HOXA13 Directly Regulates Epha6 and Epha7 Expression in the Genital Tubercle Vascular Endothelia

Home , EPHA6, EPHA7, Ephrin A1, Ephrin B2

DEVELOPMENTAL DYNAMICS 236:951–960, 2007

RESEARCH ARTICLE

HOXA13 Directly Regulates EphA6 and EphA7 Expression in the Genital Tubercle Vascular Endothelia

Carley A. Shaut,1† Chie Saneyoshi,1† Emily A. Morgan,1 Wendy M. Knosp,2 Diane R. Sexton,3 and H. Scott Stadler1,3*

Hypospadias, a common defect affecting the growth and closure of the external genitalia, is often accompanied by gross enlargements of the genital tubercle (GT) vasculature. Because Hoxa13 homozygous mutant mice also exhibit hypospadias and GT vessel expansion, we examined whether genes playing a role in angiogenesis exhibit reduced expression in the GT. From this analysis, reductions in EphA6 and EphA7 were detected. Characterization of EphA6 and EphA7 expression in the GT confirmed colocalization with HOXA13 in the GT vascular endothelia. Analysis of the EphA6 and EphA7 promoter regions revealed a series of highly conserved cis-regulatory elements bound by HOXA13 with high affinity. GT chromatin immunoprecipitation confirmed that HOXA13 binds these gene-regulatory elements in vivo. In vitro, HOXA13 activates gene expression through the EphA6 and EphA7 gene-regulatory elements. Together these findings indicate that HOXA13 directly regulates EphA6 and EphA7 in the developing GT and identifies the GT vascular endothelia as a novel site for HOXA13-dependent expression of EphA6 and EphA7. Developmental Dynamics 236:951–960, 2007. © 2007 Wiley-Liss, Inc.

Key words: Hoxa13; Hypospadias; genital tubercle; vascular endothelia; EphA6; EphA7

Accepted 4 January 2007

INTRODUCTION 2003). One phenotype commonly as- and Eichmann, 2005; Davy and So- sociated with hypospadias is the en- riano, 2005). Indeed, Eph–ephrin Hypospadias, a defect affecting the largement of blood vessels supplying signaling is essential for the pattern- growth and closure of the external the glans or prepuce (Baskin et al., ing of multiple tissues and cell types, genitalia, is highly prevalent in the birth populations of industrialized 1998; Baskin, 2000). At present, no including vascular endothelial cell nations, including the United States, molecular link between hypospadias assembly, cell migration, mesenchy- United Kingdom, Sweden, and Ja- and vessel enlargement in the geni- mal cell condensation, vascular bed pan (Giwercman et al., 1993; Pau- talia has been identiﬁed; however, formation, tumor neovasculariza- lozzi et al., 1997; Gallentine et al., studies examining perturbations in tion, and the closure of the external 2001). While the frequency of hypo- Eph–ephrin signaling may provide genitalia (Wang et al., 1998; Ogawa spadias ranges as high as 1 in 125 important clues toward understand- et al., 2000; Stadler et al., 2001; live births, the molecular mecha- ing the pathology of hypospadias and Chan et al., 2001; Dravis et al., 2004; nisms underlying this defect are its associated vascular malforma- Davy et al., 2004; Marquardt et al., poorly understood (Svensson et al., tions (reviewed by Eichmann et al., 2005; Egea et al., 2005). 1997; Paulozzi et al., 1997; Stadler, 2005a,b; Hinck, 2004; Klagsbrun Recently, we and others have

1Department of Molecular and Medical Genetics, Oregon Health and Science University, Portland, Oregon 2UCSF-Mission Bay Campus, San Francisco, California 3Shriners Hospital for Children Research Division, Portland, Oregon Grant sponsor: the National Institutes of Health; Grant number: HSS R01 DK66539; Grant sponsor: the American Heart Association. †Drs. Shaut and Saneyoshi contributed equally to this work. *Correspondence to: H. Scott Stadler, Shriners Hospital for Children Research Division, Portland, OR 97239. E-mail: [email protected] DOI 10.1002/dvdy.21077 Published online 15 February 2007 in Wiley InterScience (www.interscience.wiley.com).

© 2007 Wiley-Liss, Inc. 952 SHAUT ET AL. shown that HOXA13 function is necessary for EphA7 expression in the developing limb (Stadler et al., 2001; Salsi and Zappavigna, 2006). Recog- nizing that Hoxa13-deficient mice also exhibit hypospadias and capillary vessel enlargement (Morgan et al., 2003), we hypothesized that HOXA13 may regulate Eph receptor expression in the genital tubercle (GT) and its vasculature. Testing this hypothesis, we report that HOXA13 directly regulates EphA6 and EphA7 expression in the GT vascular endothelia. Analysis of the EphA6 and EphA7 promoter regions revealed a conserved series of DNA sequences bound with high affinity by the HOXA13 DNA binding domain (A13). In vivo, direct interactions between HOXA13 and the EphA6 and EphA7 promoter elements were detected in the GT using Hoxa13-directed ChIP. In vitro, HOXA13 can use the bound gene- regulatory elements in the EphA6 and EphA7 promoters to direct gene expression. Together these findings indicate that EphA6 and EphA7 are direct transcriptional targets of HOXA13 in the GT vascular endothelia, providing new insight into the cell-signaling mechanisms functioning during the growth and development of the external genitalia.

RESULTS Fig. 1. Hoxa13-deficient mice display enlarged blood vessels in the developing genital tubercle. A–L: Hematoxylin and eosin staining of sectioned genital tubercles from wild-type (A–D), homozygous Vessel Expansion Is Present mutant male (E–H), and homozygous mutant female (I–L) embryos at embryonic days (E) 12.5–15.5. Throughout GT Development Arrows denote a typical example of normal vessel diameters in wild-type embryos (A–D) or enlarged vessels in homozygous mutants (E–L). UPE, urethral plate epithelium. Scale bar ϭ 100 ␮m. in Hoxa13–Green Fluorescent Protein Homozygous Mutants Analysis of the distal GT from embryonic days (E) 12.5 to 15.5 revealed average vessel diameters of 15 ␮m(Ϯ 0.5 ␮m; n ϭ 12 independent samples) in Hoxa13-GFP (GFP, green fluorescent protein) wild-type and heterozygous mutants. In contrast, the GT vasculature of age- matched homozygous mutants exhibited greatly enlarged vessels with average diameters of 85 ␮m(Ϯ 15 ␮m; n ϭ 12 independent samples; Fig. 1). Of interest, embryonic sex Fig. 7. Genital tubercle (GT) vessel diameter and endothelial cell identity are maintained in EphA7 did not influence the presentation of homozygous mutants. A: Typical GT vessel diameters and endothelial cell adhesion molecule-1 (PECAM-1) expression in embryonic day (E) 13.5 wild-type control embryos. B,C: The GT vascu- the vascular phenotype as homozy- lature of EphA7 homozygous mutants (B) does not exhibit vessel enlargement or alteration in gous mutant male and female em- PECAM-1 expression when compared with the enlarged GT vasculature of Hoxa13-GFP homozy- bryos exhibited similar increases in gous mutants (C). Scale bar ϭ 50 ␮m. HOXA13 REGULATES EPHA6,7 EXPRESSION IN THE GT 953

Fig. 2. Immunohistochemical localization of EPHA6 and EPHA7 in the genital tubercle (GT) of E13.5 Hoxa13-green ﬂuorescent protein (GFP) mice. A,C: EPHA6 and EPHA7 (red signal) are expressed in the GT mesenchyme (MES) as well as the endothelial cells lining the GT vasculature (arrows) of heterozygous Hoxa13-GFP embryos. B,D: EPHA6 and EPHA7 expression is reduced in the enlarged GT vessels of age-matched Hoxa13 homozygous mutants (arrows). E,G: Colocalization (yellow signal) of Hoxa13-GFP with EPHA6 and EPHA7 in the GT vascular endothelia of heterozygous embryos. Arrows denote colocalization in the GT endothelia. F,H: Reduced expression of EPHA6 and EPHA7 in the homozygous mutant vascular endothelia. Arrows denote the enlarged vasculature. I–L: Higher magniﬁcation images of the heterozygous control and homozygous mutant vessels depicted in E–H. M,N: Platelet endothelial cell adhesion molecule-1 (PECAM-1) expression (red) is maintained in the GT vascular endothelia of Hoxa13-GFP (green) heterozygous mutant controls and age-matched homozygous mutants. UPE, urethral plate epithelium; MES, mesenchymal tissue; R, red blood cells. For clarity, only male GTs are shown, although similar vascular defects were observed in mutant female GTs. Scale bars ϭ 140 ␮m in A–H, 20 ␮m in I–L, 70 ␮m in M–P.

GT vessel diameter (Fig. 1, compare with perturbations in Eph–ephrin both EPHA6 and EPHA7 were con- E–H with I–L). signaling, we examined whether the sistently reduced in the GT vascular affected GT vasculature exhibited endothelia compared with heterozy- EphA6 and EphA7 Are changes in the expression of Eph re- gous mutant controls (Fig. 2A–D), Reduced in the GT Vascular ceptors or their ephrin ligands. which do not exhibit a GT phenotype Endothelia of Hoxa13 While no changes in EPHA2, (Morgan et al., 2003). In the vascu- EPHA4, EPHA5, EPHB2, EPHRIN lature, HOXA13-GFP, EPHA6, and Homozygous Mutants A2, or EPHRIN A5 expression were EPHA7 were strongly colocalized in Recognizing that defects in vascular detected in the Hoxa13-GFP ho- the endothelial layer of heterozygous patterning are strongly associated mozygous mutants (data not shown), controls, whereas the reduced levels 954 SHAUT ET AL. of EPHA6 and EPHA7 in the homozygous mutants minimized our detection of colocalization in the ex- panded GT vasculature (Fig. 2E–L). Platelet endothelial cell adhesion molecule-1 (PECAM-1) expression was also present in the affected GT vasculature, suggesting that endothelial cell identity was not affected by the loss of HOXA13 function (Fig. 2M–N). Semiquantitative reverse transcriptase-polymerase chain reaction (RT- PCR) analysis confirmed the levels of EphA6 and EphA7 expression detected by immunohistochemistry. In particu- Fig. 3. Quantitation of Hoxa13, EphA6, and EphA7 expression in the genital tubercle (GT) mesenchyme and vascular endothelia. A: Semiquantitative RT-PCR of EphA6 and EphA7 tran- lar, a uniform level of EphA6 and scripts revealed no differences in EphA6 and EphA7 expression in the whole distal GT. B: Hoxa13, EphA7 expression was detected EphA6, and EphA7 are coexpressed in the GT vascular endothelium as determined by semiquan- throughout the GT mesenchyme in titative reverse transcriptase-polymerase chain reaction on platelet endothelial cell adhesion mol- Hoxa13-GFP heterozygous and ho- ecule (PECAM) -positive endothelial cell isolates. Furthermore, both EphA6 and EphA7 expression mozygous mutant embryos (Fig. 3A). levels were reduced in vascular endothelial cells of Hoxa13 homozygous mutant GT compared with Hoxa13 heterozygous controls. The glyceraldehydes-3-phosphate dehydrogenase (GAPDH) Next, because cell-specific changes in housekeeping gene was used as a cDNA control. gene expression are often not detectable in whole tissue RNA samples, we examined whether endothelial cells purified ens), rat (Rattus norvegicus), and opos- creasing the HOXA13 DNA binding GT vasculature exhibited reductions in sum (Monodelphis domestica) with peptide concentration to 2 ␮M produced EphA6 and EphA7 expression, as ob- complete identity in several A-T–rich a single higher molecular weight prod- served by immunohistochemistry. RT- regions where HOXA13 is thought to uct for the EphA6 and EphA7 promoter PCR analysis of GT endothelial cell iso- bind (Knosp et al., 2004; McCabe and elements, indicating saturation of all lates confirmed that EphA6 and EphA7 Innis, 2005; Salsi and Zappavigna, Hoxa13 binding sites (Fig. 4C). are expressed in the endothelial compo- 2006; Fig. 4B). For the EphA7 frag- Next, to establish whether the nent of the GT vessels along with ment, strong sequence conservation HOXA13 DNA binding domain peptide Hoxa13 and PECAM-1 in Hoxa13-GFP was also detected between mouse, hu- binds DNA as monomer or as higher heterozygous mutants (Fig. 3B). In con- man, rat, opossum, and several addi- order complex, we performed a sedi- trast, the expression of EphA6 and tional mammalian species, including mentation equilibrium analysis. Analy- EphA7 was consistently reduced (n ϭ 3 dog (Canis familiarus), armadillo sis of the sedimentation the HOXA13 independent assessments) in the puri- (Asypus novemcinctus), and elephant DNA binding peptide bound to DNA re- fied GT vascular endothelia of Hoxa13- (Loxodonta africana), with strong iden- vealed a single complex of approxi- GFP homozygous mutants, although no tity within the putative HOXA13 bind- mately 16,900 daltons, which correlates difference in PECAM-1 expression was ing site (Fig. 4B). to a 1:1 stoichiometric ratio of peptide observed (Fig. 3B). Electrophoretic mobility shift assay (8,137 daltons) to DNA 8,100 (daltons), (EMSA) analysis of the EphA6 region suggesting that the staggered bands de- HOXA13 Binds Discrete (Ϫ2410 to Ϫ2067) revealed consistent tected by EMSA most likely reflect the Regions of the EphA6 and binding by the HOXA13 DNA binding concentration-dependent saturation of peptide, which could be competitively multiple binding sites by the mono- EphA7 Promoters removed using identical nonlabeled meric protein, rather than higher order DNA sequence analysis of the EphA6 DNA fragments (Fig. 4C). Similar bind- complexes. and EphA7 promoter regions revealed ing was also observed for the 325-bp several A-T–rich regions matching se- EphA7 promoter region (Ϫ839 to Ϫ514) HOXA13 Binds the EphA6 quences we previously identified as as well as an additional HOXA13-regu- and EphA7 cis-Regulatory sites bound by HOXA13 (Knosp et al., lated region previously characterized by Elements In Vivo 2004; Fig. 4A). Because key gene-regu- Salsi and Zappavigna (2006; data not latory regions are often conserved be- shown; Fig. 4C). Interestingly, both the Next, to determine whether HOXA13 tween species, we examined whether EphA6 and EphA7 regions exhibited a directly interacts with the EphA6 and the EphA6 and EphA7 regions contain- staggered series of mobility-shifted EphA7 promoter sequences in vivo, we ing the clustered HOXA13 binding sites bands when incubated with 0.2 ␮M A13 examined whether GT-specific chroma- are conserved using the UCSC Genome peptide, suggesting that multiple bind- tin containing the conserved EphA6 Browser (Kent et al., 2002). For the ing sites are present in the EphA6 and and EphA7 regions was immunopre- EphA6 element, a high degree of con- EphA7 conserved regions, which re- cipitated with a HOXA13 antibody servation was observed between mouse quire a higher protein concentration to (␣A13). Previous characterization of the (Mus musculus), human (Homo sapi- be completely saturated (Fig. 4C). In- HOXA13 antibody confirmed that it can HOXA13 REGULATES EPHA6,7 EXPRESSION IN THE GT 955

Fig. 5. HOXA13 associates with the EphA6 and EphA7 promoter fragments in vivo in the developing genital tubercle (GT). A,B: Positive polymerase chain reaction (PCR) amplification of the EphA6 and EphA7 conserved regions after HOXA13 chromatin immunoprecipitation Fig. 4. HOXA13 binds discrete regions of the EphA6 and EphA7 promoters. A: Sequences of the (ChIP) confirms that wild-type HOXA13 binds murine EphA6 and EphA7 promoter regions. Underlined nucleotides denote regions exhibiting high the EphA6 (A) and EphA7 (B) promoter regions conservation among multiple species. Lower case nucleotides denote primer sites used to amplify in the embryonic day (E) 12.5 GT. Note that the promoter region for electrophoretic mobility shift assay (EMSA) analysis and chromatin immu- parallel ChIP assays of GT chromatin from ho- noprecipitation in Figure 5. B: Species conservation of the EphA6 and EphA7 promoter regions mozygous mutants did not detect these same bound by the HOXA13 DNA binding domain peptide (A13). Sequence conservation and alignment DNA, suggesting that HOXA13Јs DNA binding were determined using BLAST analysis software and the UCSC Genome Browser Database (Kent function is necessary for the in vivo binding of et al., 2002, 2005; Karolchik et al., 2003). C: EMSA analysis of the EphA6 and EphA7 promoter HOXA13 to the conserved EphA6 and EphA7 regions. At 0.2 ␮M, the HOXA13 DNA binding domain peptide (A13) exhibits nonsaturating levels regions. NC, negative PCR control; IgG, immu- of binding to several Hoxa13 binding sites present in the EphA6 and EphA7 promoter fragments, noglobulin antibody control; Input, positive whereas at 2.0 ␮M of A13, all binding sites present in the EphA6 and EphA7 promoter fragments control confirming the presence of the EphA6 appear bound, producing a single reduced mobility DNA fragment. Note that the addition of 0.2 ␮M and EphA7 conserved regions in the chromatin unlabeled EphA6 or EphA7 promoter fragments was sufficient to competitively displace the A13 samples before immunoprecipitation. peptide from the radiolabeled promoter fragments. bind both HOXA13 wild-type and mu- EphA6 and EphA7 promoter frag- by HOXA13 may function as an en- tant proteins and facilitate the immu- ments was assessed in NG108-15 hancer (Fig. 6). noprecipitation of gene-regulatory ele- cells. In the absence of the EphA6 or ments directly bound by wild-type EphA7 promoter fragments, the Loss of EphA7 Is Not HOXA13 (Knosp et al., 2004). PCR am- empty pGL4.1 luciferase plasmid pro- Sufficient to Cause Gross plification of wild-type ␣A13-chromatin duced insignificant amounts of lucif- immunoprecipitates consistently de- erase when cotransfected with a Enlargement of the GT tected the EphA6 and EphA7 promoter Hoxa13 expression plasmid (pCM- Vasculature ϭ fragments (n 5 independent assess- VHoxa13; Fig. 6). Cotransfection of The individual function of EphA7 was ments; Fig. 5). In homozygous mutants, pCMVHoxa13 with pGL4.1 plasmid assessed in the developing GT vascu- the EphA6 and EphA7 promoter frag- containing the forward orientation of lature using a null EphA7 allele ments could not be detected in the the EphA6 promoter fragment pro- (Holmberg et al., 2000). A comparison ␣ A13-immunoprecipitated chromatin duced nearly a fourfold increase in lu- of the GT vessel diameters between ϭ (n 5 independent assessments; Fig. ciferase expression (Fig. 6). Of inter- wild-type and EphA7 homozygous Ј 5), suggesting that HOXA13 s DNA est, cotransfection of pCMVHoxa13 mutant embryos revealed no signifi- binding function, which is absent in the with the pGL4.1 plasmid containing cant expansions of the GT vessels mutant HOXA13-GFP protein is neces- the reverse orientation of EphA6 pro- (Fig. 7, page 952), suggesting that the sary for the immunoprecipitation of moter fragment produced no increase expression of EphA6 in the GT vascu- these gene-regulatory sequences. in luciferase expression, suggesting lature may be sufficient to compensate the EphA6 gene-regulatory region for the loss of EphA7 function, or that In Vitro Utilization of the functions in an orientation-specific EphA7 does not play a role in regulat- EphA6 and EphA7 Promoter manner. In contrast, activation of the ing vessel wall diameter. The effect of Elements by Full-Length EphA7 promoter fragment by a combinatorial loss of EphA6 and HOXA13 facilitated luciferase expres- EphA7 in the GT vasculature could HOXA13 sion at levels three- to fourfold higher not be assessed as EphA6 null muta- The capacity of full-length HOXA13 to than controls independent of its orien- tions are early embryonic lethal regulate gene expression through the tation, suggesting the sequence bound (Brown et al., 2000). 956 SHAUT ET AL.

consequences of perturbations in Eph regulate EphA7 expression (Salsi and receptor signaling indicate that Zappavigna, 2006). This finding pro- changes in cell migration, morphol- vides a possible explanation for the ogy, and adhesion are the major phe- maintenance of EphA6 and EphA7 ex- notypes in tissues lacking one or more pression in the GT mesenchyme of Eph receptor or ephrin ligand Hoxa13 homozygous mutants as both (Shamah et al., 2001; Wahl et al., Hoxa13 and Hoxd13 are coexpressed 2000; reviewed by Pasquale, 2005; strongly in this region (Warot et al., Cooke et al., 2005). In the developing 1997, Scott et al., 2005). More impor- vasculature, functional studies of tantly, among the group 13 HOX pro- EphA2, ephrin A1, ephrin B2, and teins, only HOXA13 has been reported EphB4 firmly establish a role for Eph– to be expressed in vascular endothelia ephrin signaling throughout the an- (this work; Warot et al., 1997; Stadler Fig. 6. HOXA13 activates gene expression giogenic process, including endo- et al., 2001). Thus, in the GT vascula- through the conserved EphA6 and EphA7 re- thelial cell proliferation, assembly, ture, the loss of HOXA13 function gion in vitro. Cotransfection of NG108-15 cells and extracellular matrix remodeling would affect EphA7 expression more with a pGL4.1 luciferase reporter plasmids con- (McBride and Ruiz, 1998; Gerety et severely, due to the absence of func- taining forward or reverse orientations of the al., 1999; Adams et al., 2001; Hunter tionally redundant factors such as EphA6 region bound by HOXA13 with a HOXA13 expression vector (pCMV-A13) re- et al., 2006). HOXD13. sulted in increased luciferase expression (3.75- A link between Hox proteins and Of interest, our chromatin immuno- fold) only in the forward orientation compared the expression of ephrin ligands and precipitation (ChIP) analysis in the with transfections with a control pCMV vector. Eph receptors has also been estab- GT did not detect HOXA13 binding to Identical transfections using the conserved EphA7 region bound by HOXA13 resulted in a lished. In the hindbrain, the loss of the site described by Salsi and Zap- 2.75- to 4-fold increase in relative luciferase Hoxa1 and Hoxb1 directly affect the pavigna (2006) in the limb, although expression, suggesting that this gene-regula- expression of EphA2 in rhombomere we did verify that the A13 DNA bind- tory region may be functioning as a HOXA13- 4, whereas in the developing micro- ing domain peptide could bind this directed enhancer of EphA7 expression. Lucif- vasculature, antisense oligos directed site in vitro. One possible explanation erase activity was normalized for transfection efficiency using a Renilla Luciferase control toward Hoxb3 also caused reductions for this difference in binding site uti- plasmid in all cotransfection assays. Bars rep- in ephrin A1 in the vascular endothe- lization is that HOX protein cofactors, resent the standard deviation of results derived lia (Chen and Ruley, 1998; Myers et such as the TALE class of DNA bind- from three independent assays. al., 2000). Finally the loss of HOXA13 ing proteins, are also differentially ex- function has also been linked to reduc- pressed and can influence which DNA tions in EphA7 in the developing limb sequences are used by a particular DISCUSSION mesenchyme and EphA7 and EphA4 HOX protein (reviewed by Moens and In humans, the loss of HOXA13 func- in the umbilical artery endothelia Selleri, 2006; Villaescusa et al., 2004; tion causes hand-foot-genital syn- (Stadler et al., 2001). Williams et al., 2005; Erickson et al., drome (HFGS), an autosomal domi- While the characterization of the 2006). For HOXA13, interactions with nant disorder that profoundly affects combinatorial functions of EphA6 and its cofactor, MEIS-1B, appears to be the development of the external geni- EphA7 in the GT vasculature await specific to the genitourinary region talia (hypospadias), uterus, vagina, the production of a conditional EphA6 and may influence which of the DNA cervix, bladder, and ureter (Stern et allele, the present body of evidence sequences are regulated by HOXA13 al., 1970; Mortlock and Innis, 1997; linking perturbations in Eph receptor in this tissue (Williams et al., 2005). Warot et al., 1997; Morgan et al., signaling to angiogenic defects sug- Alternatively, DNA accessibility may 2003; Stadler, 2003). An important gests that the combined reduction of also vary in limb versus genitourinary step toward understanding the devel- EphA6 and EphA7 in the GT vascula- chromatin, which could also account opmental basis for the genitourinary ture may cause a change in cell adhe- for the differential binding of HOXA13 defects associated with HFGS is the sion, which under vascular load could to gene-regulatory sequences in a tis- identification of the genes directly reg- affect the overall diameter of the GT sue-specific manner (Vashee et al., ulated by HOXA13. In this report, we vessels. Functionally, the loss of 1998; Kodadek, 1998). identify EphA6 and EphA7 as direct EphA7 in the embryo causes defects in Of interest, a common theme transcriptional targets of HOXA13 in cell adhesion and anencephaly, al- emerging from studies of HOXA13-de- the GT. The in vivo association of though no defects in the developing ficient genitourinary tissues is that HOXA13 with gene-regulatory ele- vasculature were reported (Holmberg epithelial lineages appear to be af- ments present in EphA6 and EphA7, et al., 2000). Furthermore, the loss of fected in a similar manner. Indeed, as well as their in vitro utilization by HOXA13 function also affects the the urethral plate epithelium (which HOXA13 to direct gene expression, morphology of the umbilical artery en- serves as a signaling center for the provides strong evidence that dothelia, which exhibit reduced levels GT), the vascular endothelia of the HOXA13 can regulate the tissue-spe- of EphA4 and EphA7 expression (Sta- umbilical arteries, and the epithelia cific expression of these receptor ty- dler et al., 2001). lining the developing bladder and ure- rosine kinases. In the limb, HOXA13 and HOXD13 ter (data not shown) all exhibit Studies examining the functional function in a redundant manner to changes in cell morphology and strat- HOXA13 REGULATES EPHA6,7 EXPRESSION IN THE GT 957 ification with the loss of functional lele, 5Ј-CAGGAGTGGCCCGGGAA-3Ј, 37°C for 30 min with occasional HOXA13 (present study, Perriton et 5Ј-CATTACACTTCCAGACCTGGG- shaking. After collagenase treat- al., 2002; Morgan et al., 2003). While AC-3Ј and 40 cycles of 94°C (30 sec), ment, the tissues were placed in di- alterations in epithelial cell morphol- 54°C (30 sec), 72°C (30 sec). All proce- gestion medium (0.1% trypsin/ethyl- ogy and stratification could reflect a dures using mice were done in accor- enediaminetetraacetic acid [EDTA, loss in cellular identity, this possibil- dance with an approved institutional Gibco], 0.2% Collagenase IV, in ity is unlikely as sonic hedgehog ex- animal protocol (A729 to H.S.S.). phosphate buffered saline [PBS]) for pression is maintained in the mutant 15 min at 37°C, using gentle pipet- urethral plate epithelium, and PE- GT Histology and ting every 5 min to dissociate the CAM-1 expression is also maintained tissue. Cell flow-through was col- Immunohistochemistry in the mutant vascular endothelia lected in 15-ml tubes with 10 ml of (this study; Morgan et al., 2003). Paraffin (Paraplast Plus, Fisher) -em- quenching buffer (15% fetal bovine Thus, altered epithelial morphology bedded Hoxa13 wild-type and ho- serum [#26140-087, Gibco] and 0.1% and stratification in the mutant GT mozygous mutant embryos (E12.5– bovine serum albumin in D-PBS vasculature is more likely to reflect a E15.5) were sectioned at 7-␮m [BP1605-100, Fisher]). Finally, to change in key matrix or cytoskeletal intervals and placed sequentially onto collect any residual cells, the components. These factors are neces- Superfrost plus slides (Fisher) and Netwell baskets were rinsed with sary for a tissue-specific morphologi- stained with hematoxylin and eosin as 0.1% BSA/PBS and added to the cell cal state, and it has been recently described by Stadler and Solursh flow-through. Cells were stored on demonstrated that they are regulated (1994). Sections containing the GT ice for 5 min and spun at 3,000 rpm by Eph–ephrin signaling (reviewed by and its associated vasculature were for 5 min, followed by an additional Cheng et al., 2002; Harbott and No- photographed using a Leica DMLB2 quenching wash and spin. Dyna- bes, 2005; Hunter et al., 2006; Mar- microscope and a Q Imaging Digital beads were coated with PECAM-1 ston and Goldstein, 2006). camera. antibody (MEC 13.3, #553369 BD E13.5 Hoxa13 heterozygous and ho- Pharmingen) as described by the mozygous mutant embryos were em- manufacturer (Dynal). Cells (2.5 ϫ EXPERIMENTAL bedded in OCT (Tissue Tek) and sec- 106) were combined with the Dyna- PROCEDURES tioned as previously described (Morgan beads–antibody complex (three et al., 2003). Antibodies specific for times more beads than cells), and Mouse Strains EphA6 (R&D Systems MAB6071), the mixture was incubated for 1 hr Hoxa13-GFP mutant embryos were EphA7 (R&D Systems MAB1495), and at 4°C on a rotating platform. The derived from heterozygous inter- PECAM-1 (BD Pharmingen 553708) bead–antibody–cell complexes were crosses as described (Stadler et al., were used at dilutions of 1:200 and in- isolated with a magnet, and the re- 2001; Morgan et al., 2003). The mu- cubated on the sections at 4°C over- maining PECAM-negative cells were tant HOXA13 allele encodes a fusion night. Secondary antibodies labeled collected as a control. The cells were protein of HOXA13 and GFP where with Cy5 were used as described by the gently washed with 0.1% BSA/PBS the last 34 amino acids of HOXA13, manufacturer (Jackson Immunologi- and collected for RNA extraction us- encoding the DNA contacting third cal). Imaging was performed on a Bio- ing RNA Stat-60 (CS-110, Tel-Test). helix (ISATTNLSERQVTIWFQNR- Rad MRC1024 confocal microscope us- GT mesenchyme RNA was isolated RVKEKKVINKLKTTS), is removed ing Kalman filtering. Identical laser in a similar manner using dissected and replaced with the enhanced GFP level, iris, and black level settings were GTs from E13.5 embryos. RNA qual- protein (Clontech). The nuclear local- used for all samples. ity was analyzed by agarose gel elec- ization, turnover, and tissue-specific trophoresis and ultraviolet spectros- expression of the HOXA13-GFP pro- GT Endothelial Cell copy. One microgram of RNA was tein appeared similar to the wild-type used for cDNA synthesis using the Isolation and RT-PCR protein (Stadler et al., 2001). Timed Superscript First-Strand Synthesis matings were used to establish embry- The distal half of the E13.5 GTs were system (Invitrogen). onic gestational age in embryonic isolated by microdissection in 1ϫ days, where E0.5 represents the first phosphate buffered saline. E13.5 Semiquantitative RT-PCR day of vaginal plug detection. EphA7 embryos were chosen because they homozygous mutant embryos were strongly express Hoxa13, EphA6, cDNAs derived from distal GT RNA or produced by intercrosses of EphA7 and EphA7 within the GT vascula- endothelial cell RNA was used for semi- heterozygous mutant mice, kindly ture as shown by immunohistochem- quantitative RT-PCR to detect the ex- provided by Jonas Friseń (Karolinska istry and in situ hybridization. Ap- pression levels of EphA6, EphA7, Gapdh, Institute, Stockholm, Sweden). EphA7 proximately 5–8 GTs of identical Pecam-1, and Hoxa13 in Hoxa13 wild- embryo genotypes were determined by Hoxa13 genotype were combined to type and homozygous mutants. For RT- PCR of yolk sac-derived DNA using gain adequate amounts of endothe- PCR, each cDNA template was diluted the following primers: EphA7 mutant lial cells and isolated RNA. The tis- 1:1 with water. The following gene- allele, 5Ј-CTAAGGTCCTATTTTGC- sues were pooled in individual specific primer sequences were used: CTG-3Ј,5Ј-CATTACACTTCCAGAC- Netwells (Costar) and treated with EphA6-For: 5Ј-GAGAGACCGTACTG- CTGGGAC-3Ј; EphA7 wild-type al- 0.2% Collagenase Type IV (Gibco) at GGAAATG-3Ј; EphA6-Rev: 5Ј-GCCTGT- 958 SHAUT ET AL.

GGTTTCTCTCCTTC-3Ј (NM007938, used to assess the binding affinity value, the molecular weight of the com- bp2901-3042); EphA7-For: 5Ј- CTCT- for the EphA6 and EphA7 promoter plex was determined to be 16,900 dal- TCGCTGCTGTTAGCAT-3Ј; EphA7-Rev: sequences as described (Knosp et al., tons, indicating that the HOXA13 DNA 5Ј-GTGATGACTCCATTGGGATG-3Ј 2004). Interspecies comparisons of binding domain peptide associates with (BC026153, bp1433-1566); Pecam1-For: the EphA6 and EphA7 regions con- DNA in a 1:1 stoichiometric ratio, con- 5Ј-CCAGTGCAGAGCGGATAAT-3Ј; and taining the HOXA13 binding sites firming monomeric binding to the Pecam1-Rev: 5Ј-GCACCGAAGTAC- were performed using the UCSC Ge- EphA6 and EphA7 gene-regulatory ele- CATTTCAC-3Ј (NM008816, bp1487– nome Browser (Kent et al., 2002). ments. 1634); and Hoxa13-For: 5Ј-CTGGAA- CGGCCAAATGTACT-3Ј; Hoxa13- Oligomerization Analysis of Ј ChIP Assays Rev: 5 -TATAGGAGCTGGCGTCTGA- the HOXA13 DNA Binding A-3Ј (NM008264, bp952–1058). ChIP was performed using a HOXA13 Peptide and DNA Gapdh primers were previously de- antibody and whole GTs dissected scribed (Shou et al., 2005). All RT- Complexes of the A13 DNA binding do- from E12.5 embryos. Multiple at- PCR primer pairs were designed to main peptide and DNA were assessed tempts to isolate sufficient quantities flank an intronic sequence to distin- for their oligomerization state using of endothelial cells for ChIP were not guish any PCR products derived from sedimentation equilibrium ultracentrif- successful as each GT yielded less genomic DNA contamination. Each ugation on a Beckman Coulter XL-1A than 1,000 endothelial cells, whereas RT-PCR analysis was tested from at Protein Analysis System as described 106 cells are required to produce suf- least three independent RNA isolates. (Huffman et al., 2001). A self-annealing ficient chromatin quantities for immu- fluorescein-labeled oligonucleotide, 5Ј- noprecipitation (Lavrrar and Farn- EphA6 and EphA7 Promoter 6-FAM-CCCATAAACCCCCCCGGTT- ham, 2004). Therefore, all PCR TATGGG-3Ј (5 ␮M) was combined with amplification of the HOXA13 bound Analysis the A13 DNA binding peptide (6 ␮Mto DNA regions were derived from whole Sequence analysis of the EphA6 pro- 10 ␮M) in a buffer containing 80 mM GT chromatin isolates. The 300-bp Ϫ moter (Ensembl: ENSMUST000000- KCl, 10 mM MgCl2, 0.2 mM EDTA, 1 fragments encompassing bases 3000 68860) identified a single region (Ϫ2410 mM dithiothreitol, and 20 mM Tris.HCl to ϩ1 of the EphA6 (Ensembl: ENS- to Ϫ2067) containing multiple T-A-A pH 7.8. The samples were centrifuged MUST00000068860) and EphA7 (En- motifs we previously identified as being at 4°C at 26,000 or 32,000 rpm in an sembl: ENSMUSG00000028289) pro- bound by HOXA13 (Knosp et al., 2004). AN-60Ti rotor equipped with 12-mm moters were examined for in vivo Analysis of the EphA7 locus (Ensembl: path length Epon double sector cells. association with HOXA13. The GTs ENSMUSG00000028289) identified the The absorbance was monitored at 480 were dissected in PBS containing 15 HOXA13 binding site characterized by nm as a function of radial distance. Mo- ␮l/ml protease inhibitor cocktail (PIC, Salsi and Zappavigna (2006) as well as lecular weights of the DNA and DNA ϩ Sigma). a novel region (Ϫ839 to Ϫ514) contain- protein complexes were calculated from Tissues were fixed in 1% formalde- ing several T-A-A motifs that we ex- a nonlinear least squares fit of the ab- hyde/PBS and rocked at room temper- amined in this report for HOXA13 sorbance data using the software sup- ature for 10 min. Protein–DNA cross- binding and gene regulation. PCR plied with the Model XL-1A system. A linking was stopped by the addition of amplification of the EphA6 and density value for the solvent of 1.005 glycine to a final concentration of EphA7 promoter regions was per- g/ml was used for the calculations. The 0.125 M for 5 min. Next, the samples formed using the following primers: partial specific volumes of the peptide– were centrifuged at low speed, and the EphA6P1F 5Ј-GATAGGCAGAAT- DNA complexes were estimated by tak- pellet was washed once with cold PBS GCCAGGTG-3Ј; EphA6P1R 5Ј- ing the mass-weighted average of the containing PIC and centrifuged at low GGAGCAAGGAAAGCTCAGAA-3Ј; partial specific volumes for the free oli- speed. The pellet was resuspended in EphA7P1F 5Ј-TGCCTCTCGAGTT- gonucleotide (8,100 daltons) and free 100 ␮L cell lysis buffer (5 mM PIPES, ACAGAACAG-3Ј; EphA7P1R 5Ј- peptide (8,137 daltons). Using the addi- pH 8.0/85 mM KCl/0.5% NP40) plus GGGAGCACTTGGCTTTTAGC-3Ј. tivity method of Cohn and Edsall (1943) PIC and incubated on ice for 10 min. HOXA13 binding to the PCR-ampli- the partial specific volume of the free Next, the lysate suspension was mi- fied EphA6 and EphA7 promoter peptide was estimated to be 0.745 ml/g. crocentrifuged at 5,000 rpm for 5 min fragments was determined using an The partial specific volume of the free at 4°C, followed by resuspension in 50 EMSA as previously described oligonucleotide was determined using a ␮l of nuclear lysis buffer (50 m Tris- (Knosp et al., 2004). Briefly, the am- separate sedimentation equilibrium HCl, pH 8.1, 10 mM EDTA, 1% so- plified PCR products were radio- analysis in identical buffers without dium dodecyl sulfate) plus PIC and labeled with T4 Polynucleotide ki- peptide. The partial specific volume of incubation for 10 min on ice. The lysed nase and assessed for HOXA13 unbound oligonucleotide was deter- nuclei were sonicated for 20 periods of binding by incubation with a mined to be 0.512 ml/g. Using the mass- 30 sec ON and 1 min OFF at 4°C using HOXA13 DNA binding domain pep- weighted partial specific volume aver- a Bioruptor (Cosmo Bio) to produce tide (A13) followed by nondenatur- ages for the peptide and DNA, a partial sheared chromatin of an average ing acrylamide gel electrophoresis. specific volume of 0.629 g/ml was calcu- length of 200–1,000 bp. The sheared Competitor DNA consisting of the lated for the samples containing peptide chromatin was microcentrifuged at identical unlabeled PCR product was and DNA in a 1:1 ratio. Using this 13,000 rpm for 10 min at 4°C and the HOXA13 REGULATES EPHA6,7 EXPRESSION IN THE GT 959 supernatant was transferred to a new turer (Promega) and plotted using Sig- EphB2 controls urorectal development. tube. maPlot 9.0 (Systat). Dev Biol 271:272–290. ChIP was performed using a ChIP Egea J, Nissen UV, Dufour A, Sahin M, Greer P, Kullander K, Mrsic-Flogel TD, Assay Kit as described by the manu- Greenberg ME, Kiehn O, Vanderhae- facturer (Upstate Biotechnologies/ ACKNOWLEDGMENTS ghen P, Klein R. 2005. Regulation of Millipore). Each chromatin superna- The authors thank Eric Steele, Hans EphA 4 kinase activity is required for a tant was precleared with 40 ␮lof Peter Ba¨chinger, and the Shriners subset of axon guidance decisions sug- salmon sperm DNA/Protein A Agarose Hospital for Children Analytical Core gesting a key role for receptor clustering in Eph function. Neuron 47:515–528. (Upstate Biotechnology). The chroma- for their assistance with the sedimen- Eichmann A, Le Noble F, Autiero M, Car- tin samples were incubated with the tation equilibrium analysis. W.M.K. meliet P. 2005a. Guidance of vascular HOXA13 antibody or IgG control an- was funded by Predoctoral Fellow- and neural network formation. Curr tibody on a rotating platform at 4°C ships from the National Institutes of Opin Neurobiol 15:108–115. for 3 hr. Washes, DNA elution, and Health, and C.A.S. was funded by the Eichmann A, Yuan L, Moyon D, Lenoble F, Pardanaud L, Breant C. 2005b. Vascular reverse cross-linking were performed American Heart Association. development: from precursor cells to as described in the Upstate ChIP As- branched arterial and venous networks. say Kit. Samples were ethanol precip- Int J Dev Biol 49:259–267. itated, resuspended in 100 ␮lofTE, REFERENCES Erickson T, Scholpp S, Brand M, Moens and DNA purified using the Qiaquick CB, Jan Waskiewicz A. 2006. Pbx pro- Adams RH, Diella F, Hennig S, Helm- teins cooperate with Engrailed to pat- PCR Purification kit (Qiagen). bacher F, Deutsch U, Klein R. 2001. The tern the midbrain-hindbrain and dience- The eluted DNA from the HOXA13 cytoplasmic domain of the ligand eph- phalic-mesencephalic boundaries. Dev antibody, control IgG, or no antibody rinB2 is required for vascular morpho- Biol 301:504–517. samples were assessed for the pres- genesis but not cranial neural crest mi- Gallentine ML, Morey AF, Thompson IM gration. Cell 104:57–69. ence of the EphA6 and EphA7 pro- Jr. 2001. Hypospadias: a contemporary Baskin LS. 2000. Hypospadias and ure- epidemiologic assessment. Urology 57: moter DNA using PCR. Primers used thral development. J Urol 163:951–956. 788–790. to amplify the ChIP-positive regions Baskin LS, Erol A, Li YW, Cunha GR. Gerety SS, Wang HU, Chen ZF, Anderson were identical to those used to amplify 1998. Anatomical studies of hypospa- DJ. 1999. Symmetrical mutant pheno- the candidate binding site regions de- dias. J Urol 160:1108–1115; discussion types of the receptor EphB4 and its spe- 1137. scribed above. cific transmembrane ligand ephrin-B2 in Brown A, Yates PA, Burrola P, Ortuno D, cardiovascular development. Mol Cell 4: Vaidya A, Jessell TM, Pfaff SL, O’Leary 403–414. DD, Lemke G. 2000. Topographic map- Epha6 and Epha7 Luciferase Giwercman A, Carlsen E, Keiding N, Skak- ping from the retina to the midbrain is kebaek NE. 1993. Evidence for increas- Assays controlled by relative but not absolute ing incidence of abnormalities of the hu- levels of EphA receptor signaling. Cell NG108-15 cells (ATCC#HB-12317) man testis: a review. Environ Health 102:77–88. Perspect 101(Suppl 2):65–71. Chan J, Mably JD, Serluca FC, Chen JN, were maintained and transfected as Harbott LK, Nobes CD. 2005. A key role for Goldstein NB, Thomas MC, Cleary JA, previously described (Knosp et al., Abl family kinases in EphA receptor-me- Brennan C, Fishman MC, Roberts TM. diated growth cone collapse. Mol Cell 2004). Transfections were performed in 2001. Morphogenesis of prechordal plate Neurosci 30:1–11. 12-well plates (Costar) using 1.5 ␮gofa and notochord requires intact Eph/eph- pGL4 plasmid containing either the for- rin B signaling. Dev Biol 234:470–482. Hinck L. 2004. The versatile roles of “axon guidance” cues in tissue morphogenesis. Chen J, Ruley HE. 1998. An enhancer ele- ward (PGL4.1-EphA6F or PGL4.1- Dev Cell 7:783–793. EphA7F) or reverse (PGL4.1EphA6Ror ment in the EphA2 (Eck) gene sufficient for rhombomere-specific expression is ac- Holmberg J, Clarke DL, Frisen J. 2000. PGL4.1-EphA7R) orientations of the tivated by HOXA1 and HOXB1 ho- Regulation of repulsion versus adhesion EphA6 or EphA7 ChIP-positive regions meobox proteins. J Biol Chem 273:24670– by different splice forms of an Eph recep- as well as 0.25 ␮g of the normalization 24675. tor. Nature 408:203–206. ␮ Cheng N, Brantley DM, Chen J. 2002. The Huffman JL, Mokashi A, Ba¨chinger HP, vector, pRL-CMV Renilla, and 0.5 g Brennan RG. 2001. The basic helix-loop- pCMV-Hoxa13-HA or an empty pCMV ephrins and Eph receptors in angiogenesis. Cytokine Growth Factor Rev 13:75– helix domain of the aryl hydrocarbon re- control plasmid per well. Cell lysates 85. ceptor nuclear transporter (ARNT) can were processed to detect luciferase ac- Cohn EJ, Edsall JT, editors. 1943. Proteins oligomerize and bind E-box DNA specif- tivity using the Dual-Glo Luciferase As- amino acids and peptides. New York: Re- ically. J Biol Chem 276:40537–40544. Hunter SG, Zhuang G, Brantley-Sieders D, say System (Promega) in OptiPlate-96F inhold Publishing Corp. p 370. Cooke JE, Kemp HA, Moens CB. 2005. Swat W, Cowan CW, Chen J. 2006. Es- black plates (Packard) as described EphA4 is required for cell adhesion and sential role of Vav family guanine nucle- (Knosp et al., 2004). Luciferase activity rhombomere-boundary formation in the otide exchange factors in EphA receptor- was detected using a Packard Fusion zebrafish. Curr Biol 15:536–542. mediated angiogenesis. Mol Cell Biol 26: Microplate Analyzer (Perkin Elmer). Davy A, Soriano P. 2005. Ephrin signaling 4830–4842. Karolchik D, Baertsch R, Diekhans M, Sample wells were read three times for in vivo: look both ways. Dev Dyn 232:1– 10. Furey TS, Hinrichs A, Lu YT, Roskin 1 sec each and averaged. Three repli- Davy A, Aubin J, Soriano P. 2004. Eph- KM, Schwartz M, Sugnet CW, Thomas cates of each transfection were per- rin-B1 forward and reverse signaling are DJ, Weber RJ, Haussler D, Kent WJ. formed, and each transfection assay required during mouse development. 2003. The UCSC Genome Browser Data- was repeated three times. Results were Genes Dev 18:572–583. base. Nucleic Acids Res 31:51–54. Dravis C, Yokoyama N, Chumley MJ, Kent WJ, Sugnet CW, Furey TS, Roskin normalized for transfection efficiency Cowan CA, Silvany RE, Shay J, Baker KM, Pringle TH, Zahler AM, Haussler D. using relative Renilla luciferase expres- LA, Henkemeyer M. 2004. Bidirectional 2002. The human genome browser at sion levels as described by the manufac- signaling mediated by ephrin-B2 and UCSC. Genome Res 12:996–1006. 960 SHAUT ET AL.

Kent WJ, Hsu F, Karolchik D, Kuhn RM, Mortlock DP, Innis JW. 1997. Mutation of meobox gene expression in the develop- Clawson H, Trumbower H, Haussler D. HOXA13 in hand-foot-genital syndrome. ing chick embryo. Dev Biol 161:251–262. 2005. Exploring relationships and min- Nat Genet 15:179–180. Stadler HS, Higgins KM, Capecchi MR. ing data with the UCSC Gene Sorter. Myers C, Charboneau A, Boudreau N. 2001. Loss of Eph-receptor expression Genome Res 15:737–741. 2000. Homeobox B3 promotes capillary correlates with loss of cell adhesion and Klagsbrun M, Eichmann A. 2005. A role for morphogenesis and angiogenesis. J Cell chondrogenic capacity in Hoxa13 mutant axon guidance receptors and ligands in Biol 148:343–351. limbs. Development 128:4177–4188. blood vessel development and tumor an- Ogawa K, Pasqualini R, Lindberg RA, Kain Stern AM, Gall JC Jr, Perry BL, Stimson giogenesis. Cytokine Growth Factor Rev R, Freeman AL, Pasquale EB. 2000. The CW, Weitkamp LR, Poznanski AK. 1970. 16:535–548. ephrin-A1 ligand and its receptor, The hand-food-uterus syndrome: a new Knosp WM, Scott V, Ba¨chinger HP, Stadler EphA2, are expressed during tumor neo- hereditary disorder characterized by HS. 2004. HOXA13 regulates the expres- vascularization. Oncogene 19:6043– hand and foot dysplasia, dermatoglyphic sion of bone morphogenetic proteins 2 6052. abnormalities, and partial duplication of the female genital tract. J Pediatr 77: and 7 to control distal limb morphogen- Pasquale EB. 2005. Eph receptor signal- 109–116. esis. Development 131:4581–4592. ling casts a wide net on cell behaviour. Nat Rev Mol Cell Biol 6:462–475. Svensson H, Reychman M, Troeng T, Kodadek T. 1998. Mechanistic parallels be- Aberg M. 1997. Staged reconstruction of tween DNA replication, recombination Paulozzi LJ, Erickson JD, Jackson RJ. 1997. Hypospadias trends in two US sur- hypospadias with chordee: outcome and and transcription. Trends Biochem Sci costs. Scand J Plast Reconstr Surg Hand 23:79–83. veillance systems. Pediatrics 100:831– 834. Surg 31:51–55. Lavrrar JL, Farnham PJ. 2004. The use of Vashee S, Melcher K, Ding WV, Johnston Perriton CL, Powles N, Chiang C, Ma- transient chromatin immunoprecipita- SA, Kodadek T. 1998. Evidence for two conochie MK, Cohn MJ. 2002. Sonic tion assays to test models for E2F1-spe- modes of cooperative DNA binding in hedgehog signaling from the urethral ep- cific transcriptional activation. J Biol vivo that do not involve direct protein- ithelium controls external genital devel- Chem 279:46343–46349. protein interactions. Curr Biol 8:452– opment. Dev Biol 247:26–46. Marquardt T, Shirasaki R, Ghosh S, An- 458. Salsi V, Zappavigna V. 2006. Hoxd13 and drews SE, Carter N, Hunter T, Pfaff SL. Villaescusa JC, Verrotti AC, Ferretti E, Hoxa13 directly control the expression of 2005. Coexpressed EphA receptors and Farookhi R, Blasi F. 2004. Expression of the EphA7 Ephrin tyrosine kinase recep- ephrin-A ligands mediate opposing ac- Hox cofactor genes during mouse ovar- tor in developing limbs. J Biol Chem 281: tions on growth cone navigation from ian follicular development and oocyte 1992–1999. distinct membrane domains. Cell 121: maturation. Gene 330:1–7. Scott V, Morgan EA, Stadler HS. 2005. 127–139. Wahl S, Barth H, Ciossek T, Aktories K, Genitourinary functions of Hoxa13 and Marston DJ, Goldstein B. 2006. Actin- Mueller BK. 2000. Ephrin-A5 induces Hoxd13. J Biochem (Tokyo) 137:671– collapse of growth cones by activating based forces driving embryonic morpho- 676. genesis in Caenorhabditis elegans. Curr Rho and Rho kinase. J Cell Biol 149:263– Shamah SM, Lin MZ, Goldberg JL, Es- Opin Genet Dev 16:392–398. 270. trach S, Sahin M, Hu L, Bazalakova M, McBride JL, Ruiz JC. 1998. Ephrin-A1 is Wang HU, Chen ZF, Anderson DJ. 1998. Neve RL, Corfas G, Debant A, Greenberg Molecular distinction and angiogenic in- expressed at sites of vascular develop- ME. 2001. EphA receptors regulate ment in the mouse. Mech Dev 77:201– teraction between embryonic arteries growth cone dynamics through the novel and veins revealed by ephrin-B2 and its 204. guanine nucleotide exchange factor receptor Eph-B4. Cell 93:741–753. McCabe CD, Innis JW. 2005. A genomic ephexin. Cell 105:233–244. Warot X, Fromental-Ramain C, Fraulob V, approach to the identification and char- Shou S, Scott V, Reed C, Hitzemann R, Chambon P, Dolle P. 1997. Gene dosage- acterization of HOXA13 functional bind- Stadler HS. 2005. Transcriptome analy- dependent effects of the Hoxa-13 and ing elements. Nucleic Acids Res 33:6782– sis of the murine forelimb and hindlimb Hoxd-13 mutations on morphogenesis of 6794. autopod. Dev Dyn 234:74–89. the terminal parts of the digestive and Moens CB, Selleri L. 2006. Hox cofactors in Stadler HS. 2003. Modelling genitourinary urogenitaltracts.Development124:4781– vertebrate development. Dev Biol 291: defects in mice: an emerging genetic and 4791. 193–206. developmental system. Nat Rev Genet Williams TM, Williams ME, Innis JW. Morgan EA, Nguyen SB, Scott V, Stadler 4:478–482. 2005. Range of HOX/TALE superclass HS. 2003. Loss of Bmp7 and Fgf8 signal- Stadler HS, Solursh M. 1994. Character- associations and protein domain require- ing in Hoxa13-mutant mice causes hypo- ization of the homeobox-containing gene ments for HOXA13:MEIS interaction. spadia. Development 130:3095–3109. GH6 identifies novel regions of ho- Dev Biol 277:457–471.