A potent photoreactive general anesthetic with novel binding site selectivity for GABAA receptors

Article abstract of DOI:10.1016/j.ejmech.2020.112261

The pentameric γ-aminobutyric acid type A receptors (GABA_ARs) are the major inhibitory ligand-gated ion channels in the central nervous system. They mediate diverse physiological functions, mutations in them are associated with mental disorders and they are the target of many drugs such as general anesthetics, anxiolytics and anti–convulsants. The five subunits of synaptic GABA_ARs are arranged around a central pore in the order β-α-β-α-γ. In the outer third of the transmembrane domain (TMD) drugs may bind to five homologous intersubunit binding sites. Etomidate binds between the pair of β – α subunit interfaces (designated as β⁺/α^–) and R–mTFD-MPAB binds to an α⁺/β^– and an γ⁺/β^– subunit interface (a β^– selective ligand). Ligands that bind selectively to other homologous sites have not been characterized. We have synthesized a novel photolabel, (2,6-diisopropyl-4-(3-(trifluoromethyl)-3H-diazirin-3-yl)phenyl)methanol or pTFD-di-iPr-BnOH). It is a potent general anesthetic that positively modulates agonist and benzodiazepine binding. It enhances GABA–induced currents, shifting the GABA concentration-response curve to lower concentrations. Photolabeling–protection studies show that it has negligible affinity for the etomidate sites and high affinity for only one of the two R–mTFD-MPAB sites. Exploratory site–directed mutagenesis studies confirm the latter conclusions and hint that pTFD-di-iPr-BnOH may bind between the α⁺/β^– and α⁺/γ^– subunits in the TMD, making it an α⁺ ligand. The latter α⁺/γ^– site has not previously been implicated in ligand binding. Thus, pTFD-di-iPr-BnOH is a promising new photolabel that may open up a new pharmacology for synaptic GABA_ARs.

Full text of DOI:10.1016/j.ejmech.2020.112261

Journal Pre-proof
A potent photoreactive general anesthetic with novel binding site selectivity for
GABA receptors
A
Abdelrahman R. Shalabi, Zhiyi Yu, Xiaojuan Zhou, Youssef Jounaidi, Hanwen Chen,
Jiajia Dai, Daniel E. Kent, Hua-Jun Feng, Stuart A. Forman, Jonathan B. Cohen,
Karol S. Bruzik, Keith W. Miller
PII:
S0223-5234(20)30228-2
DOI:
https://doi.org/10.1016/j.ejmech.2020.112261
EJMECH 112261
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 13 February 2020
Revised Date: 18 March 2020
Accepted Date: 19 March 2020
Please cite this article as: A.R. Shalabi, Z. Yu, X. Zhou, Y. Jounaidi, H. Chen, J. Dai, D.E. Kent, H.-J.
Feng, S.A. Forman, J.B. Cohen, K.S. Bruzik, K.W. Miller, A potent photoreactive general anesthetic with
novel binding site selectivity for GABA receptors, European Journal of Medicinal Chemistry (2020), doi:
A
https://doi.org/10.1016/j.ejmech.2020.112261.
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.
©
2020 Published by Elsevier Masson SAS.

A Potent Photoreactive General Anesthetic with Novel Binding Site Selectivity
for GABA Receptors.
A
d
b
a
a
a
Abdelrahman R. Shalabi , Zhiyi Yu , Xiaojuan Zhou , Youssef Jounaidi , Hanwen Chen ,
a
a,c
a
a
b
Jiajia Dai , Daniel E. Kent , Hua-Jun Feng , Stuart A. Forman , Jonathan B Cohen , Karol S.
d
a,*
Bruzik , Keith W. Miller .
Institutions
a
Department of Anesthesia, Critical Care and Pain Medicine, Massachusetts General Hospital,
Harvard Medical School, 32 Fruit Street, Boston, MA, 02114, USA
b
Department of Neurobiology, Harvard Medical School, 220 Longwood Avenue, Boston, MA,
0
2115, USA
c
Department of Health Science, Northeastern University, 360 Huntington Ave, Boston, MA,
2115, USA
0
d
Department of Medicinal Chemistry and Pharmacognosy, University of Illinois at Chicago, 833
South Wood Street, Chicago, IL, 60612, USA
*
Corresponding author:
Keith W Miller, k_miller@helix.mgh.harvard.edu
Department of Anesthesia, Critical Care & Pain Medicine (EDW 505),
Massachusetts General Hospital,
Blossom Street
Boston
MA 02114
USA
Present addresses
Zhiyi Yu: School of Pharmaceutical Sciences, Shandong, University, Jinan, China
zhiyi_yu@sdu.edu.cn
Hanwen Chen, Department of Anesthesiology, Shunde Hospital of Southern Medical
University, China
hanwen.chen@qq.com
Jiajia Dai, Department of Anesthesiology, Xiangya Hospital, Central South University, China
Jiajiadai123@hotmail.com
1

Highlights
•
•
•
•
•
pTFD-di-iPr-BnOH is a potent general anesthetic and enhances GABA’s actions
It photo-incorporates into intersubunit sites in the GABA(A)R’s transmembrane domain
+
–
+
–
Photo-protection shows it binds at either the γ /β or the α /β subunit interface
+
–
Similarly, it does not bind in the β /α subunit interfaces (etomidate sites)
+
–
Site–directed mutagenesis suggests the α /γ interface is another site
Graphical abstract
Abbreviations
BnOH, benzyl alcohol;
di-iPr-BnOH, (2,6-diisopropylphenyl)methanol;
ECD, extracellular domain;
nH, Hill coefficient;
pTFD-BnOH, (4-(3-(trifluoromethyl)-3H-diazirin-3-yl)phenyl)methanol;
pTFD-di-iPr-BnOH, (2,6-diisopropyl-4-(3-(trifluoromethyl)-3H-diazirin-3-yl)phenyl)methanol;
R–mTFD-MPAB, (5-allyl-1-methyl-5-(3-(3-(trifluoromethyl)-3H-diazirin-3-
yl)phenyl)pyrimidine-2,4,6(1H,3H,5H)-trione);
TBAF, tetra-n-butylammonium fluoride;
TMD, transmembrane domain;
2

ABSTRACT
The pentameric γ-aminobutyric acid type A receptors (GABA Rs) are the major inhibitory
A
ligand-gated ion channels in the central nervous system. They mediate diverse physiological
functions, mutations in them are associated with mental disorders and they are the target of many
drugs such as general anesthetics, anxiolytics and anti–convulsants. The five subunits of synaptic
GABA
A
Rs are arranged around a central pore in the order β-α-β-α-γ. In the outer third of the
transmembrane domain (TMD) drugs may bind to five homologous intersubunit binding sites.
+
–
Etomidate binds between the pair of
β
–
α
subunit interfaces (designated as β /α ) and R–mTFD-
+
–
+
–
–
MPAB binds to an
α
/β
and an
γ
/β
subunit interface (a
β
selective ligand). Ligands that bind
selectively to other homologous sites have not been characterized. We have synthesized a novel
photolabel, (2,6-diisopropyl-4-(3-(trifluoromethyl)-3H-diazirin-3-yl)phenyl)methanol or pTFD-
di-iPr-BnOH). It is a potent general anesthetic that positively modulates agonist and
benzodiazepine binding. It enhances GABA–induced currents, shifting the GABA concentration-
response curve to lower concentrations. Photolabeling–protection studies show that it has
negligible affinity for the etomidate sites and high affinity for only one of the two R–mTFD-
MPAB sites. Exploratory site–directed mutagenesis studies confirm the latter conclusions and
+
–
+
–
hint that pTFD-di-iPr-BnOH may bind between the
α
/β
and
site has not previously been implicated in ligand binding. Thus,
pTFD-di-iPr-BnOH is a promising new photolabel that may open up a new pharmacology for
α /γ subunits in the TMD, making
+
+
–
it an
α ligand. The latter α /γ
A
synaptic GABA Rs.
Keywords:
GABA-A Receptor
Transmembrane domain
3

Positive allosteric modulator
General Anesthetics Binding Sites
Photolabeling
Diazirine
4

1
. INTRODUCTION
The family of -aminobutyric acid type-A receptors (GABA Rs) are the major inhibitory
γ
A
ligand-gated ion channels in the central nervous system. They are located in both synaptic and
extrasynaptic neuronal membranes, respectively generating phasic and tonic currents. GABA R
A
activity affects many neurophysiological phenomena, including consciousness, learning and
memory, sedation, anxiety, epilepsy, and mood [1]. GABA R mutations are associated with
A
epilepsy, depression, Down’s syndrome, schizophrenia and autism [2-4]. They are also targets
for many neuromodulatory drugs such as general anesthetics, anticonvulsants and anxiolytics [5,
6
].
GABA R are members of the pentameric ligand-gated ion channel superfamily, composed of
A
five homologous subunits arranged pseudo–symmetrically around a central chloride conducting
pore (Figure 1). Mammals possess genes for 19 such homologous subunits and about two dozen
receptor isoforms have been identified. Typical synaptic receptors consist of
α, β, and γ subunits
arranged anti–clockwise in the order viewed from the extracellular space [7, 8]. The
α–γ–β–α–β
various isoforms occupy distinct locations in the CNS where they may mediate distinct behaviors
and pharmacological effects [9, 10]. This rich complexity has spurred researchers to identify
isoform–selective modulators that may target specific CNS functions (reviewed in [5]).
In typical synaptic receptors, GABA binds to a pair of nearly identical sites in the
+
–
extracellular domain (ECD) between the
β
and
site (Figure 1A; [7]). Derivatives that target specific
γ ECD subunit interfaces have been developed and have proved useful together with
α
subunits (
β
/α
sites). Benzodiazepines act at
+
–
the single homologous
α
/γ
α
isotypes at
+
–
α
(1-6) /
knock–in mice in implicating distinct
α
isotypes in the sedative, amnestic, and anxiolytic effects
5

of classical benzodiazepine [1, 11].
In the transmembrane domain (TMD), where anesthetics and their derivatives bind, selective
targeting of allosteric inter-subunit sites also occurs. Thus, etomidate and neurosteroid hypnotics
target the two
β+/α– interfaces at adjacent sites located near the extracellular and intracellular
ends of the TMD, respectively, whereas the mephobarbital derivative R–mTFD-MPAB (5-allyl-
5
-(3-(3-(trifluoromethyl)-3H-diazirin-3-yl)phenyl)pyrimidine-2,4,6(1H,3H,5H)-trione)
+
–
+
–
selectively targets both the
γ
/β
and the
α
/β
outer TMD interfaces (reviewed in [12]).
3
3
Photolabeling protection experiments using [ H]azietomidate and [ H] R–mTFD-MPAB have
been used to classify the selectivity of drugs for the outer TMD interfaces [13]. Notably,
propofol occupies all four TMD sites with little selectivity [14], whereas another propofol
analog, 4-benzoyl-propofol, has high selectivity between these sites, only having high affinity for
one of the two R–mTFD-MPAB interfaces [15]. Meanwhile, no drug has been found that binds
in the fifth homologous site in the α+/γ– outer TMD interface [16].
Thus, characterization of the pharmacology of the TMD as a drug target remains far from
complete [17]. We and others have explored photolabels related to propofol, but this has proved
difficult because the simplest derivatives are either relatively unstable, have limited efficacy as
A
GABA R modulators or have limits as photolabels [18-20]. Such chemical instability may be
related to the presence of the phenolic hydroxyl. We previously used a 4-(trifluoromethyl)-
diazirine derivative of benzyl alcohol to characterize an alkanol–binding cleft on luciferase [21].
Although that photolabel has low potency, we have now synthesized a new propofol derivative,
pTFD-di-iPr-BnOH, that is both a potent and an efficacious GABA R modulator. Preliminary
A
experiments indicate it displays novel selectivity for the outer TMD subunit interfaces on the
GABA R. It is also an effective and stable photolabel whose photoincorporation into GABA Rs
A
A
6

is enhanced by GABA.
7

2. RESULTS
2
.1. Synthesis of pTFD-di-iPr-BnOH.
The photoprobe 8 was synthesized as described below. Bromoaniline 1 was diazotized and
subjected to a Sandmeyer reaction with cuprous cyanide to yield the substituted benzonitrile 2.
The subsequent reduction of the cyano group with DIBAL and hydrolysis of the imine provided
the corresponding substituted benzaldehyde 3. Aldehyde 3 was reduced with sodium borohydride
into an alcohol 4 and protected with tert-butyldimethyl silyl chloride into the silyl ether 5.
Compound 5 was lithiated with n-BuLi and acylated using ethyl trifluoroacetate into the
corresponding substituted trifluoroacetophenone 6. The conversion of 6 into the aziridine was
performed using an earlier described sequence [22] including oximation, formation of oxime
tosylate, reaction with ammonia and the oxidation of the incipient diaziridine to afford the
corresponding diazirine 7 with satisfactory yield. Finally, the silyl group protection was removed
by the treatment with TBAF to provide the photoprobe 8.
i: (a) HCl/NaNO2, 0°C; (b) CuCN-KCN, 70°C; ii: DIBAL-H; iii: NaBH4; iv: (a) t-
butyldimethylsilyl chloride – 4-dimethylaminopyridine, 40°C; v: (a) n-BuLi, -78°C; (b) ethyl
trifluoroacetate, -78°C; vi: hydroxylamine-HCl, 80°C; vii: tosyl chloride, 4-
8

dimethylaminopyridine, diisopropylethylamine; viii: NH3/methanol; ix: I2, triethylamine; x:
3
tetra-n-butylammonium fluoride,; xi: Dess-Martin periodinane; xii: NaB H .
4
Scheme 1. Synthesis of propofol analog 8.
Synthesis of the tritiated 8 was carried out by the sequence of periodinane oxidation of
3
alcohol 8 into a corresponding aldehyde 9 and its reduction with NaB H
4
by Vitrax Corporation
3
into tritiated alcohol H-8 with specific radioactivity of 15 Ci/mmol.
2
.2. Anesthetic potency.
The fraction of tadpoles showing loss of righting reflexes (LoRR) increased with the
concentration of Di-iPr-BnOH ((2,6-diisopropylphenyl)benzyl alcohol) between 3 – 100
analysis yielded an EC50 of 16 ± 1.4 M; number of animals, N = 30. Further adding a 4-
trifluoromethydiazirine group to (2,6-diisopropylphenyl)benzyl alcohol gave pTFD-di-iPr-
BnOH, which produced LoRR in tadpoles with an EC50 of 2.5 ± 0.6 M (N = 73; concentration
range 0.5 – 10 µM), a 6–fold increase in potency. pTFD-di-iPr-BnOH had a low therapeutic
index in tadpoles with an LC50 of 3.6 ± 0.6 M. At 2, 3, and 10 μΜ, the fraction of tadpoles that
did not recover from anesthesia and were therefore excluded from the analysis for anesthesia was
μ
M, and
μ
μ
μ
4
/7, 6/9, and 7/10, respectively.
3
2
.3. Modulation of [ H]muscimol binding.
We next examined the ability of the benzyl alcohol derivatives to modulate the binding of
3
[
H]muscimol to α1β3γ2L receptors in native HEK cell plasma membranes. The base structure,
benzyl alcohol (BnOH) caused no modulation up to 50 mM. However, adding either two
isopropyl groups or a p-trifluorodiazirine group to BnOH conferred potency (Table 1) with the
di-isopropyl addition being particularly effective with an EC50 of 40 ± 6 μM (mean ± standard
deviation). Adding a pTFD group to Di-iPr-BnOH caused a further 6 fold increase in potency,
whereas adding the two 4,6-isopropyl groups to pTFD-BnOH increased potency nearly 100–fold
to 7.4 ± 0.7
μ
M (Figure 2A), a value that was both comparable to that of propofol and of
9

sufficient potency to encourage its development as a photolabel.
In 3 GABA Rs, pTFD-di-iPr-BnOH’s EC50 for modulating [ H]muscimol binding was
slightly lower than in
3
α1β
A
α1
β3γ
2 receptors at 2.9 ± 0.7
μM, showing that the switch from a γ2
subunit to a
β3 subunit in the pentamer does not attenuate its action. We compared modulation of
3
[
H]muscimol binding by pTFD-di-iPr-BnOH with that by propofol, a ligand that binds in both
the etomidate and R–mTFD-MPAB binding sites. Substituted cysteine modification protection
experiments show that M286C is protected from modification by propofol whereas A291C is
not [23]. We found that in 3(M286C), propofol’s EC50 was dramatically increased from 7.0
1.7 in WT to >500 M in the mutant, while leaving that of pTFD-di-iPr-BnOH unchanged at
.1 ± 0.4 M, an observation consistent with pTFD-di-iPr-BnOH not binding in the
interfaces. In
EC50s of pTFD-di-iPr-BnOH and propofol were only modestly increased to 13 ± 3 and 32 ± 10
M respectively. In all these experiments two independent concentration-response curves were
determined and the combined 18 points fitted to Equation 1 (see Materials and Methods).
β
α
α1β
±
μ
+
–
3
μ
β /α
+
–
α
1(A291C)β3 receptors that contain a mutation on α1 M3 in the α /β interface, the
μ
+
–
We also asked if pTFD-di-iPr-BnOH binds in the benzodiazepine site in the
α /γ interface in
3
the extracellular domain. To the contrary, pTFD-di-iPr-BnOH enhanced [ H]flunitrazepam
binding over the concentration range 0.03 to 300
μ
M with no inhibition evident at the highest
M, a Hill
H
, of 2.7 ± 1.0 and maximum modulation of 198 ± 4.4 %, comparable to modulation
concentration (Figure 2A). This effect was characterized by an EC50 of 11 ± 1
μ
coefficient, n
3
of [ H]muscimol binding (EC50 = 7.4 ± 0.6 µM, n = 1.9 ± 0.27 and maximal modulation of 260
H
3
±
46%). Similar Hill coefficients were observed in [ H]muscimol binding experiments using
α
1β3 receptors (n = 2.3 ± 0.76).
H
10

2
.4. Enhancement of GABA
Synaptic type 2L receptors were expressed in Xenopus oocytes and receptor activation
was measured using two-microelectrode voltage–clamp electrophysiology. Currents elicited by a
low concentration of GABA (3 M ~EC ) were enhanced in a concentration–dependent manner
Figure 2B) when GABA was co-applied with pTFD-di-iPr-BnOH, plateauing at 30 to 100
The concentration-response curve with the combined data from 5 oocytes was fitted to equation
A
receptor currents by pTFD-di-iPr-BnOH.
α1β3γ
μ
0
2
(
μM.
2
1
(see Materials and Methods) to yield a maximum enhancement of 30 ± 1 fold, an EC50 of 10 ±
M and a Hill coefficient of 2.0 ± 0.4. No surge currents were seen with pTFD-di-iPr-BnOH
μ
concentrations up to 100 µM.
2
.5. Leftward shift of the GABA concentration-response curve.
In the presence of a fixed hypnotic concentration of pTFD-di-iPr-BnOH (4 μM), EC50s
derived from GABA concentration-response curves in
α1β3γ2L receptors were reduced about 5–
fold, from 58 ± 5 M to 11 ± 2 M (Figure 2C). At 1 mM GABA, pTFD-di-iPr-BnOH enhanced
μ
μ
maximal GABA-activated currents by 1.13 ± 0.07 fold (n = 4). Thus, pTFD-di-iPr-BnOH’s is a
potent positive allosteric modulator, and its action resembles that of other potent intravenous
general anesthetics.
At high concentrations, most general anesthetics that modulate GABA receptors also directly
A
activate receptors in the absence of GABA. In contrast, application of 100 µM pTFD-di-iPr-
BnOH without GABA to oocytes expressing α1β3γ2L GABA receptors resulted in no
A
detectable current greater than the background noise (mean ± SD = 0.008 ± 0.013 pA), while 1
mM GABA elicited currents from every oocyte tested (3.2 ± 1.5 µA; n = 5).
11

2
.6. Effect of mutations in previously characterized general anesthetic binding pockets.
To ascertain if interfacial transmembrane sites on GABA receptors mediate modulation by
A
pTFD-di-iPr-BnOH, we introduced mutations that are known to reduce the actions of etomidate
or R–mTFD-MPAB (Figure 3A–F). Mutations were first introduced on the positive side of all
three subunits at the homologous M2-15´ positions (
Additionally, mutations were introduced on the negative side of
position (numbered from the conserved arginine at the start of M1;
α
1 S270I,
β
3 N265M, and
3 and 2 subunits at the M1-11´
3 M227W and 2 I242W) to
interfaces. Enhancement of currents activated
2L receptors containing each of the mutations
above was compared to that in wild type receptors in the presence of etomidate (5 M), R–
mTFD-MPAB (8 M) and pTFD-di-iPr-BnOH (10 M).
The anesthetic enhancement results for wild-type receptors is shown in Figure 3A. In the
etomidate sites, the M2-15´ mutation 3 N265M (Figure 3E) eliminated etomidate enhancement,
γ2 S280W).
β
γ
β
γ
+
–
+
–
differentiate effects mediated by
α /β and α /γ
with subsaturating GABA concentrations in
α1β3γ
μ
μ
μ
β
consistent with prior reports [24, 25], whereas modulation by both R–mTFD-MPAB and pTFD-
di-iPr-BnOH was greater than that in wild-type. In the two nonidentical R–mTFD-MPAB sites,
M2-15´ mutations
mTFD-MPAB modulation, as expected [26]. In
enhancement was comparable to that in wild-type, whereas enhancement by pTFD-di-iPr-BnOH
was reduced more than 50%. However, in 2(S280W) receptors, enhancement by etomidate
α
1S270I (Figure 3D) and
γ
2 S280W (Figure 3F) both significantly reduced R–
α
1(S270I) 2L receptors, etomidate
β3γ
α1β3γ
and pTFD-di-iPr-BnOH was maintained or increased.
To test whether mutations on the negative side of the
+
–
+
–
α
/
β
and
3(M227W) 2L receptors (Figure 3B),
modulation by R–mTFD-MPAB and pTFD-di-iPr-BnOH was much lower than in wild-type, but
α /γ interfaces were effective,
–
–
we introduced
β
and
γ
mutations at M1-11´. In
α
1β
γ
12

etomidate’s action was unaffected. Modulation of
and R–mTFD-MPAB was similar to wild-type, but that of pTFD-di-iPr-BnOH was strongly
attenuated (Figure 3C). These actions of etomidate and R–mTFD-MPAB on the 3 M227W and
2 I242W mutations agree with a prior report [16]. Thus, we can conclude that both
α1β3γ2(I242W) receptors by both etomidate
β
–
–
γ
β and γ
mutations at M1-11´ strongly reduce pTFD-di-iPr-BnOH’s enhancing action of GABA–induced
currents.
To more readily compare the pattern of mutant effects on different drugs, the enhancement
ratios with each drug were normalized to its mean wild-type value (Figure 4) and the results
analyzed using two-way ANOVA, which revealed highly significant effects of mutant [F(5, 130)
=
26.2; P < 0.0001] and drug [F(2,130) = 7.15; P = 0.0008] and factor interaction [F(10,130) =
3
3.2; P < 0.0001]. Normalized enhancement ratio patterns across the three M2-15´ mutations for
etomidate and R–mTFD-MPAB are very similar to recently reported allosteric shifts for the
same drugs and mutants [26]. We infer adjacent drug binding if a mutation reduces normalized
enhancement with a significance value of P < 0.01 adjusted for 5 wild-type vs. mutant
comparisons per drug (i.e. P < 0.01 = 0.05/5). The analysis of the M2-15´ mutations is consistent
+
–
with their established selectivity [12], with etomidate selectively binding in the two
β
/α
sites,
+
–
+
–
while R–mTFD-MPAB binds only in the
α /β and γ /β sites. The analysis also indicates that
+
–
+
–
pTFD-di-iPr-BnOH binds near 1S270, which contributes to the
α
α
/β
and α /γ transmembrane
anesthetic binding sites, but not elsewhere.
–
+
–
+
–
The
β
mutation
β3 M1 M227W, which abuts both
α
/β
and
γ
/
β
interfacial sites where R–
mTFD-MPAB binds, reduces modulation by both R–mTFD-MPAB and pTFD-di-iPr-BnOH.
–
Notably, the
γ
mutation,
γ2 M1 I242W, significantly weakened modulation by pTFD-di-iPr-
BnOH, but had no effect on modulation by either etomidate or R–mTFD-MPAB. These results
13

+
–
+
–
are consistent with pTFD-di-iPr-BnOH binding at both the
sites. Until now there has been no evidence that anesthetics can bind in the latter interface [12,
6], making pTFD-di-iPr-BnOH a possible first ligand to do so. Thus it could prove to be a
useful new addition to the photolabel arsenal for characterizing sites.
α /β and the α /γ intersubunit TMD
1
3
3
2
.7 Inhibition of [ H]azietomidate and [ H]R–mTFD-MPAB photolabeling.
In order to conduct photolabeling experiments, it is necessary to purify and reconstitute the
GABA Rs into asolectin:CHAPS (200
μM:5 mM) micelles. Under these conditions, pTFD-di-
A
3
iPr-BnOH still modulated [ H]muscimol binding strongly (216 ± 6% in micelles vs. 273 ± 14%
in native membranes). The EC50 in micelles was about double that in native membranes (Table
1
) at 13 ± 2
μM (3 independent concentration-response curves and 22 separate points between 1
and 300 M pTFD-di-iPr-BnOH).
μ
Competition of photolabeling was used to determine the relative potencies of the benzyl
3
alcohol anesthetics as inhibitors of photolabeling by [ H]azietomidate in the
α
subunit (primarily
3 M1
3
into
α
1 M1 M236 [27]) and by [ H]R–mTFD-MPAB in the
β
subunit (primarily into
β
A
M227 [13]) (Figures 5A & B respectively). GABA Rs were photolabeled in the presence of
3
GABA, and after photolabeling, subunits were resolved by SDS-PAGE, and H incorporation
was determined by liquid scintillation counting. None of the benzyl alcohols inhibited
3
[
H]azietomidate photolabeling in the α subunit at concentrations comparable to those that
modulate GABA R function. Di-iPr-BnOH was 3–4-fold more potent than either pTFD-BnOH
A
or pTFD-di-iPr-BnOH, but more than 10-fold less potent than propofol (Table 1). In contrast,
3
inhibition of [ H]R–mTFD-MPAB photoincorporation in the
β
subunit by pTFD-di-iPr-BnOH
was 10-fold more potent than di-iPr-BnOH or pTFD-BnOH, but comparable in potency to
propofol (Table 1). However, even at the highest concentrations, pTFD-di-iPr-BnOH only
14

3
inhibited [ H]R–mTFD-MPAB photolabeling of the
β
subunit by 60% in contrast to the ~90 %
M) itself (Figure 5B). Such
inhibition seen in the presence of excess R–mTFD-MPAB (60
μ
partial inhibition indicates that pTFD-di-iPr-BnOH binds with high affinity to only one of R–
–
mTFD-MPAB’s two
β
intersubunit anesthetic sites.
3
2
.8. [ H]pTFD-di-iPr-BnOH photolabeling of
α
1β3
A
γ2L GABA Rs.
3
We first compared photolabeling of GABA Rs by [ H]pTFD-di-iPr-BnOH in the resting and
A
desensitized states, that is in the presence of either the inverse agonist bicuculline or the agonist
3
GABA. GABA R subunits were separated by SDS-PAGE, and covalent incorporation of H into
A
receptor subunits was characterized by fluorography (Figure 6A) and by liquid scintillation
3
counting (Figure 6B). Based upon fluorography, H was incorporated primarily into the
β
subunit, and photolabeling was strongly enhanced in the presence of GABA compared to
bicuculline, showing that an allosteric site had been photolabeled.
We next examined the pharmacology of the pTFD-di-iPr-BnOH sites. In the presence of
3
GABA, photoincorporation of [ H]pTFD-di-iPr-BnOH into the
β
subunit was reduced to levels
similar to those in the presence of bicuculline by non-radioactive pTFD-di-iPr-BnOH, propofol
or R–mTFD-MPAB, whereas etomidate and allopregnanolone (3α, 5α–P) caused little inhibition
3
of photolabeling (Figure 6A). When H photoincorporation was quantified by liquid scintillation
counting (Figure 6B), R–mTFD-MPAB (60
reducing it to the level seen in the presence of bicuculline, whereas 100
di-iPr-BnOH reduced the photolabeling by 55%. Although a similar pattern of
photoincorporation is seen in the subunit, the pharmacologically specific photolabeling was 3–
μM) inhibited total
β
subunit photolabeling by 70%,
μ
M propofol and pTFD-
α
fold less than that in the
β
subunit rendering quantification less secure.
3
Based upon the radiochemical specific activity of [ H]pTFD-di-iPr-BnOH (15 Ci/mmol) and
15

A
the amount of GABA R loaded on the gel (4 pmol), the 3,900 cpm of R-mTFD-MPAB-
inhibitable
β
subunit photolabeling indicated specific photolabeling of ~6 % of GABA R
β
A
3
subunits, a photolabeling efficiency similar to that seen for [ H]R–mTFD-MPAB [22].
3
We determined the concentration dependence of inhibition of [ H]pTFD-di-iPr-BnOH
photoincorporation in the presence of GABA only in the GABA
enunciated above (Figure 6C). At higher concentrations (300 M) than used in Figure 6B, pTFD-
di-iPr-BnOH and propofol now each inhibited photolabeling to the same extent as the
background labeling, which was determined with a combination of 60 M R–mTFD-MPAB and
00 M etomidate and is denoted by the horizontal dashed line. The IC50s for pTFD-di-iPr-
A
R’s β subunit for reasons
μ
μ
3
μ
BnOH and propofol were18 ± 2 and 32 ± 6 μM respectively, and similar in each case to their
3
values for inhibition of [ H]R–mTFD-MPAB photolabeling (Table 1). R–mTFD-MPAB also
3
inhibited [ H]pTFD-di-iPr-BnOH photoincorporation with an IC50 value of 0.4 ± 0.1
similar to its affinity in the presence of GABA for its reported binding sites at the two
interfaces [13].
μ
M,
–
β
subunit
16

3
. DISCUSSION
3
.1 Does pTFD-di-iPr-BnOH bind in novel sites?
A
In the search for new allosteric sites in the TMD of the GABA R, we have synthesized a
novel photoactivatable benzyl alcohol agent. It is a potent general anesthetic and a positive
allosteric modulator of GABA Rs. After tritiation, it photo-incorporated in both the and the
subunits and that photoincorporation was positively modulated in the presence of GABA (Figure
α
β
A
6
). For technical reasons, we cannot rule out photoincorporation in the
γ subunit.
Three lines of evidence support the assertion that pTFD-di-iPr-BnOH does not bind in the
+
–
well–characterized etomidate site in the two
β
/α
subunit interfaces of the transmembrane
3
domain. First, pTFD-di-iPr-BnOH only protected against photolabeling by [ H]azietomidate
weakly and at high concentrations (Figure 5A). Second, site–directed mutagenesis of a residue in
+
–
3
the
β
/α
interface (
β
3 M3 M286C) attenuated propofol’s action on [ H]muscimol binding but
not that of pTFD-di-iPr-BnOH. Third, a mutation in
β3 M2 15’ (N265M), which strongly
attenuated etomidate’s ability to enhance currents elicited by subsaturating concentrations of
GABA [28], had no effect on pTFD-di-iPr-BnOH’s action (Figure 3 & 4).
+
–
+
–
Interaction with the two R–mTFD-MPAB sites, one in the
γ /β and the other in the α /β
interface, was more complex. Although, pTFD-di-iPr-BnOH protected against
3
photoincorporation by [ H]R–mTFD-MPAB with high affinity, it only protected against half of
3
the photolabeling (Figure 5B). Furthermore, R–mTFD-MPAB inhibited [ H]pTFD-di-iPr-BnOH
photoincorporation into the
β subunit to the same extent as non-radioactive pTFD-di-iPr-BnOH
(
Figure 6B and C). Taken together, these two results imply that pTFD-di-iPr-BnOH only
interacts with one of the two R–mTFD-MPAB sites, and that this site accounts for the
3
pharmacologically specific [ H]pTFD-di-iPr-BnOH photolabeling in the
β subunit. In the α
17

subunit a similar pharmacological pattern was observed, but photoincorporation was at less than
5% the efficiency of that in the subunit. Thus, direct identification of the amino acids
photolabeled in the subunit will be necessary to determine whether or not they contribute to an
2
β
α
R–mTFD-MPAB site. Our photolabeling results left two unresolved questions. First, which of
the two R–mTFD-MPAB sites does pTFD-di-iPr-BnOH interact with and, second, is there an
additional pTFD-di-iPr-BnOH binding site?
Because it was particularly important to assess whether pTFD-di-iPr-BnOH was a photolabel
that would reveal new sites, and therefore if it was worth pursuing in further detail, we conducted
a rapid survey using site–directed mutagenesis. Although mutagenesis studies have well known
limitations, there is a sufficient body of experience that, when used carefully, they can point
towards potential binding sites [16, 26]. We assumed that a robust criterion for identifying a
binding site was that a drug’s action was affected by adjacent mutations on both the positive and
the negative side of a given subunit interface. Applying this criterion, our observations suggest,
+
–
+
–
but do not prove, that pTFD-di-iPr-BnOH binds in the
α
/
β
and not the
γ
/β
interface. Thus, on
3 M1 M227W
+
–
the one hand, mutations on each side of the
α
/β
interface (
α
1 M2 S270I and β
respectively, see Figure1 for location of all residues mentioned here) both attenuate pTFD-di-iPr-
BnOH and R–mTFD-MPAB’s enhancing actions on GABA currents (Figures 3B & D; Figure 4).
+
–
On the other hand, a mutation in the
γ /β interface (γ2 M2 15’ S280W) that attenuated R–
mTFD-MPAB’s enhancing action was ineffective on pTFD-di-iPr-BnOH’s action (Figure 3F;
Figure 4).
Next, we tested the hypothesis that the second site is in the remaining homologous interface,
+
–
namely the
α /γ interface, which has formerly been termed the orphan (or “undruggable”) site
because photolabeling and substituted cysteine modification protection experiments have failed
18

to find ligands that binds in it [12, 16]. We placed a mutation on the negative side of this
interface ( 2 M1 I242W, which is homologous to 3 M1 M227W), and found that it attenuated
only the action of pTFD-di-iPr-BnOH and certainly not that of R–mTFD-MPAB (Figure 3C;
Figure 4). Considering that the 1 M2 15’ mutation on the other side of this interface also
attenuated pTFD-di-iPr-BnOH’s action (Figure 3D; Figure 4), our observations point to the
γ
β
α
+
–
second pTFD-di-iPr-BnOH site being in the
α /γ interface. Definitively answering these
questions will take much more detailed and time consuming experiments using photolabeling
and sequencing in purified reconstituted receptors and substituted cysteine modification
protection experiments in intact cells
+
–
If it is confirmed that pTFD-di-iPr-BnOH binds in the
α /γ interface, it would be the first
photolabel to bind in this, so–called, orphan interface. In that case a third type of action and
pharmacology in the transmembrane domain is added to those previously established. First,
+
pTFD-di-iPr-BnOH would be the first agent that binds in the pair of
α
interfaces, whereas R–
–
+
mTFD-MPAB binds in the two
β
interfaces and etomidate in the two
β
interfaces. Thus, pTFD-
di-iPr-BnOH promises to be an important new tool to aid the development of further subunit–
interface selective pharmacological agents.
3
.2. In vivo potency of benzyl alcohols and their sites of action
BnOH and p-TFD-BnOH caused loss of righting reflexes in tadpoles at some 20-times lower
3
concentrations than they modulated [ H]muscimol binding to GABA Rs in HEK cell
A
membranes (Table 1), suggesting that they may have other sites of action most likely on
glutamate receptors [29]. On the other hand, all the ligands with 2,6-isopropyl groups, had
comparable general anesthetic and GABA
A
R modulatory potencies. This is consistent with the
isopropyl group conferring a strong association with the latter receptors, a conclusion that is
19

backed by other studies [30, 31].
.3. Structure activity relationships of benzyl alcohols acting on GABA Rs
3
A
A
BnOH had no action on GABA Rs but addition of either a pTFD or two ortho isopropyl
groups conferred potency, with the latter substitution being some twenty fold more effective than
the former, again consistent with a special role for the isopropyl group. When these two
substitutions were combined the net increase in potency over BnOH was greater than one
thousand fold.
The photolabeling–protection data summarized in Table 1 make it possible to study the
structure–activity relationship in each of the homologous interfaces in the transmembrane
+
–
domain. At the etomidate site in the
β /α interface, these studies showed that any substitution on
BnOH’s aromatic ring conferred potency, but the most notable finding was that the addition of
two ortho isopropyl groups to pTFD-BnOH did not enhance potency, whereas in the absence of
the p-TFD substitution potency was moderately good. On the other hand, at the R–mTFD-MPAB
site neither of the above substitutions in BnOH produced a high affinity interaction, but addition
of a pTFD group to di-iPr-BnOH increased potency some tenfold, whereas the same addition at
the etomidate sites decreased potency some fivefold. This site was also slightly less sensitive to
the replacement of propofol’s phenolic hydroxyl with a –CH –OH group, which reduced potency
2
fivefold, half as much as at the etomidate site.
The overall effect of the differing apparent affinities for the etomidate and R–mTFD-MPAB
sites is that: (1) pTFD-BnOH binds unselectively to both sites with low affinity; (2) di-iPr-BnOH
favors the etomidate site 2–3-fold and has moderate affinity for it, and (3) pTFD-di-iPr-BnOH
has high affinity for one of the R–mTFD-MPAB sites but low affinity for the etomidate sites.
Thus, it is the combination of the two substitutions that confers binding selectivity for one of the
20

two R–mTFD-MPAB sites without having much influence at the etomidate sites.
The role of the pTFD- group in conferring selectivity to the R–mTFD-MPAB sites may be
compared to that in a series of 4-substituted propofols in a recent study [15]. Propofol itself (4-
H) and 4-Cl-propofol favored the etomidate site by 5.5- and 4-fold respectively, but the
following larger substitutions favored the R–mTFD-MPAB sites: Me–CO– by 1.5-fold; Ph–
C(OH)– by 5.3 fold, and t-Bu– by 8-fold.
Unlike pTFD-di-iPr-BnOH, none of the above compounds distinguished between the two R–
mTFD-MPAB sites but it has been reported that 4-benzoylpropofol does so [15]. Its low affinity
R–mTFD-MPAB site was favored over the etomidate sites by only 3-fold, whereas its high
affinity R–mTFD-MPAB site was favored over the etomidate site by 200-fold, about ten-fold
more than pTFD-di-iPr-BnOH, suggesting that the rigidity of this derivative of propofol confers
an advantage.
21

2
7 Jan 2020
4
. CONCLUSIONS
The aim of this study was to devise a new anesthetic photolabel that would identify novel
binding sites on GABA receptors. Our findings show that pTFD-di-iPr-BnOH fulfills this goal.
A
It is a positive allosteric modulator of synaptic GABA Rs that at hypnotic concentrations does
A
+
–
not bind to the etomidate sites in two
β
/α
interfaces, but does interact with one of the two R–
sites and with another site. This additional site is most likely in the
α /γ interface, for which there is currently no known ligand. Thus, pTFD-di-iPr-
+
–
+
–
mTFD-MPAB
α /β and γ /β
+
–
homologous
BnOH promises to be a useful new tool for characterizing the pharmacology of the five
homologous intersubunit drug binding sites in the outer third of the transmembrane domain of
GABA Rs. It will complement the two well–established tools, azietomidate and R–mTFD-
A
MPAB.
22

5
. MATERIALS AND METHODS
.1. Materials.
-Bromo-2,6-diisopropylaniline and anhydrous grade solvents used in synthesis were from
5
4
Aldrich, and were not further dried or purified. Common chemicals, etomidate, asolectin, FLAG
peptide and polyethyleneimine were from Sigma. Buffer chemicals, CHAPS and DDM were
3
from Fisher–Anatrace. pTFD-BnOH was obtained from TCI America. R–mTFD-MPAB, [ H]R–
3
mTFD-MPAB (38 Ci/mmol, 26
μ
M in ethanol, and [ H]azietomidate (19 Ci/mmol, 53
μ
M in
3
ethanol ) were synthesized and tritiated previously [22, 32]. [ H]Muscimol and
3
[
H]flunitrazepam were from Perkin Elmer (Cat. # NET 574 250UC and NET 567250UC
respectively).
The human GABA
A
Rs used for the biochemical and photolabeling studies described herein
2L or 3 had the composition N-FLAG– 2L–C–(GGS)3GK–
3 respectively and were expressed in tetracycline-inducible HEK293 cells
as previously described [33, 34]. They were used as native membranes or after solubilization and
purification on a FLAG antibody column, from which they were eluted in micelles of 200
asolectin and 5 mM CHAPS with 100 g/mL (~100 M) FLAG peptide. Membranes and
and designated as
α1
β3
γ
α
1
β
α1β3γ
1
D4 or N-FLAG–α1β
μM
μ
μ
reconstituted receptors were stored at (–80°) until needed. In electrophysiological studies, the
human subunits lacked the purification tags (see below).
5
.2. Analytical Chemistry.
Analytical Chemistry. The methods used were as previously described [35]. H, C and F
1
13
19
NMR spectra were recorded on a Bruker AVANCE spectrometer at 400 MHz, 100 MHz and 376
MHz, respectively, unless otherwise noted. The NMR chemical shifts were referenced indirectly
23

1
13
19
3
to TMS for H and C, and to CFCl for F NMR. High resolution mass spectrometry was
performed with a Q-TOF-2TM (Micromass). TLC was performed using Merck 60 F254 silica
gel plates. Purity of the final compounds was assessed by HPLC analysis with a Synergy Hydro-
RP column (4
% methanol to 99% methanol over 32 min, followed by isocratic elution. Elution was
monitored by UV at 254 nm. These HPLC analyses indicated purity > 96%.
-Bromo-2,6-diisopropylbenzonitrile (2). Hydrochloric acid (37%, 1.5 mL) was added at
μm, 4.60 x 150 mm) using a methanol and methanol-water gradient running from
1
4
room temperature to a stirred suspension of 4-bromo-2,6-diisopropylaniline (1) (1.57 g, 6.13
mmol) in H O (5 mL), and the mixture was allowed to stir at room temperature for 30 min. A
2
solution of sodium nitrite (0.31 g, 4.49 mmol) in H O (1 mL) was added in a dropwise manner at
2
0
°C (ice-bath), and the mixture was allowed to stir at 0°C for 30 min. The mixture was
neutralized by saturated aqueous solution of NaHCO and added in portions at 70°C to a stirred
solution of CuCN (0.43 g, 4.80 mmol) and KCN (0.65 g, 9.98 mmol) in H O (1 mL). The
3
2
reaction mixture was allowed to stir at 70°C for 30 min, cooled to room temperature, and
extracted with toluene (15 mL). The organic layer was washed with water (5 mL), brine (5 mL),
dried with Na SO , and evaporated to dryness under reduced pressure to yield a crude product
2
4
which was purified by flash column chromatography (silica gel, hexane-ethyl acetate, 97:3) to
1
afford 0.72 g (44%) of 4-bromo-2,6-diisopropylbenzonitrile (2) as a brown solid. H NMR
1
3
(
CDCl3):
CDCl ):
δ
δ
1.32 (d,
J
= 6.8 Hz, 12H, CH ), 3.35−3.45 (m, 2H, CH), 7.37 (s, 2H, ArH). C NMR
3
(
3
154.66, 128.33, 126.84, 116.46, 110.24, 32.61, 23.15.
4-Bromo-2,6-diisopropylbenzaldehyde (3). Diisobutylaluminum hydride (1M in toluene,
6
.5 mL) was added in a dropwise manner at 0°C (ice-bath) to a stirred solution of 4-bromo-2,6-
diisopropylbenzonitrile (2) (0.734 g, 2.76 mmol) in CH Cl (1 mL) under an argon atmosphere.
2
2
24

The ice bath was removed, and the reaction mixture was allowed to stir at room temperature
overnight, cooled to 0°C (ice-bath), and quenched by the addition of HCl (1.5M, 5 mL). The
reaction mixture was heated at reflux for 30 min, cooled to room temperature, and extracted with
CH Cl (3 x 5 mL). The combined organic portion was washed with brine (5 mL), dried with
2
2
Na SO , and concentrated to dryness under reduced pressure to yield a crude product which was
2
4
purified by flash column chromatography (silica gel, hexane-ethyl acetate, 97:3) to afford 0.631
1
g (85%) of 4-bromo-2,6-diisopropylbenzaldehyde (3) as a brown oil. H NMR (CDCl
3
):
δ
1.28
(d, J = 6.8 Hz, 12H, CH
3
), 3.48−3.58 (m, 2H, CH), 7.40 (s, 2H, ArH), 10.64 (s, 1H, CHO).
Compound 3 was previously synthesized from 5-bromo-2-iodo-1,3-diisopropylbenzene through
a Bouveault reaction [36]; however, the reported route was not followed to avoid lithiation of the
aromatic bromo group by
n-butyllithium used during the reaction.
(
4-Bromo-2,6-diisopropylphenyl)methanol (4). A solution of NaBH (0.243 g, 6.42 mmol)
4
in EtOH (5 mL) was added to 4-bromo-2,6-diisopropylbenzaldehyde (3) (0.631 g, 2.34 mmol)
under an argon atmosphere, and the reaction mixture was allowed to stir at room temperature for
3
h, cooled to 0°C (ice-bath), and quenched by the addition of HCl (1.5M, 5 mL). The solvent
was evaporated under reduced pressure. Water (10 mL) was added, and the mixture was
extracted with ethyl acetate (3 x 5 mL). The combined organic portion was washed with brine (5
mL), dried with Na SO , and concentrated to dryness under reduced pressure to yield a crude
2
4
product which was purified by flash column chromatography (silica gel, hexane-ethyl acetate,
6
0:40) to afford 0.496 g (78%) of 4-bromo-2,6-diisopropylbenzyl alcohol (4) as a yellowish
1
white solid. H NMR (CDCl
3
):
δ
1.28 (d,
J
= 6.7 Hz, 12H, CH
3
), 3.33−3.40 (m, 2H, CH), 4.77 (s,
1
3
2
2
H, CH OH), 7.31 (s, 2H, ArH). C NMR (CDCl ):
δ
150.59, 132.96, 126.55, 123.22, 57.03,
2
3
9.37, 24.34.
25

((4-Bromo-2,6-diisopropylbenzyl)oxy)(tert-butyl)dimethylsilane (5). A solution of 4-
bromo-2,6-diisopropylbenzyl alcohol (4) (0.496 g, 1.83 mmol), imidazole (0.200 g, 2.94 mmol),
and tert-butyldimethylsilyl chloride (0.400 g, 2.65 mmol) in CH Cl (5 mL) was allowed to stir
2
2
at room temperature under argon atmosphere for 40 h. 4-Dimethylaminopyridine (0.400 g, 3.27
mmol) was added, and the reaction mixture was stirred at room temperature for 24 h, heated at
reflux for 22 h, cooled to room temperature, washed with water (5 mL), brine (5 mL), dried with
2 4
Na SO , and concentrated to dryness under reduced pressure to yield a crude product which was
purified by flash column chromatography (silica gel, hexane-ethyl acetate, 97:3) to afford 0.570
g (86%) of ((4-bromo-2,6-diisopropylbenzyl)oxy)(tert-butyl)dimethylsilane (5) as a colorless oil.
1
H NMR (CDCl3):
δ
0.15 (s, 6H, CH ), 0.93 (s, 9H, CH ), 1.26 (d,
J
= 6.8 Hz, 12H, CH3),
150.51,
3
3
1
3
3
1
.28−3.38 (m, 2H, CH), 4.72 (s, 2H, CH O), 7.28 (s, 2H, ArH). C NMR (CDCl ):
δ
2
3
33.20, 126.19, 122.58, 57.15, 29.26, 25.89, 24.22, 18.28, -5.35.
1-(4-(((tert-Butyldimethylsilyl)oxy)methyl)-3,5-diisopropylphenyl)-2,2,2-trifluoroethan-
1
-one (6). -Butyllithium (1.6M in hexane, 1.6 mL) was added in a dropwise manner over 10
n
min at -78°C into a stirred solution of ((4-bromo-2,6-diisopropylbenzyl)oxy)(tert
-
butyl)dimethylsilane (5) (0.476 g, 1.24 mmol) in THF (5 mL) under an argon atmosphere, and
the reaction mixture was allowed to stir at -78°C for 2 h. A solution of ethyl trifluoroacetate
(
0.358 g, 2.52 mmol) in THF (1.5 mL) was added in a dropwise manner over 10 min at -78°C,
the reaction mixture was allowed to stir at -78°C for 1 h and quenched by the addition of
saturated aqueous solution of NaHCO (5 mL). Water (5 mL) was added, and the mixture was
extracted with ethyl acetate (3 x 10 mL). The combined organic portion was washed with brine
2 x 10 mL), dried with Na SO , and concentrated to dryness under reduced pressure to yield a
3
(
2
4
crude product which was purified by flash column chromatography (silica gel, hexane-ethyl
26

acetate, 95:5) to afford 0.330 g (66%) of 1-(4-(((tert-butyldimethylsilyl)oxy)methyl)-3,5-
1
diisopropylphenyl)-2,2,2-trifluoroethan-1-one (6) as a pale yellow oil. H NMR (CDCl ):
δ
0.17
3
(
s, 6H, CH ), 0.94 (s, 9H, CH ), 1.31 (d,
J
= 6.9 Hz, 12H, CH ), 3.36−3.47 (m, 2H, CH), 4.81 (s,
3
3
3
1
3
2
1
H, CH O), 7.88 (s, 2H, ArH). C NMR (CDCl ):
δ
180.60 (q, J = 34.3 Hz), 149.33, 142.17,
2
3
1
9
29.55, 124.92, 116.87 (q, J = 289.5 Hz), 57.21, 29.33, 25.82, 24.15, 18.26, -5.41. F NMR
(
CDCl3):
-(4-(((tert-Butyldimethylsilyl)oxy)methyl)-3,5-diisopropylphenyl)-3-(trifluoromethyl)-
-diazirine (7). Pyridine (99.8%, 1.5 mL) was added at room temperature to a stirred solution
δ -71.02.
3
3
H
of the substituted acetophenone 6 (0.330 g, 0.82 mmol) in EtOH (3 mL). Solid hydroxylamine
hydrochloride (0.103 g, 1.48 mmol) was added, and the reaction mixture was heated at 80°C for
4
h and cooled to room temperature. Water (30 mL) was added, and the mixture was extracted
with ethyl acetate (3 x 30 mL). The combined organic portion was washed with brine (3 x 30
mL), dried with Na SO , and concentrated to dryness under reduced pressure to yield a crude
product which was purified by flash column chromatography (silica gel; CH Cl ) to afford 0.267
g of an oxime as a white solid which was subjected to further steps. Tosyl chloride (0.206 g. 1.08
mmol) was added at 0°C (ice-bath) to a stirred solution of oxime (0.267 g, 0.64 mmol),
2
4
2
2
N,N-
diisopropylethylamine (0.151 g, 1.17 mmol), and 4-dimethylaminopyridine (0.011 g, 0.09 mmol)
in CH Cl (5 mL). The ice bath was removed, and the reaction mixture was allowed to stir at
2
2
room temperature overnight. Water (20 mL) was added, and the mixture was extracted with
CH
2
Cl
2
(3 x 30 mL). The combined organic portion was washed with brine (25 mL), dried with
Na
2
SO
4
, and concentrated under reduced pressure to yield a crude product which was purified by
flash column chromatography (silica gel, hexane-ethyl acetate, 50:50) to afford the
corresponding tosylate (0.359 g) as a colorless oil. Ammonia solution (7M in methanol, 3 mL)
27

was added at -78°C into a stirred solution of the tosylate (0.359 g, 0.63 mmol) in THF (4 mL)
under argon. The reaction mixture was stirred at room temperature for 24 h and concentrated
under reduced pressure. Dichloromethane (5 mL) was added, and the mixture was filtered, and
the collected solid was washed with CH Cl . The filtrate was concentrated under reduced
2
2
pressure to yield the corresponding diaziridine (0.250 g) as an orange oil. Triethylamine (99%,
.25 mL) was added to a stirred solution of diaziridine (0.250 g, 0.60 mmol) in methanol (5 mL).
0
Iodine powder was added in portions until the color of iodine disappeared. The reaction mixture
was allowed to stir at room temperature for 1 h. An aqueous solution of sodium thiosulfate
2
solution (10% in H O) was added in a dropwise manner until the color of iodine was discharged.
The solvent was evaporated under reduced pressure. Water (10 mL) was added, and the mixture
was extracted with CH Cl (3 x 15 mL). The combined organic portion was washed with brine (5
2
2
mL), dried with Na SO , and concentrated to dryness under reduced pressure to yield a crude
2
4
product which was purified by flash column chromatography (silica gel, hexane-ethyl acetate,
9
6:4) to afford diazirine 7 (0.177 g, 52%) as a colorless oil.
1
H NMR (CDCl3):
δ
0.15 (s, 6H, CH ), 0.93 (s, 9H, CH ), 1.25 (d,
J
= 6.8 Hz, 12H, CH3),
148.95,
35.85, 128.62, 123.65, 120.97, 57.12, 29.23, 25.88, 24.18, 18.30, -5.38. F NMR (CDCl ):
3
3
1
3
3
1
.31−3.41 (m, 2H, CH), 4.74 (s, 2H, CH O), 6.95 (s, 2H, ArH). C NMR (CDCl ): δ
2 3
1
9
δ
3
−
65.04.
-(4-(((tert-Butyldimethylsilyl)oxy)methyl)-3,5-diisopropylphenyl)-3-(trifluoromethyl)-
H-diazirine (8). Tetrabutylammonium fluoride solution (1M in THF, 0.73 mL) was added to a
3
3
stirred solution of silyldiazirine 7 (0.177 g, 0.43 mmol) in THF (4 mL), and the reaction mixture
was allowed to stir at room temperature for 1 h. The mixture was concentrated under reduced
pressure and a crude product which was purified by flash column chromatography (silica gel,
28

1
hexane-ethyl acetate, 94:6) to afford 0.093 g (73%) of desilylated alcohol 8 as a white solid. H
NMR (CDCl3): 1.28 (d, = 6.8 Hz, 12H, CH ), 3.35−3.45 (m, 2H, CH), 4.81 (d, = 4.6 Hz,
H, CH OH), 6.99 (s, 2H, ArH). C NMR (CDCl ): 149.10, 135.60, 129.24, 123.58, 121.31,
−65.09.
-diazirin-3-yl)benzaldehyde (9). Dess-Martin
, 0.5 mL) was added to a stirred solution of alcohol 11 (0.026 g,
.09 mmol) in CH Cl (1.5 mL), and the reaction mixture was allowed to stir at room
δ
J
J
3
1
3
2
1
δ
2
3
1
9
20.86, 56.96, 29.37, 24.30. F NMR (CDCl ):
δ
3
2
,6-Diisopropyl-4-(3-(trifluoromethyl)-3
H
periodinane (0.3 M in CH Cl
2
2
0
2
2
temperature for 1 h and diluted with ethyl ether (12 mL). A solution of Na S O (0.5 M in
2
2
3
saturated aqueous NaHCO , 3 mL) was added. The mixture was stirred at room temperature until
3
a clear biphasic solution was obtained. The organic portion was washed with saturated aqueous
solution of NaHCO (5 mL), brine (5 mL), dried with Na SO , and concentrated to dryness under
3
2
4
1
reduced pressure to yield 0.019 g (75%) of aldehyde 9 as a pale yellow oil. H NMR (CDCl
.27 (d, = 6.8 Hz, 12H, CH
), 3.44−3.51 (m, 2H, CH), 7.04 (s, 2H, ArH), 10.68 (s, 1H, CHO).
C NMR (CDCl ): 195.14, 150.12, 134.11, 132.72, 121.46, 121.02, 120.90, 29.00, 23.90. F
NMR (CDCl3): −64.87.
3
): δ
1
J
3
1
3
19
3
δ
δ
Tritiation procedure for alcohol 8. A solution of NaBH (0.003 g, 0.079 mmol) in EtOH (1
4
mL) was added to aldehyde 9 (0.004 g, 0.013 mmol) under argon atmosphere, and the reaction
mixture was allowed to stir at room temperature for 4 h, quenched by the careful addition of HCl
(1M, 2.5 mL), and extracted with ethyl ether (6 mL). The organic portion was dried with
Na SO , and concentrated under reduced pressure to yield a crude product which was purified by
2
4
flash column chromatography (silica gel, hexane – ethyl acetate, 97:3 v/v) to afford alcohol 8
0.003 g, 100%). Aldehyde 9 undergoes rapid addition of water to form a hydrate which is
resistant to reduction by sodium borohydride. To ensure successful tritiation, aldehyde 9 was
(
29