BioSystem$, 28 (1992) 139-151 Elsevier Scientific Publishers Ireland Ltd.
139
An analysis of partial 28S ribosomal RNA sequences suggests early radiations of sponges B~n~dicte L a f a y a, Nicole B o u r y - E s n a u l t b, J e a n V a c e l e t b a n d R i c h a r d C h r i s t e n a aURA 671 CNRS and Unit~rsi~ Pierre et Marie Curie, Station Zoologique, Observatoire Oc~anologique, ViUefranche sur nwr 06530 and bCcnt~ d'Ov~anologie de MarseiUe, Univevsib¢ Aix,-Ma~seille H URACNRS $1, Station Marine d'Endoume, Ma~ 13007 (F~'ance)
Sequences from the 5' end terminal part of 28S ribosomal R N A were obtained and compared for 22 animals belonging to all diploblasticphyla and for a large number of representativesof triploblasticMetazoa and protists.Phylogenetic analyses undertaken using differentmethods showed deep radiationsof phyla such as Ctenophora, Cnidaria and Placozoa but also for groups of Porifera of low taxonomic rank. Short internodes between these radiations suggested an early rapid diversification of diploblasts. A long internal branch preceding the diversification of all triploblasts analyzed could be explained either by a long period with a single ancestor or by the extinction of the earliest triploblastic radiations. Finally some unexpected relationships were revealed among Porifera. K~ywovde: Metazoa; Phylogeny; Porifera; Ribosomal RNA.
Introduction
Extant representatives of the 'lower' metazoan phyla share a very limited number of morphological homologies; on the other hand, the palaeontological record of the Precambian era is scarse and provides little indication about the transition from Protists to Metazoa. As a result, phylogenetic relationships between early metazoan phyla such as sponges, ctenophores, cnidarians, placozoans and some of the simplest forms of triploblasts remain unknown. Phylogenetic information can now be derived from the comparison of sequences from molecules that are strictly homologous in the various phyla, ribosomal RNA (rRNA) having proved to be one of the most valuable tools for such approaches. Phylogenies derived from the 5S molecule are now thought to be less reliable because of an inappropriate rate of evolution Covrvspondcn~ to: Bent~licte Lafay, URA 671 CNRS and Universlte Pierre et Marie Curie, Station Zoologique, Observatoire Oc~anologique, Villefranche sur met 06230, France.
(Halanych, 1991), but the 18S and 28S molecules have already provided information for phylogenetic reconstruction (Baroin et al., 1988; Christen et al., 1991; Field et al., 1988; KellyBorges et al., 1991; Lake 1990, Perasso et al., 1989; Sogin et al., 1986). In this study, we have investigated the problem of the relationships among the different diploblastic animals by analyzing partial 28S rRNA for representatives of several classes of sponges, cnidarians, ctenophores, placozoans and the major triploblastic phyla. Special effort has been directed toward sponges, since they are usually considered either to represent the deepest radiations among Metazoa or as having of a separate Origin.
Materials and methods
Sponges were collected locally from the Mediterranean sea, either near the marine station of Villefranche sur mer or near the marine station of Endourne. Other sequences are from Christen et al., 1991. RNA was isolated from
0303-2647/92/$05.00 © 1992 Elsevier Scientific Publishers Ireland Ltd. Printed and Published in Ireland
140
1 g of fresh or liquid nitrogen frozen tissue, homogenized with an ultra-turrax homogenizer in 5 ml of guanidinium thiocyanate (4 M), Tris-HC1 (50 raM, pH 7.6), EDTA (4 mM), Nlauryl-sarkozyl (2%), 2-mercaptoethanol (1%). Total RNA was separated from protein by phenol extraction repeated three times, followed by two chloroform washes. Total RNA was ethanol precipitated, suspended in sterile distilled water for measurement at 260/280 nm, reprecipitated with sodium acetate and ethanol and finally resuspended in sterile water at 2 ~g/~l. The quality of the RNA extract was examined by ethidium bromide staining of 1% agarose gels. RNA sequencing was carried out using a Sanger method modified to accommodate reverse transcriptase in place of DNA polymerase (Qu et al., 1983), with subsequent modifications. DNA synthesis was carried out in two steps: a labeling step involving primer extension was obtained using limited concentrations of 35S dATP, which was followed by a classical chain-termination step using dideoxynucleotides. Three synthetic primers complementary to evolutionary conserved domains located in the 5' end of the large subunit ribosomal RNA were used (Baroin et al., 1988). Sequences were aligned by eye, starting with the most conserved regions and progressively adding more divergent regions. Trees were derived from the molecular data using the parsimony computer program PAUP (Swofford, 1990), the distance matrix neighbor-joining method (Saitou and Nei, 1987) and the maximum likelihood program DNAML (Felsenstein, 1990); bootstrapping was usually done using the heuristic option of the PAUP program. When several representatives of a monophyletic group were available, bootstrapping was also carried out on species instead of nucleotide positions. Results
cies), placozoan (the single representative known) and porifera (11 species). Figure 2 summarizes the result of a phylogenetic analysis that included all diploblastic phyla. The topology shown was obtained using a neighbor-joining analysis, but congruent topologies were obtained using parsimony or maximum likelihood. The parsimony analysis led to a single tree (length 512, consistency index 0.609) that revealed the same monophyletic units as those shown in Fig. 2; the results of a bootstrap (maximum parsimony) analysis for the robustness of the data are shown on Fig. 2 as % indicated above each branch. A maximum likelihood analysis also revealed the same groupings; the branch lengths statistically significant at P < 0.01, are indicated on Fig. 1. Two monophyletic units corresponding to the phyla Cnidaria and Ctenophora can be observed and the Placozoa Trichoplax adhaerens lies close to the related Cnidaria and Ctenophora, but with a deeper origin. On the other hand, sponges are split into three groups, Dictyonella and associated species, Clathrina and Petrobiona, Reniera and Petrosia. The exact branching orders between the Cnidaria, Placozoa, Ctenophora and the three groups of Porifera remain largely uncertain, as shown by the differences observed between the various methods of tree reconstruction and the results of the bootstrap analyses. As discussed below, this is the result of short internodes separating long branches leading to extant species. Phylogenetic relationships between members of the Dictyonella group were also examined in more detail, by studying the Porifera alone. The branching orders of species within the D/ctyonella group could not be clearly determined (data not shown); however two species, Agelas oroides and Axinella damicornis, showed strong affinities. The exact relationship between the other species as well as the exact rooting position for this group remained uncertain.
Phylogenetic relationships among diploblasts We have obtained partial 28S rRNA sequences (-400 nucleotides) for 22 diploblastic Metazoa representing all diploblastic phyla (Fig. 1): ctenophores (3 species), cnidarians (7 spe-
Relationships between diploblasts and triploblasts In order to study the phylogenetic relationships between triploblasts and diploblasts,
141
M.musculus
CGCGACCUCA
P.ficiformis
NA
GAUCAGACGU GAG--A
GG*CGACCCG
--UUC ..........
CUGAAUUUAA
C ..........
GCAUAUUAGU
C-A-
R. f u l v a
NA
GAG--A
--UUC ..........
C ..........
C-A-
R.mucosa
NA--C .........
GAG--A
-ACUU ..........
C ..........
C-A-
D.incisa
NNNNN.
G-G--A
-A-UU.
A.oroides
NNG.
G--A
-A--U---U
A.damicornis
NNN .............
G--A
-A-UC---U-
C-A......
C ..........
C-AC-A-
S.genitrix
NNNA
G--A
-A-UC .....
C-A-
A.tenacior
NNGA
G--A
-A-UC .....
C .... A-
C.crambe
NNNN
G--A
-A-UC .....
-A-
C.cerebrum
NNN ..... G .......
U-AA
-A--U.
P.massiliana
NNU ..... G .......
U--A
-A--U .
T.adhaerens
NNN,
B.mitrata
NNNNNN .
B.ovata
NNN.
-G-AA
.
.
.
.
A.
.
.
.
.
.
***-
-A--U.
C ..........
A -A--U.
C ..........
C-AC-A-
-A - A - - U .
C ..........
C-A-
C ..........
C-A-
C.veneris
NNN.
.G--A
-A--U
E.strlcta
NNNN .............
G-AA
-A-UU
C-
L.octona
NNNN ..............
AA
-A--U.
A-
G.proboscldalis
NNNN ..............
AA
-A--U
A-
V.vellela
NNN ...............
AA
-A--U .......................
A-
A.uvaria
NNU ...............
NA
-A--U .......................
A-
H.hippopus
NNU.
.NA - A - - U
A-
F.edwardsl
NNU,
.AA - A - - U
A-
M.musculus
CAGCGGAGGA
P. f i c i f o r m i s
A
AAAGAAACUA
ACCAGGAUUC
CCUCAGUAAC
GGCGAGUGAA
R. f u l v a
A .... C--*-
R. m u c o s a
A
D.incisa
A
A .........
C
C---
A.oroides
A---A
A .........
C
C---
A.damicornis
A---N
A ..........
CU ............
S.genitrix
A
A
C.crambe
A
A ..........
C ..............
C.cerebrum
A
A ..........
CU
P.massiliana
A
A .........
CU
G ...... C-
--AG"
C---
-G ...... C-
--AG"
C---
G ...... C-
--AG ......................
T.adhaerens
A
C---
C---
U
N
A.tenacior
C---
C---
U
B.mitrata
A
B.ovata C.veneris E.stricta
A
A .........
C ...... U
L.octona
A
A .........
CU CU
C---
C---
U ..... U
..... AC---
A
U ..... U
..... AC---
A
U ..... U
..... AC---
G.proboscidalis
A
A .........
V.vellela
A
A .........
A.uvaria
N ................
H.hippopus
A
N .........
CU
F.edwardsi
A
A .........
C .......
*-N
--N ..... *-
. . . . . A.
CU --CU---N
Fig. 1. Sequences of diploblastie metazoans used in the paper. Only nucleotides that differ from those of the mouse are indicated (identities are denoted by hyphens, deletions by stars and nucleotide positions that could not be identified by 'N'). The portions of sequences used for the tree shown in Fig. 2 are overlined. Other sequences of protists and triploblastic metazoans were previously published.
142
M.musculus
CAGGGA***A
GAGCCCAGCG
CC*G*AAUCC
CCGCCGCGCG
UCGCG*****
P.ficlformls
GU .... AGA-
-U ...... GU
-U--A .... U
---GUCA-GC
-U*UAGUUAU
R.fulva
GU .... AG--
UU ...... GU
-U--A .... U
---GUCA-GC
-UCGAGUUGU
R.mucosa
GU .... UG--
CU ...... GU
-U-UU .... U
---AUCG-GC
-ACU*GUUGU
D.inclsa
GC ........
C--UU-GAGC
-U--A"
-**A--U-U*
***** .....
A.oroides
GU
A---U-GAGC
-U--A .... U
-U-G-AGUU-
AU*** .....
A.damicornis
GC ............
U-GAGC
-U-AA .... U
-U-G-AGUU-
AU*** .....
S.genltrix
GU ............
U-GAGC
-U--A-
-**G-AGU-A
CU*** .....
A.tenacior
GC ........
UU--U-GAGC
-U--A ........
C.crambe
GU
U---U-GAGC
-U--A .... *
C---U-GAAU
UU-AA .... U
G--GU--UU-
***** .....
-U--A .... U
GA-G---UU-
***** ..... ***** .....
C.cerebrum
GC ........
P.massiliana
GC ............
T.adhaerens
GU
U--AGC
GGC-CU*
***** .....
***G-AUU-C
AU*** .....
U---U--AGC
UU-*A .... U
-*CGAAGCUU
B.mitrata
C---U--AAC
UU-UA .... U
-*-A-CGCUU
***** .....
B.ovata
C---U--AAU
UU-UA .... U
-*-G-CGCUU
***** .....
C.veneris
C---U--AAU
UU-UA .... U
-*-GAUGCUU
***** .....
C---U--A*C
UUGAA .... U
-*CGUUGUGU
-**** .....
U---U--AAC
UU-AA .... U
-*UGUUGCUU
***** .....
UUCAG---U-
-*CUUCGCU*
***** .....
C---U--AGC
UU-AA .... U
-*CGUUGCUU
***** .....
A.uvaria
GC ....... U A---U--AAC
UU-AA .... U
-**GUUGCUU
***** .....
H.hippopus
GC ....... C
C---U--AAC
UU-AA .... U
-**GUUGCUU
***** .....
F.edwardsi
GC ....... C
C---U--AAC
UU-AA .... U
-**GUUGCUU
***** .....
M.musculus
**GC**GUGG
GAAAUGUGGC
GUACGGAAGA
CCCACUCCCC
*GGCGCCGCU
P.flciformis
U--A-CCG-C
---U ..... U
C-GGA--GUC
GAACUGU---
---UC-ACUG
E.stricta
G ........
U
L.octona
GC ........
G.proboscidalis
GC ...........
V.vellela
GC ........
U---A*C
R.fulva
UG .... CG-C
---U ..... U
C-GGA--GUC
GAACUGU---
--AUC-ACUG
R.mucosa
C--AU-CG-C
---U ..... U
C-GGA--G-C
AAACUGU---
--AUC-AC-G
D.incisa
....... C G - A
---U ......
CGGGA--G-U
AG-UUGUG-*
U-CGUUGC**
A.oroides
...... CA-C
---U ......
CGGGA--G-C
AG-UGGA---
U-CGUAGC**
A.damicornis
...... CA-C
---U ......
UGGGA--G-C
AG-UGGA---
U---CGGU**
S.genitrix
...... CA-A
---U ......
CGGGA--G-C
AG-UAGG---
U-A-AGAC
A.tenacior
.... GUCG-A
---U ......
CGGGA--G-C
AG-CUGUG--
U--GCUG-U*
C.crambe
...... CG-A
---U ......
CGGGA--G-C
AG-U*GUG--
C--ACUG-U*
C.cerebr~
--CAUC--CA
---U ..... U
U--AA .... U
GUUU**U---
C--GAUGU-C
P.massiliana
--CGUCU-CA
---U---N-U
UG-GA .... U
G-UU**U-A-
C--GA-GA-G U---AGUAUG
**
T.adhaerens
.... UUCG-C
---U---A-U
C-GU ..... U
GUU-**U**-
B.mitrata
.... GGCC-C
---U---A-U
U-CUA---AC
GUU-**U---
CAU*-AU-UG
B.ovata
.... AGCC-C
---U---A-U
U-CUA---AC
GUU-**U---
C-UGA-GA-G
C.veneris
.... GUCC-C
---U---A-U
U-C-A---AC
GUU-**U---
C-CGAGG--G
E.stricta
.... AACG-C
---U---A-U
U-CGA .... U A-UU**U*--
C ..... AU-G
L.octona
.... AACA-C
---U---A-U
C-CGA .... C
A-AU-GAUA-
GUUU**U*-A
G.proboscidalis
.... AACG-C
---U---A-U
C-CGA .... C
GUUU**U*-A
A-A--GAUA-
V.vellela
----AACGUC
---U---A--
C-CGA .... C
GUUU**U*AA
~C--GAUGC
A.uvaria
.... AACG-C
---U---A-U
C-CGA .... C
GUUU**U*-A
A .... GAUGC
H.hippopus
.... AACGCC
---U---A-U
C--GA--UUC
GUUU**U*-A
A .... AAUGC
F.ec~ardsi
.... AACG-C
---U---A-U
C-CGA .... C
GUUU**U*-A
A .... AAUGC
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
M.musculus
CGUGG*****
**GGGGCCCA
GAUCGAGGCC
CAGCCCGUGG
P.ficiformis
-UCCACU-GU
UC---C--G
.....
GAC-UG
--A--GC-UG
-****-UGA-
R.fulva
-UCCUCUUGU
UU---C--G
.....
GAC-UG
--A--GC-UG
-****-UGA-
R.mucosa
-UCCUUU-GU
UU---C--G
.....
GAC-UG
--A--GC-UG
-****-U'A-
Fig. 1. ~ontinue~
AGUCC*UUCU
143 D.incisa
***** .........
C---G ....
UGACCUG
--AG-GCA-G
U****--GA-
A.oroides
***** ........
CU-U-G ....
UGACCUG
--AA-GCA-G
U****-AGA-
A.damicornis
***** ........
CU-U-G ....
UGACCUG
--AA-GCA-N
U****-AGA-
S.genltrix
***** ........
CU-U-G ....
UGACCUG
--AA-GCA-G
U****--AA-
A.tenacior
*****
CC-U-G ....
GGACCUG
--AA-GCA-G
U****-AGA-
C.crambe
***** ........
CU-U ......
UGACCUG
--AA-GCA-G
-****-AGA-
C.cerebrum
G**** .......
CU .... U ....
NG-C-UG
--A-AGCA-G
U****-A-A-
P.massiliana
G**** .......
AU .... U ....
NG-C-UG
--A-AGCA-G
U****-A-A-
T.adhaerens
UU-** .......
U---U-U ....
UG-CCUG
--A--GCA-G
U****-A-A-
B.mitrata
ACGA* .......
UC-NN-U ....
UG-C-UG
--A-AGCAUA
U****--GA-
B.ovata
A**** .......
CC--U ......
UG-C-UG
--A-AGCAUA
U****--AA-
C.veneris
A**** .......
CC--U-U ....
UG-C-UG
--A-AGCAUG
U****-CGA-
E.stricta
G-*** .......
CAU---U ....
UG-C-UG
--A-AGCA-G
U****-ACA-
L.octona
AC*** .......
A-U--U .....
UG-C-UG
--A--GCA-A
U****---A-
G.proboscidalis
UU*** .......
A-U--U .....
U--C*UG
--A--GCA-A
-****-AA--
V.vellela
AU*** .......
U-U--UU ....
UG-C-AG
--A-AGCA-A
U****---A-
A.uvaria
GC*** .......
A-UA-UU
....
*G-C-UG
--A--GCA-A
U****---A-
H.hippopus
GC*** .......
A-UA-UU
....
UG-C-UG
--A--GCA-A
U****---A-
F.edwardsi
GC*** .......
A-UA-UU
....
UG-C-UG
--A--GCA-A
U****---A-
M.musculus
ACGGUGUGAG
GCCGGUAG**
CGGCCCCC**
GGCGCGCCGG
GCUCGGGUCU
P.ficiformis
-G .... A---
C--C .... GG
-AAGGAAACC
--GAGCGGC-
-GA--A-GG-
R.fulva
-G .... A---
C--N .... GG
-AAGGAAACC
--GAGCGGC-
-GA--A-GG-
R.mucosa
-G .... A---
N--C .... GG
-ANGGAAACC
--GAGCGGC-
-GG--A-AG-
D.inclsa
-G .... AC-A
C--C--GUG
A.oroides
-G .... AC--
C--C--GUG-
--A-A--GCC
--A**-GGCC
A-CACU ....
A.damlcornis
-G .... AC--
C--C--GCU-
--A-A-UGCU
--ACA-GGUC
---ACU ....
S.genitrlx
-G .... A--A
C--C--GUG-
--A-A--GUU
-U-CU-GGCC
---ACU-*--
A.tenaclor
-G .... A---
C--C--GUG-
A-A-G-UGCC
--GC--GGCC
-U-GCU-*--
C.crambe
-G .... A---
C--C--GCGU
*-A-A-UGCC
--UCU-GGNC
---ACU ....
C.cerebrum
-G--N-A--A
C--C--CUGU
G-CU-GGUGG
ANNCUUGGUA
UGAU-CACU-
P.massiliana
-G .... A--A
U--C--CUGU
G-UUGGU-GG
UCGCUCGGNA
UGAU-CACU-
T.adhaerens
-G .... AU-A
C--C--CUUU
*-A-UUA-AA
CAGCU-U-AA
UGAG-C-CU-
B.mitrata
-G .... A--A
U--C---UUU
*-A--GGACG
UC-CAUGGNA
CGAG-C-NU-
B.ovata
-G .... A--A
U--C---UUC
* .... GGACG
UC-CAUGGCA
CGAG-C--U-
C.veneris
-G .... A--A
U--C--UUUG
*-A--GGUCG
UC-CGCGGCA
CGAG-C--U-
E.stricta
-G .... AC-A
C--C---CGU
*---AA-GCU
C-A .... UC-
CGAU-U-CU-
L.octona
-G .... AC-A
U--C---CGU
*---A-UGUA
UA-UGUU*CA
CGAU-C-CU-
G.proboscidalls
-G .... AC--
U--C--UUGU
*---A-UGAA
UA-UGUU*CA
CGAU-C-CU-
V.vellela
-G .... AU-A
U--C---CGC
*--UA-AGUG
CA-NNNU*CA
UGAU-C-CU-
A. u v a r i a
-G .... AC-A
U--C---CGU
*--UA-UGUG
UAUCGUU*CA
CGAU-C-CU-
H. h i p p o p u
-G .... AC-A
U--C---CGU
*--UA-UGUG
CAUCGUU*CA
CGAU-C--U-
F. e d w a r d s
-G .... AC-A
U--C---CGU
*--UA-UGUG
CAUCGCU*CA
UGAU-C-CU-
M. m u s c u l u s
UCCCG*****
*GAGUCGGGU
UGCUUGGGAA
UGCAGCCCAA
AGCGGGUGGU
P. f i c i f o r m i s
--GUACUCGU
U
--U ...... U
R. f u l v a
--GUACUCGU
A
--U ...... U
........
R. m u c o s a
-UGU-CUCGU
A
--U ...... U
....................
D. i n c l s a
--GGA-
A. o r o i d e s
--NGA .................
A. d a m i c o r n i s
--AGA ..............
S. g e n i t r i x
--GGA-
Fig. 1. (continued)
......
U-G ....
G---AGCC
C
--U ...................
C
---GCU-*--
C ........... U .......
U ...................
U ........
--U ...................
U ...... N
--U ...................
U .......
144 A.tenacior
--GGA
--U-C
C.crambe
--GGA
--U ......
C.cerebrua
--GAA
--U ...............
P.massiliana
--GAA
N
T.adhaerens
---GA
..............
C
B.mitrata
--GAA
..............
N
--U .....
B.ovata
--GAA
C
--U ......
U
C.veneris
--GAA
.............
--U ......
U
E.stricta
--GAA
.................
L.octona
--UAU
....................................
AU---A---
G.proboscidalis
- -UAU
....................................
AU .......
V.vellela
N-UAU
..............
AU---A-
A.uvaria
--UNU
....................................
AU---A---
AC
..............
N
U
--U ........ --U .......
U-
-UU .......
--U' ..............
NU-
-AU .......
--U---A--G
U--
....... NN
......
NN--
-UU ....... -AU-C -AU-C
.....
U .....
U ..................
A--C
..... ..... ......
A
N
--
H.hlppopus
--UAU
......................
U .............
AUU--A---
F.edwardsi
--UAU
......................
U
AU---A---
M.musculus
AAACUCCAUC
UAAGGCUAAA
P.ficiformis
-G
C ......
G .....
A--UG'
R.fulva
-G
C ......
G .....
A--UG'
R.mucosa
-G
C ....
D.incisa
.............
A ........
A.oroides
........
A.
N ....
A.damicornis S.genitrlx
N-G .....
U--UG
AGACCGAUAG
UCAACAAGUA CGG ....... CGG .......
............
CG'
AU ................ --UG
CG.
................
CG ........
-A---G .....
A .... N .........
A .........
U ................
CG ........
U ..........
NNNNNNNNNN
.............
---A---G
A.tenacior
UACCGGCACG
.....
N-
NNNNNN
CNNNNNNNNN
C.crambe
.............
A-
C.cerebrun
.............
A ........
UU-C
....
CG ........
P.massiliana
.............
A ........
UU-C
..............
CG---NNNNN
T.adhaerens
-G ...........
A ........
UU-A-G
B.mitrata
....
C--CC
A ......
C-UU
................
CG ........
B.ovata
....
C---CN
......
C-UU
................
CG ........
C--CC
--UU
.... ---A ....
CG ........
C.veneris
....
E.stricta
.............
AA ........
L.octona
.......
G. proboscidalis
.............
U .....
V.vellela
U .....
............
CG ........
................
CG ........
UA ................
CG ........
A ........
UU ................
CG ........
A---G
UU ................
CG---G
....
UU ................
CG---N
....
....
A ........
-N-A---N--
N ......
U--
H. hippopus
.......
U .....
A
UU---N--
CG ........
F. edwardsl
.......
U .....
A ........
UU---N--
CG---N
M.musculus
CCGUAAGGGA
AAGU
P. ficiformis
....
G--NNN
NNNN
R. fulva
....
G ........
R. mucosa
....
G ......
D. incisa
....
G ........
A. oroides
....
G---C
A. damicornis
NNNNNNNNNN
N
....
S. genitri
--A-G
A. tenacior
NNNNNNNNNN
C. crambe
....
G ........
C. cerebrum
....
G ........
P. massiliaD~
NNNNNNNNNN
Fig. 1. (continued)
N NNN
N
NNNN
.........
N
NNNN N A NNNN
-NUU---N
..........
N-
CG---N---N
A.uvaria
....
145 T. adhaerens
.... G ........
A
B. mitrata
--A-G ........
A
B. ovata
--A-G ........
A
C. venerls
--A-G ........
A
E. stricta
.... G ........
A
L. octona
.... G ........
A
G. proboscidalis
.... G ........
A
V.vellela
.... G ........
N
A. uvaria
.... G ........
A
H. hippopus
.... G ........
A
F. edwardsi
.... G .....
---A
analyses were undertaken that included two representatives of each monophyletic unit revealed in Fig. 2 (but for the Placozoa for which only one representative was available) and representatives of some of the major triploblastic phyla. This approach was necessary because only a limited number of species could be reasonably analyzed at one time by the methods of parsimony, maximum likelihood and bootstrapping. The results of a neighbor-joining analysis are shown in Fig. 3. The same results were obtained following parsimony or maximum likelihood analyses. Parsimony led to two most parsimonious trees (length 335, consistency index: 0.573). Concerning the radiations of the diploblastic groups, one of the trees was identical to Fig. 3, the other topology differed in having Eunicella stricta forming a monophyletic unit with the two other cnidarians. Results of the bootstrap analysis are shown as a percentage indicated above each branch in Fig. 3, and revealed that none of the internal nodes was extremely robust. Maximum likelihood showed a slightly different order of branching and a decrease in the confidence limits exactly when the data from bootstrap were below 55% (see Fig. 3). Identical results were obtained when different representatives of the various groups were used (not shown). In conclusion, the monophyletic units revealed in Fig. 2 and characterized by rather long internal branches were observed as solid groupings, but the phylogenetic relations between these different groups were difficult to analyze, because they were separated by short internodes.
Rooting the metazoan tree When trying to root the metazoan tree using protistan sequences, the different methods (neighbor-joining, PAUP, DNAML) resulted in incongruent topologies: the branching order of the different metazoan phyla varied according to the method used {data not shown). Using the neighbor-joining method, diploblasts and triploblasts formed a monophyletic group that excluded any protist. In contrast analysis by parsimony or maximum likelihood resulted in topologies that intermingled protists among diploblastic phyla. Long periphal branches, indicative of a higher rate of evolution for triploblastic Metazoa, could have a negative effect (Smith et al., 1992) that could explain the difficulty of rooting the metazoan tree. We have thus tried to root the diploblastic species alone. A first analysis was undertaken using a few representatives of the Metazoa and a large number of protistan species. Following each analysis, protistan species that were clearly distantly related to Metazoa were removed and the data were reanalyzed. This approach is necessary to remove random noise resulting from distantly related species and that obscures real phylogenetic information between the related species. A small number of protists were identified as closely related to the Metazoa; they included the fungi, Parphyridium purpureun and the cryptomonads. This restricted set of protistan sequences was then used in conjunction with sequences representative of the different diploblastic groups identified above. Results obtained using the various methods are represented in Fig. 4. The exact rooting point of the diploblastic tree could not be determined with certainty. Placozoa, Ctenophora, Cnidaria and two Porifera subgroups belong to early radiations that probably diverged within a very short time. Discussion
There is a general agreement that Porifera can be grouped in a single phylum subdivided into two subphyla: Symplasma with class Hexactinellida, and Cellularia with classes Calcarea
146 Crambe crambe
Anchinoe tenacior 5pongosorltes genitrix Agelas oroides
I
89%
Sponges
I
Axinella damicornis Dictyoneila inctsa edwardsi
,•ikalia
oo::2::::::'°°°°-
Vellela vellela
99% ,
Cnldarlans
Gertonia probos~dalis
ilif Leuckartiara octona Euntcella strtcta
I I
i-I i I i I i I ||
97%
|
Ctenophoree
I00% |
!
Cestus veneris
i i!
! ! ! ! ! ! !
!
Clathrina cerebrum
99%
I
Sponges
Petroblonamassiliane Trichoplax aOhaerens
100~
I00~
Placozoan
Petrosia ficiformis Reniera fulva
Sponges
~N
Reniera mucosa
Fig. 2. Phylogenetic relationships among diploblastic Metazoa. Partial 28S rRNA sequences were analyzed as described in the text. The figure shown is the result of a neighbor-joining analysis; parsimony and maximum likelihood gave similar topologies (not shown). The most consistent branches identified by parsimony are indicated by numbers above each branch that correspond to the percentages of consistency in a bootstrap analysis. The confidence index of the maximum likelihood analysis on the same set of data are also indicated below the branches. **Indicates a branch statistically different from zero at P < 0.01). Dotted lines have been used to indicate when the branching orders are not robust according to one of the criteria discussed in the text. The tree is unrooted.
and Demospongiae (see Table I). However it has been suggested that Hexactinellida which have not been analyzed in this study may represent a distinct phylum with independent origin (Bergquist, 1985).
Palaeontological data suggest that hexactinellids and demosponges were already present in the early Cambrian. Demospongiae underwent major radiations in Ordovician possibly from the middle Cambrian Hazelia (Finks, 1970;
147
Petrosia riciformis
I00~ I gg~ ~N
Renieramucosa Sponges
Crambe crambe
I ; i
:
Dictyonella incisa Trichoplax adhaerens i
:
l
ioo
Beroe ovata
Cestus veneris
[.i : II
Placozoan
I Ctenophores
Clathrina cerebrum Petrobionamassiliana
I Sponges
Forskalia edwardsi L
Leuckartiara octona
I Cnldarlans
EuniceHastricta Mytllusedulis
Paracentrotus lividus r'lus muscuhJs Bombyxmori Caenorhabditis elegans Onchocerca gibsoni
Hymenolepsisdiminuta Fig. 3. Diploblast-triploblastrelationships.Sequences were analyzed with a neighbor-joiningmethod and verifiedby parsimony and maximum likelihoodas in Fig. 1. Triploblastsand diploblastsare separated in two well defined groups, but the branching orders of the diploblasticunits cannot be ascertained with good confidence (represented by dotted lines).Only the Petros/a-Ren/era radiation seems to be well separated from the other diploblasts.The tree is unrooted.
Bergquist, 1978). Calcarea could be present in the Cambrian if the Heteractinida were indeed Calcarea. However, Heteractinida which have polyactinal spicules made of calcium carbonate are very different from the modern Calcarea and are best considered as a separate extinct class (Finks, 1970). Isolated spicules that could be of calcarean origin have however been found in the early Cambrian (James et al., 1983), but the major calcarean radiations appear only in the Carboniferous or Permian (Finks, 1970). At low taxonomic ranks, molecular data provide interesting confirmations for the affinity between Agelas and Azinella (Bergquist, 1985) and between Haposclerida and Petrosida. All analyses showed that Renie'ra fulva is more
closely related to Petvosia than it is to Reniera mucosa, a result that brings doubts about the separation of these species into two separate orders (Haplosclerida and Petrosida, see Bergquist, 1980). In agreement with morphological interpretations, there is confirmation of the division between the two classes Calcarea and Demospongiae; however, rRNA sequences suggest a separation earlier than usually thought. Our data also suggest an early divergence between Haplosclerida and Petrosida on the one side and Poecilosclerida, Halichondrida, Axinellida and Agelasida on the other. Each of these two groups seems to form well separated monophyletic units. These data are clearly in
148 Chilomonas paramecium Cryptomonas ovata Porphyrldium purpureum 5accharomyces cerevlslae 5chizosaccharomyces pombe Neurospora crassa Beroe ovata
100% I "'"
,"*I i i
I I
i i
I I
; i
I
'1
!
'
I
Sponges
Petrobiona massiliana Forskaha edwardsi
,00%
l
Ctenophores
Clathrina cerebrum II
I I
I
, 52%
!
g4%
I Cestus venerls
Leuckartiara
38%
CnldaMans
octona
Eunlcella stricta Crambe crambe DIctyonella iNcisa Sponges Petrosia ficiformis
L
100%
~J 27% l
!
Trichoplax adhaerens
RenJera mucosa Placozoan
Fig. 4. Originof the diploblasts.Sequencesfromrepresentativesof the majordiploblasticgroupingsand fromrelatedprotists were analyzedto determinethe order of radiationfor the diploblasticphyla.This figure showsthe result of a neighbor-joining analysis, the conclusionsof whichwere in agreementwith parsimony(bootstrap % shown aboveeach branch) and maximum likelihood. **Representingbranch length positive at P < 0.01). Allphylaand two groupsof sponges showdeepbranchings,with short internodesby comparisonto the externalbranches: their exact branchingorder is thus difficultto assess precisely(domainsof uncertaintyrepresented by dotted lines).
contradiction with the classical taxonomy of demosponges, in particular because the coherence of the sub-class Ceractinomorpha is well accepted (L~vi, 1956; Bergquist, 1978, Bergqulst et al., 1983). If the respective places of Axinellida and Halichondrida have always been a matter of discussion, the presence of Haplosclerida and Poecilosclerida within the same subclass is of general agreement and recent cladistic analyses place Haplosclerida as sister group to Poecilosclerida (Van Soest, 1991). Finks (1970) suggested that from the middle
Cambrian primitive Hazelia, may have evolved on the one side the axial skeleton of Saccospongia followed by Axinellida and Ceractinomorpha and on the other side Hindia with a radial skeleton, followed by Tetractinellida, Spirophorida and Desmophorida. It is worth noting that Hazelia has a reticulated skeleton of oxeas similar to Haplosclerida-Petrosida, which could correlate with the early radiation observed in this study. Deep branchings and calculated genetic distances between the three different groups of Porifera identified in this study, as well as be-
149 Table I. Classical taxonomy of sponges. Phylum PORIFERA
Subphylum Cellularia Calcarea Calcinea C/athr/na c ~ n ~ m Calcaronea P ~ b / o n a ma~///ana Demospongiae Homoscleromorpha Homosclerophorida Tetracfinomorpha Axinellida Azi~ dam/corn/s Agelasida Age/as oro/des Astrophorida Spirophorida Hadromerida Desmophorida Ceractinomorpha Poecilosclerida Crambe ~ambe An~hinoe tvnaciov
Halichondrida Spongosorites genitrix Dictyonella incisa
Haplosclerida Reniera mu¢osa Reniera fidva
Petrosida Petrosia ficifo~ais
Dietyoceratida Dendroceratida Verongida Subphylum Symplasrna Hexactinellida Species are indicated in italicswhen their sequences have been analysed in this study.
tween each of these groups and the phyla Cnidaria, Ctenophora and the Placozoa suggest that each of these poriferan groups could be attributed a higher taxonomic rank. This is clearly in contrast with current classical views and it is importance that new molecular data be acquired for important orders that were not represented in our study (see Table I) or from a different molecule to investigate this problem furthermore.
Finally, the apparent relationship of the Calcarea with the cnidarian and ctenophore phyla in some figures of the present study must be taken with caution, since these radiations occurred within an area of poor resolution of the molecular methods: the genetic distances of the peripheral branches are long while the internodes are short and the order of branching for the internal nodes cannot yet be described with precision. In conclusion, a large agreement now exists among molecular phylogenies concerning the position of Metazoa within the protistan tree and the fact that triploblasts show a monophyletic origin well supported by a long internal branch (Christen et al., 1991). The general phylogeny of Metazoa is yet unresolved: 18S rRNA data have been considered for only a few diploblastic representatives (two cnidarians) which have been placed in a monophyletic unit with triploblasts (Patterson, 1989; Lake, 1990). Shorter sequences of 28S rRNA obtained for many diploblastic representatives (including Placozoa and Porifera), suggested that triploblasts are an early monophyletic radiation, sister group to the diploblasts (Christen et al., 1991). Regardless of which interpretation is correct, it is now necessary to further investigate two specific problems: what are the phylogenetic relationships among the earliest metazoan phyla and which protists are the closest relatives to Metazoa? A two-step approach is suggested by the results of a study that used echinoderms as a model system to compare palaeontological data and molecular phylogenies and which suggested that it was necessary first to derive an unrooted topology for closely related species before attempting to root this topology using only very closely related outgroups (Smith et al., 1992). Our data show that among Porifera, the different classes display deep branchings that probably place them as the earliest radiations among Metazoa; in order to resolve the basis of the metazoan tree, it will thus be necessary to obtain molecular data from the various poriferan classes and more particularly from hexactinellids. A major difference between diploblasts and
150 triploblasts is that triploblasts are united by a long internal branch to all other eukaryotes: the monophyly of triploblasts appears as a robust feature. The existence of this long internal branch could be explained (i) by the complete extinction of the earliest triploblastic radiations, or (ii) by a long period during which triploblasts were represented by a single lineage. These two hypotheses have major bearings on the interpretations of the fossil record. The extinction hypothesis predicts that triploblastic fossils from the early Cambrian and perhaps much later, are possibly only distantly related to any living animal; the two hypotheses mean that it would be difficult to reconstruct the morphology of the primitive triploblasts from data of extant triploblasts and that a careful study combining palaeontology, morphology and molecular phylogenies will be required. Concerning the rooting of the metazoan tree, only a few closely related protists are known and rooting involves resolving short internal nodes followed by long branches leading to extant species. In the presence of different rates of molecular evolution in the different lineages, all methods have a tendency to root the tree in order to make the apparent rates of evolution more equal: the root is then artificially placed on the longest branch. A correct rooting is possible when the most closely related outgroups are used; invariant positions in these outgroups can then be used to determine the ancestral states and derive the rooting through a polarized parsimony approach (Smith et al., 1992). It is not yet feasible to use this approach to root the metazoan tree. If we have now located a number of protists that are indeed closely related to Metazoa (fungi, cryptomonads, Porphyridium...), none of these groups has been studied to such an extent that invariant positions and thus ancestral states can be derived. Also, within Metazoa, groups that show deep radiations should now be sequenced for a large number of species in order to derive their internal topologies. Except for hexactinellids, for which we have as yet no data, our study has shown deep radiations for the classes Calcarea and Demospongiae; long internal branches also appeared for the few examined orders of
Demospongiae. These degrees of divergence could lead to the recognition of higher taxonomic ranks. The study of the origin of Metazoa and their relationships to the closest protists should now clearly include representatives of all classes and orders of sponges, especially since our results suggest that taxonomic ranks within Porifera are still uncertain. We are now investigating this problem by sequencing longer stretches of the various poriferan representatives as well as the other diploblastic phyla and the closely related protists revealed by rRNA analyses (Baroin et al., 1988; Christen et al., 1991; Perasso et al., 1989). Acknowledgements
This work was supported by the CNRS and the Universities Paris VI and Aix-Marseille II, and grants from the Association Recherche et Partage as well as the Conseil G~n~ral des Alpes Maritimes. References Baroin, A., Perasso, R., Qu, L.H., Brugerolle, G.,
Bachellerie, J.P. and Adoutte, A., 1988. Partial phylogenyof the unicellulareukaryotesbased on rapid sequencing of a portion of 28S ribosomal RNA. Proc. Natl. Acad. Sci. U.S.A. 83, 3474-3478. Bergqulst, P.R., 1978. Sponges(HutchinsonPress, London) pp. 1-268. Bergqulst, P.R., 1980. The ordinal and subclass classification of the Demospongiae(Porifera); appraisal of the present arrangement and proposal for a new order. N.Z.J. Zool. 7, 1-16. Bergqulst, P.R. 1985. Poriferan relationships, in: The Origins and Relationships of Lower Invertebrates, S. Conway Morris et al. (eds.) SystematicsAssoc. Special Vol., 28, 14-27. Bergquist, P.R. and WellsmR.J., 1983. Chemotaxonomyof the Porifera:the developmentand current status of the field, in: Marine Natural Products, P. Scheuer (ed.) (AcademicPress, New York)pp. 1- 50. Christen, R., Ratto, A., Baroin, A., Perasso, R., Grell, K.G. and Adoutte,A., 1991. An analysisof metazoansorigin, using comparisons of partial sequences of the 28S ribosomal RNA, reveals an early emergence of triploblasts. EMBO J. 10, 499-503.
Felsenstein,J., 1990.PHYLIPManualVersion3.3. (University Herbarium, University of California, Berkeley, California). Field, K.G., Olsen, G.J., Lane, D.J., Giovannoni, S.J.,
151 Ghisolin, M.T., Raft, E.C., Pace, N.R. and Raft, R.A., 1988. Molecular phylogeny of the animal kingdom. Science 239, 748-753. Halanych, K.M. 1991. 5S ribosomal RNA sequences inappropriate for phylogenetic reconstruction. Mol. Biol. Evol. 8, 249-253. Finks, R.M. 1970. The evolution and ecologic history of sponges during palaezoic times. Syrup. Zool. See. London 25, 3-22. James, N.P. and Klappa, C.F., 1983, Petrogenesis of early Cambrian reef limestones, Labrador, Canada. J. Sedim. Petrol. 53, 1051-1096. Kelly-Borges, M., Bergquist, P.R. and Bergquist, P.L., 1991, Pbylogenetic relationships within the order Hadromerida (Porifera, Demospongiae, Tetractinomorpha) as indicated by ribosomal RNA sequence comparisons. Biochem. Syst. Ecol. 19, 117-125. Lake, J.A., 1990, Origin of the metazoa. Proc. Natl. Acad. Sci. U.S.A. 87, 763-766. I.~viC., 1956, Etude des Ha//sarca de Roscoff. Embryologie et syst~matique des D~mosponges. Arch. Zool. Exp. G~n. 93, 1-181. Patterson, C., 1989, Phylogenetic relationsof major groups: conclusions and prospects, in: The Hierarchy of Life. Molecules and morphology in phylogenetic analysis. Nobel Symposium 70. B. Fernholm, K. Bremer and H. JSrnvall (eds.) (Biomedical Division, Elsevier Science Publishers B.V., The Netherlands).
Qu, L.H., Michot, B. and Bachellerie, J.P., 1983. Improved method for structure probing in large RNAs: a rapid heteroiogens sequencing approach is coupled to the direct mapping of nuclease accessible sites. Application to the 5' terminal domain of eukaryotic 28S rRNA. Nucleic Acids Res. 11, 5903-5920. Perasso, R., Baroin, A., Qu, L.H., Bachellerie, J.P. and Adoutte, A., 1989, Origin of the algae. Nature. 339, 142 - 144. Saitou, N. and Nei, M., 1987, The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4, 406-425. Smith, A.B., Lafay, B. and Christen, R., 1992. Comparative variation of morphological and molecular evolution through geologic time: 28S ribosomal RNA versus morphology in echinoids. Phil. Trans. R. Soc., London B., in press. Soest van R.W.M., 1991, Demosponge higher classification re-examined, in: Fossil and Recent Sponges. J. Reitner and H. Keupp (eds.) (Springer- Verlag, Berlin Heidelberg) pp. 54- 71. Sogin, M.L., Elwood, H.J. and Gunderson, J.H., 1986. Evolutionary diversity of eukaryotic small subunit rRNA genes. Proc. Natl. Acad. Sci. U.S.A. 83, 1383-1387. Swofford, DL., 1990. PAUP: phylogenetic analysis using parsimony, version 3.0. (Illinois Natural History Survey, Champaign, IL).
139
An analysis of partial 28S ribosomal RNA sequences suggests early radiations of sponges B~n~dicte L a f a y a, Nicole B o u r y - E s n a u l t b, J e a n V a c e l e t b a n d R i c h a r d C h r i s t e n a aURA 671 CNRS and Unit~rsi~ Pierre et Marie Curie, Station Zoologique, Observatoire Oc~anologique, ViUefranche sur nwr 06530 and bCcnt~ d'Ov~anologie de MarseiUe, Univevsib¢ Aix,-Ma~seille H URACNRS $1, Station Marine d'Endoume, Ma~ 13007 (F~'ance)
Sequences from the 5' end terminal part of 28S ribosomal R N A were obtained and compared for 22 animals belonging to all diploblasticphyla and for a large number of representativesof triploblasticMetazoa and protists.Phylogenetic analyses undertaken using differentmethods showed deep radiationsof phyla such as Ctenophora, Cnidaria and Placozoa but also for groups of Porifera of low taxonomic rank. Short internodes between these radiations suggested an early rapid diversification of diploblasts. A long internal branch preceding the diversification of all triploblasts analyzed could be explained either by a long period with a single ancestor or by the extinction of the earliest triploblastic radiations. Finally some unexpected relationships were revealed among Porifera. K~ywovde: Metazoa; Phylogeny; Porifera; Ribosomal RNA.
Introduction
Extant representatives of the 'lower' metazoan phyla share a very limited number of morphological homologies; on the other hand, the palaeontological record of the Precambian era is scarse and provides little indication about the transition from Protists to Metazoa. As a result, phylogenetic relationships between early metazoan phyla such as sponges, ctenophores, cnidarians, placozoans and some of the simplest forms of triploblasts remain unknown. Phylogenetic information can now be derived from the comparison of sequences from molecules that are strictly homologous in the various phyla, ribosomal RNA (rRNA) having proved to be one of the most valuable tools for such approaches. Phylogenies derived from the 5S molecule are now thought to be less reliable because of an inappropriate rate of evolution Covrvspondcn~ to: Bent~licte Lafay, URA 671 CNRS and Universlte Pierre et Marie Curie, Station Zoologique, Observatoire Oc~anologique, Villefranche sur met 06230, France.
(Halanych, 1991), but the 18S and 28S molecules have already provided information for phylogenetic reconstruction (Baroin et al., 1988; Christen et al., 1991; Field et al., 1988; KellyBorges et al., 1991; Lake 1990, Perasso et al., 1989; Sogin et al., 1986). In this study, we have investigated the problem of the relationships among the different diploblastic animals by analyzing partial 28S rRNA for representatives of several classes of sponges, cnidarians, ctenophores, placozoans and the major triploblastic phyla. Special effort has been directed toward sponges, since they are usually considered either to represent the deepest radiations among Metazoa or as having of a separate Origin.
Materials and methods
Sponges were collected locally from the Mediterranean sea, either near the marine station of Villefranche sur mer or near the marine station of Endourne. Other sequences are from Christen et al., 1991. RNA was isolated from
0303-2647/92/$05.00 © 1992 Elsevier Scientific Publishers Ireland Ltd. Printed and Published in Ireland
140
1 g of fresh or liquid nitrogen frozen tissue, homogenized with an ultra-turrax homogenizer in 5 ml of guanidinium thiocyanate (4 M), Tris-HC1 (50 raM, pH 7.6), EDTA (4 mM), Nlauryl-sarkozyl (2%), 2-mercaptoethanol (1%). Total RNA was separated from protein by phenol extraction repeated three times, followed by two chloroform washes. Total RNA was ethanol precipitated, suspended in sterile distilled water for measurement at 260/280 nm, reprecipitated with sodium acetate and ethanol and finally resuspended in sterile water at 2 ~g/~l. The quality of the RNA extract was examined by ethidium bromide staining of 1% agarose gels. RNA sequencing was carried out using a Sanger method modified to accommodate reverse transcriptase in place of DNA polymerase (Qu et al., 1983), with subsequent modifications. DNA synthesis was carried out in two steps: a labeling step involving primer extension was obtained using limited concentrations of 35S dATP, which was followed by a classical chain-termination step using dideoxynucleotides. Three synthetic primers complementary to evolutionary conserved domains located in the 5' end of the large subunit ribosomal RNA were used (Baroin et al., 1988). Sequences were aligned by eye, starting with the most conserved regions and progressively adding more divergent regions. Trees were derived from the molecular data using the parsimony computer program PAUP (Swofford, 1990), the distance matrix neighbor-joining method (Saitou and Nei, 1987) and the maximum likelihood program DNAML (Felsenstein, 1990); bootstrapping was usually done using the heuristic option of the PAUP program. When several representatives of a monophyletic group were available, bootstrapping was also carried out on species instead of nucleotide positions. Results
cies), placozoan (the single representative known) and porifera (11 species). Figure 2 summarizes the result of a phylogenetic analysis that included all diploblastic phyla. The topology shown was obtained using a neighbor-joining analysis, but congruent topologies were obtained using parsimony or maximum likelihood. The parsimony analysis led to a single tree (length 512, consistency index 0.609) that revealed the same monophyletic units as those shown in Fig. 2; the results of a bootstrap (maximum parsimony) analysis for the robustness of the data are shown on Fig. 2 as % indicated above each branch. A maximum likelihood analysis also revealed the same groupings; the branch lengths statistically significant at P < 0.01, are indicated on Fig. 1. Two monophyletic units corresponding to the phyla Cnidaria and Ctenophora can be observed and the Placozoa Trichoplax adhaerens lies close to the related Cnidaria and Ctenophora, but with a deeper origin. On the other hand, sponges are split into three groups, Dictyonella and associated species, Clathrina and Petrobiona, Reniera and Petrosia. The exact branching orders between the Cnidaria, Placozoa, Ctenophora and the three groups of Porifera remain largely uncertain, as shown by the differences observed between the various methods of tree reconstruction and the results of the bootstrap analyses. As discussed below, this is the result of short internodes separating long branches leading to extant species. Phylogenetic relationships between members of the Dictyonella group were also examined in more detail, by studying the Porifera alone. The branching orders of species within the D/ctyonella group could not be clearly determined (data not shown); however two species, Agelas oroides and Axinella damicornis, showed strong affinities. The exact relationship between the other species as well as the exact rooting position for this group remained uncertain.
Phylogenetic relationships among diploblasts We have obtained partial 28S rRNA sequences (-400 nucleotides) for 22 diploblastic Metazoa representing all diploblastic phyla (Fig. 1): ctenophores (3 species), cnidarians (7 spe-
Relationships between diploblasts and triploblasts In order to study the phylogenetic relationships between triploblasts and diploblasts,
141
M.musculus
CGCGACCUCA
P.ficiformis
NA
GAUCAGACGU GAG--A
GG*CGACCCG
--UUC ..........
CUGAAUUUAA
C ..........
GCAUAUUAGU
C-A-
R. f u l v a
NA
GAG--A
--UUC ..........
C ..........
C-A-
R.mucosa
NA--C .........
GAG--A
-ACUU ..........
C ..........
C-A-
D.incisa
NNNNN.
G-G--A
-A-UU.
A.oroides
NNG.
G--A
-A--U---U
A.damicornis
NNN .............
G--A
-A-UC---U-
C-A......
C ..........
C-AC-A-
S.genitrix
NNNA
G--A
-A-UC .....
C-A-
A.tenacior
NNGA
G--A
-A-UC .....
C .... A-
C.crambe
NNNN
G--A
-A-UC .....
-A-
C.cerebrum
NNN ..... G .......
U-AA
-A--U.
P.massiliana
NNU ..... G .......
U--A
-A--U .
T.adhaerens
NNN,
B.mitrata
NNNNNN .
B.ovata
NNN.
-G-AA
.
.
.
.
A.
.
.
.
.
.
***-
-A--U.
C ..........
A -A--U.
C ..........
C-AC-A-
-A - A - - U .
C ..........
C-A-
C ..........
C-A-
C.veneris
NNN.
.G--A
-A--U
E.strlcta
NNNN .............
G-AA
-A-UU
C-
L.octona
NNNN ..............
AA
-A--U.
A-
G.proboscldalis
NNNN ..............
AA
-A--U
A-
V.vellela
NNN ...............
AA
-A--U .......................
A-
A.uvaria
NNU ...............
NA
-A--U .......................
A-
H.hippopus
NNU.
.NA - A - - U
A-
F.edwardsl
NNU,
.AA - A - - U
A-
M.musculus
CAGCGGAGGA
P. f i c i f o r m i s
A
AAAGAAACUA
ACCAGGAUUC
CCUCAGUAAC
GGCGAGUGAA
R. f u l v a
A .... C--*-
R. m u c o s a
A
D.incisa
A
A .........
C
C---
A.oroides
A---A
A .........
C
C---
A.damicornis
A---N
A ..........
CU ............
S.genitrix
A
A
C.crambe
A
A ..........
C ..............
C.cerebrum
A
A ..........
CU
P.massiliana
A
A .........
CU
G ...... C-
--AG"
C---
-G ...... C-
--AG"
C---
G ...... C-
--AG ......................
T.adhaerens
A
C---
C---
U
N
A.tenacior
C---
C---
U
B.mitrata
A
B.ovata C.veneris E.stricta
A
A .........
C ...... U
L.octona
A
A .........
CU CU
C---
C---
U ..... U
..... AC---
A
U ..... U
..... AC---
A
U ..... U
..... AC---
G.proboscidalis
A
A .........
V.vellela
A
A .........
A.uvaria
N ................
H.hippopus
A
N .........
CU
F.edwardsi
A
A .........
C .......
*-N
--N ..... *-
. . . . . A.
CU --CU---N
Fig. 1. Sequences of diploblastie metazoans used in the paper. Only nucleotides that differ from those of the mouse are indicated (identities are denoted by hyphens, deletions by stars and nucleotide positions that could not be identified by 'N'). The portions of sequences used for the tree shown in Fig. 2 are overlined. Other sequences of protists and triploblastic metazoans were previously published.
142
M.musculus
CAGGGA***A
GAGCCCAGCG
CC*G*AAUCC
CCGCCGCGCG
UCGCG*****
P.ficlformls
GU .... AGA-
-U ...... GU
-U--A .... U
---GUCA-GC
-U*UAGUUAU
R.fulva
GU .... AG--
UU ...... GU
-U--A .... U
---GUCA-GC
-UCGAGUUGU
R.mucosa
GU .... UG--
CU ...... GU
-U-UU .... U
---AUCG-GC
-ACU*GUUGU
D.inclsa
GC ........
C--UU-GAGC
-U--A"
-**A--U-U*
***** .....
A.oroides
GU
A---U-GAGC
-U--A .... U
-U-G-AGUU-
AU*** .....
A.damicornis
GC ............
U-GAGC
-U-AA .... U
-U-G-AGUU-
AU*** .....
S.genltrix
GU ............
U-GAGC
-U--A-
-**G-AGU-A
CU*** .....
A.tenacior
GC ........
UU--U-GAGC
-U--A ........
C.crambe
GU
U---U-GAGC
-U--A .... *
C---U-GAAU
UU-AA .... U
G--GU--UU-
***** .....
-U--A .... U
GA-G---UU-
***** ..... ***** .....
C.cerebrum
GC ........
P.massiliana
GC ............
T.adhaerens
GU
U--AGC
GGC-CU*
***** .....
***G-AUU-C
AU*** .....
U---U--AGC
UU-*A .... U
-*CGAAGCUU
B.mitrata
C---U--AAC
UU-UA .... U
-*-A-CGCUU
***** .....
B.ovata
C---U--AAU
UU-UA .... U
-*-G-CGCUU
***** .....
C.veneris
C---U--AAU
UU-UA .... U
-*-GAUGCUU
***** .....
C---U--A*C
UUGAA .... U
-*CGUUGUGU
-**** .....
U---U--AAC
UU-AA .... U
-*UGUUGCUU
***** .....
UUCAG---U-
-*CUUCGCU*
***** .....
C---U--AGC
UU-AA .... U
-*CGUUGCUU
***** .....
A.uvaria
GC ....... U A---U--AAC
UU-AA .... U
-**GUUGCUU
***** .....
H.hippopus
GC ....... C
C---U--AAC
UU-AA .... U
-**GUUGCUU
***** .....
F.edwardsi
GC ....... C
C---U--AAC
UU-AA .... U
-**GUUGCUU
***** .....
M.musculus
**GC**GUGG
GAAAUGUGGC
GUACGGAAGA
CCCACUCCCC
*GGCGCCGCU
P.flciformis
U--A-CCG-C
---U ..... U
C-GGA--GUC
GAACUGU---
---UC-ACUG
E.stricta
G ........
U
L.octona
GC ........
G.proboscidalis
GC ...........
V.vellela
GC ........
U---A*C
R.fulva
UG .... CG-C
---U ..... U
C-GGA--GUC
GAACUGU---
--AUC-ACUG
R.mucosa
C--AU-CG-C
---U ..... U
C-GGA--G-C
AAACUGU---
--AUC-AC-G
D.incisa
....... C G - A
---U ......
CGGGA--G-U
AG-UUGUG-*
U-CGUUGC**
A.oroides
...... CA-C
---U ......
CGGGA--G-C
AG-UGGA---
U-CGUAGC**
A.damicornis
...... CA-C
---U ......
UGGGA--G-C
AG-UGGA---
U---CGGU**
S.genitrix
...... CA-A
---U ......
CGGGA--G-C
AG-UAGG---
U-A-AGAC
A.tenacior
.... GUCG-A
---U ......
CGGGA--G-C
AG-CUGUG--
U--GCUG-U*
C.crambe
...... CG-A
---U ......
CGGGA--G-C
AG-U*GUG--
C--ACUG-U*
C.cerebr~
--CAUC--CA
---U ..... U
U--AA .... U
GUUU**U---
C--GAUGU-C
P.massiliana
--CGUCU-CA
---U---N-U
UG-GA .... U
G-UU**U-A-
C--GA-GA-G U---AGUAUG
**
T.adhaerens
.... UUCG-C
---U---A-U
C-GU ..... U
GUU-**U**-
B.mitrata
.... GGCC-C
---U---A-U
U-CUA---AC
GUU-**U---
CAU*-AU-UG
B.ovata
.... AGCC-C
---U---A-U
U-CUA---AC
GUU-**U---
C-UGA-GA-G
C.veneris
.... GUCC-C
---U---A-U
U-C-A---AC
GUU-**U---
C-CGAGG--G
E.stricta
.... AACG-C
---U---A-U
U-CGA .... U A-UU**U*--
C ..... AU-G
L.octona
.... AACA-C
---U---A-U
C-CGA .... C
A-AU-GAUA-
GUUU**U*-A
G.proboscidalis
.... AACG-C
---U---A-U
C-CGA .... C
GUUU**U*-A
A-A--GAUA-
V.vellela
----AACGUC
---U---A--
C-CGA .... C
GUUU**U*AA
~C--GAUGC
A.uvaria
.... AACG-C
---U---A-U
C-CGA .... C
GUUU**U*-A
A .... GAUGC
H.hippopus
.... AACGCC
---U---A-U
C--GA--UUC
GUUU**U*-A
A .... AAUGC
F.ec~ardsi
.... AACG-C
---U---A-U
C-CGA .... C
GUUU**U*-A
A .... AAUGC
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
M.musculus
CGUGG*****
**GGGGCCCA
GAUCGAGGCC
CAGCCCGUGG
P.ficiformis
-UCCACU-GU
UC---C--G
.....
GAC-UG
--A--GC-UG
-****-UGA-
R.fulva
-UCCUCUUGU
UU---C--G
.....
GAC-UG
--A--GC-UG
-****-UGA-
R.mucosa
-UCCUUU-GU
UU---C--G
.....
GAC-UG
--A--GC-UG
-****-U'A-
Fig. 1. ~ontinue~
AGUCC*UUCU
143 D.incisa
***** .........
C---G ....
UGACCUG
--AG-GCA-G
U****--GA-
A.oroides
***** ........
CU-U-G ....
UGACCUG
--AA-GCA-G
U****-AGA-
A.damicornis
***** ........
CU-U-G ....
UGACCUG
--AA-GCA-N
U****-AGA-
S.genltrix
***** ........
CU-U-G ....
UGACCUG
--AA-GCA-G
U****--AA-
A.tenacior
*****
CC-U-G ....
GGACCUG
--AA-GCA-G
U****-AGA-
C.crambe
***** ........
CU-U ......
UGACCUG
--AA-GCA-G
-****-AGA-
C.cerebrum
G**** .......
CU .... U ....
NG-C-UG
--A-AGCA-G
U****-A-A-
P.massiliana
G**** .......
AU .... U ....
NG-C-UG
--A-AGCA-G
U****-A-A-
T.adhaerens
UU-** .......
U---U-U ....
UG-CCUG
--A--GCA-G
U****-A-A-
B.mitrata
ACGA* .......
UC-NN-U ....
UG-C-UG
--A-AGCAUA
U****--GA-
B.ovata
A**** .......
CC--U ......
UG-C-UG
--A-AGCAUA
U****--AA-
C.veneris
A**** .......
CC--U-U ....
UG-C-UG
--A-AGCAUG
U****-CGA-
E.stricta
G-*** .......
CAU---U ....
UG-C-UG
--A-AGCA-G
U****-ACA-
L.octona
AC*** .......
A-U--U .....
UG-C-UG
--A--GCA-A
U****---A-
G.proboscidalis
UU*** .......
A-U--U .....
U--C*UG
--A--GCA-A
-****-AA--
V.vellela
AU*** .......
U-U--UU ....
UG-C-AG
--A-AGCA-A
U****---A-
A.uvaria
GC*** .......
A-UA-UU
....
*G-C-UG
--A--GCA-A
U****---A-
H.hippopus
GC*** .......
A-UA-UU
....
UG-C-UG
--A--GCA-A
U****---A-
F.edwardsi
GC*** .......
A-UA-UU
....
UG-C-UG
--A--GCA-A
U****---A-
M.musculus
ACGGUGUGAG
GCCGGUAG**
CGGCCCCC**
GGCGCGCCGG
GCUCGGGUCU
P.ficiformis
-G .... A---
C--C .... GG
-AAGGAAACC
--GAGCGGC-
-GA--A-GG-
R.fulva
-G .... A---
C--N .... GG
-AAGGAAACC
--GAGCGGC-
-GA--A-GG-
R.mucosa
-G .... A---
N--C .... GG
-ANGGAAACC
--GAGCGGC-
-GG--A-AG-
D.inclsa
-G .... AC-A
C--C--GUG
A.oroides
-G .... AC--
C--C--GUG-
--A-A--GCC
--A**-GGCC
A-CACU ....
A.damlcornis
-G .... AC--
C--C--GCU-
--A-A-UGCU
--ACA-GGUC
---ACU ....
S.genitrlx
-G .... A--A
C--C--GUG-
--A-A--GUU
-U-CU-GGCC
---ACU-*--
A.tenaclor
-G .... A---
C--C--GUG-
A-A-G-UGCC
--GC--GGCC
-U-GCU-*--
C.crambe
-G .... A---
C--C--GCGU
*-A-A-UGCC
--UCU-GGNC
---ACU ....
C.cerebrum
-G--N-A--A
C--C--CUGU
G-CU-GGUGG
ANNCUUGGUA
UGAU-CACU-
P.massiliana
-G .... A--A
U--C--CUGU
G-UUGGU-GG
UCGCUCGGNA
UGAU-CACU-
T.adhaerens
-G .... AU-A
C--C--CUUU
*-A-UUA-AA
CAGCU-U-AA
UGAG-C-CU-
B.mitrata
-G .... A--A
U--C---UUU
*-A--GGACG
UC-CAUGGNA
CGAG-C-NU-
B.ovata
-G .... A--A
U--C---UUC
* .... GGACG
UC-CAUGGCA
CGAG-C--U-
C.veneris
-G .... A--A
U--C--UUUG
*-A--GGUCG
UC-CGCGGCA
CGAG-C--U-
E.stricta
-G .... AC-A
C--C---CGU
*---AA-GCU
C-A .... UC-
CGAU-U-CU-
L.octona
-G .... AC-A
U--C---CGU
*---A-UGUA
UA-UGUU*CA
CGAU-C-CU-
G.proboscidalls
-G .... AC--
U--C--UUGU
*---A-UGAA
UA-UGUU*CA
CGAU-C-CU-
V.vellela
-G .... AU-A
U--C---CGC
*--UA-AGUG
CA-NNNU*CA
UGAU-C-CU-
A. u v a r i a
-G .... AC-A
U--C---CGU
*--UA-UGUG
UAUCGUU*CA
CGAU-C-CU-
H. h i p p o p u
-G .... AC-A
U--C---CGU
*--UA-UGUG
CAUCGUU*CA
CGAU-C--U-
F. e d w a r d s
-G .... AC-A
U--C---CGU
*--UA-UGUG
CAUCGCU*CA
UGAU-C-CU-
M. m u s c u l u s
UCCCG*****
*GAGUCGGGU
UGCUUGGGAA
UGCAGCCCAA
AGCGGGUGGU
P. f i c i f o r m i s
--GUACUCGU
U
--U ...... U
R. f u l v a
--GUACUCGU
A
--U ...... U
........
R. m u c o s a
-UGU-CUCGU
A
--U ...... U
....................
D. i n c l s a
--GGA-
A. o r o i d e s
--NGA .................
A. d a m i c o r n i s
--AGA ..............
S. g e n i t r i x
--GGA-
Fig. 1. (continued)
......
U-G ....
G---AGCC
C
--U ...................
C
---GCU-*--
C ........... U .......
U ...................
U ........
--U ...................
U ...... N
--U ...................
U .......
144 A.tenacior
--GGA
--U-C
C.crambe
--GGA
--U ......
C.cerebrua
--GAA
--U ...............
P.massiliana
--GAA
N
T.adhaerens
---GA
..............
C
B.mitrata
--GAA
..............
N
--U .....
B.ovata
--GAA
C
--U ......
U
C.veneris
--GAA
.............
--U ......
U
E.stricta
--GAA
.................
L.octona
--UAU
....................................
AU---A---
G.proboscidalis
- -UAU
....................................
AU .......
V.vellela
N-UAU
..............
AU---A-
A.uvaria
--UNU
....................................
AU---A---
AC
..............
N
U
--U ........ --U .......
U-
-UU .......
--U' ..............
NU-
-AU .......
--U---A--G
U--
....... NN
......
NN--
-UU ....... -AU-C -AU-C
.....
U .....
U ..................
A--C
..... ..... ......
A
N
--
H.hlppopus
--UAU
......................
U .............
AUU--A---
F.edwardsi
--UAU
......................
U
AU---A---
M.musculus
AAACUCCAUC
UAAGGCUAAA
P.ficiformis
-G
C ......
G .....
A--UG'
R.fulva
-G
C ......
G .....
A--UG'
R.mucosa
-G
C ....
D.incisa
.............
A ........
A.oroides
........
A.
N ....
A.damicornis S.genitrlx
N-G .....
U--UG
AGACCGAUAG
UCAACAAGUA CGG ....... CGG .......
............
CG'
AU ................ --UG
CG.
................
CG ........
-A---G .....
A .... N .........
A .........
U ................
CG ........
U ..........
NNNNNNNNNN
.............
---A---G
A.tenacior
UACCGGCACG
.....
N-
NNNNNN
CNNNNNNNNN
C.crambe
.............
A-
C.cerebrun
.............
A ........
UU-C
....
CG ........
P.massiliana
.............
A ........
UU-C
..............
CG---NNNNN
T.adhaerens
-G ...........
A ........
UU-A-G
B.mitrata
....
C--CC
A ......
C-UU
................
CG ........
B.ovata
....
C---CN
......
C-UU
................
CG ........
C--CC
--UU
.... ---A ....
CG ........
C.veneris
....
E.stricta
.............
AA ........
L.octona
.......
G. proboscidalis
.............
U .....
V.vellela
U .....
............
CG ........
................
CG ........
UA ................
CG ........
A ........
UU ................
CG ........
A---G
UU ................
CG---G
....
UU ................
CG---N
....
....
A ........
-N-A---N--
N ......
U--
H. hippopus
.......
U .....
A
UU---N--
CG ........
F. edwardsl
.......
U .....
A ........
UU---N--
CG---N
M.musculus
CCGUAAGGGA
AAGU
P. ficiformis
....
G--NNN
NNNN
R. fulva
....
G ........
R. mucosa
....
G ......
D. incisa
....
G ........
A. oroides
....
G---C
A. damicornis
NNNNNNNNNN
N
....
S. genitri
--A-G
A. tenacior
NNNNNNNNNN
C. crambe
....
G ........
C. cerebrum
....
G ........
P. massiliaD~
NNNNNNNNNN
Fig. 1. (continued)
N NNN
N
NNNN
.........
N
NNNN N A NNNN
-NUU---N
..........
N-
CG---N---N
A.uvaria
....
145 T. adhaerens
.... G ........
A
B. mitrata
--A-G ........
A
B. ovata
--A-G ........
A
C. venerls
--A-G ........
A
E. stricta
.... G ........
A
L. octona
.... G ........
A
G. proboscidalis
.... G ........
A
V.vellela
.... G ........
N
A. uvaria
.... G ........
A
H. hippopus
.... G ........
A
F. edwardsi
.... G .....
---A
analyses were undertaken that included two representatives of each monophyletic unit revealed in Fig. 2 (but for the Placozoa for which only one representative was available) and representatives of some of the major triploblastic phyla. This approach was necessary because only a limited number of species could be reasonably analyzed at one time by the methods of parsimony, maximum likelihood and bootstrapping. The results of a neighbor-joining analysis are shown in Fig. 3. The same results were obtained following parsimony or maximum likelihood analyses. Parsimony led to two most parsimonious trees (length 335, consistency index: 0.573). Concerning the radiations of the diploblastic groups, one of the trees was identical to Fig. 3, the other topology differed in having Eunicella stricta forming a monophyletic unit with the two other cnidarians. Results of the bootstrap analysis are shown as a percentage indicated above each branch in Fig. 3, and revealed that none of the internal nodes was extremely robust. Maximum likelihood showed a slightly different order of branching and a decrease in the confidence limits exactly when the data from bootstrap were below 55% (see Fig. 3). Identical results were obtained when different representatives of the various groups were used (not shown). In conclusion, the monophyletic units revealed in Fig. 2 and characterized by rather long internal branches were observed as solid groupings, but the phylogenetic relations between these different groups were difficult to analyze, because they were separated by short internodes.
Rooting the metazoan tree When trying to root the metazoan tree using protistan sequences, the different methods (neighbor-joining, PAUP, DNAML) resulted in incongruent topologies: the branching order of the different metazoan phyla varied according to the method used {data not shown). Using the neighbor-joining method, diploblasts and triploblasts formed a monophyletic group that excluded any protist. In contrast analysis by parsimony or maximum likelihood resulted in topologies that intermingled protists among diploblastic phyla. Long periphal branches, indicative of a higher rate of evolution for triploblastic Metazoa, could have a negative effect (Smith et al., 1992) that could explain the difficulty of rooting the metazoan tree. We have thus tried to root the diploblastic species alone. A first analysis was undertaken using a few representatives of the Metazoa and a large number of protistan species. Following each analysis, protistan species that were clearly distantly related to Metazoa were removed and the data were reanalyzed. This approach is necessary to remove random noise resulting from distantly related species and that obscures real phylogenetic information between the related species. A small number of protists were identified as closely related to the Metazoa; they included the fungi, Parphyridium purpureun and the cryptomonads. This restricted set of protistan sequences was then used in conjunction with sequences representative of the different diploblastic groups identified above. Results obtained using the various methods are represented in Fig. 4. The exact rooting point of the diploblastic tree could not be determined with certainty. Placozoa, Ctenophora, Cnidaria and two Porifera subgroups belong to early radiations that probably diverged within a very short time. Discussion
There is a general agreement that Porifera can be grouped in a single phylum subdivided into two subphyla: Symplasma with class Hexactinellida, and Cellularia with classes Calcarea
146 Crambe crambe
Anchinoe tenacior 5pongosorltes genitrix Agelas oroides
I
89%
Sponges
I
Axinella damicornis Dictyoneila inctsa edwardsi
,•ikalia
oo::2::::::'°°°°-
Vellela vellela
99% ,
Cnldarlans
Gertonia probos~dalis
ilif Leuckartiara octona Euntcella strtcta
I I
i-I i I i I i I ||
97%
|
Ctenophoree
I00% |
!
Cestus veneris
i i!
! ! ! ! ! ! !
!
Clathrina cerebrum
99%
I
Sponges
Petroblonamassiliane Trichoplax aOhaerens
100~
I00~
Placozoan
Petrosia ficiformis Reniera fulva
Sponges
~N
Reniera mucosa
Fig. 2. Phylogenetic relationships among diploblastic Metazoa. Partial 28S rRNA sequences were analyzed as described in the text. The figure shown is the result of a neighbor-joining analysis; parsimony and maximum likelihood gave similar topologies (not shown). The most consistent branches identified by parsimony are indicated by numbers above each branch that correspond to the percentages of consistency in a bootstrap analysis. The confidence index of the maximum likelihood analysis on the same set of data are also indicated below the branches. **Indicates a branch statistically different from zero at P < 0.01). Dotted lines have been used to indicate when the branching orders are not robust according to one of the criteria discussed in the text. The tree is unrooted.
and Demospongiae (see Table I). However it has been suggested that Hexactinellida which have not been analyzed in this study may represent a distinct phylum with independent origin (Bergquist, 1985).
Palaeontological data suggest that hexactinellids and demosponges were already present in the early Cambrian. Demospongiae underwent major radiations in Ordovician possibly from the middle Cambrian Hazelia (Finks, 1970;
147
Petrosia riciformis
I00~ I gg~ ~N
Renieramucosa Sponges
Crambe crambe
I ; i
:
Dictyonella incisa Trichoplax adhaerens i
:
l
ioo
Beroe ovata
Cestus veneris
[.i : II
Placozoan
I Ctenophores
Clathrina cerebrum Petrobionamassiliana
I Sponges
Forskalia edwardsi L
Leuckartiara octona
I Cnldarlans
EuniceHastricta Mytllusedulis
Paracentrotus lividus r'lus muscuhJs Bombyxmori Caenorhabditis elegans Onchocerca gibsoni
Hymenolepsisdiminuta Fig. 3. Diploblast-triploblastrelationships.Sequences were analyzed with a neighbor-joiningmethod and verifiedby parsimony and maximum likelihoodas in Fig. 1. Triploblastsand diploblastsare separated in two well defined groups, but the branching orders of the diploblasticunits cannot be ascertained with good confidence (represented by dotted lines).Only the Petros/a-Ren/era radiation seems to be well separated from the other diploblasts.The tree is unrooted.
Bergquist, 1978). Calcarea could be present in the Cambrian if the Heteractinida were indeed Calcarea. However, Heteractinida which have polyactinal spicules made of calcium carbonate are very different from the modern Calcarea and are best considered as a separate extinct class (Finks, 1970). Isolated spicules that could be of calcarean origin have however been found in the early Cambrian (James et al., 1983), but the major calcarean radiations appear only in the Carboniferous or Permian (Finks, 1970). At low taxonomic ranks, molecular data provide interesting confirmations for the affinity between Agelas and Azinella (Bergquist, 1985) and between Haposclerida and Petrosida. All analyses showed that Renie'ra fulva is more
closely related to Petvosia than it is to Reniera mucosa, a result that brings doubts about the separation of these species into two separate orders (Haplosclerida and Petrosida, see Bergquist, 1980). In agreement with morphological interpretations, there is confirmation of the division between the two classes Calcarea and Demospongiae; however, rRNA sequences suggest a separation earlier than usually thought. Our data also suggest an early divergence between Haplosclerida and Petrosida on the one side and Poecilosclerida, Halichondrida, Axinellida and Agelasida on the other. Each of these two groups seems to form well separated monophyletic units. These data are clearly in
148 Chilomonas paramecium Cryptomonas ovata Porphyrldium purpureum 5accharomyces cerevlslae 5chizosaccharomyces pombe Neurospora crassa Beroe ovata
100% I "'"
,"*I i i
I I
i i
I I
; i
I
'1
!
'
I
Sponges
Petrobiona massiliana Forskaha edwardsi
,00%
l
Ctenophores
Clathrina cerebrum II
I I
I
, 52%
!
g4%
I Cestus venerls
Leuckartiara
38%
CnldaMans
octona
Eunlcella stricta Crambe crambe DIctyonella iNcisa Sponges Petrosia ficiformis
L
100%
~J 27% l
!
Trichoplax adhaerens
RenJera mucosa Placozoan
Fig. 4. Originof the diploblasts.Sequencesfromrepresentativesof the majordiploblasticgroupingsand fromrelatedprotists were analyzedto determinethe order of radiationfor the diploblasticphyla.This figure showsthe result of a neighbor-joining analysis, the conclusionsof whichwere in agreementwith parsimony(bootstrap % shown aboveeach branch) and maximum likelihood. **Representingbranch length positive at P < 0.01). Allphylaand two groupsof sponges showdeepbranchings,with short internodesby comparisonto the externalbranches: their exact branchingorder is thus difficultto assess precisely(domainsof uncertaintyrepresented by dotted lines).
contradiction with the classical taxonomy of demosponges, in particular because the coherence of the sub-class Ceractinomorpha is well accepted (L~vi, 1956; Bergquist, 1978, Bergqulst et al., 1983). If the respective places of Axinellida and Halichondrida have always been a matter of discussion, the presence of Haplosclerida and Poecilosclerida within the same subclass is of general agreement and recent cladistic analyses place Haplosclerida as sister group to Poecilosclerida (Van Soest, 1991). Finks (1970) suggested that from the middle
Cambrian primitive Hazelia, may have evolved on the one side the axial skeleton of Saccospongia followed by Axinellida and Ceractinomorpha and on the other side Hindia with a radial skeleton, followed by Tetractinellida, Spirophorida and Desmophorida. It is worth noting that Hazelia has a reticulated skeleton of oxeas similar to Haplosclerida-Petrosida, which could correlate with the early radiation observed in this study. Deep branchings and calculated genetic distances between the three different groups of Porifera identified in this study, as well as be-
149 Table I. Classical taxonomy of sponges. Phylum PORIFERA
Subphylum Cellularia Calcarea Calcinea C/athr/na c ~ n ~ m Calcaronea P ~ b / o n a ma~///ana Demospongiae Homoscleromorpha Homosclerophorida Tetracfinomorpha Axinellida Azi~ dam/corn/s Agelasida Age/as oro/des Astrophorida Spirophorida Hadromerida Desmophorida Ceractinomorpha Poecilosclerida Crambe ~ambe An~hinoe tvnaciov
Halichondrida Spongosorites genitrix Dictyonella incisa
Haplosclerida Reniera mu¢osa Reniera fidva
Petrosida Petrosia ficifo~ais
Dietyoceratida Dendroceratida Verongida Subphylum Symplasrna Hexactinellida Species are indicated in italicswhen their sequences have been analysed in this study.
tween each of these groups and the phyla Cnidaria, Ctenophora and the Placozoa suggest that each of these poriferan groups could be attributed a higher taxonomic rank. This is clearly in contrast with current classical views and it is importance that new molecular data be acquired for important orders that were not represented in our study (see Table I) or from a different molecule to investigate this problem furthermore.
Finally, the apparent relationship of the Calcarea with the cnidarian and ctenophore phyla in some figures of the present study must be taken with caution, since these radiations occurred within an area of poor resolution of the molecular methods: the genetic distances of the peripheral branches are long while the internodes are short and the order of branching for the internal nodes cannot yet be described with precision. In conclusion, a large agreement now exists among molecular phylogenies concerning the position of Metazoa within the protistan tree and the fact that triploblasts show a monophyletic origin well supported by a long internal branch (Christen et al., 1991). The general phylogeny of Metazoa is yet unresolved: 18S rRNA data have been considered for only a few diploblastic representatives (two cnidarians) which have been placed in a monophyletic unit with triploblasts (Patterson, 1989; Lake, 1990). Shorter sequences of 28S rRNA obtained for many diploblastic representatives (including Placozoa and Porifera), suggested that triploblasts are an early monophyletic radiation, sister group to the diploblasts (Christen et al., 1991). Regardless of which interpretation is correct, it is now necessary to further investigate two specific problems: what are the phylogenetic relationships among the earliest metazoan phyla and which protists are the closest relatives to Metazoa? A two-step approach is suggested by the results of a study that used echinoderms as a model system to compare palaeontological data and molecular phylogenies and which suggested that it was necessary first to derive an unrooted topology for closely related species before attempting to root this topology using only very closely related outgroups (Smith et al., 1992). Our data show that among Porifera, the different classes display deep branchings that probably place them as the earliest radiations among Metazoa; in order to resolve the basis of the metazoan tree, it will thus be necessary to obtain molecular data from the various poriferan classes and more particularly from hexactinellids. A major difference between diploblasts and
150 triploblasts is that triploblasts are united by a long internal branch to all other eukaryotes: the monophyly of triploblasts appears as a robust feature. The existence of this long internal branch could be explained (i) by the complete extinction of the earliest triploblastic radiations, or (ii) by a long period during which triploblasts were represented by a single lineage. These two hypotheses have major bearings on the interpretations of the fossil record. The extinction hypothesis predicts that triploblastic fossils from the early Cambrian and perhaps much later, are possibly only distantly related to any living animal; the two hypotheses mean that it would be difficult to reconstruct the morphology of the primitive triploblasts from data of extant triploblasts and that a careful study combining palaeontology, morphology and molecular phylogenies will be required. Concerning the rooting of the metazoan tree, only a few closely related protists are known and rooting involves resolving short internal nodes followed by long branches leading to extant species. In the presence of different rates of molecular evolution in the different lineages, all methods have a tendency to root the tree in order to make the apparent rates of evolution more equal: the root is then artificially placed on the longest branch. A correct rooting is possible when the most closely related outgroups are used; invariant positions in these outgroups can then be used to determine the ancestral states and derive the rooting through a polarized parsimony approach (Smith et al., 1992). It is not yet feasible to use this approach to root the metazoan tree. If we have now located a number of protists that are indeed closely related to Metazoa (fungi, cryptomonads, Porphyridium...), none of these groups has been studied to such an extent that invariant positions and thus ancestral states can be derived. Also, within Metazoa, groups that show deep radiations should now be sequenced for a large number of species in order to derive their internal topologies. Except for hexactinellids, for which we have as yet no data, our study has shown deep radiations for the classes Calcarea and Demospongiae; long internal branches also appeared for the few examined orders of
Demospongiae. These degrees of divergence could lead to the recognition of higher taxonomic ranks. The study of the origin of Metazoa and their relationships to the closest protists should now clearly include representatives of all classes and orders of sponges, especially since our results suggest that taxonomic ranks within Porifera are still uncertain. We are now investigating this problem by sequencing longer stretches of the various poriferan representatives as well as the other diploblastic phyla and the closely related protists revealed by rRNA analyses (Baroin et al., 1988; Christen et al., 1991; Perasso et al., 1989). Acknowledgements
This work was supported by the CNRS and the Universities Paris VI and Aix-Marseille II, and grants from the Association Recherche et Partage as well as the Conseil G~n~ral des Alpes Maritimes. References Baroin, A., Perasso, R., Qu, L.H., Brugerolle, G.,
Bachellerie, J.P. and Adoutte, A., 1988. Partial phylogenyof the unicellulareukaryotesbased on rapid sequencing of a portion of 28S ribosomal RNA. Proc. Natl. Acad. Sci. U.S.A. 83, 3474-3478. Bergqulst, P.R., 1978. Sponges(HutchinsonPress, London) pp. 1-268. Bergqulst, P.R., 1980. The ordinal and subclass classification of the Demospongiae(Porifera); appraisal of the present arrangement and proposal for a new order. N.Z.J. Zool. 7, 1-16. Bergqulst, P.R. 1985. Poriferan relationships, in: The Origins and Relationships of Lower Invertebrates, S. Conway Morris et al. (eds.) SystematicsAssoc. Special Vol., 28, 14-27. Bergquist, P.R. and WellsmR.J., 1983. Chemotaxonomyof the Porifera:the developmentand current status of the field, in: Marine Natural Products, P. Scheuer (ed.) (AcademicPress, New York)pp. 1- 50. Christen, R., Ratto, A., Baroin, A., Perasso, R., Grell, K.G. and Adoutte,A., 1991. An analysisof metazoansorigin, using comparisons of partial sequences of the 28S ribosomal RNA, reveals an early emergence of triploblasts. EMBO J. 10, 499-503.
Felsenstein,J., 1990.PHYLIPManualVersion3.3. (University Herbarium, University of California, Berkeley, California). Field, K.G., Olsen, G.J., Lane, D.J., Giovannoni, S.J.,
151 Ghisolin, M.T., Raft, E.C., Pace, N.R. and Raft, R.A., 1988. Molecular phylogeny of the animal kingdom. Science 239, 748-753. Halanych, K.M. 1991. 5S ribosomal RNA sequences inappropriate for phylogenetic reconstruction. Mol. Biol. Evol. 8, 249-253. Finks, R.M. 1970. The evolution and ecologic history of sponges during palaezoic times. Syrup. Zool. See. London 25, 3-22. James, N.P. and Klappa, C.F., 1983, Petrogenesis of early Cambrian reef limestones, Labrador, Canada. J. Sedim. Petrol. 53, 1051-1096. Kelly-Borges, M., Bergquist, P.R. and Bergquist, P.L., 1991, Pbylogenetic relationships within the order Hadromerida (Porifera, Demospongiae, Tetractinomorpha) as indicated by ribosomal RNA sequence comparisons. Biochem. Syst. Ecol. 19, 117-125. Lake, J.A., 1990, Origin of the metazoa. Proc. Natl. Acad. Sci. U.S.A. 87, 763-766. I.~viC., 1956, Etude des Ha//sarca de Roscoff. Embryologie et syst~matique des D~mosponges. Arch. Zool. Exp. G~n. 93, 1-181. Patterson, C., 1989, Phylogenetic relationsof major groups: conclusions and prospects, in: The Hierarchy of Life. Molecules and morphology in phylogenetic analysis. Nobel Symposium 70. B. Fernholm, K. Bremer and H. JSrnvall (eds.) (Biomedical Division, Elsevier Science Publishers B.V., The Netherlands).
Qu, L.H., Michot, B. and Bachellerie, J.P., 1983. Improved method for structure probing in large RNAs: a rapid heteroiogens sequencing approach is coupled to the direct mapping of nuclease accessible sites. Application to the 5' terminal domain of eukaryotic 28S rRNA. Nucleic Acids Res. 11, 5903-5920. Perasso, R., Baroin, A., Qu, L.H., Bachellerie, J.P. and Adoutte, A., 1989, Origin of the algae. Nature. 339, 142 - 144. Saitou, N. and Nei, M., 1987, The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4, 406-425. Smith, A.B., Lafay, B. and Christen, R., 1992. Comparative variation of morphological and molecular evolution through geologic time: 28S ribosomal RNA versus morphology in echinoids. Phil. Trans. R. Soc., London B., in press. Soest van R.W.M., 1991, Demosponge higher classification re-examined, in: Fossil and Recent Sponges. J. Reitner and H. Keupp (eds.) (Springer- Verlag, Berlin Heidelberg) pp. 54- 71. Sogin, M.L., Elwood, H.J. and Gunderson, J.H., 1986. Evolutionary diversity of eukaryotic small subunit rRNA genes. Proc. Natl. Acad. Sci. U.S.A. 83, 1383-1387. Swofford, DL., 1990. PAUP: phylogenetic analysis using parsimony, version 3.0. (Illinois Natural History Survey, Champaign, IL).
