Identification of a novel 2-pyridyl-benzensulfonamide derivative, RQ-00203078, as a selective and orally active TRPM8 antagonist

Article abstract of DOI:10.1016/j.bmcl.2014.10.074

A novel series of 2-pyridyl-benzensulfonamide derivatives have been identified as selective and orally active TRPM8 antagonists via high throughput screening (HTS). Exploration of the structure-activity relationships of compound 1 has led to the identification of RQ-00203078 (compound 36) as a highly selective, potent and orally available TRPM8 antagonist. RQ-00203078 demonstrated excellent in vivo activity in a dose dependent manner with an ED₅₀ value of 0.65 mg/kg in the icilin-induced wet-dog shakes model in rats after oral administration and may become an important pharmacological tool for fully assessing the potential therapeutic use of the targets activated by cold stimulation.

Full text of DOI:10.1016/j.bmcl.2014.10.074

Bioorganic & Medicinal Chemistry Letters 24 (2014) 5364–5368
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Identification of a novel 2-pyridyl-benzensulfonamide derivative,
RQ-00203078, as a selective and orally active TRPM8 antagonist
⇑
Masashi Ohmi , Yuji Shishido, Tadashi Inoue, Kazuo Ando, Akiyoshi Fujiuchi, Akiko Yamada,
⇑
Shuzo Watanabe, Kiyoshi Kawamura
Research and Development, RaQualia Pharma Inc., 5-2 Taketoyo, Aichi 470-2341, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
A novel series of 2-pyridyl-benzensulfonamide derivatives have been identified as selective and orally
active TRPM8 antagonists via high throughput screening (HTS). Exploration of the structure–activity rela-
tionships of compound 1 has led to the identification of RQ-00203078 (compound 36) as a highly selec-
tive, potent and orally available TRPM8 antagonist.
RQ-00203078 demonstrated excellent in vivo activity in a dose dependent manner with an ED50 value
of 0.65 mg/kg in the icilin-induced wet-dog shakes model in rats after oral administration and may
become an important pharmacological tool for fully assessing the potential therapeutic use of the targets
activated by cold stimulation.
Received 4 September 2014
Revised 20 October 2014
Accepted 23 October 2014
Available online 29 October 2014
Keywords:
TRPM8 antagonist
Sulfonamide
RQ-00203078
Wet-dog shakes
Cold allodynia
Ó 2014 Elsevier Ltd. All rights reserved.
The transient receptor potential melastatin 8 (TRPM8) is a
member of the TRP melastatin sub-group (TRPM1–8), comprising
of seven groups; TRP canonical (TRPC), TRP vanilloid (TRPV), TRP
melastatin (TRPM), TRP polycystin (TRPP), TRP mucolipin (TRPML),
TRP ankyrin (TRPA) and TRP no mechanoreceptor potential C
(TRPN) within the TRP superfamily channels. TRPM8 is a calcium
(Ca2+)-permeable, nonselective cation channel, expressed on pri-
mary sensory neurons (A-delta and C-fibers in the dorsal root gan-
glia), and is activated by cold temperature in both the innocuous
(15–28 °C) and noxious (<15 °C) range, and by chemical agonists
such as menthol and icilin.1 TRPM8 channel is known to play a role
in cold hyperalgesia and cold allodynia induced by disease condi-
tions such as chemotherapy-induced peripheral neuropathy (CIPN,
using the agents oxaliplatin, paclitaxel, or vincristine), diabetic
neuropathy, migraine and overactive bladder (OAB).2 A number
of small molecule TRPM8 antagonists3 have been disclosed and
evaluated in menthol-induced calcium influx assays and the ici-
lin-induced wet-dog shakes (WDS) model (Fig. 1). Studies with
these antagonists have confirmed the role of TRPM8 antagonist
in various preclinical animal models of cold allodynia and hyper-
sensitivity. Recently, Pfizer has reported that PF-05105679, a selec-
tive TRPM8 antagonist, is analgesic in an experimental model of
cold pain in humans.3f Therefore, small molecule antagonists of
the TRPM8 channel are expected to provide important pharmaco-
logical tools for fully assessing the therapeutic potential of block-
ing the TRPM8 channel.
An HTS campaign with in-house library compounds was imple-
mented in order to identify hit compounds. Among the hits, our
group focused on sulfonamide 1, shown in Figure 2. Potential
issues associated with compound 1 were a moderate intrinsic
⇑
Corresponding authors at present address: Discovery Research, RaQualia
Pharma Inc., 5-2 Taketoyo, Aichi 470-2341, Japan.
E-mail addresses: masashi.ohmi@raqualia.com (M. Ohmi), kiyoshi.kawamura@
raqualia.com (K. Kawamura).
Figure 1. Reported TRPM8 antagonists.
http://dx.doi.org/10.1016/j.bmcl.2014.10.074
0960-894X/Ó 2014 Elsevier Ltd. All rights reserved.

M. Ohmi et al. / Bioorg. Med. Chem. Lett. 24 (2014) 5364–5368
5365
Figure 2. The hit compound in HTS and strategies for improving on several issues.
Scheme 1. Reagents and conditions: (a) NaH then sulfonyl chloride, THF, 0 °C—rt; (b) Cs2CO3, NaI, DMF, 80 °C; (c) NaOH, H2O, THF, 50 °C; (d) RR0NH, EDCI, HOBt, Et3N, CH2Cl2,
rt; (e) Et3N, CH2Cl2, 0 °C—rt; (f) Cs2CO3, MeCN, microwave irradiation, 160 °C, 10 min; (g) LiAlH4, THF, rt.
clearance (Clint = 28.7 ml/min/kg) in human liver microsomes
Table 1
(HLM) and poor equilibrium solubility (<0.8 l
M).4 However, we
Initial SAR of phenylsulfonylamide core (A-region)
were aware that similar TRPM8 antagonists of the sulfonamide
derivatives with a carboxylic acid group had been disclosed inde-
pendently by the Janssen group (e.g., compd 306). While the intro-
duction of a carboxylic acid group generally increases albumin
binding,5 it also tends to improve the stability in liver microsomes
and equilibrium solubility.6 This report summarizes our structure
activity relationship (SAR) pursuit from the original hit compound
1 to the more advanced compound 36, focusing on the improve-
ments in liver microsome stability and equilibrium solubility
through the introduction of a carboxylic acid group.
The sulfonamide derivatives 6 were prepared through the
sequences A or B as shown in Scheme 1.7 The sulfonamides 4 were
obtained by treatment of the commercially available aniline deriv-
atives 3 with sodium hydride and reaction with the corresponding
benzenesufonyl chloride derivatives. The sulfonamides 4 were
then converted to the carboxylic acid derivatives 6 by N-alkylation
with an alkyl halide followed by hydrolysis of the ester derivatives
5 (route A). In route B, the sulfonamides 9, prepared from commer-
cially available amines 8 and sulfonyl chlorides, were converted to
the carboxylic acid derivatives 6 by SNAr reaction with the aryl
halide followed by hydrolysis of the ester derivatives 5. The amide
and alcohol derivatives 7 and 10 were prepared by amidation using
the coupling reagent (EDCI) and reduction with lithium aluminum
hydride (LiAlH4), respectively.
Compd R3
hTRPM8
HLM Clint
Equilibrium
IC50 (nM)a (ml/min/kg) Solubility (
l
M)b
1
H
183
15,974
406
61
28.7
N.D.c
<7
<0.8
N.D.
285
185
35
<2.2
20
37
11
2-COOH
12
3-COOH
13
4-COOH
<7
14
4-CH2OH
4-CONH2
4-CONHMe
4-CONMe2
4-CONHCH2CH2OH
42
70
57
67
>168
27.2
85.4
>168
77.7
15a
15b
15c
15d
216
123
IC50 values based on inhibition of menthol (30 l
M) induced Ca2+ influx in
a
HEK293 cells.
Equilibrium solubility measured in a high throughput automated method.4
N.D. = Not determined.
b
c
times as potent as compound 1, while compounds 11 and 12 bear-
ing a 2- or 3-carboxylic acid group were less potent than the
nonsubstituted compound 1, as shown in Table 1. As expected,
12 and 13 exhibited improvements both in liver microsome
As a result of the introduction of a carboxylic acid group on the
A-region, compound 13 bearing a 4-carboxylic acid group was 3

5366
M. Ohmi et al. / Bioorg. Med. Chem. Lett. 24 (2014) 5364–5368
Table 2
stability and in equilibrium solubility. The amide derivatives (15a–
15c) also had IC50 values less than 100 nM in the human TRPM8
calcium influx assays, but the metabolic stability and solubility of
these compounds were not improved compared to compound 13.
Similarly, the benzyl alcohol derivative 14, whilst having good
potency with an IC50 value of 42 nM, did not show improvements
in metabolic stability and solubility. Based on the SAR on the A-
region of the molecule (Fig. 2), we selected compound 13 as a lead
compound, and shifted our efforts to the SAR exploration on the
pyridine (B-region) and benzyl (C-region) moieties.
Initial SAR of the pyridine core (B-region)
Compd
R1
R2
hTRPM8
HLM Clint
(ml/min/kg)
IC50 (nM)a
In order to better understand the role of each region and substi-
tuent as pharmacophores, molecular overlay analysis was con-
ducted using the compound 13 and compd 306. As one can
imagine from the similarity in the two-dimensional structures of
the molecules, minimized conformers of the two molecules were
easily superimposed (Fig. 3). The trifluoromethyl group in com-
pound 13 overlapped well with the benzene moiety of the benzo-
thiophene ring in compd 306, indicating that these two moieties
behave as hydrophobic groups in the pharmacophore. Further-
more, the molecular overlay suggested that there would be space
for accommodating substituents on the benzyl moiety of com-
pound 13. With this hypothesis in mind, we set out to explore
the SAR in the B- and C-regions.
13
16
17
18
19
20
21
22
23
24
25
CF3
H
Cl
Cl
Cl
Cl
Cl
Cl
Cl
F
Me
CN
H
61
9859
201
23,782
707
122
213
736
<7
N.D.
<7
N.D.
>168
85
b
CN
Me
Ph
CF3
CF3
CF3
CF3
Cl
<7
N.D.
N.D.
N.D.
N.D.
2472
2983
1562
CF3
a
IC50 values based on inhibition of menthol (30
HEK293 cells.
l
M) induced Ca2+ influx in
b
N.D. = Not determined.
As shown in Table 2, the substitution effect of R1 on the pyridine
ring was significant when R2 was held constant as chlorine.
Lipophilic atoms or groups such as chloro (17) and phenyl (20)
were associated with strong antagonistic activity, while the hydro-
gen (16) and cyano (18) substituents reduced potency by 160- and
400-fold, respectively, compared with compound 13 bearing a
trifluoromethyl group. Methyl derivative 19 showed moderate
activity, with a 10-fold drop in potency compared to compound
13. These results suggested that lipophilicity, rather than electro-
negativity, plays a key role in the antagonistic activity on TRPM8,
which is in good accordance with the pharmacophore hypothesis
suggested by the molecular overlay discussed above.
Table 3
Initial SAR of benzyl core (C-region)
Compd
R5
hTRPM8
HLM Clint
(ml/min/kg)
IC50 (nM)a
The effect of R2 on the pyridine ring was then investigated with
R1 group being held constant as a trifluoromethyl group. The fluoro
21 and methyl 22 derivatives were 4- and 10-fold less potent,
respectively, than the chloro-substituted compound 13. R2 groups
such as cyano 23 and hydrogen 24 led to a drop in potency, with a
40-fold increase of the IC50 compared to compound 13. Compound
25, where the R1 and R2 groups were switched compared to com-
pound 13, had moderate activity, which suggested that there
may be some restriction to the amount of lipophilicity and steric
bulk that can be introduced at the R2 position.
13
26
61
<7
762
N.D.b
27
28
29
49
104
81
>168
<7
8419
Taken altogether, we concluded that the combination of
R1 = CF3 or Cl and R2 = Cl are optimum substituents for the
B-region, and shifted to the SAR exploration of the benzyl moiety
(C-region).
30
3720
<7
a
IC50 values based on inhibition of menthol (30
HEK293 cells.
l
M) induced Ca2+ influx in
b
N.D. = Not determined.
In an initial SAR study of the R5 substituent, the phenylpropyl
and cyclohexylmethyl derivatives (27 and 28) proved similar to
or more potent than the benzyl derivative 13 in the in vitro assay
of human calcium influx, but the phenethyl derivative 26 displayed
a 10-fold drop in potency (Table 3). Additionally, the metabolic sta-
bility of these compounds decreased, as shown by a high clearance
in HLM. Introduction of heteroaryl and heterocyclic moieties such
as pyridyl and morpholyl (29 and 30) also resulted in a moderate
activity, despite the improved metabolic stability. Therefore, as
for the optimization of group, we selected the benzyl scaffold for
R.5
Finally, we proceeded to optimize the substitution on the ben-
zyl moiety (C-region), focusing on the in vitro antagonistic activity,
Figure 3. Molecular overlay of 13 (brown) and compd 306 (green).8

M. Ohmi et al. / Bioorg. Med. Chem. Lett. 24 (2014) 5364–5368
5367
Table 4
SAR optimization of benzyl core (C-region)
ED50: 0.65 mg/kg
Compd
R1
R6
hTRPM8
HLM Clint
(ml/min/kg)
IC50 (nM)a
Compd 306
—
CF3
—
Ph–
6.2
61
27
41
8.9
96
57
8.3
40
14
0.96
8.5
<7
<7
<7
<7
<7
<7
<7
<7
13
31
32
33
34
35
36
37
38
39
2-CF3APh–
3-CF3APh–
4-CF3APh–
2-CF3OAPh–
3-CF3OAPh–
4-CF3OAPh–
4-FAPh–
Figure 4. Dose dependent inhibitory effect on wet-dog shakes of compound 36
(RQ-00203078) Compounds 36 (RQ-00203078) and vehicle (0.5% methyl cellulose
(MC)) were administered to fasted rats 1 h before icilin injection. After intraperi-
toneal injection of icilin (1 mg/kg), the number of wet-dog shakes (WDS) were
counted over 30 min. The number of animals per group was 6.
3-CF3-4-FAPh–
4-tBuAPh–
<7
>168
40
41
42
43
3.3
8.0
26.7
>168
<7
Table 6
In vitro activities for TRP channels of compound 36 (RQ-00203078)
F3C
Cl
N
N
O S O
OCF3
67
COOH
18,216
N.D.b
Compound
36
hTRPM8 (menthol)
rTRPM8 (menthol)
h/rTRPA1a
8.3 nM
5.8 nM
>10 lM/>10 lM
44
45
46
46
15
18
53
hTRPV4b
>3
lM
h/rTRPV1c
>10
lM/>3
lM
7.5
8.6
a
IC50 values based on inhibition of allyl isothiocyanate (AITC) (100
l
M) induced
Ca2+ influx in HEK293 cells.
b
IC50 values based on inhibition of hypotonic solution (90 mMD-manni-
tol + 1.26 mM CaCl2) induced Ca2+ influx in HEK293 cells.
IC50 values based on inhibition of capsaicin (h/r:600 nM/300 nM) induced Ca2+
c
47
48
Cl
3-CF3-4-FAPh–
4-CF3APh–
16
11
<7
<7
influx in HEK293 cells.
a
IC50 values based on inhibition of menthol (30
HEK293 cells.
l
M) induced Ca2+ influx in
demonstrated excellent potency with an IC50 value of 0.96 nM, a
60-fold improvement in antagonistic activity compared to that of
compound 13 (IC50 61 nM), but its stability in HLM was much
decreased with an intrinsic Clint of more than 168 ml/min/kg.
Accordingly, we decided to replace the tert-butyl group by trifluo-
romethyl-isopropyl and methylcyclopropyl moieties,9 which are
considered to be bioisosteres of the tert-butyl group. Compounds
40 and 41 also showed strong activity with IC50 values of 3.3 nM
and 8.0 nM, respectively, but the metabolic stability of the com-
pounds was still unsatisfactory with HLM clearances of 26.7 ml/
min/kg and >168 ml/min/kg, respectively. Bicyclic fused ring sys-
tems such as indazole and benzofuran derivatives (44–46) demon-
strated a 1.5- to 4-fold improvement in potency, compared to the
compound 13, with good to moderate metabolic stability. Further-
more, the derivatives 47 and 48, where the R1 substituent is now
chlorine, also showed a dramatic increase of activity by the intro-
duction of trifluoromethyl group at the 4-position of the C-region,
compared to the corresponding analog 17 (IC50 = 201 nM). Having
thus optimized in vitro activity, liver microsome stability and equi-
librium aqueous solubility, we selected compounds 36, 37, 46 and
48 for in vivo evaluation in the wet-dog shakes (WDS) model in
rats.
b
N.D. = Not determined.
Table 5
Rat wet-dog shakes assaya
Compd
ED50 (mg/kg)
rTRPM8
IC50 (nM)
RLM Clint
(ml/min/kg)
rPPB
(%fu)b
36
37
46
48
0.65
7.5
8.8
5.8
43
28
<10.4
<10.4
41.1
0.31
1.4
1.2
1.3
9.0
<10.4
1.1
a
Compounds 36, 48 (0.3, 1, 3 and 10 mg/kg), compounds 37, 46 (1, 3 and 10 mg/
kg) and vehicle (0.5% methyl cellulose (MC)) were administered to fasted rats 1 h
before icilin injection. After intraperitoneal injection of icilin (1 mg/kg), the number
of wet-dog shakes (WDS) was counted over 30 min. The number of animals per
group was 6.
b
Rat plasma protein binding (percent fraction unbound).
hoping to achieve a drastic improvement of potency by introducing
bulky lipophilic groups at the 4-position on the benzene ring
(Table 4). In the event, compound 39 bearing a 4-tert-butyl group

5368
M. Ohmi et al. / Bioorg. Med. Chem. Lett. 24 (2014) 5364–5368
2. (a) Fernández-Peña, C.; Viana, F. Open Pain J. 2013, 6, 154; (b) Chaudhari, S. S.;
The 4 compounds were evaluated in the icilin-induced wet-dog
Kadam, A. B.; Khairatkar-Joshi, N.; Mukhopadhyay, I.; Karnik, P. V.; Raghuram,
A.; Rao, S. S.; Vaiyapuri, T. S.; Wale, D. P.; Bhosale, V. M.; Gudi, G. S.; Sangana, R.
R.; Thomas, A. Bioorg. Med. Chem. 2013, 21, 6542; (c) Lashinger, E. S. R.;
Steiginga, M. S.; Hieble, J. P.; Leon, L. A.; Gardner, S. D.; Nagilla, R.; Davenport, E.
A.; Hoffman, B. E.; Laping, N. J.; Su, X. Am. J. Physiol. Renal Physiol. 2008, 295,
F803.
shakes model in rats (Table 5),10 and all 4 compounds attenuated
the shaking behavior with ED50’s of less than 10 mg/kg (p.o.), thus
demonstrating that these sulfonamide derivatives are orally active
TRPM8 antagonists. In particular, compound 36 showed significant
efficacy with an ED50 of 0.65 mg/kg in a dose dependent manner
(Fig. 4). In addition, excellent oral exposure of compound 36 was
confirmed independently in rat PK studies at 3 mg/kg (p.o.) admin-
istration, with a Cmax value of 2300 ng/mL and 86% bioavailability.
Taken altogether, compound 36 was selected as an advanced com-
pound with the code name RQ-00203078.
3. (a) Branum, S. T.; Colburn, R. W.; Dax, S. L.; Flores, C. M.; Jetter, M. C.; Liu, Y.;
Ludovici, D.; Macielag, M. J.; Matthews, J. M.; McNally, J. J.; Reaney, L. M.;
Russell, R. K.; Qin, N.; Wells, K. M.; Youells, S. C.; Youngman, M. A.
WO2009012430, 2009.; (b) Tamayo, N. A.; Bo, Y.; Gore, V.; Ma, V.;
Nishimura, N.; Tang, P.; Deng, H.; Klionsky, L.; Lehto, S. G.; Wang, W.;
Youngblood, B.; Chen, J.; Correll, T. L.; Bartberger, M. D.; Gavva, N. R.; Norman,
M. H. J. Med. Chem. 2012, 55, 1593; (c) Horne, D. B.; Tamayo, N. A.; Bartberger,
M. D.; Bo, Y.; Clarine, J.; Davis, C. D.; Gore, V. K.; Kaller, M. R.; Lehto, S. G.; Ma, V.
V.; Nishimura, N.; Nguyen, T. T.; Tang, P.; Wang, W.; Youngblood, B. D.; Zhang,
M.; Gavva, N. R.; Monenschein, H.; Norman, M. H. J. Med. Chem. 2014, 57, 2989;
(d) Parks, D. J.; Parsons, W. H.; Colburn, R. W.; Meegalla, S. K.; Ballentine, S. K.;
Illig, C. R.; Qin, N.; Liu, Y.; Hutchinson, T. L.; Lubin, M. L.; Stone, D. J.; Baker, J. F.;
Schneider, C. R.; Ma, J.; Damiano, B. P.; Flores, C. M.; Player, M. R. J. Med. Chem.
2011, 54, 233; (e) Calvo, R. R.; Meegalla, S. K.; Parks, D. J.; Parsons, W. H.;
Ballentine, S. K.; Lubin, M. L.; Schneider, C.; Colburn, R. W.; Flores, C. M.; Player,
M. R. Bioorg. Med. Chem. Lett. 1903, 2012, 22; (f) Winchester, W.; Gore, K.; Glatt,
S.; Petit, W.; Gardiner, J. C.; Conlon, K.; Postlethwaite, M.; Saintot, P. P.; Roberts,
S.; Gosset, J. R.; Matsuura, T.; Andrews, M. D.; Glossop, P. A.; Palmer, M. J.; Clear,
N.; Collins, S.; Beaumont, K.; Reynolds, D. S. J. Pharmacol. Exp. Ther. 2014, 351,
259; (g) DeFalco, J.; Duncton, M. A. J.; Emerling, D. Curr. Top. Med. Chem. 2011,
11, 2237.
Finally, compound 36 was highly selective over other TRP chan-
nels (Table 6).
In summary, structure–activity relationship studies initiating
from hit compound 1 have led to the identification of an advanced
compound 36 (code name: RQ-00203078). RQ-00203078 is a novel
and highly potent TRPM8 antagonist with human and rat IC50 val-
ues of 8.3 nM and 5.8 nM, respectively, in the menthol-induced
calcium influx assay, and demonstrated excellent activity in vivo
in a dose dependent manner with an ED50 value of 0.65 mg/kg in
the icilin-induced wet-dog shakes model in rats after oral adminis-
tration. Pharmacological studies using RQ-00203078 are now
underway in several pharmacological models other than the pain
model,11 and the results will be reported in the near future.
4. Sugano, K.; Kato, T.; Suzuki, K.; Kako, K.; Sujaku, T.; Mano, T. J. Pharm. Sci. 2006,
95, 2115.
5. Ghuman, J.; Zunszain, P. A.; Petitpas, I.; Bhattacharya, A. A.; Otagiri, M.; Curry, S.
J. Mol. Biol. 2005, 353, 38.
6. Gleeson, M. P.; Hersey, A.; Hannongbua, S. Curr. Top. Med. Chem. 2011, 11, 358.
7. (a) Inoue, T.; Ohmi, M.; Kawamura, K.; Ando, K.; Shishido, Y. WO2010125831,
2010.; (b) Kawamura, K.; Shishido, Y.; Ohmi, M. WO2012042915, 2012.
8. The molecular overlay was carried out by using Discovery StadioÒ 3.5.
9. Tanaka, H.; Shishido, Y. Bioorg. Med. Chem. Lett. 2007, 17, 6079.
10. (a) Wei, E. T. Psychopharmacology 1983, 81, 111; (b) Wei, E. T.; Sied, D. A. J.
Pharm. Pharmacol. 1983, 35, 110; (c) Tse, S. Y. H.; Wei, E. T. Psychopharmacology
1986, 90, 322.
Acknowledgments
The authors thank Masaki Sudo for helpful discussion. Also we
are grateful to Tetsuya Tamura, Kaori Muraji, and Sayaka Kodashi-
ma for performing the in vitro ADMET assays as well as Yoshiko
Sakaguchi for conducting the wet-dog shakes assay. Finally, we
wish to express sincere gratitude to Shinichi Koizumi for scientific
advice on this research program.
11. Okamoto, Y.; Ohkubo, T.; Ikebe, T.; Yamazaki, J. Int. J. Oncol. 2012, 40, 1431.
References and notes
1. (a) McKemy, D. D.; Neuhausser, W. M.; Jullus, D. Nature 2002, 416, 52; (b) Peier,
A. M.; Moqrich, A.; Hergarden, A. C.; Reeve, A. J.; Andersson, D. A.; Story, G. M.;
Earley, T. J.; Dragoni, I.; McIntyre, P.; Bevan, S.; Patapoutian, A. Cell 2002, 108,
705.