Binding of the volatile general anesthetics halothane
and isoflurane to a mammalian b-barrel protein
Jonas S. Johansson
1,2,4
, Gavin A. Manderson
1
, Roberto Ramoni
5
, Stefano Grolli
5
and Roderic G. Eckenhoff
1,3
1 Department of Anesthesia, University of Pennsylvania, Philadelphia, PA, USA
2 Department of Biochemistry and Biophysics, University of Pennsylvania, Philadelphia, PA, USA
3 Department of Physiology, University of Pennsylvania, Philadelphia, PA, USA
4 Johnson Research Foundation, University of Pennsylvania, Philadelphia, PA, USA
5 Dipartimento di Produzioni Animali, Biotecnologie Veterinarie, Qualita
`
e Sicurezza degli Alimenti, Universita
`
di Parma, Parma, Italy
A molecular understanding of volatile anesthetic
mechanisms of action will require structural descrip-
tions of anesthetic–protein complexes. Because the
in vivo sites of action remain to be determined, the
structural features of anesthetic binding sites on
proteins are being explored using well-defined model
systems, such as the serum albumins and four-a-helix
bundle proteins [2,3]. Studies with these model systems
have suggested that volatile general anesthetics prefer-
entially bind to pre-existing appropriately sized

0.43 ± 0.12 mm determined using fluorescence quenching and competitive
binding with 1-aminoanthracene, respectively. Isothermal titration calori-
metry revealed that halothane and isoflurane bound with K
d
values of
80 ± 10 lm and 100 ± 10 lm, respectively. Halothane and isoflurane
binding resulted in an overall stabilization of the folded conformation of
the protein by )0.9 ± 0.1 kcalÆmol
)1
. In addition to indicating specific
binding to the native protein conformation, such stabilization may repre-
sent a fundamental mechanism whereby anesthetics reversibly alter protein
function. Because porcine odorant binding protein has been successfully
analyzed by X-ray diffraction to 2.25 A
˚
resolution [1], this represents an
attractive system for atomic-level structural studies in the presence of
bound anesthetic. Such studies will provide much needed insight into how
volatile anesthetics interact with biological macromolecules.
Abbreviation
AMA, 1-aminoanthracene.
FEBS Journal 272 (2005) 573–581 ª 2004 FEBS 573
packing defects, or cavities, within the protein matrix
[4,5]. In addition, favorable polar interactions with
hydrophobic core side chains can further enhance
anesthetic binding affinity [6,7].
Previous work has demonstrated that volatile anes-
thetics bind to a-helical proteins such as bovine serum
albumin [8–10] and the synthetic four-a-helix bundles
[4,6,7,11,12]. Helical proteins are known to be able to

evaluate the effect of bound halothane and isoflurane
on global protein dynamics, with the goal of further
defining a potential mechanism of volatile general
anesthetic action.
Results
Binding of the volatile anesthetic halothane to
the hydrophobic core of porcine odorant binding
protein
The binding of halothane to the porcine odorant bind-
ing protein hydrophobic core was followed by trypto-
phan fluorescence quenching [10] as shown in Fig. 2.
Halothane causes a concentration-dependent quench-
ing of the intrinsic Trp16 fluorescence, without chan-
ging the emission maximum, indicating that halothane
binding in the cavity does not alter the local dielectric
environment of the indole ring. Furthermore, the lack
of a red-shift in the tryptophan fluorescence emission
maximum upon halothane binding suggests that the
anesthetic does not promote unfolding of the protein,
which would lead to increased water exposure of the
indole ring. Figure 3 (curve a) shows a plot of
the Trp16 fluorescence as a function of the halothane
concentration. Fitting the data using Eqn (1) yields a
K
d
¼ 0.99 ± 0.06 mm with a Q
max
¼ 0.27 ± 0.01,
indicating that the fluorescence of the single trypto-
phan residue in the porcine odorant binding protein is

)1
) because it is comparable to the effect of ha-
lothane on free N-acetyl-tryptophanamide fluorescence
(K
sv
¼ 25±1m
)1
) [7]. This indicates that halothane
does not bind in close proximity to Trp16, but rather in
the vicinity of one of the five tyrosine residues. Halo-
thane is able to quench tyrosine fluorescence with the
same efficiency as tryptophan fluorescence [7]. Of the
five tyrosine residues, Tyr82 is located within the por-
cine odorant binding protein cavity (Fig. 1), and the
fluorescence quenching results in Fig. 3 suggest that this
may be one of the residues that halothane binds to
adjacently. Subtraction of the collisional quenching
contribution to the decrease in Trp16 fluorescence inten-
sity results in curve c in Fig. 3, which yields a K
d
of
0.46 ± 0.10 mm and a Q
max
of 0.17 ± 0.01.
Binding of the volatile anesthetics halothane and
isoflurane to the porcine odorant binding protein
as determined by isothermal titration calorimetry
Representative calorimetric titrations at pH 7.0 of por-
cine odorant binding protein with halothane and iso-
flurane are shown in Figs 4 and 5. Each peak in the

separate samples with error bars representing the SD. For curves
(a) and (c) the lines through the data points has the form of
Eqn (1). Error bars are omitted from curve (c) for clarity.
J. S. Johansson et al. General anesthetic binding to a b-barrel protein
FEBS Journal 272 (2005) 573–581 ª 2004 FEBS 575
phan fluorescence quenching, supporting the validity of
the results. Isoflurane binds to porcine odorant binding
protein with a K
d
¼ 100 ± 10 lm. The other thermo-
dynamic parameters underlying halothane and isoflura-
ne binding to the porcine odorant binding protein are
given in Table 1.
Halothane displaces 1-aminoanthracene (AMA)
bound to the internal cavity in the hydrophobic
core of porcine odorant binding protein
Figure 6 shows that halothane can displace AMA from
the porcine odorant binding protein cavity. The
competition curve was treated as a two parameter
hyperbolic decay (R ¼ 0.97) and gave an EC
50
of
0.86 ± 0.24 mm. The true dissociation constant ( K
d,
true
), calculated using Eqn (2) resulted in a value of
0.43 ± 0.12 mm, in agreement with the results
obtained using Trp16 fluorescence quenching and iso-
thermal titration calorimetry.
Effect of bound halothane and isoflurane on the

flurane, showing the calorimetric response as successive injections
of ligand are added to the reaction cell. The lower panel depicts the
binding isotherm of the calorimetric titration shown in the upper
panel. The continuous line represents the least-squares fit of the
data to a single-site binding model.
Table 1. Dissociation constants and thermodynamic data for the
binding of halothane and isoflurane to the porcine odorant binding
protein. The entropy unit (eu) is calÆmol
)1
Æ K
)1
.
Anesthetic K
d
(lM) DG° (kcalÆmol
)1
) DH° (kcalÆmol
)1
) DS° (eu)
Halothane 80 ± 10 )5.5 ± 0.1 )1.4 ± 0.1 14.0
Isoflurane 100 ± 10 )5.4 ± 0.1 )2.4 ± 0.1 10.3
Fig. 6. Competition between halothane and 1-aminoanthracene
(AMA) for binding to porcine odorant binding protein. The fluores-
cence intensity at 480 nm (a measure of bound AMA) is plotted as
a function of the halothane concentration. See text for details.
General anesthetic binding to a b-barrel protein J. S. Johansson et al.
576 FEBS Journal 272 (2005) 573–581 ª 2004 FEBS
calorimetry indicates that halothane binds to porcine
odorant binding protein with a K
d

with an average K
d
value of 0.36 ± 0.03 mm using
19
F-NMR spectroscopy and gas chromatography [32].
Finally, isoflurane was shown to bind to the four-
a-helix bundle (Aa
2
-L38M)
2
with a K
d
¼ 140 ± 10 lm
using isothermal titration calorimetry [12]. The affinity
of the interaction of isoflurane with porcine odorant
binding protein (K
d
¼ 100 ± 10 lm) is therefore com-
parable to the findings in the latter two studies. For
both anesthetics, the free energy of binding (D G °)
exceeded the heat of binding (DH°) by more than a
factor of two (Table 1), indicating that binding to por-
cine odorant binding protein is entropy driven, in con-
trast to the results obtained with the four-a-helix
bundle (Aa
2
-L38M)
2
[12].
AMA has been shown to bind to porcine odorant

a-helix bundle scaffold stability was examined using
chemical denaturation with guanidinium chloride as
shown by circular dichroism spectroscopy [27]. The
bound anesthetic stabilized the native bundle confor-
mation by )1.8 kcalÆmol
)1
at 25 °C, and increased the
m-value (the slope of the unfolding transition) from
1.6 ± 0.2 to 2.0 ± 0.1 kcalÆmol
)1
Æ m
)1
. The latter
effect is compatible with improved hydrophobic core
packing [36], and supports anesthetic binding to the
cavity in the core of (Aa
2
)
2
. Using hydrogen exchange
[6], halothane was also shown to stabilize the folded
conformation of the four-a-helix bundle (Aa
2
-L38M)
2
by approximately )0.9 kcal Æmol
)1
. Thus, binding of
anesthetic to the four-a-helix bundle scaffolds is associ-
ated with a stabilization of the folded conformation of

[2]. No high-resolution structure that involves any of
the modern halogenated ether anesthetics has yet been
published. However, a 2.4 A
˚
resolution X-ray crystal
structure of human serum albumin with several bound
halothane molecules has recently been reported [38].
Six of the binding sites involve a combination of ali-
phatic and charged residues, such as arginine or lysine,
with the remaining two composed of aliphatic and
somewhat polar residues such as serine, phenylalanine,
and asparagine. The crystallographic results are in
accord with earlier solution studies using fluorescence
spectroscopy and photoaffinity labeling that indicated
that halothane bound in close proximity to Trp214 and
Tyr411 in human serum albumin [10,28].
Because porcine odorant binding protein has been
successfully crystallized and analyzed by X-ray diffrac-
tion to 2.25 A
˚
resolution [1], the current results suggest
that it represents an attractive system for atomic-level
structural studies in the presence of bound anesthetic.
Such studies will provide much needed insight into
how volatile anesthetics interact with biological macro-
molecules, and will provide guidelines regarding the
general architecture of binding sites on central nervous
system proteins.
Experimental procedures
Protein purification

As described previously [10], the quenched fluorescence
(Q) is a function of the maximum possible quenching
(Q
max
) at an infinite halothane concentration ([Halothane])
and the affinity of the anesthetic for its binding site (K
d
)in
the vicinity of the tryptophan residue. From mass law con-
siderations, it then follows that
Q ¼
ðQ
max
[Halothane]Þ
ðK
d
þ [Halothane]Þ
ð1Þ
Halothane displacement of bound AMA
The dissociation constant of the complex between halot-
hane and porcine odorant binding protein was determined
using a competitive binding assay with the fluorescent lig-
and AMA [21,22]. The approach has previously been
employed for the determination of the dissociation con-
stants for other ligands [23] shown crystallographically to
occupy the internal cavity of the protein [24]. Briefly, por-
cine odorant binding protein samples (1 lm), containing a
fixed amount of AMA (1 lm), were incubated overnight at
4 °C in the presence of increasing concentrations of halot-
hane in 20 mm Tris ⁄ HCl buffer, pH 7.8. The displacement

d,AMA
of the AMA–porcine odorant binding protein
complex (1 lm).
General anesthetic binding to a b-barrel protein J. S. Johansson et al.
578 FEBS Journal 272 (2005) 573–581 ª 2004 FEBS
The stock solution of halothane contains the stabilizing
agent thymol, which can also bind to porcine odorant bind-
ing protein. However, control experiments showed that thy-
mol alone, at the concentrations present in the experiments
(< 0.0001% or < 5 pm), was unable to displace AMA
from the porcine odorant binding protein. In addition,
halothane (at concentrations less than 200 mm) does not
directly quench AMA fluorescence.
Isothermal titration calorimetry
Isothermal titration calorimetry was performed using a
MicroCal VP-ITC titration microcalorimeter (Northamp-
ton, MA, USA) at 20 °C. Porcine odorant binding protein
at a concentration of 87 lm in 130 mm NaCl, 20 mm
sodium phosphate, pH 7.0, was placed in the 1.4 mL calori-
meter cell, and anesthetic (5 mm in 130 mm NaCl, 20 mm
sodium phosphate, pH 7.0) was added sequentially in
10 lL aliquots (for a total of 29 injections) at 5 min inter-
vals. The heat of reaction per injection (microcalories per
second) was determined by integration of the peak areas
using the origin v5.0 software (http://www.microcal.com/).
This software provides the best-fit values for the heat of
binding (DH°), the stoichiometry of binding (n), and the
association constant (K
a
) from plots of the heat evolved per

K
a
and DH° values to be minimized. Allowing all three var-
iables to float simultaneously during the curve-fitting proce-
dure may be associated with more variable results because
of the potential for multiple minima [26].
Hydrogen exchange
Porcine odorant binding protein (3–5 mg) was dissolved in
1mLof1m guanidinium chloride and 50 mm sodium phos-
phate, pH 8.5, with 40 lL
3
HOH added (100 mCiÆmL
)1
,
ICN, Costa Mesa, CA, USA), and allowed to equilibrate
overnight at 20 °C to permit complete exchange-in of tritium.
The porcine odorant binding protein solutions were then
passed through a PD-10 gel filtration column (Sigma Chem-
ical Co, St Louis, MO, USA) to remove free
3
HOH, and to
switch to the exchange-out buffer (50 mm sodium phosphate,
pH 7.0). The protein fraction was collected and immediately
placed in gas-tight Hamilton syringes prefilled with
exchange-out buffer, with or without 4.0 mm halothane or
7.0 mm isoflurane. The syringe contents were mixed with
microstir bars, and 100 lL aliquots were precipitated with
2 mL 20% trichloroacetic acid at regular intervals, immedi-
ately filtered through Whatman (Hillsboro, OR, USA) GF ⁄ F
filters, and washed with 8 mL 2% trichloroacetic acid. Filters

7918.
J. S. Johansson et al. General anesthetic binding to a b-barrel protein
FEBS Journal 272 (2005) 573–581 ª 2004 FEBS 579
2 Eckenhoff RG & Johansson JS (1997) Molecular inter-
actions between inhaled anesthetics and proteins. Phar-
macol Rev 49, 343–367.
3 Johansson JS & Eckenhoff RG (2002) Protein models.
In Neural Mechanisms of Anesthesia (Antognini, J, Car-
stens, E & Raines, D, eds), pp. 395–412. Humana Press,
Totowa, NJ, USA.
4 Johansson JS, Gibney BR, Rabanal F, Reddy KS &
Dutton PL (1998) A designed cavity in the hydrophobic
core of a four-a-helix bundle improves volatile anes-
thetic binding affinity. Biochemistry 37, 1421–1429.
5 Manderson GA, Michalsky SJ & Johansson JS (2003)
Effect of four-a-helix bundle cavity size on volatile anes-
thetic binding energetics. Biochemistry 42, 11203–11213.
6 Johansson JS, Scharf D, Davies LA, Reddy KS &
Eckenhoff RG (2000) A designed four-a-helix bundle
that binds the volatile general anesthetic halothane with
high affinity. Biophys J 78, 982–993.
7 Manderson GA & Johansson JS (2002) The role of
aromatic side chains in the binding of volatile general
anesthetics to a four-a-helix bundle. Biochemistry 41,
4080–4087.
8 Dubois BW, Cherian SF & Evers AS (1993) Volatile
anesthetics compete for common binding sites on bovine
serum albumin: a
19
F-NMR study. Proc Natl Acad Sci

labeling with a neuroactive steroid analog. 6-Azi-preg-
nanolone labels voltage-dependent anion channel-1 in
rat brain. J Biol Chem 278, 13196–13206.
18 Xi J, Liu R, Asbury GR, Eckenhoff MF & Eckenhoff
RG (2004) Inhalational anesthetic binding proteins in
rat neuronal membranes. J Biol Chem 279, 19628–
19633.
19 Eckenhoff RG (1998) Do specific or nonspecific inter-
actions with proteins underlie inhalational anesthetic
action. Mol Pharmacol 54, 610–615.
20 Burova TV, Choiset Y, Jankowski CK & Haertle
´
T
(1999) Conformational stability and binding properties
of porcine odorant binding protein. Biochemistry 38,
15043–15051.
21 Paolini S, Tanfani F, Fini C, Bertoli E & Pelosi P
(1999) Porcine odorant-binding protein: Structural sta-
bility and ligand affinities measured by Fourier-trans-
form infrared spectroscopy and fluorescence
spectroscopy. Biochim Biophys Acta 1431, 179–188.
22 Ramoni R, Vincent F, Ashcroft AE, Accornero P,
Grolli S, Valencia C, Tegoni M & Cambillau C (2002)
Control of domain swapping in bovine odorant-binding
protein. Biochem J 365, 739–748.
23 Vincent F, Ramoni R, Spinelli S, Grolli S, Tegoni M &
Cambillau C (2004) Crystal structures of bovine odor-
ant-binding protein in complex with odorant molecules.
Eur J Biochem 271, 3832–3842.
24 Vincent F, Spinelli S, Ramoni R, Grolli S, Pelosi P,

nuclear magnetic resonance investigation of stereoselec-
tive binding of isoflurane to bovine serum albumin.
Biophys J 70, 532–538.
General anesthetic binding to a b-barrel protein J. S. Johansson et al.
580 FEBS Journal 272 (2005) 573–581 ª 2004 FEBS
31 Johansson JS, Zou H & Tanner JT (1999) Bound vola-
tile general anesthetics alter both local protein dynamics
and global protein stability. Anesthesiology 90, 235–245.
32 Xu Y, Seto T, Tang P & Firestone L (2000) NMR
study of volatile anesthetic binding to nicotinic acetyl-
choline receptors. Biophys J 78, 746–751.
33 Franks NP & Lieb WR (1984) Do general anaesthetics
act by competitive binding to specific receptors? Nature
310, 599–601.
34 Ishizawa Y, Pidkiti R, Liebman PA & Eckenhoff RG
(2002) G protein-coupled receptors as direct targets of
inhaled anesthetics. Mol Pharmacol 61, 945–952.
35 Fersht AR (1993) Protein folding and stability – The
pathway of folding of barnase. FEBS Lett 325, 5–15.
36 Pace CN (1986) Determination and analysis of urea and
guanidine hydrochloride denaturation curves. Methods
Enzymol 131, 266–280.
37 Eckenhoff RG & Tanner JW (1998) Differential
halothane binding and effects on serum albumin and
myoglobin. Biophys J 75, 477–483.
38 Bhattacharya AA, Curry S & Franks NP (2000) Bind-
ing of the general anesthetics propofol and halothane
to human serum albumin. J Biol Chem 275, 38731–
38738.
J. S. Johansson et al. General anesthetic binding to a b-barrel protein