Linking agonist binding to histamine H
1
receptor activation
Aldo Jongejan
1,2
, Martijn Bruysters
1,2
, Juan A Ballesteros
3
, Eric Haaksma
4
, Remko A Bakker
1
, Leonardo Pardo
5
& Rob Leurs
1
G protein–coupled receptors (GPCRs) constitute a large and
functionally diverse family of transmembrane proteins. They
are fundamental in the transfer of extracellular stimuli to
intracellular signaling pathways and are among the most
targeted proteins in drug discovery. The detailed molecular
mechanism for agonist-induced activation of rhodopsin-like
GPCRs has not yet been described. Using a combination
of site-directed mutagenesis and molecular modeling, we
characterized important steps in the activation of the human
histamine H
1
receptor. Both Ser3.36 and Asn7.45 are
important links between histamine binding and previously
proposed conformational changes in helices 6 and 7. Ser3.36
acts as a rotamer toggle switch that, upon agonist binding,
initiates the activation of the receptor through Asn7.45. The
proposed transduction involves specific residues that are
conserved among rhodopsin-like GPCRs.
Rhodopsin-like, family-A G protein–coupled receptors (GPCRs) bind
a diverse set of extracellular ligands, ranging from small neurotrans-
mitters to large hormones. Each subfamily shares the common
heptahelical transmembrane (TM) domain architecture
1,2
but has
developed specific structural motifs to accommodate and respond to
its cognate ligand
3
. Upon agonist binding, the chemical signal is
propagated to selected intracellular amino acids of the TM bundles by
mechanisms that are not yet well characterized at the atomic level.
Despite the structurally diverse types of extracellular ligands, the
conservation in primary structure of the middle and the cytoplasmic
ends of the TM helices of rhodopsin-like GPCRs
4
suggests that signal
propagation occurs by common mechanisms.
Like many aminergic GPCRs
5
, the histamine H
1
receptor (H
1
R), a
typical family-A GPCR, binds its cognate agonist through a conserved
aspartic acid residue, Asp3.32 (numbering according to the Balles-
teros-Weinstein numbering scheme
6
; Methods) in TM 3 (refs. 7–10).
Mutational studies have further identified Lys5.39 (refs. 9–13), Thr5.42
(refs. 7,14), Asn5.46 (refs. 7,9,14,15) and Phe6.55 (ref. 9) in TM 5 and
TM 6 as binding partners for the imidazole ring of histamine. For the
binding of H
1
R antagonists, Trp4.56 (ref. 11) and Phe6.52 (refs. 9,11)
are also implicated. The latter amino acid is part of a cluster of highly
conserved aromatic residues in the top region of TM 6, the Cys-Trp-X-
Pro-Phe-Phe (CWxPFF) motif, which is considered critical in GPCR
activation
5
. Conformational rearrangements of Trp6.48 and Phe6.52
have been associated with structural changes of the proline kink in TM
6 (ref. 16), whereas in the b
2
adrenergic receptor, alteration of the
configuration of this cluster through mutation of Cys6.47 results in
GPCR activation
17
. Conformational changes of Trp6.48 upon GPCR
activation have received direct biophysical support in the structure of
metarhodopsin I, as determined by electron crystallography
18
.Rear-
rangement of the aromatic cluster decreases the proline kink of TM 6,
moving the cytoplasmic end of TM 6 away from TM 3 (ref. 16), and
disrupts the proposed ionic lock between TM 6 (Asp/Glu6.30) and
Arg3.50 of the highly conserved (Asp/Glu)-Arg-Tyr ((DE)RY) motif in
TM 3 (refs. 19,20), aided by the protonation of (Asp/Glu3.49) (refs.
21–23). These large conformational changes of TM 3 and TM 6 are
considered to be an important step in the process of GPCR activation
and have received experimental support from various biophysical and
mutational studies
24–26
. Besides the aromatic cluster in the top region
of TM 6, Asn7.49 of the Asn-Pro-X-X-Tyr (NPxxY) motif at the
bottom of TM 7 is also implicated in GPCR activation
27–29
. In current
models, Asn7.49 is restrained in the inactive g+ conformation, point-
ing toward TM 6 (refs. 27–29). Upon receptor activation, Asn7.49
adopts the t conformation to interact with Asp2.50 in TM 2 and
putatively with Arg3.50 in TM 3 (ref. 28).
Despite detailed insights into the structural changes occurring in
these TM microdomains upon GPCR activation, it remains unclear
how agonist binding in the top region of the TM helices triggers signal
propagation through the TM bundles. An H
1
Rmodelbasedonthe
structural data for the inactive state of rhodopsin (Fig. 1a)showsthat
the highly conserved sequence motifs in TMs 2, 3, 6 and 7 are placed
around the residues predicted to form the histamine binding site
consisting of Asp3.32, Lys5.39, Thr5.42 and Asn5.46 (refs. 7–15).
Although the rhodopsin X-ray structure may have shortcomings
for modeling of the activated state(s) of GPCRs, recent structural
data
18
show that the early phase of rhodopsin activation (metarho-
dopsin I) involves local side chain relocations and no large rigid-body
Published online 19 June 2005; doi:10.1038/nchembio714
1
Leiden/Amsterdam Center for Drug Research, Division of Medicinal Chemistry, Faculty of Science, Vrije Universiteit Amsterdam, De Boelelaan 1083, 1081 HV
Amsterdam, The Netherlands.
2
These authors contributed equally to this work.
3
Novasite Pharmaceuticals Inc., 11095 Flintkote Avenue, San Diego, California 92121,
USA.
4
Dept. of Medicinal Chemistry, Boehringer Ingelheim Austria GmbH, Dr. Boehringergasse 5-11, 1121 Vienna, Austria.
5
Laboratori de Medicina Computacional,
Unitat de Bioestadı´stica and Institut de Neurocie`ncies, Facultat de Medicina, Universitat Auto`noma de Barcelona, 08193 Bellaterra, Spain. Correspondence should be
addressed to R.L. (r.leurs@few.vu.nl).
98 VOLUME 1 NUMBER 2 JULY 2005 NATURE CHEMICAL BIOLOGY
LETTERS

2
0
0
5

N
a
t
u
r
e

P
u
b
l
i
s
h
i
n
g

G
r
o
u
p


h
t
t
p
:
/
/
w
w
w
.
n
a
t
u
r
e
.
c
o
m
/
n
a
t
u
r
e
c
h
e
m
i
c
a
l
b
i
o
l
o
g
y

movements of helices. The subsequent formation of metarhodopsin II
is characterized by substantial conformational changes
30
.
Ab initio geometry optimization of histamine binding to the
agonist-binding pocket in the H
1
R (see Methods) indicated that the
protonated amine of histamine interacts with both Asp3.32 (the
distance between the closest heteroatoms, d,is2.5A
˚
) and Ser3.36
(d¼ 3.1 A
˚
), whereas the imidazole ring of histamine is accommodated
between Lys5.39 (d ¼ 2.9 A
˚
) and Asn5.46 (d ¼ 3.6 A
˚
; Fig. 1b). In
addition, Tyr3.33 and Thr5.42 interact with Lys5.39 and Asn5.46,
respectively. This model predicted that Ser3.36 participates in agonist
binding in the (w
1
g+) rotamer conformation, similar to that described
for serotonin binding to the 5-HT
2A
receptor
31
. The presence of
Ser3.36 within the binding pocket was validated by the substituted
cysteine accessibility method (SCAM)
32
. Although the wild-type H
1
R
was not sensitive to the sulfhydryl-reactive MTSEA, the S3.36C
mutant was accessible to MTSEA, as indicated by the decrease in
[
3
H]mepyramine binding (Fig. 2a). Moreover, an S3.36A mutation
resulted in a decrease in affinity for histamine
by a factor of 12 (Fig. 2b), whereas no
difference in antagonist binding was observed
(Supplementary Table 1 online). These data
therefore strongly suggest that Ser3.36 is pre-
sent in the binding pocket and indeed inter-
acts with histamine.
The H
1
R showed considerable agonist-
independent, constitutive signaling
33
(Fig. 2c), which was strongly reduced by
the inverse agonist mepyramine
33
and was
increased 3.7 7 0.8–fold by histamine. To
our surprise, the S3.36A mutation inhibited
the basal H
1
R signaling (Fig. 2c), indicating
that Ser3.36 has an important stabilizing role
in the active state. Previously, we have
reported on the rotamer preference of serine
residues in a-helical proteins
34
. A statistical
analysis of a-helices in membrane proteins
showed that serine preferred the w
1
¼ g+ over
the g– and t conformation (52%, 20% and
28%, respectively). To probe the Ser3.36
rotamer conformation in the active H
1
R
conformation, we evaluated S3.36C and
S3.36T mutant H
1
Rs. Mutant H
1
Rs were equally expressed in COS-7
cells (750% of wild type) and did not show any important differences
with respect to ligand affinities (Supplementary Table 1 online).
However, both S3.36T and S3.36C mutant H
1
Rs showed a large
increase in constitutive activity (five- and three-fold, respectively;
Fig. 2c). The two mutant H
1
Rs were almost as active as the
histamine-stimulated wild-type H
1
R and could not be further acti-
vated by histamine (Fig. 2c), but were inhibited by the inverse agonist
mepyramine. The S3.36A mutant could still be activated by histamine,
although at much higher (100-fold) concentration (Fig. 2c,d).
In an a-helix, threonine is essentially restricted to the g+ conforma-
tion, although some g– conformation is present because of a steric
clash between the methyl group and the backbone carbonyl in the t
conformation. Therefore, in the S3.36T H
1
R mutant, threonine is
forced to adopt the g+ conformation, which is linked to a strong
constitutive H
1
R activity (Fig. 2c). The elevated constitutive activity of
the S3.36T H
1
R mutant therefore suggests that the g+ conformation
0
25
50
75
100
log[histamine]
B
i
n
d
i
n
g

(
%
)
W
T
S
3
.
3
6
A
S
3
.
3
6
C
S
3
.
3
6
T
M
o
c
k
0
1
2
3
4
5
6
7
8
R
e
l
a
t
i
v
e

a
c
t
i
v
a
t
i
o
n
1.5 × 10
5
1.0 × 10
5
0.5 × 10
5
0
WT
S3.36A
log[histamine]
R
L
U
–6 –5 –4 –3 –2 –1 –8 –5–6–7 –4 –3 –2 –1 –8–9–10 –5–6–7 –4
0
25
50
75
100
125
log[MTSEA]
R
e
s
i
d
u
a
l

b
i
n
d
i
n
g

(
%
)
WT
S3.36A
WT
S3.36A
S3.36C
bcd
Figure 2 Characterization of wild-type (WT) and Ser3.36 mutant H
1
Rs. (a) The inhibition of specific [
3
H]mepyramine binding to COS-7 cells transiently
expressing WT (K), S3.36C (m) or S3.36A (n)H
1
Rs by 5-min incubation with various concentrations of MTSEA. (b) Representative displacement curve
of [
3
H]mepyramine binding to WT (K) and S3.36A (n)mutantH
1
Rs by histamine. (c) Receptor activity of WT and Ser3.36 mutant H
1
Rs as measured by
NF-kB activation; shown are basal activity (black bars) and activity after stimulation with 10

4
M histamine (gray bars) or 10

5
M mepyramine (white bars).
Results are normalized to the basal activity of WT receptors. (d) Representative dose-response curves of histamine at WT (K) and S3.36A (n)mutantH
1
Rs,
as measured by NF-kB activation. In a,c,themean7 s.e.m. of at least three independent experiments is shown, each performed in triplicate; in b,d,the
average and s.e.m. of triplicate measurements are shown.
S3.36
Y3.33
D3.32
N5.46
T5.42
K5.39
TM 5TM 6
TM 4
TM 3
TM 7
TM 2
TM 1
P7.50
D2.50
P6.50
D3.32
R3.50
N5.46
T5.42
K5.39
ab
Figure 1 Modeling of the human H
1
R. (a)FinalmodelofH
1
R obtained by homology modeling
(Methods). The C
a
traces of TMs 2 (golden red), 3 (dark red), 6 (orange) and 7 (blue) are colored. Highly
conserved sequence motifs are shown as sticks for TMs 2 (Asp2.50), 3 (Asp-Arg3.50-Tyr), 6 (Cys-Trp-X-
Pro6.50) and 7 (Asn-Pro7.50-X-X-Tyr), and residues predicted to bind to aminergic ligands are shown as
sticks (carbon atoms green, nitrogen blue, oxygen red and polar protons white). (b) Computational model
of the binding pocket of the H
1
R obtained after ab initio calculations (Methods).
LETTERS
NATURE CHEMICAL BIOLOGY VOLUME 1 NUMBER 2 JULY 2005 99

2
0
0
5

N
a
t
u
r
e

P
u
b
l
i
s
h
i
n
g

G
r
o
u
p


h
t
t
p
:
/
/
w
w
w
.
n
a
t
u
r
e
.
c
o
m
/
n
a
t
u
r
e
c
h
e
m
i
c
a
l
b
i
o
l
o
g
y