Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

The third intracellular loop of GPCR has been identified as a region that is crucial for receptor/G protein interactions (Strader et al., 1987; Wong et. al., 1990; Cotecchia et al., 1992). To discern the structure-activity relationships of the different domains of the alpha-1B adrenergic receptor, a series of mutations were introduced into the amino acid sequence of this receptor (Cotecchia et al., 1990). It was observed that mutations in the third intracellular loop produced profound changes in the apparent affinity of epinephrine for the receptor. Furthermore, the mutant receptors displayed an increase in basal "constitutive" activation of IP production. A single mutation in amino acid residue 293 was identified as being capable of conferring both the increased agonist affinity and constitutive activity (Kjelsberg et al., 1992). However, the muscarinic M₅ receptor was mutated in the same region of the third intracellular loop analogous to amino acid 293 in the alpha-1B adrenergic receptor and was devoid of constitutive activity (Burstein et al., 1995). Thus, this region may not be a critical determinant in receptor function in some GPCRs and/or not all of the members of GPCR family may exhibit the property of constitutive activity.

Recently, several GPCR have exhibited constitutive activity in vivo. Naturally occurring amino acid mutations in the luteinizing hormone receptor (Shenker et al., 1993), parathyroid hormone receptor (Schipani et al., 1995), melanocyte-stimulating hormone receptor (Robbins et al., 1993), thyrotropin receptor (Parma et al., 1993) and rhodopsin receptor (Rao et al., 1994) result in constitutive activation of second messenger production and are thought to play a role in disease pathophysiology. In vitro, amino acid substitutions have produced constitutively active adrenergic receptors (Kjelsberg et al., 1992; Samama et al., 1993; Ren et al., 1993), muscarinic receptors (Spalding et al., 1995), dopamine D₅ receptors (Tiberi and Caron, 1994), serotonin 5-HT_2C receptors (Herrick-Davis et al., 1997), angiotensin II receptors (Groblewski et al., 1997), prostaglandin E receptors (Hasegawa et al., 1996) and thrombin receptors (Nanevicz et al., 1996).

Because the serotonin 5-HT_2A receptor has been implicated in schizophrenia, depression, anxiety and suicide (Meltzer et al., 1989; Cowen, 1991; Meltzer and Nash, 1991), we were particularly interested in determining whether the 5-HT_2A receptor could be rendered constitutively active by amino acid mutation. This receptor mediates many effects of 5-HT in the central nervous system and periphery through stimulation of phospholipase C (Conn and Sanders-Bush, 1984; de Courcelles et al., 1985).

We aligned the amino acid sequences of the alpha-1B adrenergic receptor and the 5-HT_2A receptor and found that a cysteine at position 322 of the 5-HT_2A receptor resides in the same critical position of the third intracellular loop as alanine 293 of the alpha-1B adrenergic receptor. We tested the hypothesis that mutation of amino acid 322 would produce a constitutively active 5-HT_2A receptor. Amino acid 322 was mutated to lysine (C322K), glutamate (C322E) and arginine (C322R) based on studies by Kjelsberg et al. (1992), which show the greatest degree of constitutive activation of the alpha-1B adrenergic receptor by these amino acids. In the present study, the mutant 5-HT_2A receptors exhibited an increase in agonist affinity and potency, and an increase in basal IP production. These studies indicate that the 5-HT_2A receptor can be rendered constitutively active and the C-terminal region of the third intracellular loop is important in 5-HT_2A receptor function.

Many antipsychotic drugs are potent 5-HT_2A receptor antagonists,which lead to the hypothesis that 5-HT_2A receptor blockade is a key mechanism in the efficacy of antipsychotic drugs (Meltzer et al., 1989; Kane et al., 1988; Meltzer, 1995). Therefore, we investigated the activity of five antipsychotic drugs on the constitutively activated C322K 5-HT_2A receptor. All antipsychotic drugs tested displayed robust inverse agonist activity.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Site-directed mutagenesis. The rat 5-HT_2A receptor cDNA (generously donated by Dr. David Julius) was ligated into the mammalian expression vector pcDNA3 (Invitrogen, San Diego, CA) with EcoR1 (Gibco, Grand Island, NY). This construct served as the native template for site-directed mutagenesis performed with Clontech's transformer kit. Mutagenic primers (Midland Certified Reagent Company, Midland, TX) were designed complementary to amino acid 318-330 of the native 5-HT_2A cDNA, whereas changing amino acid 322 from cysteine (TGC) to lysine (AAG), glutamate (GAG) or arginine (AGG). The C322K primer introduced a unique SCAI restriction site at amino acid 324 by changing the third base in the codon for valine from GTG to GTA. Similarly, the C322E primer introduced an RSA 1 site at amino acid 319 and the C322R created an Mnl1 site at amino acid 324. The selection primer, complementary to bases 4871 to 4914 of the pcDNA3 vector, was designed to remove a unique PVUI site by changing base 4891 from G to T. Phosphorylated primers (0.1 µg) were annealed to 10 ng of alkaline-denatured plasmid template by heating to 65°C for 5 min and cooling slowly to 37°C. Mutant DNA was synthesized with T4 DNA polymerase and ligase (Clontech, Palo Alto, CA) for 1 hr at 37°C, followed by digestion with PVU1 (Gibco) and transformation of BMH71-18 mutS Escherichia coli (Clontech). Plasmid DNA was purified with use of the Wizard miniprep kit (Promega, Madison, WI), digested with PVU1 and used to transform DH5 E. coli (Gibco). Individual bacterial colonies were screened for C322K, C322E or C322R by digestion with Sca1, Rsa1, or Mnl1, respectively. DNA sequencing (Sequenase version 2.1 USB; ³⁵S d-ATP, New England Nuclear, Boston, MA) was performed to confirm the mutations at amino acid 322. Sequencing reactions were run on a 5% acrylamide/bis (19:1) gel (Bio-Rad, Hercules, CA) for 2 hr at 50°C, dried for 2 hr at 80°C and exposed on Kodak Biomax MR film for 24 hr at -80°C.

Cell culture and transfection. COS-7 cells were grown in DMEM (Sigma, St. Louis, MO) with 10% fetal bovine serum (Sigma) in 5% CO₂ at 37°C and subcultured 1:8 twice a week. Twenty-four hours before transfection, cells were seeded at 80% confluence in 100-mm dishes for radioligand binding assays or at 10⁵ cells per well in 24-well cluster plates for IP production assays. Cells were transfected with native or mutant 5-HT_2A cDNA by Lipofectamine (Gibco). This was accomplished by combining 20 µl of Lipofectamine with 2.5 µg of plasmid per 100-mm dish or 2 µl of Lipofectamine with 0.25 µg of plasmid per well. Transfections were performed in serum-free DMEM for 4 hr at 37°C.

Radioligand binding. Thirty-six hours after transfection, membranes were prepared from COS-7 cells by scraping in 50 mM Tris-HCl/5 mM MgCl₂/0.5 mM ethylenediaminetetraacetic acid, pH 7.4 (assay buffer) and centrifugation at 10,000 × g for 30 min. Membranes were resuspended in assay buffer, homogenized and centrifuged again. After resuspension in assay buffer, 1-ml membrane aliquots (~10 µg of protein measured by bicinchoninic acid assay) were added to 1 ml of assay buffer with 0.5 nM [³H]ketanserin and competing drugs. Spiperone (10 µM) was used to define nonspecific binding. Saturation experiments were performed with [³H]ketanserin (0.1-5.0 nM). Samples were incubated at 37°C for 30 min, filtered on a Brandel cell harvester and counted in Ecoscint cocktail (National Diagnostics, Atlanta, GA) in a Beckman liquid scintillation counter at 40% efficiency.

PI hydrolysis. IP production was measured as described previously (Herrick-Davis et al., 1997). Twenty-four hours after transfection cells were washed with phosphate-buffered saline and labeled with 0.25 µCi/well of myo-[³H]inositol (New England Nuclear) in inositol-free/serum-free DMEM (Gibco) for 18 hr at 37°C. High-performance liquid chromatography analysis of this culture medium, after incubation, was reported to contain <10¹⁰ M 5-HT (Barker et al., 1994). After labeling, cells were washed with phosphate-buffered saline and preincubated in inositol-free/serum-free DMEM with 10 mM LiCl and 10 µM pargyline (assay medium) for 10 min at 37°C. When antagonists were used, they were added during the 10-min preincubation period. Agonists were added to each well and incubation continued for an additional 30 min. Assay medium was removed and cells were lysed in 200 µl of stop solution (1 M KOH/18 mM sodium borate/3.8 mM ethylendiaminetetraacetic acid) and neutralized by adding 200 µl of 7.5% HCl. The contents of each well were extracted with 3 volumes of chloroform/methanol (1:2), centrifuged 5 min at 10,000 × g and the upper layer loaded onto a 1-ml AG1-X8 resin (100-200 mesh, Bio-Rad) column. Columns were washed with 10 ml of 5 mM myoinositol and 10 ml of 5 mM sodium borate/60 mM sodium formate. Total IPs were eluted with 3 ml of 0.1 M formic acid/1 M ammonium formate. Radioactivity was measured by liquid scintillation counting in Ecoscint cocktail.

Data analysis. Data were analyzed by GraphPad Prism. Statistical analyses were performed by the Student's t test and one-way analysis of variance.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

To determine if the 5-HT_2A receptor could be rendered constitutively active, amino acid 322 (analogous to amino acid 293 in the alpha-1B adrenergic receptor) was mutated from cysteine to C322K, C322E and C322R. These three amino acids were chosen based on the large increase in basal activity produced by the same amino acid substitutions in the alpha-1B adrenergic receptor (Kjelsberg et al., 1992). Figure 1 shows 5-HT competition binding curves for both native and mutant 5-HT_2A receptors radiolabeled with H-ketanserin. Whereas the native receptor demonstrated a relatively low affinity for 5-HT (K_i, 293 nM), the three mutant receptors displayed high affinity for 5-HT. The C322K mutant exhibited a 12-fold increase in 5-HT affinity (K_i, 25 nM), the C322R mutant had a 27-fold higher affinity for 5-HT (K_i, 11 nM) and the C322E mutant had a 3-fold higher affinity for 5-HT (K_i, 86 nM). Two-site analysis revealed that the native 5-HT_2A receptor was best-fit to a two-site model with 38% of the receptors in a high-affinity state with a Hill coefficient of 0.64 (K_H, 60 nM; K_L,1.21 µM). The C322E mutant was also best fit to a two-site model with 66% of the receptors in a high-affinity state with a Hill coefficient of 0.80 (K_H, 50 nM; K_L, 1.92 µM). The C322K and C322R mutant 5-HT_2A receptors were best fit to a one-site model with 90 to 100% of the receptors in a high-affinity state with a Hill coefficient close to unity (1.0).

View larger version (20K): [in this window] [in a new window]   Fig. 1.   5-HT competition for 3H-ketanserin binding to native and mutant 5-HT2A receptors. 3H-ketanserin (0.5 nM) was used to label native and mutant receptors transiently expressed in COS-7 cells. The data represent the mean ± S.E.M. of three separate experiments performed in triplicate.

To determine whether other 5-HT_2A agonists would display a similar increase in affinity for the mutant 5-HT_2A receptors, DOB and DOM were examined at native and C322K 5-HT_2A receptors. Similar to 5-HT, DOB and DOM displayed increased affinity for C322K receptors (table 1). These data are consistent with previous studies demonstrating constitutively active mutant alpha-1B adrenergic receptors with higher affinity for epinephrine and other adrenergic agonists (Kjelsberg et al., 1992).

C322K and C322E mutant 5-HT_2A receptors demonstrated robust increases in basal IP production relative to the native 5-HT_2A receptor, which exhibited very low levels of basal activity (fig. 2). The C322K mutant demonstrated an 8-fold increase in basal activity compared with native, whereas the C322E mutant was less active, exhibiting a 3.7-fold increase in basal activity over the native 5-HT_2A receptor. The basal activity of the C322K mutant was 48% of the native 5-HT_2A receptor maximally stimulated with 10 µM 5-HT, whereas the C322E mutant had a basal activity that was 31% of the maximally stimulated 5-HT_2A receptor. On the addition of 10 µM 5-HT, both the native and mutant receptors produced a further increase in IP production. However, there was no significant difference in maximal 5-HT stimulation of the native and mutant 5-HT_2A receptors.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   IP production in COS-7 cells expressing native or mutant 5-HT2A receptors. Total IP production was measured in COS-7 cells expressing vector, native 5-HT2A receptors or mutant 5-HT2A receptors. Basal IP production in cells transfected with vector only was subtracted from basal IP production in cells transfected with native or mutant 5-HT2A receptors. Basal activity produced by vector alone was typically 400 dpm. Data represent the mean ± S.E.M. of three to six experiments performed in triplicate. *P < .01 vs. native basal; **P < .01 vs. C322E basal; ***P < .01 vs. C322K basal.

Saturation binding studies were performed to ensure that the increased basal activity was not caused by increased expression of the mutant receptors (fig. 3). B_MAX values for cells expressing native 5-HT_2A receptors were not significantly different from B_MAX values for mutant receptors (native, 193 ± 37(S.E.M.) fmol/mg; C322K, 218 ± 31 fmol/mg; C322E, 274 ± 25 fmol/mg; C322R, 131 ± 20 fmol/mg).

View larger version (19K): [in this window] [in a new window]   Fig. 3.   Saturation analyses of 3H-ketanserin-labeled native and C322K mutant receptors. Data represent the mean ± S.E.M. of one experiment performed in triplicate represent three separate experiments. Protein levels were determined by the BCA assay (Pierce).

To determine whether the mutant 5-HT_2A receptors would exhibit an increase in agonist potency relative to the native 5-HT_2A receptor, 5-HT stimulation of the native and mutant 5HT_2A receptors was measured by use of an IP production assay. Figure 4 reveals that the C322K and C322E mutant receptors exhibited a leftward shift in the 5-HT dose-response curve, which indicates that there was an increase in 5-HT potency for the mutant receptors. 5-HT had EC₅₀ values of 25 ± 6.1 nM, 61 ± 1.8 nM and 152 ± 1.9 nM for C322K, C322E and native 5-HT_2A receptors, respectively.

View larger version (18K): [in this window] [in a new window]   Fig. 4.   5-HT stimulation of IP production in COS-7 cells expressing native or mutant 5-HT2A receptors. Cells were labeled with myo-[3H]- inositol and challenged with increasing concentrations of 5-HT. Total IP production was measured by anion-exchange chromatography. The data are expressed as a percentage of maximal IP production stimulated by 10 µM 5-HT. Data are the mean ± S.E.M. of three separate experiments.

5-HT_2A receptor antagonists were examined to determine whether these drugs would exhibit a change in binding affinity for C322K receptors. Previous studies reported that certain antagonists with inverse agonist activity have lower affinity at constitutively active receptors (Samama et al., 1993). In the present study, there was a decrease in the binding affinity of methysergide and mianserin for C322K receptors, but no change in binding affinity for spiperone and ketanserin (table 1). Haloperidol and clozapine demonstrated 3-fold lower and 2-fold lower binding affinities for C322K receptors, whereas chlorpromazine and risperidone displayed no change in binding affinity.

To determine whether 5-HT_2A antagonists possess inverse agonist properties, the ability of two classical antagonists, spiperone and ketanserin, to reverse the basal activity of the C322K mutant was examined. Ketanserin (1 µM) and spiperone (1 µM) decreased C322K basal IP production by 80% and 65%, respectively (data not shown). These two drugs had no effect on basal activity of the native 5-HT_2A receptor.

To determine whether antipsychotic drugs display inverse agonist activity, the ability to inhibit the basal activity of C322K mutant receptors was evaluated (table 2). Five antipsychotic drugs were tested, each at a 1 µM concentration. All the drugs tested displayed inverse agonist activity by significantly reducing basal IP production. Maximal inhibition of basal activity ranged from 59% to 82%. However, although all the antagonists tested had robust inverse agonist activity, these drugs produced maximal levels of inhibition that were less than 100%. In parallel, recombinant cells were transfected with the native 5-HT_2A receptor to determine whether antipsychotic drugs could reduce the basal activity produced by the native 5-HT_2A receptor. Antipsychotic drugs apparently had no effect on basal IP production in recombinant cells transiently expressing native 5-HT_2A receptors (data not shown). However, basal levels of IP production by native receptors were very low, which makes it difficult to assay for inverse agonist activity.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

Constitutive receptor activity has been demonstrated for several GPCRs. The alpha-1B adrenergic, alpha-2 adrenergic and beta-2 adrenergic receptors have been mutated into forms that demonstrate constitutive activity in stimulating PI metabolism, inhibiting adenylate cyclase activity and stimulating adenylate cyclase activity, respectively (Kjelsberg et al., 1992; Ren et al., 1993; Samama et al., 1993). In vivo, constitutive activity of mutant forms of receptors such as the thyrotropin receptor, rhodopsin receptor and the luteinizing hormone receptor have been demonstrated through elevated basal levels of adenylate cyclase activity (Parma et al., 1993; Rao et al., 1994; Shenker et al., 1993). The native D_1B dopamine, 5-HT_2C serotonin, and muscarinic receptors also demonstrated some degree of constitutive activity (Soejima and Noma, 1984; Tiberi and Caron, 1994; Barker et al., 1994). In these cases, the constitutive activity was demonstrated through the actions of inverse agonists that were able to reduce the basal activity of the native receptors.

In this study, the mutant 5-HT_2A receptors reveal radioligand binding properties that are indicative of a constitutively active receptor. The 5-HT_2A receptor/G-protein complex generally displays higher affinity for agonists than the free, uncoupled receptor (Teitler et al., 1990). Our data show that 5-HT and other agonists display an increased affinity for the mutant receptors. Computer-assisted analyses of serotonin competition curves indicate that the major effect of the mutation was to shift the conformation of the receptor into an agonist high-affinity state. Our data support the modified allosteric ternary complex model put forth by Samama et al. (1993) in which GPCR exist in an equilibrium between inactive and active conformational states. Further experiments need to be performed to determine whether the constitutively active 5-HT_2A receptor can exist in multiple high-affinity states as seen in the rhodopsin receptor and was theorized for other GPCR (Hamdorf and Kirshfeld 1980; Kenakin, 1995).

The C322K and C322E mutant receptors that were created produced significant increases in basal IP production compared with the native 5-HT_2A receptor, which indicates that they exhibited constitutive activity (fig. 2). There appears to be a relationship between the binding affinities of 5-HT and the increased basal activity of the respective mutant 5-HT_2A receptors. The C322E mutant had the smallest increase in binding affinity for 5-HT and apparently is the least constitutively active mutant. Therefore, the cysteine in amino acid position 322 apparently is critical as a constraint for the native 5-HT_2A receptor, and when this constraint is removed by mutation, varying degrees of constitutive activity will result. In the three mutant 5-HT_2A receptors that were assayed, it appears that the C322K and C322R mutants produced receptors that most closely resembled the receptor stimulation by an agonist, or the "active conformation." Whereas the C322K produced the highest levels of IPs of the mutant 5-HT_2A receptors that were tested, when 5-HT was added, this mutant receptor produced the smallest increase in IPs relative to the native 5-HT_2A receptor. Because these receptors are already in a highly favorable energetic conformation, the addition of an agonist only causes a small increase in IP levels compared with the native receptor, which produces a 10-fold increase in IP levels on 5-HT addition. It appears that the C322K and C322R mutations cause these receptors to exist in a conformation that is more amenable to interaction with a G protein than the C322E mutant or the native receptor. On stimulation with 5-HT, all of the receptors then undergo the full conformational change allowing for maximal interaction with the G protein.

Radioligand binding analyses indicate that the constitutive activity of the mutant 5-HT_2A receptors is not caused by an overexpression of mutant receptors because there was no difference in B_MAX values between native and mutant receptors (fig. 3). Our results were not caused by an increase in affinity of the mutant 5-HT_2A receptors for residual serotonin because all the studies were performed in serum-free media to remove any exogenous 5-HT. Barker et al. (1994) have reported that the concentration of 5-HT present in the absence of serum is less than 10¹⁰ M as determined by high-performance liquid chromatography analysis. This low concentration of 5-HT could not account for the constitutive activity that was seen in this study.

5-HT exhibited dramatic increases in potency at C322K and C322E mutant 5-HT_2A receptors (fig. 4). The shifts in the dose-response curves were similar to the changes in binding affinity found in radioligand binding studies. The C322E mutant had the smallest change in 5-HT binding affinity and exhibited a smaller shift in 5-HT potency than the C322K mutant.

Because of the involvement of the 5-HT_2A receptor in the clinical effectiveness of antipsychotic drugs, we determined the activity of antipsychotic drugs at the activated state of the 5-HT_2A receptor. As shown in table 2, all the antipsychotic drugs tested decreased the basal activity of C322K receptors, which indicates that they all were inverse agonists. Although these drugs appear to be potent inverse agonists, they did not produce a complete inhibition of the constitutive IP production.

The inverse agonists that were tested did not appear to inhibit the basal IP production of the native 5-HT_2A receptor. Although figure 2 shows that there was a small increase in native 5-HT_2A receptor basal IP levels above the vector control, this increase was barely detectable in our system because of the low level of radioactivity represented by the native receptor basal activity (approximately 300 dpm). Therefore, it was difficult to measure a decrease in basal IP levels produced by inverse agonists at the native receptor. It may be possible to see small changes in IP levels at low levels of basal activity with use of a more sensitive indicator of IP levels, such as the IP₃ radioreceptor assay.

Preliminary studies have indicated that mutant, constitutively active forms of GPCR may be responsible for several disease states including thyroid adenomas (Parma et al., 1993) and night blindness (Rao et al. 1994). The prevalence and significance of different forms of GPCRs, either native or mutant, exhibiting different degrees of constitutive activity appears to be a largely unexplored area of research. If it is found that constitutively active receptors are expressed as part of the normal genome or because of mutations in the genome, drugs that have inverse agonist properties may have significant therapeutic potential by reversing deleterious effects produced by constitutively active GPCR. Because of the numerous physiological functions and pathophysiological conditions involving the 5-HT_2A receptor in the central nervous system and periphery, it is possible that activating mutations of the 5-HT_2A receptor may be responsible for disease states. In a related study, we showed that the 5-HT_2C receptor can be mutated to a constitutively active state by site-directed mutagenesis (Herrick-Davis et al., 1997). The results of that study and the results presented herein indicate that the members of the 5-HT₂ receptor family are sensitive to amino acid mutations resulting in drastic alterations in receptor activity.

In summary, the present study indicates that the 5-HT_2A receptor can be mutated to a constitutively activated form by a single amino acid mutation. The increased activity of the mutated receptor can be used to reveal inverse agonist activity of drugs that appear to be neutral antagonists at native receptors in recombinant cell systems. These studies show that the antipsychotic drugs tested have inverse agonist activity at the activated state of the 5-HT_2A receptor, which may contribute to their unique therapeutic profile.