Molecular determinants for the interaction of the valvulopathic anorexigen norfenfluramine with the 5-HT2B receptor

Article abstract of DOI:10.1124/mol.104.009266

S-(+)-Norfenfluramine (SNF)-an active metabolite of the now-banned anorexigen fenfluramine-has been implicated in the drug's appetite-suppressing actions and its life-threatening cardiovascular side effects. SNF reduces appetite through serotonin 5-HT_2C receptor activation; it causes cardiopulmonary side effects through 5-HT_2B receptor activation. Thus, we attempted to identify molecular determinants of SNF binding to 5-HT_2B receptors distinct from those underlying SNF-5-HT _2C/2A receptor interactions. Mutagenesis implicated Val2.53 in SNF binding to 5-HT_2B receptors. Ligand docking simulations suggested both Val2.53 γ-methyl groups form stabilizing van der Waals' (vdW) interactions with the α-methyl group of SNF. A V2.53L mutation induced a 17-fold decrease in affinity; molecular dynamics (MD) simulations suggested that this decrease resulted from the loss of one 2.53-α-methyl group vdW interaction. Supporting this, 1) the binding of norfenfluramine (NF) analogs lacking an S-(+) α-methyl group (RNF and α-desmethyl-NF) was less sensitive to the V2.53L mutation, and 2) a V2.53A mutation decreased SNF affinity 190-fold, but decreased RNF and α-desmethyl-NF affinities only 16- and 45-fold, respectively. We next addressed whether the α-methyl group of SNF contributes to 5-HT_2C/2A receptor affinity. Removal of the α-methyl group (RNF and α-desmethyl-NF), which reduced 5-HT _2B receptor binding 3-fold, did not affect 5-HT_2C/2A receptor binding. An α-ethyl substituent (α-ethyl-NF), which decreased 5-HT_2B receptor affinity 46-fold, reduced 5-HT_2C and 5-HT_2A receptor binding by 14- and 5-fold, respectively. Finally, we determined that residue 2.53 affects SNF potency and efficacy at 5-HT_2B receptors but not at 5-HT_2C and 5-HT_2A receptors. In conclusion, vdW interactions between residue 2.53 and the α-methyl group of SNF contribute to the ligand's 5-HT₂ receptor subtype-selective pharmacology. Copyright

Full text of DOI:10.1124/mol.104.009266

0
026-895X/05/6801-20–33$20.00
MOLECULAR PHARMACOLOGY
Copyright © 2005 The American Society for Pharmacology and Experimental Therapeutics
Mol Pharmacol 68:20–33, 2005
Vol. 68, No. 1
9266/3039404
Printed in U.S.A.
Molecular Determinants for the Interaction of the Valvulopathic
Anorexigen Norfenfluramine with the 5-HT Receptor
2
B
Vincent Setola, Malgorzata Dukat, Richard A. Glennon, and Bryan L. Roth
Departments of Biochemistry (V.S., B.L.R.), Psychiatry (B.L.R.), and Neurosciences (B.L.R.), Case Western Reserve University
School of Medicine, Cleveland, Ohio; and Department of Medicinal Chemistry, School of Pharmacy,
Virginia Commonwealth University, Richmond, Virginia (M.D., R.A.G.)
Received November 12, 2004; accepted April 14, 2005
ABSTRACT
S-(ϩ)-Norfenfluramine (SNF)—an active metabolite of the now- analogs lacking an S-(ϩ) ␣-methyl group (RNF and ␣-des-
banned anorexigen fenfluramine—has been implicated in the methyl-NF) was less sensitive to the V2.53L mutation, and 2) a
drug’s appetite-suppressing actions and its life-threatening V2.53A mutation decreased SNF affinity 190-fold, but de-
cardiovascular side effects. SNF reduces appetite through se- creased RNF and ␣-desmethyl-NF affinities only 16- and 45-
rotonin 5-HT2C receptor activation; it causes cardiopulmonary fold, respectively. We next addressed whether the ␣-methyl
side effects through 5-HT2B receptor activation. Thus, we at- group of SNF contributes to 5-HT2C/2A receptor affinity. Re-
tempted to identify molecular determinants of SNF binding to moval of the ␣-methyl group (RNF and ␣-desmethyl-NF), which
5
-HT2B receptors distinct from those underlying SNF-5-HT2C/2A reduced 5-HT2B receptor binding 3-fold, did not affect
receptor interactions. Mutagenesis implicated Val2.53 in SNF 5-HT2C/2A receptor binding. An ␣-ethyl substituent (␣-ethyl-
binding to 5-HT2B receptors. Ligand docking simulations sug- NF), which decreased 5-HT2B receptor affinity 46-fold, reduced
gested both Val2.53 ␥-methyl groups form stabilizing van der 5-HT2C and 5-HT2A receptor binding by 14- and 5-fold, respec-
Waals’ (vdW) interactions with the ␣-methyl group of SNF. A tively. Finally, we determined that residue 2.53 affects SNF
V2.53L mutation induced a 17-fold decrease in affinity; molec- potency and efficacy at 5-HT2B receptors but not at 5-HT2C and
ular dynamics (MD) simulations suggested that this decrease 5-HT2A receptors. In conclusion, vdW interactions between
resulted from the loss of one 2.53-␣-methyl group vdW inter- residue 2.53 and the ␣-methyl group of SNF contribute to the
action. Supporting this, 1) the binding of norfenfluramine (NF) ligand’s 5-HT receptor subtype-selective pharmacology.
2
Serotonin (5-hydroxytryptamine, 5-HT) is a biogenic amine (Meltzer, 1989; Roth et al., 2004), most hallucinogens (Glen-
that regulates a broad spectrum of processes in both the non et al., 1984; Roth et al., 2002; Nichols, 2004), and some
central nervous system and in the periphery via no fewer appetite suppressants (Moses and Wurtman, 1984; Neill and
than 15 plasma membrane receptors divided into seven fam- Cooper, 1989; Vickers et al., 2001).
ilies (Kroeze et al., 2002, 2003). All but one family belong to
The appetite suppressant fenfluramine, which is very ef-
the class A, rhodopsin-like G protein-coupled receptor super- fective in the treatment of obesity (Weintraub et al., 1992),
family. The 5-HT2 family of receptors, comprising the was withdrawn from the U.S. marketplace because of its
5
-HT2A, 5-HT2B, and 5-HT2C subtypes, represents one of the association with valvular heart disease (VHD) (Connolly et
best characterized groups of 5-HT receptors (Roth et al., al., 1997). Fenfluramine was also known to be associated
998; Kroeze et al., 2002). These receptors are the major with pulmonary hypertension (PH) (Pouwels et al., 1990;
targets/sites of action of atypical antipsychotic medications Brenot et al., 1993; Connolly et al., 1997; Simonneau et al.,
998). Fenfluramine-induced VHD and PH result from the
1
1
activation of mitogenic 5-HT2B receptors on heart valve and
pulmonary artery interstitial cells (Fitzgerald et al., 2000;
K02-MH01366, and the National Institutes of Mental Health Psychoactive Rothman et al., 2000; Launay et al., 2002; Setola et al., 2003),
This work was supported by an American Heart Association predoctoral
fellowship (to V.S.) and by National Institutes of Health grants R01-MH57635,
Drug Screening Program (to B.L.R.).
Article, publication date, and citation information can be found at
http://molpharm.aspetjournals.org.
doi:10.1124/mol.104.009266.
leading to the formation of proliferative fibromyxoid plaques
that compromise tissue integrity and function (Steffee et al.,
1999; Tomita and Zhao, 2002). Fenfluramine-induced an-
ABBREVIATIONS: 5-HT, 5-hydroxytryptamine (serotonin); VHD, valvular heart disease; PH, pulmonary hypertension; SNF, S-(ϩ)-norfenfluramine;
NF, norfenfluramine; MD, molecular dynamics; vdW, van der Waals; LSD, d-lysergic acid diethylamide ; FBS, fetal bovine serum; DMEM,
Dulbecco’s modified Eagle’s medium; RNF, R-(Ϫ)-norfenfluramine; 3-D, three-dimensional; CSA, computer-simulated annealing; WT, wild-type;
␣-H, ␣-carbon hydrogen substituent.
2
0

Residues Involved in Norfenfluramine Binding to 5-HT2B Receptors
21
orexia, in contrast, seems to be mediated by activation of hydrochloride salt (m.p., 161–163°C after recrystallization from
hypothalamic 5-HT2C receptors (Vickers et al., 2001; Heisler 2-propanol). The assigned structure was consistent with proton NMR
data, and the salt analyzed (Atlantic Microlab, Norcross, GA) to
within 0.4% of theory for C, H, and N.
Chemicals. Optically pure S-(ϩ)- and R-(Ϫ)-NF were provided by
Richard B. Rothman (National Institute on Drug Abuse, Baltimore,
et al., 2002). Thus, novel anorexigens that activate 5-HT2C
receptors and are devoid of 5-HT2B receptor activity are
expected to be safe and effective treatments for obesity; the
elucidation of 5-HT2B receptor-unique SNF-receptor intermo-
lecular interactions could facilitate the rational design of
such 5-HT2C receptor-selective agents.
MD). LSD, pergolide, dihydroergotamine, serotonin (5-HT), and
␣
-methyl-5-HT were all purchased from Sigma-Aldrich (St. Louis,
MO).
Cell Culture and Transfection. Human embryonic kidney 293T
In the present report, we identify residue 2.53 in 5-HT2
receptors as playing an important role in the subtype selec- cells were maintained at 37°C, 5% CO2 in 75-cm flasks (Costar,
tive in vitro pharmacology of S-(ϩ)-norfenfluramine (SNF). Cambridge, MA) with Dulbecco’s modified Eagle’s medium contain-
Site-directed mutagenesis, SNF and NF congener binding ing 100 mM sodium pyruvate, 100 U/ml penicillin, 100 ␮g/ml strep-
tomycin (DMEM) and 10% fetal bovine serum (DMEM/10% FBS)
studies, molecular modeling, and ligand docking and MD
(
all from Invitrogen, Carlsbad, CA). At confluence, cells were
simulations all suggest that both terminal methyl groups of
Val2.53 in the 5-HT2B receptor form stabilizing van der
Waals’ (vdW) interactions with the ␣-methyl group of SNF.
In addition, we provide functional data demonstrating that
these interactions are also important for SNF-mediated ac-
tivation of the 5-HT2B receptor. We also present evidence
trypsinized, harvested, and split 1:5 into four 10-cm plates contain-
ing 10 ml of DMEM/10% FBS (for transfection), and one 75-cm flask
containing 15 ml of DMEM/10% FBS (for cell maintenance). One day
after seeding into 10-cm dishes (Costar) (i.e., at ϳ90% confluence),
the medium was removed and replaced with 10 ml of OptiMEM
(Invitrogen). Each 10-cm plate was transfected with 24 ␮g of receptor
that the role of residue 2.53 on SNF binding to, and activa- plasmid using Lipofectamine 2000 (Invitrogen) exactly as specified
tion of, the 5-HT2B receptor is unique among 5-HT2 re- by the manufacturer. Twenty-four hours after transfection, cells
ceptors.
were processed for either radioligand binding assays or inositol phos-
phate accumulation assays as described below.
Radioligand Binding Assays. Twenty-four hours after transfec-
tion, the medium was removed and replaced with DMEM containing
Materials and Methods
Receptor Constructs and Site-Directed Mutagenesis. The 5% dialyzed FBS (Invitrogen). After 24 h, the cells were incubated
human 5-HT2A, 5-HT2B, and 5-HT2C (INI editing isoform) receptors overnight in serum-free DMEM. The next day, the cells were har-
having been described previously (Rothman et al., 2000) were sub- vested by scraping, pelleted, and resuspended in lysis buffer (50 mM
cloned into the NotI site of pUniversal-Signal, a pIRES-neo (BD Tris-HCl, pH 6.9). The membranes were then pelleted by centrifu-
Biosciences Clontech)-based expression vector containing the follow- gation and, after removal of the supernatant, the membrane fraction
ing elements: a Kozak sequence followed by a cleavable membrane was frozen at Ϫ80°C (if not used immediately). Radioligand binding
targeting sequence, followed by a FLAG epitope tag, all in frame with
assays were set up in 24 wells of a 96-well plate (1 ml/well capacity)
3
each other (Bhatnagar et al., 2004). Site-directed mutagenesis was as follows: 25 ␮l of 10 ␮M [ H]LSD (PerkinElmer Life and Analytical
performed using the QuikChange kit (Stratagene, La Jolla, CA) Sciences, Boston, MA), 25 ␮l of membrane pellet that had been
according to the manufacturer’s instructions. Mutagenized receptor resuspended in 700 ␮l of ice-cold binding buffer (50 mM Tris-HCl, 5
cDNAs were sequenced by automated, PCR-based sequencing mM MgCl
, and 0.5 mM EDTA, pH 7.4), 25 ␮l of vehicle (binding
2
(
Cleveland Genomics) to verify the introduction of the desired mu- buffer) or 10ϫ test compound dissolved in binding buffer (i.e., SNF,
tation and the absence of PCR-induced mutations.
RNF, ␣-desmethyl-NF, ␣-R,S-(Ϯ)-ethyl-NF, 5-HT, ␣-methyl-5-HT,
Chemical Synthesis of Norfenfluramine. Analog ␣-des- dihydroergotamine, or pergolide) at various concentrations spanning
methyl-NF (2-[(3-trifluoromethyl)phenyl]ethylamine hydrochloride) 7 orders of magnitude (two wells/concentration), and 200 ␮l of bind-
3
was prepared by condensation of 3-(triflurormethyl)benzaldehyde ing buffer, such that the final [ H]LSD and test compound concen-
and nitromethane in the presence of ammonium acetate. The crude
trations were 1ϫ. Reactions were equilibrated in the dark for 1 h at
product was purified by Kugelrhor distillation (90°C at 1 mm Hg) to room temperature (ϳ22°C), and then receptor-ligand complexes
give a semisolid material. Although the nitrostyrene has been re- were harvested by vacuum filtration onto Whatman (Florham Park,
ported to be a solid (Menicagli and Samaritani, 1996), the product
NJ) GF/C filters (presoaked in 0.3% polyethylenimine), then washed
failed to crystallize. Reduction of the nitrostyrene intermediate with three times with ice-cold 50 mM Tris-HCl, pH 6.9, using a Brandel
lithium aluminum hydride in dry tetrahydrofuran afforded a 47% 24-well harvester. The filters were dried overnight, and each was
yield of the target compound as its free base after Kugelrohr distil- added to a 6-ml vial, into which 4 ml of EcoScint liquid scintillation
lation. The hydrochloride salt was prepared and recrystallized from cocktail (National Diagnostics, Atlanta, GA) was then added. The
3
2
-propanol (m.p., 100°C). Structure assignment was consistent with total [ H]LSD binding present on each filter was quantified by liquid
proton NMR data, and the hydrochloride salt analyzed (Atlantic scintillation counting using a PerkinElmer Wallac (Gaithersburg,
Microlab, GA) to within 0.4% of theory for C, H, and N. Dengel et al.
MD) liquid scintillation counter. The log KD (for LSD) or log K (for all
i
(
1975) have previously prepared the free base of ␣-desmethyl-NF. other compounds) and Bmax values were determined using Prism 4.0
Branch et al. (2002) have reported the synthesis of the hydrochloride
(GraphPad Software, San Diego, CA) by fitting the average total
3
salt (using a different route of synthesis) in the patent literature, but [ H]LSD binding (in disintegrations per minute) from several inde-
the salt was characterized only by low-resolution mass spectrometry. pendent experiments, plotted as a function of the log [compound], to
The ␣-ethyl homolog of NF has been mentioned only once before a homologous (for LSD) or heterologous (for all other compounds)
(
Fuller et al., 1982), and this report described only the pharmacology competition model that takes into account ligand depletion (Graph-
of the agent, not the synthesis or physicochemical characterization. Pad) and shares the log KD or log K among all data sets.
i
The ␣-ethyl homolog, 1-[(3-trifluoromethyl)phenyl]-2-aminobutane
Inositol Phosphate Accumulation Assays. Twenty-four hours
hydrochloride, was prepared by a method similar to that employed after transfection, cells in 10-cm plates were harvested by trypsiniza-
for the synthesis of ␣-desmethyl-NF. The condensation step used tion and resuspended in 25 ml of DMEM, 5% dialyzed FBS,
nitropropane in place of nitromethane to produce the nitrostyrene; and seeded (1 ml of cell suspension/well) in a 24-well poly-L-lysine-
the latter was converted to the oily free base of the target amine by coated plate. The next day, the medium was removed and replaced
3
lithium aluminum hydride reduction, purified by Kugelrohr distilla- with BME (Invitrogen) containing
tion to obtain a 53% yield of the product, and converted to its
1
␮Ci/ml [myo- H]inositol
(PerkinElmer Life and Analytical Sciences) (500 ␮l/well). Twenty-

2
2
Setola et al.
four hours later, the medium was replaced (1 ml/well) with 1ϫ an exponential temperature-versus-time ramping method. Snap-
Hanks’ balanced salt solution (Invitrogen) containing 25 mM sodium shots were taken every 5 fs. During the MD simulations, the distance
bicarbonate, 11 mM glucose, and 10 mM LiCl (inositol phosphate
between the protonated amine group of the ligand and the nearest
buffer). Agonists (5-HT, SNF, and RNF) were diluted in inositol carboxylate oxygen of Asp3.32, the ligand atoms, and the ␣-carbon
phosphate buffer to 100ϫ and added (10 ␮l/well) to the cells. After a backbone of the receptor, and all atoms greater than 8 Å away from
1
-h incubation at 37°C, the buffer was removed and replaced with 1 the mutated side chain were constrained. For MD calculations dur-
ml of 1 mM formic acid to extract the cytosolic fraction. The [ H]- ing CSA, an AMBER7 FF99 force field was applied with an 8-Å
inositol phopshates accumulated in the cytosol were isolated by nonbonded cut-off and a distance-dependent dielectric constant of 4
3
loading the cytosolic fraction onto Dowex anion exchange resin (for-
mate form) columns, washing the columns twice with water (first
using a nonbonded cut-off. The resulting structure from each round
of CSA was examined with regard to the orientation of the mutated
with 2 ml/column, then with 10 ml/column), then washing with 10 ml residue’s side chain relative to the ligand.
of wash buffer (5 mM sodium borate and 50 mM sodium formate)/
column, and finally eluting with 10 ml of elution buffer (100 mM
formic acid and 200 mM sodium formate)/column into 30-ml vials
Results
containing 10 ml/vial 3a70B liquid scintillation cocktail (Research
Products International, Mt. Prospect, IL). [ H]Inositol phosphate
accumulation in each sample was quantified by liquid scintillation
counting using a PerkinElmer Wallac liquid scintillation counter.
Effect of Point Mutations on Ligand Affinity. To iden-
3
tify residues involved in the subtype selective binding of the
anorexigen SNF to 5-HT2 family receptors, we constructed
every possible 5-HT2B-to-5-HT2C/2A receptor mutant residing
near a putative ligand binding pocket. To do so, we first
generated a molecular model of the human 5-HT2B receptor
The log EC50 and Emax values were determined using Prism 4.0
3
(
GraphPad) by fitting the average total [ H]inositol phosphate from
several independent experiments, plotted as a function of the log
[
agonist], to a three-parameter logistic equation (sigmoidal concen- by homology model building using our bovine rhodopsin-
tration-response) and sharing the log EC50 and Emax among all data based (Palczewski et al., 2000), refined, validated 3-D homol-
sets.
ogy model of the rat 5-HT2A receptor (Shapiro et al., 2002) as
Modeling, Ligand Docking Simulations, and Molecular Dy- a template. To verify that our rhodopsin-based model was
namics Simulations. Three-dimensional models of the human
similar in predicted structure to the published coordinates of
5-HT2B, 5-HT2A, and 5-HT2C receptors were constructed using our
bovine rhodopsin (Palczewski et al., 2000), we superimposed
our 5-HT2B receptor homology model onto 1HZXA (the bovine
rhodopsin PDB file). All absolutely conserved transmem-
brane helix residues aligned and were oriented very similarly
relative to each other (data not shown). We next identified
putative ligand binding residues in the 5-HT2B receptor (i.e.,
rhodopsin-based (Palczewski et al., 2000), refined, validated 3-D
homology model of the rat 5-HT2AR (Shapiro et al., 2002) as a
template. The raw sequence of each receptor was first fit onto the
template using the DeepView (GlaxoSmithKline, Research Triangle
Park, NC) modeling program’s Magic Fit function, then uploaded to
the Swiss Model server (http://swissmodel.expasy.org//SWISS-MOD-
EL.html). The returned Protein Data Bank files were loaded into those residues in the extracellular half of the transmembrane
Sybyl 6.91 modeling and MD software (Tripos, Inc., St. Louis, MO) helices that have their side chains oriented into the inter-
and, after atom typing and calculation of charges, energy-minimized helical space) that were not conserved in linear alignments of
using the Powell method under an AMBER7 FF99 force field with an
the 5-HT receptor transmembrane helices (Fig. 1). We then
2
8-Å nonbonded cut-off and a distance-dependent dielectric constant
measured the affinity constant (K ) of SNF at wild-type and
i
of 4. The local minimum was identified using a Wolfe line search for
successive iterations that differed by less than 0.05 kcal/mol ⅐ Å. Ten
such models were generated, and the structures were averaged,
resulting in an apo receptor model. To generate a reference structure
for ligand docking simulations, 5-HT was manually placed in the
interhelical space of the 5-HT2B receptor model according to three
point mutant 5-HT2B receptors in which one nonconserved,
putative ligand binding residue was mutated to its 5-HT2C
and/or 5-HT2A receptor analog. As shown in Fig. 2A and
Table 1, only the V2.53L and the M5.39V mutations caused a
decrease in SNF affinity: the V2.53L (5-HT2B/2C-to-5-HT2A
)
conserved features of 5-HT binding the receptor: 1) the protonated mutation decreased SNF affinity 17-fold, whereas the
amine was 2 to 3 Å away from Asp3.32, 2) the 5Ј-OH of the ligand M5.39V (5-HT2B-to-5-HT2C/2A) mutation led to a 10-fold de-
was 2 to 3 Å away from Ser3.36, and 3) the aromatic moiety was near crease in ligand affinity. The two mutations in tandem de-
the “aromatic box” delineated by Trp3.28, Trp6.48, Phe6.51, and
creased SNF affinity 37-fold (Table 1). Because Val2.53 was
Phe6.52 (Roth et al., 1997; Manivet et al., 2002). To open up the
predicted by our model to reside in the putative binding
binding pocket, the receptor side chains within 6 Å of the manually
pocket, we chose to investigate further its potential, direct
docked 5-HT were energy minimized as above. The interatomic dis-
role in the 5-HT2B receptor-selective binding of SNF.
tances in the resulting structure were in good agreement with pre-
In the 5-HT2C receptor, residue 2.53 is valine. However,
the V2.53L mutation in the 5-HT2C receptor caused only a
-fold decrease in SNF affinity (i.e., half the decrease caused
vious descriptions of biogenic amine ligand binding to 5-HT -family
receptors (Roth et al., 1997; Kristiansen et al., 2000; Shapiro et al.,
2
9
2000; Manivet et al., 2002). The ligand was then extracted from the
minimized receptor, which was used as the receptor description file by the mutation in the 5-HT2B receptor) (Fig. 2B and Table
for the simulated docking of SNF and RNF by the FLEXX module of 1). The inverse L2.53V mutation in the 5-HT2A receptor had
Sybyl6.91 (Tripos, Inc.). The ligand binding site was defined as those no effect on SNF affinity (Fig. 2C and Table 1). Thus, the role
residues within 6 Å of Asp3.32. Only those solutions that were
consistent with the conserved features described above were re-
tained for consideration.
of residue 2.53 in SNF binding to 5-HT receptors is subtype-
selective; it is most important in the 5-HT2B receptor, of
moderate importance in the 5-HT2C receptor, and not at all
2
To explore the possible effects of a mutation on SNF or NF analog
important in the 5-HT2A receptor. This mirrors the rank
binding, the mutation was computationally introduced into selected
order of potency of SNF at the 5-HT family receptors (i.e., K
2
i
ligand docking simulation solutions, and 10 rounds of computer-
simulated annealing (CSA) were performed to explore possible side
chain orientations relative to the bound ligand. Docked solutions
ϭ 22, 170, and 1900 nM, respectively).
To ensure that the apparent decrease in SNF affinity
bearing the mutation to be studied were heated to and held at 700 K caused by the V2.53L mutation did not result from altered
for 1 ps, then cooled (0.5 K/fs) to 200 K during a 1-ps interval using affinity for the radioligand ([ H]LSD), we measured the K of
3
i

Residues Involved in Norfenfluramine Binding to 5-HT2B Receptors
23
LSD at wild-type and V2.53L 5-HT2B receptors. As shown in explain our mutagenesis data, we computationally intro-
Fig. 3, A and B, and Table 2, the V2.53L mutation had little duced the V2.53L mutation into each solution, and then did
effect on LSD affinity. Indeed, none of the mutations studied 10 rounds of computer-simulated annealing (CSA) to exam-
herein altered LSD affinity by more than 2-fold (Fig. 3, Table ine several possible orientations of the V2.53L side chain
2
, and data not shown), suggesting that none of the muta- relative to the ligand (Fig. 4, E–H). For the solution 1-based
tions dramatically alters receptor folding/topology. We also MD simulations, the average distances Ϯ S.E.M. between the
examined the effect of the V2.53L mutation on four other terminal ␦-methyl group carbons of V2.53L and the ␣-methyl
5
-HT2B receptor agonists: 5-HT, ␣-methyl-5-HT, dihydroer- group carbon of SNF in the 10 lowest energy structures were
gotamine, and pergolide (Fig. 3, A and B, and Table 2). The 3.35 Ϯ 0.05 Å and 4.6 Ϯ 0.1 Å (Fig. 4, E and F). We compared
V2.53L mutation had less or no effect on the binding of the the V2.53L terminal ␦-methyl group-ligand ␣-methyl group
other agonists assayed; the largest effect was a 5-fold de- intercarbon distances with those between the terminal
crease in affinity for dihydroergotamine (Fig. 3B and Table ␥-methyl group carbons of Val2.53 and the ␣-methyl group
2
). The preceding observations demonstrate that the V2.53L carbon of SNF in solution 1 (i.e., 3.73 Å and 3.76 Å). Assum-
mutation markedly and uniquely affects SNF binding to ing that favorable vdW methyl group-methyl group interac-
-HT2B receptors. tions occur at intercarbon distances of 3 to 4 Å, our MD
Modeling, Ligand Docking Simulations, and MD simulations suggested that the net effect of the V2.53L mu-
5
Simulations of Ligand Binding to 5-HT2B Receptors. To tation resulted from the loss of a stabilizing vdW interaction
investigate the possible atomic interactions by which Val2.53 between the 2.53 side chain and the ␣-methyl group of the
in the 5-HT2B receptor contributes to high-affinity SNF bind- ligand. For solution 2-based MD simulations, the V2.53L
ing, we performed ligand docking simulations. Each of the 30 terminal ␦-methyl group-ligand ␣-methyl group intercarbon
solutions generated was inspected to determine whether 1) distances Ϯ S.E.M. were 3.57 Ϯ 0.05 Å and 4.71 Ϯ 0.09 Å
the protonated amine nitrogen of the ligand was close enough (Fig. 4, G and H). Comparing these receptor side chain-ligand
to Asp3.32 to form the conserved, anchoring ionic interaction intercarbon distances with those between the terminal
and 2) whether the aromatic group of the ligand was near one ␥-methyl group carbons of Val2.53 and the ␣-methyl group
or more of the conserved aromatic residues known to stabilize carbon of SNF in solution 2 (3.58 Å and 4.57 Å) revealed very
biogenic amine ligand binding (Choudhary et al., 1993; Roth little difference. Therefore, solution 2 did not predict a
et al., 1997; Manivet et al., 2002). Only 16 of the 30 docked change in the number of stabilizing vdW interactions be-
structures met the first criterion (i.e., the protonated amine tween the 2.53 side chain and the ␣-methyl group of SNF
nitrogen of the ligand was less than 3.0 Å away from one of upon mutation of Val2.53 to leucine. Thus, in relation to the
the carboxylate oxygens of Asp3.32). Of these, only two V2.53L mutation, solution 1 best modeled our experimental
placed the ligand close enough to Val2.53 for the latter to observations.
have an affect on SNF affinity, as was indicated by our
As for Met5.39, in none of the 30 solutions was the residue
mutant receptor binding data (Fig. 4, A–D). The lower energy predicted to be close enough to SNF to directly influence
solution (solution 1, total score ϭ Ϫ10.58 kcal/mol) placed the ligand binding. Furthermore, in none of the MD simulations
␣-methyl group carbon of SNF 3.73 Å and 3.76 Å away from was the Met5.39 side chain predicted to be within less than 4
the terminal ␥-methyl group carbons of Val2.53 (Fig. 4, A and Å of bound SNF. Nevertheless, our data demonstrate a role
B). The other solution (solution 2, total score ϭ Ϫ10.48 kcal/ for Met5.39 in SNF binding that our current model suggests
mol) placed the ␣-methyl group carbon of SNF 3.68 Å and is indirect.
4
.57 Å away from the terminal ␥-methyl group carbons of
To further test the predictive power of our model, we per-
formed identical MD simulations and analyses after intro-
Val2.53 (Fig. 4, C and D).
To determine which of the two solutions was best able to ducing a conservative V2.53I mutation into solution 1 (Fig.
Fig. 1. A 3-D molecular model of the human 5-HT2B recep-
tor showing putative ligand binding residues that are non-
conserved among 5-HT2 family receptors. Side chains
shown by mutagenesis and radioligand binding assays to
be involved in SNF binding are labeled in green; those
shown not to affect SNF binding are labeled in white. Atom
color code: white, carbon; blue, nitrogen; red, oxygen; cyan,
hydrogen; and yellow, sulfur. The ␣-carbon backbone of the
seven helices is shown in purple. See Materials and Meth-
ods for details regarding generation of the model.

2
4
Setola et al.
5
). The average distances Ϯ S.E.M. between the terminal ␥- affinity by approximately 10-fold, whereas the affinity for
and ␦-methyl group carbons of V2.53I and the ␣-methyl DHE, pergolide, and LSD (the radioligand) were affected less
group carbon of SNF in the 10 lowest energy structures were than 3-fold by the mutation (Fig. 3C and Table 2).
3
.91 Ϯ 0.05 Å and 3.8 Ϯ 0.1 Å, respectively. Comparing these
Assuming that the R-enantiomer of norfenfluramine binds
values with those for the terminal ␥-methyl group carbons of in the same orientation as SNF, our model suggested that the
Val2.53 and the ␣-methyl group carbons of SNF in solution 1 ␣-H of RNF would appose Val2.53 and that the ␣-methyl
(
3.73 Å and 3.76 Å) revealed very little difference. Thus, group would project ‘down’ into the interhelical space, away
solution 1 predicted that the conservative V2.53I mutation from Val2.53. Hydrogen atoms, because of their smaller
would not alter vdW interactions with the ␣-methyl group of atomic radius, are less efficient than bulkier methyl groups
SNF. Based on these predictions, we made the V2.53I mutant at forming attractive vdW, and the ␣-H of RNF would be
receptor and measured its affinity for SNF, which we ex- more than 4 Å away from Val2.53; therefore, our model
pected to be similar to that of the wild-type receptor. As predicted that RNF would bind to the wild-type 5-HT2B re-
predicted by our model, the affinity of the V2.53I mutant ceptor with a lower affinity than SNF. As shown in Table 3,
receptor for SNF was not altered by the mutation (Table 1). RNF displayed a 3-fold decrease in affinity compared with
Solution 1, therefore, accurately modeled the effects of both SNF at the 5-HT2B receptor. Furthermore, because the R-(Ϫ)
the V2.53L and the V2.53I mutations. It is interesting that ␣-methyl group of RNF would be more than 4 Å away from
the V2.53I mutation did decrease 5-HT and ␣-methyl-5-HT Val2.53 and therefore unlikely to form stabilizing vdW inter-
Fig. 2. Representative competition binding isotherms for
SNF at wild-type and mutant 5-HT receptors. The per-
2
3
centage total binding of ϳ1 nM [ H]LSD remaining in the
presence of the indicated concentration of SNF is shown for
WT, V2.53L, and M5.39V 5-HT2B receptors (A), WT and
V2.53L 5-HT2C receptors (B), WT and L2.53V 5-HT2A re-
ceptors (C), and WT and V2.53I 5-HT2B receptors (D). Data
are presented as the mean Ϯ S.E. of at least three inde-
pendent experiments measured in duplicate. To obtain log
K values from these data, the data were fit to a heterolo-
i
gous competition model of radioligand binding to one class
of receptor sites that takes ligand depletion into account
(
GraphPad Prism 4.0). Nonspecific radioligand binding
never exceeded 20% of the total radioligand binding.

Residues Involved in Norfenfluramine Binding to 5-HT2B Receptors
25
actions, our model predicted that the V2.53L mutation would SNF and ␣-desmethyl-NF affinity at the V2.53L 5-HT2B re-
have a smaller effect on RNF affinity than it did on SNF ceptor: removal of the S-(ϩ)-methyl group, with which one of
affinity. As shown in Table 3, the affinity for RNF was de- the V2.53L terminal ␦-methyl groups forms a stabilizing vdW
creased only 3-fold by the V2.53L mutation, compared with a interaction, allows V2.53L to project into the binding pocket,
1
7-fold decrease in SNF affinity caused by the mutation. We permitting a stabilizing vdW interaction between a V2.53L
also measured the affinity of the wild-type and V2.53L mu- terminal ␦-methyl group and the ␣-carbon of ␣-desmethyl-
tant 5-HT2B receptors for ␣-desmethyl-NF. As shown in Ta- NF.
ble 3, ␣-desmethyl-NF, like RNF, exhibited a 3-fold reduction
Given our mutagenesis data and modeling predicting
in affinity for the 5-HT2B receptor compared with SNF. The that the ␣-methyl group of SNF is within 4 Å of Val2.53, we
affinity for ␣-desmethyl-NF was reduced 5-fold by the V2.53L hypothesized that an ␣-carbon substituent bulkier than a
mutation (Table 3)—a reduction very similar to that observed methyl group (i.e., an ethyl group) would reduce NF bind-
for RNF, suggesting that the decrease in RNF affinity caused ing to the wild-type 5-HT2B receptor. As shown in Table 3,
by the V2.53L mutation did not result from altered interac- ␣-R,S-(Ϯ)-ethyl-NF exhibited a 46-fold lower affinity for
tions with the ␣-methyl group. The affinity of the NF analogs the wild-type 5-HT2B receptor compared with SNF. Accord-
lacking an S-(ϩ) ␣-methyl group (RNF and ␣-desmethyl-NF) ing to our model, the reduction in affinity caused by the
was very similar to that of SNF at the V2.53L 5-HT2B recep- ␣-ethyl substitution was most probably caused by 1) steric
tor, suggesting a lack of productive interactions between the hindrance between the ␣-ethyl group of the ligand and
V2.53L side chain and the ␣-methyl group of SNF. Our Val2.53 or 2) the adoption of a different binding orienta-
model, however, suggested that the V2.53L side chain forms tion caused by nontolerance of the ␣-ethyl group by
one vdW interaction with the ␣-methyl group of SNF. As Val2.53. In either case, our results with ␣-R,S-(Ϯ)-eth-
such, we would have expected the V2.53L 5-HT2B receptor to yl-NF corroborated our model implicating interactions be-
have lower affinity for ␣-desmethyl-NF than for SNF. To tween the ␣-methyl group of SNF and the ␥-methyl groups
investigate potential interactions between the V2.53L and of Val2.53 in the ligand’s high-affinity binding to 5-HT2B
␣-desmethyl-NF, we replaced the ␣-methyl group of SNF receptors. At the V2.53L 5-HT2B receptor, ␣-R,S-(Ϯ)-eth-
with a hydrogen atom in our V2.53L 5-HT2B receptor model, yl-NF affinity was disrupted less than 2-fold compared
then performed computer-simulated annealing to explore with SNF. Because leucine is bulkier than valine, we had
several possible orientations of the V2.53L side chain relative initially predicted the V2.53L mutant 5-HT2B receptor to
to ␣-desmethyl-NF. The results of our MD simulations pre- be more sensitive to the bulkier ␣-carbon substituent of
dicted that in the absence of an S-(ϩ) ␣-methyl group, at least ␣-R,S-(Ϯ)-ethyl-NF. To explore possible rationales for the
one of the terminal ␦-methyl groups of V2.53L was close lack of differences between SNF and ␣-R,S-(Ϯ)-ethyl-NF
enough to the ␣-carbon of ␣-desmethyl-NF to form a stabiliz- affinity at the V2.53L 5-HT2B receptor, we replaced the
ing vdW interaction (Fig. 6, A and B). With SNF docked, ␣-methyl group of SNF with an ethyl group in our model
however, the V2.53L side chain was predicted to be oriented and then performed computer-simulated annealing to ex-
such that only one of the terminal ␦-methyl groups apposed plore several possible orientations of the V2.53L side chain
the ligand—an orientation that accommodates the S-(ϩ) relative to the ␣-ethyl group. As depicted in Fig. 6, C and
␣-methyl group (Fig. 4, A and B). Thus, our MD simulations D, our MD simulations predicted the V2.53L side chain
provided a plausible explanation for the lack of difference in and the ␣-ethyl group of the ligand to be oriented such that
one V2.53L terminal ␦-methyl group could form a favorable
vdW interaction with the ␣-carbon-proximal methylene
group of the ␣-ethyl substituent; the terminal methyl
TABLE 1
Affinity constants (K ) for SNF binding to wild-type and mutant 5-HT
receptors
p values were obtained from F tests comparing curve fits of competition binding
isotherms for two receptors. Where applicable, values are presented as mean Ϯ S.E.
i
2
group of the ␣-ethyl substituent projected down into the
binding pocket, away from the V2.53L side chain. Thus,
our MD simulations suggested that for SNF, ␣-desmethyl-
NF, and ␣-R,S-(Ϯ)-ethyl-NF binding to the V2.53L 5-HT2B
receptor, the V2.53L side chain forms one stabilizing vdW
interaction with each, consistent with the observation that
the compounds have similar affinity (Table 3).
If indeed the decrease in SNF affinity in the V2.53L mu-
tant was caused by the loss of one stabilizing vdW interaction
between the 2.53 side chain and the ␣-methyl group of SNF—
and not steric hindrance as a result of the bulkiness of the
V2.53L side chain—we expected that a V2.53A mutation
would result in the loss of two stabilizing vdW interactions
and, thus, an even larger decrease in SNF affinity than that
caused by the V2.53L mutation. To test this prediction, we
made the V2.53A mutant receptor and measured its affinity
for SNF. As shown in Fig. 7 and Table 1, and as predicted by
our model, the V2.53A mutation caused a dramatic 190-fold
decrease in SNF affinity. The affinities of other agonists were
altered much less by the V2.53A mutation (Table 2), suggest-
ing an SNF-specific effect. These observations strongly sug-
gest that the effect of the V2.53L mutation on SNF affinity
Receptor
pK
i
i
S-NF (K S-NF)
nM
5
-HT2B WT
7.66 Ϯ 0.02 (22)
7.81 Ϯ 0.07 (15)
a
A1.35P
A1.35S
L1.38S
7.77 Ϯ 0.07 (17)
7.76 Ϯ 0.06 (17)
a
I1.39T
V1.42I
V2.53L
V2.53I
V2.53A
L3.29I
M5.39V
V2.53L, M5.39V
E7.36N
S7.45C
5
5
5
5
7.84 Ϯ 0.07 (14)
7.82 Ϯ 0.09 (15)
a
6.43 Ϯ 0.04a (370)
7.45 Ϯ 0.04 (35)
a
5.48 Ϯ 0.08 (3,300)
7.65 Ϯ 0.07 (22)
a
6.65 Ϯ 0.06 (220)
abc
6.09 Ϯ 0.08 (810)
7.80 Ϯ 0.06 (16)
7.65 Ϯ 0.05 (22)
5.73 Ϯ 0.04 (1900)
5.74 Ϯ 0.04 (1800)
6.77 Ϯ 0.05 (170)
-HT2A WT
HT2A L2.53V
-HT2C WT
-HT2C V2.53L
a
5.8 Ϯ 0.1 (2000)
a
b
c
P Ͻ 0.05 compared with the appropriate wild type 5-HT
P Ͻ 0.05 compared with the V2.53L 5-HT2B receptor.
P Ͻ 0.05 compared with the M5.39V 5-HT2B receptor.
2
receptor.

2
6
Setola et al.
was most probably caused by a decrease in vdW stabilization methyl-NF, and ␣-R,S-(Ϯ)-ethyl-NF at the two receptors.
rather than steric hindrance. As shown in Table 3, the affinities of ␣-desmethyl-NF at
Having established the importance of interactions be- the 5-HT2C and 5-HT2A receptors were not different from
tween Val2.53 and the ␣-methyl group of SNF in ligand SNF and RNF affinities, suggesting that the ␣-methyl
binding to 5-HT2B receptors, we sought to determine group of SNF does not contribute significantly to ligand
whether the ␣-methyl group of SNF contributes to the binding to those receptors. In contrast, NF analogs lacking
ligand’s binding to 5-HT2C and 5-HT2A receptors. To do so, an S-(ϩ) ␣-methyl group (i.e., RNF and ␣-desmethyl-NF)
we measured the affinity of the SNF analogs RNF, ␣-des- exhibited a 3- to 4-fold reduction in binding affinity at
Fig. 3. Competition binding isotherms for several 5-HT2BR agonist ligands at wild-type and V2.53L 5-HT2B receptors. The percentage total binding
3
of ϳ1 nM [ H]LSD remaining in the presence of LSD (A), 5-HT (B), ␣-methyl-5-HT (C), pergolide (D), and dihydroergotamine (DHE, E) is indicated
for WT and V2.53L 5-HT2B receptors. Data are presented as the mean Ϯ S.E. of at least three independent experiments measured in duplicate. To
obtain log K values from these data, the data were fit to a heterologous competition model of radioligand binding to one class of receptor sites that
i
takes ligand depletion into account (GraphPad Prism 4.0). Nonspecific radioligand binding never exceeded 20% of the total radioligand binding. The
V2.53A and V2.53I 5-HT2B receptors were assayed and analyzed simultaneously (see Table 2).
TABLE 2
Affinity constants (K ) for other 5-HT receptor agonist ligands at wild-type and V2.53 mutant 5-HT2B receptors
i
2B
Average Bmax values, fit directly and adjusted for radioligand specific activity and protein content, were as follows: WT, 3.7 Ϯ 0.7 pmol/mg; V2.53L, 9 Ϯ 2 pmol/mg; V2.53A,
Ϯ 1 pmol/mg; V2.53I, 2.4 Ϯ 0.2 pmol/mg. Where applicable, values are presented as mean Ϯ S.E.
7
i i
pK (K )
Drug
WT
V2.53L
V2.53A
V2.53I
nM
5
-HT
7.3 Ϯ 0.1 (50)
7.8 Ϯ 0.1 (20)
7.63 Ϯ 0.08 (23)
7.68 Ϯ 0.09 (21)
8.59 Ϯ 0.05 (2.6)
7.17 Ϯ 0.08 (68)
7.3 Ϯ 0.1*(50)
6.9 Ϯ 0.2* (120)
7.5 Ϯ 0.1 (30)
6.5 Ϯ 0.2* (300)
6.4 Ϯ 0.2* (400)
7.3 Ϯ 0.2 (50)
6.6 Ϯ 0.1* (200)
8.41 Ϯ 0.05 (3.9)
6.2 Ϯ 0.4* (600)
6.7 Ϯ 0.1* (200)
7.2 Ϯ 0.1* (60)
7.3 Ϯ 0.1* (50)
8.61 Ϯ 0.05 (2.4)
␣
-Methyl-5-HT
DHE
Pergolide
LSD
8.34 Ϯ 0.04* (4.6)
*
P Ͻ 0.05 from F tests comparing curve fits of competition binding isotherms for wild-type and mutant 5-HT2BR.

Residues Involved in Norfenfluramine Binding to 5-HT2B Receptors
27
wild-type 5-HT2B receptors, reflecting productive receptor- SNF; the wild-type 5-HT2B receptor exhibited a dramatic
S-(ϩ) ␣-methyl group interactions. The affinities of the 46-fold decrease in affinity for the compound compared
5
-HT2C and 5-HT2A receptors for ␣-R,S-(Ϯ)-ethyl-NF were with SNF. Thus, a bulkier ␣-carbon substituent had the
reduced by 14- and 5-fold, respectively, compared with most dramatic effect (ϳ50-fold) on NF binding to 5-HT2B
Fig. 4. A, 3-D molecular models showing
the results of ligand docking simulations
that are consistent with conserved features
of biogenic amine ligand binding. Simu-
lated docking of SNF to the model shown in
Fig. 3 was performed using the FLEXX
module of Sybyl 6.91 software. The solu-
tions consistent with conserved features of
biogenic amine ligand binding are shown.
A (solution 1), C (solution 2), predicted ori-
entation of SNF relative to conserved res-
idues important for biogenic amine ligand
binding. B (solution 1), D (solution 2),
space fill representation of predicted dock-
ing of SNF relative to Val2.53, in which the
terminal methyl groups of Val2.53 and the
␣
-methyl group of SNF are indicated by
arrowheads and an arrow, respectively.
Atom color code: white, carbon; blue, nitro-
gen; red, oxygen; cyan, hydrogen; and yel-
low, sulfur. The ␣-carbon backbone of the
seven helices is shown in purple. B, repre-
sentative energy-minimized structures
from ten rounds of computer-simulated an-
nealing of solution 1 (A and B) and solution
2
(C and D) after insertion of the V2.53L
mutation. The results predict, in both
cases, one stabilizing van der Waals’ inter-
action between the ␣-methyl group of SNF
(arrows) and the terminal ␦-methyl groups
of V2.53L (arrowheads). Atom color code:
white, carbon; blue, nitrogen; red, oxygen;
cyan, hydrogen; and yellow, sulfur. The
␣
-carbon backbone of the seven helices is
shown in purple.

2
8
Setola et al.
receptors, a much lesser (ϳ10-fold) effect on NF binding to 5-HT2C receptors increased potency to a similar extent:
-HT2C receptors, and a modest (5-fold) effect on NF bind- approximately 2-fold (Fig. 8, A and B, and Table 4). The
ing to 5-HT2A receptors. L2.53V mutation in the 5-HT2A receptor decreased RNF
5
Effect of Point Mutations on Agonist Potency and potency by 30-fold (Table 4), an effect similar to that ob-
Efficacy. For each of the wild-type and point mutant served for SNF potency. With respect to RNF relative
5
-HT2B receptors, we also measured SNF potency (EC50
and relative efficacy compared with the full agonist 5-HT mutation in the 5-HT2B receptor affected the parameter,
Emax) (Fig. 8 and Table 4). As was true for ligand affinity, decreasing it by a dramatic 65% (Fig. 8, A and B, and Table
)
efficacy, as with SNF relative efficacy, only the V2.53L
(
agonist potency was decreased only by the V2.53L and 4). However, although the M5.39V mutation had no effect
M5.39V mutations (Fig. 8, A–C, and Table 4). In contrast on agonist efficacy by itself, in tandem with the V2.53L
to its effect on SNF affinity at 5-HT2B receptors, the mutation, it eliminated the effect of the latter mutation
V2.53L mutation in the 5-HT2C receptor did not alter ag- effect on SNF efficacy and increased its effect on RNF
onist potency (Table 4). This is in line with our finding that efficacy by ϳ1.2-fold (Fig. 8E and Table 4).
SNF affinity was less sensitive to the V2.53L mutation in
the 5-HT2C receptor than in the 5-HT2B receptor. The
L2.53V mutation in the 5-HT2A receptor caused a dramatic
Discussion
2
0-fold decrease in the potency of SNF (Table 4)—a change
The main finding reported herein is that Val2.53 in the
in the opposite direction than would have been predicted 5-HT2B receptor contributes to the high-affinity binding of
based on a role for residue 2.53 in contributing to the rank SNF through vdW interactions between both Val2.53 termi-
order of potency of SNF at 5-HT2 receptors. Only the nal ␥-methyl groups and the ␣-methyl group of the ligand. Of
V2.53L mutation in the 5-HT2B receptor altered the ago- the 11 5-HT2B-to-5-HT2C/2A mutations tested, only the
nist relative efficacy: SNF efficacy was increased 36% by V2.53L (5-HT2B/2C-to-5-HT2A) and M5.39V (5-HT2B-to-5-
the mutation (Fig. 7, A and B, and Table 4). In terms of HT2C/2A) mutations resulted in substantial changes in SNF
RNF potency, the V2.53L mutation in both the 5-HT2B and affinity (Fig. 2 and Table 1). Residue 2.53 in the 5-HT2C
Fig. 5. Representative energy-minimized structure from 10 rounds of computer-simulated annealing of solution 1 after insertion of the V2.53I
mutation. The result predicts two stabilizing vdW interactions between the ␣-methyl group of SNF (arrow) and the terminal methyl groups of V2.53I
(
arrowheads). A, stick representation; B, space fill representation. Atom color code: white, carbon; blue, nitrogen; red, oxygen; cyan, hydrogen; and
yellow, sulfur. The ␣-carbon backbone of the seven helices is shown in purple.
TABLE 3
Affinity constants (K ) for SNF and congeners at wild-type and V2.53L 5-HT2B receptors and wild-type 5-HT2C and 5-HT2A receptors
i
p Values were obtained from F tests comparing curve fits of competition binding isotherms for two receptors. Where applicable, values are presented as mean Ϯ S.E.
i i
pK (K )
Drug
WT
V2.53L
WT2C
WT2A
nM
d
SNF
RNF
7.76 Ϯ 0.05 (17)
7.21 Ϯ 0.04 (62)
6.64 Ϯ 0.07 (230)
7.16 Ϯ 0.07 (69)
7.07 Ϯ 0.08 (85)
7.06 Ϯ 0.07 (87)
5.74 Ϯ 0.08 (1,800)
5.85 Ϯ 0.07 (1,400)
5.87 Ϯ 0.06 (1,300)
d
6.84 Ϯ 0.07 (140)
a
bd
␣
␣
-Desmethyl-NF
-Ethyl-NF
7.26 Ϯ 0.06 (55)
6.58 Ϯ 0.05 (260)
abc
abcd
abc
abc
6.11 Ϯ 0.08 (780)
6.44 Ϯ 0.08
(360)
6.0 Ϯ 0.1 (1,000)
5.0 Ϯ 0.1 (10.000)
a
b
c
P Ͻ 0.05 compared with S-(ϩ)-NF.
P Ͻ 0.05 compared with R-(Ϫ)-NF.
P Ͻ 0.05 compared with ␣-desmethyl-NF.
P Ͻ 0.05 compared with WT.
d

Residues Involved in Norfenfluramine Binding to 5-HT2B Receptors
29
receptor is also a valine; however, the V2.53L mutation in the the preceding results demonstrate that the effect of the
-HT2C receptor diminished SNF binding only half as much V2.53L mutation is selective for SNF.
as it did in the 5-HT2B receptor (Fig. 2 and Table 1). The To explore how Val2.53 might contribute to SNF affinity at
5
inverse L2.53V mutation in the 5-HT2A receptor did not 5-HT2B receptors, we performed ligand docking simulations.
affect SNF affinity (Fig. 2 and Table 1), suggesting that Of the 30 solutions generated, only two were both consistent
interactions between Val2.53 and SNF are more important in with conserved features of biogenic amine ligand binding and
5
-HT2B receptors than in 5-HT2C receptors and that Leu2.53 predicted SNF to be close enough to Val2.53 for the residue to
plays no role in SNF binding to the 5-HT2A receptor. Given directly affect binding (i.e., within 4 Å) (Fig. 4). In both of
the rank order of affinity of SNF for 5-HT receptors (5-HT these, the ␣-methyl group of the ligand was predicted to be
2
2B
Ͼ 5-HT2C Ͼ 5-HT2A), our data suggest that interactions near both (solution 1) or one (solution 2) of the Val2.53
between Val2.53 in the 5-HT2B receptor and the ␣-methyl terminal ␥-methyl groups. After introduction of the V2.53L
group of the ligand contribute to subtype selectivity.
mutation into each solution and computer-simulated anneal-
To ascertain whether the V2.53L mutation affected the ing to explore possible side chain orientations relative to the
binding of other agonists, we measured the affinity of two ␣-methyl group of SNF, both solutions predicted only one
tryptamines (5-HT and ␣-methyl-5-HT) and three ergolines vdW interaction (Fig. 4). For solution 2, then, Val2.53 and
(
dihydroergotamine, pergolide, and LSD) at 5-HT2B recep- V2.53L were predicted to interact similarly with the ligand,
tors. The data presented in Table 2 demonstrate that 2.53 each contributing one stabilizing vdW interaction. For solu-
mutations in the 5-HT2B receptor did not globally perturb tion 1, Val2.53 was predicted to contribute two stabilizing
agonist binding. It is noteworthy that LSD (i.e., the radioli- vdW interactions and V2.53L was predicted to contribute
gand) affinity was altered less than 2-fold. Taken together, only one. The energetic cost (i.e., the ⌬⌬G calculated from the
Fig. 6. Representative energy-minimized structure from 10 rounds of computer-simulated annealing of solution 1 bearing the V2.53L after addition
or removal of SNF ␣-carbon substituents. A and B, the result predicts a stabilizing van der Waals’ interaction between a terminal ␦-methyl group of
V2.53L (arrowheads) and the ␣-carbon of ␣-desmethyl-NF (arrow). C and D, the result predicts a stabilizing vdW interaction between a terminal
␦
-methyl group of V2.53L (arrowheads) and the ␣-carbon-proximal methylene group of the ␣-ethyl substituent of ␣-ethyl-NF; the terminal ethyl group
of the ␣-ethyl substituent points “down” away from V2.53L. Atom color code: white, carbon; blue, nitrogen; red, oxygen; cyan, hydrogen; and yellow,
sulfur. The ␣-carbon backbone of the seven helices is shown in purple.

3
0
Setola et al.
data in Table 1) of the V2.53L mutation on SNF binding was vdW radius of the apposing ␣-H would render RNF much less
ϩ 1.66 kcal/mol. Others have measured the contribution of sensitive to the V2.53L mutation. As shown in Table 3, RNF
methyl group-methyl group interactions to ligand binding affinity was indeed decreased only 3-fold by the V2.53L mu-
energy, with values ranging from ϩ0.4 kcal/mol to ϩ1.4 kcal/ tation, compared with a 17-fold effect on SNF affinity. Like-
mol (Bigler et al., 1993; Huang et al., 1995; Morton et al., wise, ␣-desmethyl-NF also exhibited only a 3-fold decreased
1
1
995; Faergeman et al., 1996; Lu et al., 1997; Oue et al., affinity at V2.53L 5-HT2B receptors compared with its affin-
997; Kawaguchi and Kuramitsu, 1998). The value calcu- ity at the wild-type receptor (Table 3). Thus, the 5-HT2B
lated from our data (ϩ1.66 kcal/mol) is thus consistent with receptor-selective effect of the V2.53L mutation on NF bind-
the loss of a 2.53 methyl group-SNF ␣-methyl group vdW ing requires the ligand to bear an ␣-methyl group in the S-(ϩ)
interaction caused by the V2.53L mutation, a prediction of configuration; the R-(Ϫ) ␣-methyl group does not seem to
our modeling and MD simulations. In addition, the ⌬⌬G for contribute to NF binding to the wild-type 5-HT2B recep-
the V2.53A mutation reveals a ϩ2.94-kcal/mol difference in tor—an observation that is consistent with the similar affin-
SNF binding energy compared with the wild type. Thus, ities of ␣-desmethyl-NF, which lacks an ␣-methyl group, and
removal of the second stabilizing methyl group-methyl group RNF. We also measured the affinity of ␣-R,S-(Ϯ)-ethyl-NF
vdW interaction costs ϩ2.94 kcal/mol Ϫ (ϩ1.66 kcal/mol) ϭ for wild-type 5-HT2B receptors. We hypothesized that if the
ϩ1.28 kcal/mol, again consistent with reported interaction ␣-methyl group of SNF is within 4 Å of either terminal
energies and further supporting our model predicting two ␥-methyl group of Val2.53, an ␣-ethyl substituent would be
stabilizing vdW interactions between Val2.53 and the poorly tolerated as a result of steric hindrance. As shown in
␣-methyl group of SNF.
Table 3, the 5-HT2B receptor exhibited a 46-fold decreased
To further test our modeling and MD simulations, we stud- affinity for ␣-R,S-(Ϯ)-ethyl-NF compared with SNF, further
ied the effect of a V2.53I mutation on 2.53-SNF interactions. supporting our model, wherein Val2.53 of the 5-HT2B recep-
Computer-simulated annealing predicted that V2.53I would, tor is within 4 Å of the ␣-methyl group of SNF—close enough
like Val2.53, form two stabilizing vdW interactions with the for both terminal ␥-methyl groups to form stabilizing vdW
␣-methyl group of SNF (Fig. 5). Thus, we predicted that the interactions with the ␣-methyl group of ligand. In addition,
mutation would not affect SNF affinity. As shown in Table 1, the 5-HT2C and 5-HT2A receptors displayed considerably
the V2.53I 5-HT2B receptor’s affinity for SNF was not signif- smaller reductions in ␣-R,S-(Ϯ)-ethyl-NF affinity—14- and
icantly different from that of the wild type, in agreement 5-fold, respectively (Table 3). One possible explanation for
with our modeling and MD simulations. This result validates the reduced sensitivity of the 5-HT2C and 5-HT2A receptors to
our computational approach and substantiates the prediction ␣-carbon mutations in the ligands is that the ligands may
that the ␣-methyl group of SNF interacts with both terminal bind in a different orientation to the 5-HT2C and 5-HT2A
methyl groups of Val2.53.
receptors than does SNF to the 5-HT2B receptor.
Given the apparent importance of the ␣-methyl group-
Our studies of NF congener binding to the V2.53L 5-HT2B
Val2.53 interaction for SNF binding to 5-HT2B receptors, we are also consistent with interactions between Val2.53 and
measured the affinity of the wild-type and V2.53L mutant the ␣-methyl group of SNF. For instance, our modeling and
receptors for NF analogs bearing ␣-carbon ‘mutations.’ As- MD simulations predicted the formation of one stabilizing
suming that RNF binds in the same orientation as SNF, our vdW interaction between V2.53L and SNF (compared with
model predicts that the ␣-methyl group of RNF will project two such interactions between Val2.53 and the ␣-methyl
down toward the cytoplasmic face of the receptor and the ␣-H group of SNF). At first glance, one would have expected
will appose Val2.53. We reasoned that both the increased ‘removal’ of the ␣-methyl group (i.e., RNF or ␣-desmethyl-
distance of the ␣-methyl group from Val2.53 and the smaller NF) to eliminate the remaining stabilizing vdW interaction,
thereby decreasing affinity. However, we observed that RNF
and ␣-desmethyl-NF bound to the V2.53L 5-HT2B receptor
with affinities similar to that of SNF. In accord with the
previous observations, our modeling and MD simulations
suggested that, absent an S-(ϩ) ␣-methyl group, both V2.53L
terminal ␦-methyl groups can project into the binding pocket
and form at least one stabilizing vdW interaction with the
␣-carbon of ␣-desmethyl-NF. Thus, our MD simulations pro-
vided a plausible molecular rationale for the observation that
SNF, RNF, and ␣-desmethyl-NF displayed similar affinities
at the V2.53L 5-HT2B receptor. Our modeling and MD sim-
ulations also provided an explanation for the similar affini-
ties of ␣-R,S-(Ϯ)-ethyl-NF and SNF for the V2.53L 5-HT2B
receptor. With ␣-R,S-(Ϯ)-ethyl-NF bound in the pocket, our
Fig. 7. Competition binding isotherms for SNF at V2.53A 5-HT recep- MD simulations predicted that the V2.53L side chain orients
2
B
3
tors. The percentage total binding of ϳ1 nM [ H]LSD remaining in the
presence of the indicated concentration of SNF is shown. Data are pre-
sented as the mean Ϯ S.E. of at least three independent experiments
such that one of the V2.53L terminal ␦-methyl groups can
form a stabilizing vdW interaction; the other terminal
measured in duplicate. To obtain log K values from these data, the data ␦-methyl group projects away from the ligand, leaving room
i
were fit to a heterologous competition model of radioligand binding to one for the bulkier ␣-carbon substituent.
class of receptor sites that takes ligand depletion into account (GraphPad
A much simpler explanation for the effect of the V2.53L
mutation on SNF binding to the 5-HT2B receptor is that the
mutation results in steric hindrance with some group(s) on
Prism 4.0). Nonspecific radioligand binding was never exceeded 20% of
the total radioligand binding. SNF congeners (RNF, ␣-desmethyl-NF,
␣
-R,S-(Ϯ)-NF) were assayed and analyzed in parallel (see Table 3).

Residues Involved in Norfenfluramine Binding to 5-HT2B Receptors
31
the ligand. However, the observations that 1) the bulky through hydrophobic and/or hydrogen bonding interactions,
V2.53I side chain did not affect SNF binding and 2) generat- extracellular loop 2 in the binding pocket in such a way that
ing a larger region of bulk tolerance via a V2.53A mutation the loop favorably interacts with bound SNF. Indeed, the
resulted in a much larger perturbation of SNF binding than analogous dopamine D2 receptor residue (Val5.39) along
did the V2.53L mutation (i.e., 46-fold versus. 17-fold) argue with extracellular loop 2 probably plays a role in the binding
against the simpler explanation. However, the fact that the of some ligands (Shi and Javitch, 2004). On the other hand,
wild-type 5-HT2B receptor displays smaller decreases in af- given its position at the extracellular surface of the receptor,
finity for RNF and ␣-desmethyl-NF than that caused by Met5.39 might participate in initial receptor-ligand interac-
V2.53L mutation on SNF affinity suggests the existence of tions that guide SNF down into the binding pocket.
additional factors.
We also determined SNF potency and efficacy at wild-type
A secondary finding of this study is that Met5.39 also and mutant 5-HT receptors (Fig. 8 and Table 4). SNF po-
2
participates in SNF binding. We attempted to explain the tency, like its binding affinity, was only decreased at V2.53L
role of Met5.39 in SNF binding, as we had for Val2.53, using and M5.39V mutant 5-HT2B receptors (Fig. 8 and Table 4). As
modeling, ligand docking simulations, and MD simulations. for agonist relative efficacy, a unique role for Val2.53 in
In none of the computer-simulated annealing results, how- 5-HT2B receptor activation became apparent (Fig. 8 and Ta-
ever, was Met5.39 predicted to be less than 10 Å away from ble 5). That is, both SNF and RNF Emax were altered only in
bound SNF. A possible explanation for our Met5.39 data that the 5-HT2B receptor and only by the V2.53L—not the
is consistent with our model is that the residue stabilizes, V2.53I—mutation.
Fig. 8. Concentration-response isotherms for agonist-stimulated inositol phosphate accumulation. Agonist- and concentration-dependent activation of
WT (A), V2.53L (B), M5.39V (C), V2.53I (D), and V2.53L,M5.39V (E) 5-HT2B receptors transiently expressed in HEK293 cells is shown. Data represent
3
the mean Ϯ S.E. of at least three independent experiments measured in duplicate. Average baseline and maximal drug-stimulated [ H]IP
accumulation (in disintegrations per minute) in wild-type 5-HT2B receptor-expressing cells were: 2700 Ϯ 400 (baseline), 13,000 Ϯ 3000 (5-HT),
1
0,000 Ϯ 2000 (SNF), and 11,000 Ϯ 2000 (RNF); corresponding values for V2.53L 5-HT2B receptor-expressing cells were: 900 Ϯ 100 (baseline),
3,000 Ϯ 5000 (5-HT), 14,000 Ϯ 5000 (SNF), and 4000 Ϯ 2000. To obtain log EC50 and Emax values from concentration-response data, the data were
1
fit to a three parameter logistic concentration-response model (GraphPad Prism 4.0). Emax is expressed as percent of the maximum response to 5-HT.

3
2
Setola et al.
TABLE 4
Potency (EC50) and relative efficacy (Emax) values for SNF and RNF at wild-type and point mutant 5-HT2 receptors
E
max is reported as % maximum response to 5-HT. p values were obtained from F-tests comparing curve fits of concentration-response isotherms for two receptors. Where
applicable, values are presented as mean Ϯ S.E.
pEC50 (EC50
)
E
max
Receptor
SNF
RNF
SNF
RNF
nM
% 5-HT max
5
-HT2B WT
7.66 Ϯ 0.07 (22)
7.7 Ϯ 0.1 (20)
6.62 Ϯ 0.07 (240)
75 Ϯ 3
N.D.
81 Ϯ 4
N.D.
N.D.
N.D.
N.D.
A1.35P
A1.35S
L1.38S
N.D.
N.D.
N.D.
N.D.
N.D.
7.69 Ϯ 0.06 (20)
7.51 Ϯ 0.08 (30)
7.50 Ϯ 0.09 (31)
7.44 Ϯ 0.08 (36)
N.D.
N.D.
I1.39T
V1.42I
V2.53L
V2.53I
L3.29I
M5.39V
V2.53L, M5.39V
E7.36N
S7.45C
N.D.
N.D.
a
a
a
a
7.09 Ϯ 0.07 (82)
6.99 Ϯ 0.07 (100)
101.7 Ϯ 0.6
28.2 Ϯ 0.5
73 Ϯ 4
N.D.
a
a
7.7 Ϯ 0.1 (20)
7.4 Ϯ 0.1 (40)
87 Ϯ 4
a
7.92 Ϯ 0.03a (12)
N.D.
N.D.
7.19 Ϯ 0.09 (64)
6.7 Ϯ 0.1 (200)
70 Ϯ 3
74 Ϯ 5
abc
b
b
abc
6.49 Ϯ 0.05 (320)
6.80 Ϯ 0.08 (160)
73 Ϯ 4
18 Ϯ 1
N.D.
7.6 Ϯ 0.1 (20)
7.8 Ϯ 0.2 (10)
6.7 Ϯ 0.2 (200)
N.D.
N.D.
N.D.
N.D.
N.D.
5
5
5
5
-HT2A WT
6.6 Ϯ 0.2 (300)
75 Ϯ 3
70 Ϯ 3
93 Ϯ 6
85 Ϯ 5
69 Ϯ 5
68 Ϯ 5
89 Ϯ 6
82 Ϯ 4
a
a
HT2A L2.53V
-HT2C WT
-HT2C V2.53L
5.3 Ϯ 0.1 (5000)
5.07 Ϯ 0.09 (8500)
7.2 Ϯ 0.2 (70)
7.2 Ϯ 0.1 (70)
6.7 Ϯ 0.1 (200)
a
6.97 Ϯ 0.8 (100)
a
b
c
2
P Ͻ 0.05 compared with the appropriate WT 5-HT receptor.
P Ͻ 0.05 compared with the V2.53L 5-HT2B receptor.
P Ͻ 0.05 compared with the M5.39V 5-HT2B receptor.
N.D., not determined.
signaling of selected G␣q-coupled protein receptors. J Biol Chem 279:34614–
4623.
Bigler TL, Lu W, Park SJ, Tashiro M, Wieczorek M, Wynn R, and Laskowski M Jr
(1993) Binding of amino acid side chains to preformed cavities: interaction of
serine proteinases with turkey ovomucoid third domains with coded and noncoded
P1 residues. Protein Sci 2:786–799.
Branch CL, Johnson CN, and Stemp G (2002) inventors; SmithKline Beecham,
assignee. Tetrahydroisoquinoline derivatives as modulators of dopamine D3 recep-
tors. U.S. patent 6,465,485. 2002 Oct 15.
Brenot F, Herve P, Petitpretz P, Parent F, Duroux P, and Simonneau G (1993)
Primary pulmonary hypertension and fenfluramine use. Br Heart J 70:537–541.
In conclusion, we have discovered that residue 2.53—a
valine—plays a role in SNF binding to, and activation of,
the 5-HT2B receptor that is unique among 5-HT2 family
receptors. Our studies suggest the existence of two stabi-
lizing vdW interactions between the terminal methyl
groups of Val2.53 and the ␣-methyl group of SNF in the
ligand-bound receptor. These interactions also have conse-
quences on receptor function, because SNF potency and
efficacy are both affected when these interactions are per-
turbed. Thus, vdW interactions between Val2.53 and the
3
Choudhary MS, Craigo S, and Roth BL (1993)
A single point mutation
125
(Phe3403Leu340) of a conserved phenylalanine abolishes 4-[ I]iodo-(2,5-
3
3
dimethoxy)phenylisopropylamine and [ H]mesulergine but not [ H]ketanserin
binding to 5-hydroxytryptamine2 receptors. Mol Pharmacol 43:755–761.
Connolly HM, Crary JL, McGoon MD, Hensrud DD, Edwards BS, Edwards WD, and
Schaff HV (1997) Valvular heart disease associated with fenfluramine-
phentermine. N Engl J Med 337:581–588.
␣
-methyl group of SNF contribute to the subtype selective
pharmacology. Targeting these interactions via ligand mu-
tations (e.g., by synthesizing SNF analogs lacking, or bear-
ing bulkier, ␣-carbon substituents in the S-(ϩ) orientation)
may result in a more selective 5-HT2C agonist. Indeed, as
we show herein, NF congeners either lacking an ␣-methyl
group or bearing a bulkier ␣-ethyl substituent exhibit re-
ductions in 5-HT2B—but not 5-HT2C or 5-HT2A—receptor
affinity. Similar ligand-receptor interactions that are
Dengel F, Ehrmann O, Friedrich L, and Zimmermann F (1975) inventors; Knoll A.G.
Chemische Fabriken, assignee. Theophylline derivatives. U.S. patent 3,926,995.
1
975 Dec 16.
Faergeman NJ, Sigurskjold BW, Kragelund BB, Andersen KV, and Knudsen J (1996)
Thermodynamics of ligand binding to acyl-coenzyme A binding protein studied by
titration calorimetry. Biochemistry 35:14118–14126.
Fitzgerald LW, Burn TC, Brown BS, Patterson JP, Corjay MH, Valentine PA, Sun
JH, Link JR, Abbaszade I, Hollis JM, et al. (2000) Possible role of valvular
serotonin 5-HT(2B) receptors in the cardiopathy associated with fenfluramine. Mol
Pharmacol 57:75–81.
Fuller RW, Snoddy HD, Clemens JA, and Molloy BB (1982) Effect of norfenfluramine
and two structural analogues on brain 5-hydroxyindoles and serum prolactin in
rats. J. Pharm. Pharmacol. 34:449–450.
Glennon RA, Titeler M, and McKenney JD (1984) Evidence for 5-HT2 involvement in
the mechanism of action of hallucinogenic agents. Life Sci 35:2505–2511.
Heisler LK, Cowley MA, Tecott LH, Fan W, Low MJ, Smart JL, Rubinstein M, Tatro
JB, Marcus JN, Holstege H, et al. (2002) Activation of central melanocortin
pathways by fenfluramine. Science (Wash DC) 297:609–611.
unique to the 5-HT2B receptor among 5-HT receptors are
2
likely to exist for SNF-induced activation as well, and a
better understanding of these could guide drug design
efforts toward an effective appetite suppressant devoid of
VHD- and PH-inducing potential.
Huang K, Lu W, Anderson S, Laskowski M Jr, and James MN (1995) Water
molecules participate in proteinase-inhibitor interactions: crystal structures of
Leu18, Ala18 and Gly18 variants of turkey ovomucoid inhibitor third domain
complexed with Streptomyces griseus proteinase B. Protein Sci 4:1985–1997.
Kawaguchi S and Kuramitsu S (1998) Thermodynamics and molecular simulation
analysis of hydrophobic substrate recognition by aminotransferases. J Biol Chem
Acknowledgments
We are grateful to Richard B. Rothman at the National Institute
on Drug Abuse and to the National Institute of Mental Health
Chemical Synthesis Program for providing optically-pure SNF and
RNF, and to David Larson, David Derrington, William Curtis, and
Michael Dolan at Tripos, Inc. for invaluable, expert assistance with
the Sybyl 6.91 software. We also thank Wesley Kroeze for assistance
with statistical analyses and for the critical reading of our manu-
script and Douglas Sheffler for help with image preparation.
273:18353–18364.
Kristiansen K, Kroeze WK, Willins DL, Gelber EI, Savage JE, Glennon RA, and Roth
BL (2000) A highly conserved aspartic acid (Asp-155) anchors the terminal amine
moiety of tryptamines and is involved in membrane targeting of the 5-HT(2A)
serotonin receptor but does not participate in activation via a “salt-bridge disrup-
tion” mechanism. J Pharmacol Exp Ther 293:735–746.
Kroeze WK, Kristiansen K, and Roth BL (2002) Molecular biology of serotonin
receptors structure and function at the molecular level. Curr Top Med Chem
2
:507–528.
References
Kroeze WK, Sheffler DJ, and Roth BL (2003) G-protein-coupled receptors at a glance.
J Cell Sci 116:4867–4869.
Bhatnagar A, Sheffler DJ, Kroeze WK, Compton-Toth B, and Roth BL (2004) Caveo-
lin-1 interacts with 5-HT2A serotonin receptors and profoundly modulates the
Launay JM, Herve P, Peoc’h K, Tournois C, Callebert J, Nebigil CG, Etienne N,

Residues Involved in Norfenfluramine Binding to 5-HT2B Receptors
33
Drouet L, Humbert M, Simonneau G, et al. (2002) Function of the serotonin
Roth BL, Willins DL, Kristiansen K, and Kroeze WK (1998) 5-Hydroxytryptamine2-
family receptors (5-hydroxytryptamine2A, 5-hydroxytryptamine2B,
5-hydroxytryptamine2C): where structure meets function. Pharmacol Ther 79:
231–257.
Rothman RB, Baumann MH, Savage JE, Rauser L, McBride A, Hufeisen SJ, and
Roth BL (2000) Evidence for possible involvement of 5-HT(2B) receptors in the
cardiac valvulopathy associated with fenfluramine and other serotonergic medi-
cations. Circulation 102:2836–2841.
Setola V, Hufeisen SJ, Grande-Allen KJ, Vesely I, Glennon RA, Blough B, Rothman
RB, and Roth BL (2003) 3,4-methylenedioxymethamphetamine (MDMA, “Ec-
stasy”) induces fenfluramine-like proliferative actions on human cardiac valvular
interstitial cells in vitro. Mol Pharmacol 63:1223–1229.
Shapiro DA, Kristiansen K, Kroeze WK, and Roth BL (2000) Differential modes of
agonist binding to 5-hydroxytryptamine2A serotonin receptors revealed by muta-
tion and molecular modeling of conserved residues in transmembrane region 5.
Mol Pharmacol 58:877–886.
5-hydroxytryptamine 2B receptor in pulmonary hypertension. Nat Med 8:1129–
1135.
Lu W, Apostol I, Qasim MA, Warne N, Wynn R, Zhang WL, Anderson S, Chiang YW,
Ogin E, Rothberg I, et al. (1997) Binding of amino acid side-chains to S1 cavities
of serine proteinases. J Mol Biol 266:441–461.
Manivet P, Schneider B, Smith JC, Choi DS, Maroteaux L, Kellermann O, and
Launay JM (2002) The serotonin binding site of human and murine 5-HT2B
receptors: molecular modeling and site-directed mutagenesis. J Biol Chem 277:
17170–17178.
Meltzer HY (1989) Clinical studies on the mechanism of action of clozapine: the dopam-
ine-serotonin hypothesis of schizophrenia. Psychopharmacology 99 (Suppl):
S18–S27.
Menicagli R and Samaritani S (1996) 1,4-Conjugate addition of the Reformatsky
reagent to ␣-nitrostyrenes: a new synthesis of ␥-nitroesters. Tetrahedron 52:1425–
1432.
Morton A, Baase WA, and Matthews BW (1995) Energetic origins of specificity of
Shapiro DA, Kristiansen K, Weiner DM, Kroeze WK, and Roth BL (2002) Evidence
for a model of agonist-induced activation of 5-hydroxytryptamine 2A serotonin
receptors that involves the disruption of a strong ionic interaction between helices
ligand binding in an interior nonpolar cavity of T4 lysozyme. Biochemistry 34:
8564–8575.
Moses PL and Wurtman RJ (1984) The ability of certain anorexic drugs to suppress
3
and 6. J Biol Chem 277:11441–11449.
food consumption depends on the nutrient composition of the test diet. Life Sci
Shi L and Javitch JA (2004) The second extracellular loop of the dopamine D2
receptor lines the binding-site crevice. Proc Natl Acad Sci USA 101:440–445.
Simonneau G, Fartoukh M, Sitbon O, Humbert M, Jagot JL, and Herve P (1998)
Primary pulmonary hypertension associated with the use of fenfluramine deriva-
tives. Chest 114:195S–199S.
Steffee CH, Singh HK, and Chitwood WR (1999) Histologic changes in three ex-
planted native cardiac valves following use of fenfluramines. Cardiovasc Pathol
35:1297–1300.
Neill JC and Cooper SJ (1989) Evidence that D-fenfluramine anorexia is mediated by
5
-HT1 receptors. Psychopharmacology (Berl) 97:213–218.
Nichols DE (2004) Hallucinogens. Pharmacol Ther 101:131–181.
Oue S, Okamoto A, Nakai Y, Nakahira M, Shibatani T, Hayashi H, and Kagamiyama
H (1997) Paracoccus denitrificans aromatic amino acid aminotransferase: a model
enzyme for the study of dual substrate recognition mechanism. J Biochem (Tokyo)
8:245–253.
121:161–171.
Tomita T and Zhao Q (2002) Autopsy findings of heart and lungs in a patient with
primary pulmonary hypertension associated with use of fenfluramine and phen-
termine. Chest 121:649–652.
Vickers SP, Dourish CT, and Kennett GA (2001) Evidence that hypophagia induced
by D-fenfluramine and D-norfenfluramine in the rat is mediated by 5-HT2C recep-
tors. Neuropharmacology 41:200–209.
Weintraub M, Sundaresan PR, Schuster B, Moscucci M, and Stein EC (1992) Long-
term weight control study. III (weeks 104 to 156). An open-label study of dose
adjustment of fenfluramine and phentermine. Clin Pharmacol Ther 51:602–607.
Palczewski K, Kumasaka T, Hori T, Behnke CA, Motoshima H, Fox BA, Le Trong I,
Teller DC, Okada T, Stenkamp RE, et al. (2000) Crystal structure of rhodopsin: A
G protein-coupled receptor. Science (Wash DC) 289:739–745.
Pouwels HM, Smeets JL, Cheriex EC, and Wouters EF (1990) Pulmonary hyperten-
sion and fenfluramine. Eur Respir J 3:606–607.
Roth BL, Baner K, Westkaemper R, Siebert D, Rice KC, Steinberg S, Ernsberger P,
and Rothman RB (2002) Salvinorin A: a potent naturally occurring nonnitrog-
enous kappa opioid selective agonist. Proc Natl Acad Sci USA 99:11934–11939.
Roth BL, Sheffler DJ, and Kroeze WK (2004) Magic shotguns versus magic bullets:
selectively non-selective drugs for mood disorders and schizophrenia. Nat Rev
Drug Discov 3:353–359.
Roth BL, Shoham M, Choudhary MS, and Khan N (1997) Identification of conserved
aromatic residues essential for agonist binding and second messenger production
at 5-hydroxytryptamine2A receptors. Mol Pharmacol 52:259–266.
Address correspondence to: Bryan L. Roth, Case Western Reserve Univer-
sity School of Medicine, Dept. of Biochemistry, Rm. W441, 2109 Adelbert Rd.,
Cleveland, OH 44106-4935. E-mail: bryan.roth@case.edu