Material
The sequence alignments in fasta and interleaved formats are available at Molecular Biology and Evolution online (http://www.mbe.oxfordjournals.org/).
Acknowledgments
Much of this work was funded by the University of
Washington National Science Foundation ADVANCE program and the Department of Biology, and we gratefully acknowledge their support. S.E.S. was funded on a Howard
Hughes Undergraduate Research Fellowship in the Swalla
lab. A.L.R. was supported by a National Institutes of Health
Developmental Biology Training Grant 5T32 HDO7183
and by the American Museum of Natural History
Lerner-Grey Marine Research Fund. John Gerhart, Marc
Kirchner, and Chris Lowe are thanked for the S. kowalevskii
collagen clone. Sequencing was performed in the Biology
Comparative Genomics Center, funded by a generous
award from the Murdock Corporation.
Literature Cited
Aouacheria, A., C. Cluzel, C. Lethias, M. Gouy, R. Garrone, and
J.-Y. Exposito. 2004. Invertebrate data predict an early emergence of vertebrate fibrillar collagen clades and an anti-incest
model. J. Biol. Chem. 279:47711–47719.
Balser, E. J., and E. E. Ruppert. 1990. Structure, ultrastructure, and
function of the preoral heart-kidney in Saccoglossus kowalevskii (Hemichordata, Enteropneusta) including new data on the
stomochord. Acta Zool. 71:235–249.
Bateson, W. 1885. The later stages in the development of B.
kowalevskyi, with a suggestion as to the affinities of the Enteropneusta. Q. J. Microsc. Sci. 25:81–122.
———. 1886. The ancestry of the Chordata. Q. J. Microsc. Sci.
26:535–571.
Benito, J., and F. Pardos. 1997. Hemichordata. Pp. 15–101 in
F. W. Harrison and E. E. Ruppert, eds. Microscopic anatomy
of invertebrates, Vol. 15. Wiley-Liss, New York.
548 Rychel et al.
Blair, J. E., and S. B. Hedges, 2005. Molecular phylogeny and
divergence times of deuterostome animals. Mol. Biol. Evol.
22:2275–2284.
Bourlat, S. J., C. Nielsen, A. E. Lockyer, D. Timothy, J. Littlewood, and M. J. Telford. 2003. Xenoturbella is a deuterostome
that eats molluscs. Nature 424:925–928.
Burdon-Jones, C. 1952. Development and biology of larva of Saccoglossus horsti (Enteropneusta). Phil. Trans. R. Soc. Lond. B
236:553–590.
Cameron, C. B. 2002. The anatomy, life habits, and later development of a new species of enteropneust, Harrimania planktophilus (Hemichordata: Harrimaniidae) from Barkley Sound.
Biol. Bull. 202:182–191.
———. 2005. A phylogeny of the hemichordates based on morphological characters. Can. J. Zool. 83:196–215.
Cameron, C. B., J. R. Garey, and B. J. Swalla. 2000. Evolution of
the chordate body plan: new insights from phylogenetic analyses of deuterostome phyla. Proc. Natl. Acad. Sci. USA
97:4469–4474.
Cole, A. G., and B. K. Hall. 2004a. Cartilage is a metazoan tissue;
integrating data from nonvertebrate sources. Acta Zool. 85:
69–80.
———. 2004b. The nature and significance of invertebrate cartilages revisited: distribution and histology of cartilage and
cartilage-like tissues within the Metazoa. Zoology 107:
261–273.
Corbetta, S., A. Bairati, and L. Vitellaro Zuccarello. 2002. Immunohistochemical study of subepidermal connective of molluscan integument. Eur. J. Histochem. 46:259–272.
Ekanayake, S., and B. K. Hall. 1987. The development of acellularity of the vertebral bone of the Japanese medaka, Oryzias
latipes (Teleostei; Cyprinidontidae). J. Morphol. 193:253–261.
———. 1988. Ultrastructure of the osteogenesis of acellular vertebral bone in the Japanese medaka, Oryzias latipes (Teleostei,
Cyprinidontidae). Am. J. Anat. 182:241.
Exposito, J-Y., C. Cluzel, R. Garrone, and C. Lethias. 2002. Evolution of collagens. Anat. Rec. 268:302–316.
Felsenstein, J. 1978. Cases in which parsimony and compatibility
methods will be positively misleading. Syst. Zool. 27:401–410.
Fleming, A., R. Keynes, and D. Tannahill. 2004. A central role for
the notochord in vertebral patterning. Development 131:
873–880.
Gerhart, J., C. Lowe, and M. Kirschner. 2005. Hemichordates and
the origin of chordates. Curr. Opin. Genet. Dev. 15:461–467.
Halanych, K. M. 1995. The phylogenetic position of the pterobranch hemichordates based on 18S rDNA sequence data.
Mol. Phylogenet. Evol. 4:72–76.
Harada, Y., H. Yasuo, and N. Satoh. 1995. A sea urchin homologue of the chordate Brachyury(T) gene is expressed in the
secondary mesenchyme founder cells. Development 121:
2747–2754.
Hillis, D. M., and J. J. Bull. 1993. An empirical test of bootstrapping as a method for assessing confidence in phylogenetic analysis. Syst. Biol. 42: 182–192.
Holland, N. D. 2003. Early central nervous system evolution: an
era of skin brains? Nat. Rev. Neurosci. 4:1–11
Holland, N. D., and L. Z. Holland. 1995. An amphioxus Pax gene,
AmphiPax-1, expressed in embryonic endoderm, but not in mesoderm: implications for the evolution of class I paired box
genes. Mol. Mar. Biol. Biotechnol. 4:206–214.
Holland, P. W., B. Koschorz, L. Z. Holland, and B. G. Herrmann.
1995. Conservation of Brachyury (T) genes in amphioxus and
vertebrates: developmental and evolutionary implications.
Development 121:4283–4291.
Huelsenbeck, J. P., and F. Ronquist. 2001. MRBAYES: Bayesian
inference of phylogeny. Bioinformatics 17:754–755.
Hyman, L. H. 1959. Phylum Hemichordata. Pp. 72–207 in The
invertebrates: smaller coelomate groups. McGraw-Hill, New
York.
Jeanmougin, F., J. D. Thompson, M. Gouy, D. G. Higgins, and
T. J. Gibson. 1998. Multiple sequence alignment with Clustal
X. Trends Biochem. Sci. 23:403–405.
Kardong, K. V. 2002. Vertebrates: comparative anatomy, function, evolution. McGraw-Hill, Boston.
Kielty, C. M., and M. E. Grant. 2002 The collagen family: structure, assembly and organisation in the extracellular matrix. Pp.
159–221 in P. M. Royce and B. Steinmann, eds. Connective
tissue and its heritable disorders. Wiley-Liss, New York.
Knight-Jones, E. W. 1952. On the nervous system of Saccoglossus
cambrensis (Enteropneusta). Phil. Trans. R. Soc. Lond. B
236:315–354.
Linsenmayer, T. F., and M. J. C. Hendrix. 1980. Monoclonal antibodies to connective tissue macromolecules: type II collagen.
Biochem. Biophys. Res. Commun. 92:440–446.
Lowe, C. J., M. Wu, A. Salic, L. Evans, E. Lander, N. StangeThomann, C. E. Gruber, J. Gerhart, M. Kirschner. 2003. Anteroposterior patterning in hemichordates and the origins of the
chordate nervous system. Cell 113:853–865.
Matthysse, A. G., K. Deschet, M. Williams, M. Marry, A. R. White,
and W. C. Smith. 2004. A functional cellulose synthase from
ascidian epidermis. Proc. Natl. Acad. Sci. USA 101:986–991.
Morgan, T. H. 1891. The growth and metamorphosis of Tornaria.
J. Morphol. 5:407–458.
Neubüser, A., H. Koseki, and R. Balling. 1995. Characterization
and developmental expression of Pax9, a paired-box-containing
gene related to Pax1. Dev. Biol. 170:701–716.
Ogasawara, M., Y. Shigetani, S. Hirano, N. Satoh, and S.
Kuratani. 2000. Pax1/Pax9-Related genes in an agnathan
vertebrate, Lampetra japonica: expression pattern of LjPax9
implies sequential evolutionary events toward the gnathostome
body plan. Dev. Biol. 223:399–410.
Ogasawara, M., H. Wada, H. Peters, and N. Satoh. 1999. Developmental expression of Pax1/9 genes in urochordate and hemichordate gills: insight into function and evolution of the
pharyngeal epithelium. Development 126:2539–2550.
Panopoulou, G., S. Hennig, D. Groth, A. Krause, A. J. Poustka, R.
Herwig, M. Vingron, and H. Lehrach. 2003. New evidence for
genome-wide duplications at the origin of vertebrates using an
amphioxus gene set and completed animal genomes. Genome
Res. 13:1056–1066.
Pardos, F., and J. Benito. 1988. Blood vessels and related structures in the gill bars of Glossobalanus minutus (Enteropneusta). Acta Zool. 69:87–94.
Peters, H., A. Neubüser, K. Kratochwil, and R. Balling. 1998.
Pax9-deficient mice lack pharyngeal pouch derivatives and
teeth and exhibit craniofacial and limb abnormalities. Genes
Dev. 12:2735–2747.
Peterson, K. J. 2004. Isolation of Hox and Parahox genes in the
hemichordate Ptychodera flava and the evolution of deuterostome Hox genes. Mol. Phylogenet. Evol 31:1208–1215.
Peterson, K. J., R. A. Cameron, K. Tagawa, N. Satoh, and E. H.
Davidson. 1999. A comparative molecular approach to mesodermal patterning in basal deuterostomes: the expression
pattern of Brachyury in the enteropneust hemichordate Ptychodera flava. Development 126:85–95.
Presnell, J. K., and M. P. Schreibman. 1997. Humason’s animal tissue techniques. Johns Hopkins University Press, Baltimore, Md.
Rahr, H. 1982. Ultrastructure of gill bars of Branchiostoma
lanceolatum with special reference to gill skeleton and blood
vessels (Cephalochordata). Zoomorphology 99:167–180.
Rånby, B. G. 1952. Physico-chemical investigations on animal
cellulose (tunicin). Ark. Kemi 4:241–248.
Chordate Evolution 549
Rannala, B., and Z. Yang. 1996. Probability distribution of molecular evolutionary trees: a new method of phylogenetic inference. J. Mol. Evol. 43:304–311.
Rigo, C., D. J. Hartmann, and A. Bairati. 2002. Electrophoretic
and immunochemical study of collagens from Sepia officinalis
cartilage. Biochim. Biophys. Acta 1572:77–84.
Ronquist, F., and J. P. Huelsenbeck. 2003. MRBAYES 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19:1572–1574.
Ruppert, E. E. 2005. Key characters uniting hemichordates and
chordates: homologies or homoplasies? Can. J. Zool. 83:8–23.
Ruppert, E. E., C. B. Cameron, and J. E. Frick. 1999. Endostylelike features of the dorsal epibranchial ridge of an enteropneust
and the hypothesis of dorsal-ventral axis inversion in chordates. Invertebr. Biol. 118:202–212.
Sasaki, A., Miyamoto Y, Satou Y, Satoh N, and Ogasawara M.
2003. Novel endostyle-specific genes in the ascidian Ciona
intestinalis. Zool. Sci. 20:1025–1030.
Schaeffer, B. 1987. Deuterostome monophyly and phylogeny.
Evol. Biol. 21:179–235.
Sicot, F. X., J. Y. Exposito, M. Masselot, R. Garrone, J. Deutsch,
and F. Gaill. 1997. Cloning of an annelid fibrillar collagen gene
and phylogenetic analysis of vertebrate and invertebrate collagens in nucleic acids, protein synthesis and molecular genetics.
Eur. J. Biochem. 590:1–9.
Singer, J. B., R. Harbecke, T. Kusch, R. Reuter, and J. A. Lengyel.
1996. Drosophila brachyenteron regulates gene activity and
morphogenesis in the gut. Development 122:3707–3718.
Smith, A. B., K. J. Peterson, G. Wray, and D. T. J. Littlewood.
2004. From bilateral symmetry to pentaradiality: the phylogeny of hemichordates and echinoderms. Pp 365-383 in J.
Cracraft and M. J. Donoghue, eds. Assembling the Tree of
Life. Oxford University Press, New York.
Smith, J. L., and G. C. Schoenwolf. 1989. Notochordal induction
of cell wedging in the chick neural plate and its role in neural
tube formation. J. Exp. Zool. 250:49–62.
Smith, S. E., R. Douglas, K. B. D. Silva, and B. J. Swalla. 2003.
Morphological and molecular identification of Saccoglossus
species (Hemichordata: Harrimaniidae) in the Pacific Northwest. Can. J. Zool. 81:133–141.
Swalla, B. J. 2001. Phylogeny of the urochordates: implications
for chordate evolution. Pp. 219–224 in H. Sawada and C.
Lambert, eds. Biology of ascidians (Proceedings of the First
International Symposium on the Biology of Ascidians).
Springer Verlag, Tokyo, Japan.
Swofford, D. L. 1999. PAUP*: phylogenetic analysis using parsimony (*and other methods). Version 4. Sinaeur Associates,
Sunderland, Mass.
Takacs, C. M., V. N. Moy, and K. J. Peterson. 2002. Testing putative hemichordate homologues of the chordate dorsal nervous system and endostyle: expression of NK2.1 (TTF-1) in
the acorn worm Ptychodera flava (Hemichordata, Ptychoderidae). Evol. Dev. 4:405–417.
Tanabe, Y., and T. M. Jessell. 1996. Diversity and pattern in the
developing spinal cord. Science 274:1115–1123.
Templeton, A. R. 1983. Phylogenetic inference from restriction
endonuclease cleavage site maps with particular reference
to the evolution of humans and the apes. Evolution 37:
221–244.
Urata M., and M. Yamaguchi 2004. The development of the enteropneust hemichordate Balanoglossus misakiensis KUWANO.
Zool. Sci. 21:533–540.
Valentine, J. W., D. Jablonski, and D. H. Erwin. 1999. Fossils,
molecules and embryos: new perspectives on the Cambrian explosion. Development 126:851–859.
Von Straaten, H. W. M., J. W. M. Hekking, E. J. L. M.
Wiertz-Hoessels, F. Thors, and J. Drukker. 1988. Effect of
the notochord on the differentiation of a floor plate area in
the neural tube of the chick embryo. Anat. Embryol.
177:317–324.
Wallin, J., H. Eibel, A. Neubüser, J. Wilting, H. Koseki, and
R. Balling. 1996. Pax1 is expressed during development of
the thymus epithelium and is required for normal T-cell maturation. Development 122:23–30.
Wilkinson, D. G., S. Bhatt, and B. G. Herrmann. 1990. Expression
pattern of the mouse T gene and its role in mesoderm formation. Nature 343:657–659.
Willey, A. 1899. Remarks on some recent work on the Protochordata, with a condensed account of some fresh observations on
the Enteropneusta. Q. J. Microsc. Sci. 42:223–244.
Winchell, C. J., J. Sullivan, C. B. Cameron, B. J. Swalla, and J.
Mallatt. 2002. Evaluating hypotheses of deuterostome phylogeny and chordate evolution with new LSU and SSU ribosomal
DNA Data. Mol. Biol. Evol. 19:762–776.
Yasuo, H., and N. Satoh. 1993. Function of the vertebrate T gene.
Nature 364:582–583.
———. 1994. An ascidian homolog of the mouse Brachyury (T)
gene is expressed exclusively in notochord cells at the fate restricted stage. Dev. Growth Differ. 36:9–18.
Zeng, L., and B. J. Swalla. 2005. Molecular phylogeny of the protochordates: chordate evolution. Can. J. Zool. 83:24–33.
Laura Katz, Associate Editor
Accepted November 2, 2005