Genome-wide analysis of the sox family in the calcareous sponge Sycon ciliatum: multiple genes with unique expression patterns
- First Online:
- 7.3k Downloads
Sox genes are HMG-domain containing transcription factors with important roles in developmental processes in animals; many of them appear to have conserved functions among eumetazoans. Demosponges have fewer Sox genes than eumetazoans, but their roles remain unclear. The aim of this study is to gain insight into the early evolutionary history of the Sox gene family by identification and expression analysis of Sox genes in the calcareous sponge Sycon ciliatum.
Calcaronean Sox related sequences were retrieved by searching recently generated genomic and transcriptome sequence resources and analyzed using variety of phylogenetic methods and identification of conserved motifs. Expression was studied by whole mount in situ hybridization.
We have identified seven Sox genes and four Sox-related genes in the complete genome of Sycon ciliatum. Phylogenetic and conserved motif analyses showed that five of Sycon Sox genes represent groups B, C, E, and F present in cnidarians and bilaterians. Two additional genes are classified as Sox genes but cannot be assigned to specific subfamilies, and four genes are more similar to Sox genes than to other HMG-containing genes. Thus, the repertoire of Sox genes is larger in this representative of calcareous sponges than in the demosponge Amphimedon queenslandica. It remains unclear whether this is due to the expansion of the gene family in Sycon or a secondary reduction in the Amphimedon genome. In situ hybridization of Sycon Sox genes revealed a variety of expression patterns during embryogenesis and in specific cell types of adult sponges.
In this study, we describe a large family of Sox genes in Sycon ciliatum with dynamic expression patterns, indicating that Sox genes are regulators in development and cell type determination in sponges, as observed in higher animals. The revealed differences between demosponge and calcisponge Sox genes repertoire highlight the need to utilize models representing different sponge lineages to describe sponge development, a prerequisite for deciphering evolution of metazoan developmental mechanisms.
The Sox genes (Sry related high mobility group, HMG box) are a family of transcription factors with important roles in regulating development and cell fate determination throughout the animal kingdom[1, 2]. The Sox proteins are characterized by the HMG DNA binding domain of 79 amino acids, resembling the mammalian testis determination factor, Sry, which was the first Sox domain identified. There are 20 Sox genes in mammals which have been classified in five groups of Sox proteins (B, C, D, E, and F). However, additional groups have been created to accommodate divergent genes with limited taxonomic distribution, for instance group J. Groups B, C, E, and F are found in all eumetazoan lineages, but group D is found only in the bilaterians.
No Sox genes are present in the sequenced genomes of the unicellular choanoflagellate, Monosiga brevicollis, or the amoeboid holozoan Capsaspora owczarzaki. Since they are present in basal metazoans like sponges (that is, the demosponge Amphimedon queenslandica)[8, 9] and placozoans (Trichoplax adhaerens), they have likely arisen in the last common ancestor to the Metazoa. There is a larger repertoire of Sox genes in cnidarians[11, 12, 13] and the ctenophore Pleurobrachia pileus than in the demosponges[8, 9, 15] and the placozoans. Previous phylogenetic analysis of cnidarian Sox genes including the species Hydra magnipapillata, Nematostella vectensis, and Clytia hemisphaerica placed some of these sequences into the previously identified groups of Sox genes; however some of these genes cannot be classified into any specific group[11, 12, 13]. The expression patterns of cnidarian Sox genes suggest that they have roles in a wide variety of developmental functions, such as germ layer formation, organ development, cell type specification, and neural development[11, 12, 13].
Previous studies on Sox genes in sponges include the two demosponges, Amphimedon queenslandica[8, 9] and Ephydatia muelleri, as well as the calcareous sponge Sycon raphanus. In Amphimedon, four Sox genes have been found, including two members of group B (AmqSoxB1 and AmqSoxB2) and single members of groups C and F. Sox genes from Ephydatia and Sycon raphanus could not be clearly classified due to incomplete domain sequences included in the phylogenetic analyses. As a consequence, the complement of Sox genes in calcareous sponges is still unclear. In addition, apart from an RT-PCR study suggesting dynamic expression of Sox genes during embryonic development in Amphimedon, no expression patterns on a cellular level are published for this or any other sponge. For this reason, more studies in sponges are required to fully understand the function of Sox genes in the phylum Porifera in comparison with the Eumetazoa. The aims of this study were to analyze the repertoire of Sox genes in the calcareous sponge Sycon ciliatum and to trace their expression during development.
Recently generated complete draft genomic sequence and extensive transcriptome resources allow us to perform whole-genome analysis of developmentally important gene families (Adamski et al., unpublished work), and established in situ hybridization protocols allow for studies of gene expression in all life stages.
Phylogenetic analysis of sycon Sox genes
Internal ID number of retrieved sequence
First hit on NCBI BlastX search
Name given after phylogenetic analysis
Accession numberSycon andLeucosolenia
SoxpB Acropora millepora
Sox21B Danio rerio
Sox8 Oncorhynchus keta
Sox18 Xenopus Silurana
HMG box Brugia malayi
Sox8 Gallus gallus
Sox17 Homo sapiens
HMG Brugia malayi
Sox13 Takifugu rubripes
Syr-box 32 Oreochromis niloticus
Sox8 Homo sapiens
Sox14 Danio rerio
Syr 9 Monodelphis domestica
SoxF Acropora millepora
SoxBb Acropora millepora
SoxF Lethenteron camtschaticum
Sox13 Ixodes scapularis
Sox similar protein Suberites domunluca
Summary ofSyconSox and SoxL genes expression
Oocytes, cleavage stage embryos, macromeres, and cruciform cells
Oocytes, cleavage stage embryos, macromeres
Choanocytes and some mesohyl cells
Choanocytes and accessory cells, some mesohyl cells
Large spindle-shaped cells around osculum
Choanocytes, pinacocytes, small cells around osculum
Ubiquitous during embryogenesis, choanocytes
Oocytes, cleavage stage embryos, cruciform cells, choanocytes
Oocytes, cleavage stage embryos, macromeres, choanocytes, small cells around osculum
We have performed phylogenetic analyses of HMG domain sequences of Sox genes using different combination of taxa and the 12 sequences of Sycon (data not shown). In the initial phylogenetic analysis, most of Sycon Sox genes did not clearly fall into the recognized Sox groups (data not shown). To test whether adding sequences from another sponge closely related to Sycon would help to resolve the phylogenetic tree, we additionally identified and included sequences of Sox genes from another calcareous sponge, Leucosolenia complicata (Adamski et al., unpublished work). Up to date, we have recovered a total of seven Sox and Sox-related sequences from Leucosolenia (Table1).
Another phylogenetic tree was then constructed including the entire repertoire of identified Sycon and Leucosolenia Sox genes (Additional file1). However, this phylogenetic analysis also resulted in a non-resolved tree with multiple long-branch attraction artefacts. To reduce long-branch attraction, the most divergent sequences from both Sycon and Leucosolenia were excluded from further analyses. The excluded genes resemble Sox genes but have a divergent conserved motif within the HMG domain: either at the motif RPMNAF (positions 5 to 10), and/or at YK/R (positions 70 to 72); we named them Sox-like genes: SciSoxL1 to SciSoxL4a/b and LcoSoxL1, LcoSoxL4a/b (Table1, Additional file2).
Notably, our analysis did not reveal orthologous relationships between Amphimedon and calcaronean sequences even in cases where members of the same subfamily are present in both sponges, such as SoxB or SoxC. As reported by Larroux and colleagues the Amphimedon SoxF gene did not cluster with other SoxF sequences in the maximum likelihood analysis. However, conserved motif analysis (see below) indicates that this gene belongs to the SoxF subfamily.
The remaining two Sycon Sox genes named SciSox6 and SciSox7 (Table1) did not fall into any known Sox group, while clustering within the Sox family (Figure2). One ortholog of SciSox6 was found in Leucosolenia, and it was named LcoSox6. In contrast, we have not found a counterpart of SciSox7 in Leucosolenia.
Motif conservation within sponge Sox genes
We compared full length Sox proteins from Sycon, Leucosolenia, and Amphimedon with their homologs from different taxa (Figure2, Additional file4) to find conserved motifs outside the HMG domain. The analysis revealed the presence of a number of motifs that are conserved between the eumetazoan and poriferan sequences. However, the motifs in sponge sequences were often quite divergent as compared to their bilaterian and cnidarian counterparts (Figure2, Additional file4). Amphimedon SoxB1 and B2, Leucosolenia SoxB, and Sycon SoxB genes contained the B-group specific motif. In contrast to the eumetazoan SoxB proteins, the B-group specific motif in sponges was not located directly next to the C-terminal of the HMG domain, but appeared in different positions within the C-terminal part of the protein. Both Amphimedon and Sycon SoxC proteins contained a slightly divergent C-group motif as compared to Homo and Acropora SoxC. Two conserved regions were found for the Sycon SoxE protein while only one region was found in Leucosolenia SoxE. Finally, the conserved short SoxF motif was also found in the three sponge proteins, but was located closer to the HMG domain, while in Acropora and Homo it is located at the C-terminal of the protein.
Sox genes are dynamically expressed during embryogenesis and cell differentiation
We have studied expression of Sox genes in adult sponges containing a wide variety of embryonic stages by whole mount in situ hybridization. Except for SciSoxL3 and SciSoxL4a/b, for which we could not amplify probes suggesting they are not significantly expressed in adult cells or during embryogenesis, all other genes displayed unique patterns during development and/or in adult cells.
The Sox gene family is significantly larger in Sycon than in Amphimedon
As previously reported by Larroux et al.[8, 9]Amphimedon has four Sox genes corresponding to groups B, C, and F. In the demosponge Ephydatia muelleri only three Sox genes could be identified. In contrast, the genome of the calcareous sponge Sycon ciliatum contains seven Sox genes and four additional Sox-related genes.
In Sycon, five Sox genes correspond to the recognized Sox subfamilies, confirming the presence of Sox genes of the groups B, C, and F in sponges, and adding SoxE to the sponge repertoire. While bootstrap support and posterior probabilities values for assigning the poriferan sequences into eumetazoan subfamilies are generally low, analysis of conserved motifs within the full length proteins consistently confirmed placement of the calcaronean sequences within the recognized subfamilies.
There are several differences between the demosponge and calcaronean Sox genes as evidenced by the comparison between Amphimedon and Sycon. For example, there is only one SoxB gene in Sycon. In contrast, the calcaronean sequences can be classified as belonging to SoxE and SoxF families; while only a single (and difficult to place in phylogenetic analysis) SoxF gene is present in the Amphimedon genome. The Amphimedon SoxE gene might have been lost, or SoxE genes might have evolved after demosponges diverged. It is impossible to differentiate between these two scenarios until the issue of sponge monophyly vs. paraphyly is resolved. On the other hand, our result indicates that SoxF genes in Sycon and Leucosolenia are likely to be a result of lineage-specific duplication.
Interestingly, the Amphimedon genome does not appear to contain the large number of Sox-related genes that we have identified in the two calcaronean genomes. It remains unclear whether this is a result of significant gene loss in Amphimedon, or rather of expansion of the Sox family in the Calcaronea. Only analysis of additional poriferan genomes representing a range of clades (especially homoscleromorphs, calcineans, and a range of demosponges) will help to shed light into this issue.
Dynamic expression of Sox genes in sycon
The expression patterns of Sycon Sox genes fall into two categories: embryonic (SciSoxB and SciSoxC) or predominantly in differentiated adult cells (SciSoxE, SciSoxF1, SciSoxF2, and SciSox6). Sox-like genes are expressed both during development and in adult tissues (Summary on Table2).
Until functional data are obtained in sponges, the specific roles of the identified genes will remain unclear. However, we can hypothesize on their putative function in Sycon and on hypothetical ancestral roles in the metazoan ancestor, by comparing the expression patterns of Sycon and the eumetazoan Sox genes. This is particularly tempting for genes belonging to subfamilies that appear to have a conserved function throughout the Eumetazoa, such as the SoxB group. At least one Sox gene belonging to Group B is expressed in the embryonic ectoderm and the neurogenic region of embryos in early development in most bilaterians (for a review see), cnidarians[12, 13], and in the ctenophore P. pileus.
Sycon SoxB expression is restricted to two cell types of the embryo, the macromeres and the cruciform cells. During settlement and metamorphosis, the macromeres become the outer cells of the post-larva and subsequently differentiate into exopinacocytes, the outer epithelium of the sponge[22, 23]. The SciSoxB expression in the macromeres provides support for the notion that the exopinacoderm of the sponges might be homologous to the ectoderm of higher metazoans.
The cruciform cells are characteristic cells of the calcaronean sponge larvae[19, 24]. They form from four cytoplasm regions segregated during cleavage and differentiate at the pre-inversion stage; they are present in the swimming larva, to later degenerate during settlement and metamorphosis. Their role is not yet clear, but these four cells are the only candidate cells suggested to play a role in larval photoreception. If the cruciform cells are indeed involved in photoreception, the SoxB expression during their differentiation would indicate conservation of SoxB functions in broadly defined neurogenesis and sensory organ formation.
The expression of Sycon SoxC is very prominent in macromeres during pre-inversion, while expression was not detected in larvae. In the cnidarians Acropora and Nematostella, SoxC is expressed during embryogenesis in cell types that are suspected to be sensory neurons[11, 12]. However in Clytia, SoxC (ChSox15) is expressed in stem cells. Therefore it appears that there is no clear conservation of expression pattern among these organisms.
While there is no strong conservation of expression for SoxE and SoxF genes, SoxE genes in bilaterian invertebrates tend to have a role in sex-specific aspects of gonad development, and SoxF genes tend to be associated with endoderm formation[21, 26]. In the cnidarians Nematostella and Acropora, SoxE and SoxF are expressed in endodermal lineages; while in Clytia SoxE is expressed in germline cells, stem cells, and nematoblasts, indicating once again no clear conservation among cnidarians within this group. However, expression in the endoderm (in Anthozoan cnidarians) and mesodermal derivatives (gonads) of bilaterians, together with the observed expression of Sycon SoxE and SoxF in choanocytes and some mesohyl cells, could be used to support a concept of homology of the choanoderm + mesohyl with endomesoderm. Otherwise, these two genes might play roles in cell differentiation in Sycon, as evidenced by the fact that expression of SoxE disappears in choanocytes that transdifferentiate into accessory cells, while expression of SoxF1 becomes stronger in these cells during the process.
Sponges are relatively simple organisms with few cell types, thus the limited number of transcription factors representing conserved metazoan families in the demosponge Amphimedon quenslandica fits neatly with the concept of a simple developmental tool kit patterning a simple body. This study demonstrates that Sycon ciliatum has multiple Sox genes which are dynamically expressed during development and in patterns consistent with governing adult cell differentiation. This indicates that Sox genes were involved in development and cell differentiation from the beginning of multicellular animal evolution. Further analyses of this and other developmental gene families in the Calcarea and in other sponge group are necessary to test whether the identified differences between Sycon and Amphimedon are indicative of global differences in the developmental toolkits. Such studies, now underway in our laboratory and in other groups, will provide insight into the evolutionary history of the animal developmental toolkit.
Identification of Sox genes in Sycon and Leucosolenia
Sox-like genes from Sycon ciliatum were retrieved by searching our recently generated genomic and transcriptome databases (Adamski et al., unpublished work) using HMG domain sequences from Nematostella and Amphimedon. Scaffolds were recovered and annotated using TBLASTN and BLASTX searches. Additionally, we searched in our on-going genome and transcriptome project of another calcaronean, Leucosolenia complicata, using the 12 identified Sycon HMG domain sequences to recover their orthologs from this species. These sequences were used in the phylogenetic analysis.
Sycon Sox genes were amplified by either RACE or RT-PCR using SMARTTM RACE Amplification kit (Clontech). Primer sequences are available upon request. The cDNA used as a template was prepared from a mixture of RNA extracted from juveniles and adult samples containing embryonic stages. PCR products were cloned into pGEM-Teasy (Promega) and sequenced using the BigDye Terminator v3.1 protocol (ABI). Purified PCR products obtained using SP6 and T7 primers during colony PCR were used to produce Dig-labeled antisense RNA probes for in-situ hybridization (see below).
Alignment and phylogenetic analysis
Alignment of HMG domains for phylogenetic analyses: MUSCLE was used for the alignment which included Sycon and Leucosolenia complete HMG domains of candidate Sox genes together with a different combination of taxa (see Additional file2). The alignment was manually modified where needed. In this final dataset, the following sequences were included: two HMG domains from Sycon Tcf genes and out-groups used for phylogenetic analysis as in Jager et al.. We did not include the sponge Sox sequences from the previous study in sponges from Jager et al. as these HMG domains contain only partial information (59 aa).
Two independent runs of PhyML  were performed. Each run searched for five random starting trees using SPR moves. The tree with the best log likelihood value was selected (Log likelihood = −5686.2). From this tree a bootstrap analysis using 100 replicates was performed.
Bayesian analysis  under LG model, with 5,000,000 generations sampled every 500 generations using four chains. Convergence was reached before 5,000,000 generations. A majority rule of consensus tree of 12,500 trees was generated and posterior probabilities values were calculated from this tree.
Finding conserved motifs within sponge Sox sequences
MEME 3.5.7 was used to find conserved motifs outside the HMG domain within Sycon and Leucosolenia Sox proteins and their closest homologues from Acropora, Homo, Nematostella, and Amphimedon. The following parameters were used for searching possible conserved motifs: minimum motif width, six; maximum width, 100; maximum motifs to find, six. Complete sequences were aligned and their motif locations were compared with previous studies[4, 12]. ‘My domain image creator’ tool included in Prosite was used to visualize the locations of motifs in Sox proteins.
Specimen collection and whole mount in-situ hybridization
Adult Sycon specimens were collected from fjords located near Bergen, Norway (+60° 27' 33", +4° 56' 1") during the reproductive season from May to September (2008 to 2011). For in-situ hybridization, samples were immediately fixed in 100 mM MOPS, pH 7.5; 0.5 M sodium chloride; 2 mM MgSO4; 4% paraformaldehyde; 0.05% glutaraldehyde over night at 4°C, stepped into and extensively washed in 70% EtOH and stored at −20°C until processing. Macro sections of sponges in 24 well plates (Nunc) were rehydrated and washed in PBS/0.1% Tween (PTw). Samples were pretreated with 7.5 μg/mL proteinase K for 10 minutes at 37°C, followed by quenching with glycine (2 mg/mL PTw). Acetylation was performed by serial treatment with 0.1 M triethanolamine containing 0, 1.5, and 3 μl/mL acetic anhydride. Re-fixation was done in 4% paraformaldehyde/0.05% glutaraldehyde in PBS for 1 h at room temperature, followed by extensive washing in PTw. Tissue was prehybridized as previously described in 2 mL-tubes for 90 to 180 min at 51°C. Probe hybridization was done with denatured RNA probe (0.1-0.3 ng/μL, approximately 1 kb) for 12 to 18 h at 51°C. Stringent washes were carried out at 55°C as following: 1 × 10 min in hybridization buffer; 2 × 10 min 50% formamide/4 × SSC/0.1%; 2 × 10 min 50% formamide/2 × SSC/0.1% Tween; 2 × 10 min 25% formamide/2 × SSC/0.1% Tween, followed by 3 × 15 min 2 × SSC/0.1% Tween at room temperature. Samples were transferred to maleic acid buffer and incubated in 2% (w/v) Blocking Reagent (Roche) for 60 min at room temperature. After overnight incubation with AP-coupled anti-Digoxigenin-Fab fragments (Sigma, 1:5,000) at 4°C, samples were washed in maleic acid buffer at least 6 × 30 min. Probe was detected using NBT/BCIP as substrate (Roche) with tissue equilibrated in alkaline phosphatase buffer (100 mM sodium chloride, 50 mM MgCl2, 100 mM Tris pH 9.5, 0.1% Tween, 1 mM Levamisole). The staining reaction (0.5 to 3 days) was stopped with PBS/0.5% Tween, samples were transferred to 100% glycerol for microscopy or ethanol-dehydrated and embedded in epoxy resin (Sigma) for sectioning. Pictures of whole mount samples and sections were taken using a Nikon DS-U3 microscope and processed in Photoshop.
This study was funded by the core budget of the Sars International Centre for Marine Molecular Biology. Sequencing has been performed at The Norwegian High-Throughput Sequencing Centre funded by the Research Council of Norway. We thank Lucas Leclère for helpful comments on the manuscript.
1Sars International Centre for Marine Molecular Biology, Thormøhlensgt. 55, Bergen 5008, Norway. 2Department of Biology and Centre for Geobiology, University of Bergen, Thormøhlensgt. 55, Bergen, 5008, Norway.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.