Urea unfolding of opsin in phospholipid bicelles

Article abstract of DOI:10.1111/j.1751-1097.2008.00503.x

Opsin is the unstable apo-protein of the light-activated G protein-coupled receptor rhodopsin. We investigated the stability of bovine opsin, solubilized in 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC)/detergent bicelles, against urea-induced unfold

Full text of DOI:10.1111/j.1751-1097.2008.00503.x

Photochemistry and Photobiology, 2009, 85: 494–500
Urea Unfolding of Opsin in Phospholipid Bicelles†
Craig McKibbin*1§, Nicola A. Farmer1, Patricia C. Edwards2, Claudio Villa2 and Paula J. Booth*1
1Department of Biochemistry, University of Bristol, Bristol, UK
2Medical Research Council, Laboratory of Molecular Biology, Cambridge, UK
Received 3 July 2008, accepted 14 October 2008, DOI: 10.1111 ⁄ j.1751-1097.2008.00503.x
structures of these receptors is of immense pharmaceutical
value for the rational design of drugs (4) yet progress in protein
ABSTRACT
Opsin is the unstable apo-protein of the light-activated G
protein-coupled receptor rhodopsin. We investigated the stability
of bovine opsin, solubilized in 1,2-dimyristoyl-sn-glycero-3-phos-
phocholine (DMPC) ⁄ detergent bicelles, against urea-induced
unfolding. A single irreversible protein unfolding transition was
observed from changes in intrinsic tryptophan fluorescence and
far-UV circular dichroism. This unfolding transition correlated
with loss of protein activity. Changes in tertiary structure, as
indicated by fluorescence measurements, were concomitant with
an approximate 50% reduction in a-helical content of opsin,
indicating that global unfolding had been induced by urea. The
urea concentration at the midpoint of unfolding was dependent
on the lipid ⁄ detergent environment, occurring at approximately
1.2 M urea in DMPC ⁄ 1,2-dihexanoyl-sn-glycero-3-phosphocho-
line bicelles, while being significantly stabilized to approximately
3.5 M urea in DMPC ⁄ 3-[(cholamidopropyl)dimethylammonio]-
1-propanesulfonate bicelles. These findings demonstrate that
interactions with the surrounding lipids and detergent are highly
influential in the unfolding of membrane protein structure. The
urea ⁄ bicelle system offers the possibility for a more detailed
understanding of the structural changes that take place upon
irreversible unfolding of opsin.
crystallization trials has been held back by difficulties in
achieving the vast quantities of overexpressed recombinant
protein required and by poor long-term membrane protein
stability in detergent and aqueous solution. These twin
problems for GPCR structural biology have been partly
addressed by promising developments in the employment of
phospholipid bicelles. Bicelles are mixtures of phospholipids
such as 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC)
and a detergent such as 3-[(cholamidopropyl)dimethylammo-
nio]-1-propanesulfonate (CHAPS) or a short-chain phos-
pholipid like 1,2-dihexanoyl-sn-glycero-3-phosphocholine
(DHPC). The constituent molecules associate to form disk-
like segregated aggregates possessing a lipid bilayer sur-
rounded by a detergent ⁄ short chain lipid rim (5). It has
proved possible to isolate milligram quantities of functional
purified 5-HT4 hydroxytryptamine receptor following over-
expression as inclusion bodies in Escherichia coli, solubilizing
the aggregated material in urea and then renaturing in
DMPC ⁄ CHAPS bicelles doped with cholesterol hemisuccinate
(6). Stabilization of an antagonist conformation of the beta2
adrenergic receptor in DMPC ⁄ CHAPS enabled the long-
awaited crystal structure of the receptor to be solved (7). These
successful applications of bicelles demonstrate that the pres-
ence of a detergent-solubilized phospholipid bilayer can
stabilize native conformations of GPCRs.
INTRODUCTION
The current understanding of the factors that determine how
a-helical membrane proteins fold is poor compared to water
soluble protein folding (1). Knowledge is limited to data on a
handful of membrane proteins, most notably bacteriorhodop-
sin, that can be refolded from a denatured state in vitro (2,3).
For many membrane proteins it is challenging to optimize the
hydrophobic environment that is required for preserving
activity upon extraction of the protein from a native mem-
brane. Knowledge of the folding and stability of membrane
proteins is vital for engineering superior solubilizing systems
and to answer the more fundamental question of how
membrane proteins adopt their native state.
Rhodopsin is composed of the apo-protein opsin covalently
bound to the light-sensitive cofactor 11-cis retinal and is a
GPCR responsible for light detection in the eye. It has been
calculated that binding interactions with 11-cis retinal impart
approximately )46 kJ mol)1 of energetic stability to rhodop-
sin (8). The crystal structure of bovine opsin has recently been
solved from native rod outer segment membranes by employ-
ing an astonishingly effective detergent extraction procedure
which promises a new method for studying functional opsin in
detergent solution (9,10). Until now it has been accepted that
detergent solubilization of opsin leads to rapid and irreversible
loss of 11-cis retinal binding activity (11,12) that can be
kinetically separated into distinct conformational events in the
case of dodecylmaltoside denaturation (13,14). However, opsin
is stabilized during membrane extraction and purification into
a bicelle-forming phospholipid ⁄ detergent mixture, DMPC ⁄
CHAPS (15) and to a lesser extent into DMPC ⁄ DHPC (16). In
these environments, DMPC lipid is thought to surround opsin
in a bilayer, preserving the unstable protein in a conformation
that can bind 11-cis retinal to regenerate rhodopsin with
The largest single category of vertebrate membrane proteins
are the G protein-coupled receptors (GPCRs). Solving the
†This paper is part of the Proceedings of the 13th International Conference on
Retinal Proteins, Barcelona, Spain, 15–19 June 2008.
§Current address: Craig McKibbin, Faculty of Life Sciences, Michael Smith
Building, Oxford Road, University of Manchester, Manchester M13 9PT, UK.
*Corresponding authors email: paula.booth@bristol.ac.uk (Paula J. Booth),
craig.mckibbin@manchester.ac.uk (Craig McKibbin)
ꢀ 2009 The Authors. JournalCompilation. The AmericanSocietyofPhotobiology 0031-8655/09
494

Photochemistry and Photobiology, 2009, 85 495
urea in the same DMPC ⁄ detergent and buffer to the opsin sample, to
avoid dilution of the DMPC ⁄ detergent concentration. Samples were
mixed thoroughly by pipetting and left to incubate for 30 min at 25ꢁC
with gentle mixing on a rotary wheel before spectral analysis was
carried out, unless indicated.
ꢀ90% efficiency over 24 h at room temperature (RT) in
DMPC ⁄ CHAPS. Thus this system enables the preservation of
opsin in the active, native state and in soluble form for
spectroscopic studies.
UV ⁄ Vis absorption spectroscopy. UV ⁄ Vis absorption spectra were
Understanding of GPCR folding would be greatly enhanced
by the establishment of in vitro refolding conditions. Unfor-
tunately opsin has proved difficult to refold. It therefore
becomes important to understand the processes involved in
unfolding to gain insight into why the unfolding process is
irreversible. In this work we investigate the urea-induced
denaturation of opsin in DMPC bicelles by the methods of 11-
cis retinal binding activity, fluorescence and circular dichroism
(CD) spectroscopy. We show that urea causes a concentration-
dependent destabilization in opsin structure which appears to
be irreversible. Furthermore, the resistance to chemical
unfolding is dependent on the type of bicelle system used to
solubilize opsin in vitro. To our knowledge, no report has been
published presenting chemical unfolding titration data for
rhodopsin or any other GPCR. We show that urea denatur-
ation in bicelles is a good method to delineate intrinsic
destabilizing features of the opsin fold.
We use the following terminology to differentiate different
isomers of retinal bound, or added, to the protein: covalently
bound 11-cis retinal refers to that bound to opsin via a Schiff
base linkage to Lys296; covalently bound all-trans retinal
refers to photoisomerized retinal bound via the Schiff base
link; hydrolyzed all-trans retinal results from breaking the
Schiff base linkage, and this all-trans retinal could be bound
noncovalently to opsin or be free in solution. Finally exoge-
nous 11-cis retinal refers to unbound retinal added to opsin,
from a solution in ethanol to regenerate rhodopsin.
recorded using
a Varian Cary 300 UV ⁄ Vis spectrophotometer
equipped with thermostated cuvette holders connected to a circulating
water bath. All spectra were recorded at 25ꢁC between 250 and 700 nm
with bandwidths of 2 nm, a 0.1 s integration time and a scan speed of
600 nm min)1. A 1 cm pathlength cuvette was used.
11-cis Retinal binding activity. The loss of opsin activity following
denaturation in urea was assessed by measuring the extent of
regeneration of rhodopsin by absorbance spectroscopy after
incubation of opsin with exogenously supplied 11-cis retinal in the
dark as described previously (16). Briefly, the rhodopsin concentration
before bleaching was determined by measuring the absorbance of the
chromophore at 500 nm, corrected for residual background light
scattering at 650 nm (where there is no absorbance of the chromo-
phore), giving A500 initial. After 30 min incubation with urea at 25ꢁC,
5 molar equivalents of 11-cis retinal were added to the sample
(approximately 5 lL of a 1 mM stock in ethanol) in the dark, mixed by
inversion and left to regenerate rhodopsin for 30 min. The addition of
exogenous retinal (from a concentrated solution in ethanol) appeared
to increase light scattering of the samples. Thus the extent of
regeneration (A500
)
was determined by calculating the
difference in absorbance at 500 nm between the regenerated sample
and light-scattering baseline. This baseline was produced by
regenerated
a
photobleaching the regenerated sample for 15 s, causing conversion
of rhodopsin to Meta II (380 nm absorbance peak) which has no
chromophore absorbance at 500 nm. We have previously shown that
both DMPC ⁄ CHAPS and DMPC ⁄ DHPC lipid bicelle environments
allow efficient photobleaching to Meta II (i.e. metarhodopsin I
[485 nm absorbance peak] does not accumulate considerably) (16).
With regenerated samples, this photobleaching is complete. The
rhodopsin regeneration yield was calculated as: ðA500 regenerated
=
A500 initialÞ ꢁ 100%.
Fluorescence. Fluorescence emission spectra were recorded on a
FluoroMax-2 (Jobin Yvon) at 25ꢁC using an excitation wavelength of
295 nm to excite intrinsic tryptophan residues on opsin and measuring
fluorescence emission between 305 and 450 nm with 2 nm excitation
and 4 nm emission bandwidths. The fluorescence intensity at the
maximum wavelength was used for plotting fluorescence intensity
against time or denaturant concentration. For urea unfolding time
course experiments, opsin samples at approximately 0.5 lM were
mixed manually by pipetting with the stated concentration of urea
from a 7 M urea stock as for the titrations above. The samples were
then transferred to the thermostated cuvette connected to a water bath
at 25ꢁC and emission spectra were taken once every 30 s with mixing
using a stirrer bar. Observed rate constants of urea-induced unfolding
were derived by fitting the data using GraFit software to a mono-
exponential decay function in the form:
MATERIALS AND METHODS
Materials. 11-cis Retinal was
a gift from R. Crouch (Medical
University of South Carolina and the National Eye Institute, National
Institutes of Health). Bovine rod outer segments (ROS) were provided
by G. Schertler (LMB, Cambridge, UK) and prepared as described by
Edwards et al. (17). 1D4 antirhodopsin monoclonal antibody was from
University of British Columbia. Sepharose 4B was from Amersham
Biosciences. 1D4 antibody was coupled to Sepharose 4B matrix
according to the manufacturer’s instructions. C9 elution peptide
TETSQVAPA corresponding to the C-terminus of rhodopsin was
synthesized by G. Bloomberg (University of Bristol, UK). DMPC and
DHPC were purchased from Avanti Polar Lipids, Inc. (Alabaster,
AL). CHAPS was from Calbiochem. Urea and 1,3-bis[tris(hydro-
xymethyl)methylamino]propane (BTP) were from Sigma-Aldrich.
Protein purification and preparation. ROS rhodopsin was solubilized
in DMPC ⁄ CHAPS or DMPC ⁄ DHPC and purified by 1D4 immun-
oaffinity chromatography under dim red light conditions as described
previously (16). Rhodopsin was eluted into a final buffer containing
10 mM BTP, 140 mM NaCl, pH 6.0 with 1% wt ⁄ vol DMPC and either
1% wt ⁄ vol CHAPS or 1% wt ⁄ vol DHPC. Protein concentration
and purity was determined spectrophotometrically. The molar extinc-
F ¼ F1eðꢂktÞ þ F2
where F is the fluorescence at time t, k is the observed rate constant
(kobs) and F1 and F2 are the total change in fluorescence and offset
fluorescence value, respectively.
There were slight differences in the experimental conditions for the
time course (see Fig. 2) and urea titration (see Fig. 3) experiments, and
thus the fluorescence intensities in the two are not directly comparable.
For the time course, samples were incubated in a thermostated
waterbath at 25ꢁC with mixing by magnetic stirrer bar and repeatedly
exposed to excitation wavelength light. The data were also compiled
from time courses of single experiments collected over several days
with different batches of protein. For the urea titration, samples were
at 25ꢁC in a temperature-controlled room with mixing on a rotary
wheel in the dark.
tion coefficient for rhodopsin at 500 nm (ꢀ500
) was taken as
40 600 M cm)1 (18) and eluted rhodopsin with an A280 ⁄ A500 ratio
of 1.6–1.8 was used within 24 h as the bicelles became unstable over
long periods of time and were not amenable to freeze-thawing. To
prepare opsin, purified rhodopsin was photobleached with a 300 W
projector lamp equipped with a >495 nm long-pass filter for 15 s,
followed by aging at RT in the light for 1 h. After this 1 h incubation
the sample possessed no absorbance peak at 500 nm and Metarho-
dopsin II (Meta II) formed from photobleaching had decayed to free
opsin plus hydrolyzed all-trans retinal (16).
)1
Circular dichroism. Circular dichroism spectra were recorded on a
Jasco J800 spectropolarimeter equipped with a Peltier PTC-423S
system. Wavelength spectra were acquired at
a scan speed of
20 nm min)1 with a response time of 1 s and averaged over four scans
at 25ꢁC between 180 and 270 nm with a 1 nm bandwidth. Protein
concentration was at 6 lM and a cuvette pathlength of 0.01 cm was
used.
Urea chemical unfolding titrations. A protein concentration of 0.5
and 6 lM was used for fluorescence and CD unfolding experiments,
respectively. Opsin samples were titrated with stock solutions of ꢀ7 M

496 Craig McKibbin et al.
RESULTS
Urea induces changes to the spectral characteristics of opsin
in DMPC ⁄ CHAPS consistent with unfolding
Upon photobleaching rhodopsin undergoes a series of rapid
conformational changes culminating in the formation of the
active signaling state Meta II. Meta II decays to free opsin by
hydrolysis of the unprotonated retinylidene Schiff base linkage
(19).
The fluorescence of intrinsic tryptophan residues of the
protein lining the retinal binding pocket are quenched by the
covalently bound all-trans retinal of rhodopsin. During Meta
II decay there is a reduction in the amount of fluorescence
quenching (20). This large effect of covalently bound 11-cis
retinal on tryptophan fluorescence intensity means that the
fluorescence is not very sensitive to small structural changes in
the protein that are linked to denaturation. This complicates
the study of rhodopsin unfolding by intrinsic fluorescence
changes.
By photobleaching rhodopsin purified in DMPC ⁄ CHAPS
or DMPC ⁄ DHPC bicelles for 15 s and then leaving to age in
the light for 1 h at RT (16) we were able to prepare fresh opsin
to study unfolding by fluorescence spectroscopy of the
sensitive intrinsic tryptophan probes, circumventing potential
complicating effects of covalently bound 11-cis retinal quench-
ing within the binding pocket.
Emission spectra of opsin intrinsic tryptophan fluorescence.
Opsin possesses five tryptophan residues. The intrinsic fluo-
rescence emission spectrum displays a maximum at 332 nm as
shown in Fig. 1A. Incubation of opsin with 4 M urea causes a
decreased emission intensity coupled with a slight redshift in
the fluorescence maximum to 336 nm. This change is consis-
tent with an enhanced quenching of tryptophan fluorescence
by solvent and increased polarity of the environment sur-
rounding these tryptophans.
Far-UV CD spectra of opsin secondary structure. The far-UV
CD spectrum of opsin in DMPC ⁄ CHAPS displays two
negative bands centered at 222 and 210 nm, characteristic of
a protein possessing a high degree of a-helical secondary
structure (Fig. 1B). The intensities of these bands are reduced
by nearly 50% upon incubation in 4 M urea, suggesting that a
considerable loss of helical content has been inflicted by urea
treatment. Together, the fluorescence and CD data show that
urea unfolds opsin.
Figure 1. Spectroscopic features of opsin in the presence of urea. (A)
Fluorescence emission spectra of opsin in DMPC ⁄ CHAPS in BTP
buffer at pH 6.0 (open circles) and subsequent to addition of urea to a
final concentration of 4 M and a 30 min incubation at 25ꢁC. (B) Far-
UV CD spectra of opsin in DMPC ⁄ CHAPS in phosphate buffer at pH
6.0 (open circles) and in the presence of 4 M urea (closed circles)
following the same protocol as in panel (A).
relationship between the observed rate constant of unfold-
ing and urea concentration (Fig. 2B). At 4 M urea,
kobs = 0.2 min)1, at 3 M urea, kobs = 0.02 min)1 and at 2 M
urea, kobs = 0.003 min)1. The behavior implies that the
process reported by changes in intrinsic tryptophan fluores-
cence is not an equilibrium unfolding event. It was attempted
to refold opsin following urea-induced unfolding by dilution of
urea in DMPC ⁄ CHAPS buffer. Dilution at any point during
the unfolding event, monitored by fluorescence, halted
further decreases in fluorescence (results not shown). However,
no recovery in fluorescence intensity, suggestive of refolding,
was observed under the various refolding conditions tested in
this work. Although not at equilibrium, after 30 min incuba-
tion the contrast between high concentrations and low
Rate of opsin unfolding is urea dependent and
does not reach equilibrium
In order to determine the effect of urea in more detail, the
time-dependentchangeinopsinfluorescenceinDMPC ⁄ CHAPS
over a range of urea concentrations was measured (Fig. 2A).
The rate of fluorescence decrease was found to increase with
increasing urea concentration. At each concentration of urea,
opsin fluorescence decreases toward a final stable level.
However, the time required to reach this new level is dependent
on the urea concentration. The decrease in fluorescence over
time at each urea concentration fitted well to a single
exponential decay function and there was a logarithmic

Photochemistry and Photobiology, 2009, 85 497
Figure 2. Kinetic analysis of opsin fluorescence changes in urea.
Fluorescence emission intensity at 332 nm recorded at 25ꢁC over time
for opsin in DMPC ⁄ CHAPS immediately after manual mixing with a
range of urea concentrations solubilized in DMPC ⁄ CHAPS to give
final urea concentrations of 1–6 M. (A) Time courses of opsin
fluorescence decrease at each urea concentration can be described by
a single exponential decay function. (B) The relationship between the
observed rate constant of unfolding (kobs) and urea concentration.
Figure 3. Correlation between urea-induced changes in opsin activity,
tryptophan environment and helical content. Freshly prepared opsin in
DMPC ⁄ CHAPS was titrated with a range of urea concentrations
between 0 and 5 M. Samples were incubated for 30 min at 25ꢁC and
then assessed by activity assay (A), fluorescence emission spectra (B) or
circular dichroism (C). (A) Five molar equivalents of 11-cis retinal
were added to each sample in the dark, followed by a further 30 min
incubation at RT and absorbance spectra were then recorded to
determine the percentage regeneration of the A500 absorbance band of
rhodopsin. (B) Protein intrinsic fluorescence emission intensity at
332 nm. (C) Far-UV CD ellipticity at 222 nm.
concentrations of urea on opsin fluorescence is pronounced
and thus serves as a suitable time duration for characterizing
properties of opsin during unfolding at different concentra-
tions of urea.
Loss of opsin activity in urea correlates with changes in intrinsic
tryptophan fluorescence and CD
In order to determine whether changes in fluorescence were
consistent with
a loss of protein activity, opsin in
DMPC ⁄ CHAPS was incubated with different concentrations
of urea for 30 min. The capacity of opsin to regenerate
rhodopsin was then assessed by recording the increase in the
absorbance at 500 nm after the addition of exogenous 11-cis
retinal in the dark.
Urea was found to denature opsin as defined by a loss of
11-cis retinal binding ability. In Fig. 3A, the percentage
regeneration of rhodopsin following denaturation of opsin in
DMPC ⁄ CHAPS is plotted against denaturant concentration.
The active state of opsin persists in the presence of urea below
a threshold concentration of ꢀ3 M, i.e. urea does not prevent
DMPC ⁄ CHAPS-solubilized opsin from binding 11-cis retinal
to regenerate rhodopsin. However after the 30 min incubation
with urea concentrations above 3 M, opsin is unable to
regenerate rhodopsin, indicating that urea has perturbed the
function of opsin above this concentration over a 30 min time
period.
The secondary and tertiary structure of opsin were found to
unfold in a co-operative manner upon urea denaturation. Urea
titrations of opsin in DMPC ⁄ CHAPS shown in Fig. 3B,C
reveal that the fluorescence and CD spectra changes induced
by urea have very similar relationships. When spectra taken at
30 min are plotted against urea, one steep transition centered
around 3.5 M urea is present in both fluorescence and CD data.
More remarkably, the same 50% decrease in signal intensity is
observed at 332 nm for fluorescence and 222 nm for CD, after

498 Craig McKibbin et al.
unfolding in urea. This strongly suggests that these two
complementary techniques follow the same unfolding event.
This unfolding event involves loss of secondary structure, as
shown by far-UV CD, together with changes in the tertiary
structure, monitored by protein fluorescence. At 1–2 M urea
concentration, both protein fluorescence and a-helicity as
measured by CD are increased, compared to opsin in the
absence of urea. Taken together with Fig. 2, which shows that
the fluorescence decreases over time in urea (at differing rates
depending on urea concentration), this implies that there is
another, rapid process occurring within the dead time of
mixing opsin with urea (ꢀ5 s) that initially increases protein
fluorescence and very likely increases a-helicity. This rapid
event probably reflects mixing of urea and bicelles.
Urea unfolding does not induce opsin aggregation
Disruption of native interactions during the unfolding reaction
in urea may be anticipated to drive nonspecific associations,
particularly of partly solvent-exposed hydrophobic surfaces,
resulting in aggregation. No opsin aggregation was observed in
this bicelle ⁄ urea unfolding system, thus precipitation and loss
of protein from solution is not the main contributor to the
observed signal changes. Samples were examined by eye for
signs of precipitation and were found to remain optically clear
at each concentration of urea. Moreover, no increase in light
scattering of the samples was found at any urea concentration.
During CD time-course studies (data not shown) there was no
change in total sample absorbance as read by voltage changes
in the photomultiplier tube (which at these far-UV
wavelengths gives a very sensitive measure of light scattering
from aggregates). We also noted no opsin concentration
dependency on the unfolding rate. In the fluorescence
experiments, opsin concentration was held at 0.5 lM whereas
for CD spectroscopy, 6 lM opsin was required. Therefore an
opsin concentration range of an order of magnitude is studied
in this work. It would be anticipated that if the spectroscopic
changes followed an aggregation process, this would occur at
a faster rate at higher protein concentration. In which case,
a greater change in the CD signal would be observed (i.e. for
the higher protein concentration) than in the fluorescence
signal (i.e. the lower protein concentration) after 30 min
at a particular urea concentration. However, the relative
changes in CD and fluorescence were independent of opsin
concentration.
Figure 4. Urea-induced unfolding of opsin in DMPC ⁄ CHAPS and
DMPC ⁄ DHPC. Protein intrinsic fluorescence emission intensity at
332 nm of opsin in DMPC ⁄ DHPC (open circles) or DMPC ⁄ CHAPS
(closed circles), titrated with a range of urea concentrations between
0 and 5 M.
DISCUSSION
Opsin is an inherently unstable membrane protein and
undergoes rapid loss of activity when solubilized in traditional
detergents (12). By studying opsin in stabilizing DMPC ⁄
CHAPS bicelles which preserve activity (15,21), we were able
to disconnect the denaturation processes resulting from the
addition of chaotropic agents, from denaturation that ensues
as a consequence of a detergent solubilization process.
Urea denaturation of opsin in bicelles causes loss of both
secondary and tertiary structure together with a loss of ability
to bind 11-cis retinal to regenerate rhodopsin. This unfolding
of opsin is not accompanied by protein aggregation, but is
irreversible with the rate of denaturation being dependent on
urea concentration (see Fig. 2). Interestingly, changes in
tertiary structure as indicated by fluorescence changes corre-
late with the loss in opsin secondary structure as indicated by
CD changes. Approximately 50% of the initial helical struc-
ture is lost upon unfolding at urea concentrations above the
transition midpoint.
Many studies of opsin have previously concluded that loss
of 11-cis retinal binding activity, i.e. denaturation, whether due
to disruption of the membrane, pH or detergent solubilization
is an irreversible process (11–13,22,23). In particular, Landin
et al. demonstrated by differential scanning calorimetry that
thermal denaturation of opsin in native membranes occurs at
around 60ꢁC by an apparently single irreversible transition
that is kinetically controlled and hence dependent on scanning
rate (24). Heating opsin in DMPC ⁄ CHAPS to 60ꢁC results in a
very similar fluorescence intensity (results not shown) to that
of urea-denatured opsin. Based on the data, we propose the
following reaction scheme for opsin unfolding in urea:
Opsin is less stable to urea in DMPC ⁄ DHPC bicelles
The effects of urea on opsin fluorescence were compared in
DMPC ⁄ CHAPS and DMPC ⁄ DHPC bicelles. Opsin was
found to be more resistant to urea denaturation in
DMPC ⁄ CHAPS than in DMPC ⁄ DHPC. Figure 4 shows
fluorescence emission intensity titration plots for opsin in
DMPC ⁄ DHPC and DMPC ⁄ CHAPS subjected to denatur-
ation by the addition of urea. In each case, the data reveal a
sudden decrease in opsin fluorescence over a short range of
denaturant concentration. A lower concentration of urea is
required to induce fluorescence decreases for opsin solubilized
in DMPC ⁄ DHPC (1.2 M urea) than in DMPC ⁄ CHAPS (3.5 M
urea). This is in accordance with the finding that opsin is
comparatively less stable in DMPC ⁄ DHPC (16).
opsinN
k½ureaꢃ opsinU
!
In this proposed reaction scheme, correctly folded opsin
(opsinN) undergoes an irreversible transition to an unfolded,
denatured form of opsin (opsinU). The rate of this unfolding
reaction is dependent on the concentration of urea.

Photochemistry and Photobiology, 2009, 85 499
The unfolding reaction of opsin solubilized in DMPC ⁄
DHPC was significantly more sensitive to urea concentration
than in DMPC ⁄ CHAPS (Fig. 4). This agrees well with our
previous work characterizing the stability of opsin
in DMPC ⁄ CHAPS and DMPC ⁄ DHPC bicelles (16). The
increased opsin stability imparted by DMPC ⁄ CHAPS may
result from decreased conformational freedom of the protein
due to CHAPS increasing the bicelle rigidity (25). Alterna-
tively, a direct interaction between opsin and CHAPS could
contribute to the enhanced stability in DMPC ⁄ CHAPS.
Cholesterol and CHAPS share a planar steroidal chemical
structure and several studies have suggested that rhodopsin
binds cholesterol (26,27). We propose that as the opsin
stability is decreased in DMPC ⁄ DHPC, the energy barrier to
the unfolded state is diminished. Hence an increased fraction
of opsin molecules undergo unfolding over time at RT at a
given concentration of destabilizing urea.
It has been reported for several helical membrane proteins
that traditional unfolding agents such as urea and guanidinium
are ineffective denaturants and this has been attributed to the
generally hydrophobic exterior of membrane proteins exclud-
ing access to polar chaotropes (28–30). However, urea has been
used to obtain functional leukotriene B4 and serotonin 5-HT4(a)
receptors from bacterial inclusion bodies (6,31). This demon-
strates that solubilization of inclusion bodies by urea can be an
effective strategy for GPCR production but the extent of
structural unfolding and refolding was not provided in this
earlier work (6). Nonetheless, the functional receptor was
recovered from the inclusion protein aggregates and while the
detergent lauryldimethylamine oxide was sufficient to refold
the leukotriene receptor, a mixed DMPC ⁄ CHAPS ⁄ cholesterol
hemisuccinate system was required for effective refolding of the
5-HT4(a) receptor. This suggests that despite the anticipated
conformational similarity within the GPCR family, require-
ments for individual receptors will need to be determined
empirically and bicelles may have an important role to play.
For opsin, we show that relatively low molar concentrations
of urea can induce unfolding, as revealed by a considerable
reduction in protein a-helical content in the CD spectrum. The
susceptibility of opsin to urea denaturation may be due to
the large globular extracellular structural domain that caps
the ends of its transmembrane helices: mutational studies of
rhodopsin folding and assembly in vivo led to the conclusion
that formation of a disulfide bond between Cys110 and Cys187
in the extracellular domain is essential for protein stability
(32), that this domain has exacting tertiary folding require-
ments (33) and its folding is coupled to correct packing of the
transmembrane domains for 11-cis retinal binding (33,34). The
solvent accessibility of this region means it could be denatured
by urea, leading to subsequent unfolding of the transmem-
brane domain.
affect the pathways of mechanical unfolding available to
rhodopsin (37). Thus the importance of the extracellular
domain has been identified as a critical structural element in
the unfolding reaction.
There is little information on the precise nature of the
unfolded state and why the denaturation is irreversible. Our
CD data suggest that close to a 50% loss in helical content is
found in the unfolded state in urea. As for the molecular
mechanism for irreversible denaturation of opsin, it seems
possible that the fold carefully adopted during integration into
the membrane is in fact kinetically controlled rather than a
true thermodynamically attained lowest energy conformation.
In the absence of covalently bound 11-cis retinal cofactor, this
‘‘trapped’’ structure is increasingly sensitive to insults such as
detergent molecules or urea, resulting in collapse to a lower
energy, inactive conformation.
Refolding a-helical membrane proteins from unfolded states
is a challenging enterprise and while the reversible refolding of
opsin at this point does not appear possible, it will be
fascinating to explore whether the polypeptide sequences of
other GPCRs retain the complete set of directions for adopting
their own native conformations. The experimental system
presented in this study may be useful in further determining
the features of the intrinsic instability of opsin.
Acknowledgements—We thank Gebhard Schertler for assistance with
this work and the Medical Research Council (studentship to C.M.),
EU (to P.J.B. as partner in EMeP) and the Wellcome Trust for
funding.
REFERENCES
1. Mayor, U., N. Guydosh, C. Johnson, J. Grossmann, S. Sato,
G. Jas, S. Freund, D. Alonso, V. Daggett and A. Fersht (2003)
The complete folding of a protein from nanoseconds to micro-
seconds. Nature 421, 863–867.
2. London, E. and H. G. Khorana (1982) Denaturation and
renaturation of bacteriorhodopsin in detergents and lipid-
detergent mixtures. J. Biol. Chem. 257, 7003–7011.
3. Booth, P. J. and P. Curnow (2006) Membrane proteins shape up:
Understanding in vitro folding. Curr. Opin. Struct. Biol. 16, 480–
488.
4. Lagerstrom, M. and H. Schioth (2008) Structural diversity of G
protein-coupled receptors and significance for drug discovery.
Nat. Rev. Drug Discov. 7, 339–357.
5. Vold, R. R., R. S. Prosser and A. J. Deese (1997) Isotropic solutions
of phospholipid bicelles: A new membrane mimetic for high-reso-
lution NMR studies of polypeptides. J. Biomol. NMR 9, 329–335.
6. Baneres, J. L., D. Mesnier, A. Martin, L. Joubert, A. Dumuis and
J. Bockaert (2005) Molecular characterization of a purified 5-HT4
receptor: A structural basis for drug efficacy. J. Biol. Chem. 280,
20253–20260.
7. Rasmussen, S., H. Choi, D. Rosenbaum, T. Kobilka, F. Thian,
P. Edwards, M. Burghammer, V. Ratnala, R. Sanishvili, R.
Fischetti, G. Schertler, W. Weis and B. Kobilka (2007) Crystal
structure of the human beta2 adrenergic G protein-coupled
receptor. Nature 450, 383–387.
8. Okada, T., O. P. Ernst, K. Palczewski and K. P. Hofmann (2001)
Activation of rhodopsin: New insights from structural and bio-
chemical studies. Trends Biochem. Sci. 26, 318–324.
9. Okada, T., K. Takeda and T. Kouyama (1998) Highly selective
separation of rhodopsin from bovine rod outer segment mem-
branes using combination of divalent cation and alkyl(thio)
glucoside. Photochem. Photobiol. 67, 495–499.
Rhodopsin has been the subject of a number of independent
studies focusing on the unfolding reaction in recent years.
Computer-simulated thermal unfolding analysis has predicted
that long-range interactions create a stable core of amino acids
in rhodopsin that are grouped around the conserved extracel-
lular Cys110–Cys187 disulfide bond (35). Furthermore, pre-
dicted destabilizing mutations of rhodopsin match well with
experimental findings (36). This suggestion gains support from
single-molecule force microscopy studies where the presence or
absence of the Cys110–Cys187 disulfide bond was found to
10. Park, J. H., P. Scheerer, K. P. Hofmann, H. W. Choe and
O. P. Ernst (2008) Crystal structure of the ligand-free G-protein-
coupled receptor opsin. Nature 454, 183–187.

500 Craig McKibbin et al.
11. Stubbs, G. W. and B. J. Litman (1978) Effect of alterations in the
amphipathic microenvironment on the conformational stability of
bovine opsin. 2. Rate of loss of opsin regenerability. Biochemistry
17, 220–225.
12. De Grip, W. J. (1982) Thermal stability of rhodopsin and opsin in
some novel detergents. Methods Enzymol. 81, 256–265.
13. Sakamoto, T. and H. G. Khorana (1995) Structure and function in
rhodopsin: The fate of opsin formed upon the decay of light-
activated metarhodopsin II in vitro. Proc. Natl Acad. Sci. USA 92,
249–253.
14. McKibbin, C., A. Toye, P. Reeves, H. Khorana, P. Edwards,
C. Villa and P. Booth (2007) Opsin stability and folding: The role
of Cys185 and abnormal disulfide bond formation in the intra-
discal domain. J. Mol. Biol. 374, 1309–1318.
15. Reeves, P. J., J. Hwa and H. G. Khorana (1999) Structure and
function in rhodopsin: Kinetic studies of retinal binding to puri-
fied opsin mutants in defined phospholipid-detergent mixtures
serve as probes of the retinal binding pocket. Proc. Natl Acad. Sci.
USA 96, 1927–1931.
25. Andersson, A. and L. Maler (2005) Magnetic resonance investi-
gations of lipid motion in isotropic bicelles. Langmuir 21, 7702–
7709.
26. Ruprecht, J. J., T. Mielke, R. Vogel, C. Villa and G. F. Schertler
(2004) Electron crystallography reveals the structure of meta-
rhodopsin I. EMBO J. 23, 3609–3620.
27. Albert, A. D., J. E. Young and P. L. Yeagle (1996) Rhodopsin-
cholesterol interactions in bovine rod outer segment disk mem-
branes. Biochim. Biophys. Acta 1285, 47–55.
28. Huang, K. S., H. Bayley and H. G. Khorana (1980) Delipidation
of bacteriorhodopsin and reconstitution with exogenous phos-
pholipid. Proc. Natl Acad. Sci. USA 77, 323–327.
29. Chen, G. Q. and E. Gouaux (1999) Probing the folding and
unfolding of wild-type and mutant forms of bacteriorhodopsin in
micellar solutions: Evaluation of reversible unfolding conditions.
Biochemistry 38, 15380–15387.
30. Nagy, J. K., W. L. Lonzer and C. R. Sanders (2001) Kinetic study
of folding and misfolding of diacylglycerol kinase in model
membranes. Biochemistry 40, 8971–8980.
16. McKibbin, C., N. Farmer, C. Jeans, P. Reeves, H. Khorana,
B. Wallace, P. Edwards, C. Villa and P. Booth (2007) Opsin sta-
bility and folding: Modulation by phospholipid bicelles. J. Mol.
Biol. 374, 1319–1332.
17. Edwards, P. C., J. Li, M. Burghammer, J. H. McDowell, C. Villa,
P. A. Hargrave and G. F. Schertler (2004) Crystals of native and
modified bovine rhodopsins and their heavy atom derivatives.
J. Mol. Biol. 343, 1439–1450.
31. Baneres, J. L., A. Martin, P. Hullot, J. P. Girard, J. C. Rossi and
J. Parello (2003) Structure-based analysis of GPCR function:
Conformational adaptation of both agonist and receptor upon
leukotriene B4 binding to recombinant BLT1. J. Mol. Biol. 329,
801–814.
32. Karnik, S. S. and H. G. Khorana (1990) Assembly of functional
rhodopsin requires a disulfide bond between cysteine residues 110
and 187. J. Biol. Chem. 265, 17520–17524.
18. Wald, G. and P. K. Brown (1953) The molar extinction of rho-
dopsin. J. Gen. Physiol. 37, 189–200.
19. Wald, G. and P. K. Brown (1956) Synthesis and bleaching of
rhodopsin. Nature 177, 174–176.
33. Doi, T., R. S. Molday and H. G. Khorana (1990) Role of the
intradiscal domain in rhodopsin assembly and function. Proc. Natl
Acad. Sci. USA 87, 4991–4995.
34. Hwa, J., P. Garriga, X. Liu and H. G. Khorana (1997) Structure
and function in rhodopsin: Packing of the helices in the trans-
membrane domain and folding to a tertiary structure in the
intradiscal domain are coupled. Proc. Natl Acad. Sci. USA 94,
10571–10576.
35. Rader, A. J., G. Anderson, B. Isin, H. G. Khorana, I. Bahar and
J. Klein-Seetharaman (2004) Identification of core amino
acids stabilizing rhodopsin. Proc. Natl Acad. Sci. USA 101, 7246–
7251.
36. Tastan, O., E. Yu, M. Ganapathiraju, A. Aref, A. Rader and
J. Klein-Seetharaman (2007) Comparison of stability predictions
and simulated unfolding of rhodopsin structures. Photochem.
Photobiol. 83, 351–362.
20. Farrens, D. L. and H. G. Khorana (1995) Structure and function
in rhodopsin. Measurement of the rate of metarhodopsin II decay
by fluorescence spectroscopy. J. Biol. Chem. 270, 5073–5076.
21. McKibbin, C., N. A. Farmer, C. Jeans, P. J. Reeves,
H. G. Khorana, B. A. Wallace, P. C. Edwards, C. Villa and P. J.
Booth (2007) Opsin stability and folding: Modulation by phos-
pholipid bicelles. J. Mol. Biol. 374, 1319–1332.
22. Radding, C. M. and G. Wald (1956) The stability of rhodopsin
and opsin; effects of pH and aging. J. Gen. Physiol. 39, 923–933.
23. Hubbard, R. (1958) The thermal stability of rhodopsin and opsin.
J. Gen. Physiol. 42, 259–280.
24. Landin, J. S., M. Katragadda and A. D. Albert (2001) Thermal
destabilization of rhodopsin and opsin by proteolytic cleavage in
bovine rod outer segment disk membranes. Biochemistry 40,
11176–11183.
37. Tanuj Sapra, K., P. S. Park, S. Filipek, A. Engel, D. J. Muller and
K. Palczewski (2006) Detecting molecular interactions that stabi-
lize native bovine rhodopsin. J. Mol. Biol. 358, 255–269.