| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
THEMES
1Department of Pharmacology and Toxicology, Dartmouth Medical School, Hanover, New Hampshire; and 2Norris Cotton Cancer Center, Dartmouth-Hitchcock Medical Center, Lebanon, New Hampshire
Submitted 4 December 2007 ; accepted in final form 28 January 2008
 |
ABSTRACT |
|---|
 TOP ABSTRACT GRANTSREFERENCES |
|---|
morphogen; review
Hh was originally discovered in a classic genetic screen performed by Nusslein-Volhard and Wieschaus, from which they identified mutations that disrupt the cuticle patterning of the Drosophila embryo (12). Mutation of Hh gives the Drosophila embryo a dramatic "spiny" phenotype reminiscent of the ubiquitous old-world mammal, the hedgehog. Three vertebrate Hh homologs were subsequently identified, Sonic hedgehog (Shh), Indian hedgehog (Ihh), and Desert hedgehog (Dhh) (12, 26). Although there is some degree of overlap, the expression of Shh is quite widespread whereas Ihh and Dhh expression appear to be more tissue specific. Shh-null animals have a multitude of defects including cyclopia, poor neural patterning, lack of limb growth, and abnormal organ and foregut development (12). Ihh-null mice are most noticeably deficient in aspects of bone development, whereas Dhh-null mice exhibit defects in spermatogenesis and in nerve myelin sheath development (12). Mice deficient for both Ihh and Shh die much earlier than mice lacking either gene product on its own, consistent with the various family members having some redundant functions. In general, the processing and signaling of Hh family members is similar across the various mammalian paralogs and is also evolutionarily conserved.
Production and Lipid Modification of Hh
Hh is initially synthesized as a preprotein, which subsequently enters the secretory pathway where the signal sequence is removed to yield an 
45-kDa precursor protein (12, 26). This full-length protein is subsequently cleaved into two discrete peptides (15) at a highly conserved three amino acid motif, Gly197Cys198Phe (human Shh numbering), to yield an amino terminal domain (HhN) and a carboxy terminal domain (HhC) (19; reviewed in Ref. 17). This internal proteolytic event results in the concomitant covalent attachment of a cholesterol moiety to a Gly197 residue in the newly formed carboxy terminus of HhN, resulting in HhNp (where "p" stands for processed) (Fig. 1) (20). This processing of full-length Hh occurs in an intramolecular manner, with the HhC domain acting as a cholesterol transferase. Both the internal cleavage and covalent modification by cholesterol occur as result of two specific nucleophilic displacement reactions (20). First, the sulfur that resides within the side chain of the conserved Cys198 attacks the carbonyl of the adjacent Gly197, causing an internal molecular rearrangement (N to S shift). This shift allows for a second nucleophilic attack to occur on the same carbonyl group by the cholesterol molecule's hydroxyl group, resulting in the cleavage of the bond between the carbonyl and sulfur with the release of HhC and cholesterol modified HhN (Fig. 2). As a result of this reaction, cholesterol is attached to the carboxy terminus of HhN through an ester linkage (20). Hh is the only known secreted ligand that is covalently modified by cholesterol. This unique modification probably results in many of the unprecedented biological mechanisms that have been described in the Hh pathway.
|
|
The majority of Hh found in vivo appears to be modified by both cholesterol and palmitate, although these analyses are far from complete (3, 18). The lipid modifications of Hh serve an important role in regulating its localization and activity (Fig. 1). Cholesterol is essential for coordinating Hh trafficking and restricts Hh movement, as it allows Hh to associate with the plasma membrane where it enriches in lipid rafts (21). The movement of HhN, which is not cholesterol modified, appears to be unrestricted, diffusing over longer distances, whereas HhNp containing the cholesterol moiety is more constrained in its movement (11, 12, 26). Cholesterol-modified HhNp is therefore concentrated and can elicit high-level responses, whereas unmodified HhN produces a more diffuse low-level response, as was predicted by a recent mathematical model of Hh spreading (11). Somewhat paradoxically, cholesterol modification of Hh is also required for its ability to move far from its site of synthesis, because many proteins required for the long-range movement of Hh only regulate the cholesterol-modified form of Hh. The palmitoylation of Hh has been implicated in regulating its activity and more recently has also been implicated in producing a form of Hh used for long-range movement (4, 10). Although it is currently difficult to distinguish between the two proposed roles palmitoylation plays in Hh biology, animals lacking Ski activity have unequivocally demonstrated the importance of palmitoylation to Hh function (3, 4).
Hh Movement
Hh signaling has been well characterized in various developmental processes, from the patterning of the neural tube, to the specification of asymmetrical digits in the embryonic limb bud (12). Each tissue has its own characteristic patterning mechanisms, but the fundamental principles remain similar. In general, Hh is produced and secreted from a specialized group of cells contained within a specific compartment of a tissue (Fig. 2). Hh moves away from the site of its production and across the developing tissue, where it is received by Hh-responsive cells, in effect setting up a morphogen gradient. A cell's position in this Hh gradient differentially specifies cell fate depending on how close the cells are to the source of Hh (12, 26). Hh can signal over short distances to adjacent cells and up to 300 µm away from its site of synthesis, as observed in developing limb buds (12). Hh binding to its cell surface receptor Patched (Ptc) initiates a signaling cascade, which results in the activation of a distinct set of target genes (12, 26). At the boundary of the Hh producing and Hh receiving compartments, where the concentration of Hh is highest, "high-level" target genes are expressed. Further from the Hh source, lower levels of Hh result in the expression of a different set of target genes (11, 12, 26). Ultimately, this differential gene expression pattern is what is thought to regulate the various cell fates within a Hh morphogenic gradient. Thus the regulation of this Hh gradient is of critical importance to the developing tissue, because changes in the shape or extent of the Hh gradient will have dramatic effects on the biological outcome (12, 26).
The movement of Hh across a field of cells is both positively and negatively modulated by a number of cell surface and extracellular players (Fig. 2). The mechanism by which this Hh movement occurs has long been a topic of speculation and controversy. The idea that a lipid-modified molecule could diffuse freely through the hydrophilic environment of the extracellular matrix appears counterintuitive, since lipid modifications usually serve as membrane anchors. Numerous models have been proposed to explain this apparent paradox and have included the transport of Hh via large lipoprotein particles, via planar transcytosis, by active transport, and by direct delivery to receiving cells via long cellular outgrowths termed cytonemes (6). A number of groups have also presented results suggesting that lipid-modified HhNp can form soluble, multimeric structures (s-HhNp) (Fig. 1) (4, 9, 29). Multimeric Hh is found enriched in the media of Hh-producing cells and appears to be asymmetrically secreted in polarized epithelial cells in vivo (9). A structural analysis of an oligomeric form of HhN suggested that a series of intermolecular protein-protein interactions serve to bury palmitate in a hydrophobic pocket within HhN (10). It was later shown that mutation of the amino acids responsible for these intermolecular Hh-Hh interactions, or the palmitate acceptor site, attenuated the formation of multimeric Hh (10). In vitro the HhNp multimeric complex, which has an apparent molecular size of 120 kDa, allows Hh to become more freely diffusible (29). Formation of multimeric Hh thus allows the normally hydrophobic HhNp to propagate a signal far from its site of synthesis. Consistent with this latter model, disruption of these multimeric Hh structures results in loss of long-range Hh signaling (4, 9). Eaton (6) has also described a large lipoprotein complex that aids Hh movement and may be similar to multimeric Hh. Furthermore, Eaton demonstrated that when lipoprotein levels are reduced, Hh accumulates at its site of synthesis and fails to signal over its normal range. These results also support the role lipoprotein particles may play in allowing Hh proteins to establish a morphogenic field.
Release of Hh and Dispatched
In order for fully processed HhNp to signal in a long-range fashion, it must first be released from the cells that produce it. The putative 12-pass membrane protein Dispatched (Disp) has been implicated in the release of cholesterol-modified HhNp (2). Disp contains a sterol-sensing domain and shares homology with a family of prokaryotic resistance-nodulation-cell division (RND) permeases (26). Hh processing occurs normally in Disp–/– cells, but it is unable to efficiently leave the cell (12, 26). Complete loss of Disp function leads to strong Hh loss of function phenotypes, indicating a positive role for Disp in Hh signaling (11, 26). Disp function is only required in Hh producing cells, consistent with its proposed role in HhNp release (Fig. 2) (11, 12, 26). It has also been suggested that Disp regulates the release of HhNp indirectly, by regulating its intracellular trafficking within Hh producing cells (26). HhNp localizes to two distinct regions of polarized Hh producing epithelial cells, near basal or apical membranes (8). This differential Hh localization is consistent with two distinct routes of exit for HhNp, because the two pools of HhNp are affected differently in various genetic backgrounds. Disp plays a pivotal role in the release of at least one of these HhNp pools, since HhNp produced from cells lacking Disp loses the ability to signal over long distances but can still affect cells closest to the HhNp producing cells (8). Expression of HhN can rescue the Disp loss of function phenotype, demonstrating that Disp normally functions to regulate the release of a cholesterol-modified form of Hh capable of signaling far from its site of synthesis (26). Although the use of HhN as a molecular tool is widespread, it is worth mentioning that there is little data to support the relevance of this noncholesterol modified form of Hh in vivo.
Hh Activity
Determining the activity of Hh is complicated by its two lipid modifications, the observation that different modified forms of Hh exhibit differential activity in various biological assays, and the controversy regarding the physiologically relevant form of Hh (Fig. 1). Recombinant HhN, expressed and purified from bacteria, lacks any lipid modifications and is the best characterized form of Hh. This form of HhN binds to Ptc with an approximate binding constant of 2 nM (28) and exhibits activity in a variety of assays, including the ability to induce different cell fates in a dose-dependent manner (23). Recombinant HhN can be acylated in vitro, and this form of acylated HhN has increased activity relative to unmodified HhN (25). The acylated recombinant HhN shows activity similar to unmodified HhN in some tissue explant assays (28), but in other assays it is up to 200 times more active than unmodified HhN (14). Because all Hh appears to be cholesterol modified in vivo, the relevance of these observations with recombinant HhN is not clear, beyond being a good indicator of what HhNp is capable of. However, HhN is palmitoylated when it is expressed in a variety of vertebrate and invertebrate cells, where it appears to exhibit more activity than recombinant unmodified HhN (25). This observation was validated in Ski-null background, which besides attenuating HhNp activity also attenuated HhN activity (3).
HhNp has also been purified from a number of eukaryotic cell lines, in a form that is stoichiometrically cholesterol modified (18, 25). Interestingly, HhNp binds to Ptc with a similar binding constant as recombinant HhN, suggesting that the lipid modifications of HhN do not contribute to its affinity for Ptc (28). However, some differential binding of HhN and HhNp has been described in vivo, implying that other accessory molecules may regulate the affinity of lipid modified HhNp in an in vivo setting. One group compared the activity of HhNp purified in two states, one in which the majority of the HhNp was palmitoylated and one in which the HhNp was only 30% palmitoylated (18). These two highly purified forms of HhNp had similar activity when determined in a limited number of assays. These results initially suggested that either palmitoylation or cholesterol modification were required for high levels of activity and that both lipids were not required for the high-activity form of Hh. The caveat to this interpretation was that this analysis was performed prior to the realization that different biological assays could discriminate Hh activity over a large range of potency. The Hh loss of function phenotypes of animals lacking Ski function support the model in which palmitoylation contributes to Hh activity, although this interpretation is also tempered by the potential role palmitoylation may play in Hh movement. The HhNp form has been shown to form multimers (4, 29) which are quite potent, exhibiting 30 times more activity than HhN expressed in the same cells (4). Multimerization of Hh appears to require both cholesterol and palmitate modification, consistent with the requirement of palmitate modification for high levels of activity (4). A formal comparison of purified multimeric Hh to the other forms of Hh has not yet been performed. However, in general it appears that the lowest activity form of Hh is recombinant HhN and the most active form of Hh is the dual-lipid-modified HhNp, which is multimeric when secreted from Hh producing cells (Figs. 1 and 2).
Extracellular Regulation of Hh
Once Hh is released from its site of synthesis, its activity and movement through a target tissue must be regulated in a temporally and spatially precise manner. A variety of accessory proteins are employed to accomplish this regulation, serving as both positive and negative modulators of Hh function (Fig. 2). Several of these accessory factors appear to be relevant only in specific animals and do not appear to be evolutionarily conserved (22, 27). However, the need for positive and negative modulators of Hh function is evolutionarily conserved, with a variety of different proteins playing similar roles across different phyla. These extracellular Hh modulators function in a distinct manner, promoting or attenuating its movement, acting as coreceptors that specifically increase Hh's potency, and as part of a Hh-inducible feedback mechanism, all of which act to modify and shape Hh's morphogenic field (12, 13, 16, 26).
Heparan sulfate proteoglycans (HSPG), which are large extracellular molecules of heparan sulfate glycosaminoglycan polymers covalently linked to various protein cores, act as regulators of Hh signaling (16, 26). HSPG affect Hh signaling in two distinct ways: indirectly, regulating the extracellular movement of HhNp by some undefined mechanism, and directly, taking part in Hh signal reception by the Hh receiving cell (16, 26). Some of the HSPG implicated in Hh signaling are attached to cell surface membranes through the glycophosphatidylinositol (GPI) anchor of a glypican core protein. Both the glypican core protein and the heparan sulfate glycosaminoglycan polymer attached to this core protein are required for efficient Hh signaling. Hh interacts with HSPG primarily through a highly conserved Cardin-Weintraub motif within its HhN domain, presumably via the electrostatic interaction between the negatively charged sulfates of the HSPG and the localized positively charged amino acids within this motif (24). The first evidence of HSPG involvement in the Hh pathway came from animals lacking members of the Exostosin family of glycosyltransferases, which are required for heparan sulfate polymerization (1). These mutant animals exhibited a Hh-like loss of function phenotype, and in a mosaic tissue analysis HhNp was unable to activate target genes in wild-type tissue if it had to pass through a morphogenic field lacking Exostosin function. Interestingly, this Exostosin requirement was specific for cholesterol modified HhNp, since noncholesterol modified HhN could bypass the Exostosin mutant tissue (16). These results suggested that HSPG are required to facilitate HhNp movement through a morphogenic field. Although some glypicans exhibited similar effects on HhNp movement, Hh receiving cells lacking glypican function also exhibit an attenuated biological response to Hh (16). This latter observation suggested that HSPG also play a direct role in the Hh signal transduction pathway of the Hh receiving cell. HSPG have also been found to interact with other extracellular proteins that mediate interactions with Hh (26). One example of such a specific HSPG regulator, the secreted protein sulfatase 1 (Sulf1), acts as a positive mediator of Hh signaling in the ventral portion of the neural tube. Sulf1 temporally regulates the sulfation state of HSPG to create HhNp binding sites, concentrating the amount of HhNp delivered to specific receiving neural progenitor cells (5).
Besides functioning as the receptor for Hh, Ptc also functions to sequester excess HhNp and target it to lysosomes for degradation (26). This second function of Ptc is part of a feedback loop that regulates the amount of HhNp entering a morphogenic field and is initiated by target cells increasing the transcription of Ptc in response to Hh (12, 26). The increased expression of Ptc is highest close to Hh producing cells, creating a barrier to limit the movement of HhNp through the rest of the morphogenic field (26). Hedgehog-interacting protein (HIP) also serves as a feedback attenuator of Hh movement, which is upregulated in response to Hh (25, 26). The Wnt-inducible growth-arrest specific gene (Gas1) was also thought to directly attenuate signaling by antagonizing the diffusion of Hh (13). However, more recent evidence also suggests a positive role for Gas1 in Hh signaling, as mice that lack Gas1 function have a phenotype consistent with reduced Shh activity (13). This positive role for Gas1 is suggested to be through its ability to enhance HhNp binding to Ptc. The so-called Interference Hedgehog proteins (Ihogs), which include CAM-related downregulated by oncogenes (CDO) and Brother of CDO (BOC) in mammals, also function as positive modulators of Hh signaling (13). This family of membrane proteins modulates Hh signaling primarily through their fibronectin type III domains, which are thought to synergize with Ptc in binding Hh (13). Megalin, which is a member of the low-density lipoprotein family of receptors, also binds to Hh with high affinity and mediates Hh internalization and signaling through an unknown mechanism (7). Consistent with megalin functioning as a positive regulator of Hh signaling, mice lacking megalin function also exhibit a holoproscencephaly-like phenotype reminiscent of Shh-null mice (7).
Concluding Remarks
Regulation of Hh signaling is determined by a multifaceted system that involves an unusual processing mechanism, dual lipid modifications, and a host of extracellular positive and negative regulators of activity and movement. Hh's ability to signal over a range of distances and affect cell fate is a function of its degree of processing and modification, as well as its interaction with these accessory factors. Although the number of biological processes requiring Hh continues to grow, Hh's mechanism of action is only slowly being uncovered, hampered by the unprecedented biology that has developed to deal with the unique covalent cholesterol modification found on Hh. As these mechanisms are worked out, the rational design of therapeutics that take advantage of these unique mechanisms will eventually allow us to specifically treat the growing list of disease states that result from excess or loss of Hh activity.
|
GRANTS |
|---|
|
TOPABSTRACTGRANTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTGRANTSREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
