nih public access yi-hsien su kuan-ting lin enhu li , and...

Gene Regulatory Control in the Sea Urchin Aboral Ectoderm:Spatial Initiation, Signaling Inputs, and Cell Fate Lockdown

Smadar Ben-Tabou de-Leon1,*, Yi-Hsien Su2,*, Kuan-Ting Lin2, Enhu Li3, and Eric H.Davidson3,*

1Department of Marine Biology, Leon H. Charney School of Marine Sciences, University of Haifa,Haifa 31905, Israel2Institute of Cellular and Organismic Biology, Academia Sinica, Nankang, Taipei 11529, Taiwan3Division of Biology 156-29, California Institute of Technology, Pasadena, CA 91125, USA

AbstractThe regulation of oral-aboral ectoderm specification in the sea urchin embryo has been extensivelystudied in recent years. The oral-aboral polarity is initially imposed downstream of a redoxgradient induced by asymmetric maternal distribution of mitochondria. Two TGF-β signalingpathways, Nodal and BMP, are then respectively utilized in the generation of oral and aboralregulatory states. However, a causal understanding of the regulation of aboral ectodermspecification has been lacking. In this work control of aboral ectoderm regulatory statespecification was revealed by combining detailed regulatory gene expression studies, perturbationand cis-regulatory analyses. Our analysis illuminates a dynamic system where different factorsdominate at different developmental times. We found that the initial activation of aboral genesdepends directly on the redox sensitive transcription factor, hypoxia inducible factor 1α (HIF-1α).Two BMP ligands, BMP2/4 and BMP5/8, then significantly enhance aboral regulatory genetranscription. Ultimately, encoded feedback wiring lock-down the aboral ectoderm regulatorystate. Our study elucidates the different regulatory mechanisms that sequentially dominate thespatial localization of aboral regulatory states.

Keywordsgene regulatory networks; cis-regulatory analysis; ectoderm specification; sea urchin;developmental control

IntroductionCell fate specification and differentiation are controlled by complex regulatory networksencoded in the genome (Davidson, 2006; Davidson, 2010). The architecture of generegulatory networks (GRNs) determines their information processing properties, and definesthe temporal order of specification events (Ben-Tabou de-Leon and Davidson, 2007). Thus

© 2012 Elsevier Inc. All rights reserved.*Correspondence: [email protected], 972-4-8288555, [email protected], 886-2-27899511, [email protected],626-395-4933.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to ourcustomers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review ofthe resulting proof before it is published in its final citable form. Please note that during the production process errors may bediscovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

NIH Public AccessAuthor ManuscriptDev Biol. Author manuscript; available in PMC 2014 February 01.

Published in final edited form as:Dev Biol. 2013 February 1; 374(1): 245–254. doi:10.1016/j.ydbio.2012.11.013.

$waterm

ark-text$w

atermark-text

$waterm

ark-text

the structure-function relations of GRNs as they progress through developmental timeprovide systems level understanding of the progressive molecular control of developmentalprocesses. The construction of experimentally based models of gene regulatory networksrequires advanced experimental tools and methodologies.

The sea urchin embryo presents key experimental and theoretical advantages for the study ofdevelopmental GRNs: a relatively simple developmental program that has been used toaddress fundamental questions in development (Horstadius, 1939), easy gene transfertechnology coupled with quantitative perturbation analysis of gene expression and a novelhigh throughput method for testing the expression level of more than a hundred cis-regulatory reporters in one experiment (Nam et al., 2010; Nam and Davidson, 2012).Predictive mathematical models were developed based on the knowledge gained from seaurchin regulatory studies, simulating the dynamics of gene regulatory circuits (Bolouri andDavidson, 2003; Ben-Tabou de-Leon and Davidson, 2009; Ben-Tabou de-Leon, 2010; Ben-Tabou de Leon and Davidson, 2010) and more recently, computing how the logic functionsutilized by GRNs give rise to unique spatial regulatory states (Peter et al., 2012). Theseexperimental and theoretical methodologies enabled the construction of one of the mostelaborate models of developmental GRNs, a model of the control system governingspecification of endomesoderm in the sea urchin embryo (Oliveri et al., 2008; Peter andDavidson, 2009; Peter and Davidson, 2011). It has the capacity to explain the molecularcontrol of various developmental phenomena, such as spatial and temporal patterning ofgene expression (Smith and Davidson, 2008; Peter and Davidson, 2011), the endoderm–mesoderm cell fate decision (Ben-Tabou de Leon and Davidson, 2010; Peter and Davidson,2011), and the activation of sets of differentiation and structural genes in a specific celllineage (Oliveri et al., 2008; Smith and Davidson, 2009). The recent computational modelmentioned above, demonstrates that the endomesoderm GRN in fact includes sufficientinformation to compute predicatively almost all known regulatory gene expression in spaceand time, in this portion of the embryo up to gastrulation (Peter et al.). Experimentally basedmodels of the GRNs governing specification of the various fate and regulatory state domainsof the sea urchin embryo ectoderm are now being constructed and carry the promise toilluminate the regulatory control of ectodermal patterning (Su et al., 2009; Li et al., 2012).

In its external developmental morphology, the sea urchin embryo ectoderm consists of fourprominent territories: the oral ectoderm, within which the mouth forms through fusion of theforegut and the ectoderm; the aboral ectoderm, which differentiates into squamousepithelium; the ciliary band, positioned at the border between oral and aboral ectodermaldomains; and the apical neurogenic domain (Fig. 1A left panel). The neurons of the seaurchin larva are mostly localized within the apical domain, in the ciliary band, and aroundthe mouth (Burke et al., 2006). Oral/aboral diversification along the secondary embryonicaxis of the sea urchin embryo forms as the result of an asymmetric redox gradient thatderives originally from uneven maternal mitochondrial distribution (Coffman and Davidson,2001; Coffman et al., 2004; Coffman et al., 2009). At the high end of this gradient, thefuture oral side of the embryo, the nodal gene is activated where its cis-regulatory controlsystem responds to a redox sensitive transcription factor (Coffman and Davidson, 2001;Coffman et al., 2004; Nam et al., 2007; Range et al., 2007; Coffman et al., 2009; Range andLepage, 2011). Nodal signaling then controls the specification and differentiation of the oralectoderm domain (Duboc et al., 2004). Targets of Nodal signaling within the oral ectoderminclude genes encoding both the ligand BMP2/4 and its antagonist Chordin (Duboc et al.;Duboc et al., 2004). The BMP ligand diffuses to the aboral ectoderm, to which its signalingactivity is confined, due to inhibition of BMP reception by Chordin within the oral ectoderm(Duboc et al., 2004; Chen et al., 2011).

de-Leon et al.

Dev Biol. Author manuscript; available in PMC 2014 February 01.

$waterm

ark-text$w

atermark-text

$waterm

ark-text

Recent work shows that the oral ectoderm contains an unexpectedly complex, bilaterallyarranged set of spatial regulatory state subdomains, (Li et al., 2012) and the same is true ofthe aboral ectoderm despite its seemingly uniform morphology (Chen et al., 2011). Prior togastrulation the aboral ectoderm specifically expresses a set of regulatory genes encodingthe transcription factors Tbx2/3, Irxa, Dlx, Msx, Hmx and Hox7 (Su et al., 2009; Chen et al.,2011). Although these genes turn on within a narrow time window (Fig. 1B (Materna et al.,2010)), their spatial expression is differential, defining regulatory state sub-regions withinthe aboral ectoderm (Chen et al., 2011). The dynamic changes in spatial subdomainexpression of these genes between early and late blastula stages are shown diagrammaticallyin Fig. 1C (Chen et al., 2011). Previous analyses revealed that the reception of BMP2/4enhances expression of the aboral ectoderm regulatory genes (Duboc et al., 2004; Su et al.,2009; Saudemont et al., 2011). However, the effect of BMP2/4 knock-down was less severethan the effect of BMP receptor knock-down (Yaguchi et al., 2010), and this suggests thatthere are additional BMP ligands operating at this time. Furthermore the aboral tbx2/3 geneis expressed before the phosphorylation of the BMP mediator, SMAD1/5/8, is detectable(Chen et al., 2011), indicating the presence of an early activator of aboral ectoderm genesnot related to BMP signaling. Thus, a redox-dependent aboral ectoderm transcription factorwas predicted to initiate the expression of tbx2/3 (Su et al., 2009). While intricate positiveregulatory interactions were suspected to exist among the aboral ectoderm transcriptionfactors (Su et al., 2009; Saudemont et al., 2011), prior to the present work temporalresolution of observations indicating such interactions was lacking, particularly for earlytime points.

Here we study the dynamic regulatory control of the aboral ectoderm specification bycombining gene expression studies, perturbation analysis, and cis-regulatory analyses. Weshow that the redox sensitive transcription factor HIF1α (Hypoxia Inducible Factor 1α) isdirectly activating the early expression of aboral ectoderm regulatory genes, possiblymediating the primordial redox gradient. Both BMP2/4 and BMP5/8 then contribute to themagnitude of expression of the aboral ectoderm regulatory genes, and their inputs wereverified for key genes at the cis-regulatory level. As typical for specification GRNs, thesystem later graduates to the use of encoded cross-regulatory feedback interactions, whichthereafter internally control its transcriptional functions.

Materials and methodsGene expression and perturbation analysis

Morpholino-substituted antisense oligonucleotides (MOs) specific to the ectodermregulatory genes were obtained from Gene Tools (Philomath, OR), the sequences of whichare shown in Supplementary Information, Table 1. For perturbation analyses, eggs wereinjected with 100–400 µM MO. RNA from uninjected control and MO injected embryoswas isolated by RNeasy Micro kit (Qiagen). RNA samples were reverse transcribed byiScript cDNA synthesis kit (BioRad) for quantitative PCR (QPCR). To monitor thequantitative effects of each perturbation, data were normalized to the amount of ubiquitinmRNA as described (Revilla-i-Domingo et al., 2004). QPCR primers used in this study arealso listed in Table S1. Whole mount in situ hybridization was performed as described in(Oliveri et al., 2006).

Microinjection and QPCR Measurement of GFP mRNA in Eggs Expressing GFP ConstructsPCR products were purified with the Qiagen Qiaquick PCR purification Kit or Zymo DNAclean and concentrator. The fragments were then microinjected into fertilized S. purpuratuseggs as described (Rast, 2000). Linearized BAC constructs were desalted by drop dialysisinto TE buffer on a 0.025 µm VSWP filter (Millipore). Approximately 500–800 molecules

de-Leon et al.


$waterm

ark-text$w

atermark-text

$waterm

ark-text

of the desired reporter construct were injected, along with a 3–6 fold molar excess ofHindIII-digested carrier sea urchin DNA per egg, in a 4 pl volume of 0.12 M KCl. A similarinjection solution was made for BAC reporters but with 100–200 copies of the BAC per 4pl. Embryos were collected at different stages for assessment of spatial activity byfluorescence microscopy, or for quantitative analysis of transcript prevalence by QPCR. ForQPCR measurements RNA and DNA of about 150 injected embryos was isolated usingQiagen Allprep kit. The RNA was then reverse transcribed to cDNA using iScript cDNAsynthesis Kit of Bio-Rad. QPCR for both DNA and cDNA was performed as described(Revilla-i-Domingo et al., 2004). The level of cDNA was computed by comparison to aninternal standard (ubiquitin) cDNA, and the level of injected DNA was computed incomparison to Nodal DNA.

ConstructsStandard PCR and fusion PCR techniques using the High Fidelity PCR Kit (Roche) andLong template High Fidelity PCR Kit (Roche) were used to build reporter constructs.Binding-site sequences were mutated by PCR, and the resulting constructs were checked bysequencing. The mutation PCR primers were designed with about 20 bp sequencessurrounding both ends of the target site. The target site was changed into a mutant form ofthe candidate transcription factor binding sites. Whole mount in situ hybridization wasperformed as described in (Oliveri et al., 2006).

ResultsEffects of BMP2/4 and BMP5/8 perturbations on sea urchin embryo development and onaboral ectoderm patterning

The gene encoding the ligand BMP2/4 is expressed in the oral ectoderm, starting at about 12hours post fertilization (hpf) (Fig. 2A; (Duboc et al., 2004; Su et al., 2009)). A second TGFβgene encoding BMP5/8 is expressed maternally at a very low level and then at much higherlevels zygotically, beginning at about 8 hpf. Its zygotic expression is uniform in theectoderm, spanning all the ectodermal territories (Figs. 2A, B). To study the developmentalroles of the different BMP ligands we injected fertilized sea urchin eggs with Morpholinoantisense oligonucleotides (MO) that block the translation or the splicing of given genes(For MO sequences see Supplemental Information, Table S1). Embryos injected withBMP2/4 MO display severely reduced aboral expansion; while skeletal rods are formed,even at 72hpf they are not extended at the aboral vertex to form the distinct structure of thepluteus larva (Fig. S1A, B; (Lapraz et al., 2009; Saudemont et al., 2011)). When theembryos are co-injected with BMP2/4 and BMP5/8 MO’s, in addition to the reduced aboralexpansion and skeletal patterning defects, the partitioning of the gut is affected (Fig. S1C).The effect of BMP5/8 MO injection is a mild reduction of the aboral expansion (Fig. S1E).Skeletal patterning depends on ectodermal signaling (Armstrong et al., 1993), and it is likelythat the BMP input is indirectly required for skeletogenic ectodermal signaling; thus theregulatory programs that it drives are ultimately important for correct morphologicalpatterning of the embryo.

To study the contribution of the BMP ligands to the establishment of the aboral ectodermregulatory state we investigated the effect of BMP2/4 and BMP5/8 MO injection on thespatial expression of tbx2/3, irxa and dlx at 19hpf and 24hpf by whole mount in-situhybridization (WMISH) (Figs. 2C and S1). Neither MO alone nor the two in combinationhas a strong visible effect on expression of irxa at the vegetal/aboral ectoderm boundary,though both obliterate irxa expression farther up in the aboral ectoderm indicating that irxais driven by additional activators in vegetal portion of its domain of aboral ectodermexpression (Fig. 2C, compare panels 1 and 4). The expression of tbx2/3 and dlx is

de-Leon et al.


$waterm

ark-text$w

atermark-text

$waterm

ark-text

significantly reduced at these times by BMP2/4 MO injection, or by co-injection of BMP2/4and BMP5/8 MO. The injection of BMP5/8 MO alone reduces the expression of dlx, but hasonly a weak effect on the expression of tbx2/3. These results agree with previouslypublished results of BMP2/4 perturbations (Saudemont et al., 2011). The additional findingis that BMP5/8 is also necessary for normal expression of both irxa and dlx.

Expression and developmental role of the hypoxia-dependent transcription factor HIF1αAs described in Introduction, maternal anisotropy of the mitochondrial distribution generatesa redox gradient along which oral-aboral polarization occurs (Coffman and Davidson, 2001;Coffman et al., 2004; Coffman et al., 2009). High mitochondrial concentration and theresulting high redox potential and oxidizing environment are correlated with oral fate(Coffman and Davidson, 2001; Coffman et al., 2004; Coffman et al., 2009). A redoxsensitive factor responding to the reducing environment on the opposite aboral side had beenpredicted to initiate the early expression of tbx2/3 (Su et al., 2009), but the responsiblefactor had not been identified. Here we suggest that the well known oxygen sensitivetranscription factor HIF1α (hypoxia-inducible factor 1α) executes this role in aboralectoderm specification early in sea urchin embryogenesis. HIF is a heterodimerictranscription factor which in vertebrates consists of one of three different redox-sensitiveHIFα subunits (HIF1α, HIF2α, and HIF3α) complexed with a common constitutivesubunit, HIFβ (Wenger et al., 2005). HIF1αβ and HIF2αβ heterodimers function astranscriptional activators of oxygen-regulated target genes. Under normoxic conditions theHIFα subunits are hydroxylated by a family of redox-sensitive prolyl hydroxylases (PHD)and the hydroxylated forms are targeted for degradation (Ivan et al., 2001). Hypoxicconditions decrease the activity of PHD, resulting in rapid HIFα subunit accumulation(Jewell et al., 2001). Recent studies have shown that a reducing environment is required tostabilize HIF-1α under hypoxic conditions (Chen and Shi, 2008; Guo et al., 2008). ThusHIF1α is a likely candidate for a transcription factor that operates specifically in conditionsof reducing environment and low redox potential. Sea urchin has a single HIF1α geneencoded in the genome which is structurally similar to the vertebrate HIF1α and HIF2α(Loenarz et al., 2011; Rytkonen et al., 2011). SpHIF1α is maternal and its transcripts arepresent in all the cells of the embryo, at least up to 12 hpf (Figs.3A, B). Its zygoticexpression starts at about 18hpf (Fig.3A). The HIF1α gene is then expressed specifically inthe non-skeletogenic mesoderm, small micromeres, and their postgastrular coelomic pouchdescendants (Fig. 3B), so only the maternal phase of expression is relevant to this study.Injection of HIF1α MO reduces aboral expansion at 72hpf, (Fig.S2) and severely reducedthe expression tbx2/3 at 19hpf, less so at 24hpf (Fig.3C, S3–S5). The effect of HIF1α MOon irxa expression was less pronounced at these times (Fig.3C, S3–S5).

Early drivers of the aboral ectoderm regulatory state, as revealed by quantitativeperturbation analysis

To study gene interactions through time, particularly those responsible for initiation of geneexpression when the levels are at first low, a sensitive quantitative analysis is required. Tothat end we injected fertilized sea urchin eggs with specific MO and measured theexpression levels of selected regulatory genes in the aboral ectoderm and in other embryonicterritories at three time points, using quantitative PCR (QPCR) as described in many earlierGRN analyses (Oliveri et al., 2008; Peter and Davidson, 2011) (See table S1 for MO andQPCR primer sequences). A summary of these perturbation results is presented in Fig. 4.The matrices in Fig. 4A–C indicate perturbations that reduced the average ratio between theexpression level of a gene and the expression level at control MO to less than 0.4, at 16hpf(Fig. 4A, S3), 19hpf (Fig. 4B, S4) or 24hpf (Fig. 4C, S5). At these early times allinteractions observed were positive. Average ratios between the expression level of theaboral ectoderm transcription factors in specific MO and control MO for all perturbations

de-Leon et al.


$waterm

ark-text$w

atermark-text

$waterm

ark-text

are presented in Figs. S3–5. The effect of all perturbations on all tested genes is presented inFig. S6.

A glance at the matrices of interaction effects at 16, 19, and 24 hpf in Fig.4 shows that thereis relatively rapid change in the inputs to given downstream genes. Tbx2/3 and HIF1α areearly drivers of the aboral ectoderm regulatory genes. Prominent effects of interference withtranslation of these factors are seen at the two earliest time points (Figs. 4A, B, S3, 4, 6). At16hpf and 19hpf HIF1α MO strongly reduces the expression level of tbx2/3, dlx, and msx,and has a significant effect on the level of hmx (Fig. 4A,B, S3,4). But by 24hpf the effect ofHIF1α MO on the expression level of the aboral ectoderm transcription factors is muchweaker (Fig. S5). Since, as we note above, only HIF1α translated from the ubiquitousmaternal mRNA (Fig.3A, B) can be present in ectodermal cells, the disappearance of thematernal message at 18h predicts this result. Tbx2/3 MO significantly reduces the levels ofdlx, hox7 and msx at 16hpf and these genes and irxa at 19hpf (Figs. 4A, B, S3,4,6). At 24hpfit affects only hox7 (Figs. 4C, S5, 6). These results indicate that the tbx2/3 gene is animportant driver of the aboral ectoderm genes, particularly in the early phase of theirexpression. The decreasing response to Tbx2/3 MO at 24h (Fig. 4C) could be due to theclearance of tbx2/3 expression from the animal domain of the aboral ectoderm at this time,as this reduces overlap of its spatial expression with other aboral ectoderm genes (Fig. 1C).Since these target genes continue to be expressed, additional regulatory inputs are evidentlycoming into play.

Roles of the BMP inputsThe effects of BMP2/4 and BMP5/8 MO, individually and together, were similarly assessedby QPCR (Figs. 4, S3–6). At 16hpf co-injection of BMP2/4 and BMP5/8 MO decreases theexpression of the aboral ectoderm regulatory genes, but the variation in the measurement arequite significant at this time except for the genes tbx2/3 and hox7 (Fig, 4A, S3). At this time,the injection of BMP2/4 MO alone does not produce any significant effect on the aboralgenes (Fig. S3) while the injection of BMP5/8 MO does reduce the expression of hox7, msxand dlx (Fig. 4A, S3C, E). This relatively weak effect of the BMP perturbation at 16hpfagrees with the weak phosphorylated SMAD1/5/8 signal at this time (Chen et al., 2011). Thestronger effect of BMP5/8 MO compared to BMP2/4 MO at this early time point correlateswith the earlier expression of BMP5/8 compared to BMP2/4 (Fig. 2A). However, at 19hpfco-injection of BMP2/4 MO and BMP5/8 MO together significantly decreases theexpression level of the aboral ectoderm genes except from hmx (Fig. 4B, S4, S6). At thistime injection of BMP2/4 MO or BMP5/8 MO alone also affects expression levels but lessso than when the two reagents are combined. Even stronger effects are seen at 24hpf (Fig.4).These results agree with the distinct phosphorylated SMAD1/5/8 signal at 18hpf and thestronger signal at 24hpf (Chen et al., 2011). In summary, the BMP pathway enhances theexpression of all the aboral ectoderm regulatory genes studied, starting at about 19hpf andsupplying a sharp boost in the expression levels at 24hpf. Apparently, the BMP pathwaydoes not reach its full activation strength before the ligand BMP2/4 is at work, at about 18–19hpf.

Positive regulatory interactionsInjections of irxa and dlx MO's show positive regulatory interactions among the aboralectoderm regulatory genes at 19hpf and 24hpf (Figs. 4B–C). The results indicate positivefeedback loops among the aboral ectoderm regulatory genes. For example we see that dlxhas a positive input to irxa while irxa also has a positive input to dlx. Nested feedbacksamong some of these genes were indicated in our initial draft GRN (Su et al., 2009) andproposed to be necessary for the maintenance of expression of the aboral ectodermregulatory genes.

de-Leon et al.


$waterm

ark-text$w

atermark-text

$waterm

ark-text

Perturbation analysis kinetics predict direct inputs of the BMP pathwayThe perturbation analysis reveals a dynamic system where different factors dominate atdifferent time points (Fig. 4). In Fig. 5A–F we portray the expression levels of the aboralectoderm transcription factors at 16hpf, 19hpf and 24hpf, at control MO, Tbx2/3 MO andBMP2/4+BMP5/8 MO. At 16hpf tbx2/3 expression is severely reduced due to theperturbation of the BMP pathway (Fig. 5A). At this time tbx2/3 perturbation significantlyaffects most of the aboral ectoderm genes (Fig. 5B–F). This raises the following question:does the BMP reception directly activate all the aboral ectoderm transcription factorsthrough the phosphorylated SMAD1/5/8, or does the BMP pathway affects the aboralectoderm genes indirectly, through the activation of key regulators, such as Tbx2/3? To gaininsights into the expected dynamic response to BMP perturbation at different networkarchitectures, we applied a mathematical model for network kinetics that was developedearlier (Davidson, 1986; Bolouri and Davidson, 2003; Ben-Tabou de-Leon and Davidson,2009; Ben-Tabou de-Leon, 2010).

We simulated the response of a downstream gene to the perturbations of the BMP pathwayand the transcription factor Tbx2/3 at direct and indirect network connectivity, (Fig. 5G–I.The model equations are presented in the supplemental information.). The model wasconstructed based on the following experimental observations: tbx2/3 is responding toHIF1α and BMP perturbations at 16hpf (Fig. S3, 6), and therefore in the model we assumethat tbx2/3 is activated by two inputs, pSMAD1/5/8 (BMP) and A1 (e.g., HIF1α). Theexpression of tbx2/3 precedes the phosphorylation of SMAD1/5/8 (Chen et al., 2011) andwe included that in the model (Fig. 5G). Additionally, the effect of BMP perturbation on theexpression level of the aboral ectoderm genes is strongest compared to other perturbations at19hpf and 24hpf (Fig. 5A–F, Fig S4, 5), and particularly on tbx2/3 expression at 16hpf (Fig.4A, S3A). Therefore we assumed that the activation strength of pSMAD1/5/8 is larger thanof the other inputs. The simulations of the effect of BMP and Tbx2/3 perturbation on adownstream gene are presented in Fig. 5H (indirect connectivity) and in Fig. 5I (directfeedforward connectivity). The simulation of direct activation of the downstream gene byBMP in a feedforward structure with Tbx2/3 agrees well with the dynamic perturbation data(Fig. 5A–F), strengthening the prediction of direct activation of these genes bypSMAD1/5/8.

Cis-regulatory control of tbx2/3 and dlx, two key activators of the aboral ectoderm GRNTo verify the predictions of the perturbation analysis (Figs 4, 5) we conducted cis-regulatoryanalyses of the genes tbx2/3 and dlx, two of the important activators in the aboral ectodermGRN (Figs. 4A–C, (Su et al., 2009)). High through-put screening analysis identifiedfunctional conserved regulatory elements in the vicinity of tbx2/3 and dlx (Nam et al.,2010). We constructed GFP reporter constructs driven by these elements and compared theirexpression profiles with those of GFP recombinant BACs containing the tbx2/3 and dlxgenes plus large genomic regions surrounding these genes (BAC and short construct maps,Fig. 6A, tbx2/3; 6D, dlx). The comparison shows similarity of spatial expression as well asof levels of expression between the recombinant BACs and the short reporter constructs.Thus the short constructs contain sufficient regulatory information to drive the expression oftbx2/3 and dlx, (Figs. 6B,C,E,F S7A,B,E,F).

Deletion of the 5' region annotated "U" from the tbx2/3 regulatory element reduces theexpression level of the reporter by about 50% at 19hpf and 24hpf indicating that this regioncontains important regulatory information (Fig. S7C,D). The intact tbx2/3 U:B:P:GFPreporter construct responds to HIF1α MO at 19hpf and 24hpf (Fig. 6H, light blue bars) andto BMP 2/4 MO at 24hpf (Fig. 6G, light green bars). Mutations of the four consensusSMAD sites and of the single HIF1α consensus site in "U", reduces the reporter expression

de-Leon et al.


$waterm

ark-text$w

atermark-text

$waterm

ark-text

at the same time points at which the response to the relevant MO is observed, (SMAD, Fig.6G, dark green bars, HIF1α, Fig. 6H, red bars. See supplemental data for the completesequences of regulatory elements; Table S2 for binding site sequences, and Table S3 formutation statistics). These results confirm that the BMP pathway and HIF1α directlyactivate tbx2/3 through their consensus target sites, in agreement with the perturbationpredictions of Fig. 4.

Deletion of the 5' regions annotated "U" and "B" from the dlx regulatory element reducesthe expression level of this reporter by about 40% at 24hpf indicating that these regionscontain important regulatory information (Fig. S7H). The intact dlx U:B:P:GFP reporterconstruct responds to BMP2/4+BMP5/8 MO and to Tbx2/3 MO at 24hpf, (BMP, Fig. 6I,light green bars; Tbx2/3, Fig. 6J, purple bars) and to HIF1α MO at 19hpf (Fig. 6J, light bluebars). Mutations of the two SMAD consensus sites in "U", the two consensus Tbx2/3 sites in"U" and consensus HIF1α sites in "U" and "B" reduce the expression of the dlx reporterconstruct at the same time points where the response to the relevant MO is observed,(SMAD, Fig. 6I, dark green bars; HIF1α, Fig. 6J, red bars; Tbx2/3 Fig. 6J, blue bars). ThusBMP, Tbx2/3 and HIF1α provide direct inputs into dlx, in agreement with the perturbationresults (Fig. 4). These direct BMP inputs were not predicted earlier (Su et al., 2009) thoughthe tbx2/3 input into dlx was, and though its identity was not known, so was a redoxsensitive activating input into tbx2/3.

The timing of the effect of mutations and perturbations on the level of reporter constructs,Fig. 6, seem to be delayed compared to the effects of the perturbations on the endogenousgenes, Fig. 4. This is quite common when comparing reporter constructs and endogenousgenes due to the experimental differences of the two assays and the mosaic incorporation ofthe reporters in the sea urchin embryos. Therefore we assume that the exact timing of theactual regulatory connection in the sea urchin embryo correspond to the timing observed bythe perturbation of the endogenous genes.

DiscussionSpecification of the aboral ectoderm is a progressive and spatially complex process. Asshown diagrammatically in Fig.1, there are distinct regulatory state domains arrayed alongthe animal-vegetal dimension. In a way this is paradoxical, as the initial studies on the aboralectoderm of sea urchin embryos depicted this territory as a uniform squamous epithelium inwhich downstream genes such as spec1, spec2a and cy3a actin are uniformly expressed(Cox et al., 1986; Hardin et al., 1988; Kirchhamer and Davidson, 1996). However, we arehere examining regulatory functions, not embryonic cell differentiation, and it is now clearfrom other domains of this embryo that a progression of spatial regulatory states precedesthe spatial resolution of the cell fates to which the territory eventually gives rise (Peter andDavidson, 2011). We may be here examining the initial stages of a complex partitioning ofwhat we have viewed as a single territory into multiple different regulatory domains; thefunctional biological significance of these regulatory domains might not be played out untillater stages of larval development.

In Fig.7, we show the network of regulatory interactions indicated among the genes studiedhere at the aboral ectoderm, at three different times, 16, 19, and 24hpf. This model showsthe significant findings of this work: These are the initial role of the redox sensitive maternalfactor HIF1α in activating the earliest regulatory genes on this side of the embryo,particularly the “pioneer” gene, tbx2/3; the continued strong driver architecture in whichTbx2/3 and BMP erect feed forward inputs into other regulatory genes; and the web ofcross-regulatory interactions that as we earlier proposed (Su et al., 2009) lock these circuitsinto place, after the HIF1α input has disappeared.

de-Leon et al.


$waterm

ark-text$w

atermark-text

$waterm

ark-text

The new discovery of the role of HIF1α suggests that both ends of the redox asymmetry setup in the egg by unequal distribution of maternal mitochondria are utilized at thetranscriptional regulatory level. As reviewed above, on the oral side, high redox state is usedto activate the nodal cis-regulatory system (Nam et al., 2007; Range et al., 2007). Here wesee that on the aboral side the redox sensitive factor, HIF1α, directly activates aboraltranscription factors (Fig 6), possibly due to HIF1α stabilization induced by the reducingenvironment on this side (Coffman and Davidson, 2001; Coffman et al., 2004; Coffman etal., 2009). This could solve the problem of how the aboral ectoderm regulatory state isinitiated in the right place in the embryo, an important implication here bolstered by directcis-regulatory analysis. Further studies of the spatial pattern of the protein HIF1α at theseearly developmental stages are expected to shed light on this patterning mechanism. TheBMP2/4/Chordin device ensures that though originating in the oral ectoderm, this activatinginput is experienced by cells only on the aboral side. But at the earliest times this input is notyet strongly functional. Most of the BMP input then comes from the ubiquitously buttransiently expressed BMP5/8. Throughout, the role of BMP signaling in the aboralectoderm is fundamentally that of a transcriptional booster and maintenance. Thisconclusion too is now based on cis-regulatory results which entirely confirm predictionsfrom BMP perturbation results and from the subcircuit architecture shown in Fig.7. Thus,the view which assigns to the oral ectoderm the role of the primary organizer of both oraland aboral ectoderm specification, via expression of the diffusible BMP2/4 ligand, isprobably incomplete. Our study implies that the cytoarchitecture of the egg might be utilizedon both oral and aboral sides of the embryo to control the differential initial regulatorystates.

As this work was in progress additional regulatory genes were discovered to be expressed inthe aboral ectoderm at these and later stages. Thus the aboral ectoderm GRNs can beexpected to become regionally diversified and progressively more complex as embryonicdevelopment proceeds. Further analysis will hopefully explain the differential spatialexpression of the aboral regulatory genes along the animal-vegetal dimension (Fig. 1C,(Chen et al., 2011)). HIF1α and the BMP pathway activate genes that are expressed in allthe spatial domains of the aboral ectoderm and therefore cannot explain this differentialexpression pattern. Neither can the positive interactions between the aboral ectodermtranscription factors explain the exclusion of hmx to the animal domain and msx and latertbx2/3 to the central and vegetal domains of the aboral ectoderm (Fig. 1C, (Chen et al.,2011)). Local repression is likely to be responsible for the spatial restriction of these genesto sub-domains within the aboral ectoderm (Fig. 7).

ConclusionsIn the most general terms this work is another step toward a goal that only a few years agowould have seemed utterly unattainable: extension of GRN models to encompass the wholeof an embryo. When this is achieved the inputs into the regulatory genes of the model willall be outputs from other genes of the model, and all signaling interactions will begin andend in the genomic regulatory apparatus. Explicit knowledge of the program forembryogenesis will thus arrive at a level of logical completeness, and this is now most likelyto be attained for sea urchin embryos.

Supplementary MaterialRefer to Web version on PubMed Central for supplementary material.

de-Leon et al.


$waterm

ark-text$w

atermark-text

$waterm

ark-text

AcknowledgmentsResearch was supported by NIH grant HD-037105 to EHD and by National Science Council grants 99-2627-B-001-003 and 101-2923-B-001-004-MY2 to YS.

ReferencesArmstrong N, Hardin J, McClay DR. Cell-cell interactions regulate skeleton formation in the sea

urchin embryo. Development. 1993; 119:833–840. [PubMed: 8187642]

Ben-Tabou de-Leon S. Perturbation analysis analyzed-mathematical modeling of intact and perturbedgene regulatory circuits for animal development. Dev Biol. 2010; 344:1110–1118. [PubMed:20599898]

Ben-Tabou de-Leon S, Davidson EH. Gene regulation: gene control network in development. AnnuRev Biophys Biomol Struct. 2007; 36:191. [PubMed: 17291181]

Ben-Tabou de-Leon S, Davidson EH. Modeling the dynamics of transcriptional gene regulatorynetworks for animal development. Dev Biol. 2009; 325:317–328. [PubMed: 19028486]

Ben-Tabou de Leon S, Davidson EH. Information processing at the foxa node of the sea urchinendomesoderm specification network. Proc Natl Acad Sci U S A. 2010; 107:10103–10108.[PubMed: 20479235]

Bolouri H, Davidson EH. Transcriptional regulatory cascades in development: initial rates, not steadystate, determine network kinetics. Proc Natl Acad Sci U S A. 2003; 100:9371–9376. [PubMed:12883007]

Burke RD, Angerer LM, Elphick MR, Humphrey GW, Yaguchi S, Kiyama T, Liang S, Mu X, Agca C,Klein WH, Brandhorst BP, Rowe M, Wilson K, Churcher AM, Taylor JS, Chen N, Murray G,Wang D, Mellott D, Olinski R, Hallbook F, Thorndyke MC. A genomic view of the sea urchinnervous system. Dev Biol. 2006; 300:434–460. [PubMed: 16965768]

Chen H, Shi H. A reducing environment stabilizes HIF-2alpha in SH-SY5Y cells under hypoxicconditions. FEBS Lett. 2008; 582:3899–3902. [PubMed: 18957292]

Chen JH, Luo YJ, Su YH. The dynamic gene expression patterns of transcription factors constitutingthe sea urchin aboral ectoderm gene regulatory network. Dev Dyn. 2011; 240:250–260. [PubMed:21181943]

Coffman JA, Coluccio A, Planchart A, Robertson AJ. Oral-aboral axis specification in the sea urchinembryo III. Role of mitochondrial redox signaling via H2O2. Dev Biol. 2009; 330:123–130.[PubMed: 19328778]

Coffman JA, Davidson EH. Oral-aboral axis specification in the sea urchin embryo. I. Axisentrainment by respiratory asymmetry. Dev Biol. 2001; 230:18–28. [PubMed: 11161559]

Coffman JA, McCarthy JJ, Dickey-Sims C, Robertson AJ. Oral-aboral axis specification in the seaurchin embryo II. Mitochondrial distribution and redox state contribute to establishing polarity inStrongylocentrotus purpuratus. Dev Biol. 2004; 273:160–171. [PubMed: 15302605]

Cox KH, Angerer LM, Lee JJ, Davidson EH, Angerer RC. Cell lineage-specific programs ofexpression of multiple actin genes during sea urchin embryogenesis. J Mol Biol. 1986; 188:159–172. [PubMed: 3723594]

Davidson, EH. Gene Activity in Eraly Development. Orlando: Academic Press, Inc.; 1986.

Davidson, EH. The regulatory genome: gene regulatory networks in development and evolution. San-Diego: Academic press; 2006.

Davidson EH. Emerging properties of animal gene regulatory networks. Nature. 2010; 468:911–920.[PubMed: 21164479]

Duboc V, Lapraz F, Saudemont A, Bessodes N, Mekpoh F, Haillot E, Quirin M, Lepage T. Nodal andBMP2/4 pattern the mesoderm and endoderm during development of the sea urchin embryo.Development. 2010; 137:223–235. [PubMed: 20040489]

Duboc V, Rottinger E, Besnardeau L, Lepage T. Nodal and BMP2/4 signaling organizes the oral-aboral axis of the sea urchin embryo. Dev Cell. 2004; 6:397–410. [PubMed: 15030762]

Guo S, Bragina O, Xu Y, Cao Z, Chen H, Zhou B, Morgan M, Lin Y, Jiang BH, Liu KJ, Shi H.Glucose up-regulates HIF-1 alpha expression in primary cortical neurons in response to hypoxia

de-Leon et al.


$waterm

ark-text$w

atermark-text

$waterm

ark-text

through maintaining cellular redox status. J Neurochem. 2008; 105:1849–1860. [PubMed:18266932]

Hardin PE, Angerer LM, Hardin SH, Angerer RC, Klein WH. Spec2 genes of Strongylocentrotuspurpuratus. Structure and differential expression in embryonic aboral ectoderm cells. J Mol Biol.1988; 202:417–431. [PubMed: 3172223]

Horstadius S. The mechanics of sea urchins development studied by operative methods. Biol. Rev.1939; 14:132–179.

Ivan M, Kondo K, Yang H, Kim W, Valiando J, Ohh M, Salic A, Asara JM, Lane WS, Kaelin WG Jr.HIFalpha targeted for VHL-mediated destruction by proline hydroxylation: implications for O2sensing. Science. 2001; 292:464–468. [PubMed: 11292862]

Jewell UR, Kvietikova I, Scheid A, Bauer C, Wenger RH, Gassmann M. Induction of HIF-1alpha inresponse to hypoxia is instantaneous. FASEB J. 2001; 15:1312–1314. [PubMed: 11344124]

Kirchhamer CV, Davidson EH. Spatial and temporal information processing in the sea urchin embryo:modular and intramodular organization of the CyIIIa gene cis-regulatory system. Development.1996; 122:333–348. [PubMed: 8565846]

Lapraz F, Besnardeau L, Lepage T. Patterning of the dorsal-ventral axis in echinoderms: insights intothe evolution of the BMP-chordin signaling network. PLoS Biol. 2009; 7 e1000248.

Li E, Materna SC, Davidson EH. Direct and indirect control of oral ectoderm regulatory geneexpression by Nodal signaling in the sea urchin embryo. Dev Biol. 2012; 369:377–385. [PubMed:22771578]

Loenarz C, Coleman ML, Boleininger A, Schierwater B, Holland PW, Ratcliffe PJ, Schofield CJ. Thehypoxia-inducible transcription factor pathway regulates oxygen sensing in the simplest animal,Trichoplax adhaerens. EMBO Rep. 2011; 12:63–70. [PubMed: 21109780]

Materna SC, Nam J, Davidson EH. High accuracy, high-resolution prevalence measurement for themajority of locally expressed regulatory genes in early sea urchin development. Gene ExprPatterns. 2010; 10:177–184. [PubMed: 20398801]

Nam J, Davidson EH. Barcoded DNA-tag reporters for multiplex cis-regulatory analysis. PLoS One.2012; 7:e35934. [PubMed: 22563420]

Nam J, Dong P, Tarpine R, Istrail S, Davidson EH. Functional cis-regulatory genomics for systemsbiology. Proc Natl Acad Sci U S A. 2010; 107:3930–3935. [PubMed: 20142491]

Nam J, Su YH, Lee PY, Robertson AJ, Coffman JA, Davidson EH. Cis-regulatory control of the nodalgene, initiator of the sea urchin oral ectoderm gene network. Dev Biol. 2007; 306:860–869.[PubMed: 17451671]

Oliveri P, Tu Q, Davidson EH. Global regulatory logic for specification of an embryonic cell lineage.Proc Natl Acad Sci U S A. 2008; 105:5955–5962. [PubMed: 18413610]

Oliveri P, Walton KD, Davidson EH, McClay DR. Repression of mesodermal fate by foxa, a keyendoderm regulator of the sea urchin embryo. Development. 2006; 133:4173–4181. [PubMed:17038513]

Peter I, Faure E, Davidson EH. Predictive computation of genomic logic processing functions inembryonic development. Proc Natl Acad Sci U S A. 2012; 109:16434–16442. [PubMed:22927416]

Peter IS, Davidson EH. Modularity and design principles in the sea urchin embryo gene regulatorynetwork. FEBS Lett. 2009; 583:3948–3958. [PubMed: 19932099]

Peter IS, Davidson EH. A gene regulatory network controlling the embryonic specification ofendoderm. Nature. 2011; 474:635–639. [PubMed: 21623371]

Range R, Lapraz F, Quirin M, Marro S, Besnardeau L, Lepage T. Cis-regulatory analysis of nodal andmaternal control of dorsal-ventral axis formation by Univin, a TGF-beta related to Vg1.Development. 2007; 134:3649–3664. [PubMed: 17855430]

Range R, Lepage T. Maternal Oct1/2 is required for Nodal and Vg1/Univin expression during dorsal-ventral axis specification in the sea urchin embryo. Dev Biol. 2011; 357:440–449. [PubMed:21782809]

Rast JP. Transgenic manipulation of the sea urchin embryo. Methods Mol Biol. 2000; 136:365–373.[PubMed: 10840725]

de-Leon et al.


$waterm

ark-text$w

atermark-text

$waterm

ark-text

Revilla-i-Domingo R, Minokawa T, Davidson EH. R11: a cis-regulatory node of the sea urchinembryo gene network that controls early expression of SpDelta in micromeres. Dev Biol. 2004;274:438–451. [PubMed: 15385170]

Rytkonen KT, Williams TA, Renshaw GM, Primmer CR, Nikinmaa M. Molecular evolution of themetazoan PHD-HIF oxygen-sensing system. Mol Biol Evol. 2011; 28:1913–1926. [PubMed:21228399]

Saudemont A, Haillot E, Mekpoh F, Bessodes N, Quirin M, Lapraz F, Duboc V, Rottinger E, Range R,Oisel A, Besnardeau L, Wincker P, Lepage T. Ancestral regulatory circuits governing ectodermpatterning downstream of Nodal and BMP2/4 revealed by gene regulatory network analysis in anechinoderm. PLoS Genet. 2011; 6 e1001259.

Smith J, Davidson EH. Gene regulatory network subcircuit controlling a dynamic spatial pattern ofsignaling in the sea urchin embryo. Proc Natl Acad Sci U S A. 2008; 105:20089–20094. [PubMed:19104065]

Smith J, Davidson EH. Regulative recovery in the sea urchin embryo and the stabilizing role of fail-safe gene network wiring. Proc Natl Acad Sci U S A. 2009

Su YH, Li E, Geiss GK, Longabaugh WJ, Kramer A, Davidson EH. A perturbation model of the generegulatory network for oral and aboral ectoderm specification in the sea urchin embryo. Dev Biol.2009; 329:410–421. [PubMed: 19268450]

Wenger RH, Stiehl DP, Camenisch G. Integration of oxygen signaling at the consensus HRE. SciSTKE. 2005; 2005:re12. [PubMed: 16234508]

Yaguchi S, Yaguchi J, Angerer RC, Angerer LM, Burke RD. TGFbeta signaling positions the ciliaryband and patterns neurons in the sea urchin embryo. Dev Biol. 2010; 347:71–81. [PubMed:20709054]

de-Leon et al.


$waterm

ark-text$w

atermark-text

$waterm

ark-text

Highlights

• Aboral ectoderm specification in the sea urchin embryo is controlled by multiplemechanisms.

• A redox-sensitive transcription factor initiates the aboral ectoderm regulatorystate.

• Two different BMP ligands provide inputs essential for continuing activity ofaboral ectoderm regulatory genes.

• The aboral ectoderm regulatory state is locked down by a positive feedbackwiring.

de-Leon et al.


$waterm

ark-text$w

atermark-text

$waterm

ark-text

Fig 1.Spatio-temporal patterning of the sea urchin aboral ectoderm. A. Diagram of sea urchinembryogenesis marking the different embryonic territories and ectodermal sub-domains.Left – lateral view, right – aboral view. B. Temporal expression profiles of the aboralectoderm regulatory genes (based on (Materna et al., 2010)). C. Spatial expression patternsof aboral ectoderm transcription factors at different time points at early development. Darkblue marks high expression level in this region, light blue marks low expression in thisregion, and gray marks regions where the gene is not detectable in WMISH (based on (Chenet al., 2011)).

de-Leon et al.


$waterm

ark-text$w

atermark-text

$waterm

ark-text

Fig 2.Expression profiles of the ligands BMP2/4 and BMP5/8 and the effect of their perturbationon aboral ectoderm gene expression. A. BMP2/4 and BMP5/8 temporal profiles (based on(Materna et al., 2010)). B. BMP5/8 Spatial expression (WMISH), lv- lateral view, vv-vegetal view. C. The effect of BMP2/4 and BMP5/8 MO on the spatial expression patternsof tbx2/3, irxa and dlx at 24h. BMP2/4 MO strongly reduces the expression of tbx2/3, irxaand dlx, (panels 2, 6, 10), but the expression of irxa at the ectoderm–endoderm borderremains (panel 2). BMP5/8 MO has weaker effect on the genes expression (panels 3,7, 11).Co-injection of BMP2/4 and BMP5/8 MO almost eliminates the expression of tbx2/3 and

de-Leon et al.


$waterm

ark-text$w

atermark-text

$waterm

ark-text

dlx, (panels 4, 8, 12) while still not affecting the vegetal ectoderm expression of irxa (panel4). All embryos are shown in lateral view where the aboral side is to the left.

de-Leon et al.


$waterm

ark-text$w

atermark-text

$waterm

ark-text

Fig. 3.HIF1α expression profiles and the effect of HIF1α perturbation on the spatial expression oftbx2/3 and irxa. A. HIF1α temporal expression measured by QPCR. B. HIF1α spatialexpression at different time points (WMISH). All embryos are shown in lateral view wherethe aboral side is to the left. C. Effect of HIF1α MO on the spatial expression pattern of irxaand tbx2/3 at19h and 24h. The expression of irxa is somewhat reduced at these times whilethe expression of tbx2/3 is strongly reduced at 19hpf and less so at 24hpf. All embryos areshown in lateral view where the aboral side is to the left.

de-Leon et al.


$waterm

ark-text$w

atermark-text

$waterm

ark-text

Fig 4.Summary of the quantitative perturbation analysis conducted. Each perturbation analysismatrix represents the interactions between regulatory genes at a given time point. Thechange in the output-gene expression after injection of a MO targeting the input gene isdenoted “+” if the average ratio between the expression level in the input gene MO and inrandom MO is decreased below 0.4 (that is, average of [output gene level at input MO/output gene level at random MO]<0.4, Figs. S3–S5). In that case, the input gene is anactivator of the output gene. A – 16hpf, B – 19hpf, C – 24hpf.

de-Leon et al.


$waterm

ark-text$w

atermark-text

$waterm

ark-text

Fig 5.Kinetics of perturbation analysis and mathematical modeling. A–F, Average expressionlevels of the aboral ectoderm transcription factors at control (random) MO, Tbx2/3 MO andBMP2/4+BMP5/8 MO at 16hpf, 19hpf and 24hpf. Expression levels were measured byQPCR. Here we included only batches where both Tbx2/3 MO and BMP2/4+BMP5/8 MOwere injected, 16h n=3 (tbx2/3, dlx and, msx) and n=4 (irxa, hox7 and hmx), 19h n=4, 24hn=3. Error bars show standard error. G–I, Mathematical model of the effect of BMP andTbx2/3 perturbation on a downstream gene (Sup. Information 1, model based on (Bolouriand Davidson, 2003; Ben-Tabou de-Leon and Davidson, 2009; Ben-Tabou de-Leon, 2010)).G. Simulated protein concentration per cell of the inputs, Tbx2/3 and pSMAD1/5/8. H.Simulated mRNA level per cell for a downstream gene, TF, that is indirectly activated bythe BMP pathway through the activation of the transcription factor Tbx2/3. TF mRNA levelin the intact circuit is plotted in blue, at BMP MO is plotted in green and at Tbx2/3 MO isplotted in red. I. Simulated mRNA level for cell for a downstream gene, TF that is directlyactivated by the BMP pathway in a feedforward structure with Tbx2/3. Similar color code asin H.

de-Leon et al.


$waterm

ark-text$w

atermark-text

$waterm

ark-text

Fig 6.Cis-regulatory analysis of tbx2/3 and Dlx. A. Sptbx2/3:GFP recombinant BAC map showingtbx2/3 exons (blue boxes), regulatory regions (orange boxes) and a map of tbx2/3 regulatoryregion U indicating the location of functional HIF1α and SMAD1/5/8 binding sites. B, C,Embryos injected with (B) SpTbx2/3 GFP recombinant BAC or (C) SpTbx2/3 U:B:P:GFPreporter construct show correct expression at the ectoderm. D. Spdlx:GFP recombinant BACmap showing dlx exons (blue boxes), regulatory regions (orange boxes) and a map of dlxregulatory regions U and B indicating the location of functional HIF1α, Tbx2/3 andSMAD1/5/8 binding sites. E, F, Embryos injected with (E) SpDlx GFP recombinant BAC or(F) SpDlx U:B:P:GFP reporter construct show correct expression at the ectoderm. G. Theeffect of the mutations of SMAD sites and BMP2/4 MO on tbx2/3 U:B:P:GFP construct attwo time points. H. The effect of HIF1α site mutation and HIF1α MO on tbx2/3 U:B:P:GFP

de-Leon et al.


$waterm

ark-text$w

atermark-text

$waterm

ark-text

construct at two time points. I. The effect of the mutations of SMAD sites andBMP2/4+BMP5/8 MO on dlx U:B:P:GFP construct at two time points. J. The effect ofHIF1α and Tbx2/3 site mutation and HIF1α MO and Tbx2/3 MO on Dlx U:B:P:GFPconstruct at two time points. Error bars show standard error, p values calculated by onetailed z-test.

de-Leon et al.


$waterm

ark-text$w

atermark-text

$waterm

ark-text

Fig 7.GRN diagrams at 16h, 19h and 24h showing the dynamic control of the aboral ectodermspatio-temporal patterning. Thick lines and blue diamonds mark the regulatory connectionsthat were verified by cis-regulatory analysis to be direct.

de-Leon et al.


$waterm

ark-text$w

atermark-text

$waterm

ark-text

nih public access yi-hsien su kuan-ting lin enhu li , and...

Documents