| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Divisions of Neuroscience and Reproductive Biology (P.M.C., A.U.-A., J.A.J.), Oregon National Primate Research Center (P.M.C., A.U.-A., J.I., J.A.J.), and Departments of Physiology and Pharmacology and Cell and Developmental Biology (P.M.C.), Oregon Health and Science University, Beaverton, Oregon; and Research Unit in Reproductive Medicine, Hospital de Ginecobstetricia "Luis Castelazo Ayala," Instituto Mexicano del Seguro Social, Mexico City, Mexico (P.M.C., A.U.-A.)
Abstract I. Introduction II. Endoplasmic Reticulum Quality Control System and Molecular Chaperones III. Physiology of the Gonadotropin-Releasing Hormone Receptor and Vasopressin Type 2 Receptor Systems A. The Human Gonadotropin-Releasing Hormone Receptor and Vasopressin Type 2 Receptor in Health and Disease B. Lessons from Comparison of the Gonadotropin-Releasing Hormone Receptor and Vasopressin Type 2 Receptor Systems: Selecting Likely Targets for Rescue by Pharmacoperones C. Relation between the Overall Receptor Structure and the Structure of the Ligand Binding Site D. Distribution of Mutations IV. Ligand and Receptor Frequency Modulation in Signaling Systems: Implications for Model Selection and the Timing of Pharmacoperone Administration in Vivo A. Frequency Modulation among Primate Gonadotropin-Releasing Hormone Receptors B. Amino Acids Associated with Control of Plasma Membrane Expression C. Amino Acid Positions Associated with Control of Ligand Binding Affinity D. Ligand Binding Affinity as a Squelch Control E. Receptor Concentration at the Plasma Membrane as a Gain Control V. Do Pharmacoperones Need to Be Present at the Time of Mutant Synthesis? VI. The Dominant-Negative Effect and Receptor Rescue VII. Will Pharmacoperone Drugs Be Species-Specific? Selecting the Correct Models for Drug Development VIII. Endogenous Chaperones as a Potential Site for Therapeutic Intervention IX. In Vitro and in Vivo Studies with Pharmacoperones: How Close Are We to the Transferring of Discoveries from the Laboratory Bench to the Bedside? X. Conclusions
 |
Abstract |
|---|
 Top Next |
|---|
|
I. Introduction |
|---|
|
TopPreviousNext |
|---|
; Sitia and Braakman, 2003). Protein folding is a complex process not only because of the proximity and diversity of proteins that are synthesized but also because the steric character of the nascent protein backbone restricts the spectrum of shapes that may be recognized by a stringent quality control system (QCS). The QCS protects against aberrant cellular activity from misfolded molecules by monitoring protein folding and accumulation (Sanders and Nagy, 2000; Ellgaard and Helenius, 2001; Sitia and Braakman, 2003; Ulloa-Aguirre et al., 2004a). This QCS involves the participation of accessory components known as chaperones. Chaperones are a heterogeneous class of proteins that promote and facilitate folding and assembly. They do so by engaging in association with nascent proteins displaying particular features, such as the unexpected presentation of a hydrophobic plate in an aqueous environment, for example. This is important to prevent aggregation and/or interactions of misfolded proteins with other molecules present in a crowded ER environment and to assist in protein targeting to the Golgi complex or to its final destination within the cell (Hartl and Hayer-Hartl, 2002; Horwich, 2002). If chaperone-assisted protein folding fails, the conformationally defective protein is then targeted for degradation through the polyubiquitination/proteasome pathway. Alternatively, misfolded proteins may aggregate, leading to potentially toxic intracellular accumulation or even to excessive protein accumulation in the plasma with extracellular amyloid deposition (Dobson, 1999; Kopito and Ron, 2000; Forloni et al., 2002; Chiti and Dobson, 2006). Thus, the ER QCS represents a potential site for therapeutic intervention in an array of diseases characterized by conformational aberrations of proteins.
It is becoming well recognized that mutations of receptors, enzymes, and ion channels frequently result in protein misfolding and subsequent retention by the cell's QCS (Tamarappoo and Verkman 1998; Burrows et al., 2000; Janovick et al., 2002; Leaños-Miranda et al., 2002; Ulloa-Aguirre et al., 2003, 2004a,b; Bernier et al., 2004a,b; Ishii et al., 2004; Conn and Janovick, 2005; Loo et al., 2005; Yam et al., 2005; Pastores and Barnett, 2005; Suzuki, 2006; Ulloa-Aguirre and Conn, 2006; Wang et al., 2006). Misfolding can result in protein molecules that retain intrinsic function yet become misrouted within the cell and, for reasons of mislocation only, cease to function normally and result in disease. This observation contrasts with the prior presumption that mutational inactivation always reflects loss of intrinsic function (i.e., a receptor that either fails to recognize a ligand or does not couple productively to its effector). Recognition of this alternate concept immediately presents the therapeutic opportunity to correct misrouting and rescue mutants, thereby restoring function and, potentially, curing disease.
The importance of G-protein coupled receptor (GPCR) trafficking and cell surface membrane expression is emphasized by the array of diseases caused by receptor misfolding (Table 1). This is the case for the autosomal dominant forms of retinitis pigmentosa, X-linked nephrogenic diabetes insipidus, and hypogonadotropic hypogonadism (HH). The functional characterization of mutants that cause retinitis pigmentosa due to ER trapping of misfolded mutant rhodopsin and that eventually lead to photoreceptor degeneration was initially described by Sung et al. (1991, 1993) and thereafter by Kaushal and Khorana (1994). These reports were followed by descriptions of mutant vasopressin type 2 receptors (V2Rs), leading to nephrogenic diabetes insipidus caused by the inability of the mutant receptors to reach the cell surface membrane (Birnbaumer et al., 1994; Tsukaguchi et al., 1995; Wenkert et al., 1996; Sadeghi et al., 1997). More recently, mutations leading to receptor misfolding and resultant misrouting of the gonadotropin hormone-releasing hormone receptor (GnRHR) in patients with HH have been described previously (Janovick et al., 2002; Leaños-Miranda et al., 2002; Ulloa-Aguirre et al., 2004b). There are other GPCRs in which mutations provoke loss of function of the receptor because of intracellular retention of the abnormal (and presumably misfolded and/or incompletely processed) receptor and, consequently, decreased or absent cell surface membrane expression. Some trafficking defective mutants of the glycoprotein hormone receptors [lutropin (LH), follitropin (FSH), and thyrotropin receptors] have been described in patients with Leydig cell hypoplasia, a rare autosomal recessive form of male pseudohermaphroditism (LH receptor) (Gromoll et al., 2002; Martens et al., 2002), in women with ovarian dysgenesis (FSH receptor) (Rannikko et al., 2002; Meduri et al., 2003), and in congenital hypothyroidism (thyrotropin receptor) (Biebermann et al., 1997; Costagliola et al., 1999; Tonacchera et al., 2000, 2004). Loss-of-function mutations in the calcium-sensing receptor due to intracellular retention of the mutant receptor have been found in patients with familial hypocalciuric hypercalcemia (D'Souza-Li et al., 2002). The melanocortin-1 receptor, a major determinant for variations in skin and hair pigmentation, has been found to be mutated at different locations in patients with skin and hair abnormalities and increased susceptibility to skin cancers (Valverde et al., 1996); among the 60 or so mutants described, at least four display decreased cell surface expression (Beaumont et al., 2005). Intracellular retention of mutants from two other melanocortin-related receptors, the melanocortin-3 and melanocortin-4 receptors, which are associated with regulation of fat deposition and energy homeostasis, respectively, have been detected in patients with morbid obesity (Ho and MacKenzie, 1999; Tao and Segaloff, 2003; Tao et al., 2006). Finally, mutations that lead to intracellular trapping in the ER of the endothelin-B receptor and the chemokine receptor 5 have been detected in patients with Hirschsprung's disease or aganglionic megacolon (Tanaka et al., 1998; Fuchs et al., 2001) and in subjects with resistance to human immunodeficiency virus infection (Rana et al., 1997), respectively. It is important to mention that in some cases, particularly in disease states with an autosomal recessive mode of inheritance, the defect in cell surface membrane expression is due to intracellular association of receptors, with a dominant-negative (DN) effect of the misfolded receptor on its wild-type counterpart (see section VI.); this DN effect may limit, or even abrogate, plasma membrane expression (PME) of the normal receptor and thus provoke a loss-of-function disease (Ulloa-Aguirre et al., 2004a). Concurrently, this information suggests that misfolding of GPCRs as a cause of disease may be actually substantially more common than previously recognized.
|
Pharmacological chaperones or "pharmacoperones" are small molecules that enter cells, bind specifically to misfolded mutant proteins, correct their folding, and allow them to escape retention by the cellular QCS (Sitia and Braakman, 2003; Schröder and Kaufman, 2005). In many cases, such molecules were initially identified as peptidomimetic antagonists selected from high thoughput screens (although a priori, they need not be antagonists). For that reason they may come from diverse chemical classes. In the case of the GnRHR, pharmacoperones coming from classes as diverse as erythromycin macrolides, indoles, and quinolones have been identified (Janovick et al., 2003a).
Proteins rescued by pharmacoperones then route to the plasma membrane (PM) (or other site) where they can function normally. In principle, the pharmacoperone rescue approach might apply to an array of human diseases that result from misfolding, among these are cystic fibrosis (Dormer et al., 2001; Galietta et al., 2001; Zhang et al., 2003; Amaral, 2006), HH (Ulloa-Aguirre et al., 2003), nephrogenic diabetes insipidus (Morello and Bichet, 2001; Bernier et al., 2004b; Bichet, 2006), retinitis pigmentosa (Noorwez et al., 2004), hypercholesterolemia, cataracts (Benedek et al.,1999), neurodegenerative diseases [Huntington's, Alzheimer's, and Parkinson's diseases (Heiser et al., 2000; Soto et al., 2000; Forloni et al., 2002; Permanne et al., 2002; Muchowski and Wacker, 2005)], and cancer (Peng et al., 2003). In the case of particular proteins (e.g., the GnRHR, the V2R, and rhodopsin), this approach has succeeded with a striking number of different mutants, supporting the view that pharmacoperones will become powerful ammunition in our therapeutic arsenal (Bernier et al., 2004b). On the other hand, it has also become clear that variable amounts of even some WT GPCRs are misrouted, presumably as a result of misfolding (Petäjä-Repo et al., 2000, 2001; Andersson et al., 2003; Janovick et al., 2003b; Lu et al., 2003, 2004; Cook et al., 2003; Pietilä et al., 2005), suggesting that this level of post-translational control may itself be amenable to pharmacological intervention and provide another level of potential therapeutic intervention (Ulloa-Aguirre et al., 2006).
We have previously reviewed the literature on diseases associated with folding in general, all potential targets for pharmacoperone therapeutics (Castro-Fernández et al., 2005), and now focus on translation of this concept to in vivo models. In this review we specifically focus on what we have learned in cell culture studies that is likely to become useful for controlling trafficking of receptors, ion channels, and enzymes in healthy and disease states. We focus on opportunities for drug development and on the lessons learned from two well characterized models of GPCRs, the GnRHR (Ulloa-Aguirre et al., 2003, 2004a, 2006) and the V2R (Morello et al., 2000, 2001; Petäjä-Repo et al., 2002; Bernier et al., 2004a,b; Bulenger et al., 2005). Before describing the particular pharmacological approaches potentially applicable to misfolded GnRHR and V2R, let us briefly review how the ER QCS works to prevent normal routing of defective proteins.
|
II. Endoplasmic Reticulum Quality Control System and Molecular Chaperones |
|---|
|
TopPreviousNext |
|---|
; Sanders and Nagy, 2000; Trombetta and Parodi, 2003). The ER QCS recognizes specific shapes resulting from protein wriggling and hence defines the routing, intracellular trafficking, and eventually the fate of the nascent protein within the cell (Ellgaard and Helenius, 2001; Cahill et al., 2002). To this end, the ER QCS employs a variety of mechanisms including a complex sorting system that identifies and separates proteins according to their maturation status and the action of specialized folding factors, escort proteins, retention factors, enzymes, and members of major molecular chaperone families. Molecular chaperones are accessory components of the ER QCS that participate in the folding process of newly synthesized proteins (Ellgaard and Helenius, 2001; Sitia and Braakman, 2003). They serve as a control mechanism recognizing, retaining, and targeting misfolded proteins for their eventual degradation. Although the steric character of the protein backbone restricts the spectrum of protein shapes that are recognized by the stringent quality control mechanisms, some features displayed by proteins including exposure of hydrophobic shapes, unpaired cysteines, or immature glycans have been identified as important in chaperone-protein association (Ellgaard and Helenius, 2001); in fact, molecular chaperones possess the ability to recognize misfolded proteins by the exposure of hidden hydrophobic domains or particular motifs (Tan et al., 2004; Dong et al., 2007). Through this association, chaperones attempt to stabilize unstable conformers of nascent polypeptides to prevent aggregation and facilitate correct folding or assembly of the substrate via binding and release cycles (Hartl and Hayer-Hartl, 2002). If the polypeptide chain fails to fold properly, then the incorrectly manufactured protein is targeted to the proteasomes for destruction (Werner et al., 1996; Schubert et al., 2000). Several GPCR interacting proteins that support trafficking to the cell surface have been identified. Nina A (neither inactivation nor afterpotential A) is a molecular chaperone whose absence in Drosophila melanogaster rhodopsins leads to rhodopsin 1 ER accumulation and degradation (Schneuwly et al., 1989; Shieh et al., 1989; Colley et al., 1991; Baker et al., 1994); its mammalian homolog RanBP2 binds red/green opsin molecules and acts as a chaperone aiding proper folding, transport, and localization of the mature receptors to the cell membrane (Ferreira et al., 1996). ODR4 is a molecular chaperone that assists in folding, ER exit, and/or targeting of the olfactory receptors ODR10 to olfactory cilia in the nematode Caenorhabditis elegans (Dwyer et al., 1998). Calnexin and calreticulin are molecular chaperones that bind a broad range of glycoproteins, including several GPCRs (e.g., the GnRHR, V2R, and LH, FSH, and thyrotropin receptors) (Helenius et al., 1997; Schrag et al., 2003; Vassilakos et al., 1998; Rozell et al., 1998; Morello et al., 2001; Brothers et al., 2006). The action of these chaperones predominantly centers on substrate N-glycans present on the newly synthesized proteins, adding hydrophobicity to the folding protein (Helenius et al., 1997; Schrag et al., 2003). When N-glycosylation or early glycan processing fails (due to mutations in the glycosylation sites of the receptor, for example), glycoproteins misfold, aggregate, and fail the QCS. This is the case, for example, of the V2R R337X mutant, in which an extended interaction between calnexin and the mutant receptor is involved in ER retention and the absence of cell surface membrane expression of the mutant receptor (Morello et al., 2001). Other molecular chaperones that aid GPCRs to reach the cell surface membrane have been described. These include RAMPs (receptor activity modifying proteins), which interact with several GPCRs [e.g., the calcitonin receptor-like receptor, the vasoactive intestinal polypeptide/pituitary adenylate cyclase-activating peptide receptor, the glucagon receptor, and the parathyroid hormone receptor, fostering transport of the associated receptor to and regulating its signaling function at the PM (Christopoulos et al., 2003); gC1q-R, receptor for globular heads of C1q, which interacts with the carboxyl terminus of the 
1B-adrenergic receptor and regulates the maturation and expression of the receptor (Xu et al., 1999); and BiP/Grp78, a chaperone involved in the protective unfolded protein response (UPR), which is a cell stress program activated when misfolded proteins accumulate in the lumen of the ER (Yang et al., 1998; Schröder and Kaufman, 2005)] (see paragraph below).
It is recognized that continuous ER stress, such as that provoked by the accumulation of unfolded proteins, results in cell death and relates to the pathogenesis of some neurodegenerative diseases (Forman et al., 2003). Accumulation and aggregation of misfolded proteins are presumably responsible for some neurodegenerative diseases such as early-onset familial Alzheimer's disease, Parkinson's disease, and prion disease (Forloni et al., 2002). In these diseases, the soluble conformations of proteins or fragments of proteins convert to insoluble fibrillar aggregates, known as amyloids, which are formed by cross-
-pleated sheet structures that accumulate intra- and/or extracellularly (Glenner, 1980; Dobson, 1999; Forloni et al., 2002). A set of interlinked molecular pathways, collectively referred to as the UPR, are activated by ER stress; overwhelming or failure of the UPR may lead to apoptosis and thus play an important role in the pathogenesis of the above-mentioned neurodegenerative disorders (Forman et al., 2003). Several quality control factors participate in the UPR, including the ER chaperone BiP/Grp78, which negatively regulates three proximal sensors, the transmembrane kinase and endoribonuclease IRE1, pancreatic ER kinase, and activating transcription factor 6 (Yang et al., 1998). When unfolding or misfolding occurs, BiP dissociates from the sensors and binds the unfolded proteins in an attempt to refold them; this dissociation releases the sensors from negative inhibition, leading to the activation of multiple signaling pathways and induction of UPR-inducible genes and decreased protein expression (Forman et al., 2003). These changes increase the folding capacity of the ER, reduce new protein translocation to the ER, and increase the degradation of the abnormally folded or unfolded proteins (Harding et al., 2002; Kaufman, 2002). Whereas prolonged UPR activation may lead to apoptosis, several proteins presumably involved in amyloid-forming disorders may promote or inhibit various steps in the UPR (Forman et al., 2003). Little attention has been paid to the potential role of the ER stress response on the pathogenesis of diseases caused by misfolding and intracellular accumulation of GPCRs, and there is no evidence to date that aggregation of misfolded GPCRs may follow the catastrophic fate observed for proteins that cause neurodegenerative diseases. Nevertheless, the observation that in vitro expression of the misfolded mutant Pro23His of rhodopsin (the most frequent rhodopsin mutation leading to retinitis pigmentosa) results in formation of aggregates due to a generalized impairment of the ubiquitin-proteasome system (which is potentially toxic to the cell) (Saliba et al., 2002; Illing et al., 2002), strongly suggests the existence of a link between the mechanisms leading to photoreceptor degeneration in retinitis pigmentosa and those that participate in the genesis of neurodegenerative diseases, including the ER UPR.
|
III. Physiology of the Gonadotropin-Releasing Hormone Receptor and Vasopressin Type 2 Receptor Systems |
|---|
|
TopPreviousNext |
|---|
The mammalian GnRHR type I (hereafter referred as GnRHR) (Fig. 1A) belongs to the superfamily of G protein-coupled receptors, specifically to the family related to the rhodopsin and -adrenergic receptors (family A). The GnRHR is located in the pituitary gonadotrope and is bathed by the circulation of the (closed) hypothalamic-pituitary portal system, which transfers pulsatile signals of the hypothalamic decapeptide, GnRH (shown). The gonadotrope cell responds with a concomitant pulsatile release of the gonadotropins, LH, and FSH (Santen and Bardin, 1973; Knobil, 1974). These enter the peripheral circulation and regulate gonadal steroidogenesis, along with maturation of eggs and sperm. Intermittent exposure of the GnRHR to the releasing hormone is important from a functional point of view; slower GnRH pulses favor release of FSH whereas faster pulses favor release of LH (Belchetz et al., 1978; Crowley et al., 1985; Hazum and Conn, 1988) (Fig. 2). Frequency modulated signals are also important to prevent desensitization (refractoriness) of the gonadotrope to a subsequent stimulus, allowing for the occurrence of distinct rates and patterns of synthesis and release of the gonadotropins that follow GnRH exposure (Belchetz et al., 1978). The GnRHR is among the smallest members of the GPCR superfamily (328 amino acid residues in the human GnRHR) and bears unique structural features, including the lack of a carboxyl-terminal intracellular tail (Millar et al., 2004). Fish, reptiles, birds, and the primate type II GnRHR (McArdle et al., 1999; Millar, 2003) do possess this carboxyl extension whose presence is associated with differential physiological receptor regulation (Lin et al., 1998); when added to the mammalian GnRHR, it dramatically increases PME levels of this receptor (Janovick et al., 2003b). Another important feature of the GnRHR is the amino acid residue in position 191, which is frequently Glu or Gly but is replaced by Lys in primates (Janovick et al., 2006); in rat and mouse GnRHR, this amino acid is absent (Arora et al., 1999). The GnRHR is coupled to the trimeric Gq/11 protein, whose activation stimulates the effector enzyme phospholipase-C, leading to phosphatidylinositol 4,5-biphosphate hydrolysis and formation of the second messengers, inositol 1,4,5-triphosphate and diacylgycerol (Conn et al., 1986). The former messenger diffuses through the cytoplasm, promoting the release of intracellular calcium and the release of both gonadotropins. More recently, coupling of the GnRHR to the cAMP pathway has been shown under conditions of sustained stimulation, which may be potentially important under intense GnRH release, such as the preovulatory GnRH surge (Larivière et al., 2007).
|
|
Gonadotropin-releasing hormone, the natural ligand of the GnRHR, interacts with several amino acid residues of the receptor located mainly in the transmembrane (TM) domains: these include Asp98, Asn102, Asp302, Trp101, Lys121, Asn212, and Tyr290; GnRH peptide agonist binding sites overlap with some natural ligand binding sites, and also may interact with Trp280 and Phe216, depending on the particular structure of the agonist (Sealfon et al., 1997; Millar et al., 2004). Peptide antagonists occupy GnRHR binding sites that differ from, but that may overlap, the agonist binding pocket, as suggested by mutational analysis and molecular dynamics simulations (Millar et al., 2004; Söderhall et al., 2005). Nevertheless, the fact that agonists, antagonists, and inverse agonists may exhibit distinct selectivities toward the active and the inactive conformation of the receptor suggests that competitive antagonism may occur without any overlap with agonist binding sites (Samama et al., 1993). Binding of nonpeptide antagonists is less known; nevertheless, studies on a few nonpeptide small, quinolone- and thienopyridine-based GnRH antagonists indicate that their binding sites partially overlap GnRHR residues important for GnRH binding (Cho et al., 1998; Cui et al., 2000); in addition, Phe313 has been proposed as a site critical for the binding of this class of antagonists to the human GnRHR (Cui et al., 2000).
Loss-of-function mutations in the GnRHR can lead to partial or complete hypogonadotropic hypogonadism, a failure of pituitary gonadotropes to respond to GnRH, which results in decreased or apulsatile gonadotropin release and reproductive failure (Ulloa-Aguirre et al., 2004a). To date, 21 inactivating mutations (including two leading to missing of large sequences) in the human GnRHR gene have been described as a cause of HH (Figs. 1A and 3). Seven homozygous and 12 heterozygous combinations of human GnRHR mutants are expressed by individuals exhibiting either partial or complete forms of HH (Beranova et al., 2001; Ulloa-Aguirre et al., 2004b). The majority (
90%) of the GnRHR mutants whose function has been examined to date (17 mutants) are trafficking-defective receptors as disclosed by mutational studies and/or responses to pharmacoperones (see section III.C.). Because reproductive failure is not life-threatening, it is likely that many cases (particularly partial HH forms) go undiagnosed and, individual mutants, if severe in the phenotype, are not passed to progeny.
|
, 2002). Agonists and antagonists (peptide and nonpeptide) seem to prefer a common V2R compartment for docking (Czaplewski et al., 1998a,b; [/oe]lusarz et al., 2006). Molecular modeling of the V2R has revealed that the receptor amino acid residues potentially important in ligand binding are located mainly in TM domains 3 to 7, including Cys112, Val115 and Lys116, Gln119, and Met123 (helix 3), Glu174 (helix 4), Val, 206, Ala210, Val213, and Phe214 (helix 5), Trp284, Phe287 and Phe288, and Gln291 (helix 6), and Phe307, Leu310, Ala314, and Asn317 (helix 7) (Czaplewski et al., 1998b).
Nearly 188 inactivating mutations of the V2R causing X-linked diabetes insipidus have been described previously (Fujiwara and Bichet, 2005; Bichet, 2006; Boson et al., 2006); the vast majority of these mutations correspond to the so-called type 2 V2R mutant receptor (Bichet, 2006; Morello and Bichet, 2001). Furthermore, 89 of these 184 mutations are missense mutations likely to result in misfolded proteins that are trapped in the ER and are unable to reach the basolateral cell surface to engage the antidiuretic hormone arginine-vasopressin (Bichet, 2006). These defective receptors accumulate in different compartments of the early secretory pathway, depending on their folding state (Hermosilla et al., 2004). The urine of patients with diabetes insipidus is not concentrated in response to (the antidiuretic hormone) arginine-vasopressin. Accordingly, the lack of concentration results in severe dehydration in the absence of adequate replacement hydration and elevated sodium levels.
B. Lessons from Comparison of the Gonadotropin-Releasing Hormone Receptor and Vasopressin Type 2 Receptor Systems: Selecting Likely Targets for Rescue by Pharmacoperones
In considering these two systems, each with their own set of mutants (sites shown as dark circles in Fig. 1, A and B) and their own chemically distinct pharmacoperones (Janovick et al., 2003a; Bernier et al., 2004a), some commonalities that may provide important clues about how to identify systems that are especially amenable to this rescue approach are evident: 1) Both the GnRHR and the V2R recognize small peptide ligands, a decapeptide and nonapeptide sequence, respectively; 2) each ligand is only slightly larger than 1000 Da in mass (Fig. 1, A and B), and both peptide ligands are somewhat hydrophobic in nature; and 3) because these are small molecules, it is likely that the compartments for docking in their corresponding receptor are also relatively small, and both receptors are believed to bind ligand mainly in their TM sections, consonant with their hydrophobic nature (Czaplewski et al., 1998a,b; Millar et al., 2004). It is reasonable to assume that significant features such as the ligand binding site would be maintained during evolution, as the general structure of both ligands has been conserved. In fact, many of the residues involved in ligand binding are relatively invariant for the GPCR superfamily (Ulloa-Aguirre and Conn, 1996; Millar et al., 2004).
Both receptors themselves are also relatively small, compared with other GPCRs; as mentioned in section III.A., the human GnRHR is 328 amino acids and the human V2R is 371 amino acids. Sequence identity between the human GnRHR and the V2R is 20%, whereas their similitude (i.e., sharing similar residues or conservative substitutions) is 39% (Fig. 3). These results may simply reflect the conserved nature of the TM domains among GPCRs. The cytoplasmic extensions of the amino termini of both receptors are quite short, as are those of the carboxyl termini—the GnRHR has none at the carboxyl end. This protein actually terminates in the cytoplasmic face of the membrane. The intra- and extracellular loops are, as the total size would predict, quite small (Fig. 1, A and B). The general size similarities between these two receptors, taken as a whole, suggest that the "correctly" folded (that is, the structure that passes the QCS) structures of small receptors with small ligands (i.e., small ligand binding site) might be quite sensitive to distortion and, accordingly, are easily recognized as defective by the QCS. This may explain why these two receptors, as well as human rhodopsin (348 amino acid residues), are the most frequently affected among the GPCR superfamily by mutations leading to ER trapping and disease (Tan et al., 2004). In addition, it is reasonable to consider the possibility that a mutation may have a proportionally larger effect on a small, compact structure than on a large, diffuse one and presents a potentially more easily rescuable pharmacoperone target. For larger receptors in which the ligand binding site is located in the (large) amino-terminal region (i.e., the calcium-sensing receptor and gonadotropin receptors), the ligands are generally less hydrophobic than those that bind in the TM region. This view is borne out in pharmaceutical development by the observation of the relatively more commonly available peptidomimetics for receptors with intramembrane ligand binding, compared with those that bind ligand at the amino terminus. It is true that there are some nonpeptide compounds that bind gonadotropin receptors and produce effects (van Straten et al., 2002), but these are not true competitors at the ligand binding site (which occurs at the large amino-terminal extension); in fact, binding occurs in or near the membrane-associated components of the receptor (Jäschke et al., 2006).
To function, pharmacoperones need to be small and hydrophobic, so that they can enter cells. Accordingly, the binding site for them is also likely to be small and hydrophobic. These characteristics are also expected to apply for small hydrophobic peptides that bind in the TM region. Accordingly, it is conceivable that these criteria (small hydrophobic ligands that bind to the TM regions of the smaller GPCRs) might circumscribe systems in which pharmacoperones might be expected to work best.
C. Relation between the Overall Receptor Structure and the Structure of the Ligand Binding Site
It is also worth remarking on the observation that chemical structures, which bind to the active site (or at least compete with the ability of agonists to do so), also serve as pharmacoperones, causing misfolded proteins to fold in a way that makes them fold correctly and escape the QCS (Morello et al., 2000; Janovick et al., 2003a; Noorwez et al., 2004; Hawtin, 2006). This observation suggests that there may be a relation between proper construction of the active site and proper overall construction of the receptor. When we (Janovick et al., 2003a) and others (Morello et al., 2000; Bernier et al., 2006; Hawtin, 2006) used pharmacoperones from multiple different chemical classes, we found that if a particular mutant could be rescued with one class, it could be rescued with pharmacoperones from other chemically unrelated classes. Mutants that rescued poorly with one class, rescued poorly with all (Janovick et al., 2003a). Accordingly, even very different peptidomimetic antagonist structures (indoles, quinolones, and erythromycin-derived macrolides), which presumably bind nonidentically to the mutants did something at (or near) the ligand binding site that resulted in rescue by "pulling" the entire molecule into a structure that passed the criteria of the QCS (Janovick et al., 2003a). Similar results were found for distinct mutant, traffic-defective V2Rs (see section III.D.), which may suggest that the active site has been defined at least in part by the overall shape of the molecule that is recognized by the chaperone system and that controls exposure of particular ER retention motifs that serve as mediators of the QCS, functioning only when the receptor is misfolded (Hermosilla and Schülein, 2001; Hermosilla et al., 2004). This appears to be a two-way relation, as placing a template at the ligand binding site also corrects the overall folding of the protein, at least when viewed from the perspective of the cellular QCS. This observation initially seems to be in contradiction with the consideration that endogenous protein chaperones recognize more general errors in the molecule, such as the exposure of a hydrophobic plate (Ellgaard and Helenius, 2001), rather than specific features, such as a defect in a ligand binding site that is specific to an individual receptor. It may be easier to understand the reason for this relation between the binding site and the overall structure of the receptor protein in light of the advantage of binding a ligand at a site that can cause substantive changes in the overall shape of a receptor. These changes would be required for receptor activation and transduction of ligand binding to effector activation, for example. In fact, it is accepted that activation of GPCRs results from agonist-provoked changes in the conformation of the receptor that drives the equilibrium between the inactive and active state in favor of the latter (Gether and Kobilka, 1998; Karnik et al., 2003). The ability of ligands to change the shape of the receptor [and conversely of receptors to stabilize in distinct active conformations in response to different ligands (Lu et al., 2005)] and thereby transfer the signal to domains involved in G protein activation suggests that allosteric GPCR modulators (May et al., 2007) may be potential candidates for pharmacological chaperoning. In this vein, cell membrane-permeant, allosteric modulators may be designed to either aid folding of mutant proteins or misfold overexpressed proteins that may potentially lead to disease, driving the misfolded protein to the degradation pathway.
It is also notable that for both receptors, the distribution of disease-causing mutations is quite broad and includes intra- and extracellular loops, the amino terminus, and the TM regions (Figs. 1 and 3). In the GnRHR, there are (to date) no naturally occurring mutations reported in TM segment 1, extracellular loop (ECL) 3, or intracellular loops 1 or 2. This piece of information alone could lead to two very different conclusions: either these regions are so important that any mutation leads to lethality (or inability to reproduce) or, alternatively, many are clinically silent and are never reported, even though present. We (Janovick et al., 2002) have constructed a large number of (non-naturally occurring) mutations, including deletions and truncations, in the human and rodent GnRH receptors and found that the vast majority can be rescued by pharmacological means. Even though these mutations interfere with the ability of the receptor to respond, they have not been found in association with disease (Janovick et al., 2002, 2006). This finding suggests that the ability of pharmacoperone rescue is remarkably broad and not limited to "hot spot" areas or particular motifs.
In reconstructing human GnRHR mutants that are associated with disease in orthologous rat or mouse receptor templates, it has been noticed that many of these constructs no longer result in misrouting. This was one of several observations that led to the conclusion that the human GnRHR is delicately balanced between the PM and retention in the ER, whereas rat and mouse GnRHRs are generally routed to the plasma membrane with much higher efficiency (Knollman et al., 2005; Janovick et al., 2006). This trend, which will be discussed in section IV.B., has evolved under substantial selective pressure by several mechanisms (Conn et al., 2006a,b) and costs energy because of the "inefficiency" of needed synthesis of unused protein. Accordingly, this inefficiency may represent a mechanism of post-translational regulation and presents an opportunity for pharmacological intervention.
Fujiwara and Bichet (2005) presented an image of the V2R showing disease-associated mutants (Fig. 1B). There are 188 putative disease-causing mutations in this receptor (sometimes more than 1 at the same site, shown by triangles in the figure). It is clear, by inspection, that the mutation sites are more densely associated with the TM region, an observation that is not surprising in light of the relative conservation of this area and the potential for disturbing the hydrophobic nature of the interaction between the receptor structure and the TM area. The same seems to be true for the GnRHR, although on the basis of far fewer mutations (Figs. 1A and 3). Given the total number of reported mutations, it is remarkable that none have been reported in the amino terminus of the V2R and only 2 have been reported for the GnRHR. It is certainly possible that mutations may occur in this region but go unreported if they are clinically silent. If true, that possibility would minimize the role of this part of the receptor (with the exception of Asn residues involved in glycosylation and intracellular trafficking) in determination of routing. Otherwise all intracellular loops and TM segments of the V2R contain mutations. For the GnRHR, the first two intracellular loops lack reported mutations. Only two misfolded V2R mutants are located at or near sites recognized as general motifs believed to be involved either in ER export [mutation at R337, involving the E(X)3LL motif in the carboxyl terminus (Schülein et al., 1998; Krause et al., 2000)] or ER retention [mutation at Glu242, near the two overlapping retinoid X receptor motif at intracellular loop 3 (Hermosilla et al., 2001)] (Figs. 1B and 3; Table 2), whereas in the human GnRHR only one mutation (Arg262Gln) is located at a potential retinoid X receptor ER retention motif (Gassmann et al., 2005) (Figs. 1A and 3). In both receptors, mutations in the highly conserved motifs E/DRY (located at the boundary of the TM helix 3 and the second intracellular loop and permutated to DRS in the GnRHR and to DRH in the V2R) and D/NPxxY (in the seventh TM helix) have been reported. These motifs have important structural and functional roles in many GPCRs (Gether and Kobilka, 1998; Rovati et al., 2007); GnRHRs and V2Rs bearing mutations in this motif are misfolded receptors that may be partially or completely rescuable by pharmacoperones (Leaños-Miranda et al., 2002; Bernier et al., 2004c,2006; Topaloglu et al., 2006). It has been shown recently that in some GPCRs (including the V2R), mutations in the E/DRY motif promote constitutive receptor endocytosis as a result of increased receptor phosphorylation and arrestin association (Shi et al., 1998; Barak et al., 2001; Wilbanks et al., 2002). Although pharmacoperone rescue of the V2R Arg137His mutant showing constitutive internalization may be attributed to inhibition of the constitutive interaction of the mutant receptor with arrestin, Bernier et al. (2004c) show that the effect of the pharmacoperone on cell surface expression and signaling efficacy of the V2R mutant was, rather, attributable to the pharmacological action of the chaperone on a portion of the receptor population that remained intracellularly trapped because of an inability to attain a conformation compatible with ER export. These findings concurrently suggest that the conserved DRY motif is also involved in proper folding and/or ER export of the receptor to the cell surface membrane. Nevertheless, the possibility that pharmacoperones may counteract the effects of mutations, leading to constitutive desensitization via stabilization of the receptor at the cell surface membrane, represents an interesting therapeutic alternative to rescue function of constitutively internalized receptors. In an earlier study by Bernier et al. (2004a), only 7 of the 38 mutants could not be rescued by pharmacoperones.
|
Because both the GnRHR and the V2R are small GPCRs in comparison to the rest of their superfamily, it is pragmatically easier to prepare mutants for study (i.e., by site-directed mutagenesis). This has added to the facility of use of these mutants for research models. In the case of the GnRHR, all but three [Ser168Arg, Ser217Arg, and the truncated L314X (stop)] of the 17 mutants tested were completely or partially rescued with pharmacoperones (Conn et al., 2002; Leaños-Miranda et al., 2002; Janovick et al., 2003a; Topaloglu et al., 2006). It has been possible to show that the Ser168Arg and Ser217Arg GnRHRs are mutants in which the thermodynamic changes leading to receptor distortion are too great to effect rescue (Janovick et al., 2006). Accordingly, even though these two mutants are not rescued by pharmacoperones, their failure to route correctly is attributable to misfolding not to an intrinsic inability to potentially participate in particular receptor functions such as receptor activation or G-protein coupling. In the case of misfolded V2Rs, it has been shown that distinct hydrophobic, cell membrane-permeable antagonists (Serradeil-Le Gal et al., 1996; Albright et al., 1998) effectively rescue function of several misfolded, trafficking-defective V2R mutants that cause diabetes insipidus in humans (Morello et al., 2000; Bernier et al., 2004a,b,c). The fact that the effect of these antagonists on mutant V2R expression could not be mimicked by a V2R impermeant antagonist and that the antagonist pharmacoperones did not rescue function of mutants that are normally expressed at the cell surface membrane indicated that the pharmacoperones acted intracellularly to promote maturation and targeting of misfolded mutants to the PM (Morello et al., 2000).
Although not a GPCR, another rescuable molecule is the cystic fibrosis TM conductance regulator (the ion channel that is defective in cystic fibrosis). Unlike the two GPCRs under discussion, in which many different mutations can lead to disease, the same cystic fibrosis TM conductance regulator mutation that results in this disease is common (approximately 80% of the time)—a deletion of the amino acid at position 508 (Lim and Zeitlin, 2001; Kerem, 2005). This observation initially suggested that the diversity of mutations in the GnRHR and V2R might complicate the search for rescue strategies of intracellularly retained mutant receptors (Oksche and Rosenthal, 1998). Nevertheless, the observations that a number of distinct GnRHR and V2R mutants could be rescued by the same pharmacoperones has clearly challenged this notion (Morello et al., 2000; Janovick et al., 2003a).
|
IV. Ligand and Receptor Frequency Modulation in Signaling Systems: Implications for Model Selection and the Timing of Pharmacoperone Administration in Vivo |
|---|
|
TopPreviousNext |
|---|
Of the V2R and GnRHR, only the latter seems to decode signals that are pulsatile in nature. Nonhuman primates, notably the rhesus macaque, are excellent models for human reproduction that have led to new drugs, medical procedures, and devices. One of the reasons is that the GnRHR of the rhesus macaque, like that of other primates, is sensitive to a complex signal, with both amplitude- and frequency-modulated components from the releasing hormone (Knobil, 1974). This is probably one of the mechanisms by which the cell is able to respond to one ligand with multiple different endpoints, all having different time constants (Crowley et al., 1985).
There is value in determining the degree of similarity of function among the GnRHRs, as this moiety is the key analog to digital transducer of neural signals that regulates the reproductive endocrine system, and the development of therapeutic approaches will require reliable animal models. Other receptors also may have the ability to transduce a complex signal that is both amplitude- and frequency-modulated and, in the case of the GnRHR, alterations in pulse frequency may provide a means of therapeutic intervention. Because an increased level of FSH over several days seems to be needed to recruit a dominant follicle for ovulation (Wildt et al., 1981), accurate sensing of pulses is important for ovulation. In patients with HH bearing GnRHR mutations, a variable profile of spontaneous pulsatile LH release, from completely apulsatile to decreased frequency and amplitude of pulsatile release, has been found, depending on the particular receptor mutations presented (de Roux et al., 1997, 1999; Beranova et al., 2001; Layman et al., 2001; Meysing et al., 2004). Thus, growing and maturation of ovarian follicles and ovulation cannot occur. The fact that some patients with HH do respond to exogenous (and endogenous) agonists (de Roux 1999; Layman et al., 2001; Meysing et al., 2004) indicates that some degree of GnRHR cell surface membrane expression is present and that the population of mutant receptors so expressed is functional.
The GnRHR itself cycles in rodents and primates (Filicori et al., 1986), and there is the possibility that this process could be neatly controlled by endogenous chaperones (Brothers et al., 2006). To set the stage for in vivo studies, we have examined the molecular aspect of the WT primate GnRHR that might cause this molecule to be amenable to regulation by pharmacoperones.
B. Amino Acids Associated with Control of Plasma Membrane Expression
As described in section III.A., a particular feature of primate GnRHRs is the presence of a Lys residue at position 191, which is located in the second ECL. The presence of Lys191 destabilizes a Cys14-Cys200 bridge (shown in Fig. 1) that is a critical determinant for primate GnRHRs to pass the QCS of the cell (Conn et al., 2006a,b; Ulloa-Aguirre et al., 2006). Failure of this bridge <
