Q1

123456789

101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566

Developmental Biology ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Contents lists available at ScienceDirect

Developmental Biology

0012-16http://d

Abbrethe kidnChIP, ch

n CorrE-m

Pleasplaye

journal homepage: www.elsevier.com/locate/developmentalbiology

A core transcriptional network composed of Pax2/8, Gata3 and Lim1regulates key players of pro/mesonephros morphogenesis

Sami Kamel Boualia a, Yaned Gaitan a, Mathieu Tremblay a, Richa Sharma a, Julie Cardin b,Artur Kania b, Maxime Bouchard a,n

a Goodman Cancer Centre and Department of Biochemistry, McGill University, 1160 Pine Ave. W., Montreal, Quebec, Canada H3A 1A3b Institut de Recherches Cliniques de Montréal, Montréal, Québec, Canada H2W 1R7

a r t i c l e i n f o

Article history:Received 23 January 2013Received in revised form27 July 2013Accepted 30 July 2013

Keywords:Kidney developmentTranscriptionPax2Gata3Lim1CAKUT

06/$ - see front matter & 2013 Published by Ex.doi.org/10.1016/j.ydbio.2013.07.028

viations: CND, common nephric duct; CAKUey and urinary tract; GDNF, glial cell line-deromatin immunoprecipitationesponding author. Fax: +1 514 398 6769.ail address: maxime.bouchard@mcgill.ca (M. B

e cite this article as: Kamel Boualia,rs of pro/mesonephros morphogene

a b s t r a c t

Translating the developmental program encoded in the genome into cellular and morphogeneticfunctions requires the deployment of elaborate gene regulatory networks (GRNs). GRNs are especiallycrucial at the onset of organ development where a few regulatory signals establish the differentprograms required for tissue organization. In the renal system primordium (the pro/mesonephros),important regulators have been identified but their hierarchical and regulatory organization is stillelusive. Here, we have performed a detailed analysis of the GRN underlying mouse pro/mesonephrosdevelopment. We find that a core regulatory subcircuit composed of Pax2/8, Gata3 and Lim1 turns on adeeper layer of transcriptional regulators while activating effector genes responsible for cell signalingand tissue organization. Among the genes directly affected by the core components are the keydevelopmental molecules Nephronectin (Npnt) and Plac8. Hence, the pro/mesonephros GRN linkstogether several essential genes regulating tissue morphogenesis. This renal GRN sheds new light onthe disease group Congenital Anomalies of the Kidney and Urinary Tract (CAKUT) in that gene mutationsare expected to generate different phenotypic outcomes as a consequence of regulatory networkdeficiencies rather than threshold effects from single genes.

& 2013 Published by Elsevier Inc.

676869707172737475767778798081828384858687

Introduction

Throughout development, the progressive diversification anddifferentiation of cells requires a constant reorganization of theirunderlying gene regulatory state. This progression is under thecontrol of gene regulatory networks (GRNs) comprised of transcrip-tion factors controlling regulatory and effector proteins, ultimatelyresulting in the activation of the terminal cellular differentiationprogram (Davidson and Erwin, 2006; Davidson, 2010). As such,GRNs effectively read and execute the developmental programencoded in the genome and convert it into concrete molecularmachinery regulating cell and tissue organization.

GRNs have been extensively studied in bacteria (Alon, 2007),Drosophila (Levine and Davidson, 2005) and sea urchin (Davidson,2009), which are readily amenable to elaborate regulatory networkstudies. They have also been characterized in vertebrates, notablyin heart field formation (Cripps and Olson, 2002), in mesodermspecification (Koide et al., 2005) and in haematopoietic stem cell

888990919293

lsevier Inc.

T, congenital anomalies ofrived neurotrophic factor;

ouchard).

S., et al., A core transcriptiosis. Dev. Biol. (2013), http:/

development (Pimanda and Gottgens, 2010). The comparativeanalysis of the GRNs characterized in these systems revealed animportant level of conservation in the topology of the subcircuits(network motifs) as well as a hierarchical organization of thesesubcircuits (Alon, 2007; Davidson, 2010). The gradual deploymentof the developmental program is indeed performed by sequentialactivation of GRNs subcircuits, each responsible for the execution ofa precise regulatory task. Subcircuits can be categorized as regula-tory and effectors to reflect the role they play in the hierarchy ofGRN deployment (Davidson, 2010; Davidson and Erwin, 2006).Among the regulatory subcircuits, kernels are critical evolutionaryconserved units found at the onset organ or structure formation.The inactivation of any of their components will typically result inmajor deficiencies in the structure they regulate (Davidson andErwin, 2006). Other regulatory units such as “plug-ins” (e.g. signal-ing pathways) and “on–off switches” linking subcircuits also per-form an important role as flexible modular units. On the other hand,effector subcircuits perform cell biology functions such as cell shapechanges or proliferation, while differentiation gene batteries com-prise the terminal differentiation molecules necessary for the cell toperform its ultimate role in the organism (Davidson, 2010; Davidsonand Erwin, 2006).

In the urogenital system, some crucial developmental regulatorshave been identified but a clear understanding of the GRNs underlying

9495

nal network composed of Pax2/8, Gata3 and Lim1 regulates key/dx.doi.org/10.1016/j.ydbio.2013.07.028i

123456789

101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566

676869707172737475767778798081828384858687888990919293949596979899

S. Kamel Boualia et al. / Developmental Biology ∎ (∎∎∎∎) ∎∎∎–∎∎∎2

the development of this organ system is still lacking. Urogenitalsystem development is initiated with the formation of the nephricduct by mesenchymal–epithelial transition and tubulogenesis ofintermediate mesoderm progenitor cells (Bouchard, 2004; Dressler,2006). Nephric lineage induction is followed by the elongation of thenephric duct that induces adjacent mesonephric tubules to form thepro/mesonephros, and ultimately connects to the cloaca (bladder-urethra primordium) in the caudal part of the embryo trunk. Together,the transcription factors Pax2 and Pax8 were identified as necessaryand sufficient for the specification of the nephric lineage andsubsequent pro/mesonephros morphogenesis (Bouchard et al., 2002).In addition to Pax2/8, the transcription factors Gata3 and Lim1 (Lhx1)are also crucial for pro/mesonephros development, as Gata3 regulatesproliferation, elongation and guidance of the nephric duct (Chia et al.,2011; Grote et al., 2006), while Lim1 is necessary for nephric ductelongation and integrity (Kobayashi et al., 2005; Pedersen et al., 2005;Shawlot and Behringer, 1995). Following nephric duct elongation, themetanephric (or adult) kidney is induced as a diverticulum of thenephric duct at the level of the hindlimbs. This crucial event isregulated by the secreted molecule Gdnf in the mesenchyme, whichsignals to the Ret receptor expressed in the nephric duct (Dressler,2006; Moore et al., 1996; Pichel et al., 1996; Sanchez et al., 1996;Schuchardt et al., 1994; Schuchardt et al., 1996).

In the past, important transcriptional and signaling regulatorsof nephric duct morphogenesis have been identified but relativelylittle is known about their regulatory organization and transcrip-tional output. However, evidence suggests that these regulatorsmay be part of a deep regulatory network conducting the differentaspects of nephric duct morphogenesis. Here, we adopt a com-prehensive approach to characterize the pro/mesonephros regu-latory network. We find that the pro/mesonephros GRN iscomposed of a kernel sub-circuit that links to several key effectorsof nephric duct morphogenesis.

100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132

Materials and methods

Mice

Pax2, Pax8, Gata3, Lim1, Emx2 and Pax2BACGFP mice were bredon a C3H/HeJ background for at least 6 generations. Genotyping ofthese mice has been described previously (Bouchard et al., 2000,2002, 2005; Grote et al., 2006; Hoyt et al., 1997; Miyamoto et al.,1997; Shawlot and Behringer, 1995; Yoshida et al., 1997). All in vivostudies were approved by the Animal Care Committee of McGillUniversity and strictly follow the guidelines from the CanadianCouncil on Animal Care.

Microarray experiments

Microarray experiments for Pax2-regulated genes was reportedpreviously (Grote et al., 2006) using either Pax2BACGFP;Pax2� /� orPax2+/+ cells from pro/mesonephros at E9.0–9.5 (13–20S). Briefly,embryo trunks were dissected and trypsinized. GFP+ cells (repre-senting the whole pro-mesonephros epithelium) were sorted fromsurrounding GFP- trunk cells by flow cytometry. The mRNAcontent of the cells was amplified by two rounds of linear RNAamplification using T7-RNA polymerase recognition sequenceinserted during reverse transcription. Microarray analysis wasdone on cDNA microarray slides containing 26,000 expressedsequence tag (EST) clones from the BMAP and NIA libraries.Real-time quantitative PCR validation was done on sorted Pax2-BACGFP;Pax2� /� or Pax2+/+ cells at 18–21 somite stage, usingGreen-to-go (Biobasic) on Realplex2 PCR machine (Eppendorf).The analysis was done in triplicates (technical) with 3 biologicalsamples of each genotype. The mouse beta-2 microglobulin (B2M)

Please cite this article as: Kamel Boualia, S., et al., A core transcriptioplayers of pro/mesonephros morphogenesis. Dev. Biol. (2013), http:/

gene was used as internal control. The primers used are listed inTable S3.

Plasmids

Evi1, Npnt, Wfdc2, Id4, Plac8 and Lim1 probes were generated byRT-PCR using primers described in Table S3 and cloned in thepGEM-T-easy or pLXSN vector. Murine Pax2 and Gata3 cDNA weresubcloned in pMSCV-HA3-IRES-GFP vector (kindly provided by Dr.Jerry Pelletier, McGill University). Murine Lim1 was subcloned inpLXSN-Flag vector.

Luciferase reporter assay

Enhancer sequences were cloned in pGL3-PolA Luciferasereporter plasmid. Site directed mutagenesis was done by PCR togenerate mutated derivatives. Transcription factor expressingconstructs (either Pax2, Gata3 or Lim1) and the Luciferase reporterconstructs were cotransfected in HEK293T cells with Lipofecta-mine 2000 (Invitrogen) according to the manufacturer's instruc-tions. Transient transactivation assays were performed on aLUMIstar Omega (BMG) using Dual-Glo Luciferase Assay System(Promega). In all samples the amount of total DNA was keptconstant with pGem4 and pCMV-Renilla vector was included fortransfection efficiency.

In situ hybridization

Embryos were dissected at E9.5 and fixed overnight in 4%paraformaldehyde at 4 1C. In situ hybridization was performed asdescribed previously (Henrique et al., 1995) on cryosections fromE9.5 embryo with digoxigenin-dUTP RNA probes against Lim1(Bouchard et al., 2002), Gata3 (Grote et al., 2006), Emx2 (Yoshidaet al., 1997), Pax2 (Bouchard et al., 2002), Pax8 (Bouchard et al.,2002), Pcdh19 (Gaitan and Bouchard, 2006), Evi1, Npnt, Wfdc2, Id4and Plac8.

Cell culture and infection

Murine inner medullary collecting duct cells (mIMCD3) (kindlyprovided by Dr. Paul Goodyer, McGill University) were cultured in10% fetal bovine serum DMEM/F12 media (Wisent). Virus produc-tion was performed in HEK293T cells by cotransfection of pMSCV-Pax2-HA3-IRES-GFP or pMSCV-Gata3-HA3-IRES-GFP, pVPack-GPand pVPack-VSV-G vectors (Stratagene). mIMCD3 cells wereinfected for 48 h and sorted for GFP expression using a FACSAria(BD-Bioscience). mIMCD3-Gata3-HA cells were also stably trans-fected with Lim1-Flag plasmid.

Bioinformatic analysis

Conserved genomic regions between mouse and human wereidentified by alignment of sequences of 50 kb upstream of thecoding regions (first ATG) using nBLAST (NCBI). Putative transcrip-tion factor binding sites were identified by probing conservedsequences for the presence of Pax2/5/8, Gata3 and Lim1 consensusDNA binding sequences using MacVector 8.0. The region coveredwas increased in the absence of putative binding sites. Theconsensus binding site sequence used are the following: Gata3,(A/T)GATA(A/G) (Orkin, 1992), Pax2/8, (A/G)N(A/C/T)CANT(C/G)A(A/T)GCGT(A/G)(A/T)(A/C) allowing three mismatches (Boualiaet al., 2011) and Lim1, the core homeodomain consensus bindingsequence ATTA (Mochizuki et al., 2000).

nal network composed of Pax2/8, Gata3 and Lim1 regulates key/dx.doi.org/10.1016/j.ydbio.2013.07.028i

123456789

101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566

6768697071727374757677787980818283848586878889909192939495969798

S. Kamel Boualia et al. / Developmental Biology ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 3

Chromatin immunoprecipitation

Murine inner medullary collecting duct cells (mIMCD3) stablyexpressing Pax2-HA, Gata3-HA and Lim1-Flag were cross-linkedwith 1% (w/v) formaldehyde for 10 min. The cells were collectedand sonicated to achieve DNA shearing to an average of 200 bp.The chromatin was pre-cleared with protein G-Agarose beads(Roche, cat.11243233001) and precipitated overnight with ananti-HA (Abcam), an anti-Pax2 antibody (Covance PRB-276P,LN#143834801), an anti-Flag antibody (Sigma) or beads alone asa negative control. The antibody was retrieved with proteinG-Agarose beads for 2 h. The HA immunoprectitations (Ab orbeads only) were washed 8�10 min with 1 ml of Buffer I (lowsalt: 0.1% SDS, 1% Triton-X, 20 mM Tris–HCl (pH8), 150 mM Nacl),8�10 min with 1 ml of Buffer II (high salt: 1% Triton-X, 20 mMTris–HCl (pH8), 2 mM EDTA, 500 mM NaCl), 1� with 1 ml ofBuffer III (250 mM LiCl, 1% Igepal, 1 mM EDTA, 1% Na-Desoxycho-late, 10 mM Tris–HCl (pH8)), 1� with 1 ml of Buffer IV (1� Tris-EDTA), and de-crosslinked at 65 1C overnight in 150 ml of Buffer V(1% SDS, 0.1 M NaHCO3). Pax2 immunoprecpitations (Ab and beadsonly) were washed 6� with Buffer I, 6� with Buffer II, 1� withBuffer III, 1� with Buffer IV and de-crosslinked overnight in BufferV. The Flag immunoprectitations (Ab and beads only) were washedwith 7� with Buffer I, 8� with Buffer II, 1� with Buffer III, 1�with Buffer IV, and de-crosslinked overnight in Buffer V. Thesamples were then treated with proteinase K (0.2 mg/ml) for 1 hat 55 1C. Chromatin was isolated using the QIAquick PCR purifica-tion kit (Qiagen cat.28106). Quantitative PCR was performed on aminimum of 3 independent ChIP samples (biological replicates)and compared to beads only experimental control. Each sampleand control was analyzed in technical triplicates. Error bars referto biological triplicate data derived from the mean of all 3 technical

Fig. 1. Genes differentially regulated by Pax2 and expressed in pro/mesonephros tissue. (as being differentially regulated by Pax2 at 13, 16 and 20 somites, upregulated between 1Pax2BACGFP mice. Ratios are expressed as control/mutant (Pax2� /�), late/early or GFPmicroarray results for the genes identified in (A), using quantitative RT-PCR analysis ontype or mutant for Pax2.

Please cite this article as: Kamel Boualia, S., et al., A core transcriptioplayers of pro/mesonephros morphogenesis. Dev. Biol. (2013), http:/

triplicates. The qPCR results were accepted between 495%o105% efficiency (E¼1–10(�1/slope)). All results were normalizedto an unrelated control region near the FoxO6 gene on chromo-some 1. The final enrichment was calculated as the ratio ofnormalized values for antibody over beads only. Primers weredesigned around conserved DNA-binding consensus sites and arelisted in Table S2.

Results

Pax2 transcriptional profile identifies molecular links to knownkidney developmental regulators

We have previously reported that the transcription factors Pax2and Pax8 are necessary and sufficient for renal lineage specifica-tion in the embryo (Bouchard et al., 2002). To better understandthe transcriptional program regulated by Pax2, we reanalyzed theexpression profile of Pax2-deficient pro/mesonephric cells (Groteet al., 2006). In these experiments, pro/mesonephric cells wereisolated using a BAC transgene expressing GFP under Pax2 control(Pax2BACGFP) and samples were compared (1) between wild typeand Pax2 mutant at 13, 16 and 20 somite stages, (2) between GFP-positive pro/mesonephric cells and GFP-negative trunk cells and(3) between 13 vs. 16 and 13 vs. 20 somite stage pro/mesonephriccells. This analysis placed the emphasis on genes differentiallyregulated by Pax2 at the 20 somite stage to reduce experimentalstringency and identify additional differentially expressed genes(Figs. 1 and S1). This analysis identified a large proportion of genesinvolved in transcription and signaling (Fig. S1; Table S1), support-ing a role for Pax2 in the early deployment of the regulatoryprogram of pro/mesonephros morphogenesis.

99100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132

A) Pax2-dependent genes were identified by cDNA microarray analysis and selected3 and 16 or 20 somites or specifically expressed in pro/mesonephros GFP+ tissue of+/GFP‐. The embryonic stages are defined as somite number(s). (B) Validation ofsorted GFP+ renal cells from 18 to 21 somite stage Pax2BACGFP embryos either wild

nal network composed of Pax2/8, Gata3 and Lim1 regulates key/dx.doi.org/10.1016/j.ydbio.2013.07.028i

123456789

101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566

676869707172737475767778798081828384858687888990

S. Kamel Boualia et al. / Developmental Biology ∎ (∎∎∎∎) ∎∎∎–∎∎∎4

A selection for genes positively regulated by Pax2 revealed anumber of transcription factor genes previously reported to play arole in pro/mesonephros morphogenesis. As these genes representnodes in the pro/mesonephros developmental program, wedecided to focus on them for the purpose of primary networkcharacterization. Among the transcription factors requiring Pax2function were Gata3 (as previously reported (Grote et al., 2006))but also Lim1 (Lhx1), Evi1, Id4 (Idb4) and Plac8 (Fig. 1). The latterwas recently identified as a Pax2-dependent transcriptional regulatorrequired for renal development in zebrafish (Bedell et al., 2012).In addition to transcription factors, we examined Pax2-dependentgenes for candidate effectors of pro/mesonephros morphogenesis.This analysis identified three genes previously associated with nephricduct morphogenesis among the best candidates downstream of Pax2,namely Nephronectin (Npnt), Pcdh19 and Wfdc2 (Fig. 1). Npnt isspecifically expressed in the nephric duct and acts as a ligand of themesenchymal alpha8/beta1 integrin complex (Brandenberger et al.,2001). Importantly, gene inactivation of either Nephronectin or alpha8integrin gene (Itga8) results in metanephric kidney agenesis (Lintonet al., 2007). Pcdh19 and Wfdc2 are potential nephric duct regulatorsbased on their strong expression previously reported in this tissue(Gaitan and Bouchard, 2006; Hellstrom et al., 2003). Together, theseresults identify important regulators of pro/mesonephros develop-ment requiring Pax2 function for normal expression.

919293949596979899

100101102

Hierarchical relationship between Pax2/8, Gata3 and Lim1

As the transcriptional profiling of Pax2-deficient pro/mesonephrosidentified the key developmental regulators Gata3 and Lim1, wesought out to clarify the genetic relationship between these factorsin the developing pro/mesonephros. We first assessed the expressionof each transcript in wild type embryos by in situ hybridization. Aspreviously reported, the expression of all three transcription factorswas robust and specific to the pro/mesonephric epithelium (Fig. 2Aand C) (Bouchard et al., 2002; Grote et al., 2006; Kobayashi et al., 2005;

Fig. 2. Genetic interaction between kidney development regulators. In situ hybridizationembryo sections of the indicated genotypes. The pro/mesonephros expression of both Gother for sustained expression. Pax2 and Pax8 are unaffected by the loss of Gata3 or Lim

Please cite this article as: Kamel Boualia, S., et al., A core transcriptioplayers of pro/mesonephros morphogenesis. Dev. Biol. (2013), http:/

Miyamoto et al., 1997; Pedersen et al., 2005). Due to functionalredundancy between Pax2 and Pax8 in renal lineage specification(Bouchard et al., 2002), we opted to study the role of Pax genes byremoving an additional allele of Pax8 on a Pax2 null background,effectively lowering Pax levels further, while still ensuring the forma-tion and morphogenesis of the pro/mesonephros. The analysis ofPax2� /�;Pax8+/� embryos confirmed the genetic dependency ofGata3 on Pax gene function and further revealed a complete loss ofLim1 in Pax-deficient embryos (Fig. 2F). Similarly, upon removal ofGata3, the expression of Lim1 was strongly reduced, while Pax2expression remained unchanged (Fig. 2H and I), confirming thatPax2 acts genetically upstream of these genes. We next analyzed theexpression of Pax2 and Gata3 transcription factors in Lim1-deficientembryos. Interestingly, Lim1 was found to be required for Gata3, butnot Pax2 expression, (Fig. 2K and L), placing it at a position equivalentto Gata3 in the transcriptional hierarchy. To validate the presence ofpro/mesonephric duct tissue in the region being investigated, adjacentsections to all in situ hybridization results described here stainedpositive for Pax8 expression (Fig. 2D,G,J and M, data not shown). Thefact that Pax8 was never found affected in these mutant backgroundsalso confirms Pax8 next to Pax2 at the onset of the transcriptionalcascade of pro/mesonephros development.

Together, these results identify a transcriptional cascadewhereby Pax2/8 acts upstream of both Gata3 and Lim1, which inturn require each other's function. From these data, we can definethe Pax2/8-Gata3-Lim1 group as the core or “kernel” of the renalprimordium gene regulatory network.

The renal primordium kernel regulates a deeper GRN

To study the depth of the renal primordium GRN, we nextassessed the regulation of additional transcription factors withreported roles in renal development. For this, we focused ontranscriptional regulators identified among Pax2-regulated genes;Evi1, whose mutant showmesonephric hypoplasia (Hoyt et al., 1997),

103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132

for Pax2 (A,H and K), Gata3 (B,E and L), Lim1 (C,F and I), Pax8 (D,G,J and M) on E9.5ata3 and Lim1 requires Pax2/8 gene function. In turn, Gata3 and Lim1 require each1.

nal network composed of Pax2/8, Gata3 and Lim1 regulates key/dx.doi.org/10.1016/j.ydbio.2013.07.028i

123456789

101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566

676869707172737475767778798081828384858687888990919293949596979899

100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132

Fig. 3. Altered expression pattern of transcriptional regulators of renal development. In situ hybridization for Evi1 (A,E,I and M), Id4 (B,F,J and N), Plac8 (C,G,K and O) andEmx2 (D,H,L and P) on E9.5 embryo sections of the indicated genotypes. Evi1 expression requires Lim1 gene function, while Pax2/8 plays a complementary role in itsregulation. Conversely, Id4 and Plac8 require Pax2/8 and Gata3, whereas Lim1 plays a secondary role. Emx2 expression is lost in Lim1 mutant embryos, reduced in Pax2/8deficient embryos and unaffected by Gata3 loss.

S. Kamel Boualia et al. / Developmental Biology ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 5

the new renal developmental regulator Plac8 (Bedell et al., 2012) andId4, previously shown to be expressed in Xenopus pronephricdevelopment (Liu and Harland, 2003). To expand this analysis, weincluded the transcription factor Emx2, another important regulatorof kidney development that was recently shown to cooperate withPax2 in pro/mesonephros morphogenesis (Boualia et al., 2011). In situhybridization analysis of Evi1 expression in wild type embryosconfirmed a specific localization of this gene in pro/mesonephrictissue (Fig. 3A). In contrast, the nephric duct expression of Evi1 wasseverely compromised in Pax2� /�;Pax8+/� and completely abro-gated in Lim1� /� embryos (Fig. 3E and M) but not significantlyaffected by Gata3 deficiency (Fig. 3I). In Pax2�/�;Pax8+/� embryos,the rostral-most region maintained some Evi1 expression (notshown), likely due to higher levels of Pax8 expression in this region.The wild type expression of Id4 and Plac8 was strong, and specific tothe nephric duct (Fig. 3B and C) but greatly reduced in Lim1� /�

embryos (Fig. 3N and O) and completely abrogated in the renalprimordium of Pax2�/� Pax8+/� and Gata3� /� embryos (Fig. 3F,G,Jand K). The analysis of Emx2 expression levels confirmed its down-regulation in Pax2/8-deficient embryos (Fig. 3H). Strikingly, Emx2expression was completely lost in Lim1-deficient nephric ducts(Fig. 3P) but unaffected by Gata3-deficiency. Of interest, the expres-sion of Pax2, Gata3 and Lim1 was found unaffected in Emx2-deficientembryos at E9.5, confirming the downstream position of thistranscription factor in the renal GRN (Fig. S2).

Together, these results point to a relatively deep structure ofthe renal primordium GRN in which the kernel regulates at leastone additional layer of transcriptional regulators of renal primor-dium development.

Effector gene expression is regulated at different levels of the renalprimordium GRN

Having established the genetic hierarchy within the transcrip-tional regulators, we wanted to further characterize their output

Please cite this article as: Kamel Boualia, S., et al., A core transcriptioplayers of pro/mesonephros morphogenesis. Dev. Biol. (2013), http:/

on potential regulators of the cellular and tissue events necessaryfor proper pro/mesonephric development. We focused on Npnt,Pcdh19, and Wfdc2 three downstream cellular effectors previouslylinked to kidney development that were identified differentiallyregulated in Pax2 transcriptional profiling analysis.

We first confirmed the mesonephric expression of all threeeffector genes, which was found specific to the pro/mesonephros(Fig. 4A and C) (Gaitan and Bouchard, 2006). Strikingly, Wfdc2expression was completely lost in Pax2�/� Pax8+/� embryos(Fig. 4E), while Npnt expression levels were strongly reduced(Fig. 4F), identifying Pax gene requirement for full activationand/or maintenance of these effectors genes. In contrast, Pcdh19was not significantly affected in Pax2/8-deficient embryos(Fig. 4D). In Gata3-/- embryos, Npnt and Pcdh19 levels of expressionwere only slightly downregulated and Wfdc2 levels stayed unaf-fected (Fig. 4G–I). As for Lim1�/� embryos, the expression ofPcdh19 and Npnt transcript expression was completely lost (Fig. 4Jand L), while Wfdc2 transcription was still detected (Fig. 4K),arguing for a critical role of Lim1 for proper Npnt and Pcdh19expression in the pro/mesonephros.

Together, these results point to a genetic regulation of specificeffector genes by single core regulators, as well as instances ofcoregulation by a combination of core transcriptional regulators.Noteworthy, the regulation of specific effector genes occurs atdifferent levels of the GRN kernel.

Pax2 directly binds the Lim1 and Plac8 regulatory regions

To assess whether the genetic regulatory interactions observedabove are direct or not, we next used a Chromatin Immunopreci-pitation (ChIP) approach. The number of cells of the renalprimordium being limiting, we turned to the mouse inner collect-ing duct cell line (mIMCD3). As it is derived from the collectingduct, this cell line is among the closest possible to nephric ductcellular fate. Accordingly, it was shown to express endogenous

nal network composed of Pax2/8, Gata3 and Lim1 regulates key/dx.doi.org/10.1016/j.ydbio.2013.07.028i

123456789

101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566

676869707172737475767778798081828384858687888990919293949596979899

100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132

Fig. 4. Altered expression pattern of cellular effectors of kidney development. In situ hybridization for Pcdh19 (A,D,G and J), Wfdc2 (B,E,H and K) and Npnt (C,F,I and L) on E9.5embryo sections of the indicated genotypes. Pcdh19 expression is mostly affected by Lim1 deficiency. Wfdc2 requires Pax2/8 but not Gata3 or Lim1. Npnt requires Lim1function, while Pax2/8 and Gata3 play a complementary role.

S. Kamel Boualia et al. / Developmental Biology ∎ (∎∎∎∎) ∎∎∎–∎∎∎6

Pax2 (Cai et al., 2005). Our analysis confirmed this expression butrevealed very low levels of Gata3 and Lim1 expression (Fig. S3A, C,D and E; mIMCD3-EV).

To identify Pax2, Gata3 and Lim1 target genes in the samesystem, we derived a stable mIMCD3 cell line expressing both HA-tagged Gata3 and Flag-tagged Lim1, as well as a Pax2-HA mIMCD3cell line (Fig. S3; mIMCD3-G3HA-L1Flag; mIMCD3-P2HA). As therewas no established target binding sites of Pax2 in renal primor-dium genes, we first created an artificial positive control bygenerating a stable derivative of mIMCD3 Pax2-HA cells harboringa concatemer of three strong Pax2/5/8 binding sites derived fromthe CD19 locus (Kozmik et al., 1992). This stable cell line generatedvery robust and reproducible enrichment of more than 10 foldcompared to beads only and was used as positive control insubsequent experiments (Fig. S4A). ChIP against the endogenousor HA-Pax2 gave similar results in this cell line.

To establish the direct regulatory role of Pax2, we focused onthe genes Gata3, Lim1, Evi1, Id4, Npnt, Wfdc2 and Plac8. For each ofthese genes, putative binding sites conserved between mouse andhuman were identified in genomic regions located within 50 kb ofthe translation start site. Each of these sites was tested either byendogenous Pax2 or Pax2-HA- mediated ChIP. Of those, we couldidentify a very strong binding of Pax2 to Lim1 and Plac8 regulatoryregions whereas other genes failed to identify direct DNA-bindinginteractions (Fig. 5B D and S4A). The Lim1 locus contained 57conserved regions, four of which contained conserved putativePax2 binding sites (Fig. 5A; P2Lim1-1 to 4). Chromatin immuno-precipitation identified the conserved sites P2Lim1-3 and P2Lim1-4 as robustly bound by Pax2 with an enrichment ratio of 5.5fold and 13.6 fold compared to control, respectively (Fig. 5B),whereas sites P2Lim1-1 and P2Lim1-2 failed to show enrichment.

Please cite this article as: Kamel Boualia, S., et al., A core transcriptioplayers of pro/mesonephros morphogenesis. Dev. Biol. (2013), http:/

To validate further these positive sites we tested the enrichment offlanking regions less than 300 bp away (P2Lim1-3UP (upstream);P2Lim1-3D (downstream); P2Lim1-4UP; P2Lim1-4D) which failedto be enriched by Pax2, re-enforcing the specificity of P2Lim1-3and P2Lim1-4 binding sites (Fig. 5B). The bioinformatics analysis of100 kb upstream and 50 kb downstream of the Plac8 locusidentified 31 conserved regions, two of which contained putativePax2 binding sites (Fig. 5D; P2Plac8-1 and P2Plac8-2). Chromatinimmunoprecipitation identified only P2Plac8-2 as robustly boundby Pax2 with an enrichment ratio of 5.2 fold compared to thecontrol (Fig. 5E). Together, these data point to a direct binding ofPax2 to Lim1 and Plac8 regulatory regions.

To assess whether the regions bound by Pax2 are in an activetranscriptional state we used the H3K4me1 histone mark. Mono-methylation of histone H3 on lysine 4 has indeed been used toidentify active enhancer sites in whole genome analyses (Heintz-man 2009; Visel 2009). As expected, we found that H3K4me1mark was specifically enriched in regions bound by Pax2 (P2Lim1-3, P2Lim1-4, P2Plac8-2) (Fig. 5C and F), demonstrating that thesesites are bona fide enhancers. As expected for this chromatin mark,the modification extended to the immediate neighboring regionsof the bound sites (“UP” and “D” primer pairs) but not to moredistant regions (Fig. 5B and F).

We next determined whether the identified Pax2 binding siteslocated in enhancer sequences were responsive to Pax2. For this, wecloned the P2Lim1-3, P2Lim1-4 and P2Plac8-2 regions in pGL3-PolAluciferase reporter vector and generated derivatives with pointmutations in Pax2 consensus binding sites. Co-transfections of theconstructs with a Pax2-expression vector in HEK293T demonstrateda strong activation by Pax2 that was significantly reduced in Pax2mutant derivatives (Fig. 5G and S8A).

nal network composed of Pax2/8, Gata3 and Lim1 regulates key/dx.doi.org/10.1016/j.ydbio.2013.07.028i

123456789

101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566

676869707172737475767778798081828384858687888990919293949596979899

100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132

Fig. 5. Direct interaction between Pax2 and downstream regulated genes. (A and D) schematic representation of sequence conservation (thick bars) and Pax2 putativebinding sites (asterisks) in the Lim1 (A) and Plac8 (D) loci. (B and E) Chromatin immunoprecipitation against Pax2 in mIMCD3‐Gata3‐HA‐Lim1‐Flag cells. Regions from theLim1 (B) and Plac8 (E) loci containing putative Pax2 binding sites, or either 300 bp upstream (UP) or 300 bp downstream (D) sequences were amplified by real-time PCR. Dataare expressed as fold enrichment relative to beads only and represent the average of experimental triplicates. (C and F) Chromatin immunoprecipitation against H3K4me1marks (active enhancer) in the Lim1 (C) and Plac8 (F) loci. (G) Functional analysis of Pax2 binding sites in the Lim1 and Plac8 enhancer regions. The HEK293T cell line wascotransfected with Pax2 expressing vector and wild type or mutant enhancer constructs. Results are expressed as luciferase activity relative to PolA promoter activity andrepresent the average7SD of triplicate of 2 independent experiments. Luciferase reporter activities were normalized to that of an internal control (pCMV-Renilla).

S. Kamel Boualia et al. / Developmental Biology ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 7

Together, these results demonstrate a direct binding andregulation of Pax2 to active enhancer sites of Lim1 and Plac8.

Gata3 binds the ret and Lim1 regulatory regions

We next turned to the identification of direct Gata3 binding sitesby ChIP. For this, we took advantage of a Gata3 binding sitepreviously associated with Ret gene expression to use as positive

Please cite this article as: Kamel Boualia, S., et al., A core transcriptioplayers of pro/mesonephros morphogenesis. Dev. Biol. (2013), http:/

control (G3Ret-2; Fig. 6A and B) (Grote et al., 2006, 2008). Reanalysisof the Ret locus with the criteria described above identified a total of4 putative Gata3 binding sites located in conserved regions. ChIPanalysis on these sites confirmed the binding of Gata3 to G3Ret-2(2.9 fold enrichment) and further identified a strong binding of Gata3to G3Ret-1 (4.0 fold enrichment) (Fig. 6B). These results validate theChIP approach for Gata3 binding sites and further support the directregulation of Ret by Gata3.

nal network composed of Pax2/8, Gata3 and Lim1 regulates key/dx.doi.org/10.1016/j.ydbio.2013.07.028i

123456789

101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566

676869707172737475767778798081828384858687888990919293949596979899

100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132

Fig. 6. Gata3 binds Ret and Lim1 regulatory regions. (A and D) schematic representation of sequence conservation (thick bars) and Gata3 putative binding sites (asterisks) inthe Ret (A) and Lim1 (D) loci. (B and E) Chromatin immunoprecipitation against Gata3-HA in mIMCD3‐Gata3‐HA‐Lim1‐Flag cells on regions containing or flanking (300 bpUP, D) conserved putative binding sites for Gata3 in the Ret (B) and Lim1 (E) loci. (C and F) Chromatin immunoprecipitation against H3K4me1 enhancer marks in the Ret(C) and Lim1 (F) loci. (G) Luciferase reporter analysis of Gata3 binding sites in the Ret and Lim1 enhancer regions by cotransfection with Gata3 expressing vector as in Fig. 5G.

S. Kamel Boualia et al. / Developmental Biology ∎ (∎∎∎∎) ∎∎∎–∎∎∎8

We then performed ChIP on conserved Gata3 binding sites ongenes differentially regulated by Pax2. This analysis identified adirect binding of Gata3 to Lim1. Bioinformatic analysis of the Lim1locus identified 4 conserved Gata3 consensus binding sites(Fig. 6D). Of those, one was robustly enriched to 5.0 fold whencompared to control conditions (Fig. 6E). This conserved site lies ina long 502 bp highly conserved stretch at 7.6 kb upstream of Lim1translation start site. The analysis of putative Gata3 binding sitesfor H3K4Me1 methylation mark confirmed the active state ofthe G3-Ret1, G3Ret2 and G3Lim1-1 enhancer regions (Fig. 6C,F).In addition, the functional analysis of these sites in HEK293T cells

Please cite this article as: Kamel Boualia, S., et al., A core transcriptioplayers of pro/mesonephros morphogenesis. Dev. Biol. (2013), http:/

demonstrated a direct role for Gata3 in the regulation of thesebinding sites (Fig. 6G and S8B). Together, these results identifyLim1 and Ret as direct regulatory targets of Gata3.

Npnt Is a direct regulatory target of Lim1

The genetic regulation experiments determined that Lim1 wasrequired for the expression of Gata3, Id4, Evi1, Npnt and Pcdh19.Detailed ChIP analysis of the conserved genomic regions of thesegenes positively identified a direct regulation of Npnt by Lim1.The bioinformatic analysis of 50 kb upstream of the Npnt locus

nal network composed of Pax2/8, Gata3 and Lim1 regulates key/dx.doi.org/10.1016/j.ydbio.2013.07.028i

123456789

101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566

676869707172737475767778798081828384858687888990919293949596979899

100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132

Fig. 7. Lim1 is a direct regulator of Npnt expression. (A) Schematic representationof sequence conservation (thick bars) and Lim1 putative binding sites (asterisks) inthe Npnt locus. (B) Chromatin immunoprecipitation against Lim1-Flag in mIMCD3‐Gata3‐HA‐Lim1‐Flag cells on regions containing or flanking (300 bp UP, D)conserved putative binding sites for Lim1 in the Npnt locus. (C) Chromatinimmunoprecipitation against H3K4me1 enhancer mark. (D) Functional analysis ofLim1 binding sites in the Npnt enhancer sequences by cotransfection with Lim1expressing vector as in Fig. 5G.

Fig. 8. Model of renal gene regulatory network organization. A core regulatorysubcircuit composed of Pax2/8, Gata3 and Lim1 turns on a deeper layer oftranscriptional regulators composed of developmental molecules, while activatingeffector genes which are responsible for cell signaling and tissue organization (seetext for details).

S. Kamel Boualia et al. / Developmental Biology ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 9

identified few conserved regions and was extended to 100 kbupstream and 50 kb downstream of the Npnt coding region. In the15 conserved regions identified, 7 Lim1 putative binding siteswere identified (using the generic homeodomain consensussequence ATTA) (Mochizuki et al., 2000), 3 of which showedsignificant enrichment (L1Npnt-1 (2.8 fold), L1Npnt-2 (2.3 fold)and L1Npnt-3 (3.5 fold)) (Fig. 7A and B). The L1Npnt-1 site islocated 14 kb upstream of the transcriptional start site, while theL1Npnt-2 and L1Npnt-3 sites are located in the 3′ untranslatedregion of Nephronectin (Fig. 7A). ChIP analysis with an anti-H3K4Me1 antibody successfully amplified all three Lim1 targetsites in the Npnt locus, indicating that these sites are in an active

Please cite this article as: Kamel Boualia, S., et al., A core transcriptioplayers of pro/mesonephros morphogenesis. Dev. Biol. (2013), http:/

enhancer conformation (Fig. 7C). The functional analysis of thesesites further confirmed that Lim1 directly regulates their tran-scriptional activity (Fig. 7D and S8C).

Together, these ChIP analyses successfully identified a cis-regulatory interaction of important renal GRN genes by the threekernel regulators. The identification of specific cis-regulatorymodules demonstrates the direct role played by kernel regulatorson downstream regulatory and effector genes and allows a firstrepresentation of the regulatory genome leading to renal GRNdeployment.

Discussion

The early morphogenesis of organ systems is orchestrated bygene regulatory networks (GRN), which drive cellular organizationand differentiation. In the urogenital system, we have previouslyidentified the Pax2/8 and Gata3 transcription factor genes ascrucial regulators of specification and morphogenesis at the ear-liest stages of organ development. We took advantage of thesedevelopmental nodes to elucidate the hierarchical organization ofa first gene regulatory network underlying pro/mesonephrosdevelopment.

Our results reveal that Pax2/8 and Gata3 form a core or“kernel” GRN together with the transcription factor Lim1 (Fig. 8).Whereas the regulation of Gata3 by Pax2/8 was previouslyreported (Grote et al., 2006), this study identified several newgene interactions, one of which being the direct regulation of Lim1by Pax2/8 genes. Following activation by Pax2/8, the regulatorymolecules Gata3 and Lim1 maintain each other's expression(Fig. 8). Once established, this core renal GRN activates theexpression of a deeper layer of gene regulatory molecules as wellas a complex set of effector genes. For the purpose of characteriz-ing a first skeleton of the renal primordium GRN, we concentratedon downstream genes with a known or presumed role in pro/mesonephros development. We could thus link the regulatory andeffector genes Evi1, Id4, Npnt, Ret, Wfdc2, Pcdh19 and Plac8 to thePax2/8-Gata3-Lim1 core renal GRN (Fig. 8).

Regulatory network of pro/mesonephros development

Gene regulatory “kernels” are defined by a strong interrelation-ship between evolutionary conserved components of a networkand a critical role of these components for structure formation(Davidson, 2010; Davidson and Erwin, 2006). They are imperviousto change as a result of high evolutionary pressure applied on thesuccessful initiation of vital structure/organ development. A num-ber of evidence support the notion that Pax2/8, Gata3 and Lim1correspond to this definition and can therefore be considered asthe “renal primordium GRN kernel”. Our data shows that both

nal network composed of Pax2/8, Gata3 and Lim1 regulates key/dx.doi.org/10.1016/j.ydbio.2013.07.028i

123456789

101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566

676869707172737475767778798081828384858687888990919293949596979899

100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132

S. Kamel Boualia et al. / Developmental Biology ∎ (∎∎∎∎) ∎∎∎–∎∎∎10

Gata3 and Lim1 require Pax2/8 function while the latter two areinterdependent for normal expression in the renal primordium.This close interrelationship is likely to lock the renal fate programfollowing lineage specification by Pax2/8 (Bouchard et al., 2002).Importantly, each of them plays a different but essential role inpro/mesonephros development. While Pax2/8 are required forrenal lineage specification and survival (Bouchard et al., 2002),Gata3 is necessary for duct guidance, proliferation control and toprevent a premature differentiation of nephric duct cells (Chiaet al., 2011; Grote et al., 2008; Grote et al., 2006). As for Lim1, it isrequired for nephric duct elongation and integrity (Kobayashiet al., 2004; Pedersen et al., 2005; Shawlot and Behringer, 1995).As a consequence, embryos mutant for either of the genes fail tocomplete pro/mesonephros development and to initiate adult(metanephros) kidney development. In addition to this, all fourtranscription factors are evolutionary conserved in terms ofexpression, function and hierarchy in the renal primordium.Evidence from chick, Xenopus or zebrafish systems identifiedPax2 and Pax8 as the earliest factors specifically expressed in therenal system (Heller and Brandli, 1999; Krauss et al., 1991; Mauchet al., 2000; Pfeffer et al., 1998; Puschel et al., 1992), while Gata3and Lim1 are expressed in the early pronephros (Carroll et al.,1999; Sheng and Stern, 1999; Wingert and Davidson, 2011).Functional studies further suggest that these genes have conservedtheir regulatory role throughout vertebrate evolution (Bouchardet al., 2002; Carroll and Vize, 1999; Majumdar et al., 2000).

Interestingly, Pax, Gata and Lim1 factors have been identifiedin other demonstrated or presumed GRN kernels. In the mouse,Pax2-deficient embryos completely lack kidneys, cochlea andcerebellum (Torres et al., 1995, 1996). In the mid-hindbrainboundary organizer, which initiates posterior midbrain and cere-bellum development, Pax2 was identified as the primary regula-tory molecule working upstream of other regulators of theorganizer (Ye et al., 2001). Along the same line, Gata3 acts as alineage commitment factor for the Th2 hematopoietic lineage(Hosoya et al., 2010; Ting et al., 1996) and Lim1 is a crucial playerof head organizer function (Shawlot and Behringer, 1995).Together, these observations suggest that the transcription factorsinvolved in GRN kernels may have functional properties generallyapplicable to other kernel subcircuits.

Effectors of the pro/mesonephros GRN

Evidence we have collected here, together with related datafrom the literature suggests that the renal GRN is relatively deep.In immediate response to pro/mesonephros GRN kernel is a waveof transcriptional regulators and regulatory effectors. Among themare the transcriptional regulators Evi1, Id4 and Plac8, as well as thePax2 regulated gene Emx2 (Boualia et al., 2011). Evi1 deficiencyhas been associated with mesonephros hypoplasia (Hoyt et al.,1997), which may reflect a role in proliferation control. Id4 is amember of the inhibitor of differentiation gene family specificallyexpressed in renal tissue (Liu and Harland, 2003). Interestingly, Idproteins have been shown to bind and inactivate Pax2/5/8 DNA-binding activity (Roberts et al., 2001). This property raises theinteresting possibility that the activation of Id4 by Pax2 constitutesa negative feedback loop that serves to modulate the Pax2transcriptional response.

A similar subcircuit structure has been described for Plac8. Thisnewly described transcriptional regulator protein is necessary forglomerular development in zebrafish (Bedell et al., 2012). Inter-estingly, Plac8 (ponzr1 in zebrafish) also requires Pax2 for normalexpression. In addition, the downregulation of ponzr1 leads to anupregulation of Pax2, pointing to a negative feedback loop reg-ulating Pax2 expression levels. Noteworthy, Plac8/ponzr1 appearedrelatively recently during evolution and is therefore part of the

Please cite this article as: Kamel Boualia, S., et al., A core transcriptioplayers of pro/mesonephros morphogenesis. Dev. Biol. (2013), http:/

“evolutionary dynamic” gene pool. Evolutionary dynamic genesare thought to provide an additional level of complexity in higherorganisms. Accordingly, ponzr1 morpholino knockdown in zebra-fish leads to a pronephros phenotype characteristic of lowervertebrates (Bedell et al., 2012). The fact that Plac8/ponzr1 expres-sion and regulation by Pax2 has been conserved between zebrafishand mammals is consistent with the establishment and main-tenance of a more sophisticated regulatory subcircuit of renaldevelopment in recent evolution.

Regulation of signaling outputs

We have previously reported that Gata3 is required for Retexpression in the nephric duct and that this subcircuit is activatedby Pax2 and maintained by Wnt-β-catenin (Grote et al., 2006,2008). More recently, we have shown that the regulation of Ret byGata3 and retinoic acid signaling is necessary for the guidance andfusion of the nephric duct to the cloaca (Chia et al., 2011). In thecurrent network analysis, we reinforce the relationship betweenGata3 and Ret by the identification of additional direct bindingsites in the Ret regulatory region. In addition to regulating Ret, thepro/mesonephros GRN kernel indirectly regulates its ligand GDNFexpressed in the metanephric mesenchyme. Our data indeedidentify Nephronectin (Npnt) as a direct target of Lim1 in thenephric duct. Npnt is the ligand of the alpha8-beta1 integrincomplex (Brandenberger et al., 2001; Morimura et al., 2001).Together, the Npnt-Alpha8 integrin signaling complex is necessaryfor Gdnf expression in the mesenchyme. Accordingly, mousemutants for either Npnt or Itga8 (alpha8 integrin gene) fail toinduce metanephric kidney development (Linton et al., 2007;Muller et al., 1997). Hence the pro/mesonephros GRN regulatesureter budding in the epithelial (Ret) as well as the mesenchymal(Gdnf) tissue compartments.

One interesting observation resulting from our analysis is thateffector genes are activated at all levels of the network and notonly as an output to a previously established GRN. For example,Wfdc2 is exclusively regulated by Pax genes, while Npnt is mostlyaffected by loss of Lim1 function. Hence, the core componentsperform two parallel functions: (1) they unfold the lineage-specifictranscriptional program to diversify and refine the transcriptionalresponse and (2) they activate the lineage-specific effector anddifferentiation genes. We believe that the coordination of bothfunctions insures a balance between morphogenetic events andprogressive cell differentiation.

A key finding of our study is the identification of directinteractions between kernel gene components and key regulatorsof kidney development. This observation underscores the fact thatgene inactivation phenotypes cannot be taken in isolation butrather as a series of additive misregulatory events. This may havedirect consequences in the understanding of developmental dis-eases affecting the kidney and urinary tract. The terminologyCongenital Anomalies of the Kidney and Urinary Tract (CAKUT)is used to describe a series of developmental defects affectingkidney and urinary tract development. It is characterized by a highdegree of intra and interfamilial variability (Weber, 2012). Thisvariability has been proposed to be the result of a threshold effectfor key regulatory molecules such that a given developmentalstage proceeds or fails depending on whether a given genefunction is sufficient or not. In the context of the pro/mesonephrosGRN, we propose that CAKUT variability would be better modeledas network dynamics deficiencies in which the activation, main-tenance and feedback relations of a number of molecules deter-mine whether or not a morphogenetic process is completedsuccessfully. Hence, one expects CAKUT to be caused by a numberof different interrelated GRN members and lead to a spectrum ofpossible phenotypes.

nal network composed of Pax2/8, Gata3 and Lim1 regulates key/dx.doi.org/10.1016/j.ydbio.2013.07.028i

123456789

101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566

676869707172737475767778

S. Kamel Boualia et al. / Developmental Biology ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 11

Acknowledgments

We would like to thank Drs. M. Busslinger, (IMP, Vienna), R.Behringer (U. Texas) and Amed Mansouri (MPI for BiophysicalChemistry, Göttingen, Germany) for providing mouse strainsand Drs. J. Pelletier (McGill University) and P. Goodyer (McGillUniversity) for reagents. This work was supported by a grant fromthe Canadian Institutes for Health Research (CIHR; MOP-84470).M.B. holds a Canada Research Chair in Developmental Genetics ofthe Urogenital System. M.T. was supported by a Dr. Gerald B. PriceFellowship (CRS), MICRTP and Fonds de Recherche Québec Santé(FRQS) postdoctoral fellowships.

79808182838485

Appendix A. Supporting information

Supplementary data associated with this article can be found inthe online version at http://dx.doi.org/10.1016/j.ydbio.2013.07.028.

8687888990919293949596979899

100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132

References

Alon, U., 2007. Network motifs: theory and experimental approaches. Nat. Rev.Genet. 8, 450–461.

Bedell, V.M., Person, A.D., Larson, J.D., McLoon, A., Balciunas, D., Clark, K.J., Neff, K.I.,Nelson, K.E., Bill, B.R., Schimmenti, L.A., Beiraghi, S., Ekker, S.C., 2012.The lineage-specific gene ponzr1 is essential for zebrafish pronephric andpharyngeal arch development. Development 139, 793–804.

Boualia, S.K., Gaitan, Y., Murawski, I., Nadon, R., Gupta, I.R., Bouchard, M., 2011.Vesicoureteral reflux and other urinary tract malformations in mice compoundheterozygous for Pax2 and Emx2. PLoS One 6, e21529.

Bouchard, M., 2004. Transcriptional control of kidney development. Differentiation72, 295–306.

Bouchard, M., Grote, D., Craven, S.E., Sun, Q., Steinlein, P., Busslinger, M., 2005.Identification of Pax2-regulated genes by expression profiling of the mid-hindbrain organizer region. Development 132, 2633–2643.

Bouchard, M., Souabni, A., Mandler, M., Neubuser, A., Busslinger, M., 2002. Nephriclineage specification by Pax2 and Pax8. Genes Dev. 16, 2958–2970.

Bouchard, M., St. Amand, J., Cote, S., 2000. Combinatorial activity of pair-ruleproteins on the Drosophila gooseberry early enhancer. Dev. Biol. 222, 135–146.

Brandenberger, R., Schmidt, A., Linton, J., Wang, D., Backus, C., Denda, S., Muller, U.,Reichardt, L.F., 2001. Identification and characterization of a novel extracellularmatrix protein nephronectin that is associated with integrin alpha8beta1 in theembryonic kidney. J. Cell Biol. 154, 447–458.

Cai, Q., Dmitrieva, N.I., Ferraris, J.D., Brooks, H.L., van Balkom, B.W., Burg, M., 2005.Pax2 expression occurs in renal medullary epithelial cells in vivo and in cellculture, is osmoregulated, and promotes osmotic tolerance. Proc. Natl. Acad. Sci.USA 102, 503–508.

Carroll, T.J., Vize, P.D., 1999. Synergism between Pax-8 and lim-1 in embryonickidney development. Dev. Biol. 214, 46–59.

Carroll, T.J., Wallingford, J.B., Vize, P.D., 1999. Dynamic patterns of gene expressionin the developing pronephros of Xenopus laevis. Dev. Genet. 24, 199–207.

Chia, I., Grote, D., Marcotte, M., Batourina, E., Mendelsohn, C., Bouchard, M., 2011.Nephric duct insertion is a crucial step in urinary tract maturation that isregulated by a Gata3-Raldh2-Ret molecular network in mice. Development 138,2089–2097.

Cripps, R.M., Olson, E.N., 2002. Control of cardiac development by an evolutionarilyconserved transcriptional network. Dev. Biol. 246, 14–28.

Davidson, E.H., 2009. Network design principles from the sea urchin embryo. Curr.Opin. Genet. Dev. 19, 535–540.

Davidson, E.H., 2010. Emerging properties of animal gene regulatory networks.Nature 468, 911–920.

Davidson, E.H., Erwin, D.H., 2006. Gene regulatory networks and the evolution ofanimal body plans. Science 311, 796–800.

Dressler, G.R., 2006. The cellular basis of kidney development. Ann. Rev. Cell. Dev.Biol. 22, 509–529.

Gaitan, Y., Bouchard, M., 2006. Expression of the delta-protocadherin gene Pcdh19in the developing mouse embryo. Gene Expr. Patterns 6, 893–899.

Grote, D., Boualia, S.K., Souabni, A., Merkel, C., Chi, X., Costantini, F., Carroll, T.,Bouchard, M., 2008. Gata3 acts downstream of beta-catenin signaling toprevent ectopic metanephric kidney induction. PLoS Genet. 4, e1000316.

Grote, D., Souabni, A., Busslinger, M., Bouchard, M., 2006. Pax2/8-regulated Gata3expression is necessary for morphogenesis and guidance of the nephric duct inthe developing kidney. Development 133, 53–61.

Heller, N., Brandli, A.W., 1999. Xenopus Pax-2/5/8 orthologues: novel insights intoPax gene evolution and identification of Pax-8 as the earliest marker for oticand pronephric cell lineages. Dev. Genet. 24, 208–219.

Please cite this article as: Kamel Boualia, S., et al., A core transcriptioplayers of pro/mesonephros morphogenesis. Dev. Biol. (2013), http:/

Hellstrom, I., Raycraft, J., Hayden-Ledbetter, M., Ledbetter, J.A., Schummer, M.,McIntosh, M., Drescher, C., Urban, N., Hellstrom, K.E., 2003. The HE4 (WFDC2)protein is a biomarker for ovarian carcinoma. Cancer Res. 63, 3695–3700.

Henrique, D., Adam, J., Myat, A., Chitnis, A., Lewis, J., Ish-Horowicz, D., 1995.Expression of a Delta homologue in prospective neurons in the chick. Nature375, 787–790.

Hosoya, T., Maillard, I., Engel, J.D., 2010. From the cradle to the grave: activities of GATA-3throughout T-cell development and differentiation. Immunol. Rev. 238, 110–125.

Hoyt, P.R., Bartholomew, C., Davis, A.J., Yutzey, K., Gamer, L.W., Potter, S.S., Ihle, J.N.,Mucenski, M.L., 1997. The Evi1 proto-oncogene is required at midgestation forneural, heart, and paraxial mesenchyme development. Mech. Dev. 65, 55–70.

Kobayashi, A., Kwan, K.M., Carroll, T.J., McMahon, A.P., Mendelsohn, C.L., Behringer,R.R., 2005. Distinct and sequential tissue-specific activities of the LIM-classhomeobox gene Lim1 for tubular morphogenesis during kidney development.Development 132, 2809–2823.

Kobayashi, A., Shawlot, W., Kania, A., Behringer, R.R., 2004. Requirement of Lim1 forfemale reproductive tract development. Development 131, 539–549.

Koide, T., Hayata, T., Cho, K.W., 2005. Xenopus as a model system to studytranscriptional regulatory networks. Proc. Natl. Acad. Sci. USA 102, 4943–4948.

Kozmik, Z., Wang, S., Dorfler, P., Adams, B., Busslinger, M., 1992. The promoter of theCD19 gene is a target for the B-cell-specific transcription factor BSAP. Mol. CellBiol. 12, 2662–2672.

Krauss, S., Johansen, T., Korzh, V., Fjose, A., 1991. Expression of the zebrafish pairedbox gene pax[zf-b] during early neurogenesis. Development 113, 1193–1206.

Levine, M., Davidson, E.H., 2005. Gene regulatory networks for development. Proc.Natl. Acad. Sci. USA 102, 4936–4942.

Linton, J.M., Martin, G.R., Reichardt, L.F., 2007. The ECM protein nephronectinpromotes kidney development via integrin alpha8beta1-mediated stimulationof Gdnf expression. Development 134, 2501–2509.

Liu, K.J., Harland, R.M., 2003. Cloning and characterization of Xenopus Id4 revealsdiffering roles for Id genes. Dev. Biol. 264, 339–351.

Majumdar, A., Lun, K., Brand, M., Drummond, I.A., 2000. Zebrafish no isthmusreveals a role for pax2.1 in tubule differentiation and patterning events in thepronephric primordia. Development 127, 2089–2098.

Mauch, T.J., Yang, G., Wright, M., Smith, D., Schoenwolf, G.C., 2000. Signals fromtrunk paraxial mesoderm induce pronephros formation in chick intermediatemesoderm. Dev. Biol. 220, 62–75.

Miyamoto, N., Yoshida, M., Kuratani, S., Matsuo, I., Aizawa, S., 1997. Defects ofurogenital development in mice lacking Emx2. Development 124, 1653–1664.

Mochizuki, T., Karavanov, A.A., Curtiss, P.E., Ault, K.T., Sugimoto, N., Watabe, T.,Shiokawa, K., Jamrich, M., Cho, K.W., Dawid, I.B., Taira, M., 2000. Xlim-1 and LIMdomain binding protein 1 cooperate with various transcription factors in theregulation of the goosecoid promoter. Dev. Biol. 224, 470–485.

Moore, M.W., Klein, R.D., Farinas, I., Sauer, H., Armanini, M., Phillips, H., Reichardt, L.F., Ryan, A.M., Carver_Moore, K., Rosenthal, A., 1996. Renal and neuronalabnormalities in mice lacking GDNF. Nature 382, 76–79.

Morimura, N., Tezuka, Y., Watanabe, N., Yasuda, M., Miyatani, S., Hozumi, N.,Tezuka Ki, K., 2001. Molecular cloning of POEM: a novel adhesion moleculethat interacts with alpha8beta1 integrin. J. Biol. Chem. 276, 42172–42181.

Muller, U., Wang, D., Denda, S., Meneses, J.J., Pedersen, R.A., Reichardt, L.F., 1997.Integrin alpha8beta1 is critically important for epithelial-mesenchymal inter-actions during kidney morphogenesis. Cell 88, 603–613.

Orkin, S.H., 1992. GATA-binding transcription factors in hematopoietic cells. Blood80, 575–581.

Pedersen, A., Skjong, C., Shawlot, W., 2005. Lim 1 is required for nephric ductextension and ureteric bud morphogenesis. Dev. Biol. 288, 571–581.

Pfeffer, P.L., Gerster, T., Lun, K., Brand, M., Busslinger, M., 1998. Characterization ofthree novel members of the zebrafish Pax2/5/8 family: dependency of Pax5 andPax8 expression on the Pax2.1 (noi) function. Development 125, 3063–3074.

Pichel, J.G., Shen, L., Sheng, H.Z., Granholm, A.C., Drago, J., Grinberg, A., Lee, E.J.,Huang, S.P., Saarma, M., Hoffer, B.J., Sariola, H., Westphal, H., 1996. Defects inenteric innervation and kidney development in mice lacking GDNF. Nature 382,73–76.

Pimanda, J.E., Gottgens, B., 2010. Gene regulatory networks governing haemato-poietic stem cell development and identity. Int. J. Dev. Biol. 54, 1201–1211.

Puschel, A.W., Westerfield, M., Dressler, G.R., 1992. Comparative analysis of Pax-2protein distributions during neurulation in mice and zebrafish. Mech. Dev. 38,197–208.

Roberts, E.C., Deed, R.W., Inoue, T., Norton, J.D., Sharrocks, A.D., 2001. Id helix-loop-helix proteins antagonize pax transcription factor activity by inhibiting DNAbinding. Mol. Cell. Biol. 21, 524–533.

Sanchez, M.P., Silos-Santiago, I., Frisen, J., He, B., Lira, S.A., Barbacid, M., 1996. Renalagenesis and the absence of enteric neurons in mice lacking GDNF. Nature 382,70–73.

Schuchardt, A., D_Agati, V., Larsson_Blomberg, L., Costantini, F., Pachnis, V., 1994.Defects in the kidney and enteric nervous system of mice lacking the tyrosinekinase receptor Ret. Nature 367, 380–383.

Schuchardt, A., D'Agati, V., Pachnis, V., Costantini, F., 1996. Renal agenesis andhypodysplasia in ret-k- mutant mice result from defects in ureteric buddevelopment. Development 122, 1919–1929.

Shawlot, W., Behringer, R.R., 1995. Requirement for Lim1 in head-organizerfunction. Nature 374, 425–430.

Sheng, G., Stern, C.D., 1999. Gata2 and Gata3: novel markers for early embryonic polarityand for non-neural ectoderm in the chick embryo. Mech. Dev. 87, 213–216.

Ting, C.N., Olson, M.C., Barton, K.P., Leiden, J.M., 1996. Transcription factor GATA-3 isrequired for development of the T-cell lineage. Nature 384, 474–478.

nal network composed of Pax2/8, Gata3 and Lim1 regulates key/dx.doi.org/10.1016/j.ydbio.2013.07.028i

123456789

10111213141516

S. Kamel Boualia et al. / Developmental Biology ∎ (∎∎∎∎) ∎∎∎–∎∎∎12

Torres, M., Gomez_Pardo, E., Gruss, P., 1996. Pax2 contributes to inner earpatterning and optic nerve trajectory. Development 122, 3381–3391.

Torres, M., Gomez-Pardo, E., Dressler, G.R., Gruss, P., 1995. Pax-2 controls multiplesteps of urogenital development. Development 121, 4057–4065.

Weber, S., 2012. Novel genetic aspects of congenital anomalies of kidney andurinary tract. Curr. Opin. Pediatr. 24, 212–218.

Wingert, R.A., Davidson, A.J., 2011. Zebrafish nephrogenesis involves dynamicspatiotemporal expression changes in renal progenitors and essential signalsfrom retinoic acid and irx3b. Dev. Dyn. 240, 2011–2027.

Please cite this article as: Kamel Boualia, S., et al., A core transcriptioplayers of pro/mesonephros morphogenesis. Dev. Biol. (2013), http:/

Ye, W., Bouchard, M., Stone, D., Liu, X., Vella, F., Lee, J., Nakamura, H., Ang, S.L.,Busslinger, M., Rosenthal, A., 2001. Distinct regulators control the expression ofthe mid-hindbrain organizer signal FGF8. Nat. Neurosci. 4, 1175–1181.

Yoshida, M., Suda, Y., Matsuo, I., Miyamoto, N., Takeda, N., Kuratani, S., Aizawa, S.,1997. Emx1 and Emx2 functions in development of dorsal telencephalon.Development 124, 101–111.

17

nal network composed of Pax2/8, Gata3 and Lim1 regulates key/dx.doi.org/10.1016/j.ydbio.2013.07.028i