Analogues of morphanthridine and the tear gas dibenz[ b, f ][1,4]oxazepine (CR) as extremely potent activators of the human transient receptor potential ankyrin 1 (TRPA1) channel

Article abstract of DOI:10.1021/jm100477n

The TRPA1 channel can be considered as a key biological sensor to irritant chemicals. In this paper, the discovery of 11H-dibenz[b,e]azepines (morphanthridines) and dibenz[b,f][1,4]oxazepines is described as extremely potent agonists of the TRPA1 receptor. This has led to the discovery that most of the known tear gases are potent TRPA1 activators. The synthesis and biological activity of a number of substituted morphanthridines and dibenz[b,f][1,4]oxazepines have given insight into the SAR around this class of TRPA1 agonists, with EC₅₀ values ranging from 1 μM to 0.1 nM. Compounds 6 and 32 can be considered as the most potent TRPA1 agonists known to date, with 6 now being used successfully as a screening tool in the discovery of TRPA1 antagonists. The use of ligands such as 6 and 32 as pharmacological tools may contribute to the basic knowledge of the TRPA1 channel and advance the development of TRPA1 antagonists as potential treatment for conditions involving TRPA1 activation, including asthma and pain.

Full text of DOI:10.1021/jm100477n

J. Med. Chem. 2010, 53, 7011–7020 7011
DOI: 10.1021/jm100477n
Analogues of Morphanthridine and the Tear Gas Dibenz[b,f][1,4]oxazepine (CR) as Extremely
Potent Activators of the Human Transient Receptor Potential Ankyrin 1 (TRPA1) Channel
‡,
Harrie J. M. Gijsen,*,† Didier Berthelot,† Mirko Zaja,†,§ Bert Brone, Ivo Geuens,‡ and Marc Mercken‡
^
†Medicinal Chemistry Department and ‡Neuroscience, Johnson & Johnson Pharmaceutical Research & Development, Turnhoutseweg 30,
2340 Beerse, Belgium. §Current affiliation: 4SC AG, Am Klopferspitz 19a, 82152 Planegg, Martinsried, Germany. Current affiliation:
BIOMED Research Institute, Hasselt University, Agoralaan, Bld C, B-3590 Diepenbeek, Belgium.
Received April 19, 2010
The TRPA1 channel can be considered as a key biological sensor to irritant chemicals. In this paper, the
discovery of 11H-dibenz[b,e]azepines (morphanthridines) and dibenz[b,f ][1,4]oxazepines is described as
extremely potent agonistsoftheTRPA1receptor. Thishasled tothe discoverythatmostofthe knowntear
gases are potent TRPA1 activators. The synthesis and biological activity of a number of substituted
morphanthridines and dibenz[b,f ][1,4]oxazepines have given insight into the SAR around this class of
TRPA1 agonists, with EC50 values ranging from 1 μM to 0.1 nM. Compounds 6 and 32 can be considered
as the most potent TRPA1 agonists known to date, with 6 now being used successfully as a screening tool
in the discovery of TRPA1 antagonists. The use of ligands such as 6 and 32 as pharmacological tools may
contribute to the basic knowledge of the TRPA1 channel and advance the development of TRPA1
antagonistsaspotential treatment for conditions involving TRPA1 activation, including asthma and pain.
Introduction
covalent bond with cysteine residues present in the ion
channel.7 The reported potencies of the various agonistsrange
from 1 to 100 μM, with the most potent members of a parti-
cular class displaying activities in the low to submicromolar
range.1b Forexample, 3 hasa reported EC50 of 19 μM while an
analogous isatin derived cinnamide, dubbed “supercinnamal-
dehyde” in the same assay, had an EC50 of 0.8 μM.7b
The transient receptor potential ankyrin 1 (TRPA1,a for-
merly named ANKTM1) receptor belongs to the transient
receptor potential (TRP) family of cation-selective channels
which have been shown to transduce mechanical, thermal,
and pain-related inflammatory signals.1 Within this family,
the TRPA1 receptor appears to play a key role as a biological
sensor and has been implicated in a growing number of
pathophysiological conditions,2 such as neuropathic or in-
flammatory pain,3 airway diseases,4 and bladder disorders.5
The TRPA1 channel is activated by a number of pungent
chemicals, such as allyl isothiocyanate (1, AITC) present in
mustard oil, diallyl thiosulfinate (2, allicin) present in garlic,
and cinnamaldehyde (3) present in cinnamon (Chart 1).6 A
comprehensive overview of reported TRPA1 agonists can be
found in a recent article1b and includes phytochemicals,
environmental derived irritants, a variety of medicines, and
endogenously produced reactive reagents. Although several
nonelectrophilic agents, such as thymol and menthol, have
been reported as TRPA1 agonists, most of the known activa-
tors are electrophilic chemicals which have been shown to
activate the TRPA1 receptor via the formation of a reversible
In a high throughput screen aimed at the identification of
TRPA1 ligands, a family of hits was discovered to behave as
moderate to potent agonists of the TRPA1 receptor. They all
belonged to a class of tricyclic 6,11-dihydro-5H-dibenz[b,e]-
azepinesor5,6-dihydromorphantridines of generalstructure4
and exemplified by compound 5 (Chart 1). During resynthesis
of compound5 for hit confirmation, the compound was found
to be rather unstable and to be partly oxidized to the corre-
sponding morphanthridine 6 on prolonged standing in solu-
tion and exposure to air. In addition, upon handling of the
oxidized compound, laboratory personnel noticed a pungent,
irritating effect on eyes and nose. Isolation of pure 6 and
subsequent in vitro testing showed the compound to be an
extremely potent agonist of the human TRPA1 receptor with
an EC50 of 0.1 nM, a 1000-fold increase in potency over the
previously known agonists such as 1.1,2 A careful reanalysis of
the original DMSO solutions used in the HTS campaign
revealed that 1-5% of compound 5 had indeed been con-
verted into 6. Also, in all of the analogous 5,6-dihydromor-
phantridine hits 4, varying amounts of oxidized product could
be detected. Obviously, with a subnanomolar potency, trace
amounts of morphanthridine present would already result in a
considerable measured activity. The parent, unsubstituted
morphanthridine 7 was also shown to be a potent agonist of
the TRPA1 receptor with an EC50 of 3 nM.
*To whom correspondence should be addressed. Phone: þ32-14-
606830. Fax: þ 32-14-5344. E-mail: hgijsen@its.jnj.com.
aAbbreviations: A/F, ampere/farad; AITC, allyl isothiocyanate; BITC,
benzyl isothiocyanate; CA (BBC), bromobenzyl cyanide; CN, 1-chloroa-
cetophenone; CR, dibenz[b,f ][1,4]oxazepine; CS, 2-chlorobenzylidene
malononitrile; FDSS, functional drug screening system; HEK293, human
embryonic kidney 293; I/Cslow, current/slow component of membrane
capacitance; ICt50, concentration that will incapacitate 50% of the
population exposed to the tear gas for a specific period; TC50, threshold
concentration with effect on 50% of population after 1 min of exposure;
TRP, transient receptor potential; TRPA1, transient receptor potential
ankyrin 1; cTRPM8, canine transient receptor potential melastatin 8;
hTRPV1, human transient receptor potential vanilloid 1.
The structural resemblanceof the morphanthridinetricyclic
moiety with dibenz[b,f ][1,4]oxazepine (8, CR), which is one of
r
2010 American Chemical Society
Published on Web 08/31/2010
pubs.acs.org/jmc

7012 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 19
Gijsen et al.
Chart 1
Table 1. Tear Gases as TRPA1 Agonists: Potencies as a Result of Fluorometric (FDSS) Studies Compared with the Human in Vivo Efficiency of the
Tear Gases9,a
a TC50: threshold concentration with an effect on 50% of population after 1 min of exposure. ICt50: concentration that will incapacitate 50% of the
population exposed to the tear gas for a specific period.
the ingredients found in currently used tear gases, also known
as riot control or incapacitating agents, has led to the dis-
covery of the mechanism of action for most of the tear gases
used today and in the past.8 For the tear gases 8 (CR),
1-chloroacetophenone (9, CN), 2-chlorobenzylidene malono-
nitrile (10, CS), and bromobenzyl cyanide (11, CA or BBC), a
good correlation has been observed between the relative in
vitro TRPA1 potencies with their reported human in vivo
efficiencies as tear gases9 as shown in Table 1. Recently,
behavioral tests in mice, either using TRPA1 deficient mice
or treating normal mice with TRPA1 antagonists, have con-
firmed the central role of TRPA1 in the involuntary nocifen-
sive reflex responses to tear gas agents.10
Most of the TRPA1 agonists known today, including the
tear gases such as 8, 9, and 10, have in common an electro-
philic carbon or sulfur that can be subject to nucleophilic
attack by thiol (cysteine) or amine (lysine) groups, present
in the TRPA1 receptor, and can form a covalent interaction
with the receptor.7 The variety of structures in this group of
electrophilic agents illustrates that chemical reactivity in
combination with a lipophilicity enabling membrane permea-
tion is crucial totheirTRPA1 receptoragonistic effect.6 In this
paper we describe the SAR around this dibenz(ox)azepine
tricyclic class of TRPA1 agonists and demonstrate that, in
addition to the requirement of an electrophilic functionality,
noncovalent interactions with the receptor also play a role in
the potency of these compounds.
Chemistry
All the final dibenz(ox)azepine compounds were prepared
via oxidation of the corresponding dihydro derivatives. The
synthesis of 6 andthe carboxylic ester substituted 11H-dibenz-
[b,e]azepines (morphanthridines) analogues is depicted in
Scheme 1.11 Condensation of o-aminobenzyl alcohol 12 with
benzaldehyde 13 gave the bromo-substituted 2-phenyl-1,4-di-
hydro-2H-benz[d][1,3]oxazines 14a-e in quantitative yield.
Reductive cleavage with NaBH4 in ethanol afforded [2-benzy-
laminophenyl]methanol intermediates 15a-e. The 3-bromo
derivative 15b was prepared via a slightly modified procedure
by performing a reductive amination between benzaldehyde
and O-acetyl-3-bromo-2-aminobenzyl alcohol, followed by

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 19 7013
Scheme 1a
a Reagents and conditions: (a) iPrOH; (b) NaBH4, EtOH; (c) CH2Cl2, H2SO4; (d) Br2, HOAc; (e) Pd(OAc)2, dppp, K2CO3, MeOH, THF, 50 atm of
CO, 125 °C; (f) MnO2, toluene or air oxidation.
Scheme 2a
a Reagents and conditions: (a) SOCl2; (b) Et3N, THF; (c)NaOH, DMF; (d) LiAlH4, or BH3, THF; (e) Pd(OAc)2, dppp, K2CO3, MeOH, THF, 50 atm
of CO, 125 °C; (f) MnO2, toluene.
hydrolysis of the acetyl protecting group (not shown). Inter-
mediates 15a-e were cyclized to the desired 6,11-dihydro-5H-
dibenz[b,e]azepines 16a-f with sulfuric acid using dichloro-
methane as cosolvent. For the 30-bromo derivative 15a, this
resulted in a mixture of 10- and 8-bromo derivatives 16a and
16b, respectively. For all other intermediates 15b-e, the cor-
responding brominated 6,11-dihydro-5H-dibenz[b,e]azepines
16c-f were obtained as single regioisomers. The 2-bromo deri-
vative 16g was prepared more easily via regioselective bromi-
nation of the readily available 6,11-dihydro-5H-dibenz[b,e]-
azepine 17. Subsequently, the bromo derivatives 16a-g were
converted into carboxylic esters 18a-i followed by oxidation
to the desired dibenz[b,e]azepines. This final oxidation was
routinely carried out using an excess of MnO2 in toluene. Upon
prolonged standing in aqueous solution, ranging from hours
to days, oxidation of the 6,11-dihydro-5H-dibenz[b,e]azepines
to the corresponding dibenz[b,e]azepines was observed. This
became apparent especially during lengthy preparative re-
versed phase HPLC purifications and was used to our ad-
vantage in the isolation of a number of final compounds, such
as 8-bromo derivative 24.
between anilines 25 and benzoic acids 26 followed by intramo-
lecular SNAr reaction installed the required tricyclic ring system.
Reduction of the dibenzo[b,f ][1,4]dioxepanones 28a-c gave
the brominated 10,11-dihydrodibenzo[b,f ][ 1,4]oxazepines
29a-c, which were converted to the required target com-
pounds 31-32 in analogy to the synthesis of the dibenz[b,e]-
azepines.
Several of the brominated dihydro intermediates 16 and
29 were also used to synthesize derivatives bearing a variety
of substituents. Direct oxidation of intermediates 8- and 10-
bromo derivatives 16b and 16a gave bromo analogues 34 and
35, respectively (Scheme 3). Cyano- and carboxamide-sub-
stitituted dibenz(ox)azepine derivatives 36-39 and 40-42,
respectively, were prepared starting from the corresponding
bromo derivatives as exemplified in Scheme 3.
Carboxylic acid derivative 46 was prepared via basic hydro-
lysis of 18a (Scheme 4). The intermediate acid 45 was directly
oxidized to 46 during the extended hydrolysis reaction time
(3 days). Carboxylic acid 45 wasalsoformedviaaCOinsertion
reaction of 16a in the presence of water. Again, intermediate 45
was not isolated but immediately reacted to form methoxy-
propylamide 47, which was subsequently oxidized to give
compound 49. The diethylamide analogue 50 was obtained
The general synthesis of the dibenz[b,f ][1,4]oxazepine car-
boxylic esters is depicted in Scheme 2. Benzamide formation

7014 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 19
Gijsen et al.
directly in low yield via a CO insertion reaction of 16a in the
presence of diethylamine. Oxidation of intermediate 48 took
place during the reversed phase HPLC purification leading
to 50.
correspond to changes in Ca2þ concentration in the cell
can kinetically be monitored with the FDSS instrument
(Hamamatsu) and are indicative for agonism toward the
TRPA1 ion channel. The results are expressed as EC50 repre-
senting the molar concentration of the compound, which
resulted in 50% of the maximal increase in Ca2þ concentra-
tion. Each quoted value is the mean of at least four experi-
ments. Further biological experiments are explained in the
next section.
Biology
Primary in Vitro Functional Assay. All compounds were
tested using Ca2þ fluorometry. Binding of an agonist to
the TRPA1 ion channel opens the ion channel which causes
a robust increase in intracellular Ca2þ concentration. For
detecting and measuring intracellular Ca2þ concentration,
hTRPA1 transfected HEK293 cells were loaded with a
Ca2þ-sensitive dye (Fluo4-AM, Molecular Probes, Invitro-
gen, Belgium). Changes in fluorescence in the cell that
Results and Discussion
Commercially available and structurally related but none-
lectrophilic, tricyclic analogues of morphanthridine (7) or 8,
such as 51, 52, and fluperlapine (53) were tested and found
to be inactive (EC50 >5 μM) (Chart 2). Substitution of the
electrophilic carbon in 8 with a methyl group, giving com-
pound 54, also sufficed to suppress the agonistic activity
completely. Incubation of 54 with benzylmercaptane as a
thiol nucleophile, mimicking the cysteine residues in the
TRPA1 receptor, did not show any formation of adduct, in
contrast to similar experiments with 8.8 The inactivity of
phenanthridine (55) illustrates further the need for an electro-
philic carbon.
Scheme 3a
The presence of an unsubstituted, electrophilic, disubsti-
tuted imine is no guarantee for activity. Benzalaniline (56) was
shown to react rapidly with thiols such as benzylmercaptan
but was totally devoid of activation of the hTRPA1 receptor.
For 56, the orientation of the two phenyl groups may prohibit
the positioning of the imine functionality in the proximity of
the cysteine thiols of the ion channel, hence preventing
covalent binding and activation of the receptor. The sulfur
analogue dibenzo[b,f ][1,4]thiazepine (57) did activate the
TRPA1 receptor withanEC50 of 9 nM. Again, this compound
has been described as being extremely irritating to the skin,
mucous membranes, and eyes.12 None of the inactive com-
pounds 51-56 were able to block the action of 6 or other
TRPA1 agonists, and they can therefore be considered also to
be devoid of hTRPA1 antagonistic activity.
a Reagents and conditions: (a) CuCN, DMF, 140 °C; (b) MnO2,
toluene; (c) H2SO4.
The data for the compounds in Charts 1 and 2 confirm
the requirement of a thiol-reactive electrophilic component,
Scheme 4a
a Reagents and conditions: (a) LiOH, THF/H2O; (b) Pd(OAc)2, dppp, THF, Et3N, H2O, 50 atm of CO, 150 °C; (c) methoxypropylamine, HBTU;
(d) Pd(OAc)2, dppp, THF, Et2NH, 50 atm of CO, 150 °C; (e) MnO2, toluene.

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 19 7015
Table 3. Activity of Substituted Dibenz(ox)azepine Analogues
Chart 2
compd
substituent
AAA
hTRPA1 EC50 (nM)
34
35
36
37
38
39
40
41
42
43
44
46
49
50
58
59
8-Br
CH2
CH2
CH2
CH2
O
33
10-Br
8-CN
10-CN
3.5
158
1.6
0.15
0.10
2.5
0.08
0.11
933
6.2
245
7.7
40
10-CN
1-CN
O
10-CONH2
10-CONH2
1-CONH2
8-COOiPr
10-COOnBu
10-COOH
10-CONH(CH2)3OMe
10-CONEt2
9-OMe
CH2
O
O
Table 2. Activity of Carboxylic Ester Substituted Dibenz(ox)azepine
Analogues
CH2
CH2
CH2
CH2
CH2
O
14
2.1
9-OH
O
compd
substituent
A
TRPA1 EC50 (nM)
from the reactive center, indicating that steric interactions
close to the reactive cysteine thiol groups of the ion channel
are the determining factor for the potency in this class of
TRPA1 activators.
7
H
CH2
CH2
CH2
CH2
CH2
CH2
CH2
O
3
19
20
21
22
23
24
31
32
6
1-COOMe
2-COOMe
3-COOMe
4-COOMe
7-COOMe
8-COOMe
9-COOMe
10-COOMe
10-COOMe
0.6
0.4
19
Further variation on substituents was investigated as
shown in Table 3. To confirm the difference in activity of
8- vs 10-substituted analogues, the 8-bromo and 8-cyano deri-
vatives were also prepared and tested. The 8-bromo derivative
34 showed a 10-fold decrease in activity compared to the
10-bromo analogue 35. For the 8- and 10-cyano derivatives 36
and 37, a 100-fold difference in activity was observed. Cyano
substituted dibenz[b,f ][1,4]oxazepines 38 and 39 were among
the most potent analogues synthesized and confirm the data
from Table 2, indicating that the 1- and 10-positions are the
optimal places for substitution, as well as the increased
potency for dibenzoxazepines compared to dibenzazepines.
An increase in size of the 8-substituent, such as in isopropyl
carboxylic ester substituted 43, resulted in a further decrease
in activity compared to methyl ester 24. This is in line with the
hypothesis that (lack of) steric hindrance is the main driver for
potency. Larger substituents on the 10-position such as in
compounds 44, 49, and50alsoresultedinreduced potencybut
still retained significant activity. This may provide a handle to
work on druglike properties such as aqeuous solubility.
As mentioned above, most of the known TRPA1 agonists
are lipophilic compounds, which is in agreement with the low
activity for carboxylic acid 46. On the other hand, carbox-
amides 40-42 displayed a similarly high potency as their
cyano analogues 37-39. The hydroxy- and methoxy-substi-
tuted morphanthridines 58 and 59 were prepared and tested,
since their irritant properties have been described before.13
The methoxy derivative 58 has been described as highly
irritant, whereas in the same study the hydroxyl analogue 59
did not show any irritant properties. Surprisingly, in our
hands, the hydroxy-substituted 59 showed a more potent in
vitro activation of the TRPA1 receptor than 58. A possible
explanation for this in vitro/in vivo discrepancy could be the
lower membrane permeability of the more polar 59 compared
to 58. Patch-clamp electrophysiology experiments have
shown that chemical activators of TRPA1 need to traverse
31
31
100
6
O
CH2
0.05
0.1
as observed in most exogenous and endogenous TRPA1
activators.1b In addition, it illustrates that this electrophilic
moiety needs to be accessible to the cysteine thiol groups in the
TRPA1 receptor. This creates the possibility to further opti-
mize the interaction of the benzo(ox)azepines class of TRPA1
agonists via structural modification.
The influence of substitution of the dibenz(ox)azepine
tricyclic core was investigated via a series of morphanthridine
and CR analogues containing various substituents on all of
the eight possible aryl carbon positions.
The potencies of a set of carboxylic methyl ester substituted
derivatives are given in Table 2. Although all of the synthe-
sized analogues were shown to activate the TRPA1 receptor,
the large differences in potencies point toward additional
interactions with the receptor. This is best illustrated by the
1000-fold potency difference between compounds 6 and 24. In
both compounds, the carboxylic methyl ester substituents are
oriented in a meta fashion to the electrophilic imine carbon
atom, thus resulting in a comparable influence on the electro-
philic nature of this carbon. The influence on the potency of
the carboxylic ester substituents on the 8- and 10-position can
therefore only be explained via additional, noncovalent inter-
actions with the TRPA1 receptor.
In general, 1-, 2-, and 10-substituted compounds showed
the highest potencies (Table 2, compounds 19, 20, 32, and 6),
with CR analogue 32 being the most potent compound dis-
covered so far, having an EC50 of 50 pM. Overall, substitution
in proximity to the electrophilic imine moiety (compounds
21-24) resulted in a lower activity than substitution away

7016 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 19
Gijsen et al.
Figure 1. Concentration response graphs of TRPA1 agonists in electrophysiological measurements. In panels A and B the hTRPA1 current
density, measured at 75 and -75 mV, respectively, was plotted as a function of the concentration of the different agonists. Panel C shows the
current density measured at a holding potential of -30 mV. The current density was calculated by normalizing the currents to the slow
component of membrane capacitance (I/Cslow), measured (ampere/farad: A/F) during the experiment. Note that the I/Cslow has negative
valueswhen measured at -30 and -75 mV becausethese areinwardcurrents. In panel D the time to peak is shown as a function of concentration.
The time to peak is the time between the application of the agonist and time point of the maximal current amplitude, with a maximum of 300 s.
Every data point represents the mean and SEM of five to nine replicates for 6 and 32 and three to six replicates for AITC (1) and BITC.
the plasma membrane to activate the ion channel.10 In the
same study, the requirement of cysteine residues present in the
Table 4. Effect of the TRPA1 Agonists on Currents Measured at
Different Membrane Potentials and on the Time to Peaka
AITC (1)
4.65 ( 0.75 4.56 ( 0.24 7.14 ( 0.20 6.90 ( 0.16
4.92 ( 0.17 4.81 ( 0.25 7.42 ( 022 7.08 ( 0.17
4.98 ( 0.12 4.62 ( 0.97 8.01 ( 0.37 7.62 ( 0.38
BITC
6
32
cytosolic, N-terminal domain of the TRPA1 receptor was
demonstrated by the lack of activation by 8 and other tear
gases of mutant TRPA1 channels.
In order to confirm the effect of the compounds on TRPA1,
whole cell voltage clamp experiments were performed for
compounds 6 and 32 in comparison to the known agonists
AITC (1) and benzylisothiocyanate (BITC), using the meth-
odology as described previously8 (Figure 1).
-75 mV
-30 mV
75 mV
time to peak 5.24 ( 0.19 4.96 ( 0.40 8.36 ( 0.14 8.13 ( 0.11
a The pEC50 values from the concentration response curves in Figure 1
are compared (-log [agonist], mean ( SEM). The pEC50 values of the
different agonists at -30 mV were significantly different (p e 0.05; F-test).
The concentration response curves at the three membrane
potentials are plotted in Figure1A-C. In addition, the time to
peakofthe samecurrent maximawas plottedasconcentration
response curves in Figure 1D. The pEC50 values obtained
from the fitting procedure are shown in Table 4. A membrane
potential of -30 mV is probably the most relevant membrane
potential because activated (i.e., open) TRPA1 channels will
force the resting membrane potential of a cell toward the
equilibrium potentialof the TRPA1 channel(i.e.,-10to0 mV
in a physiological environment).
These results confirm that 6 and 32 are potent TRPA1
agonists. The potencies of AITC and BITC are hard to
determine because of solubility problems at higher concentra-
tions. The plateau phase of the concentration response rela-
tionships has not been reached, but we can at least conclude
that their EC50 values will be above the micromolar range.
To check the selectivity of this class of TRPA1 agonists, we
investigated the cross-selectivity on hTRPV1 (which is the target
for the capsaicin-containing pepper sprays) and cTRPM814
on compounds 6, 7, and 8.15 No agonism was found on either
hTRPV1 or cTRPM8, confirming their TRPA1 selectivity
(data not shown).
Upon incubation of the ester derivatives 6 and 32 with
human liver microsomes at a 5 μM, complete conversion was
observed to the corresponding acids within 15 min of treat-
ment. After prolonged (7 h) incubation under similar condi-
tions but in the absence of liver microsomes, no degradation
was observed, demonstrating the chemical stability of com-
pounds 6 and 32. The chemical stability also became apparent
upon accidental, minimal contamination of the hands with
these compounds, which 24 h later was still capable of
inducing strong lachrimation upon hand-eye contact.
Incubation with liver microsomes in the presence of glu-
tathion and subsequent analysis of the mixture for the pre-
sence of glutathion adducts are a commonly used technique
in drug discovery to detect the formation of reactive inter-
mediates.16 When compounds 6, 8, and 32 were added to a
solution containing liver microsomes and glutathione, no
parent or metabolite-glutathione adducts were observed
after 1 h of incubation for 6 or 32 while at least five different
adducts could be detected for 8.17 No glutathione adduct
formation was observed in the absence of human liver micro-
somes under the same incubation conditions, indicating the

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 19 7017
need for metabolic activation before covalent binding of 8
toglutathione. Thelackofformationofreactiveintermediates
or observation of parent-glutathion adducts for 6 and 32
could result in a reduced toxicity profile of these compounds
compared to 8. The use of ester-substituted analogues of 8
as incapacitating (tear gas) agents may therefore lead to an
improved safety profile. The currently used tear gases CR and
CS have been described as environmentally persistent chemi-
cals, which can also be countered via the introduction of
carboxylic acid ester substituents by increasing aqueous solu-
bility and hydrolytic degradation. Further research into the
structure-activity relationship of these TRPA1 agonists may
leadtomore safeandenvironmentallyfriendlyalternatives for
the currently used tear gases.
The work described in this paper has led to the identifica-
tion of the most potent class of TRPA1 agonists knowntoday.
The results confirm the observation of thiol-reactive electro-
philes being able to activate the TRPA1 receptor. In addition,
the inactivity of benzalaniline (56) and the SAR around the
benzo(ox)azepines demonstrate the importance of an optimal
fit of the electrophiles within the TRPA1 receptor in order to
achieve maximal potency.
time of 0.02 s. The capillary needle voltage was 3.5 kV, and the
source temperature was maintained at 140 °C. Nitrogen was used
as the nebulizer gas. Sensitive reactions were performed under
nitrogen. Commercial solvents were used without any pretreat-
ment. Preparative HPLC purifications were carried out using a
Shandon Hyperprep column C18 BDS (base deactivated silica),
8 μm, 250 g, 5 cm i.d., eluting with a gradient of three subsequent
mobile phases: phase A consisting of a 0.25% NH4HCO3 solu-
tion in water; phase B consisting of CH3OH; phase C consisting
of CH3CN).
In general, final compounds obtained as solids were triturated
with DIPE. Alternatively, some compounds were precipitated
and isolated as the HCl salt, by dissolving them in diethyl ether
followed by the addition of a 1 N HCl solution in diethyl ether.
Allyl isothiocyanate (AITC), benzyl isothiocyanate (BITC),
acrolein, 1-chloroacetophenone (CN), ethyl bromoacetate, and
compounds 51, 52, 55, 56 were purchased from Sigma-Aldrich.
Fluperlapine (53) was purchased from Sanver Tech. 2-Chloro-
benzylidene malononitrile (CS), 54, and 57 were purchased
from Scientific Exchange. Bromobenzyl cyanide19 morphanth-
ridine,20 CR,21 58,13 and 5913 were synthesized according to
literature procedures. All compounds were dissolved in dime-
thyl sulfoxide (DMSO) as 5, 10, or 100 mM stock solutions
which were used within 48 h of their preparation.
CAUTION: All active TRPA1 agonist compounds men-
tioned in this article should be handled with care in a fumehood,
using gloves, because of their strong irritant and lachrymatory
properties.
Compound 6 has now been used successfully as a screening
tool in the discovery of TRPA1 antagonists.18 Data obtained
using 6 were comparable to the use of the more commonly
used but less potent isothiocyanate TRPA1 agonists such as
benzyl isothiocyanate, with the advantage of giving a more
robust signal. The SAR results and the use of ligands such as 6
and 32 as pharmacological tools may contribute to the basic
knowledge of the TRPA1 channel and facilitate the discovery
of new treatments for diseases such as asthma and pain.
5H-Dibenzo[b,e]azepine-10-carboxylic Acid Methyl Ester
(6). Step 1. 2-(3-Bromophenyl)-1,4-dihydro-2H-benzo[d][1,3]oxazine
(14a).A mixtureof2-(hydroxymethyl)aniline (9.0 g, 73 mmol) and
3-bromobenzaldehyde (13.5 g, 73 mmol) in 2-propanol (100 mL)
was stirred for 3 h at room temperature. The solvent was eva-
porated, and the residue (20.5 g, 97%) was used as such in the next
step. Part (3 g) of the residue was crystallized from hexane. The
precipitate was filtered off and dried in vacuo, yielding 1.37 g of
Experimental Section
1
intermediate 14a; mp 53.6 °C. H NMR (360 MHz, DMSO-d6)
Chemistry. Melting points were determined with a DSC823e
(Mettler-Toledo) and were measured with a temperature gradient
δ ppm 4.76 (d, J=14.6 Hz, 1 H), 4.99 (d, J=14.6 Hz, 1 H), 5.59
(d, J=3.3 Hz, 1 H), 6.43 (d, J=2.9 Hz, 1 H), 6.68 (d, J=7.7 Hz,
2 H), 6.91 (d, J=7.3 Hz, 1 H), 7.01 (t, J=7.7 Hz, 1 H), 7.39 (t, J=
7.9 Hz, 1 H), 7.53 (d, J=8.1 Hz, 1 H), 7.60 (d, J=8.1 Hz, 1 H),
7.70 (s, 1 H).
1
of 30 °C/min. The reported values are peak values. H NMR
spectra were recorded with Bruker Avance DPX 400 and 360
spectrometers, and chemical shifts (δ) are expressed in parts per
million (ppm) with TMS as internal standard. Mass spectral data
were obtained via GC-MS analysis, using an Agilent 6890 series
GC combined with a 5973N mass selected detector. High resolu-
tion mass spectra were acquired on a Acquity UPLC/QToF
Premier system (Waters). The mass spectrometer was equipped
with an ESI source and was operated in the positive ion mode.
Leucine enkephalin was used as external lock spray. The source
temperature was 120 °C, desolvation temperature was 350 °C,
and the cone voltage was set to 30 V. Elemental analyses were
performed with a Carlo-Erba EA1110 analyzer. Combustion
analytical results were within 0.4% of the theoretical values
except when noted otherwise. For all tested and final compounds
where no analytical purity is mentioned, compounds were con-
firmed to be >95% pure via HPLC methods. The analytical LC
measurement was performed using an Acquity UPLC (Waters)
system comprising a binary pump, a sample organizer, a column
heater (set at55 °C), a diode-array detector (DAD), and a bridged
ethylsiloxane/silica hybrid (BEH) C18 column (1.7 μm, 2.1 mm ꢀ
50 mm; Waters Acquity) with a flow rate of 0.8 mL/min. Two
mobile phases (mobile phase A consisting of 0.1% formic acid in
H2O/methanol95/5;mobile phaseB consisting ofmethanol) were
used to run a gradient condition from 95% A and 5% B to 5% A
and 95%Bin1.3minandheldfor 0.2min. Aninjectionvolume of
0.5 μL was used. Cone voltage was 10 V for positive ionization
mode and 20 V for negative ionization mode. Flow from the
column was split to a mass spectrometer. The MS detector was
configured with an electrospray ionization source. Mass spectra
were acquired by scanning from 100 to 1000 in 0.18 s using a dwell
Step 2. [2-(3-Bromobenzylamino)phenyl]methanol (15a). So-
dium borohydride (4.4 g, 0.12 mol) was added slowly to a
mixture of 14a (17.1 g, 59 mmol) in dry ethanol (200 mL) under
a nitrogen atmosphere. The reaction mixture was stirred and
heated at reflux for 1 h. The mixture was cooled on an ice-water
bath, quenched with a 20% aqueous NH4Cl solution, and ex-
tracted with CH2Cl2. The organic layer was dried over MgSO4,
filtered, and concentrated under reduced pressure to give 14.8 g
(86%) of crude 15a. 1H NMR (360 MHz, DMSO-d6) δ ppm 4.37
(d, J=6.2 Hz, 2 H), 4.49 (d, J=5.5 Hz, 2 H), 5.17 (t, J=5.5 Hz, 1
H), 5.83 (t, J=6.0Hz, 1 H), 6.40 (d, J=8.1 Hz, 1 H), 6.54 (t, J=7.3
Hz, 1 H), 6.98 (dd, J=15.4, 1.5 Hz, 1 H), 7.10 (d, J=7.3 Hz, 1 H),
7.28 (d, J=8.1 Hz, 1 H), 7.37 (d, J=7.7 Hz, 1 H), 7.41 (dd, J=8.1,
1.83 Hz, 1 H), 7.54 (s, 1 H).
Step 3. 10-Bromo-6,11-dihydro-5H-dibenzo[b,e]azepine (16a)
and 8-Bromo-6,11-dihydro-5H-dibenzo[b,e]azepine (16b). A so-
lution of 15a (52.6 g, 0.18 mol) in CH2Cl2 (50 mL) was added
over a 1 h period to a cooled (-10 to -20 °C) solution of con-
centrated H2SO4 (500 mL). Then the ice bath was removed, and
the mixture was stirred for 1 h at room temperature. The reac-
tion mixture was added to ice-water, cooled on ice, and
alkalized with a 50% aqueous NaOH solution. The resulting
mixture ((3 L) was extracted with CH2Cl2. The organic layer
was separated, dried over MgSO4, filtered, and the filtrate was
concentrated in vacuo. The residue (47 g) was partly (8 g)
purified via supercritical fluid chromatography (SFC, column
Diacel AD-H 30 mm ꢀ 250 mm, mobile phase of 55% MeOH/
45% CO2 þ 0.2% isopropylamine, 40 °C, 100 bar) to give 2 g

7018 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 19
Gijsen et al.
(24%) of 16a and 4.65 g (55%) of 16b. Data for 16a: mp
100.9 °C. H NMR (360 MHz, DMSO-d6) δ ppm 4.33 (s, 2
129.75 (d), 132.27 (d), 132 29 (s), 145.57 (s), 146.45 (s), 160.26 (d),
165.95 (s); MS (EIþ) m/z (rel intensity) 251 (100%, Mþ), 236
(3%), 220 (22%), 192 (35%), 165 (26%); ESI-HRMS, found
252.1035, calculated for C16H14NO2 (M þ H)þ 252.1025. Anal.
(C16H13NO2 0.4 H2O) C, H, N.
1
H), 4.47 (d, J=5.1 Hz, 2 H), 5.98 (t, J=5.1 Hz, 1 H), 6.35-6.42
(m, 2 H), 6.84 (td, J=7.7, 1.5 Hz, 1 H), 6.89 (dd, J=7.7, 1.5 Hz, 1
H), 7.13 (t, J=8.1 Hz, 1 H), 7.25 (dd, J=7.7, 1.5 Hz, 1 H), 7.51
(dd, J=8.1, 1.1 Hz, 1 H). Data for 16b: mp 135.5 °C. 1H NMR
(360 MHz, DMSO-d6) δ ppm 4.09 (s, 2 H), 4.39 (d, J=5.1 Hz, 2
H), 5.92 (t, J=5.1 Hz, 1 H), 6.38 (t, 2 H), 6.82 (td, J=7.7, 1.5 Hz,
1 H), 6.88 (d, J=7.3 Hz, 1 H), 7.20 (d, J=8.1 Hz, 1 H), 7.39 (dd,
J=8.1, 2.2 Hz, 1 H), 7.46 (d, J=2.2 Hz, 1 H).
Dibenzo[b,f ][1,4]oxazepine-4-carboxylic Acid Methyl Ester
(32). Step 1. 3-Bromo-2-fluoro-N-(2-hydroxyphenyl)benzamide
27a. A mixture of 3-bromo-2-fluorobenzoic acid (5.0 g, 22
mmol) and SOCl2 (20 mL) was stirred and refluxed for 2 h.
The reaction mixture was concentrated in vacuo and coevapo-
rated twice with toluene. The crude acid chloride residue was
dissolved in THF (25 mL) and was added dropwise to a mixture
of 2-aminophenol (2.4 g, 22 mmol) and triethylamine (4.5 g,
44 mmol) in THF (75 mL) at 0 °C. The reaction mixture was
allowed to warm to room temperature and stirred overnight.
The reaction mixture was poured into water (400 mL) and
acidified to pH 4-5 with a 1 N aqueous HCl solution. The
precipitate was filtered off and washed with a 1 N aqueous HCl
solution and water, dried in vacuo, yielding 6.4 g (93%) of 27a;
mp 219.3 °C. 1H NMR (360 MHz, DMSO-d6) δ ppm 6.83 (t, J=
7.6 Hz, 1 H), 6.92 (d, J=8.8 Hz, 1 H), 7.02 (t, J=7.6 Hz, 1 H),
7.30 (t, J=7.9 Hz, 1 H), 7.78 (t, J=7.1 Hz, 1 H), 7.89 (t, J=7.4
Hz, 1 H), 7.94 (d, J=7.9 Hz, 1 H), 9.61 (d, J=4.9 Hz, 1 H), 9.96
(s, 1 H).
Step 2. 4-Bromo-10H-dibenzo[b,f ][1,4]oxazepin-11-one (28a).
A mixture of 27a (3.8 g, 12 mmol) and sodium hydroxide (490
mg, 12 mmol) in DMF (60 mL) was stirred and refluxed for 5 h.
The reaction mixture was poured onto 800 mL of ice-water and
the resulting precipitate filtered off and washed with a 1 N
aqueous sodium hydroxide solution and water, then dried in
vacuo, yielding 2.5 g (70%) of 28a; mp 272.3 °C. 1H NMR (360
MHz, DMSO-d6) δ ppm 7.13 (t, J=7.7 Hz, 1 H), 7.18 (dd, J=
8.1, 1.8 Hz, 1 H), 7.37 (td, J=7.7, 1.1 Hz, 1 H), 7.44 (dd, J=7.7,
1.8 Hz, 1 H), 7.47 (dd, J=8.4, 1.1 Hz, 1 H), 7.67 (td, J=8.1, 1.8
Hz, 1 H), 7.81 (dd, J=7.7, 1.8 Hz, 1 H), 10.72 (s, 1 H).
Step 3. 4-Bromo-10,11-dihydrodibenzo[b,f ][1,4]oxazepine (29a).
To a suspension of 28a in THF (80 mL) was added a 1 M BH3
solution in THF (24 mL, 24 mmol), and the reaction mixture was
stirred at room temperature for 2 weeks. The reaction mixture was
cooled on ice, and 100 mL of a 1 N aqueous HCl solution was
added. The mixture was partially concentrated in vacuo and then
basified with solid NaHCO3 (pH ∼7). The aqueous layer was
extracted with 2 ꢀ 150 mL of CH2Cl2. The separated organic layer
was dried over MgSO4, filtered, and the filtrate was concentrated.
The residue was purified via reversed phase HPLC, yielding 2.1 g
(85%) of 29a; mp 254 °C. 1H NMR (360 MHz, DMSO-d6) δ ppm
4.43 (d, J=4.3 Hz, 2 H), 6.17 (t, J=4.3 Hz, 1 H), 6.52 (td, J=7.6,
1.6 Hz, 1 H), 6.57 (dd, J=8.1, 1.6Hz, 1 H), 6.82 (td, J=7.4, 1.7 Hz,
1 H), 7.07 (t, J=7.7 Hz, 1 H), 7.12 (dd, J=7.7, 1.5 Hz, 1 H), 7.35
(dd, J=7.5, 1.3 Hz, 1 H), 7.58 (dd, J=8.1, 1.5 Hz, 1 H).
Step 4. 10,11-Dihydrodibenzo[b,f ][1,4]oxazepine-4-carboxylic
Acid Methyl Ester (30a). A 75 mL stainless steel autoclave was
charged under nitrogen atmosphere with 29a (725 mg, 2.6
mmol), Pd(OAc)2 (10 mg, 0.04 mmol), 1.3-bis(diphenylphos-
phino)propane (40 mg, 0.1 mmol), KOAc (750 mg, 7.9 mmol),
methanol (20 mL), and THF (20 mL). The autoclave was closed
and pressurized to 50 atm. CO and the reaction was carried out
for 16 h at 125 °C. The reaction mixture was filtered and con-
centrated in vacuo. The residue was partitioned between CH2Cl2
and water. The separated organic layer was dried over MgSO4,
filtered, and the filtrate was concentrated. The residue was
purified by silica gel flash chromatography (eluent heptane/
CH2Cl2 70/30 to 0/100), yielding 550 mg (82%) of 30a as a light
yellow oil. 1H NMR (360 MHz, DMSO-d6) δ ppm 3.89 (s, 3 H),
4.43 (d, J=4.2 Hz, 2 H), 6.13 (t, J=4.2 Hz, 1 H), 6.51 (td, J=7.7,
1.5 Hz, 1 H), 6.56 (dd, J=8.1, 1.5 Hz, 1 H), 6.80 (td, J=7.7, 1.5
Hz, 1 H), 7.04 (dd, J=8.1, 1.5 Hz, 1 H), 7.22 (t, J=7.7 Hz, 1 H),
7.55 (dd, J=7.7, 1.8 Hz, 1 H), 7.65 (dd, J=7.7, 1.8 Hz, 1 H).
Step 5. Dibenzo[b,f ][1,4]oxazepine-4-carboxylic Acid Methyl
Ester, HCl Salt (32). A mixture of 30a (510 mg, 2.0 mmol) and
Step 4. 6,11-Dihydro-5H-dibenzo[b,e]azepine-10-carboxylic
Acid Methyl Ester (18a). A mixture of 16a (2.2 g, 8.0 mmol),
KOAc (4 g, 41 mmol), Pd(OAc)2 (40 mg, 0.2 mmol), and 1,3-
bis(diphenylphosphino)propane (0.16 g, 0.4 mmol) in methanol
(100 mL) and THF (100 mL) was placed in a pressure reactor
and pressurized with CO gas up to 50 bar. The reaction mixture
was heated at 125 °C for 16 h, then cooled, filtered over dicalite,
and the solvent was evaporated. The residue was partitioned
between CH2Cl2 and water. The organic layer was dried over
MgSO4, filtered, and then the filtrate was concentrated. The
residue was purified by flash chromatography over silica gel
(eluent, CH2Cl2). The desired fractions were collected and the
solvent was evaporated, yielding 1.86 g (92%) of 18a;mp105.7°C.
1H NMR (400 MHz, CDCl3) δ ppm 3.93 (s, 3 H), 4.58 (s, 3 H),
4.61 (s, 2 H), 6.37 (d, J=8.1 Hz, 1 H), 6.60 (t, J=6.8 Hz, 1 H), 6.94
(t, J=8.4 Hz, 1 H), 7.12 (d, J=7.3 Hz, 1 H), 7.23 (t, J=7.7 Hz, 1
H), 7.34 (d, J=7.3 Hz, 1 H), 7.77 (dd, J=8.1, 1.5 Hz, 1 H). Anal.
(C16H15NO2) C, H, N.
Step 5. General MnO2 Oxidation Procedure: 11H-Dibenzo[b,e]-
azepine-10-carboxylic Acid Methyl Ester (6). A mixture of 18a
(1.86 g, 74 mmol) and manganese oxide (3.2 g, 37 mmol) in toluene
(20 mL) was stirred at 90 °C for 4 h. The reaction mixture was
filtered over a silica gel pad (eluent, toluene and then CH2Cl2). The
organic layer was concentrated. The residue was purified by flash
chromatography over silica gel (eluent, heptane/EtOAc 100/0 to
70/30). The product fractions were collected, and the solvent was
evaporated. The residue was triturated with diisopropyl ether,
filtered off, and dried, yielding 1.35 g (73%) of 6; mp 86.7 °C. 1H
NMR (360 MHz, CDCl3) δ ppm 4.01 (3 H, s), 4.06 (2 H, br s),
7.25-7.33 (2 H, m), 7.36 (1 H, t, J=7.7 Hz), 7.40-7.44 (1 H, m),
7.49-7.53 (1 H, m), 7.65 (1 H, dd, J=7.7, 1.1 Hz), 8.01 (1 H, dd,
J=7.9, 1.3 Hz), 8.96 (1 H, s); 13C NMR (90 MHz, DMSO-d6) δ
ppm 33.05 (t), 52.39 (q), 125.52 (d), 126.22 (d), 127.04 (d), 127.07
(d), 127.98 (d), 128.24 (s), 131.82 (d), 131.97 (s), 132 24 (d), 133.32
(s), 139.83 (s), 145.21 (s), 159.84 (d), 167.14 (s); MS (EIþ) m/z (rel
intensity) 251 (99%, Mþ), 236 (66%), 220 (15%), 208 (5%), 191
(100%), 165 (22%); ESI-HRMS, found 252.1041, calculated for
C16H14NO2 (M þ H)þ 252.1025. Anal. (C16H13NO2) C, H, N.
6,11-Dihydro-5H-dibenzo[b,e]azepine-8-carboxylic Acid Methyl
Ester (18b) and 11H-dibenzo[b,e]azepine-8-carboxylic Acid Methyl
Ester (24). A mixture of 16b (2.74 g, 10 mmol), 1,10-(1,3-pro-
panediyl)bis[1,1-diphenylphosphine(80mg, 0.2mmol), Pd(OAc)2
(20 mg, 0.08 mmol), and triethylamine (4 mL) in methanol
(50 mL) and THF (50 mL) was stirred in an autoclave at 125 °C
for 16 h under 50 atm of carbonmoxide pressure, then for 4 h
at 150 °C. Thereaction mixture wascooled and filtered over dica-
lite. The solvent was evaporated, and the residue was purified
via reversed phase HPLC. The desired fraction was collected and
the solvent was evaporated, yielding 830 mg (33%) of 18b and
1
380 mg (15%) of 24. Data for 18b: mp 155 °C. H NMR (400
MHz, CDCl3) δ ppm 3.90 (s, 3 H), 4.09 (br s, 1 H), 4.22 (s, 2 H),
4.54 (s, 2 H), 6.42 (dd, J=8.1, 0.7 Hz, 1 H), 6.61 (td, J=7.3, 1.1
Hz, 1 H), 6.94-7.00 (m, 2 H), 7.30 (d, J=7.8 Hz, 1 H), 7.86 (d, J=
1.5 Hz, 1 H), 7.90 (dd, J=7.7, 1.8 Hz, 1 H). Anal. (C16H15NO2)
C, H, N. Data for 24: mp 137.9 °C. 1H NMR (360 MHz, CDCl3)
δ ppm 3.76 (2 H, s), 3.93 (3 H, s), 7.24-7.27 (2 H, m), 7.28-7.33
(1 H, m), 7.39 (1 H, d, J=8.1 Hz), 7.42 (1 H, d, J=7.7 Hz), 8.10 (1
H, dd, J=8.1, 1.8 Hz), 8.18 (1 H, d, J=1.5 Hz), 8.94 (1 H, s); 13
C
NMR (90 MHz, DMSO-d6) δ ppm 37.97 (t), 52.65 (q), 126.46
(d), 127.68 (d), 127.71 (d), 128.01 (d), 128.44 (s), 128.50 (d),

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 19 7019
MnO2 (521 mg, 6.0 mmol) in toluene (20 mL) was stirred at
80 °C for 3 h. Thereaction mixture wasfiltered over a dicalite pad
(eluent toluene, then CH2Cl2). The organic layer was concen-
trated, and the residue was purified by silica gel flash chroma-
tography (eluent CH2Cl2/CH3OH(NH3) 100/0 to 99/1). The
pure fractions were collected, and the solvent was evaporated.
The residue was purified further via reversed phase HPLC. The
product was dissolved in Et2O (10 mL), a 2 N HCl solution in
Et2O was added, and the resulting precipitate was filtered off and
dried, yielding 305 mg (53%) of 32 as a yellow solid; mp 205.5 °C.
1H NMR (360 MHz, DMSO-d6) δ ppm 3.96 (3 H, s), 7.27-7.33
(1 H, m), 7.35-7.41 (3 H, m), 7.44 (1 H, t, J=7.9 Hz), 7.84 (1 H,
dd, J=7.7, 1.5 Hz), 7.94 (1 H, dd, J=7.9, 1.6 Hz), 8.75 (1 H, s);
13C NMR free base (90 MHz, DMSO-d6) δ ppm52.91 (q), 122.26
(d), 124.33 (s), 125.84 (d), 126.63 (d), 128.69 (s), 128.93 (d),
129.40 (d), 134.35 (d), 134.69 (d), 140.68 (s), 152.17 (s), 157.74 (s),
160.72 (d), 165.30 (s); MS (EIþ) m/z (rel intensity) 253 (100%,
Mþ), 222 (56%), 194 (11%), 166 (14%); ESI-HRMS, found
254.0833 calculated for C15H12NO3 (M þ H)þ 254.0817. Anal.
Gibco) supplemented with 14.6 g/L L-glutamine (200 mM), 5 g/L
penicillin/streptomycin (5 ꢀ 10-6 IU/L), 5.5 g/L pyruvic acid,
5 mM HEPES (Gibco), 5 mL of insulin-transferrin-selenium-x
(Gibco), and 10% fetal calf serum (heat inactivated for 30 min at
56 °C). The cells were seeded in poly-D-lysine-coated 384-well
round-bottom polypropylene plates (Costar Corning, Data Pack-
aging) at 12 000 cells/well. Then 50 ng/mL tetracycline was added
to induce the hTRPA1 expression 24 h before the experiment. The
cells were loaded with 5 mg/L Fluo4-AM (Molecular Probes,
Invitrogen, Belgium) dissolved in HBSS seeding medium supple-
mented with 0.7 g/L probenecid (Sigma) and incubated for 1 h
at 37 °C and subsequently at 20 °C for 1-2 h. The fluorescence
was measured in the FDSS 6000 imaging based plate reader
(Hamamutsu Photonics K.K., Japan). The excitation wavelength
was 488 nm and the emission wavelength 540 nm. After a control
period of 12 s the TRPA1 agonists were added and their effect was
measured within 14 min after application. The mean and SEM of
four repeats at each concentration were plotted as a function of the
agonist concentration, and the concentration response data were
fitted to obtain the EC50 value.
(C15H11NO3 HCl) C, H, N.
3
Electrophysiologal Measurements. For patch clamp experi-
ments the cells were resuspended in culture medium without
Geneticin and blasticidin but with 500 ng/mL tetracycline to
induce the hTRPA1 expression. The cells were seeded in Nun-
clon Petri dishes (35 mm, Nunc, Denmark) at 30 000 cells/mL
and incubated in a humidified incubator at 37 °C with 5% CO2
for 2-5 days before the experiments. Whole cell patch clamp
recordings were performed as described8 with an EPC9 or
EPC10 amplifier (HEKA Electronik, Germany). Gigaseals were
obtained, and the series resistance was compensated for 90%.
The membrane potential was clamped at -30 mV throughout
the experiment, and fast solution changes were done by means of
a Warner SF-77B fast step superfusion system (Warner Instru-
ments LLC). Currents were sampled at 10 kHz and filtered at
1 kHz. Data acquisition was done with Pulse/Pulsefit (HEKA
Electronik, Germany). The extracellular solution contained (in
mM) 152 NaCl, 3 KCl, 10 glucose, 10 HEPES, 1 MgCl2, and
2 CaCl2, and the pH was adjusted to 7.4 with NaOH. The
intracellular solution contained (in mM) 130 CsCl,10 EGTA,10
HEPES, 2 MgATP, 0.2 Na3GTP, and 6.5 CaCl2, and the pH was
adjusted to 7.2 with CsOH. The free Ca2þ concentration of the
intracellular solution was 311 nM (calculated with Maxchelator,
free at http://www.stanford.edu/∼cpatton/maxc.html). A good
control of the intracellular Ca2þ concentration was necessary
Dibenzo[b,f ][1,4]oxazepine-4-carbonitrile (38). Step 1. 10,1-
Dihydrodibenzo[b,f ][1,4]oxazepine-4-carbonitrile (33a). A mix-
ture of 29a (750 mg, 2.7 mmol) and CuCN (608 mg, 6.8 mmol) in
DMF (15 mL) was degassed and then shaken under nitrogen
atmosphere at 140 °C overnight. The reaction mixture was
cooled to room temperature, and a 0.2 N aqueous sodium
hydroxide solution (200 mL) was added. The mixture was
extracted twice with ethyl acetate (2 ꢀ 100 mL). The combined
organic layers were washed with water and brine, dried over
MgSO4, concentrated in vacuo, yielding 255 mg (42%) of inter-
mediate 33a. 1H NMR (360 MHz, DMSO-d6) δ ppm 4.45 (d, J=
4.1 Hz, 2 H), 6.23 (t, J=4.0 Hz, 1 H), 6.60 (td, J=8.1, 1.5 Hz, 1
H), 6.65 (dd, J=8.1, 1.5 Hz, 1 H), 6.87 (td, J=7.7, 1.5 Hz, 1 H),
7.04 (dd, J=8.1, 1.5 Hz, 1 H), 7.30 (t, J=7.7 Hz, 1 H), 7.69 (dd,
J=7.7, 1.5 Hz, 1 H), 7.77 (dd, J=7.7, 1.5 Hz, 1 H).
Step 2. Dibenzo[b,f ][1,4]oxazepine-4-carbonitrile (38). Start-
ing from 33a, standard MnO2 oxidation gave compound 31 in
36% yield: mp 109.3 °C. 1H NMR (360 MHz, DMSO-d6) δ ppm
7.25 (dd, J=7.3, 1.8 Hz, 1 H), 7.34 (td, J=7.3, 1.8 Hz, 1 H),
7.39-7.44 (m, 2 H), 7.53 (t, J=7.7 Hz, 1 H), 7.95 (dd, J=7.7, 1.8
Hz, 1 H), 8.08 (dd, J=7.7, 1.8 Hz, 1 H), 8.71 (s, 1 H). Anal.
(C14H8N2O) C, H, N.
Dibenzo[b,f ][1,4]oxazepine-4-carboxylic Acid Amide (41). A
mixture of 38 (60 mg, 0.27 mmol) and concentrated H2SO4
(2 mL) was stirred at a temperature between 0 and 5 °C for 2 h.
The mixture was kept stirring at room temperature for 3 days.
The reaction mixture was poured into 100 mL of ice-water and
basified with a concentrated aqueous NH3 solution. The mix-
ture was extracted with CH2Cl2. The organic layer was washed
with brine, dried over MgSO4, filtered, and the filtrate was con-
centrated. Yield 64 mg (99%) of 41; mp 193 °C. 1H NMR (360
MHz, DMSO-d6) δ ppm 7.21-7.29 (m, 2 H), 7.31-7.38 (m, 3 H),
7.66 (dd, J=7.7, 1.8 Hz, 1 H), 7.69 (dd, J=7.7, 1.8 Hz, 1 H), 7.74
(br s, 1 H), 7.96 (br s, 1 H), 8.68 (s, 1 H). Anal. (C14H10N2O2)
C, H, N.
because of the direct activation of hTRPA1 channels by Ca2þ 22
.
At the end of the experiments 50 μM TRPA1 blocker ruthenium
red was applied to evaluate the non-TRPA1 component of the
holding current and as a consequence the quality of the experi-
ment. For electrophysiological experiments the stock solutions
of the compounds (10 or 100 mM) were further diluted in
DMSO. The final DMSO concentration in the extracellular
solution was 0.1%. Ruthenium red was dissolved at 50 μM in
extracellular solution containing 0.1% DMSO.
Acknowledgment. We thank Sebastiaan van Liempt for
carrying out the glutathione trapping experiments.
Biology. Cells and Culture. T-REx-293 cells (Invitrogen,
Belgium) were stably transfected with the tetracycline-inducible
human TRPA1 gene and cultured as previously described.8
Briefly, hTRPA1/T-REx-HEK293 cells (referred to as hTRPA1-
HEK cells in the text) were maintained under standard sterile
cell culture conditions. The culture medium for the hTRPA1-
HEK cells was Dulbecco’s modified Eagle’s medium (DMEM,
Gibco BRL, Invitrogen, Belgium) supplemented with 0.5 g/L
Geneticin (G148, Gibco), 5 mg/L blasticidin (Invitrogen), 14.6
g/L L-glutamine (200 mM, Gibco), 5 g/L penicillin/streptomycin
(5 ꢀ 10-6 IU/L, Gibco), 5.5 g/L pyruvic acid (Gibco), and 10%
fetal calf serum (Hyclone).
Supporting Information Available: Synthesis and analytical
data of all intermediates and compounds not described in the
Experimental Section, description of the electrophysiological
effect of 6 on the hTRPA1 current to illustrate how dose
response curves in Figure 1 have been obtained, and experi-
mental procedures on metabolic stability and reactive metabo-
lite formation experiments. This material is available free of
charge via the Internet at http://pubs.acs.org.
Ca2þ Fluorometry. For the fluorometric Ca2þ measurements
hTRPA1-HEK cells were resuspended in seeding medium:
Hanks’ balanced salt solution (HBSS with CaCl2 and MgCl2;
References
(1) Recent reviews: (a) Patapoutian, A.; Tate, S.; Woolf, C. J.
Transient receptor potential channels: targeting pain at the source.

7020 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 19
Gijsen et al.
Nat. Rev. Drug Discovery 2009, 8, 55–68. (b) Baraldi, P. G.; Preti, D.;
Materazzi, S.; Geppetti, P. Transient receptor potential ankyrin 1
(TRPA1) channel as emerging target for novel analgesics and anti-
inflammatory agents. J. Med. Chem. 2010, 53, 5085–5107.
Sipido, V. K. Synthesis and structure-activity relationship of 2-
(aminoalkyl)-2,3,3a,8-tetrahydrodibenzo[c,f ]isoxazolo[2,3-a]azepine
derivatives: a novel series of 5-HT(2A/2C) receptor antagonists. Part
2. Bioorg. Med. Chem. Lett. 2002, 12, 243–248.
(2) Viana, F.; Ferrer-Montiel, A. TRPA1 modulators in preclinical
development. Expert Opin. Ther. Pat. 2009, 19, 1787–1799.
(3) (a) Cai, X. A new tr(i)p to sense pain: TRPA1 channel as a target for
novel analgesics. Expert Rev. Neurother. 2008, 8, 1675–1681.
(b) Bang, S.; Hwang, S. W. Polymodal ligand sensitivity of TRPA1
and its modes of interactions. J. Gen. Physiol. 2009, 133, 257–262.
(4) (a) Bessac, B. F.; Jordt, S. E. Breathtaking TRP channels: TRPA1
and TRPV1 in airway chemosensation and reflex control. Physiol-
ogy 2008, 23, 360–370. (b) Caceres, A. I.; Brackmann, M.; Elia, M. D.;
Bessac, B. F.; del Camino, D.; D'Amours, M.; Witek, J. S.; Fanger,
C. M.; Chong, J. A.; Hayward, N. J.; Homer, R. J.; Cohn, L.; Huang, X.;
Moran, M. M.; Jordt, S. E. A sensory neuronal channel essential for
airway inflammation and hyperreactivity in asthma. Proc. Natl. Acad.
Sci. U.S.A. 2009, 106, 9099–9104.
(12) Hunziker, F.; Kunzle, F.; Schindler, O.; Schmutz, J. Seven-
membered heterocycles. III. Dibenzazepines and oxazepines, and
dibenzodiazepines and -thiazepines. Helv. Chim. Acta 1964, 47,
1163–1172.
(13) Brewster, K.; Clarke, R. J.; Harrison, J. M.; Inch, T. D.; Utley, D.
Preparation of the eight monohydroxydibenz[b,f ][1,4]oxazepin-
11(10H)-ones. J. Chem. Soc., Perkin Trans. 1 1976, 1286–1290.
(14) Liu, Y.; Lubin, M. L.; Reitz, T. L.; Wang, Y.; Colburn, R. W.;
Flores, C. M.; Qin, N. Molecular identification and functional
characterization of a temperature-sensitive transient receptor po-
tential channel (TRPM8) from canine. Eur. J. Pharmacol. 2006,
530, 23–32.
(15) hTRPV1 and cTRPM8 agonism was determined using a modifica-
tion of the methods described in the following: Behrendt, H.-J.;
Germann, T.; Gillen, C.; Hatt, H.; Jostock, R. Characterization of
the mouse cold menthol receptor TRPM8 and vanilloid receptor
type-1 VR1 using a fluorometric imaging plate reader (FLIPR)
assay. Br. J. Pharmacol. 2004, 141, 737–745.
(16) (a) Baillie, T. A. Metabolic activation and drug design: challenges
and opportunities in chemical toxicology. Chem. Res. Toxicol.
2006, 19, 889–893. (b) Ma, S.; Subramanian, R. Detecting and
characterizing reactive metabolites by liquid chromatography/tandem
mass spectrometry. J. Mass Spectrom. 2006, 41, 1121–1139.
(17) Wen, B.; Fitch, W. L. Screening and characterization of reac-
tive metabolites using glutathione ethyl ester in combination
with Q-trapmass spectrometry. J. Mass Spectrom. 2009, 44, 90–
100.
(18) Gijsen, H. J. M.; Berthelot, D. J.; De Cleyn, M. A. J. 3,4-
Dihydropyrimidine Derivatives as TRPA1 Antagonists and Their
Preparation, Pharmaceutical Compositions and Use in the Treat-
ment of Diseases. WO2009147079, 2009.
(19) Molina, P.; Lopez-Leonardo, C.; Llamas-Botia, J.; Foces-Foces,
C.; Fernandez-Castano, C. Unexpected Staudinger reaction of
[alpha]-azidoacetonitriles [alpha]-phenyl substituted with triphe-
nylphosphine. Preparation, X-ray crystal and molecular structures
of a phosphazine, an aminophosphonium carbanion salt and a
phosphazide, with (Z)-configuration. Tetrahedron 1996, 52, 9629–
9642.
(5) Everaerts, W.; Gevaert, T.; Nilius, B.; De Ridder, D. On the origin
of bladder sensing: Tr(i)ps in urology. Neurourol. Urodyn. 2008, 27,
264–273.
(6) (a) Bautista, D. M.; Movahed, P.; Hinman, A.; Axelsson, H. E.;
Sterner, O.; Hogestatt, E. D.; Julius, D.; Jordt, S. E.; Zygmunt,
P. M. Pungent products from garlic activate the sensory ion
channel TRPA1. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 12248–
12252. (b) Jordt, S. E.; Bautista, D. M.; Chuang, H. H.; McKemy, D. D.;
Zygmunt, P. M.; Hogestatt, E. D.; Meng, I. D.; Julius, D. Mustard oils
and cannabinoids excite sensory nerve fibres through the TRP channel
ANKTM1. Nature 2004, 427, 260–265. (c) Macpherson, L. J.;
Geierstanger, B. H.; Viswanath, V.; Bandell, M.; Eid, S. R.; Hwang,
S.; Patapoutian, A. The pungency of garlic: activation of TRPA1 and
TRPV1 in response to allicin. Curr. Biol. 2005, 15, 929–934.
(7) (a) Hinman, A.; Chuang, H. H.; Bautista, D. M.; Julius, D. TRP
channel activation by reversible covalent modification. Proc. Natl.
Acad. Sci. U S A 2006, 103, 19564–19568. (b) Macpherson, L. J.;
Dubin, A. E.; Evans, M. J.; Marr, F.; Schultz, P. G.; Cravatt, B. F.;
Patapoutian, A. Noxious compounds activate TRPA1 ion channels
through covalent modification of cysteines. Nature2007, 445, 541–545.
^
(8) Brone, B.; Peeters, P. J.; Marrannes, M.; Mercken, M.; Nuydens,
R.; Meert, T.; Gijsen, H. J. M. Tear gasses CN, CR, and CS are
potent activators of the human TRPA1 receptor. Toxicol. Appl.
Pharmacol. 2008, 231, 150–156.
(9) Blain, P. G. Tear gases and irritant incapacitants 1-chloroaceto-
phenone, 2-chlorobenzylidene malononitrile and dibenz[b,f ]-1,4-
oxazepine. Toxicol. Rev. 2003, 22, 103–110.
(10) Bessac, B. F.; Sivula, M.; von Hehn, C. A.; Caceres, A. I.; Escalera,
J.; Jordt, S.-E. Transient receptor potential ankyrin 1 antagonists
block the noxious effects of toxic industrial isocyanates and tear
gases. FASEB J. 2009, 23, 1102–1114.
(20) Wardrop, A. W. H.; Sainsbury, G. L.; Harrison, J. M.; Inch, T. D.
Preparation of some dibenz[b,f ][1,4]oxazepines and dibenz[b,e]-
azepines. J. Chem. Soc., Perkin Trans. 1 1976, 1279–1285.
(21) Fakhraian, H.; Nafary, Y.; Yarahmadi, A.; Hadj-Ghanbary, H
Improved etherification procedure for the preparation of dibenz[b,f ]-
[1,4]oxazepine. J. Heterocycl. Chem. 2008, 45, 1469–1472.
(22) Zurborg, S.; Yurgionas, B.; Jira, J. A.; Caspani, O.; Heppenstall,
(11) Andres, J. I.; Alcazar, J.; Alonso, J. M.; Diaz, A.; Fernandez, J.;
P. A. Direct activation of the ion channel TRPA1 by Ca2þ
.
Gil, P.; Iturrino, L.; Matesanz, E.; Meert, T. F.; Megens, A.;
Nat. Neurosci. 2007, 10, 277–279.