Dicyema Pax6 and Zic: tool-kit genes in a highly simplified bilaterian
© Aruga et al. 2007
Received: 04 July 2007
Accepted: 25 October 2007
Published: 25 October 2007
Skip to main content
© Aruga et al. 2007
Received: 04 July 2007
Accepted: 25 October 2007
Published: 25 October 2007
Dicyemid mesozoans (Phylum Dicyemida) are simple (8–40-cell) cephalopod endoparasites. They have neither body cavities nor differentiated organs, such as nervous and gastrointestinal systems. Whether dicyemids are intermediate between Protozoa and Metazoa (as represented by their "Mesozoa" classification) or degenerate species of more complex metazoans is controversial. Recent molecular phylogenetic studies suggested that they are simplified bilaterians belonging to the Lophotrochozoa. We cloned two genes developmentally critical in bilaterian animals (Pax6 and Zic), together with housekeeping genes (actin, fructose-bisphosphate aldolase, and ATP synthase beta subunit) from a dicyemid to reveal whether their molecular phylogeny supported the "simplification" hypothesis, and to clarify evolutionary changes in dicyemid gene structure and expression profiles.
Genomic/cDNA sequence analysis showed that 1) the Pax6 molecular phylogeny and Zic intron positions supported the idea of dicyemids as reduced bilaterians; 2) the aa sequences deduced from the five genes were highly divergent; and 3) Dicyema genes contained very short introns of uniform length. In situ hybridization analyses revealed that Zic genes were expressed in hermaphroditic gonads, and Pax6 was expressed weakly throughout the developmental stages of the 2 types of embryo and in the hermaphroditic gonads.
The accelerated evolutionary rates and very short and uniform intron may represent a part of Dicyema genomic features. The presence and expression of the two tool-kit genes (Pax6 and Zic) in Dicyema suggests that they can be very versatile genes even required for the highly reduced bilaterian like Dicyema. Dicyemids may be useful models of evolutionary body plan simplification.
The original classification of dicyemids as Mesozoa reflects their intermediate position between the Protozoa (unicellular eukaryotic organisms) and Metazoa (multicellular animals) in body organization. The phylogeny of the dicyemids is controversial, and some researchers consider that dicyemids represent truly primitive multicellular organisms. However, several zoologists regard the simple body plan of dicyemids as the result of specialization of parasitism [references in ]. Recent molecular studies suggest that dicyemids are not truly primitive animals. 18s rDNA phylogenetic analyses by Katayama et al.  showed that the dicyemids belong among the bilateria, and from the structure of the amino acid (aa) sequence in the carboxy-terminal flanking region of the Antennapedia protein in Dicyema orientale, Kobayashi et al.  suggested their affinity to the Lophotrochozoa. A limited number of genes are available for phylogenetic analysis, and the phylogenetic relationships of the dicyemids will need further evaluation when the sequences of more genes become available.
If reduction of body plan complexity secondary to parasitism truly happened in dicyemids, we expect to find genomic features in the dicyemid genome that are associated with the adaptive simplification of body organization. However, such features have not been described in dicyemids. In broader terms, the genomic basis of the simplification that occurs during the course of evolution is poorly understood.
To determine the genomic changes that may occur during the simplification of body organization, we focused on Pax6 and Zic (both of which encode proteins that regulate developmentally critical processes in diverse bilaterians) in dicyemids. Pax6 plays key roles in the development of sensory organs and the nervous system [reviewed in ]. In particular, Pax6 is a master gene in eye development in both vertebrates and Drosophila. In contrast, Zic family proteins have diverse roles [reviewed in [6, 7]]. In vertebrates, they are crucial for embryonic patterning as well as neural tube formation, neural crest generation, and mesodermal segmentation. In urochordates, Zic proteins are required for cell fate decisions for differentiation into various cell lineages, including nerve, muscle, and notochord tissues. In protostomes, Zic genes are essential for embryonic segmentation, midgut morphogenesis, adult head formation in Drosophila [8–10], and vulva formation in Caenorhabditis . Both Pax6 and Zic have structurally related genes in cnidarians; the possible involvement of these cnidarian genes in developmental processes analogous to those controlled by their triploblast homologs has been proposed [12, 13].
Here, we investigated the structures of the dicyemid Pax6 and Zic genes and 3 housekeeping genes [actin, ATP synthase beta subunit (ATPS), and fructose-bisphosphate aldolase (aldolase)], together with the Pax6 and Zic expression patterns. Our results should add to knowledge of the phylogenetic position, common genomic features of the dicyemids, and the role of Pax6 and Zic in their ontogeny.
We identified two Dicyema actin genes that encode actin1 (332 aa) and actin2 (376 aa) proteins (data not shown). The difference in sequence length reflects the presence or absence of an N terminal region, and the homology in the overlapping region was 89% at the aa level.
In total, there were 16 spliceosomal introns in the protein coding regions of the 7 genes (Fig. 2). The size of the introns was a surprising feature, as they were generally short and distributed over a narrow range (mean length, 26.2 ± 1.9 nucleotides, n = 16; Fig. 2). This intron length may be representative of the major population of intron sizes in dicyemids.
The positions of introns in the conserved domains of ZicA, ZicB, and Pax6 were identical to those previously identified as evolutionarily conserved intron positions in each gene family [15–18]. The Dicyema ZicA and ZicB genes both possessed a single intron in the center of the evolutionarily conserved ZF domain (Figs. 2, 4). The positions of these introns corresponded to that of the A-intron, which was conserved in all 32 Zic genes from 7 different bilaterian phyla but was not found in the 8 cnidarian Zic genes evaluated . Dicyema Pax6 had 7 introns (a-g) in the putative protein-coding region (Figs. 2 and 3). Of the 7 introns, one (a) was located in PD and 2 (c, d) were in HD. The positions of these 3 introns matched those of the introns of evolutionarily conserved Pax genes from a broad range of eumetazoans [17–19].
We next tested whether the Dicyema sequences obtained are useful to understand the phylogenetic position of Dicyema. All of the deduced aa sequences were subjected to molecular phylogenetic analyses based on the neighbour-joining (NJ), Bayesian inference (BI), and maximum parsimony (MP) methods. A support for the bilaterian origin was obtained in the case of the Pax6 molecular phylogeny.
Shimodaira-Hasegawa test for Pax6 BI trees in Fig. 5A Paired domain tree
Shimodaira-Hasegawa test for Pax6 BI trees in Fig. 5B Paired domain + Homeodomain tree
To confirm the robustness of the trees, we replaced the Dicyema branch to other places in the BI tree, and performed the likelihood ratio test developed by Shimodaira and Hasegawa [21, 22]. The results indicated that the placement of the Dicyema branch to other bilaterian clades did not significantly worsen the tree except placing in urochodate clade whereas its relocation to non-Pax6 clades worsen the likelihood scores (Tables 1 and 2).
The Dicyema Pax6, ZicA, and ZicB aa sequences had divergence rates 3.6, 2.1, and 2.3 times higher than the means of these sequences in the other bilaterians, supporting the hypothesis that the sequence divergence rate is increased among the proteins encoded by Dicyema genes. However, in the case of Zic and Pax6, the divergence rates varied among the bilaterian phyla. The variance in divergence rates over all of the common animal taxa differed strongly between Zic and the 3 housekeeping proteins [actin, P = 0.0056 (n = 10); ATPS, P = 0.057 (n = 8); aldolase, P = 0.0126 (n = 9) in F-tests].
In contrast to the highly restricted expression of Zic genes, actin1 was expressed widely: in the agamete, hermaphroditic gonad, and the 2 types of developing embryo (Fig. 8F,G). actin2 showed an expression pattern very similar to that of actin1 (data not shown). In the hermaphroditic gonads (Fig. 8F), actin1 expression was detected in the sperm and the early-stage primary oocytes. In the developing infusoriform embryos, actin1 transcript was detected in blastomeres in the vegetal hemisphere of the embryo from the early to middle stages of development (Fig. 8F). In the late developmental stages, expression was restricted to the urn cells that consisted of the internal mass of the embryo (Fig. 8F). In the vermiform embryos, actin1 was expressed in the presumptive peripheral cells and agametes within the axial cell of the embryo (Fig. 8G).
Recent phylogenetic analysis of Pax genes revealed five main classes of Pax genes (Pax4/6, Pax2/5/8/B, PaxA/C/Pox-neuro, Pax1/9, and Pax3/7/D) [12, 20]. In a comprehensive analysis , Pax6 and Pax4 homologues are limited to bilaterians and mammalians, respectively, even though a cnidarian genome project is completing . Based on these backgrounds, we think that the grouping of the Dicyema Pax homologue with bilaterian Pax6 genes in this study supports the bilaterian origin of Dicyema.
In addition to the Pax6 phylogeny, the presence of introns at bilaterian-specific positions in the ZicA and ZicB ZF-encoding domains also favors the idea that dicyemids have bilaterian ancestry. Taken together with the results of previous molecular phylogenetic studies (18S rDNA , Antennapedia orthologue ), we can conclude that dicyemids are members of the Bilateria, and that dicyemids are simplified bilaterian animals. In fact, the infusoriform larvae of dicyemids have a bilateral body plan (Fig. 1). This feature seems to represent the true level of organization in dicyemids, because these larvae are free-swimming, whereas the adult stages are restricted to the renal organs of host cephalopods . The bodies of the adult stages might have been simplified as a reflection of their specialization to their parasitic habitat – the tubular spaces of cephalopod renal organs .
Although we tried to delineate in more detail the origin of Dicyema by using other Dicyema sequences, including concatenated ones, the attempt was unsuccessful (data not shown). The difficulty seems to lie mainly in the highly diverged nature of the Dicyema sequences. Molecular phylogenetic analyses using other sets of slowly evolving sequences may improve our understanding.
The five Dicyema sequences obtained in this study generally showed higher aa sequence diversification rates than the mean rates of other bilaterian orthologs. In a recent study, Thomas et al.  revealed that marked rate variations exist among the Echinodermata, Arthropoda, Mollusca, Annelida, and Platyhelminthes. They suggested that the existence of a strict molecular clock cannot be assumed for the Metazoa. The fact that we found general acceleration of the divergence rate in Dicyema may support this idea.
There are two circumstances under which lineage-specific differences in divergence rates may arise : (1) alteration in the mutation rate of the species (e.g. a higher metabolic rate and shorter generation time may result in more mutations); and (2) differences in the proportion of mutations that become fixed in a population (e.g., species with smaller population sizes are expected to have faster rates of molecular evolution). In the case of dicyemids, both remain plausible, because only self-fertilization occurs in dicyemids [28, 29]. However, we can at least speculate that the presence of an asexual reproductive cycle may accelerate the accumulation of mutations, because a recent study showed that transition to asexuality leads to an excess number of aa substitutions in a microcrustacean, Daphnia pulex . Other possibilities, including shorter generation time and higher metabolic rate, remain open.
Another critical finding about the Dicyema genome was the consistently short length of the introns in dicyemid genes, especially compared with other eumetazoans. Size distribution analyses have revealed that the mean lengths of human, Ciona, and Drosophila introns peak at 87, 60, and 57 bp, respectively . Dicyema introns were even smaller than the average introns of fungi (Schizosaccharomyces pombe, 93 bp; Aspergillus, 72 bp) . A recent study showed that the ciliated protozoan Paramecium tetraurelia has 20- to 35-bp introns (average size, 24.8 bp; n = 1061) . This is very close to our mean intron size for Dicyema (26.2 bp). Although the intron size varies in the eukaryotes, Dicyema may belong to the lowest group.
There is currently no definitive answer to the biological significance of the smallness and uniformity of these intron sizes. However, we noticed that more extreme cases of intron size reduction appear in the nucleomorph of the alga Bigelowella natans . The nucleomorph is a vestige of the eukaryotic cell nucleus after an endosymbiotic process in which one eukaryotic cell is engulfed by another. The sizes of the 852 introns in the Bigelowella natans nucleomorph genome are limited to either 18, 19, 20, or 21 bp. Considering the analogous situation between endosymbiosis and parasitism, the shortening of the introns in dicyemids may reflect a high level of selective pressure to reduce unnecessary expense (such as long introns) during the course of adaptive simplification. The size reduction may be facilitated by a lack of the need for the presumptive cis-regulatory elements essential for adapting to various environments.
The actin sequences showed particularly strong diversification between Dicyema and other eumetazoa. This finding surprised us, because actin genes are widely conserved in eukaryotes, including in fungi and plants, in which Pax6 or Zic homologs have not yet been identified. Dicyema actin was expressed in the agamete, hermaphroditic gonad, and the developmental stages of the 2 types of embryo. Actin bundles play a role in many cellular processes, such as cell division. This suggests that more transcripts of actin may be required for cell division and cleavage during embryogenesis than in the external structure of the worms (calotte and peripheral cells). The actin bundles also contribute to cell movement; actin expression was therefore detected in unflagellated sperm in the dicyemids. Considering the essential role of actin in determining cell shape in multicellular animals, it is tempting to speculate that a reduction in somatic cell number, as is seen in Dicyema, may reduce the evolutionary constraint on actin protein structure.
Our major interest was in the features of Dicyema Pax6 and Zic, which are utilized in the developmental processes of many bilaterians. In this context, they can be regarded as "toolkit genes ," like the genes encoding Hox, parahox, and other developmentally critical transcription factors. Our most important finding may be the presence of both genes in the dicyemid. These genes could have been lost in the course of evolution. In particular, Pax6 is widely involved in photosensing organ development in bilaterians . However, Pax6 PD/HD and Zic ZF were fairly well conserved, and their diversification rates were comparable to those of ATPS and aldolase and lower than that of actin.
Based on their expression profiles, the products of ZicA and ZicB may have a similar role in the hermaphroditic gonad within the axial cell. Dicyema Zic genes may therefore be associated with gametogenesis or the differentiation of gametes.
Dicyema Pax6 was detected throughout the developmental stages of the 2 types of embryo and in the hermaphroditic gonads. The location of this expression suggests that Dicyema Pax6 has few specific roles, such as the genetic control of eye development and neurogenesis. In this regard, previous studies showed various roles Pax6 in eumetazoan development. In Caenorhabditis elegans, its Pax6 homologue (vab-3) is required for the cell fate decision of a male specific blast cell lineage , and the control of hermaphrodite gonad size and shape . In planarians, Pax6A is expressed not only in their nervous system but also in a non-cephalic parenchyma that gives rise to ventral marginal adhesive zone and is dispensable for eye regeneration . Together with our results, the phylogenetic role of Pax6 may not be limited to the nervous system and photosensing organs.
The maintenance of the two toolkit genes in Dicyema that lack photosensing organs or nervous systems suggests that these genes possess unexpected versatility in the development of multicellular animals. The common expression of Pax6 and Zic in Dicyema hermaphroditic gonads suggests that they can be involved even in the most fundamental and indispensable function in multicellular animals, i.e. gametogenesis.
The results in this study can be summarized as follows: 1) the Pax6 molecular phylogeny and Zic intron positions supported the idea of dicyemids as reduced bilaterians; 2) the aa sequences deduced from the 7 Dicyema genes were highly divergent in comparison to those from bilaterians; 3) dicyemid genes contained very short introns of uniform length; and 4) ZicA and ZicB were expressed in hermaphroditic gonads, and Pax6 was expressed weakly throughout the developmental stages of the 2 types of embryo and in the hermaphroditic gonads. The accelerated evolutionary rates and very short and uniform intron may represent a part of Dicyema genomic features. The presence and expression of the 2 tool-kit genes (Pax6 and Zic) in Dicyema suggests that they can be very versatile genes even required for the nervous-system-or-photosensing-organ-lacking bilaterian.
Dicyema acuticephalum was obtained from the renal organs of Octopus vulgaris purchased at a fish market in Osaka, Japan. The bodies of D. acuticephalum, which are 200–800 μm long, consist of a central cylindrical cell (the axial cell) and a single layer of 18 ciliated external cells (the peripheral cells; Fig. 1). In the anterior region, 8 peripheral cells form the calotte, which inserts the anterior region of the body into the crypts or folds in the renal organs of the host cephalopod. Species identity was confirmed by 18S rDNA phylogeny (AB266027, data not shown).
Dicyemids were collected from the renal organs of the octopus with pipettes. Contaminating host tissues in the dicyemid suspension were removed with pipettes under a stereoscopic microscope. The isolated dicyemids were washed several times with artificial seawater.
Genomic DNA and total RNA were extracted as described . For both nucleic acids, glycogen was used during precipitation to increase recovery. The cDNA fragments for Pax6, Zic, and actin were amplified by PCR using universal primers [15, 40, 41]. In a pilot molecular phylogeny analysis, it was clear that these genes were not orthologs of any other paired domain-containing genes (e.g. Pax2/5/8) or ZF motif-containing genes (e.g. Gli and Glis) (data not shown). On the basis of the sequences initially identified, 3' RACE (rapid amplification of cDNA-ends) PCR was done with a 3'-Full RACE Core Set (Takara BIO, Shiga, Japan). In the course of above PCR analyses, we obtained ATP synthase beta subunit and fructose-bisphosphate aldolase as PCR products of mispriming. Their nucleotide sequences were determined and deposited in DDBJ database. We considered that the cloned ZicA, ZicB, Pax6, and actin1 and actin2 cDNAs represented major molecular species in D. acuticephalum, because all of the PCR fragments (at least 15 for each gene) generated by universal primers contained sequences nearly identical to the first fragment we obtained. The 5' flanking genomic sequences were cloned with a DNA Walking SpeedUp Kit (Seegene, Seoul, Korea). Putative exons were identified by the open reading frame search program in DNASISPro (Hitachi Software Engineering, Tokyo, Japan) and by their similarity to homologs in other bilaterian species. Finally, entire coding regions were amplified from both cDNA and genomic DNA by using a pair of PCR primers that amplified the deduced open reading frames of the genes. The PCR products were cloned into pGEMT Easy Vector (Promega, Madison, WI). The nucleotide sequences were determined on the basis of the sequencing of at least 4 independent clones. 18S ribosomal DNA was cloned as described . The sources of the cloned sequences were verified by highly stringent Southern blot analysis using genomic DNAs from both D. acuticephalum and the cephalopod Octopus ocellatus (data not shown).
Deduced aa sequences were subjected to phylogenetic analysis. The other sequences were collected from NCBI database. The lists of the sequences with their species identity and accession number are indicated in Tables (see Additional file 2).. Sequences were aligned by using CLUSTALW  implemented in MEGA3 . Regions with insertions or deletions resulting from the program were omitted by visual inspection. BI analysis was done with MrBayes 3.1.2 [44, 45]. In BI analysis, we used an empirical model (JTT matrix, ) with Inv+gamma, alpha shape parameters, and aa frequencies estimated from the data. We ran 1,000,000 generations with 1 cold and 3 incrementally heated Markov chains, random starting trees for each chain, and sampling of trees every 100 generations. 50% major-rule consensus trees were constructed from the samples of stabilized trees. NJ and MP analyses were done with MEGA3.1 . NJ tree was based on the distance calculation with JTT matrix  after removing position containing gaps (complete deletion option) for genes other than Zic genes, or after removing only in pairwise sequence comparisons (Pairwise deletion option) for Zic genes. For the NJ trees and MP tree analyses, the statistical significance of the branches obtained was calculated according to a bootstrap test  with 1000 repetitions. The alternative tree analysis was done as described . Briefly, alternative trees were constructed by RETREE program, and the log-likelihood values were calculated by PROML program, taking into account site-specific rate differences using a gamma correction. Both programs are implemented in PHYLIP program package . The statistical significance was determined by Shimodaira-Hasegawa test [21, 22] implemented in PHYLIP.
For the aa substitution rate analysis, ancestral aa sequences for each protein were deduced by the ANCESCON program, a distance-based program that gives more accurate ancestral sequence reconstruction than do PAML, PHYLIP, and PAUP* at large evolutionary distances  with all of the bilaterian sequences. We used MEGA3.1 to calculate the pairwise distance. For the test of heterogeneity of variance (F-test), the aa substitution frequencies were first normalized by [(total aa numbers in the sequence) × (1-Pinv)].
Parts of the cDNA fragments were ligated into pBluescript II KS vector, and we used this plasmid to synthesize a Dig (digoxygenin)-labeled RNA probe. Dig-labeled RNA probes were made by in vitro transcription from the linearized plasmid with a Dig RNA labeling kit (Roche). T7 or T3 RNA polymerase (Roche) was used in the in vitro transcription to synthesize an antisense RNA probe for hybridization.
In each step, it took 15 min to form a drop of dicyemids in the bottom of each microtube. Dicyemids were washed 5 times and then fixed overnight at 4°C in 4% paraformaldehyde in 0.5 M NaCl and 0.1 M MOPS buffer. After fixation, the dicyemids were dehydrated in an ethanol series (30%, 50%, 70%) for 10 min each and the dicyemids were stored in a 70% ethanol up-series at -30°C. Dicyemids were hydrated in an ethanol down-series (70%, 50%, 30%) for 5 min each, and were washed 3 times in PBT (phosphate buffer saline [PBS] with 0.1% Tween 20). The hydrated dicyemids were partly digested by protease K (0.5 mg/mL in PBS) for 15 min at 37°C.
After digestion, the dicyemids were washed in PBT and then fixed in 4% para- formaldehyde in PBS for 60 min at room temperature. After fixation, the dicyemids were washed 4 times in PBT for 5 min at room temperature. After washing, they were incubated in the following serial steps: in 50% hybridization buffer (5 × SSC, 1% SDS, 50% formamide) in PBT for 10 min at room temperature; in hybridization buffer for 10 min at room temperature; and in hybridization buffer for 1 h at 50°C. After incubation, the Dig-labeled RNA probe was added to the hybridization buffer and the mixture incubated overnight at 50°C.
The dicyemids were washed in the following serial steps: in 5 × SSC – 50% formamide – 1% SDS for 20 min at 50°C; in 2 × SSC – 50% formamide – 1% SDS for 20 min at 50°C; in 2 × SSCT (2 × SSC and 0.1%Tween) for 15 min at 50°C twice; in 1 × SSCT (1 × SSC and 0.1% Tween) for 15 min at 50°C twice; in 0.5 × SSCT (0.5 × SSC and 0.1% Tween) for 15 min at 50°C twice; and in 0.2 × SSCT (0.2 × SSC and 0.1% Tween) for 15 min at 50°C twice. After these washes, the dicyemids were washed for 5 min in a blocking buffer containing 2% BM blocking reagent (Roche); they were incubated in the blocking buffer for 60 min and then placed at 4°C for 10 min before the addition of antibody. An alkaline-phosphatase-conjugated anti-Dig antibody (Roche) was added at 1/2000 volume of the suspension of dicyemids. The antibody reaction was conducted overnight at 4°C. The dicyemids were then washed in PBT for 15 min at room temperature 4 times. After this, they were washed twice in a detection buffer (100 mM Tris-HCl [pH 9.5], 100 mM NaCl, 50 mM MgCl2) for 15 min at room temperature. After these washes, the dicyemids were immersed in BM purple solution (Roche) to detect the alkaline phosphatase activity.
We also conducted control experiments in the absence of probe or in the presence of a probe [DDBJ accession no. AB299857] that gave a spatially restricted signals. The specificity of the signals was checked with these two controls in each hybridization analysis.
synthase (ATPS), ATP synthase beta subunit
paired box domain
zinc finger domain
We thank Yayoi Nozaki and the Research Resource Center of RIKEN Brain Science Institute for their skilful assistance, Takahiko J. Fujimi and Hirokazu Takahashi for their helpful comments on the manuscript, and Mark Q. Martindale (University of Hawaii) for a Pax nexus file. This work was supported by grants-in-aids from the Japanese Ministry of Education, Science, Sports, and Technology; the Inamori Foundation; and the Japan Society for the Promotion of Science (research grant no. 18570087).
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.