Pharmacologic Rescue of Conformationally-Defective Proteins: Implications for the Treatment of Human Disease

Authors

  • Alfredo Ulloa-Aguirre,

    1. Oregon National Primate Research Center, Oregon Health & Science University, Beaverton, OR, USA
    2. Research Unit in Reproductive Medicine, Instituto Mexicano del Seguro Social, Mèxico D.F., Mexico
    Search for more papers by this author
  • Jo Ann Janovick,

    1. Oregon National Primate Research Center, Oregon Health & Science University, Beaverton, OR, USA
    Search for more papers by this author
  • Shaun P. Brothers,

    1. Oregon National Primate Research Center, Oregon Health & Science University, Beaverton, OR, USA
    2. Departments of Physiology and Pharmacology, and
    Search for more papers by this author
  • P. Michael Conn

    Corresponding author
    1. Oregon National Primate Research Center, Oregon Health & Science University, Beaverton, OR, USA
    2. Research Unit in Reproductive Medicine, Instituto Mexicano del Seguro Social, Mèxico D.F., Mexico
    3. Departments of Physiology and Pharmacology, and
    4. Cell and Developmental Biology, Oregon Health & Science University, OR, USA
    Search for more papers by this author

P. Michael Conn, connm@ohsu.edu

Abstract

The process of quality control in the endoplasmic reticulum involves a variety of mechanisms which ensure that only correctly folded proteins enter the secretory pathway. Among these are conformation-screening mechanisms performed by molecular chaperones that assist in protein folding and prevent non-native (or misfolded) proteins from interacting with other misfolded proteins. Chaperones play a central role in the triage of newly formed proteins prior to their entry into the secretion, retention, and degradation pathways. Despite this stringent quality control mechanism, gain- or loss-of-function mutations that affect protein folding in the endoplasmic reticulum can manifest themselves as profound effects on the health of an organism. Understanding the molecular, cellular, and energetic mechanisms of protein routing could prevent or correct the structural abnormalities associated with disease-causing misfolded proteins. Rescue of misfolded, “trafficking-defective”, but otherwise functional, proteins is achieved by a variety of physical, chemical, genetic, and pharmacological approaches. Pharmacologic chaperones (or “pharmacoperones”) are template molecules that may potentially arrest or reverse diseases by inducing mutant proteins to adopt native-type-like conformations instead of improperly folded ones. Such restructuring leads to a normal pattern of cellular localization and function. This review focuses on protein misfolding and misrouting related to various disease states and describes promising approaches to overcoming such defects. Special attention is paid to the gonadotropin-releasing hormone receptor, since there is a great deal of information about this receptor, which has recently emerged as a particularly instructive model.

Orderly cellular progress depends on both the absolute number and the appropriate location of a diverse array of proteins; among these are receptors, ion channels and enzymes. The synthetic processes are tightly regulated at the transcriptional, translational and post-translational levels by multiple signaling pathways. Synthesis and processing of this cellular machinery occur in association with the endoplasmic reticulum (ER). This cellular organelle has the daunting task of synthesizing and assembling nearly 100 000 proteins and provides the specialized environment necessary for folding, glycosylation, oxidation and oligomeric assembly of proteins prior to their translocation to other domains of the cell. As proteins are synthesized in the ER, they fold and adopt distinct conformations that provide a stable structure compatible with ER export (1–7). Protein folding is, of necessity, a complex process, given the proximity and diversity of proteins that accumulate in the cytosol at a nominal concentration of 100 mg/mL. According to current models of protein folding (4,8), the dihedral angles of the protein backbone fold synchronously in groups of four or more (“protein wriggling”), thereby avoiding dramatic steric clashes. The steric character of the protein backbone restricts the spectrum of protein shapes that are recognized by the stringent quality control mechanisms, which protect against aberrant cellular activity from misfolded molecules and promote the proper balance of synthesis, maturation and degradation that is crucial for cell function and survival. These mechanisms exist to check for proper folding, processing and structural integrity of nascent proteins, which in turn ensures proper intracellular trafficking and metabolic fate of the protein within the cell. The quality control system also prevents accumulation of defective proteins that potentially interfere with normal cell function (7–12).

It has been recognized that point mutations resulting in protein sequence variations may result in the production of misfolded and disease-causing proteins that are transcribed and translated at normal levels but are unable to reach their functional destinations in the cell (4,5,13,14). Misfolding can also be triggered by other factors including protein overexpression, temperature, oxidative stress and activation of signaling pathways associated with protein folding and quality control (13,15–17). In particular cases, misfolding results in loss-of-function of the conformationally defective protein via degradation through the polyubiquitination/proteasome pathway (18–20). Alternatively, misfolded proteins may aggregate, leading to potentially toxic intracellular accumulation or even to protein accumulation in the plasma with extracellular amyloid deposition (2,13,21–23). In the past decade, extraordinary efforts have been made to understand how abnormal folding relates to certain pathologies and to design therapeutic interventions that could prevent or correct the structural abnormality of disease-causing misfolded proteins. In this regard, rescue of misfolded “trafficking-defective” proteins by pharmacological chaperones is emerging as one of the most promising therapeutic strategies for such disorders (24–26). This review focuses on protein misfolding and misrouting in disease states and describes how the rescue of misfolded proteins can be effected pharmacologically. The gonadotropin-releasing hormone receptor (GnRHR) (Figure 1) is used as an example of how conformationally defective molecules are sensitive to functional rescue by a variety of experimental interventions (27–31). This particular system has been a useful model for defining the common characteristics of the mutants, the pharmacology of rescue, as well as for describing the dominant–negative actions of the mutants on wild-type (WT) receptors.

Figure 1.

Figure 1.

A) Sequence of the human gonadotropin-releasing hormone (GnRH) receptor (GnRHR) and location of the inactivating (loss-of-function) mutations identified to date (31). A member of the G protein-coupled receptor superfamily, the GnRHR consists of a single polypeptide chain that traverses the cell surface membrane seven times, forming characteristic transmembrane helices, interconnected by alternating extracellular (EC) and intracellular (IC) loops. The GnRHR is a ligand-activated switch that activates the heterotrimeric Gq/11 guanosine nucleotide-binding protein. This receptor exhibits several unique features compared to other members of the superfamily, including the reciprocal exchange of the conserved D and N residues in the transmembrane domains II and VII (from left to right), the replacement of Y with S in the highly conserved DRY motif located in the junction of the transmembrane domain III and the second IC loop (residues 138–140), and the lack of the cytosolic carboxyl-terminal extension into the cytosol, which plays an important role in receptor cell surface expression, internalization and desensitization (32–34). Following GnRH binding, the GnRHR activates phospholipase Cβ as a result of coupling to the membrane associated trimeric Gq/11 protein associated with the intracellular domains of the receptor, leading to phosphatidylinositol 4,5-biphosphate hydrolysis and formation of inositol 1,4,5-triphosphate (IP) (34,35). B) Counterclockwise orientation of a prototypic G protein-coupled receptor from transmembrane domains I to VII (with the cytoplasmic side at bottom). The closed loop structure is representative of receptors for peptide ligands such as GnRH. In this arrangement, the core is composed mainly of transmembrane domains II, III, VI and VII, whereas domains I, IV and V are peripherally sequestered. GnRH interacts with the second EC loop (at N102), transmembrane domain III (at K121), and the third EC loop-transmembrane domain VII junction (D302) of the GnRHR (34).

Molecular Chaperones as Key Constituents of the ER Quality Control System

The ER quality control system operates at several levels, employing a variety of mechanisms that include a complex sorting system to identify and separate proteins according to their maturation status; it also includes the action of specialized folding factors, escort proteins, retention factors, enzymes and members of major molecular chaperone families (3,36–39). Chaperones promote folding and assembly through associations with proteins displaying particular features such as exposure of hydrophobic shapes, unpaired cysteines or immature glycans (11). In general, molecular chaperones are ER-resident proteins that bind to and stabilize unstable conformers of nascent polypeptides to facilitate the correct folding or assembly of the substrate polypeptide through regulated binding and release cycles (39). In addition to providing assistance with the folding of nascent proteins, molecular chaperones also prevent aggregation and/or incorrect interactions of misfolded proteins with other molecules in a crowded and viscous ER environment, thereby preventing their export to other cellular compartments (3,6,9,38,39). Thus, molecular chaperones guard nascent polypeptide chains against potentially unproductive and even toxic interactions that may occur during the different stages of the folding process (Figure 2). If the polypeptide chain fails to fold properly, then the incorrectly manufactured protein is targeted for destruction by proteasomes (18–20).

Figure 2.

Figure 2.

Quality control in the endoplasmic reticulum (ER). Newly synthesized polypeptides are translocated to the lumen of the endoplasmic reticulum (step 1). Folding is facilitated by interaction of the nascent polypeptide with molecular chaperones (rod-like structures) (step 2). Misfolded and misassembled products are retained in the ER and exposed to resident chaperones to attempt folding (step 3). Eventually, misfolded proteins may be dislocated into the cytoplasm for proteosomal degradation after dissociation of the molecular chaperones (step 4). Alternatively, defective proteins may be exported to and retained by the Golgi, and retrotranslocated to the ER, where correct folding is again attempted (step 5), or diverted to lysosomes for degradation (step 6). Proteins properly folded are translocated to the Golgi (step 7) where processing (e.g. glycosylation) and maturation of the protein molecule is completed (step 8). Mature products are then exported to their final destination (e.g. the cell surface membrane) (step 9). In the presence of pharmacological chaperones (e.g. IN3), folding and assembly of a defective (or mutated) protein is facilitated early during its biosynthesis, so that it can escape degradation, exit the ER and be properly localized (27–30). Molecular chaperones dissociate in the ER, whereas pharmacological chaperones may remain associated (37). Once the newly synthesized protein (e.g. the GnRHR) reaches its destination, the pharmacoperone can dissociate from the rescued molecule (step 10) to allow interaction of the ligand with its binding site at the receptor protein (step 11).

Molecular chaperones are also involved in surface expression of newly synthesized or recycled receptors (40). For example, in the nematode Caenorhabditis elegans, the odr-4 gene encodes for a regulatory molecule specifically expressed in chemosensory neurons. In this system, the chaperone assists in folding and/or targeting of the ODR 10 receptor to olfactory cilia (41). In Drosophila melanogaster, absence of the specific trans-acting factor Nina A (neither inactivation nor afterpotential A) leads to rhodopsin 1 ER accumulation and, eventually, to degradation (42–45). Nina A and its mammalian homolog RanBP2 bind specific opsins and act as chaperones, aiding proper folding, transport and localization of the mature receptors to the cell membrane (46). It is well known that the presence of N-glycans within glycoproteins favors folding and transport of the glycoprotein (47,48). A number of glycoproteins, including several members of the GPCR superfamily (49,50), interact with the ER lectin, calnexin, and its soluble homolog, calreticulin, both molecular chaperones that bind a broad range of proteins. The cycle of these chaperones predominantly centers on substrate N-glycans present on the newly synthesized proteins, adding hydrophobicity to the folding protein (51,52). The calnexin/calreticulin cycle depends on the concerted action of carbohydrate-modifying enzymes (glucosidases I and II). The action of the ER glucosidase I leads to formation of monoglucosylated oligosaccharide structures that interact with the chaperones, whereas the action of glucosidase II removes the remaining glucose residue from the oligosaccharide and terminates the association. At this point, the glycoprotein (already in its native conformation) is exported to the Golgi complex. Failure to achieve its native conformation promotes the addition of glucose to the defective glycoprotein by the folding sensor UDP-glucose:glycoprotein glucosyltransferase, allowing successive interactions of the glucose-tagged glycoprotein with calnexin and calreticulin to ensure that only properly folded molecules reach the cell surface or target organelle (11,53–56). When N-glycosylation does not occur (due to mutation in particular glycosylation sites of the protein molecule, for example), glycoproteins misfold, aggregate, and fail the quality control system. An abnormal interaction between calnexin and the R337X mutant of the arginine-vasopressin (AVP) V2 receptor (AVP VsR) is apparently responsible for the absence of cell-surface expression of the mutant molecule, which leads to X-linked nephrogenic diabetes insipidus in humans (50).

Protein Misfolding as Disease Etiology

Protein overexpression and/or mutations, oxidative stress or activation of signaling pathways linked to protein folding and quality control machinery may result in misrouting of newly synthesized proteins and consequently their premature or inappropriate processing and rapid degradation. In principle, production of defective proteins could also result in abnormal accumulation in the ER. Accumulation and aggregation of misfolded proteins are presumably responsible for some neurodegenerative diseases such as early onset familial Alzheimer's disease, Parkinson's disease, prion disease (14,16,57,58), as well as for early onset cataracts, α1-antitrypsin deficiency, and type II diabetes mellitus (59–62). In these diseases, proteins or fragments of proteins convert from their normally soluble conformations to insoluble, well-structured fibrillar aggregates, known as amyloids (e.g. β-amyloid in Alzheimer's disease and α-synuclein in Parkinson's disease), which are formed by cross-β-pleated sheet structures that accumulate intra- and/or extracellularly (14,63–66). The pathogenesis of amyloid-forming diseases is complex and multifactorial. Ineffectiveness of the quality control mechanisms to refold and/or degrade misfolded precursors (which may lead to formation of amyloid fibrils), the particular kinetics of misfolding and misassembly of the defective proteins, and the cytotoxic effects of the misfolded proteins or their fibrillar aggregate derivatives are likely the means by which such disease states develop (13,21, 22,24,67). In this regard, it has been proposed that direct or indirect activation of the so-called unfolded protein response (UPR) may be implicated in the pathogenesis of amyloid-forming diseases (12,16). The UPR is a regulatory mechanism which may be activated by diverse mechanisms, including the accumulation of aberrant proteins in the ER (which provokes ER stress). Activation of the UPR leads to the coordinated synthesis of ER-resident chaperones and enzymes. Several quality control factors participate in this response. The ER chaperone BiP/Grp78 negatively regulates three proximal sensors: the transmembrane kinase and endoribonuclease IRE1, pancreatic ER kinase, and activating transcription factor 6. 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, induction of UPR-inducible genes and decreased protein expression (16). 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 (68,69). While prolonged UPR activation may lead to apoptosis, several molecules presumably involved in amyloid-forming disorders (e.g. presenilin 1 in Alzheimer's disease or Parkin in juvenile parkinsonism) may promote or inhibit various steps in the UPR (16). In this regard, it has been shown that BiP binds the amyloid precursor protein in healthy cells, thereby limiting production of β-amyloid (70); mutations in presenilin 1 associated with familial Alzheimer's disease may alter the transmembrane kinase and endoribonuclease sensor IRE1 and reduce BiP levels, leading to ER stress and apoptosis or alternatively to enhanced β-amyloid production (71–73). Protein aggregation and fibrillar formation may occur later in the secretory pathway or even outside the cell, after secretion or membrane attachment. Although extracellular amyloid and aggregates can be degraded by immune system pathways, including catabolism by macrophages, clearance of protein aggregates may be a slow process, particularly in disease states, leading to the accumulation of misfolded and misassembled proteins responsible for diseases such as familial amyloidoses. In principle, some of the factors responsible for the development of these protein aggregates may be targets for therapeutic interventions.

Many diseases associated with misfolding and impaired intracellular trafficking involve membrane-associated proteins. Diseases caused by cell-surface protein mislocalization of otherwise functionally competent molecules include forms of familial hypercholesterolemia (74,75), retinitis pigmentosa (76–80), cystic fibrosis (81,82), and diabetes insipidus (83–85). Several mutations in the low-density lipoprotein (LDL) receptor involve defects in trafficking and/or processing that lead to accelerated degradation of the receptor (74,75). In cystic fibrosis, the ΔF508 mutation (found in ∼70% of patients with this condition) leads to chaperone-mediated ER retention and rapid degradation of the incompletely processed (albeit functional) cystic fibrosis transmembrane conductance regulator (CFTR) by the proteasome; this prevents cell surface expression of the chloride channel protein and consequently loss of cyclic adenosine-5′-monophosphate (AMP)-regulated chloride transmembrane conductance (82). In nephrogenic diabetes insipidus, urine concentration is defective due to resistance of the kidney to AVP or to defects involving the arginine-vasopressin-responsive aquaporin-2 water channel (84–86). When expressed in vitro, most (∼70%) AVP V2 receptor mutations exhibit intracellular trapping of the receptor molecules that are then unable to reach the cell membrane (85). Similarly, mutations of aquaporin-2 water channel can cause misrouting of the protein, preventing its cell surface expression (86). In retinitis pigmentosa (a disease characterized by retinal degeneration and, ultimately, total blindness), mutations in the gene encoding rhodopsin result in defective molecules that misfold and accumulate in the ER (87,88). Mutations in the carboxyl-terminus of rhodopsin cause defects in receptor trafficking to the outer segment of the rod cell (80). Finally, mutations in the gene of another GPCR, the GnRHR, may cause misfolding of the receptor protein, leading to impaired gonadal function (30,31).

Strategies for Stabilizing Misfolded Proteins

Several approaches have been applied to salvage defective proteins. Among these are the use of physical methods (81,89–91), nonspecific, low molecular weight protein stabilizing compounds (such as polyols) (92), genetic modification of mutant proteins (“genetic rescue”) (93–95) and use of template molecules or pharmacological chaperones (“pharmacoperones”) that correct errors in folding and restore activity by reestablishing correct protein routing (“pharmacological rescue”) (30,91,96,97). Although the mechanisms by which exogenous chaperones function are not fully understood, it is thought that they do one or several of the following: stabilize a specific conformation of the misfolded protein, reduce aggregation, prevent nonproductive interactions with other resident proteins and/or modify the activity of endogenous chaperones. Such effects serve to increase the efficiency of ER export and promote the proper trafficking of the protein to its correct destination (Figure 2).

Studies addressing the biosynthesis and localization of CFTR ΔF508 have shown that at 37 °C the mutant protein is not processed correctly and, accordingly, it is trapped in the ER and not delivered to the cell surface plasma membrane (81,82). Nevertheless, expression of this mutant in Xenopus oocytes and in Sf9 insect cells (which are usually maintained at lower temperatures than mammalian cells) has led to the detection of chloride channel activity, based on the sensitivity of mutant protein processing to temperature (81). Moreover, incubation at reduced temperatures (20–30 °C) caused the processing of the temperature-sensitive folding CFTR mutant to be reverted towards the native-type receptor configuration, allowing the cAMP-regulated chloride channel to be expressed at the cell surface membrane (81). Similarly, increased expression of WT GnRHR and several conformationally defective GnRHRs bearing different point mutations results from incubating transfected cells at lower (32 °C) temperatures (Figure 3). Thus, it appears that for certain temperature-sensitive, defectively folded proteins, lower temperatures prevent aggregation in the ER and permit the misfolded protein to escape the quality control mechanisms, allowing them to reach their final destination (81,98,99). Stable CFTR ΔF508 transfectants also exhibit cell surface expression of the mutant protein by employing protein stabilizing compounds such as glycerol, trimethylamine N-oxide and deuterated water (see Table 1). These substances may promote ER export of the mutant protein (even at nonpermissive temperatures) to the plasma membrane, and increase the functional activity of the CFTR ΔF508 protein including whole cell chloride conductance (89,92). Similar results were obtained employing 4-phenylbutyric acid, a low molecular weight fatty acid (100). In principle, some chemical chaperones or stabilizing agents are not selective and therefore can potentially assist folding and increase secretion of many different proteins in various cellular compartments, leading to inappropriate changes in the levels and/or secretion of many proteins that may be extremely undesirable. It has been observed, however, that some agents such as glycerol, 4-phenylbutyric acid and trimethylamine N-oxide (a highly potent osmolyte), selectively increase the secretion efficiency of α1-antitrypsin without influencing the secretion efficiency of other proteins or decreasing proteasomal degradation (25,107). The mechanism(s) whereby the former chaperones can selectively influence secretion is unknown, although up-regulation of the chaperone system has been proposed to explain why misfolded variants of α1-antitrypsin, but not other wild-type proteins, are secreted more efficiently (26). Nevertheless, while such physical and chemical approaches can rescue incompletely processed mutants, these are in general nonspecific, and are therefore of limited therapeutic value. Further, in some situations high concentrations of chemical chaperones might alter folding of some molecules in a manner that they may potentially promote polymerization/aggregation of certain conformationally defective proteins such as β-amyloid (108).

Figure 3.

Figure 3.

Effect of temperature on WT and modified GnRHR expression. In this experiment, 0.1 μg cDNA (WT or mutant) was transfected in COS-7 cells. Five hours after transfection, 20% fetal calf serum (FCS) was added to the cells and then placed at 32 °, 37 °, or 39 °C for 22 h. Cells were washed and then preloaded with 3H-inositol for 18 h, then stimu- lated for 2 h with 10−7 m GnRHn agonist (Buserelin) and total inositol phosphate production (IP) was measured. Note that the expression of rat WT GnRHR, is also increased by lowering the incubation temperature to 32 °C.

Table 1. Chemical chaperones studied for conformationally defective, disease-causing proteins
Target diseaseProtein involvedChemical chaperoneReference
  1. *CFTR: cystic fibrosis transmembrane conductance regulator. DMSO: dimethyl sulfoxide. TMAO: trimethylamine N-oxide. PBA: 4-phenylbutyric acid. D2O: deuterated water. AVP V2R: arginine-vasopressin V2 receptor. BCKD: mitochondrial branched-chain α-ketoacid dehydrogenase. MNK: Menkes protein.

Cystic fibrosisCFTR*Glycerol, DMSO, TMAO, PBA,D2O25,89,92,100
Nephrogenic diabetes insipidusAVP V2R, aquaporinGlycerol101,102
Emphysema and liver diseaseα1-antitrypsinGlycerol26
Maple syrup urine diseaseBCKD complexTMAO103
Menkes diseaseMNKGlycerol, copper104
Machado-Joseph diseaseAtaxin-3Glycerol, DMSO, TMAO105
Cancerp53, pp60, ubiqitin-activating enzyme E1Glycerol, TMAO, D2O25,89
 glucocorticoid receptorGlycerol106

Genetic approaches in which modifications are introduced to an already defective protein have been used to rescue the function of conformationally abnormal molecules. These approaches either overexpress or stabilize molecules rendered unstable by genetic defects or attempt to correct trafficking by adding specific cellular compartment targeting sequences. Rescue of the mislocalized CFTR ΔF508 mutant, which otherwise retains significant phosphorylation-regulated Cl channel activity, can be achieved by overexpression of the CFTR ΔF508 mutant regulator, an effect that results in the escape of limited amounts of the mutant protein to the plasma membrane (93). Introduction of additional cysteine residues in the extracellular domains of GPCRs may potentially allow formation of additional disulfide bonds or impair formation of a conserved bond, leading to binding- or trafficking-defective receptors (27,87,94). With binding-defective mutants it is possible to rescue receptor function by introducing a second site-suppressor mutation that abrogates formation of additional disulfide bonds (94). Another example of genetic rescue is the GnRHR E90K loss-of-function mutant whose plasma membrane localization and function are rescued to the WT level by deleting the amino acid K191, a residue absent in rodent GnRHRs and associated with increased cell surface membrane expression (Figure 4) (95). Genetic techniques are, by their nature, limited as the basis of therapeutic strategies. They are of considerable academic interest, however, since they led to the concept that the mutation itself does not significantly alter intrinsic molecular function (e.g. ligand binding, effector coupling), particularly in the case of misrouted cell-surface membrane receptors.

Figure 4.

Figure 4.

Mutational rescue of human GnRH receptors. Expression of both WT human GnRHR and human GnRHR E90K, a mutation in the second transmembrane domain (TMD-II), is increased by deleting the amino acid K191. In the experiment shown, COS-7 cells transiently expressing either nonmodified or modified (ΔK191) human WT GnRHR or the E90K mutant were exposed to GnRH analog (10−7 m Buserelin). Inositol phosphate (IP) production was determined after 2 h of incubation in the absence (green bars) or presence (red color) of the agonist. Activation of hGnRHR E90K-mediated intracellular signal was restored by the mutational approach (95).

The observation that trafficking-defective proteins that are retained by the ER quality control system may be functionally rescued by physical and chemical interventions has initiated the search for pharmacological agents that can rescue abnormally folded proteins (including receptor molecules having intrinsically low maturation efficiencies) (30,31) or prevent fibrillar aggregate formation (26). In general, desirable characteristics of molecules that could function as pharmacoperones for conformationally defective proteins include:

  • • cell-permeability;
  • • ability to reach the ER and remain undegraded long enough to stabilize the target mutant or change the protein misfolding energetics of the abnormal molecule (26,109);
  • • specificity for the aggregate precursor or the target protein being rescued.

In the case of misfolded, trafficking defective proteins, the rescuing agent should be able to dissociate from the target molecule (or, at least, not compete with the natural ligand binding site) after its localization at the correct cellular destination [e.g. the plasma membrane (Figure 2)] (see Table 2 for examples of pharmacoperones employed in experimental conditions). Short β-sheet breaker peptides have been designed for blocking the conformational changes and aggregation undergone by β-amyloid (24). These synthetic mini-chaperones, which have a structure homologous to the central hydrophobic region of the fibril aggregate, inhibit and dissolve β-amyloid aggregates both in vitro and in vivo (110,111). In transthyretin (TTR) amyloidogenesis, several small molecules bind with high affinity to the unoccupied binding sites within the TTR molecule, stabilizing the native state of the protein and decreasing the concentration of the intermediate species (109). Decreasing the concentration of aggregation-prone species is important in the process of aggregation because amyloid formation is a high-order kinetic process (26).

Table 2. Pharmacological chaperones experimentally employed to correct conformational diseases
Target diseaseProtein involvedChemical chaperoneReference
  1. * PS1: Presenilin 1. PS2: Presenilin 2. APP: amyloid precursor protein. PrP: Prion protein. HERG K+ channel: ether-a-gogo-related gene. IprP13: 13-residue β-sheet breaker peptide.

Misfolding/aggregation
 Alzheimer's diseasePS1, PS2, APP*β-Sheet breaker peptides24,110–112
 Transthyretin amyloid diseaseTransthyretinTransthyretin amyloidosis inhibitors109
 Prion diseasePrPβ-Sheet breaker peptides (IPrP13), quinacrine,chorpromazine24,113,114
 Gaucher's diseaseβ-glucosidaseN-(n-nonyl) deoxynojirimycin115
 β-Galactosidosisβ-galactosidaseGalactonojirimycin derivatives116
 Long QT syndromeHERG K+ channelCisapride, E-4031, astemizole117
 Retinitis pimentosaRhodopsin11-cis-7-ring retinal118
Misfolding/degradation/traffic defective
 Cystic fibrosisCFTRBenzoquinolizinium lead compounds119,120
 Fabry's diseaseβ-galactosidase A1-deoxy-galactonojirimycin121,122
 Hypogonadotropic hypogonadismGnRHRIndoles, erythromycin-derivatives, quinolones27–29
 Nephrogenic diabetes insipidusAVP V2RSR121463, VPA-98596
 Hyperinsulinemic hypoglycemiaSulfonylurea receptor-1Diazoxide123
 Drug resistanceP-glycoproteinCyclosporine, capsaicin, vinblastine, verapamil124,125
 Immunoglobulin secretionAnti-phenylphosphocholineHapten p-nitrophenylphosphocholine126
 Painδ-Opioid receptorOpioid receptor antagonists97
 Menkes diseaseMNKcopper104
 TumorigenesisSmoSmo antagonists127

Given that small ligand molecules may influence folding and organelle targeting of some proteins, and that antagonists may stabilize the catalytic center of an enzyme or prevent down-regulation of plasma membrane receptors, it was thought that small, cell-permeant antagonists may exert favorable effects by stabilizing the corrected conformation of mutant proteins, preventing their degradation, and facilitating translocation through the cell. The competitive α-galactosidase A inhibitor,1-deoxy-galactonojirimycin, facilitates ER export of the R301Q mutant form of this enzyme (whose retention in the ER leads to the lysosomal storage disease, Fabry's disease in humans) and transportation to its normal destination within the lysosome in vivo, thus restoring galactosidase activity in several organs (121). Similar results have been obtained by treating fibroblasts from patients with Gaucher's disease (which results from mutations in lysosomal β-glucosidase, leading to the accumulation of glycosylceramide in macrophages) with the enzyme inhibitor N-(n-nonyl)deoxynojirimycin (115).

G protein-coupled receptors represent the largest family of pharmacological targets. In these receptors, several compounds acting as pharmacoperones have been shown to be highly effective in experimental conditions. In vitro, treatment of cells expressing mutant AVP V2Rs with two nonpeptidic V2R antagonists (SR121463A and VPA-985) promoted proper folding, Golgi-type maturation and targeting of mutant receptors to the cell surface (96). For the human δ opioid receptor, for which only a fraction (∼40%) of synthesized protein is normally transferred to the cell membrane (128,129), both membrane permeable agonists and antagonists facilitated post-translational processing and increased export of the ligand-stabilized receptor from the ER to the cell surface (97). Another example of disease-causing GPCR mutants that are rescued by pharmacological chaperones is the GnRHR (Figure 1). Resistance to GnRH by mutations in its cognate receptor leads to impaired gonadal function due to hypogonadotropic hypogonadism, a disease characterized by delayed pubertal development and sexual immaturity (31,130,131). Expression of the majority of these GnRHR mutants in heterologous systems results in cells that neither bind GnRH agonist nor activate effector (131). Studies employing pharmacological chaperones have indicated that protein misfolding and resultant GnRHR misrouting of otherwise functional receptors is a mechanism that alone may explain the molecular etiology of this particular disorder (Figure 1) (27–29). In fact, for 12 of the 14 naturally occurring GnRHR mutants tested to date, in vitro exposure to particular indole derivatives resulted in both specific G protein coupling and ligand binding (Figure 5) (27–29,31). The observation that peptide antagonists that are membrane impermeable yet bind to the same WT receptor are unable to rescue human or rodent GnRHR mutants, suggests that interaction of the permeable peptidomimetics and the misfolded receptor occurs intracellularly, presumably at the ER (see next section) (27–29). Pharmacoperone treatment was also able to rescue function of some laboratory designed GnRHRs with terminal truncations, internal deletions or mutations at sites in which a cysteine residue normally appears (Figure 5 and Table 3). In these abnormally configured receptors, differences in the ability to effect rescue by IN3 depended either on the steric freedom or constraints associated with the amino acid substitution or the extent of the truncation (27).

Figure 5.

Figure 5.

Pharmacological rescue of WT and mutant human GnRH receptors by three cell membrane-permeant peptidomimetics. The figure on the right shows inositol phosphate production (IP) by COS-7 cells transiently expressing each receptor in the presence of 10−7 m Buserelin. Hydrophobic peptidomimetics are able to penetrate cells and interact specifically with protein targets; one such antagonist of GnRH is the indole, IN3 (132–134). In the experiment shown, cells were exposed to a 1.78 μm concentration of each pharmacoperone at the time of transfection. The pharmacological chaperones tested were the indoles IN30, IN31b, and IN3 (left figure). For 12 of 13 naturally occurring human GnRHR mutants tested in these studies, exposure to IN3 has resulted in both Gq/11 coupling and specific ligand binding (27–29,31), demonstrating that the effect of the mutational error leading to intracellular retention is either completely or partially corrected by this approach. Two of the mutants, S168R and S217R (Figure 1) that could not be rescued by this approach are only about 20 Å apart in the WT molecule. Conceivably, some mutants exhibiting either low responsiveness or refractoriness to pharmacoperone treatment may bear structural defects involving microdomains critical for ligand binding, receptor activation and/or effector coupling (34), but not those essential for proper protein folding or intracellular routing. All peptidomimetics studied with an IC50 value for the human GnRHR of 2.3 nm [including quinolones and erythromycin-derived macrolides (not shown in the figure)] displayed a measurable efficacy in rescuing GnRHR mutants, and within a single chemical class, this ability correlated to these IC50 values (29). The results from four laboratory manufactured loss-of-function rat GnRHRs [des325–327, des237–241 and des260–265 rat GnRHRs (three nonfunctional deletion mutants), and C278A rat GnRHR (a nonfunctional Cys mutant)] are also shown. Reproduced from [29], with permission from The American Society for Pharmacology and Experimental Therapeutics.

Table 3. Pharmacological rescue of laboratory manufactured rat GnRHRs bearing mutations in cysteine residues, truncations and deletions. Reproduced from [27] with permission from the Endocrine Society
   IN3 1.78 μmIN3 4.45 μm
 Vehicle(1 μg/mL)(2.5 μg/mL)
 MediumBuserelinMediumBuserelinMediumBuserelin
  1. The indicated mutant receptors were transiently transfected into COS-7 cells and the pharmacological chaperone IN3 was included at 0 (vehicle), 1.78 μm or 4.45 μm concentrations. Inositol phosphate production (shown in counts per minute) was determined in the absence (Medium) or presence of 10−7 m GnRH agonist (Buserelin). When varied substitutions are made at the same locus (i.e. C278A, C278V, C278T, C278M), differences in the ability to effect rescue by IN3 were found: the mutants resulting from the more bulky substituents (C278V and C278T) are apparently unable to reach a conformation necessary for rescue with low concentrations (27). Therefore, the expression of such variants on the plasma membrane also varies in the absence of IN3. Substitution of the bulky W279 (adjacent to C278) with alanine (W279A) yields an inactive mutant receptor that is rescued by IN3 (27). Of particular interest are the double mutants, C278W/W279C (exchange of WT sequence), C278V/W279V and C278A/W279A. The first two are modestly rescuable by high IN3 concentration, whereas the latter can be restored to full activity. The difference, again, appears to reflect the significant steric constraints associated with the larger amino acid residues: valine, tryptophan and cysteine. Alanine, in contrast, allows for more steric freedom owing to its smaller size; it better accepts the adjustment in the folding template configuration, resulting in the highest level of IP production. The ability of IN3 to rescue a carboxyl-terminal truncation rat GnRHR mutant (des325–327), which shows no response to GnRH when expressed in vitro, is also shown; IN3 rescued this mutant receptor. Removing a larger (12-amino acid) sequence from the carboxyl-terminus yields a mutant (des316–327) that can not be pharmacologically rescued, although as many as four amino acids can be removed from the third intracellular loop (des237–241 GnRHR) and then rescued, albeit partially (see Figure 5) (27,29).

Cys mutants and adjacent mutants
 C278A13227213820931472216
 C278V1611441536711431519
 C278T1441461428261641102
 C278M17223215413291321169
 W279A15216713510741491609
Double mutants
 C278W/W279C148147155152146255
 C278V/W279V161147153155148475
 C278A/W279A1591481602191482077
Deletion mutants
 Des (237–241)104106110319121452
 Des (316–327)112103105103111107
 Des (325–327)1081229910311091605

Taken together, the data allow some predictions regarding the expected difficulty of mutant rescue by small permeable peptidomimetics.

  • • The abnormal protein should not bear mutations in residues essential for ligand binding (or substrate/cofactor binding in the case of enzymes, and ion binding in the case of ion channels) or interaction with other effectors (such as G proteins); in these situations, the abnormal receptor (enzyme, ion channel), even when targeted to its proper destination within the cell, would be, predictably, nonfunctional.
  • • Large sequence omissions (deletions and/or truncations) may have a dramatic effect on the protein structure and thus make the protein unrescuable since nothing could reestablish the higher order structure.
  • • Loss or gain of a cysteine residue may potentially disrupt required bridges or form inappropriate bridges that may be so significantly disruptive to the structure of the protein that rescue cannot occur.
  • • Loss or gain of a proline residue may limit or even impede pharmacological rescue since the occurrence of this amino acid is associated with a forced turn in the protein sequence; in some proteins, an abrupt turn is likely a requisite for correct structure and cannot be corrected by pharmacoperones.
  • • Amino acid substitutions that impede or promote hydrogen bond formation may reduce the capacity of the pharmacoperone to effect rescue due to the inability of the protein to establish correct interactions between its different domains (31).
  • • In contrast to the significant constraints associated with substitutions involving replacement by larger amino acid residues (valine, tryptophan, threonine and cysteine), replacement with smaller residues (glycine or alanine) allow for more steric freedom, owing to their smaller size and, accordingly, greater acceptance of the folding template's adjustment of configuration; the latter situation may potentially yield a high efficiency of pharmacological rescue.

One interesting observation is that both incubation at lower temperatures (Figure 3) and treatment with the pharmacoperone IN3 increase the expression level of WT human GnRHR (27–29). Further, addition of either the hemagglutinin influenza virus epitope tag or the carboxyl-terminal tail to the WT human GnRHR, or deletion of lysine 191, result in a significant increase in functional expression of the receptor (Figure 6) (28,85,135,136). Thus, as with other membrane receptors [(e.g. the human δ opioid receptor (see above; 97)], the use of pharmacoperones has revealed that a portion of the WT hGnRHR is inefficiently processed by the cell in natural conditions, retained in the ER, and eventually degraded (Figure 6).

Figure 6.

Figure 6.

A large portion of the WT human GnRHR is inefficiently processed by the cell, retained, and eventually degraded (A). Experimental evidence for this phenomenon comes from studies showing increased expression of the WT human GnRHR (hGnRHR) by treatment with the pharmacoperone IN3 or by genetic modifications of the receptor protein [addition of the catfish carboxyl-terminal tail (cfCtail) or deletion of the amino acid K191 (ΔK191)] (B and C). B): Specific binding of [125I]-labeled agonist (Buserelin) to COS-7 cells transiently expressing each GnRHR receptor. C): Buserelin-stimulated (10−7 m) inositol phosphate (IP) production in cells transfected with the WT and the genetically modified receptors. As shown, the use of pharmacoperones has revealed that a portion of the WT hGnRHR is inefficiently processed by the cell in natural conditions, retained in the ER and eventually degraded. Incompletely processed receptors may function as a reserve pool of molecules that can be called upon when required. This view is supported by the observation that pharmacoperones do not increase the functional expression of GnRH receptors with high intrinsic maturation efficiency and membrane expression levels [such as human GnRHR ΔK191, catfish GnRHR carboxyl-terminal tail/human GnRHR chimera, rat GnRHR (in which K191 is absent), and rat GnRHR chimeras bearing the catfish GnRHR carboxyl-terminal tail or several carboxyl-terminal tail truncated fragments] (136), suggesting that the WT human GnRHR is less efficiently routed to the functionally active site on the plasma membrane than is the rodent or piscine counterpart. The lower intrinsic expression of the human GnRHR may reflect a relatively new evolutionary development for this receptor, as deletion of the primate-specific K191 or addition of the premammalian carboxyl-terminal tail increases plasma membrane expression, yet produces a modified receptor that does not increase plasma membrane expression by pharmacological means.

Pharmacoperones Reverse the Dominant–Negative Effects of Misfolded Human GnRHR Mutants on Wild-Type Receptor Expression

Receptor dimerization or oligomerization and interactions with accessory proteins have been well documented and proposed as important determinants of GPCR activity (137,138). It appears that GPCRs approach the issue of oligomerization differently, just as some receptors are phosphorylated (whereas others are not) and some receptors bind ligand in the amino terminus (whereas others bind ligands in the lateral plane of the membrane) (130). One reason for this might be related to the opportunity for receptor cross-talk affected by heterologous interactions of receptors (139). Clearly, some receptors are monomeric in the membrane and oligomerize upon ligand binding (Figure 7), whereas others oligomerize as they are synthesized in the ER, an apparent requisite for correct targeting to the cell surface (137,138). Intracellular association of GPCRs as homo- or heterodimers could lead, in principle, to either cell surface targeting (a dominant–positive effect (148–150)) or to intracellular retention of the complex (dominant–negative effect (148–150)). Further, truncated variants of the AVP V2R (148), the human GnRHR (151) and the D2 and D3-dopamine receptors (152,153) may interfere with the cell surface expression of their corresponding WT receptors, by their association in the ER and misrouting of the resulting complex. In this regard, it has been shown that a number of human GnRHR mutants bearing folding defects may affect WT GnRHR function (154). Further, the function of the WT receptor complexed with pharmacoperone-sensitive mutants (E90K, C200Y, or L266R) (Figures 1 and 5), recovered to levels above those observed for the WT receptor alone, indicating that the pharmacoperone interacted with and successfully rescued both the mutant and the WT receptor species (154).

Figure 7.

Figure 7.

Multimerization or oligomerization of the GnRHR in the plasma membrane is among the first measurable events leading to receptor activation by an agonist. The GnRHR was, in fact, the first GPCR shown to become activated upon oligomerization (140–144), an event that seems to be a general feature shared by members of the superfamily of GPCR membrane receptors (137,138). In these figures, cell surface membrane expression and interactions between rat GnRH receptors in GH3 cells are shown. A) Plasma membrane expression of rat GnRHR-catfish carboxyl-terminal tail-green fluorescent protein (GFP) (a) and rat GnRHR-catfish carboxyl-terminal tail-red fluorescent protein (RFP) (b) chimeras as measured by confocal microscopy. Both fluorophores are expressed on the plasma membrane, and significant amounts of RFP can be found inside the cytoplasm; this material may represent misfolded copies or molecules otherwise targeted for destruction. GnRHR chimeras were functional and elicited inositol phosphate (IP) production in response to GnRH agonist (c). B) The ratio between red and green fluorescence measured for the same area on the membrane increases with time after addition of the GnRH agonist. Small areas on the membrane excluding intracytoplasmic organelles were measured on images acquired at various times after addition of GnRH agonist to a live cell: green and red fluorescence at time 0 (a and c) and after 10 min (b and d). The decrease of green fluorescence accompanied by the increase in red fluorescence for the same area is an indication that fluorescence resonance energy transfer (FRET) occurs when GFP- and RFP-tagged receptors carrying different fluorophores come into appropriate physical proximity (i.e. receptor microaggregation in cells expressing both receptor chimeras causes the increase in red fluorescence at the expense of green fluorescence) (e). The time course of the ratio of red : green increase includes the rate of GFP photobleaching as well as the rate of formation of GnRHR-GFP and GnRHR-RFP oligomers. Green and red fluorescence intensities are normalized to their values at time 0. The graph shows a small but consistent increase in red fluorescence, increasing ∼ 10% after 20 min. The intensity of green fluorescence drops robustly, partly due to both energy transfer to RFP and photobleaching. Reproduced from [143], with permission from the American Society for Biochemistry and Molecular Biology.

As described above, a number of mutant receptors may potentially act as negative or positive regulators of membrane receptor expression and function via formation of oligomers that affect proper delivery and trafficking of the complexed receptors to the cell surface (148–153,155,156). Certainly, the fate of the resulting complex will depend on the particular conformation adopted by the associated proteins (148,150,153,157). In this vein, the use of confocal microscopy of fluorescently labeled WT GnRHR has shown that the dominant–negative effect of GnRHR mutants results from WT receptor retention in the ER by mislocalized mutants (158). Pharmacoperone treatment increases plasma membrane localization and restores receptor-mediated intracellular signaling of both mislocalized mutants sensitive to pharmacological rescue and the wild type receptor, thus reversing the dominant–negative effect (Figure 8). The observations that a number of G protein-coupled receptors, including the human GnRHR, have evolved to restrict the amount of cell surface membrane expressed receptors and that conformationally defective variants of receptors may exert dominant–negative effects on WT receptor function, concurrently suggest that aberrant ER intracellular retention of proteins normally expressed in the plasma membrane might be a common feature in human disease.

Figure 8.

Figure 8.

Localization of green fluorescent protein (GFP)-tagged wild-type (WT) human GnRHR when coexpressed in COS-7 cells with dominant–negative disease-causing human GnRHR mutants [(E90K, L266R and S168R (see Figure 1)], before and after treatment with the pharmacoperone IN3. A to H: Confocal micrographs of cells coexpressing the GFP-tagged WT GnRHR (green) and empty vector (A and B), GnRHR E90K (C and D) , GnRHR L266R (E and F) or GnRHR S168R (G and H) and stained with ER-Tracker dye (blue). Micrographs numbered 1 are single confocal sections, whereas those numbered 2 are overlay projections of all sections through the cell. Upon pharmacoperone exposure, the GFP-tagged WT receptor exhibits greater plasma membrane localization when expressed alone (B) or with the E90K mutant (a fully rescuable receptor mutant) (D). GFP-tagged WT GnRHR coexpressed with the L266R mutant shows a modest re-localization of the fluorescent protein to the plasma membrane (F), reflecting the partial ability of IN3 to rescue this particular mutant. IN3-treated cells coexpressing the GFP-tagged WT receptor and the (temperature and pharmacoperone) unrescuable S168R mutant, show GFP retained in the endoplasmic reticulum and not localized in the plasma membrane (H). I–L) Inositol phosphate (IP) production in cells individually transfected with each human GnRHR with and without IN3. The failure of the pharmacoperone to ablate the dominant–negative effect of the unrescuable mutant on the WT receptor function suggests that proteins in a multimeric complex must be in the correct conformation to be properly exported from the ER. Reproduced from [158] with permission from the Endocrine Society.

Concluding Remarks

Since the studies on the CFTR ΔF508 mutant (89,92) and the energy-dependent transporter P-glycoprotein (124,125), the ability to rescue misfolded mutant proteins by chemical or pharmacological means has forced the reexamination of the widely held view that all mutations result in molecules that can not recognize ligands. This has led to the conclusion that many products of genetic mutations are simply misrouted within the cell. Acceptance of this view opens the door to the concept that pharmacological chaperones might be identified that can be used therapeutically. Molecules that stabilize mutants, of course, do not need to compete with agonists or antagonists, although it has been convenient to utilize peptidomimetic hormone antagonists for proof of principle. For this reason, it is not only possible, but likely, that pharmaceutic archives contain pharmacoperones that have been missed in standard agonist and antagonist screens.

While protein rescue may be a significant therapeutic venue, one could also imagine molecules that interact with native structures, altering their folding kinetics so that they undergo misfolding and removal; such an approach, which we have chosen to call ‘protein shipwrecking’ might be effective in cancer therapeutics or the development of the mechanism of contraception.

Mutant rescue and protein shipwrecking provide new approaches to dealing with defective or undesirable gene products without the need for in vivo genetic engineering. A companion review that focuses more generally on rescue of disease associated mutants is available (159).

Acknowledgments

The work described from our laboratory was supported by NIH Grants HD-19899, RR-00163, HD-18185, and TW/HD-00668. We thank Drs. Jim Dias, Jon Hennebold, and Sergio Ojeda for their useful comments on the manuscript.

Ancillary