In the development of most arthropods, the caudal region of the elongating germ band (the growth zone) sequentially produces new segments. Previous work with the spider Cupiennius salei suggested involvement of Delta-Notch signaling in segmentation. Here, we report that, in the spider Achaearanea tepidariorum, the same signaling pathway exerts a different function in the presumptive caudal region before initiation of segmentation. In the developing spider embryo, the growth zone becomes morphologically apparent as a caudal lobe around the closed blastopore. We found that, preceding caudal lobe formation, transcripts of a Delta homolog, At-Delta,are expressed in evenly spaced cells in a small area covering the closing blastopore and then in a progressively wider area of the germ disc epithelium. Cells with high At-Delta expression are likely to be prospective mesoderm cells, which later express a twist homolog, At-twist, and individually internalize. Cells remaining at the surface begin to express a caudal homolog, At-caudal, to differentiate as caudal ectoderm. Knockdown of At-Delta by parental RNA interference results in overproduction of At-twist-expressing mesoderm cells at the expense of At-caudal-expressing ectoderm cells. This condition gives rise to a disorganized caudal region that fails to pattern the opisthosoma. In addition, knockdown of Notch and Suppressor of Hairless homologs produces similar phenotypes. We suggest that, in the spider, progressive activation of Delta-Notch signaling from around the blastopore leads to stochastic cell fate decisions between mesoderm and caudal ectoderm through a process of lateral inhibition to set up a functional caudal lobe.
Segments consisting of ectoderm and mesoderm are characteristic of arthropod embryos. In development of most arthropods, segments in the posterior half of the future body are sequentially produced at the caudal region of the elongating germ band, which is often referred to as `the growth zone' (Peel et al., 2005). Similar modes of segmentation are seen outside the phylum Arthropoda (e.g. in Chordata and Annelida). Whether segmentation in different bilaterian phyla occurs by common or independent origins is the subject of controversy(Davis and Patel, 1999; Peel and Akam, 2003). Recent work in spider has suggested that spider segmentation and vertebrate somitogenesis share a mechanism involving Delta and Notch, which favors a common origin model (Stollewerk et al.,2003; Schoppmeier and Damen,2005). However, it is still not clear whether the mesodermal somites of vertebrates can be compared to arthropod segments, which consist of ectoderm and mesoderm (Patel,2003). Understanding of developmental processes occurring before segmentation may be essential to resolve these issues.
Increasing numbers of studies in arthropods have focused on the growth zone. In the well-studied insect Drosophila melanogaster, all segments form almost simultaneously during early stages of embryogenesis. Thus, the Drosophila embryo shows no area corresponding to the growth zone at any developmental stage. Nonetheless, mechanisms regulating posterior and terminal patterning in Drosophila(St Johnston and Nüsslein-Volhard,1992) have facilitated analysis of the development of the growth zone. In diverse arthropod species, the growth zone is characterized by specific expression of homologs of Drosophila caudal (cad)(Mlodzik et al., 1985; Macdonald and Struhl, 1986; Moreno and Morata, 1999; Schulz et al., 1998; Akiyama-Oda and Oda, 2003; Copf et al., 2004; Shinmyo et al., 2005). cad genes were shown to be essential for sequential production of the posterior segments in some species (Copf et al., 2004; Shinmyo et al.,2005). In addition to the cad genes, homologs of Drosophila maternal terminal genes, namely torso and torso-like, are required to set up or maintain a functional growth zone in the beetle Tribolium(Schoppmeier and Schröder,2005). Despite these significant studies, the cellular and molecular events leading to establishment of the growth zone and the diversity of mechanisms underlying its development among arthropods are poorly understood.
Spiders, which are chelicerate arthropods, phylogenetically distant from the insects, are classical organisms that have been used for developmental study (Montgomery, 1909; Holm, 1952; Seitz, 1966; Yoshikura, 1975). Modern molecular techniques have been applied to two spider species, Cupiennius salei and Achaearanea tepidariorum. In Achaearaneadevelopment, the growth zone becomes morphologically apparent as a caudal lobe during stages when the germ disc is converted to a germ band(Montgomery, 1909; Yamazaki et al., 2005). These stages overlap with appearance of mesoderm cells expressing a homolog of twist (twi), At-twi(Yamazaki et al., 2005) and caudal ectoderm cells expressing a homolog of cad, At-cad(Akiyama-Oda and Oda, 2003). Once formed, the caudal lobe is composed of an ectodermal and a mesodermal cell layer (Montgomery, 1909). Work with Cupiennius has revealed dynamic expression of Delta among other genes at the caudal region(Damen et al., 2000; Stollewerk et al., 2003),reminiscent of the molecular oscillation proposed for vertebrate somitogenesis(Pourquié, 2003). RNA interference (RNAi)-mediated knockdown of several components of the Delta-Notch signaling pathway, including Delta, was shown to cause defects in the pattern of the opisthosomal segments(Stollewerk et al., 2003; Schoppmeier and Damen, 2005). However, as double-stranded RNA (dsRNA) was applied to Cupienniusembryos around the blastula stage by injection in to the perivitelline space(Schoppmeier and Damen, 2001),it is unclear whether target genes were silenced during stages before initiation of segmentation.
In this study, we examined the cellular and molecular events giving rise to a caudal lobe in Achaearanea, by taking advantage of parental RNAi,which can suppress target gene function during early stages of embryogenesis(Akiyama-Oda and Oda, 2006). We found that transcripts of a Delta homolog are expressed at high levels in what are likely to be prospective mesoderm cells arising from around the blastopore. Based on gene expression data and phenotypes produced by RNAi-mediated knockdown of Delta and other components of the Notch signaling pathway, we propose that Delta-Notch signaling is essential for caudal lobe formation in Achaearanea, preceding initiation of segmentation. Our findings reveal that formation of mesoderm and caudal ectoderm is a single event that sets up a functional caudal lobe in the Achaearaneaembryo.
MATERIALS AND METHODS
An Achaearanea germ-band stage embryo cDNA library(Akiyama-Oda and Oda, 2003) was used to generate more than 9000 expressed sequence tags (ESTs), which were sequenced from the 5′ end. These include two sequences derived from a single gene closely related to Drosophila Delta [clone identification(ID), At_eW_009_P16 and At_eW_018_P16]. A full-length cDNA clone for this gene was isolated by re-screening the library, its nucleotide sequence was determined, and the deduced amino acid sequence of the Delta-related cDNA analyzed by BLASTP (hit against Drosophila Delta with an E-value of 1e-156), BLAST 2 sequences, ClustalW version 1.81 and PHYLIP version 3.5. Analysis confirmed that the gene is the closest relative of known Delta genes, and it was designated At-Delta. The EST collection also included Notch (clone ID, At_eW_027_D04) and Suppressor of Hairless [Su(H)] (clone ID,At_eW_016_A20) homologs, designated At-Notch and At-Su(H), respectively. The nucleotide sequences of a partial At-Notch and a partial At-Su(H) cDNA were determined. The deduced amino acid sequence of the At-Notch cDNA was hit against Drosophila Notch with an E-value of 4e-147 and that of the At-Su(H) cDNA against Drosophila Su(H) with an E-value of 9e-35. A short fragment of a homolog of Drosophila hedgehog (hh), designated At-hh, was isolated by PCR with the following primers: forward, 5′-GARGARGGNACNGGNGCNGA-3′,and reverse, 5′-ACCCARTCRAANCCNGCYTC-3′. A full-length At-hh cDNA clone was isolated by cDNA library screening. Phylogeny of Hh proteins including At-Hh was analyzed in a previous study(Simonnet et al., 2004). Sequences are available from GenBank under the accession numbers: At-Delta, AB287420; At-Notch, AB287421; At-Su(H), AB287422; and At-hh, AB125742.
A DNA fragment encoding the carboxyl-terminal 182 amino acids of Achaearanea Forkhead (At-Fkh)(Akiyama-Oda and Oda, 2003) was amplified by PCR and inserted between the EcoRI and SalI sites of pET-28a(+) vector (Novagen, Madison, WI). Expression of the fusion protein was induced in BL21 (DE3) cells, and cell lysates were fractionated by SDS-PAGE and electroeluted from the gel. Purified protein was employed as an antigen to immunize two guinea pigs. As antiserum from one of the two animals specifically reacted to bacterially expressed At-Fkh fused to maltose-binding protein (see Fig. S1 in the supplementary material) and showed nuclear staining consistent with the expression patterns of At-fkhtranscripts previously reported(Akiyama-Oda and Oda, 2003; Akiyama-Oda and Oda, 2006), it was used for experiments. A commercially available rabbit anti-β-catenin antiserum (C2206, Sigma-Aldrich, St Louis, MO) was used.
Staining of embryos
For immunostaining, embryos were fixed as described(Oda et al., 2005). Anti-At-Fkh and anti-β-catenin antisera were diluted 1:1000. Donkey anti-guinea pig IgG labeled with Cy3 (Chemicon, Temecula, CA) and donkey anti-rabbit IgG labeled with Cy5 (Chemicon) were used as secondary antibodies diluted 1:200. Samples were counterstained with 1 U/ml phalloidin-fluorescein(Molecular Probes, Eugene, OR) and 0.5 mg/ml DAPI (Sigma). Stained samples were examined with an Olympus IX71 microscope equipped with a cooled CCD camera (CoolSNAP HQ, Roper Scientific, Tucson, AZ) controlled by MetaMorph version 6.1 (Universal Imaging, Downington, PA).
Single- and dual-color whole-mount in situ hybridizations were performed as described (Lehmann and Tautz,1994; Akiyama-Oda and Oda,2006), followed by DAPI staining. All stained samples except for those shown in Fig. 1L,L′were flat-mounted after removal of excess yolk. Embryos were photographed using a stereomicroscope (SZX12, Olympus) equipped with a color CCD camera(C7780-10, Hamamatsu Photonics, Hamamatsu, Japan), and an Axiophot2 (Carl Zeiss). Cell death was visualized using a TACS 2 TdT-DAB kit (Trevigen,Gaithersburg, MD).
The 1054 bp (nucleotide [nt] 136-1189) and 965 bp (nt 1406-2370) regions of At-Delta cDNA were used to synthesize At-DeltaEBand At-DeltaHH dsRNAs. Control dsRNA was synthesized using the entire coding region of jellyfish green fluorescent protein(gfp) gene (Quantum), At-Notch dsRNA with a 1190 bp region(nt 30-1219) of a partial At-Notch cDNA, and At-Su(H) dsRNA with a 580 bp region (nt 1-580) of a partial At-Su(H) cDNA. Synthesis and injection of dsRNA were performed as described (Akiyama-Oda and Oda, 2006). dsRNAs were used at concentrations of 1.5-2.0μg/μl for injection. dsRNA solution (1-2 μl) was repeatedly introduced into the opisthosoma of each female at 2-3 day intervals. The injection was repeated five times for At-DeltaEB, gfp and At-Su(H)dsRNA, three (the individuals #7 and #8) or four(the individual #9) times for At-DeltaHH dsRNA, and six times for At-Notch dsRNA.
Time-lapse recording of live embryos
After dechorionation with 100% bleach, embryos were attached to glass coverslips with double-sided sticky tape and covered with halocarbon oil 700(Sigma). Images were taken every 5 minutes with the above color CCD camera and processed in MetaMorph version 6.1, ImageJ version 1.32 and Adobe Premiere version 6.0 to produce the movie.
Characterization of five distinct cell types that internalize during early stages
Examination of embryos fixed at successive stages with several probes allowed us to identify five distinct cell types that were situated below the surface epithelial cell layer during the period between stages 4 and 7(Fig. 1A). The center of the forming germ disc during stage 4 (Fig. 1B) was the site at which cells expressing transcripts and proteins of At-fkh internalized(Fig. 1C-F)(Akiyama-Oda and Oda, 2003). This site of cell internalization is referred to as the blastopore(Fig. 1B,D)(Montgomery, 1909). We designated At-fkh-positive cells situated below the central area of the germ disc epithelium as central endoderm (cEND) cells. Cumulus mesenchymal(CM) cells, characterized by decapentaplegic (dpp)expression (Akiyama-Oda and Oda,2003), appeared to originate as a subpopulation of cEND cells. In accordance with this, strong β-catenin concentration at specific sites of cell-cell contact was observed in the cEND cell cluster at late stage 4(Fig. 1F, arrow) and in the migrating CM cell cluster during stage 5 (data not shown). After CM cells left the center of the germ disc, remaining cEND cells were scattered(Fig. 1H,J,J′,K,K′).
The rim of the germ disc (Fig. 1G) was another site of cell internalization. At mid stage 5, but not earlier, some cells expressing At-fkh transcripts were found below the germ disc epithelium close to the rim(Fig. 1I,I′). These internalized cells, which did not express At-twi transcripts(Fig. 1L,L′), were designated peripheral endoderm (pEND) cells. At mid- to late stage 5, some At-fkh-expressing cells located at the rim of the germ disc began to express At-twi transcripts (Fig. 1L,L′). These cells did not enter the inside until CM cells almost reached the rim of the germ disc (late stage 5). At approximately the beginning of stage 6, when the extraembryonic area begins to differentiate(Fig. 1M), an increasing number of At-twi-expressing cells was seen at and near the rim of the germ disc (Fig. 1N). All of these peripheral At-twi-positive cells except cells at the rim were located below the surface epithelium (data not shown). At-twi-positive cells situated below the peripheral epithelium were designated peripheral mesoderm(pMES) cells. After internalization, pMES cells appeared to migrate centripetally (Fig. 1N,T-W).
As previously described (Yamazaki et al., 2005), some cells around the center of the germ disc, which were apparently distinct from cEND cells, began to express At-twitranscripts starting at the beginning of stage 6(Fig. 1N-Q,O′-Q′). These cells showed an evenly spaced distribution, although many were seen in pairs. Visualization of nuclei showed that At-twi-positive cells were at the surface or internalizing (Fig. 1O,O′,P,P′). At-twi-positive cells that situated below the central epithelium were designated central mesoderm (cMES)cells. New cMES cells continued to arise from an area adjacent to the growing At-cad expression domain (Fig. 1R-W). Our previous studies showed that in the formed caudal lobe,in addition to ectoderm cells, some mesoderm cells express At-cadtranscripts and that expression of At-twi transcripts disappears(Akiyama-Oda and Oda, 2003; Yamazaki et al., 2005).
Molecular characterization of a Delta homolog in Achaearanea
EST analysis of Achaearanea germ-band-stage embryos identified a homolog of Drosophila Delta, designated At-Delta. Subsequent in situ hybridization analysis revealed that the At-Delta expression pattern was, in part, similar to that of At-twi at around stages 5 and 6 (see below).
The deduced amino acid sequence of At-Delta showed a Delta/Serrate/LAG-2(DSL) domain and nine EGF-like repeats(Fig. 2A,B). The overall protein structure more closely resembled Drosophila Delta than Drosophila Serrate, and phylogenetic analysis using unambiguously alignable amino acids in the DSL domain and in EGF-like repeats 1-3 confirmed this observation (Fig. 2C). It also showed that At-Delta most closely resembled Cs-Delta1, one of two Delta homologs reported in another spider, Cupiennius salei(Fig. 2B,C)(Stollewerk, 2002).
Expression patterns of At-Delta transcripts
Expression patterns of At-Delta transcripts were examined by whole-mount in situ hybridization. From before mid-stage 4, At-Deltatranscripts were detected in cells located near, but not at, the rim of the forming germ disc (Fig. 3A-C). No expression was detectable around the center of the germ disc at mid-stage 4(Fig. 3A). In a late-stage 4 embryo, however, several cells located at the rim of the forming germ disc and all non-germ-disc surface cells expressed At-Delta(Fig. 3D). In the same embryo,some surface cells around the blastopore distinct from the cEND cells expressed At-Delta (Fig. 3D-F). During stage 5, the area displaying interspersed At-Delta expression expanded (Fig. 3G). At-Delta-expressing cells, many of which were seen in pairs, were located at the same level as the remaining cells in the epithelium (Fig. 3H). About a quarter of the surface cells at the central area displayed high-levels of At-Delta expression. Both paired and unpaired At-Delta-positive cells were evenly spaced.
By early stage 6, At-Delta expression began to fade from around the center of the germ disc (Fig. 3I). This process was followed by the appearance of At-twi transcripts, probably in prospective cMES cells(Fig. 3L,M). In support of this idea, cells positive for low levels of both At-Delta and At-twi were found in that area as At-Delta expression faded(Fig. 3M, arrows). Expression of At-cad was observed in the growing At-Delta-negative domain at mid-stage 6 but not early stage 6(Fig. 3I,J). Changing expression patterns of At-Delta, At-fkh, At-twi and At-cadpreceding caudal lobe formation are summarized schematically in Fig. 3N.
A second wave of At-Delta expression was observed in the nascent caudal lobe at late stage 6/early stage 7(Fig. 3K). Similar to previous observations for Cupiennius Delta genes(Stollewerk et al., 2003), an increase in the number of stripes of At-Delta expression was observed in the developing opisthosomal region (Fig. 3O,P). In addition, one relatively broad band of At-Deltaexpression became prominent in the prosomal region during early germ band stages (Fig. 3O,P). In the limb-extending germ band (stage 9), At-Delta expression was observed in neural precursors (Fig. 3Q),as reported for Cs-Delta1(Stollewerk, 2002).
Parental RNAi against At-Delta results in defects in caudal lobe formation
To examine the function of At-Delta in early spider embryos, we used parental RNAi. Two types of dsRNA were prepared from non-overlapping regions of At-Delta cDNA, designated At-DeltaEBand At-DeltaHH dsRNAs(Fig. 2A). Females injected with At-DeltaEB or At-DeltaHH dsRNA,but not with control gfp dsRNA, produced embryos displaying morphological defects starting at early stage 6(Fig. 4A-E; see Movie 1 in the supplementary material). In affected embryos, an unusual large invagination was formed around the center of the germ disc, resulting in abnormal thickening of cells at the emerging caudal region(Fig. 4B). For most females injected with At-DeltaEB or At-DeltaHHdsRNA, the penetrance of this `invagination' phenotype was 100% (excluding embryos showing non-specific developmental defects) 2 to 3 weeks after the first injection (Fig. 4D,E). This condition lasted for another few weeks. Neither the frequency of egg-laying nor the quality of eggs was significantly affected.
Embryos displaying invagination phenotypes at stage 6 developed into various forms. Many formed either the correct or a reduced number of limbs,but formation of the germ band and extension of limbs were markedly delayed(see Movie 1 in the supplementary material). Phenotypes of late embryos were categorized as normal or consisting of three defective classes (classes 1 to 3; Fig. 5). Despite early defects, embryos of the normal class recovered a normal form(Fig. 5E). The class 1 phenotype showed a reduced opisthosoma with the prosoma developing rather normally (Fig. 5B). In many class 1 embryos, extension of limbs corresponding to the third and fourth walking legs was delayed. The class 2 phenotype displayed a loss of the entire opisthosoma, with the prosomal limbs frequently reduced in number(Fig. 5C). The class 3 phenotype showed a significant reduction of the prosoma with a disorganized germ band configuration, in addition to loss of the entire opisthosoma(Fig. 5D). Based on late phenotypes, the effect of At-DeltaHH dsRNA injection was more significant than that of At-DeltaEB dsRNA injection(Fig. 5E).
To confirm that defects result from suppression of At-Deltaexpression, the At-Delta transcripts were examined in stage 5 embryos derived from females injected with gfp, At-DeltaEB or At-DeltaHH dsRNA by whole-mount in situ hybridization. It was found that injection of At-DeltaEB and At-DeltaHH dsRNA but not gfp dsRNA disrupted the interspersed pattern of At-Delta expression in the germ disc(Fig. 6A,C,E) and resulted instead in a smeared pattern. High magnification of central cells of the affected germ discs revealed that there were fewer At-Deltatranscripts in the cytoplasm, whereas most, if not all, nuclei displayed two dots of At-Delta transcripts possibly related to homologous chromosomes (Fig. 6B,D,F). This observation implies that At-Delta transcription is active, but At-Delta transcripts are not stable in the cytoplasm. Combined with the fact that the two dsRNAs yielded similar phenotypes, we conclude that defects seen following injection of At-DeltaEB or At-DeltaHH dsRNA result from specific degradation of At-Delta transcripts. Altered distribution of At-Deltatranscripts among central cells of the germ disc were probably due to suppressed At-Delta function, as discussed below. Embryos in which cytoplasmic At-Delta transcripts were depleted after injection of At-DeltaEB or At-DeltaHH dsRNA are hereafter referred to generally as At-Delta RNAi embryos and specifically as At-DeltaEB or At-DeltaHH RNAi embryos.
Mesoderm and caudal ectoderm phenotypes of At-Delta RNAi embryos
We further investigated the cause of caudal lobe defects in At-Delta RNAi embryos using molecular markers of early cell types. In At-Delta RNAi embryos, the At-fkh-expressing cEND cells,including CM cells, formed and behaved normally(Fig. 7A,B). However, the number of At-fkh-expressing cells at the periphery of the stage 5 germ disc was markedly reduced in At-DeltaHH RNAi embryos(Fig. 7A,B), but less so in At-DeltaEB RNAi embryos (data not shown). At stage 6,there was no significant difference in the expression of At-fkhbetween control and At-DeltaEB RNAi embryos(Fig. 7C,D). By contrast,expression patterns of At-twi and At-cad were grossly defective in At-DeltaEB RNAi embryos at late stage 6(Fig. 7E-H). All cells participating in abnormal invagination seen in At-Delta RNAi embryos expressed high levels of At-twi(Fig. 7F). Conversely, no cells in the emerging caudal region of At-Delta RNAi embryos expressed significant levels of At-cad transcripts(Fig. 7G,H). Finally, At-Delta RNAi embryos at stage 6 did not show At-twitranscripts in the peripheral region, as did control embryos(Fig. 7E,F).
Most At-DeltaEB RNAi embryos formed a bilaterally symmetrical germ band of reduced length. Such germ bands displayed a rather normal arrangement of At-twi-expressing cells in most of the prosomal region, but persistently exhibited a large mass of At-twi-expressing cells in the caudal region (Fig. 7J). This was in contrast to the normal germ band, in which At-twi expression is absent from the caudal lobe(Fig. 7I)(Yamazaki et al., 2005). In addition, staining for At-hh showed that At-DeltaEB RNAi embryos did not generate opisthosomal segments despite the rather normal segmentation seen in the prosomal region(Fig. 7K,L). These observations accounted for the class 2 phenotype. Moreover, TUNEL staining showed that there were many dying cells in the caudal region of At-Delta RNAi embryos, but only a few in the caudal lobe of control embryos(Fig. 7M,N).
Parental RNAi against Notch and Su(H) homologs results in similar phenotypes
To examine whether other components of the Delta-Notch signaling pathway function in Achaearanea caudal lobe formation, we performed parental RNAi against At-Notch and At-Su(H). As seen with At-Delta dsRNAs, injection of At-Notch or At-Su(H)dsRNA resulted in embryos displaying unusual invagination around the center of the germ disc and subsequent thickening of cells at the emerging caudal region(Fig. 8A,B). However, the penetrance of the invagination phenotype in individual egg sacs after the injection of At-Notch and At-Su(H) dsRNA was not 100% (see Fig. S2 in the supplementary material). Nonetheless, expression patterns of At-Delta, At-twi and At-cad transcripts were affected in both At-Notch and At-Su(H) RNAi embryos derived from egg sacs with the high penetrance of the invagination phenotype(Fig. 8C-H). In these embryos,the interspersed pattern of At-Delta expression around the center of the germ disc was partially or completely disrupted, although levels of At-Delta transcripts were higher than in At-Delta RNAi embryos (Fig. 8C,D). At later stages, there was an increased number of At-twi-expressing cells and a reduced number of At-cad-expressing cells at the caudal region. This phenotype was similar to that seen in At-Delta RNAi embryos. In addition, in At-Notch but not At-Su(H) RNAi embryos, many At-twi-expressing cells probably corresponding to pMES cells were observed outside the caudal region (Fig. 8E,F), indicative of a difference between At-Delta and At-Notch RNAi embryos.
We have shown that formation of the caudal lobe in the spider Achaearanea tepidariorum, which precedes initiation of segmentation,requires three major components of the canonical Delta-Notch signaling pathway, namely, Delta, Notch and Su(H).
High levels of Delta transcripts are expressed in prospective mesoderm cells arising from around the blastopore
We described five distinguishable cell types that internalize during stages 4-7, namely cEND, CM, pEND, pMES and cMES cells. The first three are categorized as endoderm, and the last two as mesoderm. CM cells are regarded as a derivative of the cEND cells. Formation of the Achaearaneablastopore is associated with cEND cell internalization. Both mesodermal cell types express At-twi, a homolog of the Drosophilamesoderm-determining gene, suggesting that Achaearanea mesoderm is evolutionarily related to Drosophila mesoderm(Yamazaki et al., 2005). At-twi-expressing cMES cells appear to originate from multiple sites around the blastopore independently of cEND cells. This idea is reinforced by observation of evenly spaced cells expressing At-Delta in the central area of the germ disc epithelium at earlier stages(Fig. 3D-I) in combination with simultaneous detection of At-Delta and At-twi transcripts(Fig. 3L,M). It is likely that in normal development, the cells expressing At-Delta begin to express At-twi and enter the mesodermal pathway.
As is evident by time-lapse microscopy (see Movie 1 in the supplementary material), most surface cells of the germ disc do not significantly change positions during stages 4 and 5. Therefore, altering At-Deltaexpression patterns in the germ disc is not the result of cell movement but of progressive activation of At-Delta transcription. This process might be governed by a relay of short-range signals, by diffusion of long-range signals, or by graded activity of a maternal transcription factor.
Delta-Notch signaling is required to restrict the number of mesoderm cells
Defects seen after injection of At-DeltaEB,At-DeltaHH, At-Notch and At-Su(H) dsRNA were similar to each other, but completely different from those previously obtained with dsRNAs of two Achaearanea genes, At-dpp and At-short gastrulation (At-sog)(Akiyama-Oda and Oda, 2006). This situation suggests that the defects reveal specific developmental processes requiring canonical Delta-Notch signaling. Like Cupiennius,however, Achaearanea might have a second Delta gene and/or a second Notch gene. The possibility that such genes might have been additionally silenced in the RNAi experiments is not excluded.
Although it is a useful marker of future cMES cells, At-Delta does not seem to be necessary for differentiation into mesoderm or internalization. Defects in At-Delta RNAi embryos suggest that Delta-Notch signaling restricts the number of cMES cells, a function reminiscent of mechanisms controlling cell fate decisions in Drosophila early neurogenesis(Campos-Ortega, 1993). The evenly spaced but non-stereotyped arrangement of At-Delta-positive cells in the germ disc epithelium as well as the ratio of At-Delta-positive to -negative cells could be explained by lateral inhibition (Meinhardt and Gierer,1974; Honda et al.,1990). As the lateral inhibition hypothesis would predict,suppression of the inhibitory Delta signal as well as of At-Notch and At-Su(H) function abolishes interspersed expression of At-Delta and concomitantly causes overproduction of a single cell type (cMES cells) at the expense of other (caudal ectoderm cells) cell types(Fig. 7F,H; Fig. 8E-H). The cells that receive the At-Delta signal may be inhibited from transcribing At-Delta and At-twi in response to a different signal and entering the mesodermal pathway. The effect of blocking Delta-Notch signaling is that all the germ disc cells in the central area are allowed to transcribe At-Delta and At-twi and take the mesodermal fate.
Although Delta-Notch signaling does not function in specification of the primary mesoderm in Drosophila, its activity is required for specification of the mesectoderm (or the ventral midline cells) flanking the mesoderm in the cellularizing embryo(Menne and Klämbt, 1994; Martín-Bermudo et al.,1995; Morel and Schweisguth,2000; Bardin and Schweisguth,2006; De Renzis et al.,2006). Thus, in both Drosophila and Achaearanea,emerging mesoderm cells influence the fate of neighboring cells via Delta-Notch signaling. It is tempting to speculate that Drosophilamesectoderm specification and Achaearanea caudal ectoderm specification share an evolutionary link. This hypothesis would require that drastic changes in the mechanisms of ventral and caudal patterning occurred after divergence of the lineages leading to Drosophila and Achaearanea (Akiyama-Oda and Oda,2006).
By analogy with the evolutionarily conserved mechanism of Drosophila and vertebrate neurogenesis(Campos-Ortega, 1993; Henrique et al., 1995), in which the proneural genes promote Delta expression to confer neural competence(Campuzano and Modolell, 1992; Kunisch et al., 1994), a mesoderm-promoting gene(s) might exist upstream of At-Delta. Interestingly, in the pre-involuted mesoderm of the Xenopus gastrula,the muscle-determining factor MyoD has been identified as a direct positive regulator of Delta-1 expression(Wittenberger et al., 1999). Another example comes from studies of sea urchin embryos, in which a double repressor system involving pmar1 activates Delta transcription(Oliveri et al., 2002; Revilla-i-Domingo et al.,2004), which is necessary for secondary mesoderm cell specification (Sweet et al.,2002). Moreover, in the stem cell zone of the chick embryo spinal cord, FGF signaling functions to induce Delta1 expression(Akai et al., 2005). Future studies of spider embryos should identify a gene(s) activating At-Delta transcription around the blastopore.
In the early Drosophila embryo, the highest nuclear concentrations of the maternal transcription factor Dorsal directly activate zygotic transcription of the mesoderm-determining gene twi(Jiang et al., 1991). Characterization of differences in Drosophila and Achaearanea gene cascades leading to activation of twitranscription should also be the focus of future studies.
Delta-Notch signaling is required to induce the caudal ectoderm
Our results show that Delta-Notch signaling is essential to induce and specify caudal ectoderm in Achaearanea embryos. Caudal ectoderm is characterized by At-cad expression. Loss of the entire opisthosoma in embryos showing potent At-Delta RNAi phenotypes may be attributed to almost complete failure to generate At-cad-expressing caudal ectoderm cells. Our data indicate that zygotic At-cad transcription is activated downstream of Delta-Notch signaling. However, as there is a significant time lag (more than 10 hours) between initial elevation of At-Delta transcription around the closing blastopore (late stage 4)and the onset of At-cad transcription at the forming caudal lobe (mid stage 6), it is unlikely that At-cad is a direct target of Delta-Notch signaling.
In severe At-Delta RNAi embryos, overproduced mesoderm cells do not form repeating structures. Despite the fact that some mesoderm cells in the normally formed caudal lobe express At-cad(Akiyama-Oda and Oda, 2003), At-cad transcripts were not expressed in overproduced mesoderm cells in At-Delta RNAi embryos. Although At-twi transcripts disappear in the normally formed caudal lobe(Fig. 7I)(Yamazaki et al., 2005),overproduced mesoderm cells continued to express At-twi in the At-Delta RNAi embryos (Fig. 7J). These data suggest that regulation of gene expression in caudal lobe mesoderm may be influenced by ectoderm.
Opisthosomal defects that we observed in At-Delta RNAi embryos are due to failure in caudal lobe formation rather than in segmentation. Curiously, in previous Cupiennius studies, investigators described morphological abnormalities of the caudal region in embryos injected with dsRNAs targeted to components of the Delta-Notch signaling pathway(Stollewerk et al., 2003; Schoppmeier and Damen, 2005). The cause of those segmentation defects is unclear, and, in addition, mesoderm development has been poorly described in Cupiennius. Potential discrepancies between the Cupiennius and Achaearanea studies may arise from developmental variations between the species and/or from different experimental approaches to performing RNAi(Schoppmeier and Damen, 2001; Akiyama-Oda and Oda, 2006).
Mesoderm and caudal ectoderm formation are a single event required to establish a normal caudal lobe
We propose that in the process of caudal lobe formation, At-Delta functions both to prevent cells from adopting the mesoderm fate and to specify caudal ectoderm fate. These two outcomes may represent interdependent outputs of Notch-mediated genetic circuits that amplify small differences between initially equivalent neighboring cells(Artavanis-Tsakonas et al.,1999). Therefore, mesoderm and caudal ectoderm formation in Achaearanea are not independent events but a single one achieved by Delta-Notch signaling progressively activated in the germ disc epithelium from around the blastopore. Previous work on Delta-Notch signaling in spider revived the hypothesis that arthropod segmentation and vertebrate somitogenesis have a common origin(Stollewerk et al., 2003; Tautz, 2004; Patel, 2003). However, an unanswered question is how formation of mesodermal somites in vertebrates is related to arthropod segments consisting of ectoderm and mesoderm. Our finding that mesoderm and caudal ectoderm form as a single event in spider embryos may be relevant to this issue.
Roles for At-Delta in other processes of Achaearanea development
At-Delta expression was also observed at the rim and outside of the germ disc, and At-Delta RNAi phenotypes potentially related to this expression include reduction of At-fkh-expressing cells at the rim of the germ disc at stage 5 (Fig. 7A,B), absence of At-twi expression outside the emerging caudal region at stage 6 (Fig. 7E,F), and delayed or defective development of the prosomal region(Figs 4, 5). Further analysis of At-Delta function in prosomal development is needed.
Evolution of early patterning in arthropods
In Drosophila embryos, cascades of maternal and zygotic transcription factors make a major contribution to anterior, posterior,terminal and dorsoventral patterning, into which germ-layer specification is integrated. These patterning systems are achieved in a syncytial environment. Similar aspects of early patterning have been reported in another insect, Tribolium (Sommer and Tautz,1994; Chen et al.,2000). By contrast, our previous and present studies with the spider Achaearanea highlight the importance of cell-cell communication mediated by Dpp and Sog in dorsoventral patterning(Akiyama-Oda and Oda, 2006) and by Delta-Notch signaling in mesoderm and caudal specification. The necessity for cell-cell communication in early spider embryos may be due to early completion of cellularization (Kondo,1969; Suzuki and Kondo,1995). Based on Achaearanea studies, the blastopore region is particularly important, because cells producing Dpp signals originate from that region and progressive activation of Delta-Notch signaling starts there. Possibly, localized sources of intercellular signals, rather than gradients of maternal transcription factors, play central roles in early patterning in spider. Further studies with Achaearanea should contribute to a better understanding of the diversity of arthropod development and its ancestral mode.
We thank A. Noda for technical assistance and other members of JT Biohistory Research Hall for discussion and support.