Mechanisms of Development 120 (2003) 429–440
www.elsevier.com/locate/modo
Early steps in the evolution of multicellularity: deep structural and
functional homologies among homeobox genes in sponges and higher
metazoans
Cristiano C. Coutinhoa,*, Rodrigo N. Fonsecaa, José João C. Mansurea, Radovan Borojevicb
b
a
Laboratory of Molecular Biology of Embryonic Development, Federal University of Rio de Janeiro, 21941-970 Ilha do Fundão, Rio de Janeiro, Brazil
Laboratory of Cell Proliferation and Differentiation, Department of Histology and Embryology, Federal University of Rio de Janeiro, 21941-970 Ilha do
Fundão, Rio de Janeiro, Brazil
Received 24 May 2002; received in revised form 11 December 2002; accepted 6 January 2003
Abstract
The sponge homeobox gene EmH-3 had not been attributed to any homeobox family. Comparative promoter and homeodomain sequence
analyses suggest that it is related to the Hox11 gene, which belongs to the Tlx homeobox family. Hox11 is highly expressed in proliferating
progenitor cells, but expression is downregulated during cell differentiation. Using reporter gene methodology, we monitored function of the
sponge EmH-3 promoter transfected into human erythroleukemia K562 cells. These cells express the Tlx/Hox11 gene constitutively, and
downregulate its expression upon differentiation. The same pattern of expression and downregulation was observed for the sponge reporter
construct. We propose that Tlx/Hox11 genes have structural and functional homologies conserved in phylogenetically distant groups, that
represent a deep homology in the regulation of cell proliferation, commitment and differentiation.
q 2003 Elsevier Science Ireland Ltd. All rights reserved.
Keywords: Tlx; Hox11; EmH-3; NKL complex; Porifera; Multicellular evolution
1. Introduction
Acquisition of the multicellular level of organization was
a major innovative step in evolution. The coordinated
integration of several cell types in a higher functioning
organism gave rise to all the extant animals. While classical
cytological studies have already shown that multicellular
organisms share cell structure characteristics with protists,
the question of which protist group(s) harbor(s) the
precursors that evolved from complex colonies to true
multicellular animals, the Metazoa, remains open (Hyman,
1940; Hadzi, 1953; Willmer, 1970). Recently, molecular
studies have shown that: (1) Metazoa are monophyletic
(Shenk and Steele, 1993; Müller, 1995); (2) sponges are the
most ancient and primitive multicellular organisms (Borchellini et al., 1998, 2001; Zrzavy et al., 1998); and (3) among
the Protista, choanoflagellates are the sister group of
* Corresponding author. Tel.: þ55-21-2562-6481/2590-8736; fax: þ 5521-2562-6483.
E-mail address: ccoutinho@hotmail.com (C.C. Coutinho).
sponges and consequently of all the Metazoa (Wainright
et al., 1993). The significance of these postulates is that
molecular mechanisms underlying cell integration in multicellular organisms arose, at least in part, during the
evolutionary step from choanoflagellates to sponges.
Multicellularity involves the ordered determination of
cell fate and spatial distribution during embryogenesis, and
homeostatic maintenance of different and cooperating cell
types in the adult. Recent studies of both phenomena have
stressed the importance of stem cells, characterized by their
self-renewal potential and capacity to give rise to several
differentiated cell types following specific induction stimuli
(Blau et al., 2001). The retention of the undifferentiated
phenotype versus engagement in differentiation is thus one
of the key-points in the regulation of tissue structure and
function in multicellular organisms. Stem cell systems are
already present in the most primitive metazoans. Previous
studies showed that sponge cell populations are organized as
a single stem cell system, in which archaeocytes can
0925-4773/03/$ - see front matter q 2003 Elsevier Science Ireland Ltd. All rights reserved.
doi:10.1016/S0925-4773(03)00007-8
430
C.C. Coutinho et al. / Mechanisms of Development 120 (2003) 429–440
produce all cell types including the germ cells (Borojevic,
1966, 1970).
In higher organisms, homeobox genes are important in
the control of cell fate, proliferation and differentiation. We
searched for homeobox sequences in sponges and, initially,
we identified EfH-1 (for Ephydatia fluviatilis homeobox
sequence) (Coutinho et al., 1994), followed by the
orthologous EmH-3 gene detected in Ephydatia muelleri
(Richelle-Maurer et al., 1998). The former gene was also
identified by Seimiya et al. (1994) and named prox2 (for
Porifera homeobox gene). Subsequently, several other
homeobox genes belonging to different homeobox subfamilies were identified in demosponges (Seimiya et al.,
1997; Hoshiyama et al., 1998; Degnan et al., 1995).
Recently, Manuel and Le Parco (2000) extended this list
to calcareous sponges, and assembled a phylogenetic
analysis of all sponge homeobox genes. They suggested
the following classification for sponge gene groups that are
related to the well-defined homeobox classes of other
animals: Msx ( prox3 – Seimiya et al., 1994), NK2 ( prox-1
– Seimiya et al., 1994; SrNkxA, SrNkxB, SrNkxC, SrNkxD –
Manuel and Le Parco, 2000), POU (spou-1 and spou-2 –
Seimiya et al., 1997), and Pax (sPax2/5/8 – Hoshiyama
et al., 1998). A group of several genes (EmH-3, prox2,
EfH1, SpoxTa1 – Coutinho et al., 1994; Degnan et al., 1995)
was considered to have no obvious orthologous relationship
to bilateralian homeobox genes.
In the present study, we have focused our attention on the
latter group of genes, and have compared their sequences
with those of a series of homeobox genes from invertebrates
and vertebrates. Here we propose the inclusion of EmH-3,
prox2, EfH-1 and SpoxTa1 into the Tlx homeobox gene
family. The homology between EmH-3 and Tlx genes was
further studied by monitoring gene function and expression
in a heterologous assay system. The human Tlx gene
Hox11/tcl3 (homeobox gene/T-cell leukemia), was originally isolated as a proto-oncogene associated with human
T-cell leukemias (Kennedy et al., 1991; Lu et al., 1991).
Oncogenic activity of Hox11 has been confirmed in bone
marrow cells in vitro following cell infection with retroviruses containing the gene (Hawley et al., 1994). The
Hox11 protein is now known to regulate the cell cycle,
interacting with cell-cycle controlling phosphatases
(Kawabe et al., 1997). Similarly, RT-PCR analysis indicates
that expression of EmH-3 is correlated with cell replication
in sponges, being most intense in an enriched population of
proliferating sponge stem cells, the archaeocytes (Richelle-Maurer et al., 1998; Richelle-Maurer and Van de Vyver,
1999). Both archaeocyte differentiation and EmH-3
expression were responsive to retinoic acid (Nikko et al.,
2001), which is a morphogen in higher animals.
Genes of the Tlx family contribute to homeobox gene
clusters, forming the vertebrate NKL and Drosophila 93DE
complexes. They encode related homeodomain primary
structures, and share the Eh1 repression domain at the
amino-terminal region (Pollard and Holland, 2000; Jagla
et al., 2001). Moreover, NKL and 93DE gene complexes
appear to share homologous functions: they are involved in
a network of gene interactions that govern progressive cell
fate decisions during mesoderm development. The fact that
the location of the Tlx locus is evolutionary conserved
indicates the importance of the relative position of coding
sequences and controlling elements within functional gene
clusters.
Considering: (1) the similarity between sponge EmH-3
and Tlx homeodomain genes from other animals; (2) the
position of the Tlx locus in gene complexes; and (3) the
regulation of gene expression in proliferating and differentiating stem cells, we expected that some features of their
promoters would also be evolutionary conserved. The
activity of the EmH-3 proximal promoter was already
tested in the murine 3T3-cell line, using a reporter gene
strategy. Positive and negative regulatory regions were
identified, suggesting the existence of common transcriptional machinery in sponge and vertebrate cells (Coutinho
et al., 1998). This heterologous approach has been now used
to test the activity of regions of the sponge EmH-3 promoter
in two cell lines: a human erythroleukemia cell line K562
(Lozzio and Lozzio, 1975) and Fedora, a murine myeloid
progenitor cell line that can be induced to differentiate into
granulocytes. K562 cell line was chosen because: (1) it
represents a proliferating blood cell progenitor that can
express characteristics of the fetal erythropoiesis (Miller
et al., 1984; Sabath et al., 1998); (2) it can be induced to
differentiate in vitro (Andersson et al., 1979); and (3) it
expresses the endogenous Hox11 gene (Brake et al., 1998).
We assessed whether the differentiation of the blood cell
progenitor that is potentially controlled by the endogenous
Hox11 gene can influence the activity of the sponge EmH-3
promoter, in which we now found conserved putative
binding domains when compared to the mammalian
promoter region. As a control, we monitored the activity
of the EmH-3 promoter in the Fedora cell line, which does
not express the Hox11 gene, can be induced with retinoic
acid to differentiate into granulocytes, and represents a later
progenitor stage in the blood cell differentiation cascade.
We found that sponge reporter constructs and mammalian
homologue genes were regulated in the same manner in
cultured cells, indicating deep structural and functional
homologies among sponge and mammalian homeobox
genes.
2. Results
2.1. Phylogenetic analysis of sponge homeobox genes
The aims of the initial comparative sequence analysis
were to identify the orthologous homeobox genes present in
sponges, and to determine which of these genes are present
in higher metazoans and situated within murine, human or
Drosophila NKL/93DE-type complexes. Homeodomains of
C.C. Coutinho et al. / Mechanisms of Development 120 (2003) 429–440
genes that are found in the NKL complex, complete sponge
homeodomain sequences, and those that have been already
reported to be closely related to at least one of the known
sponge homeodomains were extracted from the NCBI
GenBank.
Sequences were aligned and a dendrogram was built by
the neighbor-joining procedure (Saitou and Nei, 1987) (Fig.
1). The branches of the dendrogram indicate distinct groups
of genes, some of which contain sponge genes. The sponge
prox1 and prox3 genes were grouped with genes of the NK3
and Msh (Msx) families, respectively, as previously
431
described by Seimiya et al. (1994). EmH-3 clustered with
the Tlx genes with a low bootstrap support (25), leaving
some doubt as to whether EmH-3 is a member of the Tlx or
Lbx homeobox gene family. Since parsimony is more
suitable for the analysis of smaller groups of closely related
sequences (Mount, 2000), we have used it to assess the
position of EmH-3. Parsimony analysis resolved the
position of EmH-3 in the Tlx group, supported by the
bootstrap value of 86 (Fig. 2).
An additional argument supporting this interpretation
comes from the alignment of protein sequences closely
Fig. 1. Neighbor-joining tree constructed from the alignment of 60 amino acids of the studied homeodomains. Bootstrap values are indicated above each
branch. The scale bar indicates the number of amino-acid substitutions per position in the sequence. The underlined names indicate homeobox genes from
sponges. The homeobox gene families are indicated at the right side.
432
C.C. Coutinho et al. / Mechanisms of Development 120 (2003) 429–440
Fig. 2. Maximum Parsimony tree constructed from the alignment of sequences presented in the Hox 11/Tlx and Lbx group homeodomains. Bootstrap values are
indicated at each branch. The underlined names denote homeobox genes from sponges. The homeobox gene families are indicated at the right side.
related to the predicted protein encoded by the sponge EmH3 gene, which reveals the presence of a conserved Eh1
repressor domain. Eh1 has been previously described in
engrailed, goosecoid and in gene families of the NKL
complex. It is thought to act as a protein-protein binding
domain mediating transcriptional repression of target genes
(Smith and Jaynes, 1996). Among the eight amino acids of
the Eh1 domains of EmH-3 and human Hox11 proteins,
60% are identical and 12% are conservative substitutions
(Fig. 3).
2.2. Promoter structural analysis
Since Tlx and EmH-3 form a group of closely related
genes, we have searched for evidence of conserved
promoter regions. DNA sequence of an approximately 1
Kb region, upstream of the EmH-3 translation-initiation site,
Fig. 3. N-terminal regions of the Hox11 protein and the homologues. The
homologous regions chosen by the MACAW program for the Eh1 region
are indicated in bold. The Hox11 homologues are prox2 (Ephydatia
fluviatilis), EmH-3 (Ephydatia muelleri), C15 (Drosophila melanogaster),
Tlx-3 (Mus musculus), Hox11 (Homo sapiens), Hox11L1 (Homo sapiens),
and Hox11L2 (Homo sapiens).
was aligned with the corresponding region of Tlx genes from
the mouse, human and Drosophila genomes. Fig. 4 shows
the position of putative elements that were identified from
the aligned promoter sequences. The Drosophila C15
promoter has the conserved elements shifted compared to
the others, suggesting a deletion in the 30 region that
occurred in the insect branch of evolution. The identified
promoter elements are putative binding sites for transcription factors of the following families: TCF-1, CAAT
binding protein, LMO2, USF-1 and Ikaros. Although a gel
shift analysis should be undertaken in order to asses the
affinity of association of these nuclear factors for the
putative binding sites, conservation of relatively long
sequences among distant animal groups is highly suggestive
of their functional relevance in control of gene expression.
2.3. Promoter functional analysis
To extend the analysis of the Tlx promoters, we
undertook a comparative functional analysis of the regulation of Tlx expression. We determined whether the EmH-3
promoter is modulated by the state of differentiation of
K562 cells that express the endogenous Hox11 gene (Brake
et al., 1998). As a control, expression of the EmH-3
promoter construct was assayed in mouse Fedora cells that
do not express the endogenous Hox11 gene (Fig. 5). As
expected, the K562 and Fedora cells differentiated when
exposed to butyrate or retinoic acid, respectively (Andersson et al., 1979). Following treatment, the K562 cells
decreased their rate of proliferation and enlarged. The
cytoplasm became more eosinophilic and contained
inclusions. In retinoic acid-treated Fedora cells, the major
C.C. Coutinho et al. / Mechanisms of Development 120 (2003) 429–440
433
Fig. 4. Putative conserved binding elements for transcription factors in Hox11 homologue gene promoters. Alignment of the 50 region upstream of the
translation-initiation codon of EmH-3 (sponge), prox2 (sponge), C15 (Drosophila), Hox11 (mouse), Hox11 (human) genes. The colored fonts indicate elements
recognized in all the promoters of this gene family: TCF-1 (green), LMO2 (blue), CCAAT (yellow), USF-1 (red), IK-2 (indigo). The flags indicate the
beginning of the EmH-3 promoter/luciferase constructs, described in the Fig. 8.
morphological change was the formation of ring-form
nuclei, typical of murine granulocyte progenitors (Fig. 6).
Semi-quantitative RT-PCR analysis confirmed that
undifferentiated proliferating K562 cells, but not differentiated ones, expressed the endogenous Hox11 gene (Brake
et al., 1998). Hox11 expression was undetectable in K562
cells treated with 1 mM sodium butyrate, which had no
effect on transcription of the constitutively expressed gapdh
gene (Fig. 7). The 1 mM concentration of butyrate was also
required for induction of the decrease of K562 cell growth
434
C.C. Coutinho et al. / Mechanisms of Development 120 (2003) 429–440
Fig. 5. RT-PCR for Hox11 on Fedora cell line. The GAPDH cDNA
corresponds to the 571 bp band, and Hox11 cDNA to the lower 460 bp band.
Lane 1 corresponds to Molecular Weight Standard – 1 kb DNA ladder
(GIBCO-BRL). Lane 2 contains RT-PCR amplified fragment correspondent to Fedora GAPDH, but not from Hox11 gene, despite the use of
primers for Hox11 cDNA. Lane 3 contains RT-PCR amplified fragments
correspondent to K562 GAPDH and Hox11 genes.
(data not shown), in accordance with the previously
published data (Andersson et al., 1979).
Since K562 cells responded to butyrate induction and
possessed the necessary components to regulate expression
of Hox11, the response of the sponge promoter was assayed
using this heterologous model. Different regions of the
EmH-3 promoter were fused with the luciferase gene of the
pGL2 vector (Fig. 8). The constructs were transiently
transfected into K562 cells and into control Fedora cells
under two different growth conditions: with and without
induction of differentiation. Enzymatic detection of luciferase activity was measured in extracts prepared from cells
under each experimental condition. Luciferase activities
dropped almost to zero when K562 cells were grown under
butyrate induction. However, no change in gene expression
was observed in similarly treated cells transfected with a
positive control that expressed the luciferase gene constitutively (Fig. 9). The modification of EmH-3 gene
expression induced by differentiation was much less
pronounced in Fedora cells. The greatest reduction of
luciferase activity following the retinoic acid treatment was
observed in cells containing a 323 bp promoter fragment,
and was only half of the control value (Fig. 10). We
conclude that the sponge EmH-3 promoter is regulated in a
similar manner to the corresponding endogenous gene, but
only in a cell line that shows expression of the mammalian
orthologue.
3. Discussion
The present study was done to shed new light on the
relationship of a group of sponge homeobox genes with
similar genes in higher animals. The neighbor-joining
followed by the maximum parsimony analysis indicated
that the sponge EmH-3 homeobox gene belonged to the
Hox11/Tlx gene family. The presence of the Eh1 repressor
domain on EmH-3 further supports that the EmH-3 gene is
homologous to the ancestral NKL complex. Additional
support for this classification was obtained by the
examination of the EmH-3 promoter sequence, in which
putative binding sites for a number of transcription factors
were found to co-localize with those found in Drosophila,
mouse and human Tlx promoter regions.
Pollard and Holland (2000) hypothesized that the
ancestor of the Drosophila/murine/human NKL cluster
contained the following loci: Msx, NK6, Hmx, Emx, Vax,
NK1, Tlx, Lbx, NK3 and NK4. Some of these loci, namely
Msh, NK3 and Tlx, have sponge homologues. We propose
that these genes belong to an ancient group from which the
NKL complex evolved (Fig. 11). Since none of the sponge
homeobox gene families are found in non-metazoans,
results of the comparative analysis suggest that the Msh,
NK3, and Tlx genes appeared in early steps of the Metazoa
evolution, more than one billion years ago. Alternatively,
but less probable, these gene families could have existed in
other groups that do not contain them now (prokaryotes,
fungi, protozoans and plants), and were subsequently lost.
The structural analogies suggest a common function, and
we have tested this hypothesis by transfecting the reportercontaining constructs of the sponge EmH-3 promoter into
mammalian cells. Not only was the sponge promoter
operational in the human intracellular environment, but it
was also expressed in a coordinate pattern with the
endogenously expressed Hox11 gene. Members of the Tlx
homeobox gene family appear to be highly expressed in
proliferating progenitor cells, and down-regulated when
cells differentiate. This is reminiscent of the previously
observed expression of EmH-3 gene in sponge archaeocytes
(Richelle-Maurer and Van de Vyver, 1999). Thus, this
pattern of regulated gene expression represents a deep
homology bridging the gap between the most primitive and
most complex multicellular organisms.
We have previously reported the presence of conserved
small regions within the EmH-3 and prox2 promoters,
suggesting that they are putative upstream promoter
elements for transcriptional control (PUPE) (Coutinho
et al., 1998). We have now extended this analysis to include
the corresponding promoter regions of related genes from
phylogenetically distant animals. Conservation of the overall nucleotide sequences was not detected. Nevertheless, the
search for PUPEs revealed the presence of distinct elements
at similar positions and in the same order in the promoters of
the sponge, Drosophila, mouse and human Tlx genes. The
distance among these elements is approximately 50– 100 bp
and they are positioned in the following sequence: Tcf1,
Lmo2, CAAT, USF and IK. All of these elements
participate in the control of proliferation and differentiation
of mesenchymal cell types, and in particular in hematopoi-
C.C. Coutinho et al. / Mechanisms of Development 120 (2003) 429–440
435
Fig. 6. Morphological changes in differentiation induced K562 and Fedora cell lines. (A, C) K562 and Fedora cells grown under standard culture conditions,
respectively. (B) K562 cells maintained for 72 h in the standard medium supplemented with 5 mM sodium butyrate. (D) Fedora cells maintained for 7 days in
the standard medium supplemented with 1 mM retinoic acid. Original magnification ¼ 1000 £ .
Fig. 7. Semiquantitative RT-PCR analysis of the HOX11 mRNAs from
K562 cultures induced with different amounts of sodium butyrate. The
gapdh cDNA corresponds to the 571 bp band, and Hox11 cDNA to the
lower 460 bp band. Samples were divided in three groups (A–C) according
to the PCR cycles numbers (25, 30 and 35, respectively). Lanes 1 and 17 –
Molecular Weight Standard – Low DNA Mass Ladder (GIBCO-BRL).
Lanes 2, 7 and 12: cells maintained in the control culture medium without
sodium butyrate. Lanes 3, 8 and 13: cultures induced with 0.001 mM
sodium butyrate. Lanes 4, 9 and 14: cultures induced with 0.01 mM sodium
butyrate. Lanes 5, 10 and 15: cultures induced with 0.1 mM sodium
butyrate. Lanes 6, 11 and 16: cultures induced with 1 mM sodium butyrate.
Fig. 8. Structure of the EmH-3 promoter – luciferase constructs. Schematic
representation of the stepwise 50 deletions of the EmH-3 promoter region,
fused upstream of the luciferase reporter gene in the pGL2 plasmid. The
size of the promoter fragment is relative to the translation start site; the
upstream position of conserved putative binding sites for transcription
factors is indicated at the top of the figure.
436
C.C. Coutinho et al. / Mechanisms of Development 120 (2003) 429–440
Fig. 9. Relative luciferase activity in K562 cells. Cells maintained under
standard culture conditions (open bars) and in the presence of sodium
butyrate (full bars) were transfected with plasmids containing the constructs
of the EmH-3 promoter fragments/luciferase, described in the Fig. 8.
Results represent mean values of three experiments done in duplicate and
standard errors.
esis, including early yolk sac hematopoiesis represented by
the K562 cell line model (Wall et al., 1996; Sabath et al.,
1998). An increasing list of transcription factors are known
to bind to these PUPEs. The transcription factors are
frequently oncogenes. Loss-of-function mutations result in
leukemia, suggesting that they function as negative
regulators of sustained progenitor cell proliferation (Hawley
et al., 1997; Visvader et al., 1997; Brake et al., 1998;
Nichogiannopoulou et al., 1999).
One of the caveats of multicellular organization is that
cells have to coordinate the intense progenitor proliferation
that generates the required cell mass with the partial or
terminal differentiation that generates diverse cell populations with specialized functions. In early mouse development, cells express only low levels of homeobox genes that
become highly expressed with the onset of hematopoiesis,
remain high in hematopoietic stem cells in both embryonic
and adult stages of hematopoiesis, but are down-regulated in
maturing blood cells (Pineault et al., 2002). In sponges, no
information on gene expression pattern in embryos is
available. Gemmules are hibernating asexual reproduction
bodies, containing only archaeocytes that are resting
totipotent cells. An increase in EmH-3 is required for
hatching of gemmules, when archaeocytes are activated to
proliferate and subsequently engage into differentiation of a
functional aquiferous system (Richelle-Maurer and Van de
Vyver, 1999). An early decrease of EmH-3 expression in
hatching gemmules, induced by retinoic acid, disturbs
reversibly the normal sponge development (Nikko et al.,
2001). We understand that the ordered activation and
downregulation of homeobox gene expression, similar to
that observed in mouse hematopoiesis, is required for the
normal sponge development.
It should be noted that the K562 cells used in this study
represent embryonic blood cell precursors, with molecular
characteristics of the yolk sac stage of hematopoiesis (Miller
et al., 1984). Besides the extraembryonic erythropoiesis,
yolk sac produces primitive macrophages, whose destiny is
not well known, but which have been proposed to be one of
the origins of resident tissue macrophages of the adult
(Yamashita, 1996). The analogy of macrophages and
sponge archaeocytes has been underlined in the seminal
study of cytology and evolution by Willmer (1970). We
have already argued that archaeocytes normally function as
active macrophages in the adult sponge, but can acquire the
function of stem cells in repair and regeneration, in
gametogenesis, or under stressful conditions (Borojevic,
1966, 1970). The molecular mechanisms that govern the
choice between proliferation and differentiation have been
apparently created during the initial major step of development of multicellularity, and remarkably conserved during
evolution.
Taken together, our study indicates that fundamental
regulatory events required for an operational multicellular
organization have appeared very early in evolution. The
nuclear factors that govern cell proliferation and differentiation have been sufficiently conserved to be functional on
the sponge promoter in the environment of a human cell
line.
4. Material and methods
4.1. Homeodomain sequence analysis
Fig. 10. Relative luciferase activity in Fedora cells. Cells maintained under
standard culture conditions (open bars) and in the presence of retinoic acid
(full bars) were transfected with plasmids containing the constructs of the
EmH-3 promoter fragments/luciferase, described in the Fig. 8. Results
represent mean values of three experiments done in duplicate and standard
errors.
Homeodomain sequences closely related to at least one
of the known sponge homeobox genes and the homeodomains from families present in NKL complex were
searched by the ‘tblastn’ option in the www BLAST server
(www.ncbi.nlm.nih.gov). The POU and Pax families were
not included in this analysis. Alignment of homeodomain
sequences was done using CLUSTAL W program as
Multiple Sequence Alignment method (Thompson et al.,
C.C. Coutinho et al. / Mechanisms of Development 120 (2003) 429–440
437
Fig. 11. Sponge genes homologous to the genes of the NKL/93DE complex. The homeobox gene families that form the NKL complex are indicated by different
shades, and their position is compared in mouse and in the 93DE complex of Drosophila. At present, three members of these gene families have been described
in sponges.
1994). Neighbor-Joining and Maximum Parsimony trees
were computed using the MEGA2 software distributed at
http://www.megasoftware.net. Bootstrap values have been
calculated from 500 replicates.
Homeodomain sequences used in comparisons analyses
are listed below, with accession numbers:
Amphivent B. floridae (Chordata), AAK58840; bagpipe
D. melanogaster (Arthropoda), P22809; BarH1 D. melanogaster (Arthropoda) B39369, BarH2 D. melanogaster
(Arthropoda), A41726; BarX1 H. sapiens (Chordata),
AAG23738; Cnox3 C. viridissima (Cnidaria), S20894;
Bsh D. melanogaster (Arthropoda), Q04787; C15 D.
melanogaster (Arthropoda), AAF55898; CEH-19 C. elegans (Nematoda), P26797; CEH-22 C. elegans (Nematoda)
P41936; CEH-9 C. elegans (Nematoda), P56407;
Csx/Nkx2-5 M. musculus (Chordata), AAG38875; distalless (Dll) D. melanogaster (Arthropoda), NP_523857; Dth1 D. tigrina (Platyhelminthes), S33701; Dth-2 D. tigrina
(Platyhelminthes), S33702; EmH-3 E. muelleri (Porifera),
AAC18965; Emx1 D. rerio (Chordata), BAA06912; Emx1
H. sapiens (Chordata), CAA48750; Emx2 D. rerio
(Chordata), BAA06913; Emx2 M. musculus (Chordata),
CAA48753; Emx2 H. sapiens (Chordata), CAA48751;
GBX-1 H. sapiens (Chordata), Q14549; GBX-2 H. sapiens
(Chordata), P52951; GHOX-7 G. gallus (Chordata),
P50223; H2.0 D. melanogaster (Arthropoda), P10035;
Hmx3 M. musculus (Chordata) NP_032283; Hmx D.
melanogaster (Arthropoda), AAF55433; Hox11 H. sapiens
(Chordata), A40855; Hox11L1 M. musculus (Chordata),
NP_033418; Hox-11L1 H. sapiens (Chordata), O43763;
ladybird early D. melanogaster (Arthropoda), CAA70056;
ladybird late D. melanogaster (Arthropoda), CAA70057;
Lbx D. rerio (Chordata), CAC15184; Lbx2 M. musculus
(Chordata), NP_034822; Msh H. vulgaris (Cnidaria),
CAB88390; Msh-2 D. melanogaster (Arthropoda),
A43561; Msh-like 3 M. musculus (Chordata),
NP_034966, Msx1 H. sapiens (Chordata), AAL17870;
Msx-2 M. musculus, (Chordata), Q03358; Nk-1 D.
melanogaster (Arthropoda), P22807; Nk2.1a D. rerio
(Chordata), AAF78912; NK2.1b D. rerio (Chordata),
AAK01120; Nkx-2.2 M. musculus (Chordata),
NP_035049; Nkx-2.3 D. rerio (Chordata), AAC05228;
Nkx-2.3 M. musculus (Chordata), P97334; Nkx3-1 M.
musculus (Chordata), NP_035051; Nkx6-1 R. norvegicus
(Chordata), NP_113925.1; Nkx-6.1 H. sapiens (Chordata)
P78426; Nkx6A H. sapiens (Chordata), XP_003468; NkxD
S. raphanus (Porifera), AAG28516; OM(1D) D. ananassae
(Arthropoda) P22544, Prox1 E. fluviatilis (Porifera),
AAA20149; Prox3 E. fluviatilis (Porifera), AAA20151;
Sax1 (Nk-1) M. musculus (Chordata), P42580; Sax2 M.
musculus (Chordata), AAB53323; Tin D. melanogaster
(Arthropoda), AAF55890; TTF-1 H. sapiens (Chordata),
P43699; Vnd D. melanogaster (Arthropoda), P22808;
XHOX-7.1(Xmsh) Xenopus laevis (Chordata), P35993;
Tlx1 M. musculus (Chordata), NP_068701; TlxA D. rerio
(Chordata), AAK98790; Tlx-3(Hox11L2) G. gallus,
AAC23901; Vax G. gallus (Chordata), BAA84282; Vax1
R. norvegicus (Chordata), NP_072158; Vax1 M. musculus
(Chordata), NP_033527; Vax2 M. musculus (Chordata),
NP_036042; Vax3 X. laevis (Chordata), AAF25692;
VENTX2 H. sapiens (Chordata), AAK83043; Vox-15 X.
laevis (Chordata), AAC59910; XHox11 X. laevis (Chordata), AAG14453; XHox11L2 X. laevis (Chordata),
AAG14452; Xom X. laevis (Chordata), CAA67093;
Xvent-1 X. laevis (Chordata), CAA63437. Xvent-2 X.
laevis (Chordata), CAA67354.
4.2. Eh1 domain
The assessment of the conserved domains among
438
C.C. Coutinho et al. / Mechanisms of Development 120 (2003) 429–440
different proteins containing homeodomains from Porifera
to Mammalia was done using MACAW software (Multiple
Alignment Construction & Analysis Workbench, Version
2.05, Jan. 20. 1995, National Center for Biotechnology
Information (NCBI, Greg Schuler).
4.3. Promoter sequence analysis
The 1 kb DNA sequence upstream to the translation
initiation site of EmH-3 and the plasmid constructs with
EmH-3 promoter fragments of 2 538, 2 498, 2 401, 2 323,
2 279, 2 231, and 2 173 were previously described
(Coutinho et al., 1998). DNA sequences from human
(Hox11 – h), mouse (Hox11 – m), Drosophila (C15) and
sponge (EmH-3) promoters (1 Kb) were aligned using
CLUSTAL W software (Version 1.8) (Thompson et al.,
1994). Each promoter sequence was independently searched
for putative binding site elements using the default
adjustment of the TFSEARCH program (http://www.cbrc.
jp/research/db/TFSEARCH.html). This program correlated
sequence fragments against TFMATRIX transcription
factor binding sites in the ‘TRANSFAC’ databases (http://
transfac.gbf.de/TRANSFAC/) (Heinemeyer et al., 1998).
TRANSFAC covers the whole range of eukaryotic cisacting regulatory DNA elements and trans-acting factors
from yeast to humans. The putative binding elements of
each promoter were manually compared. The elements that
were located at the same position in the aligned promoters
of all species were pointed out and defined as putative
evolutionary conserved elements. Fig. 8 presents a correlation between the putative conserved elements of the EmH3 promoter and different fragments of the EmH-3 promoter
fused with the luciferase gene that were used in the present
study.
4.4. Semi-quantitative RT-PCR for Hox11 on normal and
induced K562 cells
Cultures of K562 cells were divided in five groups:
controls and cultures induced with 0.001, 0.01, 0.1 and 1
mM butyrate during 72 h. RNA was extracted by standard
protocols using Trizol Reagent (Gibco BRL, Gaithersbourgh, MD, USA). The total RNA was dissolved in water
and quantified. Five mg RNA were reverse-transcribed
into cDNA, and amplified by PCR with cycle numbers
increasing from 25 to 30 and 35. The glyceraldehyde
3-phosphate dehydrogenase (gapdh) primer sequences
were:
50 ATCACCATCTTCCAGGAGCG30
and
0
5 CCTGCTTCACCACCTTCTTG30 that amplify a 571 bp
fragment. Hox11 PCR primer sequences were
5 0 AACAACCTCACTGGCTCAC30
and
50 TGATTTTGGTCGAGTCGTCA30 that amplify a 460 bp
fragment. Routinely, the amplification was done using one
initial denaturation step (948C, 10 ), 35 rounds of amplification (948C, 10 /570 – 1/72 2 10 ) and a final extension step
(728C, 70 ). The products were separated in 2% agarose gel,
stained with ethidium bromide and visualized by ImageMaster VDS (Pharmacia Biotech, Uppsala, Sweden).
4.5. RT-PCR for Hox11 on native Fedora cell line
Fedora RNA was extracted by standard protocols using
Trizol Reagent (Gibco BRL, Gaithersbourgh, MD, USA).
The total RNA was dissolved in water, quantified and stored
at 2 208C. Five mg RNA were reverse-transcribed into
cDNA, and amplified by PCR with the same primers used
for K562 cell cDNA and the same amplification program.
4.6. Promoter activity assay in native and induced K562 and
Fedora cell lineages
K562, the human erythroleukemia cells and Fedora, the
murine granulocyte precursors, were obtained from the Rio
de Janeiro Cell Bank (PABCAM, Federal University, Rio de
Janeiro, RJ, Brazil). They were maintained in RPMI 1640
medium (SIGMA Chemical Company, St Louis, MO, USA)
with 10% fetal bovine serum (FBS, GIBCO), 0.1 mg/ml
penicillin and 100 U/ml streptomycin. The proliferating
undifferentiated K562 cells were used for transient transfection experiments and a fraction of cells was separated to be
induced to differentiate. The Fedora culture was divided in
two groups 7 days before transfection: the controls and the
cells induced to differentiate by addition of 1 mM of alltrans-retinoic acid (Sigma, Aldrich). Transient cotransfections in K562 and Fedora cells were performed by lipofectin
using different luciferase reporter constructs in association
with normalizing pCMV-b-galactosidase vector (Clontech,
Palo Alto, CA, USA). Two wells were used for every
sample in each experiment. Five mg of each plasmid and 10
ml lipofectin were dissolved in 100 ml medium each
(without FBS and antibiotics). Subsequently, the two
solutions were mixed and incubated for 30 min at room
temperature. The cell culture medium was replaced by 2 ml
medium without FBS and induction factors. After 30 min at
378C this medium was replaced by the DNA/lipofectin
mixture, which was further diluted with 0.8 ml medium.
Sixteen h later, the transfection medium was replaced with
normal medium with or without differentiation factor (5 mM
sodium butyrate for K562 and 2 mM all-trans-retinoic acid
for Fedora). After 24 h, the medium of differentiating K562
was replaced by RPMI supplemented with 10% FBS and 5
mM of sodium butyrate. After a further incubation of 48 h,
the cells were lysed using 250 ml reporter lysis buffer
(Promega Corporation, Madison, WI, USA). In order to
confirm the cell differentiation, aliquots were harvested
before and after the butyrate and retinoic acid induction,
cytosmears were prepared with a cytocentrifuge, stained by
the standard May-Grünwald and Giemsa solutions and
analyzed under microscope.
Luciferase activity was measured using the Luciferase
Assay system (Promega), in a luminometer (TD-20/20,
Turner Designs Instrument, Sunnyvale, CA, USA). B-
C.C. Coutinho et al. / Mechanisms of Development 120 (2003) 429–440
galactosidase activity was determined by spectrophotometry, using o-nitrophenyl-b-D -galactopyranoside as
substrate (Rocancourt et al., 1990). The luciferase activity
was calculated by dividing the relative light units (RLU)
value by the optical density (OD) of the b-Gal colorimetric
reaction (in order to normalize for transfection efficiency).
The activity was determined in relation to the expression of
the pGL-2 control plasmid (positive control) in cell extracts
normalized for b-galactosidase activity. One positive and
two negative controls were included: (1) transfection with
the pGL2-basic vector, containing the luciferase gene, but
lacking any promoter and enhancer; (2) transfection with the
pCMV-b-galactosidase plasmid only, and (3) transfection
with pGL2-control containing the SV40 promoter and
enhancer.
Acknowledgements
A publication of the Millennium Institute of Tissue
Bioengineering. This study has received support from CNPq
and FINEP grants of the Brazilian Ministry of Science and
Technology, FAPERJ grant of the Rio de Janeiro State
Government, and the Fundação Universitária José Bonifácio
(FUJB).
References
Andersson, L.C., Jokinen, M., Gahmberg, C.G., 1979. Induction of
erythroid differentiation in the human leukaemia cell line K562. Nature
278, 364 –365.
Blau, H.M., Brazelton, T.R., Weimann, J.M., 2001. The evolving concept
of a stem cell: entity or function? Cell 105, 829–841.
Borchellini, C., Boury-Esnault, N., Vacelet, J., Le Parco, Y., 1998.
Phylogenetic analysis of the Hsp70 sequences reveals the monophyly of
Metazoa and specific relationship between animals and fungi. Mol.
Biol. Evol. 15, 647–655.
Borchellini, C., Manuel, M., Alivon, E., Boury-Esnault, N., Vacelet, J., Le
Parco, Y., 2001. Sponge paraphyly and the origin of Metazoa. J. Evol.
Biol. 14, 171–179.
Borojevic, R., 1966. Etude expérimentale de la différentiation de cellules de
l’Eponge au cours de son développement. Dev. Biol. 14, 130– 153.
Borojevic, R., 1970. Différenciation cellulaire dans l’embryogenèse et la
morphogenèse chez les Spongiaires. In: Fry, W.G., (Ed.), The Biology
of the Porifera, Academic Press, London, pp. 467–490.
Brake, R.L., Kees, U.R., Watt, P.M., 1998. Multiple negative elements
contribute to repression of the HOX11 proto-oncogene. Oncogene 17,
1787–1795.
Coutinho, C.C., Vissers, S., Van de Vyver, G., 1994. Evidence of
homeobox genes in the fresh water sponge Ephydatia fluviatilis. In:
van Soest, K.B., (Ed.), Sponges in Time and Space, Balkema,
Rotterdam, Holland, pp. 385–388.
Coutinho, C.C., Seack, J., Van de Vyver, G., Borojevic, R., Muller, W.E.G.,
1998. Origin of the metazoan bodyplan: characterization and functional
testing of the promoter of the homeobox gene EmH-3 from the
freshwater sponge Ephydatia muelleri in mouse 3T3 cells. Biol. Chem.
379, 1243–1251.
Degnan, B.M., Degnan, S.M., Giusti, A., Morse, D.E., 1995. A hox/hom
homeobox gene in sponges. Gene 155, 175– 177.
439
Hadzi, J., 1953. An attempt to reconstruct the system of animal
classification. Syst. Zool. 2, 145 –154.
Hawley, R.G., Fong, A.Z., Lu, M., Hawley, T.S., 1994. The HOX11
homeobox-containing gene of human leukemia immortalizes murine
hematopoietic precursors. Oncogene 9, 1– 12.
Hawley, R.G., Fong, A.Z., Reis, M.D., Zhang, N., Lu, M., Hawley, T.S.,
1997. Transforming function of the HOX11/TCL3 homeobox gene.
Cancer Res. 57, 337–345.
Heinemeyer, T., Wingender, E., Reuter, I., Hermjakob, H., Kel, A.E., Kel,
O.V., Ignatieva, E.V., Ananko, E.A., Podkolodnaya, O.A., Kolpakov,
F.A., Podkolodny, N.L., Kolchanov, N.A., 1998. Databases on
transcriptional regulation: TRANSFAC, TRRD, and COMPEL.
Nucleic Acids Res. 26, 364–370.
Hoshiyama, D., Suga, H., Iwabe, N., Koyanagi, M., Nikoh, N., Kuma, K.,
Matsuda, F., Honjo, T., Miyata, T., 1998. Sponge Pax cDNA related to
Pax-2/5/8 and ancient gene duplications in the Pax family. J. Mol. Evol.
47, 640–648.
Hyman, L.H., 1940. The Invertebrates, 1. McGraw-Hill, New York.
Jagla, K., Bellard, M., Frasch, M., 2001. A cluster of Drosophila homeobox
genes involved in mesoderm differentiation programs. Bioessays 23,
125– 133.
Kawabe, T., Muslin, A.J., Korsmeyer, S.J., 1997. HOX11 interacts with
protein phosphatases PP2A and PP1 and disrupts a G2/M cell-cycle
checkpoint. Nature 385, 454 –458.
Kennedy, M.A., Gonzalez-Sarmiento, R., Kees, U.R., Lampert, F., Dear,
N., Boehm, T., Rabbitts, T.H., 1991. HOX11, a homeobox-containing
T-cell oncogene on human chromosome 10q24. Proc. Natl. Acad. Sci.
USA 88, 8900–8904.
Lozzio, C.B., Lozzio, B.B., 1975. Human chronic myelogenous leukemia
cell-line with positive Philadelphia chromosome. Blood 45, 321 –334.
Lu, M., Gong, Z.Y., Shen, W.F., Ho, A.D., 1991. The tcl-3 proto-oncogene
altered by chromosomal translocation in T-cell leukemia codes for a
homeobox protein. EMBO J. 10, 2905–2910.
Manuel, M., Le Parco, Y., 2000. Homeobox gene diversification in the
calcareous sponge, Sycon raphanus. Mol. Phylogen. Evol. 17, 97–107.
Miller, C.W., Young, K., Dumenil, D., Alter, B.P., Schofield, J.M., Bank,
A., 1984. Specific globin mRNAs in human erythroleukemia (K562)
cells. Blood 63, 195–200.
Mount, D.W., 2000. Phylogenetic prediction, Bioinformatics: Sequence
and Genome Analysis Cold Spring Harbor, New York, pp. 237–280,
(Chapter 6).
Müller, W.E.G., 1995. Molecular phylogeny of Metazoa (animals):
monophyletic origin. Naturwissenschaften 82, 321–329.
Nichogiannopoulou, A., Trevisan, M., Neben, S., Friedrich, C., Georgopoulos, K., 1999. Defects in hemopoietic stem cell activity in Ikaros
mutant mice. J. Exp. Med. 190, 1201–1214.
Nikko, E., Van de Vyver, G., Richelle-Maurer, E., 2001. Retinoic acid
down-regulates the expression of EmH-3 homeobox-containing gene in
the freshwater sponge Ephydatia muelleri. Mech. Ageing Dev. 122,
779– 794.
Pineault, N., Helgason, C.D., Lawrence, J., Humphries, R.K., 2002.
Differential expression of Hox, Meis1, and Pbx1 genes in primitive cells
throughout murine hematopoietic ontogeny. Exp. Hematol. 30, 49– 57.
Pollard, S.L., Holland, P.W., 2000. Evidence for 14 homeobox gene
clusters in human genome ancestry. Curr. Biol. 10, 1059–1062.
Richelle-Maurer, E., Van de Vyver, G., 1999. Temporal and spatial
expression of EmH-3, a homeobox-containing gene isolated from the
freshwater sponge Ephydatia muelleri. Mech. Ageing Dev. 109,
203– 219.
Richelle-Maurer, E.G., Van de Vyver, G., Vossers, S., Coutinho, C.C.,
1998. Homeobox-containing genes in freshwater sponges: characterization, expression, and phylogeny. In: Muller, W.E.G., (Ed.), Molecular
Evolution: Evidence for Monophyly of Metazoa; Progress in Molecular
and Subcellular Biology 19, Springer-Verlag, Berlin – Heidelberg, pp.
157– 175.
Rocancourt, D., Bonnerot, C., Jouin, H., Emerman, M., Nicolas, J.F., 1990.
Activation of a beta-galactosidase recombinant provirus: application to
440
C.C. Coutinho et al. / Mechanisms of Development 120 (2003) 429–440
titration of human immunodeficiency virus (HIV) and HIV-infected
cells. J. Virol. 64, 2660– 2668.
Sabath, D.E., Koehler, K.M., Yang, W.Q., Phan, V., Wilson, J., 1998.
DNA-protein interactions in the proximal zeta-globin promoter:
identification of novel CCACCC- and CCAAT-binding proteins.
Blood Cells Mol. Dis. 24, 183 –198.
Saitou, N., Nei, M., 1987. The Neighbor joining method: a new method of
reconstructing phylogenetic trees. Mol. Biol. Evol. 4, 406–425.
Seimiya, M., Ishiguro, H., Miura, K., Watanabe, Y., Kurosawa, Y., 1994.
Homeobox-containing genes in the most primitive metazoa, the
sponges. Eur. J. Biochem. 221, 219–225.
Seimiya, M., Watanabe, Y., Kurosawa, Y., 1997. Identification of POUclass homeobox genes in a freshwater sponge and the specific
expression of these genes during differentiation. Eur. J. Biochem.
243, 27–31.
Shenk, M.A., Steele, R.E., 1993. A molecular snapshot of the metazoan
‘Eve’. T. Biochem. Sci. 18, 459–463.
Smith, S.T., Jaynes, J.B., 1996. A conserved region of engrailed, shared
among all en-, gsc-, Nk1-, Nk2-, and msh-class homeoproteins,
mediates active transcriptional repression in vivo. Development 122,
3141–3150.
Thompson, J.D., Higgins, D.G., Gibson, T.J., 1994. CLUSTAL W:
improving the sensitivity of progressive multiple sequence alignment
through sequence weighting, position-specific gap penalties and weight
matrix choice. Nucleic Acids Res. Nov 11 22 (22), 4673–4680.
Visvader, J.E., Mao, X., Fujiwara, Y., Hahm, K., Orkin, S.H., 1997. The
LIM-domain binding protein Ldb1 and its partner LMO2 act as negative
regulators of erythroid differentiation. Proc. Natl. Acad. Sci. USA 94,
13707–13712.
Wainright, P.O., Hikle, G., Sogin, M.L., Stickel, S.K., 1993. Monophyletic
origins of the metazoan: an evolutionary link with fungi. Science 260,
340 –342.
Wall, L., Destroismaisons, N., Delvoye, N., Guy, L.G., 1996. CAAT/
enhancer-binding proteins are involved in beta-globin gene expression
and are differentially expressed in murine erythroleukemia and K562
cells. J. Biol. Chem. 271, 16477–16484.
Willmer, E.N., 1970. Cytology and Evolution, Academic Press, New York
– London.
Yamashita, A., 1996. Role of yolk sac endodermal cells with special
reference to the fetal macrophage differentiation. J. Leukoc. Biol. 59,
139 –144.
Zrzavy, J., Mihulka, S., Kepka, P., Bezdek, A., 1998. Phylogeny of the
Metazoa based on morphological and 18S ribosomal DNA evidence.
Cladistics 14, 49–285.