Regulation of G Protein-coupled Receptor Trafficking by Inefficient Plasma Membrane Expression

Despite the prevalence of G protein-coupled receptors as transducers of signals from hormones, neurotransmitters, odorants, and light, little is known about mechanisms that regulate their plasma membrane expression (PME), although misfolded receptors are recognized and retained by a cellular quality control system (QCS). Convergent evolution of the gonadotropin-releasing hormone (GnRH) receptor (GnRHR) progressively decreases inositol phosphate production in response to agonist, validated as a measure of PME of receptor. A pharmacological chaperone that optimizes folding also increases PME of human, but not of rat or mouse, GnRHR because a higher percentage of human GnRHRs are misfolded structures due to their failure to form an apparent sulfhydryl bridge, and they are retained by the QCS. Bridge formation is increased by deleting (primate-specific) Lys191. In rat or mouse GnRHR that lacks Lys191, the bridge is non-essential and receptor is efficiently routed to the plasma membrane. Addition of Lys191 alone to the rat sequence did not diminish PME, indicating that other changes are required for its effects. A strategy, based on identification of amino acids that both 1) co-evolved with the Lys191 and 2) were thermodynamically unfavorable substitutions, identified motifs in multiple domains of the human receptor that control the destabilizing influence of Lys191 on a particular Cys bridge, resulting in diminished PME. The data show a novel and underappreciated means of posttranslational control of a G protein-coupled receptor by altering its interaction with the QCS and provide a biochemical explanation of the basis of disease-causing mutations of this receptor.


EXPERIMENTAL PROCEDURES
Materials-pcDNA3.1 (Invitrogen), the GnRH analog, D-tert-butyl-Ser 6 -des-Gly 10  Mutant Receptor-Rodent and human WT and mutant GnRHR cDNAs for transfection were prepared as reported (4); the purity and identity of plasmid DNAs were verified by dye terminator cycle sequencing (PerkinElmer).
Transient Transfection-Cells were cultured in growth medium (DMEM, 10% fetal calf serum, 20 g/ml of gentamicin) at 37°C in a 5% CO 2 humidified atmosphere. For transfection of WT or mutant receptors into COS-7 cells, 5 ϫ 10 4 cells were plated in 0.25 ml of growth medium in 48-well Costar cell culture plates. 24 h after plating, the cells were washed once with 0.5 ml of OPTI-MEM and then transfected with 5 ng/well of WT or mutant receptor with 95 ng of pcDNA3.1 (empty vector) to keep the total amount of DNA at 100 ng/well. Lipofectamine was used according to the manufacturer's instructions. 5 h after transfection, 0.125 ml of DMEM with 20% fetal calf serum and 20 g/ml of gentamicin were added. 23 h after transfection the medium was replaced with 0.25 ml of fresh growth medium. Where indicated, 1 g/ml of IN3 in 1% dimethylsulfoxide ("vehicle") was added for 4 h in respective media to the cells and then removed 18 h before agonist treatment (7).
Inositol Phosphate (IP) Assays-27 h after transfection, cells were washed twice with 0.50 ml of DMEM/0.1% bovine serum albumin/20 g/ml of gentamicin and "pre-loaded" for 18 h with 0.25 ml of 4 Ci/ml myo-[2-3 H(N)]-inositol in inositol-free DMEM. The cells were washed twice with 0.30 ml of DMEM (inositol free) containing 5 mM LiCl and treated for 2 h with 0.25 ml of a saturating concentration of Buserelin (10 Ϫ7 M) in the same medium. Total IP was determined (8). This assay has been validated as a sensitive measure of PME for functional receptors when expressed at low amounts of DNA and stimulated by excess agonist (1-5, 7, 9 -13).
Binding Assays-Cells were cultured and plated in growth medium as described except that 10 5 cells in 0.5 ml of growth medium were added to 24-well Costar cell culture plates (cell transfection and medium volumes were doubled accordingly). 23 h after transfection, the medium was removed and replaced with 0.5 ml of fresh growth medium. 27 h after transfection, cells were washed twice with 0.5 ml of DMEM containing 0.1% bovine serum albumin and 20 g/ml of gentamicin, and 0.5 ml of DMEM was added. After 18 h, cells were washed twice with 0.5 ml of DMEM/0.1% bovine serum albumin/10 mM HEPES, and a range of concentrations of 125 I-Buserelin (1.25 ϫ 10 5 to 4 ϫ 10 6 CPM/ml) (4) in 0.5 ml of the same medium were added to the cells. Cells incubated at room temperature for 90 min (14). After 90 min, the media were removed and radioactivity was measured as previously described (15).
To determine nonspecific binding, the same concentrations of radioligand were added to similarly transfected cells in the presence of 10 M unlabeled GnRH. Saturation binding curve fits and calculations were computed using SigmaPlot 8.02 (Jandel Scientific Software, Chicago, IL); a non-linear one-site binding model was used to calculate B max values (16), which are displayed as percentages of respective WT levels.
Statistics-Data (n Ն3) were analyzed with one-way analysis of variance and then paired Student's t test (SigmaStat 3.1; Jandel Scientific Software); p Ͻ0.05 was considered significant. Fig. 1 shows the hGnRHR, indicating residues that differ in the rat (ovals; the rat substitution is shown first in each pair of letters). Lys 191 (absent in rat and mouse GnRHR) is associated with diminished PME of the human sequence and is shown by an enlarged circle in ECL2. The thermodynamic "favorability" of the interspecific substitutions (www.russell.embl-heidelberg.de/aas/aas.html, for membrane proteins) is indicated by numbers. Changes marked with zero are neutral substitutions, and positive numbers are favored. The three negative numbers (disfavored substitutions) are shown in bold and are summarized for other species in the table. Proximity suggests (9) that two Cys bridges may form in the rat and hGnRHR, connecting the NH 2 -FIGURE 1. Diagram of the human WT GnRHR, emphasizing those residues that differ compared with rat WT GnRHR (shown Rat/Human in black circles). An enlarged circle shows the position of Lys 191 in ECL2. The "favorability" of substitutions (www.russell.embl-heidelberg.de/aas/ aas.html, for membrane proteins) of the interspecific differences is indicated by numbers in parentheses adjacent to the changes. Zero is neutral; positive numbers are favored; negative numbers are disfavored.
The Cys 114 -Cys 195/196 , but Not Cys  , Bridge Is Required for Optimal Routing of the Rat and Mouse GnRHR, whereas Both Are Needed for hGnRHR-We assessed the potential bridge requirements by converting each of the four Cys to Ala (Fig. 2) in the mouse, rat, or human sequence. For all sequences, the bridge between ECL1 and ECL2 (Cys 114 -Cys 195/196 ) appears essential because replacement of either Cys residue by Ala resulted in loss of IP in response to 10 Ϫ7 M Buserelin (a saturating concentration of a metabolically stable agonist of the GnRHR) (10). Under the conditions described under "Experimental Procedures," the IP response of a functional receptor to a saturating concentration of agonist is a sensitive measure of PME (1-5, 7, 9 -13). Pharmacoperone IN3 did not effect rescue (shown only for human, Fig.  2, inset). Replacement of either Cys in the bridge between the NH 2terminal and the ECL2 resulted in loss of IP production in the human but had a more modest effect in the mouse sequence and virtually no effect in the rat sequence. The human mutants (C14A or C200A) were rescued by IN3, suggesting that the absence of this bridge resulted in a rescuable misfolded protein.
The Requirement for the Potential Cys  Bridge Is Lost in the Absence of Lys 191 and Requires an Additional Motif Not Found in Rat or Mouse GnRHR-Deletion of Lys 191 from the human WT sequence elevated PME, but insertion of Lys 191 had little effect on the rodent recep-tors (which lack Lys 191 ), suggesting that a more complex motif was responsible for the decreased efficiency of PME (Fig. 3A). Pharmacoperone IN3 increased PME of the hGnRHR (Fig. 2, inset), revealing that WT PME is limited by the presence of a proportion of misfolded receptors. Fig. 3B shows that removal of Lys 191 from the hGnRHR sequence also rescues mutants (C14A and C200A), indicating a relation between these two sites and that the presence of Lys 191 diminishes the probability of this association. Interestingly, the effect of the Lys 191 is not wholly an effect of charge because converting it to uncharged Ala 191 , negatively charged Glu 191 , or modestly positively charged Gln 191 also produces inefficient PME of human sequence compared with hGnRHR lacking Lys 191 , although to somewhat different levels (Fig. 3C).
Particular Amino Acids Co-evolved with the Appearance of Lys 191 or Other Insertions at Position 191-The table in Fig. 1 compares the most strongly, thermodynamically disfavored amino acid substitutions among GnRHR sequences for various species (amino acids 7, 168, and 203 (202 in rat and mouse)). Amino acid 189 was included because of its physical proximity to Lys 191 . Of the 40 amino acid differences between rodent and human GnRHR, it was striking that three of the four sites identified in the   Fig. 4 shows the effect of mutating amino acids in the human second extracellular loop (hECL2) to the rat sequence (rendering the human ECL2 sequence more rat-like from amino acids 186 to 203), as well as several individual or multiple changes selected based on proximity to this region (Fig. 4A). The difference between the rat (Ala) and human (Thr) at position 190 was ignored because of the close conservation of these amino acids and the observation that mouse, guinea pig, canines, and ungulates are identical to humans at this site. The data show that replacement of the rat sequence by hECL2 enhances PME whether or not Lys 191 is present (Fig. 4B, left). The reverse combination in which hECL2 is replaced by the rat ECL2 significantly blunts the effect of removal of Lys 191 and diminishes the overall level of PME (Fig. 4B, right). These results were unexpected because the rat WT normally expresses at a higher level than the human WT. Such mutants (including the human GnRHR with mutation to the rat sequence at ECL2 alone (Fig. 4B), 168/ECL2, 161/168/ECL2, 182/ ECL2, 189, 186, 203, or 189/203, Fig. 4D) are rescued with pharma-coperone IN3 (not shown) to levels above that observed for WT receptor. Moreover, deletion of Lys 191 (Fig. 4, B and D), a modification associated with increased routing to the plasma membrane, also rescues these mutants. These observations confirm that the effects of mutating ECL2 are due to misfolding and misrouting rather than altered interactions with effectors or altered levels of mRNA synthesis. Also surprising, replacement of the rodent ECL2 with the human sequence did not restore the ability of Lys 191 to interfere with folding or routing and actually increased the level of PME.

Exchange of Human and Rat ECL2 Appears to Influence the Proximity between Cys 14 and Cys 200 -
Mutations in Amino Acids within ECL2 or in Structures Flanking It Also Regulate the Relation between the NH 2 -terminal and ECL2-The effects of mutating individual amino acids within and flanking ECL2 were examined (Fig. 4C, rat; Fig. 4D, human) because these might impact the relation between Cys 14  In the rat, mutations at 161,168, 182, 186, and 189 increased WT PME (Fig. 4C). Less dramatic increases were seen with mutation in position 202 of rat GnRHR. When combined mutants (189/202, 168/ECL2, or 182/ECL2) were made, the effect of the mutant that exacerbated PME appeared to predominate. Also in the rat sequence, several mutants into which Lys 191 was inserted showed higher PME (161, 168, 189, 202); 182 and 186 showed more modest effects. The proximity required for Cys bonds to form (1-2 Å) explains why exceedingly precise alignment is needed for Cys 14 -Cys 200 and why twisting of the support sequences (transmembrane segments 4,5) or those that abut on that area (i.e. ECL1, 3) also impacts bridge formation.
Other Sites Impact on the Net Level of Receptor PME and Allow the Effect of Lys 191 -Modifications of the hGnRHR at positions 4, 7, and 10 resulted in PME above the level of WT and amplification of the effect of deletion of Lys 191 (Fig. 5A). Mutation at these same sites in the rat GnRHR did not have a measurable effect on WT PME, although modifications at 4, 7, or 10 in the presence of hECL2 enhanced PME.
Mutations in human positions 24 and 27 (Fig. 5A) decreased PME of WT GnRHR and blunted the effect of deletion of Lys 191 . In the rat, homologous mutations at residues 24 and 27 increased WT and WT(ϩLys 191 ) PME, perhaps due to the replacement of a Thr (C ␤branched, presenting bulkiness near the backbone) with an amphipathic Met 24 that allows the side chain to be buried and increases the chance of bridging the NH 2 -terminal with ECL2.
We created an additional series of interspecific chimeras. Fig. 5, B (rat) and C (human), shows the dramatic effect of modifications at amino acids 112, 207/208, 224/225, 299/300, 301/302, as well as selected combinations of these sites. Unlike the majority of mutations associated with ECL2, these are all steric in nature, because hydrophobicity and charge are largely conserved. In the rat sequence, mutations were made at 299 alone or in combinations. For all mutants PME was blunted. Insertion of Lys 191 with 299 mutants also had diminished PME. Mutations at 207, 224, 301 were also associated with elevated PME, with or without Lys 191 .
Of note, substitutions at positions 161, 207 224, 207/224, and 229/301 (Figs. 4 and 5) each allowed destabilization by Lys 191 in the rat sequence, presenting the possibility that contributions from multiple sites were important. Individual homologous mutations in the human sequence did not totally ablate the ability of the deletion of Lys 191 to decrease PME; some actually increased the effect, again consistent with the view that several different sites contribute to the final effect of Lys 191 . In addition, mutations at 300 and 302 dramatically increased overall PME in the human sequence, with or without Lys 191 , and decreased the same in the rat sequence.

Breaking the Cys 14 -Cys 200 Bridge Allows Assessment of the Stability of the Association of the NH 2 -terminal and ECL2 in Human and Rat
Mutants-To determine the stability of the NH 2 -terminal-ECL2 relation and the persistence of destabilization by Lys 191 created by the transfer of human substitutions to the rat sequence, we prepared "broken bridge" mutants (i.e. C14A, Fig. 5D) for rat and human sequences indicated by brackets (Fig. 4, B-D, and Fig. 5, B and C).
The human sequence in which position 189 was converted to rat (Q189P) results in modestly more PME than hWT and a considerable blunting of the effect of removal of Lys 191 (Fig. 4D). When the bridge is broken in this sequence, the ability to create a functional receptor is lost and not rescuable by deletion of Lys 191 . This observation means that formation of the bridge occurs at a time when the association of these regions is intimate. The observation that all rat sequences in which we successfully transferred the destabilizing effect of Lys 191 with human substitutions lose this effect when the bridge is unable to form (Fig. 5D, C14A) suggests that failure to create the bridge during synthesis results in a lost opportunity for Lys 191 to modulate this effect. The impact of Lys 191 can only occur at the time when the bridge is able to form (likely at the time of synthesis) and not once the molecule is complete.
The human with rat substitutions at positions 112, 208, 300, and 302 is remarkable, as these amino acids flank and appear to stabilize the association of positions 14 and 200 even in the absence of the Cys bridge (i.e. when C14A is present). In this curious molecule, which is primarily human, we observe rat-like PME levels (rat WT GnRHR typically reaches the plasma membrane with about twice the efficiency of the human WT) and lack of requirement for the Cys 14 -Cys 200 bridge yet maintenance of sensitivity to the effect of destabilization by Lys 191 .
Quantification of Plasma Membrane GnRHRs by Radioligand Binding-To confirm that the results measured in the functional assay (IP production) reflected changes in actual receptor numbers (in contrast to changes in the efficiency of coupling to G proteins, for example), we used radioligand binding to quantify receptor numbers of selected WTs and mutants (Fig. 6). We included rat and human WT GnRHR, as well as human WT (which normally contains Lys 191 ) with this residue deleted, and rat WT (which normally does not contain Lys 191 ) with this residue added. We also included representative protein-encoding plasmids with the exchange of the human and rat ECL2 for both species, as well as representatives of human sequences with C14A. This modifica- tion results in breaking the apparent bridge between positions 14 and 200, a modification that is without effect in the rat. We also included human mutants in which the rat characteristics (less requirement for the Cys bridge and higher expression levels) were conveyed (112/208/ 300/302). The data confirm radioligand binding and IP production produce similar conclusions, notably that deletion of Lys 191 from the hWT markedly increases PME, but adding this residue to the rat has little effect. Further, breaking the apparent Cys 14 -Cys 200 bridge with the mutation C14A is associated with loss of PME in the human unless the 112/208/300/302 mutation is present.

DISCUSSION
Lys 191 in the hGnRHR sequence (328 amino acids) appears to raise the folding criteria and structural requirements for the hGnRHR compared with its rat/mouse counterparts that lack this residue (327 amino acids) or hGnRHR from which Lys 191 was deleted (1)(2)(3)(4). The view that Lys 191 increases the percentage of misfolded/misrouted receptors is supported by the observation that a pharmacoperone, which corrects folding errors, enabling mutants to pass through the cellular QCS, also increases the PME of the human WT, but not rat GnRHR. This observation also allowed us to exclude ligand binding, effector coupling, or level of mRNA expression as a cause of reduced mutant PME (10 -11).
Radioligand binding data and studies in which pharmacoperones are used to correct folding and sequential routing indicate that, at the low cDNA levels transfected, IP production is a good measure of functional receptors at the plasma membrane.
Removal of Lys 191 or use of pharmacoperone obviates the requirement for an apparent Cys 14 -Cys 200 bridge in the hGnRHR required for correct folding. Because inserting Lys 191 alone into the rat or mouse sequence (Ͼ88% homologous with the human) did not impart the requirement for this bridge between the NH 2 -terminal and ECL2, we sought to identify the other components of the requisite motif.
We examined thermodynamically unfavorable amino acid substitutions after recognizing that these frequently co-evolved with the appearance of the "extra" amino acid at position 191 and were proximal to it and to the bridge. An extra amino acid at position 191, not necessarily Lys, is present in all non-rat/mouse mammalian species that have been sequenced to date.
Insertion of Lys 191 or Glu 191 is associated with substitution of Leu 7 for Pro 7 , except in the guinea pig in which case a less favorable change is made (Ser 7 ). Furthermore, Lys 191 is associated with a change from Ile 168 to Ser 168 in primates, whereas Glu 191 is associated with Ser 168 in the guinea pig and ungulates but Gly 168 in the case of canines. Such substitutions are always unfavored. Ile is C ␤ branched with two non-H substituents attached to the ␤ carbon, presenting bulkiness near the protein backbone and restricting conformations that the main chain can adopt.
Insertions at position 191 also correlate with a change of the Pro 189 to Gln 189 , neutral for membrane proteins but unfavored (Ϫ1) in the extracellular space. Only primates show the change from Pro 202/203 to Ser 203 .
Glu 191 is found in the sequence of many pre-primate mammalian GnRHs (Fig. 1, table). Although it too can destabilize bridge formation, it is markedly less effective than Lys 191 found in primates. This further supports the progressive trend of limiting GnRHR expression at the plasma membrane in higher evolved species.
The effect of mutations associated with human disease emphasizes the importance of the Cys 14 -Cys 200 bridge. Mutants S168R and S217R are in a previously reported "zone of death" (1) and cannot be rescued by any of several different chemical classes (indoles, quinolines, and erythromycin macrolides) of pharmacoperones tried, which is a rare circumstance because the vast majority of mutants are rescuable by all classes (5). This observation and the physical relation between transmembrane segments 4 and 5 and ECL2 make it attractive to consider that (chargealtering) mutations in these two residues exert their influence by regulating the position of ECL2 and the intimacy of Cys 14 and Cys 200 . Because of charge considerations, the unfavored exchange of Ser and Arg likely moves the ECL2 into a position from which the formation of a Cys bridge is improbable and the mutant never passes the cellular QCS even in the presence of pharmacoperones. The dramatic effect of the mutation at position 217 supports the view that the Cys 14 -Cys 200 bridge is not created until the transmembrane segment 5 is formed.
The importance of position 217 is emphasized by recognizing that the homologous residue in rodents (amino acid 216) is associated with substantial differences in both trafficking and dominant negative action between the rat (Ser 216 ) and mouse (Gly 216 ) GnRHR (10), despite Ͼ96% homology. The substitution Ser 217 with a highly flexible Gly 217 results in loss of PME of the mutant, an effect that is rescuable by deletion of Lys 191 (10).
C200Y is another disease-associated mutation (12)(13)(17)(18)(19) that results in loss of the Cys 14 -Cys 200 bridge and decreased PME; rescue by a pharmacoperone (10 -13) supports the significance of the bridge for creating a correctly folded moiety. Consistent with that view, the homologous mutation in the rat or mouse GnRHR is a fully functional receptor. Indeed, the observation that many human disease-associated mutations recreated in rodent templates are functional (10) emphasizes the increased significance of the QCS in the human. It appears that a balance is created (in the human) between retention by the QCS and routing to the membrane. Accordingly, the human, but not rat or mouse, GnRHR is extremely sensitive to naturally occurring point mutations (10) that perturb this balance. . PME of GnRHR and mutants from protein-encoding plasmids was assessed in a radioligand binding assay using 125 I-Buserelin, a GnRH agonist. Plasmids expressing rat and human sequences are shown as percent of WT for the respective template species. Plasmids for rat and human WT GnRHR, as well as human WT (which normally contains Lys 191 ) with this residue deleted and rat WT (which normally does not contain Lys 191 ) with this residue added, were expressed as described under "Results." Representative protein-encoding plasmids with exchanged human and rat ECL2 for both species were included, as well as representatives of human sequences containing C14A. This modification results in breaking the apparent bridge between positions 14 and 200, a modification that is without effect in the rat. We also included human mutants in which the rat characteristics (less requirement for the Cys bridge and higher expression levels) were conveyed (112/208/300/302).
Of rodents, animals with large litters, only the guinea pig is known to have added an amino acid (Glu) at position 191 of the GnRHR. Interestingly, the guinea pig, a hystricomorph that diverged very early in rodent evolution, has a long luteal phase (a primate characteristic). Most non-rodent mammals such as cows, sheep, pigs, dogs, and horses also contain Glu 191 , suggesting that the loss of an amino acid in the homologous position is a specialization associated with very short reproductive cycles in rats and mice. Unlike all other reported mammalian sequences, the opossum (a non-placental mammal that places fetuses in a marsupium) has an uncharged, racemic Gly 191 that may reflect the early divergence of this group. In hGnRHR, substitution of uncharged (Ala 191 ) or negatively charged (Glu 191 ) amino acids appears to have similar effects as the Lys 191 , both in the epitope-tagged (20) and in the WT hGnRHR (present study), a finding that suggests that neither charge nor specific steric considerations totally explain the action of this amino acid but supports the view that the extra amino acid provokes crowding and a change in the relative positioning of amino acids that normally leads to Cys bridge formation. Noting that various species have inserted Lys 191 , Gly 191 , or Glu 191 suggests convergent evolutionary pressure for a means to use this position as a regulatory mechanism.
The observation that residues that regulate the efficiency of trafficking continue to evolve, even among the upper primates, suggests that selective pressure remains for this regulatory approach. Ser 203 and Leu 300 are uniquely human as other mammals have Pro 203 or Val 300 . Asp 302 is found in mammals with an extra amino acid at position 191. In rodents the corresponding amino acid is Glu 301 , believed to interact with the Arg 8 of the natural ligand, GnRH (21). Position 112 is Phe in rat, mouse, opossum, dog, and pig, but the human sequence contains Leu 112 at this position. This is not a case of uniqueness, because cows, sheep, horses, and even pre-mammals (fruit fly, chicken, and Xenopus) share Leu at this site.
In considering the molecule as a whole, the modifications associated with orientation of ECL2 (positions 7, 168, 189, and 202/203) all involve the gain or loss of Pro or Ser. Pro forms a five-membered nitrogencontaining ring, a feature that causes it to be found in very tight turns in protein structures (i.e. where the polypeptide chain must change direction). Ser, with a slightly polar nature, small size, and propensity of the side-chain hydroxyl oxygen to H-bond with the protein backbone, also causes it to be found in association with tight turns of the protein structure.
Several conservative modifications were present in regions that might impact the putative Cys 14 -Cys 199/200 bridge. These modifications appear to participate in recreating the bridge-destabilizing effect of Lys 191 in the rodent, suggesting that the effect in human is the impact of multiple evolutionary changes. Among the conserved modifications, Leu 224/225 was changed to Phe 225 only in canines, whereas only in humans was Val 300 converted to Leu. There also appeared to be a relation between Asp 301/302 because this residue is Glu 301 in rodents lacking an insertion at position 191. The ability of particular residues to have an impact at such an apparent distance is both a tribute to the interactive nature of this G protein-coupled receptor and the failure of two-dimensional representations to portray accurately the proper spatial relations.
The progressive decrease of PME within mammals began in premammalian classes. The GnRHR in fish and birds (species that produce large numbers of offspring at low metabolic energy per unit and low survival and have a single gonadotropic hormone) contains a cytoplasmic carboxyl-terminal tail that increases net expression and effector coupling efficiency of the receptor at the plasma membrane (11,22). Truncation of the tail in these species results in diminished PME, whereas production of a chimeric structure of a rodent GnRHR con-taining the catfish tail increases PME (11,22). Piscine and avian GnRHRs also lack the cysteines needed to bridge the NH 2 -terminal region and the ECL2; such residues are absolutely conserved among mammalian GnRHRs. Because the number of receptors per cell expressed on the plasma membrane (23,24) cycles and thereby controls gonadotropin levels and in turn ovulation, it may be that the advent of complex primate ovarian cycling requires the ability to regulate promptly the GnRHR without the requisite time for translation and transcription.
The observation that hGnRHR is inefficiently expressed at the plasma membrane is intriguing, because this is the result of a progressive and convergent evolutionary trend toward diminished PME of the GnRHR, adds metabolic cost to create unused receptors, and creates an increased sensitivity to mutation. Potentially, this mechanism provides a source of GnRHR needed for rapid availability without transcription or translation. A similar mechanism appears to regulate the human ␦ opioid receptor because permeable agonists and antagonists also facilitate post-translational processing and increased export of the ligand-stabilized receptor from the endoplasmic reticulum to the cell surface (25). Other reports indicate that other receptors (GluR1, ␣ 1 D adrenoreceptor, odorant, and Luteinizing hormone receptors) are likewise inefficiently expressed at the plasma membrane (26 -30); this suggests that restricted trafficking may be a more commonly occurring means of regulating protein availability than presently appreciated.