Identification of Chaetognaths as Protostomes Is Supported by the Analysis
of Their Mitochondrial Genome
Daniel Papillon,* Yvan Perez, Xavier Caubit,à and Yannick Le Parco*
*Centre d’Océanologie de Marseille, Marseille, France; Laboratoire de Biologie Animale (Plancton),
Université Aix-Marseille I, Marseille, France; and àInstitut de Biologie du Développement de Marseille,
Laboratoire de Génétique et Physiologie du Développement, Campus de Luminy, Marseille, France
Determining the phylogenetic position of enigmatic phyla such as Chaetognatha is a longstanding challenge for
biologists. Chaetognaths (or arrow worms) are small, bilaterally symmetrical metazoans. In the past decades, their
relationships within the metazoans have been strongly debated because of embryological and morphological features
shared with the two main branches of Bilateria: the deuterostomes and protostomes. Despite recent attempts based on
molecular data, the Chaetognatha affinities have not yet been convincingly defined. To answer this fundamental question,
we determined the complete mitochondrial DNA genome of Spadella cephaloptera. We report three unique features: it is
the smallest metazoan mitochondrial genome known and lacks both atp8 and atp6 and all tRNA genes. Furthermore
phylogenetic reconstructions show that Chaetognatha belongs to protostomes. This implies that some embryological
characters observed in chaetognaths, such as a gut with a mouth not arising from blastopore (deuterostomy) and
a mesoderm derived from archenteron (enterocoely), could be ancestral features (plesiomorphies) of bilaterians.
Introduction
Chaetognaths (commonly named arrow worms) are
transparent marine metazoans, ranging in size from 2 to
120 mm. They are mostly planktonic and display a very
simple body plan, divided into three regions separated by
two transverse septa: the head, the trunk, and the tail. In
addition to the nervous system and muscles, the main
internal organs observed are the gut and ovaries localized
in the trunk and the testes in the tail. In the past century,
Chaetognatha relationships within the metazoans have
not been determined on the basis of their embryological
and morphological features (Hyman 1959; Beklemishev
1969; Ghirardelli 1994). Indeed, chaetognaths display
some embryological characters considered typical of deuterostomes (a gut with a mouth not arising from blastopore
and a mesoderm derived from the archenteron by enterocoely), whereas their morphology recalls the organization of the protostomes (ventral nerve cord and chitinous
structures) (Nielsen 2001). Such a division of characters
strongly isolates the chaetognaths from the majority of
bilaterians (Hyman 1959) and allowed Beklemishev (1969)
to put the chaetognaths and the brachiopods in a separate
group, as he thought that both of them belonged neither
to protostomes nor to deuterostomes. Despite several proposed relatives (Ghirardelli 1968), it has been traditionally
espoused that chaetognaths are allied to deuterostomes on
the basis of their embryological features (for review see
Willmer [1990]). Comparison of their photoreceptive cells
with those of the chordates corroborated such a phylogenetic position (Eakin and Westfall 1964).
However, the affinities with the deuterostomes have
not been supported by most recent authors. Indeed, in the
following years, Casanova (1986) has revived the idea of
a relationship with mollusks, after the discovery of
a hermaphrodite gland in deep benthoplanktonic species
Key words: Chaetognatha, mitochondria, gene loss, evolution,
phylogeny, Spadella cephaloptera.
E-mail: [email protected].
Mol. Biol. Evol. 21(11):2122–2129. 2004
doi:10.1093/molbev/msh229
Advance Access publication August 11, 2004
belonging to the Archeterokrohnia genus. Finally, Nielsen
(2001) has regarded the phylum as a sister group of
the gnathostomulids and the rotifers on the basis of the
chitinous spines surrounding the mouth (hooks) and the
innervations of the muscles from a vestibular ganglion that
resemble the ganglion in rotifers and gnathostomulids.
More recently, the metazoan phylogeny surveys using
molecular data, sometimes combined with morphological
characters, suggested various affinities, mostly among
protostomes: within lophotrochozoans with intermediate
filament sequences (Erber et al. 1998) or with ribosomal
RNA 18S sequences (18S) within ecdysozoans (Halanych
1996; Littlewood et al. 1998; Zrzàvy et al. 1998), basal
ecdysozoans (Peterson and Eernisse 2001), between
deuterostomes and protostomes (Giribet et al. 2000), or
as an early offshoot of the bilaterian lineage (Telford and
Holland 1993; Wadah and Satoh 1994). However, the
sequences used in these works are extremely divergent,
and most of these authors admitted that the phylogenetic position of Chaetognatha remains dubious because
of long-branch attraction artifacts. Landmark studies of
18S in nematodes (Aguinaldo et al. 1997) and acoels
(Ruiz-Trillo et al. 1999) clearly excluded Chaetognatha sequences from the analyses, and this long-branch attraction
problem is well documented and reviewed in the study by
Mallatt and Winchell (2002). Finally, the Hox gene survey
we performed emphasized a basal position among Bilateria
(Papillon et al. 2003), whereas other nuclear markers we
isolated (myosin, elongation factor, and slow-evolving 18S
from several genera [unpublished data]) did not allow us to
propose an unambiguous position.
The mitochondrial DNA (mtDNA) genome is now
widely used to infer deep phylogenetic relationships
(Boore 1999). It is, then, of particular interest in
deciphering the Chaetognatha phylogenetic position, and
it gives us the opportunity to add phylogenetically useful
genes to a growing metazoan data base. Metazoan mtDNA
genome is usually considered a small extrachromosomal
circular molecule ranging in size from about 13 to 18 kb
and almost always containing the same 37 genes: two
genes coding for rRNAs, 22 coding for tRNAs, and
Molecular Biology and Evolution vol. 21 no. 11 Ó Society for Molecular Biology and Evolution 2004; all rights reserved.
Chaetognaths Are Protostomes 2123
13 coding for proteins. This gene content has been shown
to vary in several metazoans (Boore 1999), such as nematodes (lack of atp8), mollusks (lack of atp8 and presence
of one extra tRNA), or cnidarians (lost of nearly all tRNA
genes and gain of genes not found in other mtDNAs).
Phylogenetic analyses of the primary sequences of mitochondrial genes (Boore and Staton 2002; Stechmann and
Schlegel 1999) support the new animal phylogeny based
on 18S (Aguinaldo et al. 1997) and Hox genes (de Rosa
et al. 1999). Moreover, mitochondrial gene arrangement is
a powerful phylogenetic tool for several reasons. First, the
gene content is almost invariant in the metazoans and,
thus, provides a unique and universal pool of information.
Second, stable structural gene rearrangements are rare
events because functional genomes must be maintained.
Third, the great number of potential gene rearrangements
makes it very unlikely that different lineages would
independently adopt the same gene order or that any gene
would move back to a previous location (Boore and Brown
1998).
Here we report the complete mitochondrial genome
sequence of Spadella cephaloptera in which we identified
three major striking and unique features: (1) the lack of
both atp8 and atp6, (2) the absence of all tRNA genes, and
(3) the smallest size among metazoan mtDNA genomes.
gene and a 10kb region, possibly spanning the almost
complete mitochondrial genome. This last fragment was
used as a template for the following PCR experiments. We
amplified the cytochrome oxydase 1 (cox1) gene with
universal degenerate primers COI5A2 (59-TAATWGGTGGNTTYGGNA-39) and COI3A (59-TCAGGRTGNCCRAARAAYCA-39). This sequence and the rrnL-nad6-cob
fragment were again used as anchor genes to isolate missing regions. Two 5-kb bands were obtained. Both fragments were cut with EcoR1 and resulting subfragments
were cloned into pBC SK1 (Promega) and sequenced by
‘‘primer-walk.’’
Materials and Methods
Specimens and DNA Extraction
We chose a broad representation of taxa from
available complete mtDNA sequences for this study (table
S1 in Supplementary Material online). Inferred amino acid
sequences of 11 mitochondrial protein genes (cox1 to
cox3, cob, nad1 to nad6, and nad4L) from 28 bilaterian
ingroup and two outgroup taxa were concatenated and
aligned using ClustalW in MacVector 7.1, and alignments
were refined by eye. Phyla with long branches were constantly excluded from the analyses (notably the nematodes and the platyhelminths). Among ecdysozoans,
except arthropods, only nematode mtDNA genomes are
fully sequenced, but these sequences are problematic for
phylogenetic reconstructions because of dramatically accelerated substitution rates (data not shown), even with the
mtDNA genome of Trichinella, considered as the less
derived of the nematodes phylum (Lavrov and Brown
2001). Thus, arthropods are the only ecdysozoan representatives. To avoid subjectivity in excluding unreliably
aligned positions from phylogenetic analysis, the program
Gblocks version 0.91b (Castresana 2000) was used with
the least-stringent settings.
Specimens of a benthic chaetognath species, Spadella
cephaloptera Busch 1851, were collected in the bay of La
Ciotat near Marseilles (South of France) on the Posidonia
oceanica bed, at a depth of 3 to 5 m. In the laboratory,
samples were kept in large aquaria containing natural
seawater and placed in a constant temperature at 218C 6
18C. DNA was prepared from adult specimens starved for
2 days in filtered seawater to prevent contamination by
ingested preys, mainly represented by crustaceans and
even small chaetognaths. Then, the selected specimens
were dried on pieces of filter paper; homogenized in lysis
buffer containing 100 mM Tris-HCl, 100 mM EDTA, and
0.1% SDS; and incubated with proteinase K at 50 mg/ml
for 4 h at 508C. After successive extractions with phenol,
phenol/chloroform, and chloroform, DNA was ethanolprecipitated in presence of 0.3 M NaAc.
Isolation of the Mitochondrial Genome
The ribosomal RNA 16S (rrnL) gene was the first
mitochondrial sequence isolated in S. cephaloptera, using
the following primers: 16SF (59-CCTGTTTANCAAAAACAT-39) and 16SR (59-GGTCCAACCAAAGATAGA39). We isolated the cytochrome b gene (cob) while
randomly sequencing a cDNA library of S. cephaloptera
(see Papillon et al. [2003]). Using rrnL and cob as anchor
genes, we tried to amplify regions of mitochondrial
genomes lying between these two genes with the
Advantage 2 PCR Kit (CLONTECH) dedicated to longrange PCR. All combinations of genome organization
were tested, and we were able to amplify a 2-kb region
containing the NADH dehydrogenase subunit 6 (nad6)
Genome Assembling
After sequencing, mtDNA genome was assembled
using MacVector version 7.1 program (Oxford Molecular
Group). Protein-encoding genes and start codons were
identified by Blast matching to other animal mtDNAs.
Transfer RNA searching was done by hand and using the
tRNA scan-SE Search Server (www.genetics.wustl.edu/
eddy/tRNAscan-SE). Sequence of the mtDNA genome of
S. cephaloptera has been submitted to GenBank under the
accession number AY545549.
Alignments
Phylogenetic Reconstruction
Sequence alignments were analyzed by NeighborJoining (NJ) (Gamma model of distances and sites pairwise
deletion) and maximum-parsimony (MP) with MEGA
version 2.1 (Kumar, Tamura, and Nei 1994). Confidence
estimates included bootstrap analysis with 1,000 replicates.
The full alignment comprised 2,511 amino acid positions,
of which 1,671 were parsimony-informative. Maximumlikelihood (ML) analysis employed 10,000 quartet-puzzling steps, an mtREV24 model of substitution, and eight
Gamma rate categories in Tree-Puzzle version 5.0 (Schmidt
et al. 2002).
2124 Papillon et al.
In another analysis, two set of trees were compared so
that support for specific phylogenetic hypotheses could be
assessed. We produced an exhaustive search of ML trees
in ProtML (Molphy 2.3b3 [Jun Adachi and Masami
Hasegawa 1992–1996]) using the mtREV24 model with
various constraints. First, we tested hypotheses of Chaetognatha being a deuterostome, a protostome, a basal bilaterian, or between protostomes and deuterostomes. Then,
four positions within protostomes were also assessed:
chaetognaths included in the lophotrochozoans, in the
ecdysozoans, as a third protostomian branch, or as a basal
protostome. Two different approaches were applied in tree
selection to obtain two sets of trees. First, we built consensus trees (PHYLIP version 3.5c [Felsenstein 1993]) for
each phylogenetic hypothesis from the entire set of generated trees or from the first 1,000 top-ranking trees when
more trees were generated. Second, we retained the best
trees for each constraint using the Approximately Unbiased (AU), Kishino-Hasegawa (KH), and ShimodairaHasegawa (SH) tests (Shimodaira 2002, and references
therein), as implemented in the CONSEL program
(Shimodaira and Hasegawa 2001). Then, to chose among
these phylogenetic hypotheses, the selected trees of each
set (‘‘consensus’’ or ‘‘best tree strategies’’) were compared
using CONSEL, as described above to test whether the
difference between the log-likelihood scores (LnL) of the
ML trees obtained was statistically significant. The ‘‘besttree’’ strategy was also conducted with a second substitution model, mtREV24-F, to test its influence on the
results.
For phylogenetic analysis of gene arrangements, two
different methods were used to construct pairwise distance
matrices. The first is based on minimum-breakpoint
analysis (Blanchette, Kunisawa, and Sankoff 1999). The
program is given a set of gene orders, and it finds the tree
and the ancestral gene orders that relate them. The
optimization criterion used is the minimal number of
breakpoints. The second method is used in the GRIMM
program (Tesler 2002), which computes the minimum
possible number of rearrangement steps and determines
a possible scenario taking this number of steps. This
method takes only into account gene inversions. Pairwise
distance matrices obtained with these two methods (figure
S2 in Supplementary Material online) were used to
construct phylogenetic trees with the NJ algorithm in
MEGA version 2.1.
Results and Discussion
The mtDNA genome of S. cephaloptera (GenBank
accession number AY545549) is very unusual both in
size and in composition (fig. 1). With 11,905 bp, it is
the smallest known metazoan mtDNA genome and
contains only 13 of the 37 genes usually found. Genome
composition, usage codon and initiation codons (9 ATG
and 2 ATA) are described in the table S3 of Supplementary Material online.
The most striking feature is the failure in detecting
any region homologous to one of the 22 bilaterian mitochondrial tRNAs. Among the usual 13 protein-encoding
genes, ATP synthetase subunit 6 (atp6) and atp8 are
FIG. 1.—The complete mtDNA gene map of the chaetognath
Spadella cephaloptera. Genesare abbreviated as in the text. Genes coding
on different strands are indicated by arrow direction. Numbers represent,
respectively, in base pairs (bp), the gene lengths and the size of the
noncoding regions separating each gene. Asterisks (*) indicate that length
of the ribosomal RNA genes rrnL and rrnS could not be precisely
determined.
lacking. The remaining 11 protein-encoding genes (cox1 to
cox3, cob, nad1 to nad6, and nad4L) use the classical
invertebrate mitochondrial genetic code. Six genes are
transcribed from one strand (nad4, 5, 6, cob, rrnL, and
cox1), and seven are transcribed from the other strand
(nad1, 2, and 3; 4L; cox2 and 3, and rrnS). No intron is
found, as expected, given that in metazoans, introns have
only been characterized in cnidarians (Boore 1999).
Intergenic spaces are not longer than 100 bp and have
a total length of only 265 bp. In these noncoding regions,
the signaling elements found in some animals (Boore
1999) cannot be identified. Although the genome is very
compact, a single overlap (2 nucleotides long) between the
neighboring genes nad4L and cox1 is observed. It has been
proposed that differences in size of mtDNA are mostly
caused by marked variations in the length and organization
of intergenic regions (Burger, Gray, and Lang 2003). This
is corroborated by the comparison of the size of Spadella
cephaloptera mtDNA genes with those of other bilaterians
(table S2 in Supplementary Material online). The length of
the genes in Spadella cephaloptera are sometimes a bit
smaller than those of mouse or fly but can be longer than
those of the smallest metazoan mtDNA genome known
until now (Taenia crassiceps, 13,503 bp), all of these
organisms having a longer mtDNA genome. This shows
that the small length of the Chaetognatha mtDNA genome
is, then, mostly caused by the lack of genes (tRNAs, atp6,
and atp8) and compact noncoding regions.
The loss of atp8 can be observed not only in
S. cephaloptera but also in a few other species of mollusks,
nematodes, and platyhelminths. In these last organisms, it
is not clear whether atp8 (1) moved to the nucleus, (2)
Chaetognaths Are Protostomes 2125
FIG. 2.—Position of Spadella cephaloptera within the bilaterians as suggested by maximum-likelihood (ML) phylogenetic analyses based on
comparisons of primary mitochondrial sequences. The tree was rooted with the cnidarian sequences. Numbers at various nodes indicate, respectively,
from left to right, percentage of quartet-puzzling replicates for ML and bootstrap support values for Neighbor-Joining (NJ) and maximum parsimony
(MP). Values are indicated only when percentage of quartet-puzzling replicates are in excess of 75% in the ML tree. The phylogenetic relationships for
which the methods give differing results are discussed in the text. Minus signs (–) indicate that a node is absent in the corresponding method; asterisks
(*) indicate that in the NJ and MP trees, S. cephaloptera is within protostomes but excluded from lophotrochozoans and ecdysozoans (figure S1a and
b in Supplementary Material online).
became dispensable entirely, or (3) had its function
coopted by one of the other ATPase subunits (Boore
1999). Unlike the atp8 loss, the absence of atp6 in
S. cephaloptera is unique among metazoan mtDNAs.
Chaetognaths are the first metazoans that could
possess an mtDNA genome with a complete absence of
tRNA genes. Total lack of tRNA genes has previously
been reported only for some protozoans (Plasmodium),
and some tRNA genes are lacking in green alga
(Pedinomonas) and angiosperm plants (Arabidopsis)
(Burger, Gray, and Lang 2003). The metazoans closest
to this situation are the cnidarians, for instance, Metridium
senile (Beagley, Okimoto, and Wolstenholme 1998) and
Acropora tenuis (Van Oppen et al. 2002), where only
Tryptophan and Methionine tRNAs are observed and the
other necessary tRNAs are imported nuclear products
(Boore 1999). As in cnidarians, paucity of mtRNA genes
in S. cephaloptera is probably a derived condition rather
than a conserved primitive state for multicellular animals.
Finally, given the very peculiar features of the
S. cephaloptera mtDNA genome, the existence of other
mtDNA molecules containing the lacking genes cannot be
excluded (linear concatemers or supplementary circular
mtDNA), as it has been reported in few unrelated
organisms (for review see Burger, Gray, and Lang [2003]).
Despite this very unusual organization, mitochondrial
data analyses based on deduced proteins sequences
allowed us to investigate the phylogenetic position of
chaetognaths. The search strategies for inferring phylogenetic trees included Neighbor-Joining (NJ), maximum
parsimony (MP), and maximum likelihood (ML). In each
analysis, deuterostome and protostome monophylies are
highly supported (fig. 2). Discrepancies among these
analyses are in the deuterostome relationships and the
monophyly of lophotrochozoans. First, B. lanceolatum,
P. marinus, and E. burgeri change phylogenetic positions,
depending on the methods used. Second, lophotrochozoans monophyly is supported in the NJ and ML trees
(support values 99% and 88%, respectively) but not in MP
analyses (39%). Nevertheless, the three methods yielded
unambiguously inclusion of Chaetognatha within the
protostomes (100% branch-support values in each tree)
either among (in ML tree [fig. 2] or outside the
lophotrochozoans (in MP and NJ tree [figure S1a and
b in Supplementary Material online]). Tree topology
presented in figure 2 also indicates that chaetognaths are
distant from arthropods. This hypothesis is supported by
a recent study in which a specific tissue marker showed
that chaetognaths are excluded from ecdysozoans (Haase
et al. 2001).
2126 Papillon et al.
To test this positioning, we compared four alternative
topologies using statistical tests: Chaetognatha within
protostomes, deuterostomes, as a third bilaterian branch,
or basal bilaterian. In both strategies used (‘‘best tree’’ and
‘‘consensus’’), the three latter phylogenetic hypotheses
were strongly rejected in favor of the proposed belonging
of Chaetognatha to protostomes (table 1). Another set of
tests was performed to precisely determine the positioning
of chaetognaths within protostomes. They showed that
relationships of Chaetognatha in the protostomian group
cannot be accurately determined. Indeed, although the
hypothesis of the Chaetognatha affinities with lophotrochozoans gives the best tree in both strategies, the
positions as basal protostome and within ecdysozoans
cannot be rejected in the ‘‘best-tree strategy’’ (only the
hypothesis as a basal protostome cannot be excluded in the
‘‘consensus-tree strategy’’) (table 1). Similar results have
been obtained in the analysis based on the other
substitution model (data not shown).
In addition, although the Chaetognatha branch is
longer than the average bilaterian branch length, longbranch attraction artifacts can be rejected for two reasons:
(1) the cnidarian long branch does not attract the
Chaetognatha branch outside the bilaterians, and (2) bootstrap supports for the belonging of Chaetognatha to the
protostomes are very strong in all phylogenetic analyses.
Finally, it is noteworthy that alignments of S. cephaloptera
mitochondrial proteins with those of various metazoans
reveal the presence of residues characteristic to protostomes in the Chaetognatha nad5 sequence (fig. 3).
As an independent source of phylogenetic information, S. cephaloptera mitochondrial gene arrangement was
compared with those of various representative metazoans.
We built pairwise distance matrices based on two methods
of genome comparison, GRIMM and minimum breakpoint, and constructed phylograms from these matrices. In
both analyses, the three main branches of bilaterians are
Table 1
Tests of Significance for Several Competing Phylogenetic
Hypotheses Within Metazoans and Within Protostomes
Hypothetical Affinities
Within Metazoans
Consensus tree strategy
Protostome
Basal bilaterian
Deuterostome
Third bilaterian branch
Best tree strategy
Protostome
Basal bilaterian
Deuterostome
Third bilaterian branch
Within Protostomes
Consensus-tree strategy
Lophotrochozoan
Basal protostome
Third protostomian
branch
Ecdysozoan
Best-tree strategy
Lophotrochozoan
Basal protostome
Ecdysozoan
Third protostomian
branch
LnL
AU
KH
SH
286647.3
286769.4
286790.9
286843.8
0.994
0.009
6e-5
9e-6
0.994
0.006
0.001
1e-5
0.999
0.007
0.001
1e-5
Best
286587.9
286733.2
286759.9
286703.2
0.098
0.002
2e-34
1e-40
0.996
0.004
0
0
0.999
0.004
0
0
Best
286609.0 0.662 0.629 0.817
286619.7 0.373 0.371 0.586
Best
286668.3 0.001 0.019 0.03
286687.4 0.006 0.009 0.017
286587.9 0.708 0.694 0.803
286601.8 0.352 0.306 0.431
286619.5 0.057 0.113 0.122
286665.3 3e-4
Best
0.001 0.001
NOTE.—LnL ¼ Log-likelihood scores; AU ¼ Approximately Unbiased test;
KH ¼ Kishino-Hasegawa test; SH ¼ Shimodaira-Hasegawa test. Boldface type
indicates hypotheses that are above the significance level. Significance level is 5%.
recovered (fig. 4a and b), and S. cephaloptera clusters
within the lophotrochozoan protostomes, supporting the
results obtained from primary sequences.
In addition, we could identify in the S. cephaloptera
mtDNA genome several gene boundaries that are likely to
be ancestral bilaterian features, because they can be
observed both in triploblasts and diploblasts. The nad1nad3-nad2 block present in chaetognaths is shared by the
FIG. 3.—Alignment of previously identified nad5 proteins of various metazoans with Spadella cephaloptera nad5 sequence. S. cephaloptera
displays several protostome signature residues.
Chaetognaths Are Protostomes 2127
FIG. 4.—Unrooted phylogram of Neighbor-Joining analysis obtained from metazoan mtDNA genome arrangements. Nine taxa were chosen among
the three main branches of bilaterians. The pairwise distance matrices were constructed using the GRIMM algorithm (a) and the minimum-breakpoint
method (b). Both of them were implemented in MEGA to reconstruct a phylogenetic tree. The two encoding protein genes, atp6 and atp8, and all
tRNAs were not considered in the GRIMM analysis, whereas only the tRNAs were excluded in the minimum-breakpoint method.
annelid/pogonophore and cnidarian clades. In addition, the
nad1-nad3 association is also present in another lophotrochozoan group, the platyhelminths. Another proposed
plesiomorphy, the nad3-nad2 block, is conserved in
several taxa, including annelids/pogonophores, mollusks,
brachiopods, and cnidarians. Neither of these blocks has
been reported in ecdysozoans or deuterostomes. The nad6cob couple can be found in representative taxa of
lophotrochozoans, ecdysozoans, cnidarians, but not in
deuterostomes. These analyses have several implications
with respect to the Chaetognatha phylogenetic position.
Indeed, there is no gene boundary in common between
S. cephaloptera and deuterostomes. Thus, in terms of gene
pair arrangements, there is no derived and specific feature
(synapomorphy) shared between Chaetognatha and deuterostomes. However, given that all common gene
boundaries observed in Chaetognatha, protostomes and
cnidarians are probable plesiomorphies, the affinities of
arrow worms within protostomes are still unclear on the
basis of mtDNA gene organization. Therefore, all
mitochondrial data (primary sequences, gene order, and
boundaries) support the identification of chaetognaths as
protostomes. However, none of these data allows a more
precise positioning within the protostomian group.
It should be noticed that the inclusion of Chaetognatha within protostomes, based on mitochondrial data, is
conflicting with previous conclusions, based on Hox genes
(Papillon et al. 2003). This could imply that the proposed
basal bilaterian feature conserved (a mosaic Hox gene that
has retained characteristics of both central and posterior
classes of Hox genes) could rather be a derived feature of
the phylum.
The way to divide the bilaterians is traditionally based
on two characters: the origin of the coelom and the fate of
the blastopore. In the new phylogeny (Adoutte et al. 2000),
several phyla initially described as deuterostomes have
been placed into protostomes. This is, for instance, the
case for brachiopods (de Rosa 2001; Cohen 1998;
Stechmann and Schlegel 1999) and phoronids (Helfenbein
and Boore 2004), even though in these species, the
2128 Papillon et al.
coelomic cavities form by enterocoely and the mouth does
not arise from the blastopore, like in deuterostomes.
Valentine (1997), supported by Peterson and Eernisse
(2001), previously advocated that traditional characters
placing the lophophorates into the deuterostomes are
plesiomorphies of bilaterians. Chaetognaths exhibit such
plesiomorphies: a complete gut with a mouth not arising
from the blastopore and coelomic cavities forming by
enterocoely. Moreover, chaetognaths do not display any of
the typical ecdysozoan and lophotrochozoan synapomorphies (possession of a molting habit and presence of
lophophores or trochophore larvae, respectively). As
Chaetognatha clearly exhibit plesiomorphies and lack
synapomorphies, they cannot be accurately placed among
the current protostomes, but this could bring them closer to
the common ancestor, excluded from lophotrochozoan and
ecdysozoan groups. Another embryological character
traditionally used to link Chaetognatha with deuterostomes
is the radial cleavage of the egg. However, recent cell fate
analysis indicates that the four-cell embryo displays
similarities with classic spiralians (Shimotori and Goto
2001), supporting the conclusion of the work presented
here. Nevertheless, more precise phylogenetic investigations on arrow worms will need further sequencing efforts,
for example, EST programs, to isolate new nuclear
molecular markers.
The acceptance that Chaetognatha are genealogically
allied to undoubted protostomes strengthens the phylogenetic validity of some morphological characters such as the
presence of a ventral nerve cord and chitinous structures
but stresses that ontological criteria (i.e., embryological
location of mouth and anus and mode of coelom
formation) that traditionally define the deuterostome and
protostome lineages can be misleading and raises once
more the question of their relevance to define a bilaterian
phylogeny.
Acknowledgments
We wish to thank Céline Brochier and André Gilles
for phylogenetic advices. Also, we are grateful to Patrick
Lemaire, Stephen Kerridge, and Laurent Fasano for
helpful discussions and corrections.
Literature Cited
Adachi, J., and M. Hasegawa. 1996. Molphy version 2.3b:
programs for molecular phylogenetics based on maximum
likelihood. Comput. Sci. Monogr. 28:1–150.
Adoutte, A., G. Balavoine, N. Lartillot, O. Lespinet, B.
Prud’homme, and R. de Rosa. 2000. The new animal
phylogeny: reliability and implications. Proc. Natl. Acad.
Sci. USA 97:4453–4456.
Aguinaldo, A. M., J. M. Turbeville, L. S. Linford, M. C. Rivera,
J. R. Garey, R. A. Raff, and J. A. Lake. 1997. Evidence for
a clade of nematodes, arthropods and other moulting animals.
Nature 387:489–493.
Beagley, C. T., R. Okimoto, and D. R. Wolstenholme. 1998. The
mitochondrial genome of the sea anemone Metridium senile
(Cnidaria): introns, a paucity of tRNA genes, and a nearstandard genetic code. Genetics 148:1091–1108.
Beklemishev, W. N. 1969. Principles of comparative anatomy of
invertebrates. Oliver and Boyd, Edinburgh.
Blanchette, M., T. Kunisawa, and D. Sankoff. 1999. Gene order
breakpoint evidence in animal mitochondrial phylogeny.
J. Mol. Evol. 49:193–203.
Boore, J. L. 1999. Animal mitochondrial genomes. Nucleic Acids
Res. 27:1767–1780.
Boore, J. L., and W. M. Brown. 1998. Big trees from little
genomes: mitochondrial gene order as a phylogenetic tool.
Curr. Opin. Genet. Dev. 8:668–674.
Boore, J. L., and J. L. Staton. 2002. The mitochondrial genome of the
Sipunculid Phascolopsis gouldii supports its association with
Annelida rather than Mollusca. Mol. Biol. Evol. 19:127–137.
Burger, G., M. W. Gray, and B. F. Lang. 2003. Mitochondrial
genomes: anything goes. Trends Genet. 19:709–716.
Casanova, J. P. 1986. Archeterokrohnia rubra n. gen., n. sp.,
nouveau Chaetognathe abyssal de l’Atlantique nord-africain :
description et position systématique, hypothèse phylogénétique. Bull. Mus. Natn. Hist. Nat., Paris, 4e sér., 8:185–194.
Castresana, J. 2000. Selection of conserved blocks from multiple
alignments for their use in phylogenetic analysis. Mol. Biol.
Evol. 17:540–552.
Cohen, B. L. 1998. Comparison of articulate brachiopod nuclear
and mitochondrial gene trees leads to a clade-based redefinition of protostomes (Protostomozoa) and deuterostomes
(Deuterostomozoa). Proc. R. Soc. Lond. B Biol. Sci. 265:
475–482.
de Rosa, R. 2001. Molecular data indicate the protostome affinity
of brachiopods. Syst. Biol. 50:848–859.
de Rosa R., J. K. Grenier, T. Andreeva, C. E. Cook, A. Adoutte,
M. Akam, C. B. Carroll, and G. Balavoine. 1999. Hox genes
in brachiopods and priapulids and protostome evolution.
Nature 399:772–776.
Eakin, R. M., and J. A. Westfall. 1964. Fine structure of the eye
of a chaetognath. J. Cell. Biol. 21:115–132.
Erber, A., D. Riemer, M. Bovenschulte, and K. Weber. 1998.
Molecular phylogeny of metazoan intermediate filament
proteins. J. Mol. Evol. 47:751–762.
Felsenstein, J. 1993. Phylogeny Inference Package (PHYLIP)
[computer program]. Version 3.5. University of Washington,
Seattle.
Ghirardelli, E. 1968. Some aspects of the biology of the
chaetognaths. Adv. Mar. Biol. 6:271–375.
———. 1994. The state of knowledge on Chaetognaths. Contrib.
Anim. Biol. 237–244.
Giribet, G., D. L. Distel, M. Polz, W. Sterrer, and W. C. Wheeler.
2000. Triploblastic relationships with emphasis on the
acoelomates and the position of Gnathostomulida, Cycliophora, Plathelminthes, and Chaetognatha: a combined approach of 18s rDNA sequences and morphology. Syst. Biol.
49:539–562.
Haase, A., M. Stern, K. Wachtler, and G. Bicker. 2001. A tissuespecific marker of Ecdysozoa. Dev. Gen. Evol. 211:428–433.
Halanych, K. 1996. Testing hypotheses of chaetognath origins:
long branches revealed by 18S rDNA. Syst. Biol. 45:223–246.
Helfenbein, K. G., and J. L. Boore. 2004. The mitochondrial
genome of Phoronis architecta—comparisons demonstrate
that Phoronids are lophotrochozoan protostomes. Mol. Biol.
Evol. 21:153–157.
Hyman, L. H. 1959. The invertebrates, Vol 5. Smaller Coelomate
groups. McGraw-Hill, New York.
Kumar, S., K. Tamura, and M. Nei. 1994. MEGA: molecular
evolutionary genetics analysis software for microcomputers.
Comput. Appl. Biosci. 10:189–191.
Lavrov, D. V., and W. M. Brown. 2001. Trichinella spiralis
mtDNA: a nematode mitochondrial genome that encodes
a putative ATP8 and normally structured tRNAs and has
a gene arrangement relatable to those of coelomate metazoans.
Genetics 157:621–637.
Chaetognaths Are Protostomes 2129
Littlewood, D. T. J., M. J. Telford, K. A. Clough, and K. Rohde.
1998. Gnathostomulida—an enigmatic metazoan phylum
from both morphological and molecular perspective. Mol.
Phyl. Evol. 9:72–79.
Mallatt, J., and C. J. Winchell. 2002. Testing the new
animal phylogeny: first use of combined large-subunit and
small-subunit rRNA gene sequences to classify the protostomes. Mol. Biol. Evol. 19:289–301.
Nielsen, C. 2001. Animal evolution: interrelationships of the
living Phyla. 2nd edition. Oxford University Press, Oxford.
Papillon, D., Y. Perez, L. Fasano, Y. Le Parco, and X. Caubit.
2003. Hox gene survey in the chaetognath Spadella cephaloptera: evolutionary implications. Dev. Genes Evol. 213:
142–148.
Peterson, K. J., and D. J. Eernisse. 2001. Animal phylogeny and
the ancestry of bilaterians: inferences from morphology and
18s rDNA gene sequences. Evol. Dev. 3:170–205.
Ruiz-Trillo, I., M. Riutort, D. T. Littlewood, E. A. Herniou, and
J. Baguna. 1999. Acoel flatworms: earliest extant bilaterian
Metazoans, not members of Platyhelminths. Science
283:1919–1923.
Schmidt, H. A., K. Strimmer, M. Vingron, and A. von Haeseler.
2002. TREE-PUZZLE: maximum likelihood phylogenetic
analysis using quartets and parallel computing. Bioinformatics
18:502–504.
Shimodaira, H. 2002. An approximately unbiased test of
phylogenetic tree selection. Syst. Biol. 3:492–508.
Shimodaira, H., and M. Hasegawa. 2001. CONSEL: for
assessing the confidence of phylogenetic tree selection.
Bioinformatics. 12:1246–1247.
Shimotori T., and T. Goto. 2001. Developmental fates of the
first four blastomeres of the chaetognath Paraspadella
gotoi: relationship to protostomes. Dev/ Growth Differ. 43:
371–382.
Stechmann, A., and M. Schlegel. 1999. Analysis of the complete
mitochondrial DNA sequence of the brachiopod Terebratulina
retusa places Brachiopoda within the protostomes. Proc. R.
Soc. Lond. B Biol. Sci. 266:2043–2052.
Telford, M. J., and P. W. H. Holland. 1993. The phylogenetic
affinities of the chaetognaths: a molecular analysis. Mol. Biol.
Evol. 10:660–676.
Tesler, G. 2002. GRIMM: genome rearrangements web server.
Bioinformatics 18:492–493.
Valentine, J. W. 1997. Cleavage patterns and the topology of
the metazoan tree of life. Proc. Natl. Acad. Sci. USA 94:
8001–8005.
Van Oppen, M. J., J. Catmull, B. J. McDonald, N. R. Hislop,
P. J. Hagerman, and D. J. Miller. 2002. The mitochondrial
genome of Acropora tenuis (Cnidaria; Scleractinia) contains a
large group I intron and a candidate control region. J. Mol.
Evol. 55:1–13.
Wadah, H., and N. Satoh. 1994. Details of the evolutionary
history from invertebrates to vertebrates as deduced from
the sequences of 18s rDNA. Proc. Natl. Acad. Sci. USA 91:
1801–1804.
Willmer, P. 1990. Invertebrate relationships: patterns in animal
evolution. Cambridge University Press, New York
Zrzàvy, J., S. Mihulka, P. Kepka, A. Bezdik, and D. Tietz. 1998.
Phylogeny of the metazoa based on morphological and 18s
ribosomal DNA evidence. Cladistics 14:249–285.
Franz Lang, Associate Editor
Accepted July 22, 2004