發(fā)信人: newsci (樂科學), 信區(qū): BioTrends 標 題: Cell:有利于細胞外信號分子Hedgehog分布的細胞膜上的通道發(fā)信站: 生命玄機站 (Fri Oct 4 03:34:43 2002) , 轉信 挺有趣的 Cell, Vol 111, 63-75, October 2002 Hedgehog-Mediated Patterning of the Mammalian Embryo Requires Transporter-like F unction of Dispatched Yong Ma1, Alfrun Erkner2, Ruoyu Gong1, Shenqin Yao1, Jussi Taipale1, Konrad Basl er2, and Philip A. Beachy1 1 Howard Hughes Medical Institute, Department of Molecular Biology and Genetics, Johns Hopkins University School of Medicine, Baltimore, MD 21205 USA 2 Institut für Molekularbiologie, Universität Zürich, Winterthurerstrasse 190, CH-8057 Zürich, Switzerland Correspondence: Philip A. Beachy 410-955-1862 (phone) 410-955-9124 (fax) [email protected] Summary The dispatched (disp) gene is required for long-range Hedgehog (Hh) signaling in Drosophila. Here, we demonstrate that one of two murine homologs, mDispA, can r escue disp function in Drosophila and is essential for all Hh patterning activit ies examined in the early mouse embryo. Embryonic fibroblasts lacking mDispA res pond normally to exogenously provided Sonic hedgehog (Shh) signal, but are impai red in stimulation of other responding cells when expressing Shh. We have develo ped a biochemical assay that directly measures the activity of Disp proteins in release of soluble Hh proteins. This activity is disrupted by alteration of resi dues functionally conserved in Patched and in a related family of bacterial tran smembrane transporters, thus suggesting similar mechanisms of action for all of these proteins. Introduction The pattern of cellular proliferation and differentiation that leads to normal d evelopment of embryonic structures often depends upon the localized production o f secreted protein signals. Cells surrounding the source of a particular signal respond in a graded manner according to the effective concentration of the signa l, and this response produces the pattern of cell types constituting the mature structure (Jessell, 2000 ; Gurdon and Bourillot, 2001 ). The importance in embry onic patterning of concentration-dependent cellular responses highlights in turn the importance of mechanisms that influence the effective concentration of the signaling protein. Such mechanisms include the interaction of a signaling protei n with extracellular matrix, the presence and action of other interacting secret ed proteins, and the sequestration and possible rerelease of the signal within t he field of responding cells (Dudley and Tabin, 2000 ; Teleman et al., 2001 ). Among secreted signaling proteins, the Hedgehog (Hh) family (Ingham and McMahon, 2001 ) is unique in that the mature active form of the signal (HhNp, for N-term inal, processed domain) is dually lipid modified, with an ester-linked carboxy-t erminal cholesteryl moiety (Porter et al., 1996b ) and an amide-linked amino-ter minal palmitate (Pepinsky et al., 1998 ). Cholesterol addition results from an a utoprocessing reaction undergone by the Hedgehog protein precursor (Lee et al., 1994 ; Porter et al., 1996b ), and palmitoylation, which is critical for signali ng activity of the Hh protein, requires action of the Skinny hedgehog (Ski) acyl transferase (Chamoun et al., 2001 ; Lee and Treisman, 2001 ). Efficient addition of palmitate depends upon prior cholesterol addition, and truncated Hh proteins that are not cholesterol modified show poor activity in embryos (Lewis et al., 2001 ) unless expressed at high levels (Porter et al., 1996a ). Despite dual lipid modification, which might be expected to firmly anchor HhNp t o the membranes of producing cells, Hh signaling extends beyond immediately adja cent cells. In the wing imaginal disc of Drosophila, for example, Hh production in the posterior compartment directly induces expression of the target genes pat ched (ptc) and decapentaplegic (dpp) in a band extending 5–10 cells beyond the boundary of Hh protein expression (Hidalgo and Ingham, 1990 ; Basler and Struhl, 1994 ; Capdevila and Guerrero, 1994 ). The range of direct Hh signaling is even greater in vertebrate embryos, where the mature form of the Sonic hedgehog protein (ShhNp), also dually lipidated, extends its influence many cells beyond its source to globally influence development throughout a tissue or structure. In the developing neural tube, for example, direct responses to Sonic hedgehog (Shh) are required for induction of diverse neuronal types throughout the ventral half of the neural tube (Roelink et al., 1995 ; Jessell, 2000 ; Briscoe et al., 2001 ), despite production of Shh protein exclusively in the notochord and floor plate (Roelink et al., 1995 ). The influence of this ventral midline source of Shh signal also extends through the paraxial mesoderm as far as the dorsal somite, where Shh directly induces Myf5 expression in the myogenic precursors of the epaxial musculature (Borycki et al., 1999 ; Gustafsson et al., 2002 ). In addition, a long-range effect of direct Shh signaling is evident in the developing limb, where Shh expression at the posterior margin induces gradients of Ptch transcription and of Gli3 transcription and proteolytic processing across the limb bud, encompassing many cell diameters (Goodrich et al., 1996 ; Marigo et al., 1996b ; Wang et al., 2000 ). The long-range action of Shh in these tissues raises the question of how a duall y lipidated protein signal can escape the membranes of cells in which it is prod uced to directly stimulate pathway activity in distant target cells. One possibl e clue to this puzzle derives from the identification of the Drosophila gene dis patched (disp) (Burke et al., 1999 ), which has a mutant phenotype similar to th at of hh, and encodes a protein that is similar in sequence and transmembrane to pology to the Ptc component of the Hh receptor (Nakano et al., 1989 ; Hooper and Scott, 1989 ; Marigo et al., 1996a ; Goodrich et al., 1996 ; Stone et al., 1996 ; Fuse et al., 1999 ). Unlike Ptc, however, mutant mosaic studies indicate a re quirement for Disp function exclusively in Hh-producing cells, with loss of Disp function resulting in accumulation of the Hh signal in the producing cells and restriction of target gene expression to those cells immediately adjacent to Hh- producing cells (Burke et al., 1999 ). These results suggest that Disp acts by s ome mechanism to present or release the Hh signal for stimulation of distant tar get cells. To investigate the role and mechanism of Disp protein action in mammalian Hh sig naling, we characterized two murine disp homologs. We demonstrate that one of th ese, mDispA, is required for all detectable manifestations of Hh signaling in th e patterning of the early mouse embryo. We also demonstrate in embyonic fibrobla sts that a normal response to the ShhNp signal does not require mDispA function, and further show that Shh processing and modification in mDispA mutant cells ar e normal. In addition, in a direct biochemical assay of Hh export by Disp, the l evels of a soluble form of Hh protein released into the medium are increased man y fold upon coexpression of mammalian or Drosophila Disp proteins. This activity is disrupted by alteration of residues functionally conserved in Patched and in a related family of bacterial transmembrane transporters, thus suggesting the p ossibility that all of these proteins act by similar mechanisms. Results and Discussion Summary Introduction Results and Discussion Conclusions Experimental Procedures References Identification of Two Murine dispatched Homologs Using ESTs identified in database searches and by cDNA cloning, we characterized coding sequences for two murine homologs of disp, mDispA, and mDispB (Figure 1A ). Based on the experimentally verified topology of the homologous protein, NPC1 (Davies and Ioannou, 2000 ), we propose a membrane topology for Disp that inclu des twelve transmembrane spans, with cytoplasmic N- and C-terminal tails (Figure 1B). The predicted amino acid sequences of Disp, mDispA, and mDispB are easily aligned, with greatest similarity noted within a continuous region that includes and extends just beyond the twelve predicted transmembrane spans (TM region; Fi gure 1A); little similarity occurs outside this region. Overall, the murine prot eins, particularly mDispA, have larger N- and C-terminal cytoplasmic domains, wh ereas the Drosophila protein has relatively larger loops, particularly the extra cellular loops between TM1 and TM2, and between TM7 and TM8. Figure 1. Characterization of Murine dispatched Homologs (A) Sequence alignment between mDispA, mDispB, and Drosophila Disp. Identical re sidues are in yellow background. The twelve transmembrane spans are underlined, and the GxxxD motifs within TM4 and TM10 are in red rectangles. The intron/exon junctions are indicated by red vertical lines. Blue brackets delimit the sequenc es used in determining the percentages of sequence identities (see text). (B) Proposed topology of Disp proteins. Note that the structure probably arose b y tandem duplication of a six transmembrane unit. Red letters indicate the three aspartate residues within the GxxxD(D) motifs in TM4 and TM10. (C) Generation of mDispA null allele. Homologous recombination of the wild-type allele with the targeting vector results in a mutant mDispA allele lacking a lar ge exon coding for 11 of the 12 transmembrane domains (blue vertical lines). The seven coding exons are represented by white rectangles, and green and red verti cal lines inside exons 1 and 7 indicate the start and stop codons, respectively. Thymidine kinase (tk, green rectangle) and LoxP (red box) -flanked Neo (gray bo x) selection cassettes are also shown. Introns of known and unknown lengths are represented by solid and dashed lines, respectively. View larger version: [In this window] [In new window] Within the membrane-spanning region, mDispA and mDispB, respectively, display se quence identities with Disp of 36% and 31%, discounting the gaps, whereas sequen ce identity in this region is 42% between mDispA and mDispB. These relationships suggest that the two murine homologs duplicated after divergence of the insect and mammalian lineages. Consistent with this scenario, the coding sequences of m DispA and mDispB are identically distributed among seven exons, whereas Drosophi la Disp coding sequences are differently distributed with only a single intron/e xon junction at a point homologous to a junction in the mDisp genes (see Figure 1A). Early Embryonic Expression and Functional Conservation of mDispA, but not mDispB Given their probable duplication from a single ancestral gene, we sought to dist inguish mDispA and mDispB functionally on the basis of their embryonic expressio n and their ability to complement disp mutations in Drosophila. By in situ hybri dization, we found that the mDispA message is detected throughout the embryo at 7.5 days of gestation (E7.5; data not shown). This nearly ubiquitous expression of mDispA is maintained throughout all stages examined, albeit with some tissue- specific variations in level. Thus, for example, the level of expression at E8 i s higher in the somites than in the rest of the embryo (Figure 2A), and a higher relative level of expression is observed in the branchial arches at E8.75 (Figu re 2B) and in the limb buds at E9.5 (Figure 2C); in contrast, little expression is observed at E8.75 in the heart. A short color reaction at E10.25 reveals an a pparent gradient of mDispA mRNA levels across the forelimb and hindlimb buds, fr om highest at the anterior to lowest at the posterior (Figure 2D), suggesting th at mDispA transcription in the limb bud may be negatively regulated by signals f rom the posterior, such as Shh. At all the stages examined before E10, the sense strand probe for mDispB yielded a stronger signal than the corresponding antise nse strand probe (data not shown), suggesting that mDispB is not expressed durin g the first half of gestation. Figure 2. Early Embryonic Expression and Functional Conservation of mDispA, but not mDispB (A–C) mDispA is nearly ubiquitously expressed during early embryonic development, as indicated by whole-mount in situ hybridization of wild-type embryos with antisense mDispA probe at E8 (A), E8.75 (B), and E9.5 (C). (D) shows the gradient of mDispA message level across the hindlimb bud of E10.25 wild-type embryo (anterior is up). (E–I) The disp mutation in Drosophila can be rescued by mDispA, but not by mDispB. (E) Diagram of a wild-type third instar Drosophila wing imaginal disc. Endogenou s Hh is expressed in the posterior (P) compartment, inducing a 5–10 cell wide band of anterior (A) compartment cells to express the ptc-lacZ reporter gene (red). Blue color indicates the pattern of en-GAL4 driven ectopic gene expression. The yellow box demarcates the region analyzed in (F?CI). (F) shows the wide band of ptc-lacZ expression in wild-type wing disc. In disp w ing discs (G), the ptc-lacZ expression is limited to a 1 or 2 cell wide region a djacent to the A-P boundary. This defect can be fully rescued by expression of m DispA (H) but not mDispB (I). disp mutants die at pupal stage (Pupal lethal), an d this lethality cannot be rescued by mDispB expression. In contrast, mDispA-exp ressing disp flies develop to adulthood with normal wing pattern (compare J and K). View larger version: [In this window] [In new window] For further assessment of mDispA and mDispB, full coding sequences were tested f or their ability to rescue a disp mutation. Although both proteins were expresse d in Drosophila cultured cells (data not shown), their biological activities in Drosophila were markedly different, as expression of mDispA in Hh-producing cell s restored full expression of the ptc-lacZ target gene in wing imaginal discs ( Figure 2H, compare to Figures 2F–2G) and produced viable adults with no patterning defects ( Figure 2K, compare to Figure 2J), whereas mDispB rescued neither target gene expression (Figure 2I) nor viability. Heart Looping and Turning Defects in mDispA Mutant Embryos As mDispB fails to rescue Drosophila disp and appears not to be expressed in ear ly mouse embryogenesis, we concentrated on mDispA as the gene most likely to pla y a role in early mammalian Hh signaling. The genomic organization of the mDispA gene (Figure 1C) reveals that the 3'-most coding exon contains 78% of the codin g sequence (1193/1521 codons), including eleven of the twelve transmembrane doma ins. We therefore eliminated this exon by targeted recombination (Figure 1C) in the expectation that such a deletion would cause complete loss of mDispA functio n. We found, using this targeted allele of mDispA, that homozygous mutants (mDispA- /-) die at or soon after E9.5 whereas heterozygotes (mDispA+/-) are phenotypical ly wild-type. The early demise of mDispA-/- embryos suggests a phenotype more se vere than that of Shh-/- embryos, which survive through most of gestation (Chian g et al., 1996 ). Indeed, a direct comparison at E9.5 reveals that mDispA-/- emb ryos (Figure 3A) are smaller than Shh-/- embryos (Figure 3C) and also carry an a bnormally inflated pericardial sac and a kink in the trunk with dorsal instead o f ventral curling of the tail. The mDispA-/- phenotype thus greatly resembles th at of Smoothened mutant (Smo-/-) embryos, which similarly display an inflated pe ricardial sac and abnormal trunk morphology (Figure 3B) due to a defect in embry onic turning. Smo is a seven transmembrane protein required for all aspects of H h signaling (van den Heuvel and Ingham, 1996 ; Alcedo et al., 1996 ; Chen et al. , 2001 ; Zhang et al., 2001 ). Embryos lacking Smo function previously were repo rted to show patterning defects equivalent to those observed in embryos doubly m utant for Shh and Indian hedgehog (Ihh; Zhang et al., 2001 ). These defects incl ude a failure to establish normal asymmetry along the left/right axis, resulting in a failure of normal embryonic turning and abnormal looping of the embryonic heart; the heart defects are probably responsible for the inflated pericardial s ac and early embryonic death. Figure 3. The Laterality Defects of mDispA-/- Embryos (A–C) Overall morphological defects of E9.5 mDispA-/- (A), Smo-/- (B), and Shh-/- (C) embryos compared to wild-type embryos in the same litters. Note the similar phenotypes of mDispA-/- and Smo-/- embryos. (D and E) Scanning electron micrographs of embryo hearts at E9.5, rostral is to the top. Compared to the normal rightward heart looping of a wild-type embryo (D ), heart looping of a mDispA-/- embryo (E) is almost completely lost. (F) Whole-mount in situ hybridization for Nodal expression at 5-6 somite stage. Whereas residual Nodal expression is detected in the node of mDispA-/- embryo (a rrowheads), the normal Nodal expression in the left lateral plate mesoderm (arro w) is completely lost in the mDispA-/- embryo. View larger version: [In this window] [In new window] We found that left/right asymmetry in mDispA-/- embryos indeed is disrupted, as indicated by defective heart looping (compare Figures 3D and 3E). At the molecul ar level left/right asymmetry defects are apparent even earlier, as the normal a symmetric expression of Nodal in left lateral plate mesoderm (Collignon et al., 1996 ; Lowe et al., 1996 ) was disrupted in 4–8 somite stage embryos, with no expression detected in seven mutant embryos examined (Figure 3F). As noted for Smo-/- embryos (Zhang et al., 2001 ), expression in the node of these same embryos was variable, with three embryos displaying little expression (not shown), three displaying a somewhat higher level of expression on the left side (Figure 3F) as is characteristic of normal embryos, and one displaying a similar level of expression on right and left sides (not shown). Early Forebrain and Branchial Arch Defects in mDispA Mutants In addition to defects in left/right asymmetry, mDispA-/- embryos also display d efects of the prospective head and face that are characteristic of loss of Hh si gnaling (Chiang et al., 1996 ). Thus, by the early head fold stage (E8.25; Figur es 4A–4D), Shh-/-, Smo-/-, and mDispA-/- embryos all display common midline defects of the neural plate in the region of the prospective forebrain. These defects include midline fusion of the anterior lips of the cephalic neural plate and a poorly defined or absent midline groove and consequent incomplete separation of the bilateral evaginations that form the optic vesicles. Soon after, at ~E8.5 (Figures 4E?C4H), the well-separated bilateral cup-like evaginations in the walls of the wil d-type forebrain are replaced in Shh-/-, Smo-/-, and mDispA-/- embryos by a deep single evagination in the floor of the forebrain. It is this evagination that p roduces the cyclopic eye in later Shh-/- embryos, and presumably also would do s o in Smo-/- and mDispA-/- embryos if their heart defects did not preclude furthe r embryonic development. Figure 4. The Forebrain and Branchial Arch Defects of Shh-/-, Smo-/-, and mDispA-/- Embryo s Examined by Scanning Electron Microscopy (A–D) Frontal views of E8.25 embryo cephalic region. (E–H) Frontal views of E8.5 embryo cephalic region. (I–L) Lateral views of E9.5 embryo cephalic region. (A, E, and I) Wild-type embryos. (B, F, and J) Shh-/- embryos. (C, G, and K) Smo-/- embryos. (D, H, and L) mDispA-/- embryos. The midline defects of E8.25 and E8.5 forebrain are marked by arrowheads and the branchial arch defects at E9.5 are marked by a rrows. View larger version: [In this window] [In new window] Defects of the branchial arches are also evident in mDispA-/- embryos. In partic ular, the mandibular component of the first branchial arch is severely reduced a nd fused in the midline ( Figure 4L, compare to Figure 4I). As seen for left/rig ht axis defects, the branchial arch defects in mDispA-/- embryos are similar to those found in Smo-/- embryos (Figure 4K), and both are more severe than the def ects in Shh-/- embryos (Figure 4J), suggesting that normal development of the br anchial arches may also involve Hh signaling activity other than that provided b y Shh. DISPA Is Located within HPE10, a Human Holoprosencephaly Locus As in Shh-/- mutants, the cephalic defects in mDispA-/- embryos are reminiscent of holoprosencephaly (HPE) in humans. Holoprosencephaly encompasses a spectrum o f brain and facial midline deficits that in extreme cases involve cyclopia and d evelopment of the forebrain or prosencephalon as a single undivided vesicle (Mue nke and Beachy, 2001 ). Although heterozygous mutations in the murine Shh gene a re silent, human SHH gene function is haploinsufficient and ~5% of HPE, generall y at the milder end of the spectrum, can be accounted for by heterozygous SHH mu tations (Roessler et al., 1996 ; Muenke and Beachy, 2001 ). Indeed, the DISPA ge ne is located in chromosomal region 1q42 (data not shown), which corresponds to the chromosomal interval 1q42-ter associated with HPE10 (Muenke and Beachy, 2001 ). It is conceivable that, despite the apparent absence of a heterozygous pheno type in the mouse, haploinsufficiency of the human DISPA gene could account for apparent dominant effects of chromosomal abnormalities associated with HPE10 (Mu enke and Beachy, 2001 ). Neural Tube Patterning Defects in mDispA-/- Mutants Embryos lacking mDispA function also display patterning defects at more caudal l evels. The ventral neural tube at the level of the presumptive spinal cord, for example, fails to take on its characteristic shape with well-differentiated floo r plate cells at a sharply defined midline, and instead develops as an undiffere ntiated epithelial tube (compare Figures 5A and 5B). To further characterize the effects of mDispA loss on patterning of the neural tube and somites, we examine d various genes known to depend on Hh signaling for their normal expression, and which in some cases serve as markers for the development of particular cell typ es. Figure 5. The Neural Tube and Somite Defects of mDispA-/- Mutants (A and B) Scanning electron micrographs of the neural tubes of E9.5 embryos. The clearly defined morphology of the floor plate in wild-type embryo (A, arrowhead ) is lost in mDispA-/- mutant (B). (C–P) Loss of long-range Hh signaling in mDispA-/- embryos. (C) In E9.5 mDispA-/- embryos, weak expression of a Hh pathway sensitive reporte r (Ptch-LacZ) is retained in tissues that express Hh, such as the notochord (arr owhead) and gut. However, LacZ staining is completely lost in tissues that are f arther from Hh sources, such as ventral neural tube and somites (arrow in wild-t ype embryo). (D) Pax7 expression in E9.25 embryos. (E and F) The crosssections marked in (D) showing the ventral expansion of Pax7 expression in both neural tube (arrow) and somites (arrowhead) in mutant embryo (F) compared to wild-type (E). (G and H) HNF3 expression at E8.25 (G) and E9.25 (H). Whereas gut and weak notoc hord (arrowhead) expression of HNF3 persists at both stages, ventral neural tube expression (arrow in wild-type embryo) is not detected in mDispA-/- embryos. (I) Shh expression is retained in E9.5 mDispA-/- embryo in the gut and the remai ning portions of the degenerating notochord (arrowhead), but not in the ventral neural tube (arrow in wild-type embryo). (J), (K), and (L) show sections through embryos as indicated in (I). In mDispA-/ - embryos, where notochord is intact (K), expression of Shh is comparable to tha t of wild-type (I), but no expression is detected in the adjacent neural tube. N o Shh expression is detected in a section in which the notochord has degenerated (L). (M) Pax1 expression in E8.75 embryos. Shh induced sclerotomal expression of Pax1 in wild-type embryo (arrow) is totally lost in a mDispA-/- mutant, whereas the Hh independent expression in the pharyngeal region (arrowhead) is not affected. (N) Myf5 expression in E9.25 embryos. (O) (wild-type) and (P) (mDispA-/-) are higher magnification images of the middl e sections of the embryos (dashed boxes in N). Shh dependent dorsal somite expre ssion of Myf5 (arrow) is abolished in mDispA-/- mutant whereas Shh independent v entral somite expression (arrowhead) is not affected. View larger version: [In this window] [In new window] Perhaps the most universal indicator of Hh signaling activity is the expression of Patched (Ptch) (Hidalgo and Ingham, 1990 ; Goodrich et al., 1996 ; Marigo et al., 1996b ), which encodes a component of the Hh receptor and is induced by Hh signaling. We employed the murine Ptch-lacZ allele (Goodrich et al., 1997 ) to m onitor pathway activity, as fusion of the E. coli lacZ gene in frame to the thir d codon of Ptch coding sequences provides a sensitive report of Ptch expression through histochemical staining for -galactosidase activity. The normal expressio n of -galactosidase from Ptch-lacZ in the ventral neural tube and somites is com pletely lost in the mDispA mutant background (Figure 5C), although continued exp ression in the notochord and portions of the gut suggests that some response to Hh signaling is maintained in tissues that themselves express Hh proteins (see b elow). We note that heterozygosity for the Ptch-lacZ mutant allele did not notic eably affect the abnormal morphology of mDispA mutant embryos. As an additional sensitive indicator of Hh protein influence we examined expression of Pax7, norm ally suppressed in the ventral neural tube by Shh that is produced in the notoch ord and floor plate (Fan and Tessier-Lavigne, 1994 ; Ericson et al., 1996 ). In mDispA-/- mutants, we find that the normal dorsally restricted domain of Pax7 ex pression extends throughout the neural tube, including the ventral midline, sugg estive of a loss of Shh signaling (Figures 5D–5F). As markers of ventral midline structures, the Shh and HNF3- genes normally are e xpressed independently of Hh signaling in the early gut and notochord and are la ter induced in the floor plate by Shh signaling from the notochord. In mDispA-/- embryos, normal initiation of Shh and HNF3- expression in the notochord and gut is observed (Figures 5G–5I), but expression does not occur in the neural tube. Later expression of these markers indicates that the notochord begins to degenerate in a discontinuous fashion, much like the degeneration observed in Shh mutants (Chiang et al., 1996 ). Sections through different levels of the trunk in these embryos (Figures 5J?C5L) indeed demonstrate that the notochord is absent in regions lacking Shh expr ession. Furthermore, these sections clearly demonstrate that no expression of Sh h occurs in the neural tube, even at levels of the trunk where notochord is pres ent and is expressing Shh RNA. Thus, in the mDispA-/- mutant background, express ion of Shh in the notochord appears to be incapable of inducing normal Shh respo nses in the adjacent neural tube. Somite Patterning Defects in mDispA-/- Mutants Shh signaling normally functions to induce sclerotome, marked by expression of P ax1 in the ventral somite (Fan and Tessier-Lavigne, 1994 ; Johnson et al., 1994 ; Fan et al., 1995 ). Furthermore, Shh also suppresses Pax7 expression in the ve ntral somite (Fan and Tessier-Lavigne, 1994 ), restricting it to the dorsal derm omyotome. In addition, Shh signaling contributes to dorsal somite expression of Myf5 in the myotomal precursors of epaxial muscles (Borycki et al., 1999 ), and this represents one of the longest range effects that can be assigned directly t o Shh signaling (Gustafsson et al., 2002 ). All three of these activities are di srupted in mDispA-/- embryos. Thus, sclerotomal Pax1 expression is lost (althoug h Hh-independent expression in the pharyngeal pouches is maintained; see Figure 5M), Pax7 expression expands ventrally throughout the entire paraxial mesoderm ( Figures 5D–5F), and the epaxial domain of Myf5 expression in the dorsal somite is lost (Figures 5N?C5P), even though a more ventral domain of Hh-independent expression of Myf5 is maintained. Response of mDispA-/- Cells to Hh Signaling Although most aspects of Hh signal response appear to be disrupted in mDispA-/- embryos, some expression of the Ptch-lacZ reporter is retained in the notochord and in the gut at E9.25 (Figure 5C); this expression is even clearer at E8.5, be fore the onset of notochord degeneration (Figure 6A). The retention of Ptch-lacZ expression in mDispA-/- embryos would appear to represent a genuine response to Shh signaling, albeit restricted to cells that themselves express Shh, as this response is absent in Smo-/- embryos (Zhang et al., 2001 ). Furthermore, homozyg osity for the Ptch-lacZ mutant allele in a DispA mutant background produced the abnormal morphology and widespread -galactosidase expression characteristic of h omozygous Ptch-lacZ mutant embryos (data not shown), indicating that mDispA func tions upstream of Ptch and of signal reception, possibly in signal production. T o further examine the role of mDispA function in Hh signaling, fibroblastic cell lines isolated from E8.75 mDispA-/- and mDispA+/- embryos were tested for respo nse to exogenously added ShhNp. As monitored by Gli-luc, a luciferase reporter w ith eight tandem Gli binding sites (Sasaki et al., 1997 ; Taipale et al., 2000 ) , heterozygous and homozygous mutant cells were both responsive to Shh signaling (Figure 6B), and this response was abrogated by the specific Hh pathway antagon ist, cyclopamine (Cooper et al., 1998 ; Incardona et al., 1998 ; Taipale et al., 2000 ). Figure 6. Cells Lacking mDispA Function Respond Normally to Shh but Fail to Efficiently St imulate Responding Cells (A) mDispA-/- cells are capable of responding to Hh signals, as indicated by the expression of a Hh-sensitive Ptch-lacZ reporter in the notochord (arrow) and gu t in a E8.5 mDispA-/- embryo. (B) mDispA-/- primary fibroblasts respond normally to exogenous Shh. mDispA-/- ( yellow bars) and mDispA+/- (red bars) primary fibroblasts were transfected with Shh-sensitive Gli-luc reporter. Luciferase expression was induced in both cell l ines in response to exogenously added ShhNp (10 nM), and the responses were bloc ked by cyclopamine (5 µM), a specific inhibitor of Shh signaling. (C) Shh autoprocessing and cholesterol modification do not require mDispA functi on. mDispA-/- cells were transfected with the expression constructs indicated, a nd after 3 days, the electrophoretic mobilities of Shh proteins were analyzed by SDS-PAGE followed by immunoblotting. Migration of Shh N-terminal domain derived from full-length Shh construct (Shh) was not affected by coexpression of mDisp proteins and was faster than that of protein derived from truncated ShhN constru ct (ShhN). Because these proteins are identical in amino-acid sequence, the fast er migration is indicative of cholesterol modification. (D) Shh-expressing mDispA-/- cells fail to efficiently stimulate Hh responsive r eporter cells. mDispA-/- cells were transfected with the expression constructs i ndicated and mixed with Shh-sensitive reporter cells (Shh-LIGHT2) in the absence or presence of cyclopamine. Transient expression of Shh alone elicited low leve l of response probably due to contact-dependent signaling. Coexpression of mDisp A increased the response of the Shh-LIGHT2 cells to a level approaching that eli cited by soluble ShhN. However, mDispB expression had no effect, consistent with the observation that mDispB cannot rescue Drosophila disp-/- mutant. Fold induc tion is normalized to the respective cyclopamine controls. View larger version: [In this window] [In new window] An intact response to Shh signaling in mDispA-/- cells indicates that the defect in embryonic patterning could be due to inappropriate cleavage and lipidation o f the Shh protein or alternatively to a defect in presentation or release of the processed Shh protein. To test the ability of mDispA-/- cells to correctly proc ess Shh, we transfected an expression construct for full-length Shh into mDispA- /- cells and compared the mobility of the protein produced to that produced by a n expression construct for ShhN. The Shh expression construct encodes a full-len gth protein that undergoes internal cleavage and modification by cholesterol at the newly formed C terminus followed by palmitate addition at the N terminus (Sh hNp). The ShhN construct contains a stop codon following the site of internal cl eavage and therefore produces a protein (ShhN) containing the same amino acid re sidues as the processed protein but lacking cholesterol modification. We found t hat Shh protein expressed in mDispA-/- cells was efficiently cleaved and that it s electrophoretic mobility was slightly greater than that of ShhN (Figure 6C), i ndicative of normal cholesterol modification (Porter et al., 1996b ). This mobil ity was not altered by cotransfection with expression constructs for mDispA or m DispB, and these results suggest that the signaling defect in mDispA-/- embryos is not due to a defect in Shh processing. Defective Export of ShhNp from mDispA-/- Cells To explore whether the mDispA-/- signaling defect is due to an inability to pres ent or release a signaling-competent form of ShhNp, we transfected mDispA-/- mut ant cells with expression constructs for full-length Shh or for ShhN and mixed t hese cells with stably transfected cells carrying the Gli-luc reporter (Shh-LIGH T2 cells; Taipale et al., 2000 ). We found that the ShhN-transfected cells induc ed a greater response in the Shh-LIGHT2 reporter cells (Figure 6D), consistent w ith previous studies demonstrating that ShhN protein is readily released from ce lls (Porter et al., 1996a ). We tested the ability of Disp proteins to increase the amount of ShhNp available for signaling by cotransfecting the Shh expression construct into mDispA-/- cells with a construct for expression of mDispA or mDi spB. We found that mDispA but not mDispB expression significantly increased the response of Shh-LIGHT2 cells (Figure 6D), suggesting that mDispA expression can increase the amount of signaling-competent Shh protein available to other cells. These experiments did not, however, distinguish between the possibilities that mDispA mediates release of ShhNp or instead mediates more efficacious presentati on to responding cells on the surface of presenting cells. To directly measure release of processed Hh proteins, we inserted coding sequenc es for Renilla luciferase in-frame into the amino-terminal signaling domains of full-length Hh and Shh coding sequences (Figure 7A). The point of insertion was a peripheral loop within the structure of ShhN (Hall et al., 1995 ), and the abi lity of Hh-Rluc and Shh-Rluc proteins to undergo processing was preserved (data not shown). We then transfected expression constructs for these modified Hh prot eins together with expression constructs for Disp proteins and assayed the relea se of Renilla luciferase activity into the culture medium. Upon testing of sever al cell lines for Disp protein enhancement of protein export, the best results f or Hh-Rluc and Shh-Rluc proteins were obtained in Drosophila S2 cells. In typica l assays (Figure 7B), Disp proteins increased export efficiency ~3–10-fold, and enhancement of export by mouse or fly Disp was greatest for the Hh protein of the corresponding species. No enhancement of export was observed for mDispB (Figure 7B). These results indicate that Disp and mDispA, but not mDispB, can mediate the release and extracellular accumulation of Hh proteins in soluble form, and further suggest that the physiological role of Disp activity is to release Hh proteins from cells. Figure 7. Transporter-Like Activity of Dispatched Exports Hh from the Producing Cell in So luble Form (A) The structure of Hh-Renilla luciferase fusion proteins used in the export as say. A secretable form of Renilla luciferase was generated as in Liu and Escher, 1999 , and inserted into homologous sites in Shh and Hh to generate the Shh- an d Hh-Renilla luciferase fusion proteins (Shh- and Hh-Rluc). (B) Hh export assay in S2 cells. Both Drosophila Disp and mDispA increase secret ion of coexpressed Hh-Rluc (blue bars) or Shh-Rluc (red bars) into the culture m edium, whereas mDispB does not. (C) Alignment of the GxxxD motif sequences of Drosophila Disp, mDispA, and mDisp B. Residues mutated in the NNN and AAA mutant forms of both Drosophila Disp and mDispA are also indicated. (D) Both the NNN and AAA mutant forms of Disp and mDispA are inactive (compare t o wild-type protein and GFP control). Results from Hh and Shh export assays are shown for Drosophila Disp (blue bars) and mDispA (red bars), respectively. (E–H) Drosophila dispAAA and dispNNN fail to rescue disp mutant phenotype. In disp wing disc, ptc-lacZ expression is limited to 1?C2 cells wide region anterior to the compartment boundary (E). en-GAL4 driven ex pression of wild-type Disp (F), but not DispNNN (G) or DispAAA (H), restores Lac Z expression to a region comparable to that found in wild-type discs (compare F to Figure 2F). View larger version: [In this window] [In new window] Functional Conservation in Disp Proteins of Residues Critical for Function of RN D Transmembrane Transporters Sequence comparison of Disp to other proteins in the database clearly shows sequ ence similarity not only to Ptc (Burke et al., 1999 ), but also to the prokaryot ic RND permease superfamily. The majority of these proteins are prokaryotic effl ux pumps involved in conferring resistance to drugs or heavy metals or in the se cretion of endogenous molecules. These proteins appear to have arisen by tandem duplication of a six transmembrane unit (Tseng et al., 1999 ) to give rise to th e twelve transmembrane spans of the full structure, including large extracellula r loops at homologous positions in the two units, between TM1 and TM2, and betwe en TM7 and TM8 (see Figure 1B). Biochemical studies have indicated that the RND superfamily proteins function as proton-driven antiporters (Nies, 1995 ; Tseng e t al., 1999 ). Members of the RND family of proteins have a conserved GxxxD moti f in the middle of TM4 (Tseng et al., 1999 ), and the aspartate residue within t his motif is important for antiporter function and has been proposed to be the p roton binding site (Goldberg et al., 1999 ). An expanded form of this motif, Gxx xDD, is present in the middle of TM4 of Disp and mDispA, and the GxxxD motif is also present within TM10, the homologous position within the presumptive intramo lecular duplication; mDispB in contrast lacks Asp within TM4 (Figure 7C). The presence of three acidic Asp residues in the middle of the bilayer in Disp a nd in mDispA and the experimentally demonstrated requirement for one of these re sidues in the function of bacterial metal efflux pumps (Goldberg et al., 1999 ) suggests the possibility that Disp and the bacterial proteins might act by a sim ilar mechanism. Consistent with this possibility, 2 of 3 of the Asp residues con served in Disp and mDispA are absent from mDispB, which is unable to complement Drosophila disp mutations and unable to release Hh proteins in cultured cell ass ays. To further test whether these conserved Asp residues are important for the function of Disp proteins, we introduced Asn or Ala substitutions at these sites and tested expression constructs encoding these altered proteins for their abil ity to export Hh proteins from S2 cells. We found that neither Disp nor mDispA i n either of their altered forms were able to increase export efficiency of Hh pr oteins when tested in the S2 cell assay (Figure 7D). To further test these alter ed proteins and to validate the S2 cell export assay, we tested both altered Dis p proteins for their ability to rescue the Drosophila disp mutation in vivo. Sim ilarly, we found that neither Disp protein carrying the Asn or the Ala substitut ions was able to rescue disp mutant function in the wing imaginal disc ( Figures 7G–7H, compare to E–F). Conclusions Summary Introduction Results and Discussion Conclusions Experimental Procedures References Perhaps the most striking and unexpected aspect of our results is the extreme na ture of the pattern disruptions in mDispA-/- embryos. The mDispA-/- mutant pheno type is more severe than that of mutations in any single gene encoding a Hh prot ein (Bitgood et al., 1996 ; Chiang et al., 1996 ; St-Jacques et al., 1999 ). In contrast, the Drosophila disp mutant phenotype is less severe than the hh phenot ype (Burke et al., 1999 ). This discrepancy appears in part due to disruption of signaling by multiple Hh proteins, as the mDispA-/- phenotype resembles that of Smo-/- and that of the Shh-/-; Ihh-/- double mutant (Zhang et al., 2001 ). Howe ver, the phenotype also appears to owe its severity, at least in part, to a dist inct balance in the relative importance of long-range and short-range signaling in Drosophila and in the mouse. In Drosophila disp mutants, short-range Hh signa ling is intact and contributes to maintenance of target gene expression and to p atterning. In mDispA-/- embryos, some signal response is retained in cells that express Shh (Ptch-lacZ is expressed in the notochord), but this response appears not to contribute to morphological pattern. The null function phenotype for the mDispA gene thus permits dissection of the relative importance of long- and sho rt-range Hh signaling in mouse embryos and reveals a near absolute dependence on long-range signaling in patterning of the mouse embryo. Mechanisms and processes proposed to play a specific role in long-range signalin g (reviewed in Teleman et al., 2001 ) include the specific transport of protein signals to distant cells by movement through intervening cells (transcytosis), t he sensing of signals produced at distant locations by long cytoplasmic extensio ns (cytonemes), and the movement of signaling proteins through tissues in membra ne fragments termed argosomes (Greco et al., 2001 ). We can not rule out a role for Disp in helping to form these structures or mediate these processes, and the se mechanisms may very well play some role in long-range action by other signali ng proteins. However, the ability of Disp proteins to release soluble Hh protein s into the medium of cultured cells presents the relatively simple alternative o f catalyzed signal release as a primary mechanism for initiating communication w ith distant cells. The severe phenotype of the mDispA-/- mutant indicates that little Hh protein is released in the absence of Disp activity. This is perhaps not surprising, as si ngle lipid modification of Hh protein fosters some association with cell membran es and dual lipid modification nearly quantitative association (Chamoun et al., 2001 ), presumably by energetically favorable insertion of the cholesteryl and p almitoyl adducts into the lipid bilayer. Membrane release of Hh protein represen ts a novel activity for bacterial RND transporters, the family to which mDispA b elongs. A precedent for transmembrane molecular transporter acting in release of lipid-modified proteins from membranes is that of the bacterial ATP binding cas sette (ABC) transporter, the LolCDE complex of E. coli (Yakushi et al., 2000 ). Membrane release of lipoproteins also represents a novel activity for ABC transp orters, as other members of this family are involved in transport of substrates across the membrane (Linton and Higgins, 1998 ). Like these other ABC transporte rs, activity of the LolCDE complex is coupled to hydrolysis of ATP, which presum ably supplies the energy required for membrane release. Transmembrane transport by RND family members also requires energy in the form of an electrochemical gra dient, and it will be interesting to learn whether release of Hh by Disp also de pends upon such a gradient, perhaps with the use of our cultured cell system tha t releases soluble Hh protein. Physical characterization of protein from this sy stem should also illuminate the question of why soluble lipidated Hh protein in the medium does not reassociate with cells and should permit comparison to other soluble forms of lipidated Hh protein (Zeng et al., 2001 ). Finally, it is interesting to note that another RND family member, the Ptc prote in, is a Hh pathway component that plays a central role not in Hh signal release , but in regulating response to the Hh signal by modulating the activity of Smo. Like Disp, Ptc proteins also contain a functionally conserved GxxxD motif withi n the predicted TM4 domain, and this motif is affected in three of the six known missense mutations that cause Basal Cell Nevus Syndrome (also Gorlin's syndrome ) (Taipale et al., 2002 ). The Ptc protein thus also appears to function by a me chanism similar to that of Disp and RND transporters, although unlike Disp, its substrate remains unclear. These two members of the RND family with apparently s imilar mechanisms thus appear to have evolved to play essential yet very differe nt roles in the Hh signaling pathway. Experimental Procedures Summary Introduction Results and Discussion Conclusions Experimental Procedures References Cloning and Sequence Alignments Expressed sequence tags (ESTs) corresponding to the two murine disp homologs wer e identified by database search, and the initially identified ESTs were used to screen mouse cDNA libraries. mDispA cDNA was isolated as a single cDNA clone fro m a mouse testis cDNA library, and the full-length mDispB cDNA was assembled fro m two independent clones isolated from a mouse lung cDNA library. The accuracy o f the full-length sequences was confirmed both by PCR reactions and intron/exon organization. The sequence alignment was generated using Pileup (GCG software pa ckage) and the identities between the sequences were marked by the SeqVu program . Drosophila Stocks and Immunostaining The disp mutant (l(3)S037707), the UAS-disp transgene, the en-Gal4 driver, and t he ptc-lacZ reporter gene are described in Burke et al., 1999. mDispA, mDispB, m yc-tagged dispNNN, and dispAAA sequences were cloned into pUAST vector (Brand an d Perrimon, 1993 ) and the corresponding UAS-transgenes were generated through P -element mediated germline transformation. disp-/- larvae for dissection were id entified by the absence of the Tubby genotype present on the balancer chromosome TM6b. Third instar wing disc fixation and fluorescence labeling was performed a s in Burke et al., 1999. Generation and Identification of mDispA-/- and Smo-/- Mutant Alleles The last coding exon of the mDispA gene was replaced by a loxP-flanked PGK-neo c assette through homologous recombination in the R1 ES cell line. Correctly targe ted ES clones were identified by genomic Southern blot (See Supplemental Data av ailable at http://www.cell.com/cgi/content/full/111/1/63/DC1) and chimeras were generated by blastocyst injection of C57/BL6 hosts. Germline transmission was co nfirmed by genotyping PCR reactions and the mutant allele was maintained in a C5 7/BL6;129/SvJ mixed genetic background. The Smo-/- mutant allele was generated b y replacing part of the first coding exon of Smo gene including the start codon and the signal peptide by a loxP site through CRE-mediated site-specific recombi nation of a conditional Smo allele (to be described elsewhere). Segregation of t he phenotypes with the genotypes was established using genotyping PCR reactions (See Supplemental Data available at above URL). Probes, In Situ Hybridization, and -Galactosidase Histochemistry In situ hybridization and -galactosidase histochemistry were performed essential ly as described (Hogan et al., 1994 ; Knecht et al., 1995 ). Ptch-lacZ allele wa s from Dr. M. Scott. The in situ hybridization probes are as follows: two probes , AA(V3-L184) and AA(R660-Y901), were used for mDispA which gave essentially the same pattern; Myf5 probe was provided by Dr. R. Emerson, and Pax1 and Pax7 prob es were from Dr. C. Fan. HNF3(3'-UTR), Shh(AA L373-437), and Nodal(3'-UTR) probe s were generated by PCR reactions with mouse genomic DNA as template. Cell Culture Based Assays mDispA-/- and mDispA+/- embryonic fibroblasts were isolated directly from E8.75 embryos and maintained in DMEM supplemented with 10% fetal bovine serum (FBS). D rosophila S2 cells were maintained in Schneider's Drosophila Medium supplemented with 10% FBS. Fibroblasts and S2 cells were transfected using Fugene 6 (Roche) and calcium phosphate coprecipitation, respectively. For reporter assays, cell c ulture medium was replaced 48 hr after transfection with fresh low serum medium (DMEM with 0.5% bovine calf serum and 5 mM HEPES buffer, [pH 7.4]) containing 10 nM ShhNp and/or 5 mM cyclopamine where indicated, and after 72 hr further incub ation, the cells were lysed and reporter activities measured (Dual Luciferase, P romega). For Hh/Shh-Rluc export assays, cells were collected by centrifugation 3 days after transfection and lysed directly in Passive Lysis Buffer (Promega). C onditioned medium was further cleared by centrifugation at 21,000 × g, and repo rter activities in the cell lysates and conditioned medium were assayed as above . Similar results were observed using ultracentrifugation (100,000 × g for 1 hr ). The relative export efficiency index was calculated using the following formu la: Acknowledgments We would like to thank Drs. S.-J. Lee, C. Blobel, C.-M. Fan, and C. Emerson for R1 ES cells, libraries, and probes; Dr. M. Scott for the Ptch-LacZ mice; M. Cowa n for blastocyst injections; and K. Young for technical assistance. We also than k Drs. S.-J. Lee, J. Nathans, G. Seydoux, and C. Machamer and members of their l aboratories as well as members of the Beachy and Basler laboratories for suggest ions. This work was supported by an NIH grant and by the Swiss National Science Foundation. P.A.B. is an investigator of the Howard Hughes Medical Institute. Accession Numbers Accession numbers are AY150698 for mDispA and AY150699 for mDispB. References Summary Introduction Results and Discussion Conclusions Experimental Procedures References Alcedo J., Ayzenzon M., Von Ohlen T., Noll M. and Hooper J.E. (1996) The Drosoph ila smoothened gene encodes a seven-pass membrane protein, a putative receptor f or the hedgehog signal. Cell, 86:221-232. [Medline][Summary][Full Text] Basler K. and Struhl G. (1994) Compartment boundaries and the control of Drosoph ila limb pattern by hedgehog protein. Nature, 368:208-214. [Medline] Bitgood M.J., Shen L. and McMahon A.P. (1996) Sertoli cell signaling by desert h edgehog regulates the male germline. Curr. Biol., 6:298-304. [Medline][BMN Summa ry][BMN Full Text] Borycki A.G., Brunk B., Tajbakhsh S., Buckingham M., Chiang C. and Emerson C.P. Jr. (1999) Sonic hedgehog controls epaxial muscle determination through Myf5 act ivation. Development, 126:4053-4063. [Medline] Brand A.H. and Perrimon N. (1993) Targeted gene expression as a means of alterin g cell fates and generating dominant phenotypes. Development, 118:401-415. [Medl ine] Briscoe J., Chen Y., Jessell T.M. and Struhl G. (2001) A hedgehog-insensitive fo rm of patched provides evidence for direct long-range morphogen activity of soni c hedgehog in the neural tube. Mol. Cell, 7:1279-1291.[Summary][Full Text] Burke R., Nellen D., Bellotto M., Hafen E., Senti K.A., Dickson B.J. and Basler K. (1999) Dispatched, a novel sterol-sensing domain protein dedicated to the rel ease of cholesterol-modified hedgehog from signaling cells. Cell, 99:803-815. [M edline][Summary][Full Text] Capdevila J. and Guerrero I. (1994) Targeted expression of the signaling molecul e decapentaplegic induces pattern duplications and growth alterations in Drosoph ila wings. EMBO J., 13:4459-4468. [Medline] Chamoun Z., Mann R.K., Nellen D., von Kessler D.P., Bellotto M., Beachy P.A. and Basler K. (2001) Skinny hedgehog, an acyltransferase required for palmitoylatio n and activity of the hedgehog signal. Science, 293:2080-2084. [Medline] Chen W., Burgess S. and Hopkins N. (2001) Analysis of the zebrafish smoothened m utant reveals conserved and divergent functions of hedgehog activity. Developmen t, 128:2385-2396. [Medline] Chiang C., Litingtung Y., Lee E., Young K.E., Corden J.L., Westphal H. and Beach y P.A. (1996) Cyclopia and defective axial patterning in mice lacking sonic hedg ehog gene function. Nature, 383:407-413. [Medline] Collignon J., Varlet I. and Robertson E.J. (1996) Relationship between asymmetri c nodal expression and the direction of embryonic turning. Nature, 381:155-158. [Medline] Cooper M.K., Porter J.A., Young K.E. and Beachy P.A. (1998) Teratogen-mediated i nhibition of target tissue response to Shh signaling. Science, 280:1603-1607. [M edline] Davies J.P. and Ioannou Y.A. (2000) Topological analysis of Niemann-Pick C1 prot ein reveals that the membrane orientation of the putative sterol-sensing domain is identical to those of 3-hydroxy-3-methylglutaryl-CoA reductase and sterol reg ulatory element binding protein cleavage-activating protein. J. Biol. Chem., 275 :24367-24374. [Medline] Dudley A.T. and Tabin C.J. (2000) Constructive antagonism in limb development. C urr. Opin. Genet. Dev., 10:387-392. [Medline] Ericson J., Morton S., Kawakami A., Roelink H. and Jessell T.M. (1996) Two criti cal periods of sonic hedgehog signaling required for the specification of motor neuron identity. Cell, 87:661-673. [Medline][Summary][Full Text] Fan C.M. and Tessier-Lavigne M. (1994) Patterning of mammalian somites by surfac e ectoderm and notochord: evidence for sclerotome induction by a hedgehog homolo g. Cell, 79:1175-1186. [Medline][Summary] Fan C.M., Porter J.A., Chiang C., Chang D.T., Beachy P.A. and Tessier-Lavigne M. (1995) Long-range sclerotome induction by sonic hedgehog: direct role of the am ino-terminal cleavage product and modulation by the cyclic AMP signaling pathway . Cell, 81:457-465. [Medline][Summary] Fuse N., Maiti T., Wang B., Porter J.A., Hall T.M., Leahy D.J. and Beachy P.A. ( 1999) Sonic hedgehog protein signals not as a hydrolytic enzyme but as an appare nt ligand for patched. Proc. Natl. Acad. Sci. USA, 96:10992-10999. [Medline] Goldberg M., Pribyl T., Juhnke S. and Nies D.H. (1999) Energetics and topology o f CzcA, a cation/proton antiporter of the resistance-nodulation-cell division pr otein family. J. Biol. Chem., 274:26065-26070. [Medline] Goodrich L.V., Johnson R.L., Milenkovic L., McMahon J.A. and Scott M.P. (1996) C onservation of the hedgehog/patched signaling pathway from flies to mice: induct ion of a mouse patched gene by hedgehog. Genes Dev., 10:301-312. [Medline] Goodrich L.V., Milenkovic L., Higgins K.M. and Scott M.P. (1997) Altered neural cell fates and medulloblastoma in mouse patched mutants. Science, 277:1109-1113. [Medline] Greco V., Hannus M. and Eaton S. (2001) Argosomes: a potential vehicle for the s pread of morphogens through epithelia. Cell, 106:633-645. [Medline][Summary][Ful l Text] Gurdon J.B. and Bourillot P.Y. (2001) Morphogen gradient interpretation. Nature, 413:797-803. [Medline] Gustafsson M.K., Pan H., Pinney D.F., Liu Y., Lewandowski A., Epstein D.J. and E merson C.P. Jr. (2002) Myf5 is a direct target of long-range Shh signaling and G li regulation for muscle specification. Genes Dev., 16:114-126. [Medline] Hall T.M., Porter J.A., Beachy P.A. and Leahy D.J. (1995) A potential catalytic site revealed by the 1.7-A crystal structure of the amino-terminal signalling do main of sonic hedgehog. Nature, 378:212-216. [Medline] Hidalgo A. and Ingham P. (1990) Cell patterning in the Drosophila segment: spati al regulation of the segment polarity gene patched. Development, 110:291-301. [M edline] Hogan B., Beddington R., Costantini F. and Lacy E. Manipulating the Mouse Embryo : A Laboratory Manual. (1994) (Second Edition Edition)Cold Spring Harbor, NY: Co ld Spring Harbor Laboratory Press Hooper J.E. and Scott M.P. (1989) The Drosophila patched gene encodes a putative membrane protein required for segmental patterning. Cell, 59:751-765. [Medline] [Summary] Incardona J.P., Gaffield W., Kapur R.P. and Roelink H. (1998) The teratogenic ve ratrum alkaloid cyclopamine inhibits sonic hedgehog signal transduction. Develop ment, 125:3553-3562. [Medline] Ingham P.W. and McMahon A.P. (2001) Hedgehog signaling in animal development: pa radigms and principles. Genes Dev., 15:3059-3087. [Medline] Jessell T.M. (2000) Neuronal specification in the spinal cord: inductive signals and transcriptional codes. Nat. Rev. Genet., 1:20-29. Johnson R.L., Laufer E., Riddle R.D. and Tabin C. (1994) Ectopic expression of s onic hedgehog alters dorsal-ventral patterning of somites. Cell, 79:1165-1173. [ Medline][Summary] Knecht A.K., Good P.J., Dawid I.B. and Harland R.M. (1995) Dorsal-ventral patter ning and differentiation of noggin-induced neural tissue in the absence of mesod erm. Development, 121:1927-1935. [Medline] Lee J.D. and Treisman J.E. (2001) Sightless has homology to transmembrane acyltr ansferases and is required to generate active hedgehog protein. Curr. Biol., 11: 1147-1152. [Medline][Summary][Full Text] Lee J.J., Ekker S.C., von Kessler D.P., Porter J.A., Sun B.I. and Beachy P.A. (1 994) Autoproteolysis in hedgehog protein biogenesis. Science, 266:1528-1537. [Me dline] Lewis P.M., Dunn M.P., McMahon J.A., Logan M., Martin J.F., St-Jacques B. and Mc Mahon A.P. (2001) Cholesterol modification of sonic hedgehog is required for lon g-range signaling activity and effective modulation of signaling by Ptc1. Cell, 105:599-612. [Medline][Summary][Full Text] Linton K.J. and Higgins C.F. (1998) The Escherichia coli ATP-binding cassette (A BC) proteins. Mol. Microbiol., 28:5-13. [Medline] Liu J. and Escher A. (1999) Improved assay sensitivity of an engineered secreted Renilla luciferase. Gene, 237:153-159. [Medline] Lowe L.A., Supp D.M., Sampath K., Yokoyama T., Wright C.V., Potter S.S., Overbee k P. and Kuehn M.R. (1996) Conserved left-right asymmetry of nodal expression an d alterations in murine situs inversus. Nature, 381:158-161. [Medline] Marigo V., Davey R.A., Zuo Y., Cunningham J.M. and Tabin C.J. (1996) Biochemical evidence that patched is the hedgehog receptor. Nature, 384:176-179. [Medline] Marigo V., Scott M.P., Johnson R.L., Goodrich L.V. and Tabin C.J. (1996) Conserv ation in hedgehog signaling: induction of a chicken patched homolog by sonic hed gehog in the developing limb. Development, 122:1225-1233. [Medline] Muenke M. and Beachy P.A. (2001) Holoprosencephaly. In: Scriver C.R., Beaudet A. L., Sly W.S., Valle D., Childs B., Kinzler K.W. and Vogelstein B. (Eds.) The Met abolic and Molecular Bases of Inherited Disease. New York: McGraw-Hill Nakano Y., Guerrero I., Hidalgo A., Taylor A., Whittle J.R. and Ingham P.W. (198 9) A protein with several possible membrane-spanning domains encoded by the Dros ophila segment polarity gene patched. Nature, 341:508-513. [Medline] Nies D.H. (1995) The cobalt, zinc, and cadmium efflux system CzcABC from Alcalig enes eutrophus functions as a cation-proton antiporter in Escherichia coli J. Ba cteriol., 177:2707-2712. [Medline] Pepinsky R.B., Zeng C., Wen D., Rayhorn P., Baker D.P., Williams K.P., Bixler S. A., Ambrose C.M., Garber E.A. and Miatkowski K. et al. (1998) Identification of a palmitic acid-modified form of human sonic hedgehog. J. Biol. Chem., 273:14037 -14045. [Medline] Porter J.A., Ekker S.C., Park W.J., von Kessler D.P., Young K.E., Chen C.H., Ma Y., Woods A.S., Cotter R.J., Koonin E.V. and Beachy P.A. (1996) Hedgehog pattern ing activity: role of a lipophilic modification mediated by the carboxy-terminal autoprocessing domain. Cell, 86:21-34. [Medline][Summary][Full Text] Porter J.A., Young K.E. and Beachy P.A. (1996) Cholesterol modification of hedge hog signaling proteins in animal development. Science, 274:255-259. [Medline] Roelink H., Porter J.A., Chiang C., Tanabe Y., Chang D.T., Beachy P.A. and Jesse ll T.M. (1995) Floor plate and motor neuron induction by different concentration s of the amino-terminal cleavage product of sonic hedgehog autoproteolysis. Cell , 81:445-455. [Medline][Summary] Roessler E., Belloni E., Gaudenz K., Jay P., Berta P., Scherer S.W., Tsui L.C. a nd Muenke M. (1996) Mutations in the human sonic hedgehog gene cause holoprosenc ephaly. Nat. Genet., 14:357-360. [Medline] Sasaki H., Hui C., Nakafuku M. and Kondoh H. (1997) A binding site for Gli prote ins is essential for HNF-3 floor plate enhancer activity in transgenics and can respond to Shh in vitro. Development, 124:1313-1322. [Medline] St-Jacques B., Hammerschmidt M. and McMahon A.P. (1999) Indian hedgehog signalin g regulates proliferation and differentiation of chondrocytes and is essential f or bone formation. Genes Dev., 13:2072-2086. [Medline] Stone D.M., Hynes M., Armanini M., Swanson T.A., Gu Q., Johnson R.L., Scott M.P. , Pennica D., Goddard A. and Phillips H. et al. (1996) The tumour-suppressor gen e patched encodes a candidate receptor for sonic hedgehog. Nature, 384:129-134. [Medline] Taipale J., Chen J.K., Cooper M.K., Wang B., Mann R.K., Milenkovic L., Scott M.P . and Beachy P.A. (2000) Effects of oncogenic mutations in Smoothened and Patche d can be reversed by cyclopamine. Nature, 406:1005-1009. [Medline] Taipale J., Cooper M.K., Maiti T. and Beachy P.A. (2002) Patched acts catalytica lly to suppress the activity of Smoothened. Nature, 418:892-896. [Medline] Teleman A.A., Strigini M. and Cohen S.M. (2001) Shaping morphogen gradients. Cel l, 105:559-562. [Medline][Full Text] Tseng T.T., Gratwick K.S., Kollman J., Park D., Nies D.H., Goffeau A. and Saier M.H. Jr. (1999) The RND permease superfamily: an ancient, ubiquitous and diverse family that includes human disease and development proteins. J Mol. Microbiol. Biotechnol., 1:107-125. van den Heuvel M. and Ingham P.W. (1996) Smoothened encodes a receptor-like serp entine protein required for hedgehog signalling. Nature, 382:547-551. [Medline] Wang B., Fallon J.F. and Beachy P.A. (2000) Hedgehog-regulated processing of Gli 3 produces an anterior/posterior repressor gradient in the developing vertebrate limb. Cell, 100:423-434. [Medline][Summary][Full Text] Yakushi T., Masuda K., Narita S., Matsuyama S. and Tokuda H. (2000) A new ABC tr ansporter mediating the detachment of lipid-modified proteins from membranes. Na t. Cell Biol., 2:212-218. Zeng X., Goetz J.A., Suber L.M., Scott W.J. Jr., Schreiner C.M. and Robbins D.J. (2001) A freely diffusible form of sonic hedgehog mediates long-range signallin g. Nature, 411:716-720. [Medline] Zhang X.M., Ramalho-Santos M. and McMahon A.P. (2001) Smoothened mutants rev