THE JOURNAL OF BIOLOGICAL CHEMISTRY
© 1995 by The American Society for Biochemistry and Molecular Biology, Inc.
Vol. 270, No. 47, Issue of November 24, pp. 28268 –28275, 1995
Printed in U.S.A.
Fluorescent Labeling of Purified b2 Adrenergic Receptor
EVIDENCE FOR LIGAND-SPECIFIC CONFORMATIONAL CHANGES*
(Received for publication, August 8, 1995, and in revised form, September 18, 1995)
Ulrik Gether‡, Sansan Lin‡, and Brian K. Kobilka‡§¶
From the ‡Howard Hughes Medical Institute and the §Division of Cardiovascular Medicine, Stanford University Medical
School, Stanford, California 94305
The purpose of the present study was to develop an
approach to directly monitor structural changes in a G
protein-coupled receptor in response to drug binding.
Purified human b2 adrenergic receptor was covalently
labeled with the cysteine-reactive, fluorescent probe
N,N*-dimethyl-N-(iodoacetyl)-N*-(7-nitrobenz-2-oxa-1,3diazol-4-yl)ethylenediamine (IANBD). IANBD is characterized by a fluorescence which is highly sensitive to the
polarity of its environment. We found that the full agonist, isoproterenol, elicited a stereoselective and dosedependent decrease in fluorescence from IANBD-labeled b2 receptor. The change in fluorescence could be
plotted against the concentration of isoproterenol as a
simple hyperbolic binding isotherm demonstrating interaction with a single binding site in the receptor. The
ability of several adrenergic antagonists to reverse the
response confirmed that this binding site is identical to
the well described binding site in the b2 receptor. Comparison of the response to isoproterenol with a series of
adrenergic agonists, having different biological efficacies, revealed a linear correlation between biological
efficacy and the change in fluorescence. This suggests
that the agonist-mediated decrease in fluorescence from
IANBD-labeled b2 receptor is due to the same conformational change as involved in receptor activation and G
protein coupling. In contrast to agonists, negative antagonists induced a small but significant increase in
base-line fluorescence. Despite the small amplitude of
this response, it supports the notion that antagonists by
themselves may alter receptor structure. In conclusion,
our data provide the first direct evidence for ligandspecific conformational changes occurring in a G protein-coupled receptor. Furthermore, the data demonstrate the potential of fluorescence spectroscopy as a
tool for further delineating the molecular mechanisms
of drug action at G protein-coupled receptors.
The b2 adrenergic receptor is a prototype member of the G
protein-coupled receptor family (1). The receptor family constitutes the largest group of plasma membrane receptors, which
are characterized by a remarkable diversity in the chemical
structure of their endogenous ligands (1–3). The receptors are
all believed to share a common topology with seven a-helical,
transmembrane segments; however, the helical arrangement
and actual three-dimensional structure of the receptors remain
* The work was supported by National Institutes of Health Grant RO
1 NS28471. The costs of publication of this article were defrayed in part
by the payment of page charges. This article must therefore be hereby
marked “advertisement” in accordance with 18 U.S.C. Section 1734
solely to indicate this fact.
¶ To whom correspondence should be addressed: Howard Hughes
Medical Institute, B157 Beckman Center, Stanford University Medical
School, Stanford, CA 94305. Tel.: 415-723-7069; Fax: 415-498-5092.
unknown (1–3). Mutagenesis studies in the b2 adrenergic receptor as well as in many other G protein-coupled receptors
have been able to assign distinct receptor functions, such as
ligand binding and G protein coupling, to specific receptor
domains (1–3). Several molecular models of these receptors
have also been generated based on the structure of bacteriorhodopsin and rhodopsin for which more detailed structural
information is available (3–5). Nevertheless, very little is
known about the molecular events and structural changes in
the receptor that provide the important link between ligand
binding and transmission of the signal across the membrane.
Drugs acting at G protein-coupled receptors are traditionally
classified in biological assays as either full agonists, which
elicit the maximal response, as partial agonists, which only
elicit a fractional response, or as antagonists, which block the
response induced by agonists (6, 7). In addition, recent data
have suggested that antagonists should be subclassified into at
least two categories: neutral antagonists, which have no effect
on basal receptor activity, and negative antagonists (also referred to as inverse agonists), which inhibit basal receptor
activity occurring in the absence of agonist (7–13). Thus, in
contrast to the conventional view, it has now become evident
even in vivo (13) that many antagonists actively inhibit receptor function rather than just passively block access of the
agonist to its binding site (8 –13). However, the structural basis
for the biological classification of drug action at G proteincoupled receptors is not yet known. Any direct evidence for the
existence of discrete ligand-specific conformational states of G
protein-coupled receptors has not been obtained. To date, the
conformational state of G protein-coupled receptors has been
assessed only by indirect methods, such as the effect of receptor
conformation on G protein GTPase activity or on the activity of
the effector enzymes.
In the present study we have developed a fluorescence spectroscopy approach to directly monitor ligand-induced conformational changes in a G protein-coupled receptor. Our approach takes advantage of the sensitivity of many fluorescent
molecules to the polarity of their molecular environment (14 –
17). Fluorescent labels incorporated into proteins can therefore
often be used as sensitive indicators of conformational changes
and of protein-protein interactions that cause changes in polarity of the environment surrounding the probe (14 –17). To
accomplish these experiments, we expressed the human b2
adrenergic receptor in SF-9 insect cells and established a purification procedure, which allowed us to obtain the required
amount of pure protein for the spectroscopy analysis. The purified receptor was labeled with the cysteine-specific and environmentally sensitive fluorescent probe N,N9-dimethyl-N-(iodoacetyl)-N9-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)ethylenediamine (IANBD).1 We found that binding of full agonists and
1
28268
The abbreviations used are: IANBD, N,N9-dimethyl-N-(iodoacetyl)-
Fluorescent Labeling of the b2 Adrenergic Receptor
strong partial agonists to the fluorescently labeled b2 adrenergic receptor produced a stereospecific, dose-dependent, and
reversible decrease in fluorescence emission, which is not observed in response to weaker partial agonists or in response to
neutral or negative antagonists. The magnitude of the response
correlated with agonist efficacy, suggesting that this change in
fluorescence emission is due to the same conformational
change as involved in agonist activation of the receptor.
EXPERIMENTAL PROCEDURES
Expression and Purification of the b2 Adrenergic Receptor—DNA
sequences encoding the human b2 adrenergic receptor, epitope-tagged
at the amino terminus with the cleavable influenza-hemagglutinin
signal-sequence followed by the “FLAG”-epitope (IBI, New Haven, CT),
and tagged at the carboxyl terminus with six histidines, were cloned
into the baculovirus expression vector pVL1392 (Invitrogen, San Diego,
CA) and expressed in SF-9 insect cells according to previously described
methods (18). For purification, the cells were grown in 1000-ml cultures
in SF 900 II medium (Life Technologies, Inc.) containing 5% fetal calf
serum (Gemini, Calabasas, CA) and 0.1 mg/ml gentamicin (Life Technologies, Inc.). Cells were routinely infected with a 1:30 – 40 dilution of
a high titer virus stock at a density of 6 – 8 3 106 cells/ml and harvested
after 48 h. Cell pellets were kept at 270 °C until used for purification.
The receptor was purified to homogeneity using a three-step purification procedure as described elsewhere (19). Briefly, lysed cell membranes were solubilized in 0.8% n-dodecyl b-D-maltoside (DbM) (CalBiochem, La Jolla, CA) followed by Ni-column chromatography using
chelating Sepharose (Pharmacia, Uppsala, Sweden). The eluate from
the Ni-column was further purified on an M1 anti-FLAG antibody
column (IBI) and finally by alprenolol-affinity chromatography. This
procedure ensured that only noncleaved and functional protein was
purified. Approximately 5 nmol of pure protein generally could be
obtained from a 1000 –1500-ml culture.
Fluorescent Labeling of Purified b2 Receptor—Purified b2 receptor
(1–1.5 nmol) was labeled with 10 –15-fold molar excess of IANBD (Molecular Probes, Eugene, OR) (150 mM) in a total volume of 100 ml of
buffer (20 mM Tris-buffer, pH 7.4, containing 100 mM NaCl and 0.05%
DbM). The reaction was allowed to proceed for 1 h at room temperature
in the dark and was quenched by addition of 1 mM cysteine. Cysteinereacted dye was removed by desalting on a Sephadex G50 gel filtration
column (0.5 cm 3 9 cm) followed by concentrating the resulting sample
to 100 ml using a Centricon-30 concentrator (Amicon, Beverly, MA). The
labeled protein was diluted approximately 100-fold in buffer for the
fluorescence measurements. Fluorescence in control samples without
receptor or with receptor incubated with cysteine-reacted IANBD was
negligible (Fig. 1b). The stoichiometry of the labeling was determined
by measuring absorption at 481 nm and using an extinction coefficient
of 21,000 M21 cm21 for IANBD and a molecular mass of 50,000 Da for
the b2 receptor. Protein concentration was determined using the BioRad DC protein assay kit (Bio-Rad). The labeling procedure resulted in
incorporation of 1.2 6 0.16 mol of IANBD per mol of receptor (mean 6
S.E., n 5 7).
Binding Assay—Binding assays on solubilized and purified b2 receptor were performed using [3H]dihydroalprenolol as radioligand (Amersham Corp.). To routinely assess total amount of receptor, solubilized
and purified b2 receptor was incubated with 10 nM [3H]dihydroalprenolol in a total volume of 100 ml of buffer (20 mM Tris-buffer, pH 7.4,
containing 100 mM NaCl and 0.05% DbM) for 1 h. In competition
binding assays, unlabeled or IANBD-labeled purified receptor was incubated with 1.2 nM [3H]dihydroalprenolol in a total volume of 100 ml of
buffer and indicated concentrations of alprenolol or isoproterenol for 1
h. The binding assay was stopped and free [3H]dihydroalprenolol separated from bound by desalting on Sephadex G50 columns (4 cm 3 0.5
cm) using ice-cold buffer. Nonspecific binding was determined in presence of 10 mM alprenolol. IC50 values were determined by nonlinear
regression analysis using Inplot 4.0 from GraphPad Software, San
Diego, CA.
SDS-Polyacrylamide Gel Electrophoresis—Purified b2 receptor labeled with IANBD was analyzed by 10% SDS-polyacrylamide gel electrophoresis according to Laemmli (20). Incorporated fluorophore was
visualized by photographing the gel under UV light. The protein was
visualized by standard Coomassie Blue staining.
Fluorescence Spectroscopy—Fluorescence spectroscopy was perN9-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)ethylenediamine; DbM, n-dodecyl b-D-maltoside.
28269
formed at room temperature on a SPEX Fluoromax spectrofluorometer
with photon counting mode using an excitation and emission bandpass
of 4.2 nm. In emission scan experiments, 30 – 60 pmol of IANBD-labeled
b2 receptor in 400 ml of buffer (20 mM Tris-buffer, pH 7.4, containing 100
mM NaCl and 0.05% DbM) were used. Excitation was 481 nm, and
emission measured from 490 to 625 nm with an integration time at 0.3
s/nm. Time course experiments were performed with 30 – 60 pmol of
IANBD-labeled b2 receptor in 500 ml of buffer under constant stirring.
Excitation was 481 nm, and emission was 523 nm. The volume of the
ligands (or water) was always 1% of total volume, and fluorescence was
corrected for this dilution in all experiments shown. Fluorescence
quenching experiments were performed by sequential addition to the
cuvette of potassium iodide (KI) from a freshly made 1 M stock containing 10 mM of Na2S2O3. The experiments were done with 50 – 60 pmol of
receptor in an initial volume of 400 ml. After each addition of KI, the
sample was thoroughly mixed, and fluorescence was recorded at 523 nm
(excitation at 481 nm). To correct for dilution/ionic strength effects on
fluorescence changes, 1 M KCl was added to control samples as described above for KI. The corrected data were plotted according to
Stern-Volmer equation (21), F0/F 5 1 1 KSV[KI], where F0/F is the ratio
of fluorescence intensity in the absence and presence of KI, and KSV is
the Stern-Volmer quenching constant. All of the compounds tested had
an absorbance of less than 0.01 at 481 and 523 nm in the concentrations
used excluding any “inner filter” effect in the fluorescence experiments
(data not shown).
Membrane Preparation from SF-9 Insect Cells—Membranes were
prepared from SF-9 cells (30-ml cultures at a density of 3 3 106 cells/ml)
infected with b2 receptor baculovirus for 24 h to obtain low receptor
density (approximately 1.2 pmol/mg of protein) and 48 h to obtain high
receptor density (approximately 7 pmol/mg of protein). Cells were
washed with phosphate-buffered saline and lysed using 25 strokes with
a Dounce homogenizer in 10 mM Tris-HCl (pH 7.4) containing 1 mM
EDTA, 0.2 mM phenylmethylsulfonyl fluoride, and 10 mg/ml benzamidin. The lysates were centrifuged for 5 min at 500 3 g, and the resulting
supernatant was centrifuged at 40,000 3 g for 30 min. The membrane
pellet was resuspended in binding buffer (75 mM Tris-HCl, pH 7.4, with
12.5 mM MgCl2 and 1 mM EDTA). Protein was determined using the
Bio-Rad DC protein assay kit (Bio-Rad). Adenylate cyclase in membranes was measured as described elsewhere (22).
RESULTS
Fluorescent Properties of IANBD and Purified b2 Receptor
Labeled with IANBD—IANBD is a highly fluorescent, cysteineselective reagent (16, 17). The fluorescence of IANBD increased
as the polarity of the solvent decreased and was more than
10-fold stronger in 1-butanol and n-hexane than in aqueous
buffer (Fig. 1a). There was a parallel blue-shift in the emission
maximum from 540 nm in aqueous buffer to 530 nm in 1-butanol and 510 nm in n-hexane (Fig. 1a). Labeling of the b2
adrenergic receptor, purified from SF-9 insect cells, with
IANBD at a stoichiometry of about 1.2 mol of IANBD per mol of
receptor revealed a strong fluorescence signal with an emission
maximum at 523 nm (Fig. 1b). The blue-shift in emission maximum, as compared to cysteine-reacted IANBD in aqueous
buffer, indicated that the modified cysteine(s) are located in an
environment that, on the average, is of lower polarity than
1-butanol but higher than n-hexane (Fig. 1b). This would likely
involve labeling of one or more of the five cysteine residues that
are located in the transmembrane, hydrophobic core of the
receptor. This was also supported by the observation that denaturation of the labeled receptor with guanidinium chloride
dramatically reduced emission around 80% and caused a
2–3-nm red-shift in the emission maximum (data not shown).
The covalent modification of the receptor was confirmed by
SDS-polyacrylamide gel electrophoresis of the labeled receptor,
and the specificity of the labeling was verified by blocking the
incorporation of IANBD with the cysteine-specific, nonfluorescent reagents, iodoacetamide and N-ethylmaleimide (Fig. 1b,
inset). The fluorescent labeling did not perturb the pharmacological properties of the receptor as illustrated in Fig. 2. Thus,
the agonist, isoproterenol, and the antagonist, alprenolol, inhibited [3H]dihydroalprenolol binding to IANBD-labeled b2 re-
28270
Fluorescent Labeling of the b2 Adrenergic Receptor
FIG. 2. Competition binding profiles of unlabeled and IANBDlabeled, purified b2 receptor. a and b, competition binding of [3H]dihydroalprenolol (1.2 nM) with isoproterenol (a) and alprenolol (b) to
unlabeled, purified b2 receptor. c and d, competition binding of [3H]dihydroalprenolol (1.2 nM) with isoproterenol (a) and alprenolol (b) to
IANBD-labeled, purified b2 receptor. Data are expressed as percent of
maximum bound [3H]dihydroalprenolol (mean 6 S.E., n 5 3).
FIG. 1. Fluorescence properties of IANBD and IANBD-labeled
b2 receptor. a, emission spectra of cysteine-reacted IANBD (0.3 mM) in
solvents of different polarity. Excitation was set at 481 nm. b, emission
spectrum of IANBD-labeled b2 receptor (0.15 mM receptor, 1.2 mol
IANBD per mol receptor). Control is emission spectrum of 0.15 mM b2
receptor “labeled” with IANBD prebound to free cysteine instead of free
IANBD to asses possible nonspecific attachment of the probe to the
receptor during labeling. Insert, 10% SDS-polyacrylamide gel electrophoresis of IANBD-labeled b2 receptor. Lane 1, 150 pmol of IANBDlabeled b2 receptor; lanes 2 and 3, 150 pmol of b2 receptor preincubated
before exposure to IANBD with iodoacetamide (lane 2) and N-ethylmaleimide (lane 3). Inset: left panel, Coomassie Blue staining of gel; right
panel, gel photographed under UV light. The weak band with an apparent molecular mass of 32.5 kDa is a degradation product of the
receptor.
ceptor with IC50 values of 1.3 6 0.5 mM (n 5 3, mean 6 S.E.)
and 2.8 6 0.6 nM (n 5 3, mean 6 S.E.), respectively, as compared to 0.8 6 0.2 mM (n 5 3, mean 6 S.E.) and 2.9 6 0.9 nM (n
5 3, mean 6 S.E.) for the unlabeled receptor. These data are in
agreement with previous binding data on purified b2 receptor
(23). Labeling of the b2 receptor with IANBD did not affect the
total number of binding sites as assessed from binding assays
using a saturating concentration of [3H]dihydroalprenolol (10
nM) (data not shown).
Stereospecificity, Dose Dependence, and Reversibility of Isoproterenol-mediated Decrease in Fluorescence from IANBD-labeled b2 Receptor—Binding of the full agonist, isoproterenol, to
IANBD-labeled b2 receptor caused a decrease in fluorescence
intensity without detectable change in the wavelength at which
maximal emission occurred (Fig. 3). The decrease was stereospecific as illustrated in Fig. 3 by comparing the effect of 30
mM of the (2)-isomer of isoproterenol with 30 mM of the less
active (1)-isomer. In addition, no response to isoproterenol was
observed with IANBD-labeled receptor denatured in guanidinium chloride (data not shown). To investigate the kinetics of
this agonist-mediated change, the fluorescence intensity was
measured as a function of time (Figs. 4 and 5). Prior to adding
ligand we observed a slight but constant decline in base-line
fluorescence (Figs. 4 and 5). This loss of fluorescence over time
is likely due to factors such as bleaching and hydrolysis of the
probe during the experiments. It is unlikely that this decline in
FIG. 3. Stereospecificity of isoproterenol induced decrease in
fluorescence from IANBD-labeled b2 receptor. a, emission spectra
of IANBD-labeled b2 receptor obtained immediately (t 5 0) and after 15
min (t 5 15) following addition of 30 mM of the less active (1)-isomer of
the agonist, isoproterenol ((1)ISO). b, Emission spectra of IANBDlabeled b2 receptor obtained immediately (t 5 0) and after 15 min (t 5
15) following addition of 30 mM of the active (2)-isomer of the agonist,
isoproterenol ((2)ISO). The experiments shown is representative of
three identical experiments. Fluorescence measurements were done as
described under “Experimental Procedures” with excitation set at 481 nm.
fluorescence is due to denaturation of the protein, since a
similar loss of fluorescence also was observed with labeled
receptor that was intentionally denatured in guanidinium chloride (see Fig. 7d). It should also be noted that the decrease over
Fluorescent Labeling of the b2 Adrenergic Receptor
FIG. 4. Time course and dose-dependence of isoproterenol induced decrease in fluorescence from IANBD-labeled b2 receptor. a, emission from IANBD-labeled b2 receptor measured over time
following stimulation with indicated concentrations of isoproterenol.
Excitation was 481 nm and emission measured at 523 nm. Isoproterenol (ISO) was added at the time indicated by the arrow. Fluorescence in
the individual traces was normalized to the fluorescence observed immediately after addition of ligand. The experiment shown is representative of four identical experiments. b, the percent change in fluorescence at t 5 15 min following addition of isoproterenol plotted against
the isoproterenol concentration and fitted to a simple hyperbolic function (F 5 F0L/Kd 1 L; F, change in fluorescence at the ligand concentration L; F0, maximum change in fluorescence; Kd, affinity constant). c,
the relative change in fluorescence at t 5 15 min following addition of
isoproterenol plotted against the logarithm of the isoproterenol concentration and fitted to a one-site sigmoid curve. The percent change in
fluorescence was calculated as the change in fluorescence relative to the
extrapolated base line at t 5 15 min after addition of isoproterenol.
Percent change is given as mean 6 S.E. of the following number of
experiments, 1000 mM, n 5 5; 300 mM, n 5 4; 100 mM, n 5 4; 30 mM, n 5
4; 10 mM, n 5 4; 3 mM, n 5 4; 1 mM, n 5 2; 0.1 mM, n 5 2, H2O, n 5 5.
Curve fittings were performed using Inplot 4.0 from GraphPad Software, San Diego, CA.
time was unaffected by addition of 0.1% bovine serum albumin,
10% glycerol, or phospholipids to the cuvette (data not shown).
The time course analysis showed that the response to isoproterenol was dose-dependent and reached a maximum amplitude below the extrapolated base line after 10 –15 min (Fig. 4a).
The relative change in fluorescence at t 5 15 min following
addition of isoproterenol could be plotted against the isoproterenol concentration and fitted to a simple hyperbolic function, describing a single binding site with a Kd of 29 mM (Fig.
4b). Similarly, the fluorescent change could be plotted against
the logarithm of the isoproterenol concentration showing the
best fit to a one-site sigmoid curve with an EC50 of 19 mM (Fig.
4c). The Kd value of 29 mM and the EC50 value of 19 mM for
isoproterenol are higher than the binding constants observed
using conventional radioligand binding techniques as shown in
Fig. 2 (;1 mM). One explanation for this apparent discrepancy
could be technical differences in the methods by which the
binding constants were obtained (the fluorescence studies were
done with more than 100-fold higher receptor concentrations
(100 nM receptor) for 15 min). Although apparent maximal
change in fluorescence was observed already after 15 min, full
equilibrium may not have been reached. Unfortunately, the
fluorescent change upon ligand binding cannot be reliably determined after 1 h of incubation due to magnification of baseline differences over a longer period of time. Another potential
explanation for the difference in apparent KD values could be
28271
FIG. 5. Reversibility of isoproterenol-induced decrease in fluorescence from IANBD-labeled b2 receptor. a, control addition of
water (H2O). b and c, reversal of the response to isoproterenol (ISO) by
the active (2)-isomer of the antagonist propranolol, (2)PROP (b), but
not by the less active (1)-isomer, (1)PROP (c). Dotted lines indicate
extrapolated base line. Excitation was 481 nm and emission measured
at 523 nm. Fluorescence in all the individual traces shown was normalized to the fluorescence observed immediately after addition of ligand.
All traces shown are representative of at least three identical
experiments.
the incomplete labeling of the cysteine in the b2 receptor that is
responsible for the agonist-induced response. The fraction of
labeled receptors may exhibit a lowered agonist affinity in
contrast to the fraction of receptors labeled at other sites.
However, this explanation is inconsistent with our radioligand
binding data, which detect only one agonist affinity site (Fig. 2).
Furthermore, we do not observe a change in the total number
of binding sites following labeling with IANBD.
We also investigated whether the response seen following
stimulation with isoproterenol could be reversed by addition of
antagonist. As shown in Fig. 5, the response to isoproterenol
could be readily reversed by the active (2)-isomer of the antagonist propranolol but not by the less active (1)-isomer. The
response to isoproterenol was similarly reversed by several
other antagonists, including alprenolol, ICI 118,551, pindolol,
and dichloroisoproterenol (data not shown).
Comparison of the Isoproterenol-mediated Response with
Other Adrenergic Agonists—A series of adrenergic agonists
were tested for their ability to affect fluorescence of IANBDlabeled b2 receptor. The efficacy of these compounds was determined by measuring adenylate cyclase activity in membranes from SF-9 cells expressing the b2 adrenergic receptor
(Fig. 6b) and was found to be in agreement with previous
studies (6, 11, 24). The full agonists, epinephrine and isoproterenol, which produced the largest increase in adenylate cyclase activity (Fig. 6b), also caused the largest decrease in
fluorescence (5– 6%) (Fig. 6a); whereas the relatively strong
partial agonist salbutamol decreased fluorescence about 2.6%
in agreement with a lower intrinsic activity in the adenylyl
cyclase assay (Fig. 6b) (24). The weaker partial agonists, dobutamine and ephedrine, did not cause a decrease in fluorescence
relative to the extrapolated base line but rather produced a
slight increase (Fig. 6b). Interestingly, plotting percent change
28272
Fluorescent Labeling of the b2 Adrenergic Receptor
FIG. 6. a, Effect of different adrenergic agonists on fluorescence from
IANBD-labeled b2 receptor. The percent change (mean 6 S.E.) in fluorescence was calculated as the change in fluorescence relative to the
extrapolated base line at t 5 15 min after addition of ligand. The ligands
and concentrations used were (number of experiments in parentheses);
H2O, water (n 5 5); EPH, 1023 M ephedrine (n 5 3); DOB, 1023 M
dobutamine (n 5 4); SAL, 1023 M salbutamol (n 5 3); ISO, 1023 M
isoproterenol (n 5 5); and EPI, 1023 M epinephrine (n 5 3). Responses
significantly different from water are indicated by *p , 0.000005 (unpaired t test). The ligand concentrations used were chosen to ensure full
saturation of the receptor with all ligands. This was confirmed by the
ability of all compounds (at the concentration used) to fully displace
[3H]dihydroalprenolol from the receptor in a binding assay (data not
shown). b, effect of adrenergic agonists on adenylate cyclase activity in
membranes from SF-9 cells. Data are maximal response to indicated
ligands (see above) expressed as percent of basal activity (mean 6 S.E.,
n 5 2) in membranes from cells infected with b2 receptor baculovirus at
a density of 1.2 pmol/mg of protein. The dotted bar shows basal activity
in an equivalent amount of membranes from non-infected cells in percent of basal activity in membranes from the infected cells. c, plot of
percent change in fluorescence at t 5 15 min against percent change in
adenylate cyclase activity. The squared correlation coefficient (r2) was
0.98.
in base-line fluorescence induced by this series of adrenergic
agonists versus percent change in basal adenylate cyclase activity revealed a linear correlation (squared correlation coefficient, r2 5 0.98) (Fig. 6c).
Effect of Antagonists on Fluorescence from IANBD-labeled b2
Receptor—We also examined the effect of a series of adrenergic
antagonists on fluorescence intensity from IANBD-labeled receptor over time. Propranolol and ICI 118,551, which both have
been described as negative antagonists (10, 11, 13), caused a
relative increase in fluorescence intensity from IANBD-labeled
b2 receptor followed by apparent stabilization of the rate at
which fluorescence decreased over time (Figs. 7, a and b, and
FIG. 7. Effect of antagonists on fluorescence from IANBD-labeled b2 receptor. a, emission from IANBD-labeled b2 receptor measured over time following stimulation with 1 mM (2)-propranolol
((2)PROP), 1 mM (1)-propranolol ((1)PROP) and water (H2O). b, emission from IANBD-labeled b2 receptor measured over time following
stimulation with 10 mM ICI 118,551 (ICI), 10 mM pindolol (PIND), and
water (H2O). c, emission from IANBD-labeled b2 receptor measured
over time following exposure to 10 mM dichloroisoproterenol (DCI),
alprenolol (ALP) and water (H2O). d, emission from guanidinium chloride denatured IANBD-labeled b2 receptor measured over time following exposure to 10 mM (2)-propranolol ((2)-PROP) and 10 mM ICI
118,551 (ICI). The compounds were added at the time indicated by the
arrows. Excitation was 481 nm, and emission was measured at 523 nm.
Fluorescence in the individual traces was normalized to the fluorescence observed immediately after addition of ligand. The experiment
shown is representative of at least three identical experiments.
8a). We did not observe any parallel change in the wavelength
at which maximal emission occurred (data not shown). The
response to propranolol was stereoselective since it was not
elicited by the less active (1)-propranolol, indicating that the
change must involve binding of propranolol to the receptor (Fig.
7a). This was also supported by the absence of response to
propranolol and ICI 118,551 when they were exposed to denatured receptor (Fig. 7d). Alprenolol, pindolol, and dichloroproterenol, which in biological assays have been shown to exhibit
either weak partial agonism, neutral antagonism or sometimes
negative antagonism (6, 10, 11, 24), induced a smaller but
nevertheless reproducible increase in fluorescence (Figs. 7, b
and c, and 8a).
The fluorescent changes were compared with the ability of
the antagonists to affect basal adenylate cyclase activity in
SF-9 cell membranes. The effect of the antagonists on basal
adenylate cyclase activity was evaluated in SF-9 cells expressing a high level of b2 receptor (approximately 7 pmol/mg of
protein) which leads to an elevated receptor-mediated basal
level of adenylate cyclase activity (Fig. 8b). In contrast, partial
Fluorescent Labeling of the b2 Adrenergic Receptor
FIG. 8. a, effect of different adrenergic antagonists on fluorescence
from IANBD-labeled b2 receptor. The percent change (mean 6 S.E.) in
fluorescence was calculated as the change in fluorescence relative to the
extrapolated base line at t 5 15 min after addition of ligand. The ligands
and concentrations used were (number of experiments in parentheses);
H2O, water (n 5 5); PIND, 1025 M pindolol (n 5 4); ALP, 1025 M
(2)alprenolol (n 5 3); DCI, 1024 M dichloroisoproterenol (n 5 3);
(2)PROP, 1025 M (2)-propranolol (n 5 3); and ICI, 1025 M ICI 118,551
(n 5 4). Responses significantly different from water are indicated by *p
, 0.0005 (unpaired t test). The ligand concentrations used were chosen
to ensure full saturation of the receptor with all ligands. This was
confirmed by the ability of all compounds (at the concentration used) to
fully displace [3H]dihydroalprenolol from the receptor in a binding
assay (data not shown). b, effect of adrenergic antagonists on adenylate
cyclase activity in membranes from SF-9 cells. Data are maximal response to indicated ligands (see above) expressed as percent of basal
activity (mean 6 S.E., n 5 2) in membranes from cells infected with b2
receptor baculovirus at a density of 7 pmol/mg protein. The dotted bar
shows basal activity in an equivalent amount of membranes from noninfected cells in percent of basal activity in membranes from the infected cells.
agonists were evaluated at a lower receptor density (approximately 1.2 pmol/mg of protein) (Fig. 6b), as partial agonists are
difficult to differentiate from full agonists in the presence of a
large receptor reserve (25). Alprenolol, pindolol, propranolol,
dichloroisoproterenol, or ICI 118,551 did not affect basal adenylate cyclase activity at the low receptor density (data not
shown). However, ICI 118,551 and propranolol, which induced
the largest increase in fluorescence, potently decreased basal
adenylate cyclase activity 40 – 60% at the high receptor density
(Fig. 8b). Alprenolol and pindolol induced an approximately
30% decrease in basal activity, whereas dichloroisoproterenol
revealed weak agonist activity at this high receptor density by
increasing activity about 30%.
Quenching of Fluorescence from IANBD-labeled Receptor—
The agonist-induced decrease in fluorescence of IANBD-labeled
b2 receptor could be due to changes in the receptor structure
that increase the exposure of one or more labeled cysteines to
the aqueous solvent. Therefore, we investigated whether the
fluorescent change was accompanied by greater accessibility of
the aqueous quencher potassium iodide to the incorporated
probe. Importantly, KI did not affect ligand binding to the
receptor in a concentration up to 250 mM (data not shown). A
strong quenching of fluorescence from IANBD-labeled b2 receptor was observed following addition of increasing concentrations of KI (Fig. 9). Stern-Volmer plots showed that the negative antagonists ICI 118,551 did not affect this quenching with
the quenching constant KSV 5 4.477 6 0.074 (mean 6 S.E., n 5
3) in absence of any ligand and KSV 5 4.436 6 0.073 (mean 6
28273
FIG. 9. Stern-Volmer plots of quenching of IANBD-labeled b2
receptor. Increasing concentrations of KI were added sequentially to
labeled receptor (open square), labeled receptor treated with 10 mM ICI
118,551 (open circle) and 100 mM (2)-isoproterenol (closed circle). Fluorescence was measured and plotted as described under “Experimental
Procedures.” The quenching constant KSV was 4.477 6 0.074 (mean 6
S.E., n 5 3) in absence of any ligand, 4.436 6 0.073 (mean 6 S.E., n 5
3) in presence of ICI 118,551 and 4.209 6 0.036 (mean 6 S.E., n 5 3) in
presence of isoproterenol. The difference between isoproterenol and
buffer and between isoproterenol and ICI 118,551 was significant (p ,
0.05) (unpaired t test).
S.E., n 5 3) in presence of ICI 118,551 (Fig. 9). However, the
presence of isoproterenol unexpectedly caused slightly less
quenching as compared to buffer and ICI 118,551 (KSV 5 4.209
6 0.036, mean 6 S.E., n 5 3). The difference was significant
with p , 0.05 both when comparing isoproterenol with buffer
and isoproterenol with ICI 118,551 (unpaired t test).
DISCUSSION
In the present study we have been able to directly monitor
conformational changes in a G protein-coupled receptor. As a
molecular reporter we have used the cysteine-selective and
environmentally sensitive, fluorescent probe, IANBD, which
can be covalently incorporated into the purified, human b2
adrenergic receptor without perturbing the pharmacological
properties of the receptor. Using the IANBD-labeled protein we
were able to detect an agonist specific and reversible decrease
in fluorescence emission from the labeled receptor protein. The
response to the full agonist, isoproterenol, was shown to be
dose-dependent and could be plotted as a simple, hyperbolic
binding isotherm demonstrating interaction with a single binding site in the receptor. The ability of several adrenergic antagonists to reverse the isoproterenol-induced response confirmed that this binding site must be identical to the well
described ligand binding site in the b2 receptor. These data
provide structural evidence in a pure system for the existence
of ligand-specific conformational states of a G protein-coupled
receptor.
The agonist induced change most likely represents changes
in the polarity of the environment surrounding one of more
labeled cysteines in the hydrophobic core of the protein. Two
possible mechanisms that could account for these changes are
outlined in Fig. 10. According to the first model, a change in the
environment around the indicated fluorophore (F) could be the
result of a ligand induced movement of a transmembrane segment perpendicular to the plane of the lipid bilayer (Fig. 10a).
This could result in the exposure of the fluorophore to the
solvent causing quenching of the fluorescence and thus a decrease in the net fluorescence from the labeled receptor. Alternatively, it could be imagined that the agonist could induce a
28274
Fluorescent Labeling of the b2 Adrenergic Receptor
FIG. 10. Illustration of two possible mechanisms that may
cause changes in polarity of the environment around the
IANBD fluorophore. a, transmembrane segments of the receptor
viewed from within the lipid bilayer. A change in the environment
around the indicated fluorophore could be the result of a ligand induced
movement of a transmembrane segment perpendicular to the plane of
the lipid bilayer. b, transmembrane segments viewed from the extracellular face of the membrane. In this model, ligands could induce a
rotation of the membrane spanning domain resulting in either movement of the fluorophore into the core of the protein, which is predicted
to be more polar, or into the more hydrophobic, membrane-embedded
shell of the protein.
rotation of the membrane spanning domain resulting in movement of the fluorophore into the core of the protein, which is
predicted to be more polar (Fig. 10b) (3–5). According to the
first model, the fluorophore should be more accessible to an
aqueous quencher like potassium iodide following agonist binding. However, we observed that the agonist isoproterenol
caused a slight decrease in quenching of the fluorescence from
the IANBD-labeled receptor (Fig. 9). This would favor the latter model in which the fluorophore is moved to a more hydrophilic pocket in the core of the protein following agonist binding. In this model aqueous quenchers might be expected to
have a more limited access to the fluorophore. It should be
noted that at this point we can only speculate on the molecular
mechanism behind the ligand-induced changes in fluorescence.
With a stoichiometry of labeling at 1.2 mol of IANBD per mol of
receptor at least two sites are being labeled in the receptor.
Thus, the isoproterenol-mediated decrease in fluorescence may
involve a different labeled cysteine than that responsible for
the isoproterenol-induced effect on KI quenching. The observed
changes may therefore be due to a more complex mechanism
than that illustrated in Fig. 10.
Our fluorescence data suggest that the b2 receptor can exist
in at least two conformational states, an unliganded state and
an agonist-bound state. It is therefore tempting to consider
these data in the context of the prevailing two-state model for
activation of G protein-coupled receptors. This model predicts
that the receptors exist in a dynamic equilibrium between two
states, an inactive (R) and active conformation (R*); and that
the biological response to a given ligand is governed by its
intrinsic ability to change the overall equilibrium between the
two states (8 –10, 13, 24, 26, 27). According to this model the
fluorescence properties of the unoccupied IANBD-labeled re-
ceptor would be expected to represent the average fluorescence
properties of the population of b2 receptors in the inactive state
(R) and the active state (R*). The fluorescent properties of the
R state alone should be observed in the presence of a negative
antagonist, while those of the R* state alone should be observed
in the presence of a full agonist (8 –10, 13, 24, 26, 27). Using our
fluorescence assay we compared a series of adrenergic agonists,
which exhibited distinct efficacies in a biological assay (Fig.
5b). We found an apparent linear correlation between the functional efficacy of these compounds and their ability to promote
a change in fluorescence from the labeled receptor (Fig. 5c).
These data can be interpreted in agreement with the two-state
model. Hence, the different functional efficacies may result
from differences in the ability of the different agonists to pull
the equilibrium toward R* and thus decrease the fluorescence
signal. Alternatively, the data also could support a multistate
model in which the biological efficacy of an agonist may be a
consequence of the magnitude of conformational change that it
induces in the receptor, rather than just affecting the equilibrium between only two states. We should note, however, that
without analyzing the fluorescently labeled receptor in a reconstituted system together with the corresponding G protein, we
cannot exclude the possibility that the agonist-mediated
change in fluorescence does not describe the structural changes
of direct importance for G protein activation. Nevertheless, the
correlation between intrinsic activity of the agonists and
change in fluorescence suggests that the change is actually
describing the conformational change in the receptor that leads
to G protein activation.
It has been suggested, but never structurally verified, that
antagonists may stabilize a conformation of the receptor that is
distinct from unliganded receptor, and thus from the R state in
the two-state model (11, 28 –32). For example, this has been
proposed to explain the unexpected observation that non-peptide antagonists of G protein-coupled peptide receptors can act
as competitive antagonists for agonist peptides without sharing apparent binding sites in the receptor (28 –31). In other
words, agonists and antagonists may be able to mutually exclude each others binding to the receptor by stabilizing different receptor conformations (28 –31). It was therefore intriguing
to observe that a series of adrenergic antagonists produced
small but very reproducible increases in base-line fluorescence
from the IANBD-labeled b2 receptor (Figs. 7 and 8). Except in
the case of dichloroisoproterenol, the changes showed a correlation with the negative intrinsic activity of the tested compounds. Thus, propranolol and ICI 188,551, which exhibited
the strongest negative intrinsic activity, caused the largest
increase in base-line fluorescence, whereas alprenolol and pindolol with a smaller negative intrinsic activity also caused a
smaller increase in fluorescence. However, given the small size
of responses observed following stimulation with this group of
compounds, it is difficult to assess the molecular significance of
the fluorescence changes at the present stage. Nevertheless,
the data are still of interest as they suggest that antagonists by
themselves may alter receptor structure. The findings are also
consistent with previous proteolysis studies on membrane
bound receptor which show that agonists and antagonists are
equivalent in protecting the b2 receptor from proteolysis (33).
This would not be expected if antagonists only changed the
conformation of the small percentage of the receptor population
which, in the absence of ligand, would be predicted to be in R*
according to the two-state model.
Summarized, our data demonstrate the sensitivity of using
fluorescence techniques for studying ligand-receptor interactions and their potential for delineating ligand-induced structural changes in G protein-coupled receptors. Importantly, if
Fluorescent Labeling of the b2 Adrenergic Receptor
site-specific fluorescent labeling of the b2 receptor can be
achieved, the approach may be useful for mapping conformational changes in the receptor structure to specific subdomains.
In this way it should be possible to more precisely define the
molecular mechanism of transmembrane signal transduction
in G protein-coupled receptors.
Acknowledgments—We thank Drs. Tae R. Ji, Jeff Jasper, Elaine
Sanders-Bush, and Mervyn Maze for critical reading of the manuscript.
Drs. Robert J. Lefkowitz, Susan Senogles, Greg Dewey, and Lubert
Stryer are thanked for helpful suggestions. Dr. Irene Sun is thanked for
help with preparation of the protein.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
Hein, L., and Kobilka, B. K. (1995) Neuropharmacology 34, 357–366
Savarese, T. M., and Fraser, C. M. (1992) Biochem. J. 283, 1–19
Schwartz, T. W. (1994) Curr. Opin. Biotechnol. 5, 434 – 444
Hibert, M. F., Trumppkallmeyer, S., Bruinvels, A., and Hoffman, J. I. E. (1991)
Mol. Pharmacol. 40, 8 –15
Baldwin, J. M. (1993) EMBO J. 12, 1693–1703
Jasper, J. R., and Insel, P. A. (1992) Biochem. Pharmacol. 43, 119 –130
Milligan, G., Bond, R. A., and Lee, M. (1995) Trends Pharmacol. Sci. 16, 10 –13
Schutz, W., and Freissmuth, M. (1992) Trends Pharmacol. Sci. 13, 376 –380
Costa, T., and Herz, A. (1989) Proc. Natl. Acad. Sci. U. S. A. 86, 7321–7325
Samama, P., Pei, G., Costa, T., Cotecchia, S., and Lefkowitz, R. J. (1994) Mol.
Pharmacol. 45, 390 –394
Chidiac, P., Hebert, T. E., Valiquette, M., Dennis, M., and Bouvier, M. (1994)
Mol. Pharmacol. 45, 490 – 499
Barker, E. L., Westphal, R. S., Schmidt, D., and Sanders-Bush, E. (1994) J.
Biol. Chem. 269, 11687–11690
28275
13. Bond, R. A., Leff, P., Johnson, T. D., Milano, C. A., Rockman, H. A., McMinn,
T. R., Apparsundaram, S., Hyek, M. F., Kenakin, T. P., Allen, L. F., and
Lefkowitz, R. J. (1995) Nature 374, 272–276
14. Cerione, R. A. (1994) Method. Enzymol. 237, 409 – 423
15. Phillips, W. J., and Cerione, R. A. (1991) J. Biol. Chem. 266, 11017–11024
16. Dunn, S. M. J., and Raftery, M. A. (1993) Biochemistry 32, 8608 – 8615
17. Gettins, P. G. W., Fan, B., Crews, B. C., Turko, I. V., Olson, S. T., and
Streusand, V. J. (1993) Biochemistry 32, 8385– 8389
18. Guan, X. M., Kobilka, T. S., and Kobilka, B. K. (1992) J. Biol. Chem. 267,
21995–21998
19. Kobilka, B. K. (1995) Anal. Biochem., in press
20. Laemmli, U. K. (1970) Nature 227, 680 – 685
21. Lakowicz, J. R. (1983) Principles of Fluorescence Spectroscopy, pp. 131–135,
Plenum Press, New York
22. Suryanarayana, S., and Kobilka, B. K. (1991) Methods 3, 193–204
23. Benovic, J. L., Shorr, R. G. L., Caron, M. G., and Lefkowitz, R. J. (1984)
Biochemistry 23, 4510 – 4518
24. Samama, P., Cotecchia, S., Costa, T., and Lefkowitz, R. J. (1993) J. Biol. Chem.
268, 4625– 4636
25. Adham, N., Ellerbrock, B., Hartig, P. H., Weinshank, R. L., and Branchek, T.
(1993) Mol. Pharmacol. 43, 427– 433
26. Lefkowitz, R. J., Cotecchia, S., Samama, P., and Costa, T. (1993) Trends
Pharmacol. Sci. 14, 303–307
27. Kenakin, T. (1994) Trends Pharmacol. Sci. 15, 408 – 409
28. Gether, U., Johansen, T. E., Snider, R. M., Lowe, J. A., Nakanishi, S., and
Schwartz, T. W. (1993) Nature 362, 345–348
29. Schambye, H. T., Hjorth, S. A., Bergsma, D., Sathe, G., and Schwartz, T. W.
(1994) Proc. Natl. Acad. Sci. U. S. A. 91, 7046 –7050
30. Elling, C. E., Nielsen, S. M., and Schwartz, T. W. (1995) Nature 374, 74 –77
31. Gether, U., Lowe, J. A., III, and Schwartz, T. W. (1995) Biochem. Soc. Trans.
23, 96 –102
32. Chidiac, P. (1995) Trends Pharmacol. Sci. 16, 83– 84
33. Kobilka, B. K. (1990) J. Biol. Chem. 265, 7610 –7618