The BMP Family Member Gdf7 Is Required for Seminal Vesicle

Developmental Biology 234, 138–150 (2001) doi:10.1006/dbio.2001.0244, available online at http://www.idealibrary.com on

View metadata, citation and similar papers at core.ac.uk brought to you by CORE The BMP Family Member Gdf7 Is Required for provided by Elsevier - Publisher Connector Seminal Vesicle Growth, Branching Morphogenesis, and Cytodifferentiation

Stephen Settle,*,1 Paul Marker,† Kyle Gurley,* Abhishek Sinha,*,2 Abigail Thacker,* Yuzhuo Wang,† Kay Higgins,*,3 Gerald Cunha,† and David M. Kingsley*,4 *Department of Developmental Biology and Howard Hughes Medical Institute, Beckman Center B300, Stanford University School of Medicine, Stanford, California 94305-5427; and †Department of Anatomy, University of California San Francisco, 513 Parnassus, San Francisco, California 94143

Epithelial–mesenchymal interactions play an important role in the development of many different organs and tissues. The secretory glands of the male reproductive system, including the prostate and seminal vesicles, are derived from epithelial precursors. Signals from the underlying mesenchyme are required for normal growth, branching, and differentiation of the seminal vesicle epithelium. Here, we show that a member of the BMP family, Gdf7, is required for normal seminal vesicle development. Expression and tissue recombination experiments suggest that Gdf7 is a mesenchymal signal that acts in a paracrine fashion to control the differentiation of the seminal vesicle epithelium. © 2001 Academic Press

INTRODUCTION mouse embryonic development, a portion of this tube grows and curves to form the seminal vesicle precursor. Epithelial–mesenchymal interactions are required for Shortly after birth, the epithelium of the seminal vesicle normal development of many different organs and tissues precursor undergoes dramatic growth and branching (Lung (Johnson and Tabin, 1997; Thesleff and Sharpe, 1997). and Cunha, 1981; Shima et al., 1990). By day 15 after birth, Previous studies have shown that epithelial–mesenchymal the epithelium of the gland begins to elaborate a character- interactions are necessary for the development of the male istic set of secretory proteins. In adult males, seminal accessory sex organs, which are a group of glands that vesicle secretions are mixed with sperm at ejaculation. The produce secretory products necessary for male reproductive major secretory proteins of the seminal vesicles comprise function (Cunha et al., 1992). These glands include the the bulk of the copulatory plug that forms after mating seminal vesicles, which are derived from the Wolffian duct, (Bradshaw and Wolfe, 1977; Fawell et al., 1987). and the prostate, which is derived from the urogenital The Wolffian duct is also the precursor for several other sinus. male reproductive structures in addition to the seminal Early in development, the Wolffian duct is a simple vesicles, including the epididymis, vas deferens, and amp- epithelial tube surrounded by mesenchyme. At day 17 of ullary gland (Lipschutz et al., 1999). All of these Wolffian duct derivatives can be distinguished by their tissue- specific morphology, histodifferentiation, and tissue- 1 Current address: CC2218 Medical Center North, Vanderbilt specific secretory proteins (Brooks and Higgins, 1980; University, Nashville, TN 37232. Fawell et al., 1986; Taragnat et al., 1988). It is thought that 2 Current address: Yale University School of Medicine, 333 inductive influences from surrounding mesenchyme act to Cedar Street, New Haven, CT 06510. 3 Current address: Howard Hughes Medical Institute, University specify the characteristic patterns of epithelial morphogen- of Utah, 15 N. 2030 E., Salt Lake City, UT 84112. esis and differentiation along the length of the Wolffian 4 To whom correspondence should be addressed. Fax: (650) 725- duct. For example, tissue recombination experiments have 7739. E-mail: [email protected]. shown that seminal vesicle mesenchyme is capable of

0012-1606/01 $35.00 Copyright © 2001 by Academic Press 138 All rights of reproduction in any form reserved. Gdf7 and Normal Seminal Vesicle Development 139 eliciting seminal vesicle differentiation from epithelium generated by cutting a 2.1-kb XbaI–KpnI subclone with XbaI and that would normally form epididymis, vas deferens, or EcoRI and isolating the 300-bp end fragment to label as a probe. The ureter (Higgins et al., 1989b; Lipschutz et al., 1996) In the neomycin resistance cassette introduces a SpeI site. Digestion of absence of mesenchyme, seminal vesicle epithelial growth, genomic DNA from targeted clones with SpeI results in a wild-type branching, and differentiation are inhibited (Cunha, 1972; (WT) band of 17 kb and a targeted band of 3 kb (Fig. 1). Higgins et al., 1989a). Thus, signals from seminal vesicle mesenchyme can instructively program Wolffian-derived Generation of Mutant Mice epithelium to adopt seminal vesicle morphology and cytodifferentiation. Multiple independently derived ES clones were injected into Two members of the fibroblast growth factor (FGF) fam- C57BL/6-J host blastocysts. Chimeric animals were backcrossed to ily, FGF7 and FGF10, previously have been implicated as C57BL/6-J and transmission to the germline monitored by coat color. Southern blots were performed to identify animals that had important mesenchymal signals in seminal vesicle devel- transmitted the targeted allele to the germline (Fig. 1). Mutant opment. FGF7 is produced by seminal vesicle mesenchyme animals were generated by intercrossing heterozygous carriers, and and can stimulate seminal vesicle growth and branching in all phenotypic analysis was performed on the C57BL/6-J ϫ 129SV/J vitro (Alarid et al., 1994). FGF10 is expressed in the mesen- background. Two independent lines that transmitted to the germ- chyme of the developing seminal vesicle (Thomson and line were analyzed initially. Both produced an identical spectrum Cunha, 1999), and FGF10 knockout mice lack seminal of phenotypes and all additional analysis was performed on a single vesicles (A. Donjacour and G.C., unpublished observations). line. The transforming growth factor ␤ (TGF␤)/Bone morphogenetic protein (BMP) superfamily has also previously been implicated in epithelial–mesenchymal interactions during Genotyping the formation of many different organs and tissues (Hogan, Genotyping of animals was performed by using two sets of PCR 1996). Members of the BMP family are expressed in mesen- primers: one set that amplifies a 148-bp fragment from the mature chyme of the developing gut, tooth, and kidney and can signaling region of Gdf7 that is removed in the knockout influence the development of the associated epithelium (5Ј-CCATCATTCAGACGCTGCTCAAC-3Ј and 5Ј-TCCACC- (Godin et al., 1999; Immergluck et al., 1990; Panganiban et ACCATGTCTTCGTACTG-3Ј), and one set that amplifies a 310-bp al., 1990; Reuter et al., 1990; Thesleff and Sharpe, 1997). fragment from the neomycin resistance cassette introduced during Ј Ј Ј TGF␤1 is expressed in the developing prostatic mesen- the targeting (5 -TGGAGAGGCTATTCGGCTATGAC-3 and 5 - TACTTTCTCGGCAGGAGCAAGG-3Ј). PCR was performed using chyme (Timme et al., 1994), and growth of the developing 35 cycles of 94°C for 30 s, 63°C for 1 min, and 72°C for 30 s (Fig. 1). prostate can be inhibited by exogenous TGF␤1 in organ culture (Itoh et al., 1998). Here, we report a combination of genetic, expression, and tissue recombination experiments In Situ Hybridization that identify a highly conserved BMP family member called Growth/differentiation factor 7 (Gdf7) (Storm et al., 1994; Antisense riboprobes were generated for Gdf7, Bmpr1A, and Bmpr1B in vitro Davidson et al., 1999) as an essential mesenchymal signal by transcription as previously described (Storm and Kingsley, 1996). The probe for Gdf7 was generated by subclon- required for normal growth, branching, and functional dif- ing a 290-bp PCR fragment amplified from the 3ЈUTR of the gene ferentiation of seminal vesicle epithelium. using the primer set 5Ј-CGCCCAGGACCCTAGCTC-3Ј and 5Ј- ACCGCCTCCAACTAAAAGG-3Ј and cloned into the PCRII vector (Invitrogen). This subclone (Gdf7e) was then linearized with MATERIALS AND METHODS SpeI and transcribed with T7 RNA polymerase. The Bmpr1A probe was generated by linearizing plasmid pTFR-EC (a 796-bp fragment Vector Construction and Homologous of the Bmpr1A gene cloned into pBlueScript II; a gift from Yugi Recombination Mishina) with XhoI and transcribing with T7 RNA polymerase (Mishina et al., 1995). The Bmpr1B probe was generated from an Genomic DNA clones containing the Gdf7 locus were isolated IMAGE cDNA clone (accession number AA350367; Research Ge- from a mouse 129SV/J genomic phage library (Stratagene). A 1.1-kb netics) (Lennon et al., 1996) by linearizing with XhoI and transcrib- fragment containing the mature signaling domain of the Gdf7 gene ing with T3 RNA polymerase. Embryonic male urogenital tissue was replaced by a positive selectable neomycin resistance cassette was frozen in OCT compound, 12-micron cryosections were col- by cloning genomic sequences flanking this fragment into the lected, and in situ hybridization was carried out as previously pPNT vector (Tybulewicz et al., 1991). A 1.3-kb XbaI fragment was described (Storm and Kingsley, 1996). Sense strand riboprobes were used as one arm and a 7-kb BamHI partial digestion fragment for used as hybridization controls. the other (Fig. 1). The pPNT vector also contains the herpes thymidine kinase gene for negative selection (Tybulewicz et al., 1991). Targeted clones were generated using R1 embryonic stem Histology cells (a gift from Janet Rossant and Andras Nagy) as previously described (Joyner, 1993). The construct was linearized with NotI Tissue was fixed overnight in 4% paraformaldehyde in PBS and and electroporated into ES cells. Colonies were selected with G418 embedded in paraffin by standard procedures. Then 7-micron for positive selection and Gancyclovir for negative selection and sections were collected, dewaxed in xylene, and stained with were then screened by Southern blot using an external probe hematoxylin and eosin.

FIG. 1. Targeting of the mouse Gdf7 locus. (A) A 1.1-kb fragment containing the entire mature region of the Gdf7 gene was targeted by cloning the neomycin resistance cassette (neo) between restriction fragments that flank the mature region (S, SpeI; X, XbaI; B, BamHI). The herpes thymidine kinase gene (tk) was used for negative selection. (B) A flanking probe (short bar in A) was used to probe a Southern blot of DNA from wild-type (lane 1), Gdf7 mutant (lane 2), and heterozygote (lane 3) mouse genomic DNA digested with SpeI. The flanking probe recognizes a 17-kb wild-type band. A 3-kb band is recognized in the targeted allele because of the introduction of a new SpeI site with the neo cassette. (C) Rapid genotyping by PCR using primers that recognize the targeted region (lower band) and neo cassette (upper band) on genomic DNA from wild-type (lane 1), Gdf7 mutant (lane 2), and heterozygote (lane 3) mice.

Immunohistochemistry developed by using the vectastain peroxidase system (Vector Lab- oratories, Burlingame, CA), and sections were counter-stained with Immunohistochemistry was performed on formalin-fixed/ hematoxylin. paraffin-embedded tissues. Sections (6 ␮m) were cut, deparaffinized in histoclear (National Diagnostics), and rehydrated through de- creasing grades of alcohol into PBS. Antigen retrieval was per- Polyacrylamide Gel Electrophoresis and Western formed by using antigen unmasking solution (Vector Laboratories) Blotting for staining with antibodies against androgen receptor (Affinity Bioreagents, Inc.), cytokeratin 8 (gift of E. B. Lane, University of Protein samples were prepared by collecting 100-␮l volumes of Dundee, UK), and cytokeratin 14 (Biogenex, Inc.). Antigen retrieval both mutant and wild-type seminal vesicle (tissue ϩ secretion) and was not performed for staining with antibodies against smooth boiling them in 200 ␮l of 10% SDS for 5 min. These samples were muscle actin (Sigma) or mouse seminal vesicle secretory proteins then spun in a microfuge and 12 ␮l of the supernatant was either (Fawell et al., 1987). Prior to staining, sections were treated for 10 mixed directly or diluted 1:10 and mixed with 8 ␮l of loading buffer min with 3% H2O2 and blocked with blocking solution (Immuno- (125 mM Tris, pH 6.8, 20% glycerol, 4.1% SDS, 2% vision). For primary antibody staining, antibody dilutions were as ␤-mercaptoethanol, and .001% bromophenol blue). Samples were follows: androgen receptor (1:50), cytokeratin 8 (undiluted boiled for 5 min and run on a 15% polyacrylamide gel by using the hybridoma-conditioned medium), and cytokeratin 14 (1:40), system of Laemmli (1970). Gels were then either immunoblotted or smooth muscle actin (1:500), and mouse seminal vesicle secretory directly stained with Coomassie blue to visualize the proteins proteins (1:5000). Secondary antibodies were anti-mouse biotin (0.25% Coomassie blue, 45% methanol, 10% acetic acid). Gels conjugated or anti-rabbit biotin conjugated (Sigma) used at dilution were transferred electrophoretically to nitrocellulose membrane by 1:200 in PBS containing 1% Sheep Serum. Immunostain was using a BioRad transblot apparatus (transfer buffer contained 25

FIG. 2. Seminal vesicle defects in Gdf7 mutant mice. Dissected organs of 12-wk-old adult wild-type (A) and mutant mice (B) show a signiﬁcant reduction in the size of the seminal vesicles (SV) in the Gdf7 mutant. Seminal vesicles cut in longitudinal section from mutant mice show absence of normal epithelial folds (G vs. H, arrows), and accumulation of large numbers of small, dark staining cells beneath abnormally large epithelial cells (I vs. J, arrowhead). In contrast, the gross and histological appearance of the testes (T), epididymis (E), and vas deferens (VD) are similar in wild-type (A, C, E) and mutant (B, D, F). P, prostate; B, bladder; small arrows, spermatozoa. Scale bars: 2 mm (A, B), 50 microns (C–F), 500 microns (G, H), 35 microns (I, J).

FIG. 3. Immunohistochemical analysis of Gdf7Ϫ/Ϫ and Gdf7ϩ/Ϫ (littermate control) seminal vesicles. Gdf7Ϫ/Ϫ (B, D, F, H) and control (A, C, E, G) adult seminal vesicle tissue from 10-wk-old mice was cut in cross section and analyzed by immunohistochemistry with four different marker antisera. Anti-smooth muscle actin antisera showed similar staining in control (A) and Gdf7Ϫ/Ϫ seminal vesicles (B), demonstrating that the smooth muscle is unaffected in the Gdf7 mutant. Androgen receptor and cytokeratin 8 are normally expressed in the seminal vesicle epithelium (C, E) and are present in Gdf7Ϫ/Ϫ epithelium, but at reduced levels (D, F). Expression of the basal cell marker cytokeratin 14 is seen in rare, scattered cells in control seminal vesicles (G), and in a much larger number of cells in the Gdf7Ϫ/Ϫ mutant (H). Scale bars: 50 microns (A–H).

Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved. Gdf7 and Normal Seminal Vesicle Development 143 mM Tris base, 192 mM glycine, 20% methanol). Following trans- targeted clones. Gdf7 mutants were generated by inter- fer, the nitrocellulose membrane was blocked for2hinTBSbuffer crossing animals heterozygous for the targeted allele. As (10 mM Tris—Cl, pH 8.0, 150 mM NaCl) plus 2% powdered milk, expected, homozygous mice were completely missing se- washed twice for 5 min at room temperature in TBST buffer (10 quences from the Gdf7 mature region (Fig. 1). Mutant mM Tris—Cl, pH 8.0, 150 mM NaCl, 0.05% Tween-20), incubated animals were present at normal Mendelian ratios (37 ϩ/ϩ; for3hinTBST plus 5% BSA containing rabbit anti-mouse SVSP 80 ϩ/Ϫ;34Ϫ/Ϫ; P ϭ 0.8459) and most showed no overt antisera diluted 1:5000 (Fawell et al., 1987), washed four times in 200 ml of TBST at room temperature for 5 min each, incubated in phenotype. A small percentage of the Gdf7 animals died by a 1:5000 dilution of HRP-conjugated secondary antibody in TBST 8 wk of age and their skulls showed indications of hydro- for 1 h, and finally washed two times for 15 min and nine times for cephaly as previously described (Lee et al., 1998). Expres- 5 min in TBST. An ECL kit was used for detection (Amersham). sion of Gdf7 was seen in the choroid plexus, which is responsible for the production of cerebrospinal fluid. Be- Organ Culture cause of the variability and reduced penetrance of the hydrocephalus phenotype, it was not further studied. Neonatal mouse seminal vesicles were cultured as previously described (Shima et al., 1990). The seminal vesicles, prostate, and a small section of urethra were dissected together from day-0 pups. Gdf7 Mutant Males Are Sterile and Have Seminal The combined tissues were then cultured on nucleopore filters Vesicle Defects supported by metal screens in 24-well tissue culture dishes containing 1 ml Dulbecco’s Modified Eagle’s Media (DMEM) high Male Gdf7 mutants were test crossed with wild-type glucose and Ham’s F-12 medium (1:1 vol:vol ratio) supplemented female mice and never produced offspring. Gross examina- with insulin (10 ␮g/ml), transferrin (10 ␮g/ml), L-glutamine (2 mM), tion of the mutant male reproductive system revealed no Dihydrotestosterone (DHT), gentamycin (50 ␮g/ml), and fungizone obvious defects in the testes, epididymis, vas deferens, or (0.25 ␮g/ml). DHT (Sigma) was dissolved in ethanol and used at a prostate. Motile sperm were collected from the epididymis final concentration of 10Ϫ8 M. Media inside the wells were replaced of mutants, and histological examination of the epididymis daily and the tissues were grown at 37°C in 5% CO2. and testes showed normal cellular structure (Figs. 2C–2F, arrows). Tissue Grafting and Recombination Seminal vesicles, on the other hand, appeared grossly abnormal in the Gdf7 mutant males. Wild-type seminal Seminal vesicles were dissected from neonatal Sprague–Dawley vesicles are normally large, branched, glandular structures rats, from 10-wk-old Gdf7Ϫ/Ϫ mice, and from 10-wk-old Gdf7ϩ/Ϫ that have visible folds and appear white due to the secretory mice (littermate controls). Rat seminal vesicle mesenchyme was separated from the seminal vesicle epithelium by using trypsin products in the lumen (Fig. 2A). The mutant seminal digestion and mechanical separation as previously described (Hig- vesicles appeared dramatically smaller than wild-type, gins et al., 1989b). Dissected tissues and tissue recombinants were lacked any sign of folding, and were more darkly colored grafted under the renal capsule of athymic male hosts for 4 wk as than wild-type due to lack of secretions in the lumen (Fig. previously described (Higgins et al., 1989b). 2B, and data not shown). Histological examination of wild-type seminal vesicles showed a complex branching network of finger-like projec- RESULTS tions consisting of columnar epithelium and scattered basal cells (Figs. 2G and 2I). The epithelial folds were greatly Generation of Gdf7-Deficient Mice reduced in the mutant. In addition, the mutant epithelium A 1.1-kb fragment containing the entire mature signaling contained scattered large cells with large nuclei and mul- region of the Gdf7 gene was deleted by using positive tiple nucleoli. Large numbers of small, darkly staining cells selection in embryonic stem cells (Fig. 1A) (Joyner, 1993). containing scant cytoplasm were observed beneath the Previous studies have shown that the activity of BMP secretory cells (Figs. 2H and 2J). These small cells, resem- molecules resides in the mature region (Massague, 1987). In bling basal cells, are normally present in the seminal vesicle addition, previous studies of other BMP family members epithelium but in small numbers. Such basal cells are have shown that mutations that disrupt only the mature thought to function as stem cells precursors of the secretory signaling region of the molecule produce the same pheno- epithelium in the prostate (Bonkhoff and Remberger, 1996). types as mutations that completely disrupt the gene (Marker et al., 1997; Storm et al., 1994). The knockout Immunohistochemical Marker Analysis of Mutant construct should thus generate a null allele of Gdf7. Epithelium The targeting construct was linearized and transfected into embryonic stem cells, and the resulting colonies were To better understand the defects in the seminal vesicles, screened by Southern blot hybridization to identify ho- we analyzed the expression of several immunohistochemi- mologous recombination events. A total of 7.5% of selected cal markers that stain luminal epithelial cells, basal epithe- colonies (9 of 120) showed the restriction fragments ex- lial cells, or mesenchymally derived smooth muscle cells. pected for integration of homologous recombination events. Immunostaining with anti-smooth muscle actin antisera Germline transmission was obtained for three independent revealed a similar staining pattern in both control and

Gdf7Ϫ/Ϫ seminal vesicles, suggesting that smooth muscle has formed correctly from the seminal vesicle mesenchyme of Gdf7 mutants (Figs. 3A and 3B). Immunostaining with anti-androgen receptor antisera showed the expected staining in control epithelial cells, but reduced staining in Gdf7Ϫ/Ϫ (Figs. 3C and 3D). Immunostaining with anti- cytokeratin 8, another marker for luminal cells, also showed the expected staining in the epithelium of control seminal vesicles, and was reduced in the epithelium of Gdf7Ϫ/Ϫ seminal vesicles (Figs. 3E and 3F). Immunostain- ing with anti-cytokeratin 14 antisera was used to identify the number and location of basal cells (Peehl et al., 1994). In control seminal vesicles, little staining is seen, reflecting the low number of basal cells normally scattered through- out the epithelium (Fig. 3G). In the Gdf7 mutant, a dramatic increase in cytokeratin 14-expressing cells was seen (Fig. 3H), suggesting a large increase in the number of basal cells. Mouse seminal vesicle epithelium normally produces large quantities of six secreted proteins, known as seminal FIG. 4. Gdf7 mutant epithelium fails to produce the normal set of vesicle secreted proteins 1–6 (SVSP 1–6), which comprise seminal vesicle secreted proteins (SVSP). Protein samples from the bulk of the copulatory plug and give the seminal vesicle seminal vesicles of 12-wk-old wild-type (a, b, e, f) or mutant (c, d, its characteristic white appearance (Bradshaw and Wolfe, g, h) mice were run undiluted (a, c, e, g) or diluted 1:10 (b, d, f, h) on 1977; Fawell et al., 1987; Higgins et al., 1989a,b). In seminal SDS-PAGE gels that were stained with Coomassie blue (a–d) or vesicle extracts from wild-type mice, all six proteins are subjected to Western blotting with anti SVSP antisera (e–h). The readily detected by Coomassie blue staining of SDS-PAGE six characteristic secreted proteins (sizes indicated) that were gels and their presence was confirmed by Western blotting clearly visible in wild type (a, b, e, f) are absent in the Gdf7 mutant with anti-SVSP antisera (Fig. 4, lanes a, b, e, and f). None of (c, d, g, h). the six bands could be detected on a Coomassie blue- stained gel of mutant seminal vesicle proteins. Western blotting with anti-SVSP antisera confirmed a dramatic reduction or complete absence of these secretory products cultures, prostate epithelium showed extensive branching (Fig. 4, lanes c, d, g, and h). in both wild-type and mutant mice (data not shown).

Gdf7 Is Required for Normal Branching Expression of Gdf7 and Bmpr1A and -1B in Morphogenesis Early in Seminal Vesicle Developing Seminal Vesicles Development To better understand the role of Gdf7 during seminal Seminal vesicle development occurs mostly after birth vesicle development, we examined the developmental ex- when the simple epithelial tube that makes up the early pression of both Gdf7 and the two BMP receptors, Bmpr1A seminal vesicle begins to branch and grow. Previous studies and Bmpr1B. Gdf7 expression was seen in the seminal have shown that the growth of the organ and the process of vesicle mesenchyme as early as 17 days postcoitum (the branching morphogenesis can be recapitulated in vitro by earliest stage examined) and as late as 9 days after birth (the using an organ culture system (Cooke et al., 1987; Shima et latest stage examined; data not shown). At 1 day postpar- al., 1990). Wild-type and Gdf7 mutant reproductive systems tum, Gdf7 is expressed in the mesenchyme adjacent to the appeared similar at birth. The seminal vesicle precursors showed no obvious defects in size or morphology (Figs. 5A epithelium, but no expression is seen in the epithelium and 5C). The seminal vesicles, prostate, and a small portion itself (Fig. 6A). By day 7, Gdf7 expression is still seen in the of attached urethra were removed intact from newborn mesenchyme, but it appears to be more concentrated to the mice, and grown together in organ culture. During 5 days in region of the mesenchyme directly adjacent to the epithe- organ culture, the reproductive tract of wild-type males lium and is still absent from the epithelium (Fig. 6D). In increased substantially in size, and the seminal vesicles contrast, the expression of the two BMP receptors, Bmpr1A precursors showed extensive branching (Fig. 5B). In con- and Bmpr1B, is restricted to the epithelium at day 1 and day trast, little epithelial branching was seen in the seminal 7 (Figs. 6B, 6C, 6E, and 6F). The ligand and receptor are thus vesicle precursor of the Gdf7 mutant in organ culture, expressed in adjacent tissues, consistent with a role for although the gland did expand in size (Fig. 5D). In the same Gdf7 in mesenchymal–epithelial interactions.

FIG. 5. Gdf7 is required for branching morphogenesis during early postnatal development. Wild-type (A, B) and Gdf7 mutant (C, D) reproductive systems were dissected on the day of birth (day 0) and grown in culture on Millipore ﬁlters for 5 days. Both wild-type (A) and Gdf7 mutant (C) are similar in size and showed no signs of branching at day 0. After 5 days in culture, the wild-type seminal vesicles showed substantial signs of growth and branching (B), while the Gdf7 mutant seminal vesicles showed a normal amount of growth, but little or no branching (D). P, prostate; SV, seminal vesicle. Scale bars: 100 microns (A–D).

Wild-Type Mesenchyme Restores Functional staining in the apical cytoplasm of luminal cells and in the Cytodifferentiation of Mutant Epithelium in Tissue adjacent lumen for the Gdf7ϩ/Ϫ seminal vesicles (Fig. 7A). Recombination Experiments Staining was not observed in Gdf7Ϫ/Ϫ seminal vesicles (Fig. 7B). Tissue of each type was grafted under the renal Although Gdf7 RNA expression is limited to mesenchyme cells, many of the effects of the mutation are seen in capsule of male athymic mouse hosts for 1 month. In the overlying epithelium, including defects in branching, addition, recombinant grafts of neonatal rSVM with Ϫ Ϫ ϩ Ϫ decrease expression of epithelial markers (androgen recep- Gdf7 / or Gdf7 / tissue were grafted. After 4 wk, grafts tor and cytokeratin 8), and lack of expression of seminal were harvested and analyzed by immunohistochemistry for vesicle secretory proteins. To test whether these epithelial the presence of mouse SVSP. Grafts of nonrecombined defects could be restored by wild-type mesenchyme, we rSVM and Gdf7Ϫ/Ϫ seminal vesicle were negative for conducted tissue recombination experiments between rat staining, while grafts of nonrecombined Gdf7ϩ/Ϫ, recom- and mouse. Neonatal rat seminal vesicle mesenchyme binant rSVM ϩ Gdf7ϩ/Ϫ seminal vesicle, and recombinant (rSVM), adult Gdf7Ϫ/Ϫ seminal vesicle, and adult rSVM ϩ Gdf7Ϫ/Ϫ seminal vesicle were strongly positive Gdf7ϩ/Ϫ (littermate control) seminal vesicle were isolated. for staining (Fig. 7C, and data not shown). The SVSP- Immunohistochemistry with anti-SVSP revealed strong positive epithelium from the rSVM ϩ Gdf7Ϫ/Ϫ seminal

Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved. 146 oyih 01b cdmcPes l ihso erdcini n omreserved. form any in reproduction of rights All Press. Academic by 2001 © Copyright

FIG. 6. Expression of Gdf7 and BMP receptors in reciprocal layers of developing seminal vesicles. Antisense RNA in situ hybridization experiments were carried out on day 1 (A–C) and day 7 (D–F). Gdf7 expression was clearly seen in mesenchymal tissue surrounding the epithelium at both stages (A, D). In contrast, Bmpr1A (C, F) and Bmpr1B (B, E) receptor expression is seen in the epithelium. A–C and D–F are near adjacent sections. Scale bars: 100 microns (A–F). etee al. et Settle Gdf7 and Normal Seminal Vesicle Development 147

FIG. 7. Wild-type mesenchyme restores secretion in the Gdf7Ϫ/Ϫ epithelium. Seminal vesicle tissue was subjected to immunocyto- chemistry using mouse speciﬁc anti-SVSP antisera. Staining was abundant in the apical cytoplasm and in the lumen adjacent to the epithelium in Gdf7ϩ/Ϫ (littermate control) seminal vesicles (A), but totally absent from Gdf7Ϫ/Ϫ mutant seminal vesicles (B). When rat seminal vesicle mesenchyme was grafted with Gdf7Ϫ/Ϫ mutant epithelium for 4 wk, expression of mouse SVSP staining was restored to normal levels in Gdf7Ϫ/Ϫ epithelium (C). The secretory epithelium of the graft in (C) was conﬁrmed to be of mouse (and thus Gdf7Ϫ/Ϫ) origin by Hoechst 33258 staining that shows a punctate nuclear pattern for mouse but not rat tissue (D, example arrow). Scale bars: 50 microns (A–C), 25 microns (D).

vesicle grafts was confirmed to be of mouse (and thus mesenchymal progress zone so that limb bud outgrowth Gdf7Ϫ/Ϫ) origin using Hoechst 33258 dye staining that can occur. In this case, both BMPs and FGFs are required for shows a punctate nuclear pattern for the secretory epithe- the reciprocal signaling between the epithelium and mes- lium (Fig. 7D). This punctate pattern is characteristic of enchyme (Johnson and Tabin, 1997). During the develop- mouse and not rat tissue (Cunha and Vanderslice, 1984). ment of the mammalian tooth, BMPs and FGFs have also Thus, wild-type mesenchyme can restore seminal vesicle been shown to act as signals in both the mesenchyme and secretion to Gdf7 mutant epithelium. epithelium (Thesleff and Sharpe, 1997). In many cases, classical embryological experiments have revealed reciprocal signaling between mesenchyme and DISCUSSION epithelium, but the signaling molecules responsible have not been identified. The male reproductive system is a good Epithelial–mesenchymal signals are important in the example of a system in which a number of classical experi- development of a number of different organs and tissues. In ments demonstrate reciprocal epithelial–mesenchymal in- some cases, the molecules involved in these signaling teractions (Cunha et al., 1992). FGF7 is expressed in the events have been identified. For example, in the developing seminal vesicle mesenchyme and seminal vesicle branch- limb bud, signals from the apical ectodermal ridge, an ing morphogenesis is blocked by anti-KGF monoclonal epithelial structure, are required to signal the underlying antibody. FGF7 is also capable of partially substituting for

Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved. 148 Settle et al. testosterone to stimulate growth and branching of seminal FGF signaling has recently been shown to control branch- vesicles grown in vitro (Alarid et al., 1994). FGF10 is also a ing morphogenesis of the Drosophila trachea (Metzger and critical mesenchymal factor in seminal vesicle develop- Krasnow, 1999). In that system, ectopic expression of FGF is ment. FGF10 is expressed in the mesenchyme of the devel- sufficient to trigger new branching of the developing epi- oping seminal vesicle (Thomson and Cunha, 1999), and thelium. Although Gdf7 is required for normal epithelial FGF10 knockout mice lack seminal vesicles (A. Donjacour growth, we have no evidence that it is also sufficient to and G.C., unpublished observations). These results impli- stimulate epithelial growth and branching morphogenesis. cate FGF7 and FGF10 as mesenchymal signals responsible In preliminary experiments, addition of recombinant GDF7 for some aspects of seminal vesicle growth and branching protein to developing seminal vesicles, either globally in morphogenesis. However, many of the other molecules the culture media or by local implantation of GDF7-soaked involved in the epithelial–mesenchymal interactions of the Affigel Blue beads, did not stimulate additional epithelial male reproductive system remain to be identified. growth in organ culture (unpublished data). Members of the Our results demonstrate that Gdf7 is a mesenchymal BMP family may normally signal as hetero- or homodimers, signal critical for the development of the seminal vesicles. and it is possible that the recombinant protein experiments The complementary expression patterns of the Gdf7 ligand have not mimicked the form of GDF7 signal actually and BMP receptors suggest that a major target of Gdf7 present in vivo. Alternatively, Gdf7 may normally be a signaling is the seminal vesicle epithelium. Although the permissive signal for seminal vesicle growth, or may only ligand is normally expressed in the mesenchyme, the epi- act in concert with several other molecules to control the thelium of the Gdf7 mutant shows abnormal growth, development of seminal vesicle epithelium. abnormal cytodifferentiation, failure of normal branching We have observed expression of Gdf7 in the mesenchyme morphogenesis, and failure to produce the normal comple- of the other Wolffian duct derivatives, including the vas ment of secretory proteins. Furthermore, the defect in deferens and the epididymis (unpublished data). No obvious secretory protein production can be rescued by wild-type morphological or histological defects were observed in mesenchyme, strongly arguing that Gdf7 acts as a paracrine either of these structures in mutant mice. It is possible that mesenchymal signal that is required for normal seminal the normal growth and differentiation of these structures is vesicle development. maintained by other molecules that can substitute for One striking aspect of the altered cytodifferentiation in Gdf7. Two good candidates for such molecules are the Gdf7 mutants is a high ratio of cytokeratin 14-positive closely related BMP family members, Gdf5 and Gdf6. Both basal epithelial cells to luminal epithelial cells. In wild-type molecules show almost 90% amino acid identity with Gdf7 mice, basal epithelial cells are rare; however, in Gdf7 in the mature signaling region, and are likely to share many mutants, a continuous layer of basal cells is found. An receptors and signaling activities (Storm et al., 1994). We increased ratio of cytokeratin 14-positive basal cells to examined the expression of both Gdf5 and Gdf6, and did cytokeratin 8-positive luminal cells, and loss of secretion of not detect significant expression in the Wolffian duct de- major seminal vesicle proteins, is also seen in castrated rats rivatives. Mice carrying mutations in either Gdf5 or Gdf6 (Hayward et al., 1996). However, androgen receptor and also showed no obvious defects in the male reproductive cytokeratin 8 expression are greatly reduced in Gdf7 mu- tract (unpublished observations). In contrast, mutations in tants, and maintained in castrated animals. It is possible the BMP receptor 1B have recently been reported to elicit that the reduced expression of major seminal vesicle secre- sterility in males (Baur et al., 2000; Yi et al., 2000). tory proteins is secondary to loss of androgen receptor Although no detailed phenotypic analysis has been re- expression in the seminal vesicle epithelium of the Gdf7 ported, it will be interesting to compare the defects in mice mutant. The lineage relationship between basal and lumi- missing Gdf7 with the defects in mice missing one of the nal epithelial cells has not been firmly established for the likely receptors for GDF signaling. seminal vesicle. It has been proposed that the basal cell Previous genetic studies have suggested that members of layer includes stem cells precursors of the secretory epithe- the Gdf5, Gdf6, and Gdf7 subgroup may function in devel- lium in the prostate (Bonkhoff and Remberger, 1996). If this opment of both skeletal and soft tissues. Loss of function is true in the seminal vesicle, the accumulation of basal mutations in Gdf5 are responsible for the classical skeletal epithelial cells and low number of luminal epithelial cells trait brachypodism, and subsequent studies in both mice in Gdf7 mutants may reflect a normal role for Gdf7 in and humans suggest that Gdf5 plays an important role in stimulating basal cells to differentiate into luminal cells. development of both joints and cartilage (Gruneberg, 1977; Alternatively, Gdf7 may normally act on a common pre- Gruneberg and Lee, 1973; Landauer, 1952; Polinkovsky et cursor to basal and luminal cells to direct precursor cells al., 1997; Storm et al., 1994; Storm and Kingsley, 1996; toward a luminal cell fate, or Gdf7 may act directly on basal Thomas et al., 1997). Ectopic expression of Gdf5, Gdf6,or cells to suppress proliferation. Although these potential Gdf7 can also stimulate cartilage and bone formation, and mechanisms cannot currently be distinguished, the mutant tendon and ligament development in vivo (Francis-West et phenotype suggests that Gdf7 normally acts to control the al., 1999; Hotten et al., 1996; Merino et al., 1999; Storm and ratio of basal to luminal epithelial cells in the seminal Kingsley, 1999; Tsumaki et al., 1999; Wolfman et al., 1997). vesicle. Although skeletal structures appear grossly normal in Gdf7

Copyright © 2001 by Academic Press. All rights of reproduction in any form reserved. Gdf7 and Normal Seminal Vesicle Development 149 mutant mice, we are currently testing to see whether more of the male urogenital tract. Role of androgens, mesenchymal- subtle defects may be present in skeletal or tendon struc- epithelial interactions, and growth factors. J. Androl. 13, 465– tures. 475. Previous studies suggest that Gdf7 is required for normal Cunha, G. R., and Vanderslice, K. D. (1984). Identification in development of particular axon subtypes within the spinal histological sections of species origin of cells from mouse, rat and human. Stain Technol. 59, 7–12. cord (Lee et al., 1998) and can stimulate the formation of Davidson, A. J., Postlethwait, J. H., Yan, Y. L., Beier, D. R., van cerebellar granule cells (Alder et al.,1999). The current Doren, C., Foernzler, D., Celeste, A. J., Crosier, K. E., and Crosier, studies add an essential role in mesenchymal–epithelial P. S. (1999). Isolation of zebrafish gdf7 and comparative genetic interactions to the growing list of functions for the Gdf5/ mapping of genes belonging to the growth/differentiation factor 6/7 subgroup of BMP signaling molecules. Further studies 5, 6, 7 subgroup of the TGF-␤ superfamily. Genome Res. 9, will be required to determine how closely related molecules 121–129. are able to stimulate a variety of distinct cellular responses Fawell, S. E., McDonald, C. J., and Higgins, S. J. (1987). Comparison in skeletal, neural, and urogenital precursors. of seminal vesicle secretory proteins of rodents using antibody and nucleotide probes. Mol. Cell. Endocrinol. 50, 107–114. Fawell, S. E., Pappin, D. J., McDonald, C. J., and Higgins, S. J. (1986). ACKNOWLEDGMENTS Androgen-regulated proteins of rat seminal vesicle secretion constitute a structurally related family present in the copulatory We thank Janet Rossant, Andras Nagy for R1 cells, Yugi-Mishina plug. Mol. Cell. Endocrinol. 45, 205–213. for the Bmpr1a probe, Vicki Rosen for recombinant GDF7 protein, Francis-West, P. H., Abdelfattah, A., Chen, P., Allen, C., Parish, J., Ken Ohgi for technical assistance with probes and plasmids, and Ladher, R., Allen, S., MacPherson, S., Luyten, F. P., and Archer, members of the Kingsley lab for useful comments on the manu- C. W. (1999). Mechanisms of GDF-5 action during skeletal script. This work was supported by a postdoctoral fellowship from development. Development 126, 1305–1315. the American Cancer Society, PF-98-153-01 (to P.C.M.), and by Godin, R. E., Robertson, E. J., and Dudley, A. T. (1999). Role of BMP National Institutes of Health Grants CA59831-07, CA64872-06, family members during kidney development. Int. J. Dev. Biol. 43, DK47517-06, and DK52708-03 (to G.R.C.); and AR42236-07 (to 405–411. D.M.K.). D.M.K. is an assistant investigator of the Howard Hughes Gruneberg, H. (1977). Brachypodism in the mouse. Birth Defects Medical Institute. Orig. Artic. Ser. 13, 221–230. Gruneberg, H., and Lee, A. J. (1973). The anatomy and development REFERENCES of brachypodism in the mouse. J. Embryol. Exp. Morphol. 30, 119–141. Alarid, E. T., Rubin, J. S., Young, P., Chedid, M., Ron, D., Aaronson, Hayward, S. W., Baskin, L. S., Haughney, P. C., Cunha, A. R., S. A., and Cunha, G. R. (1994). Keratinocyte growth factor Foster, B. A., Dahiya, R., Prins, G. S., and Cunha, G. R. (1996). functions in epithelial induction during seminal vesicle develop- Epithelial development in the rat ventral prostate, anterior ment. Proc. Natl. Acad. Sci. USA 91, 1074–1078. prostate and seminal vesicle. Acta Anat. 55, 81–93. Alder, J., Lee, K. J., Jessell, T. M., and Hatten, M. E. (1999). Higgins, S. J., Young, P., Brody, J. R., and Cunha, G. R. (1989a). Generation of cerebellar granule neurons in vivo by transplanta- Induction of functional cytodifferentiation in the epithelium of tion of BMP-treated neural progenitor cells. Nat. Neurosci. 2, tissue recombinants. I. Homotypic seminal vesicle recombi- 535–540. nants. Development 106, 219–234. Baur, S. T., Mai, J. J., and Dymecki, S. M. (2000). Combinatorial Higgins, S. J., Young, P., and Cunha, G. R. (1989b). Induction of signaling through BMP receptor IB and GDF5: Shaping of the functional cytodifferentiation in the epithelium of tissue recom- distal mouse limb and the genetics of distal limb diversity. binants. II. Instructive induction of Wolffian duct epithelia by Development 127, 605–619. neonatal seminal vesicle mesenchyme. Development 106, 235– Bonkhoff, H., and Remberger, K. (1996). Differentiation pathways 250. and histogenetic aspects of normal and abnormal prostatic Hogan, B. L. (1996). Bone morphogenetic proteins in development. growth: A stem cell model. Prostate 28, 98–106. Curr. Opin. Genet. Dev. 6, 432–438. Bradshaw, B. S., and Wolfe, H. G. (1977). Coagulation proteins in Hotten, G. C., Matsumoto, T., Kimura, M., Bechtold, R. F., Kron, the seminal vesicle and coagulating gland of the mouse. Biol. R., Ohara, T., Tanaka, H., Satoh, Y., Okazaki, M., Shirai, T., Pan, Reprod. 16, 292–297. H., Kawai, S., Pohl, J. S., and Kudo, A. (1996). Recombinant Brooks, D. E., and Higgins, S. J. (1980). Characterization and human growth/differentiation factor 5 stimulates mesenchyme androgen-dependence of proteins associated with luminal fluid aggregation and chondrogenesis responsible for the skeletal de- and spermatozoa in the rat epididymis. J. Reprod. Fertil. 59, velopment of limbs. Growth Factors 13, 65–74. 363–375. Immergluck, K., Lawrence, P. A., and Bienz, M. (1990). Induction Cooke, P. S., Young, P. F., and Cunha, G. R. (1987). A new model across germ layers in Drosophila mediated by a genetic cascade. system for studying androgen-induced growth and morphogen- Cell 62, 261–268. esis in vitro: The bulbourethral gland. Endocrinology 121, 2161– Itoh, N., Patel, U., Cupp, A. S., and Skinner, M. K. (1998). 2170. Developmental and hormonal regulation of transforming growth Cunha, G. R. (1972). Epithelio-mesenchymal interactions in pri- factor-␤1 (TGF␤1), -2, and -3 gene expression in isolated prostatic mordial gland structures which become responsive to androgenic epithelial and stromal cells: Epidermal growth factor and TGF␤ stimulation. Anat. Rec. 172, 179–195. interactions. Endocrinology 139, 1378–1388. Cunha, G. R., Alarid, E. T., Turner, T., Donjacour, A. A., Boutin, Johnson, R. L., and Tabin, C. J. (1997). Molecular models for E. L., and Foster, B. A. (1992). Normal and abnormal development vertebrate limb development. Cell 90, 979–990.

Joyner, A. L. (1993). “Gene Targeting: A Practical Approach.” IRL tive growth factors in the visceral mesoderm of Drosophila Press, Oxford. embryos. Development 110, 1031–1040. Laemmli, U. K. (1970). Cleavage of structural proteins during the Shima, H., Tsuji, M., Young, P., and Cunha, G. R. (1990). Postnatal assembly of the head of bacteriophage T4. Nature 227, 680–685. growth of mouse seminal vesicle is dependent on 5 Landauer, W. (1952). Brachypodism: A recessive mutation of house- ␣-dihydrotestosterone. Endocrinology 127, 3222–3233. mice. J. Hered. 43, 293–298. Storm, E. E., Huynh, T. V., Copeland, N. G., Jenkins, N. A., Lee, K. J., Mendelsohn, M., and Jessell, T. M. (1998). Neuronal Kingsley, D. M., and Lee, S. J. (1994). Limb alterations in patterning by BMPs: A requirement for GDF7 in the generation brachypodism mice due to mutations in a new member of the ␤ of a discrete class of commissural interneurons in the mouse TGF -superfamily [see comments]. Nature 368, 639–643. spinal cord. Genes Dev. 12, 3394–3407. Storm, E. E., and Kingsley, D. M. (1996). Joint patterning defects Lennon, G., Auffray, C., Polymeropoulos, M., and Soares, M. B. caused by single and double mutations in members of the bone (1996). The I.M.A.G.E. Consortium: An integrated molecular morphogenetic protein (BMP) family. Development 122, 3969– 3979. analysis of genomes and their expression. Genomics 33, 151–152. Storm, E. E., and Kingsley, D. M. (1999). GDF5 coordinates bone Lipschutz, J. H., Fukami, H., Yamamoto, M., Tatematsu, M., and joint formation during digit development. Dev. Biol. 209, Sugimura, Y., Kusakabe, M., and Cunha, G. (1999). Clonality of 11–27. urogenital organs as determined by analysis of chimeric mice. Taragnat, C., Berger, M., and Jean, C. (1988). Preliminary charac- Cells Tissues Organs 165, 57–66. terization, androgen-dependence and ontogeny of an abundant Lipschutz, J. H., Young, P., Taguchi, O., and Cunha, G. R. (1996). protein from mouse vas deferens. J. Reprod. Fertil. 83, 835–842. Urothelial transformation into functional glandular tissue in situ Thesleff, I., and Sharpe, P. (1997). Signalling networks regulating by instructive mesenchymal induction. Kidney Int. 49, 59–66. dental development. Mech. Dev. 67, 111–123. Lung, B., and Cunha, G. R. (1981). Development of seminal vesicles Thomas, J. T., Kilpatrick, M. W., Lin, K., Erlacher, L., Lembessis, P., and coagulating glands in neonatal mice. I. The morphogenetic Costa, T., Tsipouras, P., and Luyten, F. P. (1997). Disruption of effects of various hormonal conditions. Anat. Rec. 199, 73–88. human limb morphogenesis by a dominant negative mutation in Marker, P. C., Seung, K., Bland, A. E., Russell, L. B., and Kingsley, CDMP1. Nat. Genet. 17, 58–64. D. M. (1997). Spectrum of Bmp5 mutations from germline Thomson, A. A., and Cunha, G. R. (1999). Prostatic growth and mutagenesis experiments in mice. Genetics 145, 435–443. development are regulated by FGF10. Development 126, 3693– Massague, J. (1987). The TGF-␤ family of growth and differentia- 3701. tion factors. Cell 49, 437–438. Timme, T. L., Truong, L. D., Merz, V. W., Krebs, T., Kadmon, D., Merino, R., Macias, D., Ganan, Y., Economides, A. N., Wang, X., Flanders, K. C., Park, S. H., and Thompson, T. C. (1994). Wu, Q., Stahl, N., Sampath, K. T., Varona, P., and Hurle, J. M. Mesenchymal-epithelial interactions and transforming growth (1999). Expression and function of Gdf-5 during digit skeletogen- factor-␤ expression during mouse prostate morphogenesis. Endo- esis in the embryonic chick leg bud. Dev. Biol. 206, 33–45. crinology 134, 1039–1045. Metzger, R. J., and Krasnow, M. A. (1999). Genetic control of Tsumaki, N., Tanaka, K., Arikawa-Hirasawa, E., Nakase, T., branching morphogenesis. Science 284, 1635–1639. Kimura, T., Thomas, J. T., Ochi, T., Luyten, F. P., and Yamada, Y. Mishina, Y., Suzuki, A., Ueno, N., and Behringer, R. R. (1995). (1999). Role of CDMP-1 in skeletal morphogenesis: Promotion of Bmpr encodes a type I bone morphogenetic protein receptor that mesenchymal cell recruitment and chondrocyte differentiation. is essential for gastrulation during mouse embryogenesis. Genes J. Cell Biol. 144, 161–173. Dev. 9, 3027–3037. Tybulewicz, V. L., Crawford, C. E., Jackson, P. K., Bronson, R. T., Panganiban, G. E., Reuter, R., Scott, M. P., and Hoffmann, F. M. and Mulligan, R. C. (1991). Neonatal lethality and lymphopenia (1990). A Drosophila growth factor homolog, decapentaplegic, in mice with a homozygous disruption of the c-abl proto- regulates homeotic gene expression within and across germ oncogene. Cell 65, 1153–1163. Wolfman, N. M., Hattersley, G., Cox, K., Celeste, A. J., Nelson, R., layers during midgut morphogenesis. Development 110, 1041– Yamaji, N., Dube, J. L., DiBlasio-Smith, E., Nove, J., Song, J. J., 1050. Wozney, J. M., and Rosen, V. (1997). Ectopic induction of tendon Peehl, D. M., Leung, G. K., and Wong, S. T. (1994). Keratin and ligament in rats by growth and differentiation factors 5, 6, expression: A measure of phenotypic modulation of human and 7, members of the TGF-␤ gene family. J. Clin. Invest. 100, prostatic epithelial cells by growth inhibitory factors. Cell Tissue 321–330. Res. 277, 11–18. Yi, S. E., Daluiski, A., Pederson, R., Rosen, V., and Lyons, K. M. Polinkovsky, A., Robin, N. H., Thomas, J. T., Irons, M., Lynn, A., (2000). The type I BMP receptor BMPRIB is required for chondro- Goodman, F. R., Reardon, W., Kant, S. G., Brunner, H. G., van der genesis in the mouse limb. Development 127, 621–630. Burgt, I., Chitayat, D., McGaughran, J., Donnai, D., Luyten, F. P., and Warman, M. L. (1997). Mutations in CDMP1 cause autosomal Submitted for publication January 3, 2001 dominant brachydactyly type C [letter]. Nat. Genet. 17, 18–19. Revised February 21, 2001 Reuter, R., Panganiban, G. E., Hoffmann, F. M., and Scott, M. P. Accepted February 21, 2001 (1990). Homeotic genes regulate the spatial expression of puta- Published online April 17, 2001