X-ray structures of general anaesthetics bound to a
pentameric ligand-gated ion channel
Hugues Nury1,2,3,4, Catherine Van Renterghem1,2, Yun Weng5, Alphonso Tran5, Marc Baaden6, Virginie Dufresne1,2,
Jean-Pierre Changeux2,7, James M. Sonner5, Marc Delarue3,4& Pierre-Jean Corringer1,2
General anaesthetics have enjoyed long and widespread use but
their molecular mechanism of action remains poorly understood.
There is good evidence that their principal targets are pentameric
ligand-gated ion channels1,2(pLGICs) such as inhibitory GABAA
anaesthetics. The bacterial homologue from Gloeobacter violaceus3
(GLIC), whose X-ray structure was recently solved4,5, is also sensitive
to clinical concentrations of general anaesthetics6. Here we describe
pre-exists in the apo-structure in the upper part of the transmem-
brane domain of each protomer. Both molecules establish van der
inside. Mutations of some amino acids lining the binding site pro-
foundly alter the ionic response of GLIC to protons, and affect its
general-anaesthetic pharmacology. Molecular dynamics simulations,
performed on the wild type (WT) and two GLIC mutants, highlight
differences in mobility of propofol in its binding site and help to
for the design of general anaesthetics and of allosteric modulators of
Understanding the mechanism of action of general anaesthetics
requires the identification of their binding site(s) within the three-
dimensional structure of pLGICs. To identify such a site, we used
the pH-gated bacterial homologue GLIC, a homopentameric member
of the pLGIC family that was recently shown to be sensitive to general
anaesthetics6and amenable to X-ray structure determination4,5.
Co-crystals ofGLIC were grown with propofoland apo-GLICcrys-
tals were equilibrated in mother liquor saturated with desflurane.
pofol) resolution. In both structures, strong densities in otherwise
empty Fourier Fo2Fcdifference maps revealed bound anaesthetics
in each subunit (mean peak height 8.860.6s for desflurane and
5.960.8s for propofol; Supplementary Fig. 1). Both molecules were
found to bind in the same region, with little change in the protein
conformation compared with apo-GLIC4, a feature observed also for
some soluble proteins complexed with general anaesthetics7–9. The
binding site is locatedintheupperhalfof thetransmembranedomain
The general-anaesthetic cavity is accessible from the lipid bilayer and
progressively narrows down towards the interior of the subunit
(Fig. 1b, c). Another cavity of comparable volume is located at the
from the outside. A narrow tunnel (less than 3A˚in diameter) links
F303, near the mouth), and from the b6–b7 loop (Y119, P120, F121,
(T255 and I258) and M2 (V242). Its oxygenatom is within hydrogen-
bonddistanceof the T255 hydroxylgroup.Significantadditionalelec-
to lipids, with an alkyl chain obstructing the cavity entrance (Fig. 1b),
as observed with the apo structure. This defines a cavity volume of
238A˚3whereas the volume of desflurane is 94A˚3. Propofol lies closer
to the entrance of the general-anaesthetic cavity and would clash with
the lipid seen in the apo and desflurane structures. Accordingly, the
T255 and Y254, byvanderWaals contacts(Fig.2).Intheorientations
that best fitthe density maps,the propofolhydroxylgroupcould form
chain, whereas desflurane is 3.5A˚away.
We have recently shown that GLIC activation is inhibited by most
general anaesthetics at clinical concentrations6. To check whether the
general-anaesthetic binding sites contribute to this inhibition, we
more bulky residues (mutants I202A,W,Y, V242M,W and T255A),
and studied the functional effect by two-electrode voltage-clamp elec-
trophysiology in oocytes.
Among the mutants tested, I202Y and T255A produce a marked
towards lower concentrations (pH5056.160.1 with Hill number
(nH)52.060.2 and pH5056.060.2 with nH51.160.2 respec-
tively), compared with WT (pH5055.060.3 with nH51.860.3)
(Fig. 3a, b). T255A shows also slower apparent rate constants for
plementary Table 2).
The inhibitory action of general anaesthetics was measured around
the one-fifth maximum effective concentration (EC20) of proton
activation. On WT, propofol and desflurane produced 100% maximal
inhibition with half-maximum inhibitory concentration (IC50) values
of 2466.3mM (nH51.160.2) and 27613mM (nH50.360.2)
respectively. Screening the mutants for inhibition by 10mM propofol
but shows that V242M and T255A produce a parallel tenfold shift of
the propofol dose–inhibition curve to lower concentration (Fig. 3c, d
and Supplementary Table 3). In contrast, V242M has no effect on
desflurane inhibition, whereas T255A produces a tenfold shift of the
desflurane inhibition curve towards higher concentrations. Measure-
ment of general-anaestheticinhibitionatdifferent pHshows thatboth
anaesthetics are more efficient at higher pH, for both WT and T255A
(Fig. 3e and Supplementary Fig. 2). Strikingly, T255A increases the
1Institut Pasteur, Groupe Re ´cepteurs-Canaux, F-75015 Paris, France.2CNRS, URA2182, F-75015 Paris, France.3Institut Pasteur, Unite ´ de Dynamique Structurale des Macromole ´cules, F-75015 Paris,
9080, F-75005 Paris, France.7Colle `ge de France, F-75005 Paris, France.
4 2 8 | N A T U R E | V O L 4 6 9 | 2 0 J A N U A R Y 2 0 1 1
Macmillan Publishers Limited. All rights reserved
inhibition by propofol but decreases the inhibition by desflurane at all
the general-anaesthetic binding site affects (1) the intrinsic ionic res-
mutation3,10, and (2) the pharmacology of general anaesthetics, illu-
strated both by V242M, which displays an increased sensitivity to pro-
pofol but not to desflurane, and T255A, which has an increased
sensitivity to propofol but a decreased sensitivity to desflurane.
ing site described here contributes to general-anaesthetic-mediated
inhibition of GLIC. For desflurane, mutagenesis data match the char-
acteristics of its binding site in the X-ray structure well, with no sig-
nificant effect when the relatively distant positions 202 and 242 are
mutated, and a strong impairing effect when mutating T255, which
extensively contacts desflurane, into an alanine. In contrast, for pro-
pofol,both positions202and255contactpropofol,butonly mutation
direct contact with propofol but V242M modifies its response. These
data suggest a significant mobility of propofol within the cavity, a
feature that may be reflected by the high B factors of general anaes-
thetics in the crystal structure (Bdesflurane5121A˚2, Bpropofol5135A˚2,
cannot be discriminated at 3.3-A˚resolution.
To examine this possibility further, we performed 30-ns molecular
dynamics simulations of propofol bound to the WT protein, T255A,
associated with (1) reduced propofol fluctuation (root mean square
fluctuation of propofol non-H atoms of 361.1A˚, 2.460.8A˚,
2.360.8A˚, 2.7A˚for the WT, T255A, V242M and I202A runs,
respectively), (2) deeper penetration inside the cavity (Fig. 4b) and
(3) more frequent interaction with residue 242 compared with the
WT and I202A (data not shown). Altogether, these simulations pro-
vide complementary interpretations to account for the higher sensiti-
vity of T255A and V242M to propofol inhibition that could not have
been deduced from the static structure alone.
apparently open conformation4,5. But general anaesthetics behave as
inhibitors of the ionic response and are therefore expected to stabilize
a closed conformation. Our data unravel a general-anaesthetic site in
timescale. This apparent contradiction can be readily explained by a
non-exclusive(differential)binding ofgeneral anaesthetics tothe open
and closed states, with general anaesthetics displaying a higher affinity
Figure 1 | Propofol and desflurane binding sites. a, General view of GLIC
from the plane of the membrane in cartoon representation with a bound
general-anaesthetic molecule in space-filling representation. The molecular
surface is represented in the insets and coloured in yellow for the binding
pocket. b, Cartoon andsurface representationof thegeneral-anaesthetic cavity
seen from the membrane (left) and from the adjacent subunit (right, M1
two structures (green and orange respectively) depicted as sticks. For this
of 0.13A˚. c, Molecular surface of the general-anaesthetic intra-subunit cavities
(yellow) and neighbouring inter-subunit cavities (pink) for the whole
pentamer. In one of the subunits, the communication tunnel between the two
cavities is depicted in orange, and its constriction indicated by an arrow in the
Y119, P120, F121
Figure 2 | Residuesofthebindingsite. Sitesforpropofol(right)anddesflurane
(left), viewed from the membrane (top panelswith M4helixremoved),andfrom
the ECD domain (lower panels with ECD removed). Residues bordering the
sticks. SigmaA weighted Fourier difference maps 2Fo2Fccontoured at 1.5s
around the anaesthetics molecules are represented as a blue mesh.
2 0 J A N U A R Y 2 0 1 1 | V O L 4 6 9 | N A T U R E | 4 2 9
Macmillan Publishers Limited. All rights reserved
of the general-anaesthetic binding site during gating, in line with its
gate, and close to the transmembrane-domain–extracellular-domain
interface. We note that the hypothetical gating mechanism4previously
a strong reorganization of the general-anaesthetic binding site as a
consequence of M2 and M3 helix tilting.
and desflurane structures and is displaced in the propofol structure.
ing within protein subsites known to be critical for the transmembrane
domain structure13,14. It is known that lipids contribute to pLGIC func-
tion15,16, and those observed in the present electron-density maps are
endogenous allosteric modulators17, lipids in this case, but also possibly
fatty acids, cholesterol and/or neurosteroids18in the case of eukaryotic
It is striking that the pharmacology of GLIC inhibition resembles
that of nicotinic acetylcholine receptors (nAChRs), which are also
anaesthetics19. The general-anaesthetic binding site described here is
Gly/GABAAreceptors are mostly potentiated by general anaesthetics.
Experimental data involving chimaeric constructs show that GlyRa1
and alcohols potentiation2,20. The general anaesthetic etomidate
labelled brain GABAAreceptors at residues a1M236 and a3M28621,
interaction between several binding sites. Overall, the interpretation
of these data using homology models based on the cryoelectron-
microscopy nAChR structure at medium resolution23–25suggests that
intra-subunit and/or inter-subunit sites in the upper part of the trans-
The sequence of GLIC can be readily aligned with that of GABAAR
and GlyR (Supplementary Fig. 3). From the GLIC three-dimensional
structure and this alignment, the three residues identified in GABAA/
Gly receptors can be seen to point towards the inter-subunit cavity
(Fig. 1c) close to the intra-subunit general-anaesthetic binding site
in the initial steps of channel closing during our 1-ms molecular
dynamics simulation of GLIC at neutral pH26(Supplementary Fig.
4). This involves transient communications between the inter- and
intra-subunit cavities caused by M2 and M3 motions, which further
to channel gating, and could suggest that a cross-talk between both
cavities might underlie general-anaesthetic-mediated potentiation.
pLGIC and particularly nAChRs are not only the target of general
anaesthetics27but also of natural and synthetic allosteric modulators
that are developed for their therapeutic potential28,29. Mutational data
suggest that ivermectine30and PNU-120596, which behave as positive
6.05.55.35.0 4.54.0 3.77.3
6.56.0 5.8 5.54.08.0
log (DSF concentration)
(log (mol l–1))
log (PPF concentration)
(log (mol l–1))
EC 23EC 59
EC 14EC 32
Figure 3 | Electrophysiological characterization of binding-site residues.
a, Traces of currents evoked by 30-s applications of low extracellular pH
separated by 30–60s wash. b, Corresponding plots for currents normalized
with respect to the value at pH 4. Mean6s.d. of 4 to 12 cells per construct.
c, Inhibition by 0.5mM desflurane (left) or 10mM propofol (right) applied for
60s during the plateau of GLIC activation by a pH near EC20(EC10–30).
d, Corresponding concentration–inhibition characteristics of desflurane (left)
or propofol (right). e, Current traces showing the effect of 10mM propofol on
GLIC currents corresponding toprotonEC3,23,59(WT, left traces) or EC14,32,57
Distance to pore centre (Å)
Figure 4 | Molecular dynamics simulation of propofol bound to GLIC.
a, View from the ECD domain depicting propofol positions as green sticks,
and final (transparent cartoon) conformation of the protein, during the 30-ns
distribution between the propofol centre of mass and the pore centre. The
distribution is shifted closer to the pore centre for the mutants showing an
4 3 0 | N A T U R E | V O L 4 6 9 | 2 0 J A N U A R Y 2 0 1 1
Macmillan Publishers Limited. All rights reserved
allosteric modulatorsof a7 nAChR, bindat a location resembling that
lators inhibiting or potentiating pLGICs.
GLIC production4, electrophysiology3and molecular dynamics26were performed
as described (full methods in the Supplementary Information). Crystals were
typically grown in 12–16% PEG 4000, 400mM NaSCN, 100mM Na-Acetate at
pH 4, in the presence of an excess of general anaesthetics.
Full Methods and any associated references are available in the online version of
the paper at www.nature.com/nature.
Received 29 April; accepted 1 November 2010.
sleep and arousal. Nature Rev. Neurosci. 9, 370–386 (2008).
Lobo, I. A. & Harris, R. A. Sites of alcohol and volatile anaesthetic action on glycine
receptors. Int. Rev. Neurobiol. 65, 53–87 (2005).
Bocquet, N. et al. A prokaryotic proton-gated ion channel from the nicotinic
acetylcholine receptor family. Nature 445, 116–119 (2007).
Bocquet, N. et al. X-ray structure of a pentameric ligand-gated ion channel in an
apparently open conformation. Nature 457, 111–114 (2009).
Hilf,R.J.C.& Dutzler, R.Structureofapotentiallyopenstateofa proton-activated
pentameric ligand-gated ion channel. Nature 457, 115–118 (2009).
Weng, Y., Yang, L., Corringer, P. J. & Sonner, J. M. Anaesthetic sensitivity of the
Bhattacharya, A. A., Curry, S. & Franks, N. P. Binding of the general anaesthetics
propofol and halothane to human serum albumin. High resolution crystal
structures. J. Biol. Chem. 275, 38731–38738 (2000).
Zhang, H., Astrof, N. S., Liu, J., Wang, J. & Shimaoka, M. Crystal structure of
isoflurane bound to integrin LFA-1 supports a unified mechanism of volatile
anaesthetic action in the immune and central nervous systems. FASEB J. 23,
284, 24176–24184 (2009).
nicotinic receptor. Nature 353, 846–849 (1991).
non-exclusive ligand binding. J. Mol. Biol. 21, 265–274 (1966).
12. Hilf, R. J. C. & Dutzler, R. X-ray structure of a prokaryotic pentameric ligand-gated
ion channel. Nature 452, 375–379 (2008).
13. Haeger, S. et al. An intramembrane aromatic network determines pentameric
assembly of Cys-loop receptors. Nature Struct. Mol. Biol. 17, 90–98 (2010).
14. Villmann, C. et al. Functional complementation of Glra1spd-ot, a glycine receptor
15. Nievas, G. A. F., Barrantes, F. J. & Antollini, S. S. Modulation of nicotinic
Chem. 283, 21478–21486 (2008).
16. da Costa, C. J. B. et al. Anionic lipids allosterically modulate multiple nicotinic
acetylcholine receptor conformational equilibria. J. Biol. Chem. 284,
17. Franks, N. P. & Lieb, W. R. Do general anaesthetics act by competitive binding to
specific receptors? Nature 310, 599–601 (1984).
18. Hosie, A. M., Wilkins, M. E., da Silva, H. M. A. & Smart, T. G. Endogenous
neurosteroids regulate GABAAreceptors through two discrete transmembrane
sites. Nature 444, 486–489 (2006).
19. Violet, J. M., Downie, D. L., Nakisa, R. C., Lieb, W. R. & Franks, N. P. Differential
sensitivitesofmammalian neuronal and muscle nicotinic acetylcholine receptors
to general anesthetics. Anesthesiology 86, 866–874 (1997).
20. Mihic, S. J. et al. Sites of alcohol and volatile anaesthetic action on GABAAand
glycine receptors. Nature 389, 385–389 (1997).
21. Li, G. et al. Identification of a GABAAreceptor anaesthetic binding site at subunit
interfaces by photolabeling with an etomidate analog. J. Neurosci. 26,
22. Li, G., Chiara, D. C., Cohen, J. B. & Olsen, R. W. Numerous classes of general
anaesthetics inhibit etomidate binding to c-aminobutyric acid type A (GABAA)
receptors. J. Biol. Chem. 285, 8615–8620 (2010).
23. Miyazawa, A., Fujiyoshi, Y. & Unwin, N. Structure and gating mechanism of the
acetylcholine receptor pore. Nature 423, 949–955 (2003).
24. Trudell, J. R. & Bertaccini, E. Comparative modeling of a GABAAa1 receptor using
three crystal structures as templates. J. Mol. Graph. Model. 23, 39–49 (2004).
25. Bali, M., Jansen, M. & Akabas, M. H. GABA-induced intersubunit conformational
experimentalbasis forhomologymodeling ofanintravenous anaesthetic binding
site. J. Neurosci. 29, 3083–3092 (2009).
in a nicotinic receptor homologue. Proc. Natl Acad. Sci. USA 107, 6275–6280
binding sites in the nicotinic acetylcholine receptor for [3H]azietomidate, a
photoactivatable general anaesthetic. J. Biol. Chem. 279, 17640–17649 (2004).
allosteric transitions and therapeutic targets in the nervous system. Nature Rev.
Drug Discov. 8, 733–750 (2009).
29. Bertrand,D.& Gopalakrishnan, M.Allosteric modulation ofnicotinic acetylcholine
receptors. Biochem. Pharmacol. 74, 1155–1163 (2007).
30. Krause, R. M. et al. Ivermectin: a positive allosteric effector of the a7 neuronal
nicotinic acetylcholine receptor. Mol. Pharmacol. 53, 283–294 (1998).
Supplementary Information is linked to the online version of the paper at
Acknowledgements This work was supported by the Commission of the European
Communities (Neurocypres project; to H.N.), the Louis D. Foundation oftheInstitut de
France, the Network of European Neuroscience Institutes (ENI-NET) and a National
gift of desflurane, the European Synchrotron Radiation Facility and Soleil staff for
assistance during data collection, and G. Brannigan for providing propofol simulation
parameters. Simulations were performed using high-performance computing
resources from the Grand Equipement National de Calcul Intensif, Institut du
De ´veloppement et des Ressources en Informatique Scientifique (GENCI-IDRIS, grant
Author Contributions Allauthorscontributed extensivelytotheworkpresentedinthis
Author Information The coordinates of models are deposited in Protein Data Bank
under accession numbers 3P50 (propofol) and 3P4W (desflurane). Reprints and
permissionsinformationisavailable atwww.nature.com/reprints.The authors declare
no competing financial interests. Readers are welcome to comment on the online
version of this article at www.nature.com/nature. Correspondence and requests for
materials should be addressed to P.-J.C. (firstname.lastname@example.org) or M.D.
2 0 J A N U A R Y 2 0 1 1 | V O L 4 6 9 | N A T U R E | 4 3 1
Macmillan Publishers Limited. All rights reserved
Protein production. The protein was produced and purified as described previ-
ously3,4,26with a few variations. The GLIC protein was overexpressed in
Escherichia coli C43 cells, and expression was induced with 0.1mM isopropyl
b-D-1-thiogalactopyranoside (IPTG) at absorbance (A600 nm)51 overnight at
20uC. Cells were mechanically lysed in buffer 1 (Tris 20mM pH 7.6, NaCl
300mM, with proteases inhibitors from Roche); membranes were isolated by
ultracentrifugation. The proteins were extracted from membranes with 2%
DDM (Anatrace) under agitation at 4uC, and the solubilized fraction was cleared
by ultracentrifugation. Solubilized proteins were first purified by affinity chro-
matography on an amylose resin. After extensive wash, in buffer 1 supplemented
with 0.1% DDM, the MBP–GLIC fusion protein bound to the resin was cleaved
eluted in buffer 1 supplemented in 0.02% DDM and concentrated. It was then
subjected to size exclusion chromatography on a Superose 6 10/300 GL column
desflurane was then added in the well (typically the well of a Linbro plate was
after the mixture of the protein solution with the mother liquor. An emulsion was
crystallization drops. Crystals were flash-frozen in the usual manner.
A great number of data sets of frozen single crystals were collected on beamlines
Facility and processed with XDS31and CCP4 (ref. 32) programs. Crystals were iso-
unit. Non-crystallographic symmetry (NCS) restraints were used throughout the
matic refinement using REFMAC34and BUSTER35. Difference Fourier Fo2Fc
maps were checked for strong signals indicating the presence of a ligand. In the
two data sets presented here, bound anaesthetics corresponded to 5–9s peaks
(Supplementary Fig. 1), depending on subunits. The peaks were present in each
subunitina region devoidofany densityinother datasets. Moreover,the electron
density for the known bound detergents in the pore4appears below or at this level
(data not shown), namely 6s for the most ordered part of the detergent, which
constitutes an intrinsic positive control for the presence of anaesthetics. At 3.2- to
3.3-A˚resolution itmay not be justified to model the boundmolecules individually
as this will result in a different orientation in each subunit. For propofol we used
fivefold NCS-averaged maps to build the molecules. For desflurane we also used
NCS maps; inaddition,bothplausibleorientations weretriedand theone with the
trifluoromethyl group at the bottom of the cavity was selected by comparing dif-
ference maps after refinement and independent docking scores (data not shown).
Lipidssurrounding the anaestheticsbindingsitewerepartlymodelled,inNCS-
averaged maps. Identification of the chemical nature of the lipids is not possible
with such maps at this resolution and thus we arbitrarily used phosphatidyl-
choline. As in the apo- protein model, the acyl chains are more ordered than
the polar heads. One small part of an acyl chain corresponding to a second layer
of lipid with no direct interaction with the protein is present. Structural analysis
and figure preparation were done with PyMOL36, VMD37and Molprobity38.
Refinement parameters for ligands were generated with the PRODRG server39.
Electrophysiology. Conditions for electrophysiological experiments were as fol-
lows (apart from a few variations6in experiments using desflurane).
Cell injection. Defolliculated, stage VI40X. laevis oocytes were obtained from a
into the nucleus through the animal pole, using a pneumatic microinjector, as a
intranuclear injection. Identified cells were kept in 96-well plates with U-shaped
bottom, in a HEPES-buffered modified Barth’s41solution (in mM: NaCl 88, KCl 1,
NaHCO32.4, HEPES 20, MgSO40.82,Ca(NO3)20.33, CaCl20.41; pH 7.4; 0.22mm
filtered), at 18uC for two days and then at 15uC. Oocytes with the T255A mutant
selected 2days after injection were recorded 2–6days after injection.
Electrophysiological recordings. Oocytes42were superfused with the animal pole
8.0 solution. The whole oocyte plasma membrane was voltage clamped
(GeneClamp 500, Axon Instruments/Molecular Devices) using two intracellular
pipettes filled with 3M KCl (0.8–1.5MV), and distinct current and voltage extra-
cellular electrodes separately bridged to the bath near suction using 5gl21agar in
room temperature (21–23C), acquired at 500Hz after low-pass filtering (200Hz),
and further filtered using 100-to 1-mean sample data reduction for figure display.
Proton concentration–response curves were established at a holding potential of
250mV, using manually controlled 30-s test-pH applications (or shorter when
desensitization/inhibition at low pH produced a peak current) separated by 30- to
60-s wash at pH 7.3, or 8.0 for the T255A GLIC mutant.
tight ground-glass syringes, and vigorously agitated. The final concentration was
spot-checked by headspace gas chromatography. Propofol (2,6 diisopropylphenol)
was obtained from Aldrich (W50,510-2). Propofol stock solution was made by
dissolving pure oily liquid propofol at 1moll21in dimethyl sulphoxide (DMSO).
0.1mM and then to lower concentrations. For most of the recordings, a single
a single application of propofol. The perfusion system was extensively cleaned and
partly replaced before going from high to low propofol concentrations.
Fit of data and statistics. Data in Fig. 3b–e are presented as mean6s.d. Plots
a sigmoid function. Parameters given in Supplementary Table 2, and in Sup-
plementary Table 3 for desflurane, were obtained by fitting for each cell a plot of
Hill-fitting a scatter plot of individual concentration/moll21versus percentage
inhibition data points obtained from all the cells tested (one data point per cell in
most cases); nHand IC50are given with the standard error of the parameters deter-
fitted with a straight line, or a simple exponential decay, to improve readability.
mined in the crystal structures presented in this work. The protonation state was
assigned similar to previous simulations on the basis of pKacalculations with the
Yasara software43to represent the most probable pattern at pH 4.6, with residues
E26, E35, E67, E75, E82, D86, D88, E177 and E243 being protonated. H277 was
sn-glycerol-phosphatidylcholine (POPC) lipid bilayer (307 lipids, approximately
44,000 water molecules) leading to an initial system size of 128A˚3125A˚3182A˚.
The net charge of the system was neutralized with 54 Na1 and 89 Cl2counterions,
achieving a salt concentrationof about 100mM. Thesesteps were performedwithin
were derived for the I202A, V242M and T255A mutants. The simulations were
performed with NAMD44using the CHARMM27 (ref. 45) force field. Parameters
for propofol were provided by G. Brannigan9.
Langevinpistonalgorithmwasusedtomaintainthepressureat 1atm. Ashort10-A˚
for van der Waals interactions between 8.5 and 10A˚. Long-range electrostatic inter-
bonded interactions, and a 4-fs step for long-range electrostatic forces. All bonds
between hydrogen atoms and heavy atoms were constrained with the SHAKE algo-
rithm. All molecular dynamics simulations were performed on Vargas, an IBM
Regatta Power6 machine at the Institut du De ´veloppement et des Ressources en
Informatique Scientifique (IDRIS) Supercomputer Center in Orsay (France).
31. Kabsch, W. Automatic processing of rotation diffraction data from crystals of 21
32. Collaborative Computational Project.. The CCP4 suite: programs for protein
crystallography. Acta Crystallogr. D 50, 760–763 (1994).
33. Emsley, P. & Cowtan, K. Coot: model-building tools for molecular graphics. Acta
Crystallogr. D 60, 2126–2132 (2004).
Macmillan Publishers Limited. All rights reserved
34. Murshudov, G. N., Vagin, A. A. & Dodson, E. J. Refinement of macromolecular
structures by the maximum-likelihood method. Acta Crystallogr. D 53, 240–255
35. Bricogne, G. et al. BUSTER, version 2.8.0. Cambridge, UK: Global Phasing (2009).
36. Schro ¨dinger, L. L. C. The PyMOL molecular graphics system, version 1.3. Æhttp://
37. Humphrey, W., Dalke, A. & Schulten, K. VMD: visual molecular dynamics. J. Mol.
Graph. 14, 33–38 (1996).
and nucleic acids. Nucleic Acids Res. 35, W375–W383 (2007).
39. Schuettelkopf, A. W. & van Aalten, D. M. PRODRG: a tool for high-throughput
crystallography of protein-ligand complexes. Acta Crystallogr. D 60, 1355–1363
40. Dumont, J. N. Oogenesis in Xenopus laevis (Daudin), I. Stages of oocyte
41. Barth, L. G. & Barth, L. J. Differentiation of cells of the Rana pipiens gastrula in
unconditioned medium. J. Embryol. Exp. Morphol. 7, 210–222 (1959).
42. Kusano, K., Miledi, R. & Stinnakre, J. Cholinergic and catecholaminergic
receptors in the Xenopus oocyte membrane. J. Physiol. (Lond.) 328, 143–170
Ewald summation. J. Mol. Graph. Model. 25, 481–486 (2006).
44. Phillips,J.C. et al. Scalable moleculardynamics with NAMD. J.Comput. Chem. 26,
45. MacKerell, A. D. Jr et al. All-atom empirical potential for molecular modeling and
dynamics studies of proteins. J. Phys. Chem. B 102, 3586–3616 (1998).
46. Darden, T., York, D. & Pedersen, L. Particle mesh Ewald: an N log(N) method for
Ewald sums in large systems. J. Chem. Phys. 98, 10089–10092 (1993).
47. Tuckerman, M., Berne, B. J. & Martyna, G. J. Reversible multiple time scale
molecular dynamics. J. Chem. Phys. 97, 1990–2001 (1992).
Macmillan Publishers Limited. All rights reserved