The relative abundance of gatae and foxa transcripts in A. miniata embryos at various times from fertilization (dark gray squares represent data points for gatae, and light gray diamonds represent data points for foxa) were determined by using QPCR. Bars, ±SE.

Fate maps showing selected stages of starfish A. miniata (Left) and sea urchin S. purpuratus (Right) development. (i) Surface views from the vegetal (posterior) pole of a sixth cleavage S. purpuratus embryo (Right) and a blastula stage A. miniata embryo (Left). (ii-iv) Lateral optical sections; animal pole (Upper), vegetal pole (Lower), which are color-coded to display fate. Blastula (ii), gastrula (iii), and early larval stages from both taxa (iv). In iv, the oral side is to right and aboral to left in each drawing. Blue with pink stippling marks the endomesodermal veg₂ lineage of S. purpuratus, which resolves into endodermal (blue) and mesodermal (pink) by seventh cleavage (i), and A. miniata before blastula stage (ii). The mesoderm lineage can be further subdivided into coelomic (purple) and other mesodermal (pink) cell types. The echinoid-specific, micromere-derived skeletogenic lineage is shown throughout in red.

The gene tbr is regulated by other endomesodermal transcription factors in the starfish A. miniata, but not in the sea urchin S. purpuratus. (A) Comparison of the effects on tbr transcript abundance caused by perturbations of otx or gatae expression in sea urchins and starfish. Transcript levels consequent on perturbations are represented as in Fig. 2. Light gray and solid bars represent ΔC_T (left ordinate) or fold change (right ordinate) in tbr expression after the indicated perturbation in sea urchins and starfish, respectively. The significance of the difference in ΔC_T between sea urchin and starfish was determined by using a t test, and the resultant P values are provided. Error bars, ±SE. Sea urchin data can be accessed at http://sugp.caltech.edu/endomes. (B and C) Expression of tbr is required for endomesodermal specification in starfish. A. miniata (48 h) zygotes were injected with either 1 mM of a random sequence (control) MASO (B) or 600 μM αTbr MASO (C). The archenteron forms correctly in B but fails to develop when levels of Tbr are diminished in C.

A diagram of inferred architecture of the GRNs in S. purpuratus (A) and A. miniata (B). Data for A. miniata depend on the observations summarized in Figs. 2 and 6; for data for S. purpuratus, see ref. 5. Short horizontal lines represent the cis-regulatory region(s) controlling expression of the gene that is named beneath. Inputs to each cis-regulatory region are either positive (arrow) or negative (bar). Connections between genes were determined through various perturbation assays, and, in some cases in S. purpuratus, have been verified by direct cis-regulatory analyses. The genes krox, otx, bra, foxa, and gatae are all expressed within the endomesoderm and are required for its correct specification. In S. purpuratus, the tbr gene, its downstream target genes, and its upstream regulators are all expressed in the micromere/skeletogenic lineage and are necessary for its normal specification. The tbr gene in A. miniata is expressed within the endomesoderm and is required for its correct specification. Dashed lines indicate a regulatory connection observed in A. miniata but not present in S. purpuratus, or vice versa. For instance, the krox gene is self-regulating only in S. purpuratus, and Otx regulates tbr only in A. miniata. The positive regulatory feedback loops among krox, otx, and gatae that are present in both taxa are bold. Thinner lines, emanating from foxa, represent regulatory connections that exist only in later development (28-32 h in A. miniata).

Quantitative effects on expression levels of A. miniata transcription factors after various specific perturbations of gene expression. Zygotes were injected with either mRNA encoding an Otx-Eng fusion (A; ref. 25) or MASOs that block the translation of krox (B), gatae (C), or foxa (D) mRNA. The abundances of various transcripts in the experimental embryos are compared with their levels in similarly injected controls of the same batch of eggs and are indicated as ΔC_T (left ordinate), and as the corresponding fold change in transcript abundance (right ordinate). ΔC_T gives the difference between control and experimental samples in the number of PCR cycles required to attain threshold. Significantly affected transcripts (ΔC_T > 1.6, i.e., fold change of more than ≈3) are shown as filled bars. Stars indicate a significant difference with respect to S. purpuratus embryos (see text). All observations shown refer to 19-24 h blastulae except the two open bars in D, which display the later (28-32 h) effect of αfoxa MASO on bra and tbr expression. For individual measurements, see Table 1.

Visualization of regulatory interactions in perturbed A. miniata embryos by using WMISH. Views are lateral (A-F, I, and L) or from vegetal pole (G, H, J, and K). Embryos (except C) are stained after WMISH to reveal the localization of the transcript indicated at the bottom on the right. Zygotes were injected with 600 μM αfoxa MASO (D and I), or one of the first two blastomeres was injected with 1 mM αgatae MASO (C, E, and K), or 0.4 pg/pl otx-eng mRNA (J), as indicated at the top on the right. (A and B) Control gatae WMISH patterns: A, normal blastula; B, normal gastrula. (C) Blastula grown from a zygote injected with an mRNA of an in-frame fusion corresponding to the 5′ gatae sequence (containing the gatae MASO target site) with GFP, followed by an injection of 1 mM αgatae MASO into one blastomere at the two-cell stage. The loss of GFP expression from the half-embryo that results from this injected blastomere demonstrates that the αgatae MASO effectively binds to its target sequence and blocks translation in vivo. Gastrulation fails to initiate in this half-embryo. Zygotes injected with both this mRNA fusion and 1 mM control (random sequence) MASO expressed GFP throughout (data not shown). (D) Effect of blocking Foxa translation on gatae expression at blastula stage (compare A). (E) Effect of blocking Gatae translation on expression of gatae gene at gastrula stage (compare left and right halves). (F) Normal foxa expression at blastula stage. (G and H) Normal tbr expression viewed from vegetal pole of blastula (G) or gastrula (H). (I) Effect of αfoxa MASO on foxa RNA expression at blastula stage (compare with F). (J) Effect of Otx-Eng fusion on tbr expression at blastula stage (compare with G). (K) Effect of αgatae MASO on tbr expression at gastrula stage (compare with H). (L) Normal tbr expression in gastrula, viewed laterally. Double-headed arrows connect the relevant control embryos provided for comparison with its perturbed partner; each such comparison confirms a regulatory connection, which was inferred from the QPCR data summarized in Figs. 2 and 5.