Facile and practical synthesis of a cannabinoid-1 antagonist via regio- and stereoselective ring-opening of an aziridinium ion

Article abstract of DOI:10.1016/j.tet.2009.09.054

A scalable synthetic strategy of a chiral, trisubstituted imidazolidinone (1), a novel cannabinoid-1 antagonist, starting from a commercially available mandelic acid (5) is described. The key step involves a regio- and stereoselective ring-opening of an a

Full text of DOI:10.1016/j.tet.2009.09.054

Tetrahedron 65 (2009) 9067–9074
Contents lists available at ScienceDirect
Tetrahedron
journal homepage: www.elsevier.com/locate/tet
Facile and practical synthesis of a cannabinoid-1 antagonist via
regio- and stereoselective ring-opening of an aziridinium ion
*
*
Edwin B. Villhauer , Wen-Chung Shieh , Zhengming Du, Kevin Vargas, Lech Ciszewski,
Yansong Lu, Michael Girgis, Melissa Lin, Mahavir Prashad
Chemical and Analytical Development, Novartis Pharmaceuticals Corporation, East Hanover, NJ 07936, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
A scalable synthetic strategy of a chiral, trisubstituted imidazolidinone (1), a novel cannabinoid-1
antagonist, starting from a commercially available mandelic acid (5) is described. The key step involves
a regio- and stereoselective ring-opening of an aziridinium ion by an aniline nucleophile (3).
A mechanistic study revealed the insight into rate amplification at a lower temperature for vicinal
diamine 12 formation via a aziridinium ion 14. Although most intermediates are not isolable by
crystallization due to their intrinsic physical properties (oil or foamy solid), the reported synthesis
furnished pure 1 without any chromatography purification throughout the entire synthesis. Employing
green chemistry principles, this novel synthesis appears to be highly efficient for the manufacturing of
multi-kilogram quantities of an optically-pure active pharmaceutical ingredient.
Received 29 July 2009
Received in revised form 10 September
2009
Accepted 11 September 2009
Available online 17 September 2009
Keywords:
Aziridinium ion
Imidazolidinone
Regioselective
Stereoselective
Vicinal diamine
Ó 2009 Elsevier Ltd. All rights reserved.
1. Introduction
2. Results and discussion
Imidazolidinone
1
is
a
structurally novel cannabinoid-1
2.1. Synthesis of
b-amino alcohol
antagonist,1a as compared to two leading analogs: Rimonabant1b
and Taranabant.1c It penetrates the blood–brain barrier and is
orally active with potential to treat obesity and diabetes. To assess
its efficacy and safety in human, we needed to develop a com-
mercially viable route for the manufacturing of this active phar-
maceutical ingredient for clinical trials. Our strategy for the
synthesis of 1 is outlined in Scheme 1, which involves an attempt
to assemble the target molecule by the condensation of vicinal
diamine 2 with a ‘carbonyl’ surrogate. To establish the stereogenic
center of 2, we considered ring-opening of a chiral, quaternary
aziridinium ion 4 with aniline 3 in a regio- and stereoselective
manner. Aziridinium salt 4 can be synthesized from a commer-
Our synthesis of a
b
-amino alcohol (10) that led to the desired
aziridinium salt (4) is described in Scheme 2. (S)-3-Trifluoro-
methylmandelic acid 5 was converted to (S)-O-acetylmandelic acid
7 according to a modified literature procedure.2 Treating 5 with
1.5 equiv of acetyl chloride in toluene at ambient temperature
cleanly furnished 7 in quantitative yield. Traces of acetyl chloride
were removed under reduced pressure until the chloride content
was determined to be less than 1 wt % as measured by a classical
titration of 7 against a 0.1 N silver nitrate solution. This quality
control was essential to minimize the formation of an undesired
acetamide from 8 with acetyl chloride in the subsequent step. As
a result, chromatography purification was avoided. To set up the
formation of a quaternary aziridinium salt a few steps later, sulfonyl
acid 63 was converted to secondary amine 8 in two steps by first
coupling with 4-methoxybenzyl amine followed by selective amide
reduction with BH3-THF. The resulting amine 8 was isolated as
a hydrochloric salt in 67% yield. Suppressing racemization during
the amide formation involving mandelic acid 7 and amine 8
hydrochloric salt turned out to be a challenging task. Employing
a variety of coupling reagents, such as DMTMM,4 CDMT,5 and IBCF,6
that had demonstrated to be effective in our lab in controlling
cially available mandelic acid 5 and sulfonyl acid 6 via a b-amino
alcohol scaffold.
* Corresponding authors.
E-mail address: wen.shieh@novartis.com (W.-C. Shieh).
0040-4020/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2009.09.054

9068
E.B. Villhauer et al. / Tetrahedron 65 (2009) 9067–9074
O
O
O
O
O
O
O
H
H
S
S
CH3
CH3
Cl
N
N
Cl
N
N
CF3
CF3
1
2
-X
O
O
R
N+
OH
S
CH3
O
O
O
O
O
S
HO
OH
CH3
&
+
NH2
Cl
CF3
CF3
3
4
5
6
Scheme 1. Retrosynthetic analysis of 1.
OH
OAc
O
O
OAc
OH
OH
O
O
O
CF3
CF3
S
N
CH3
7
5
CF3
O
O
HCl
HN
S
O
CH3
O
O
OMe
S
9
HO
CH3
6
OMe
8
OH
O
O
S
N
CH3
CF3
OMe
10
Scheme 2. Synthesis of chiral
b-amino alcohol 10.
racemization for amide syntheses, we observed significant race-
mization in the range of 8–30% (Table 1, entries 1–5). DMTMM was
found to be more effective when used at lower temperature. High
vulnerability to racemization was presumably caused by the
acid-catalyzed formation of benzylic cation could also contribute to
the racemization. Based on our previous studies to suppress ra-
cemization, we observed that increasing the rate of coupling re-
action between an ‘activated ester’ and an amine reduces
racemization.6 We noticed that a potential equilibrium between
salt 8 and free base 8a in the system could diminish the avail-
ability of 8a, which is responsible for the attack to the activated
ester 11a. A lower concentration of 8a at any given time could
ultimately contribute to a slower coupling rate and higher race-
mization. This hypothesis led us to replace HCl salt 8 with free
base 8a for the amide synthesis employing DMTMM and did
minimize racemization from 8 to 1.5% (Fig. 2). Free base 8a can be
generated either through neutralization-extraction procedure or
in situ with Cs2CO3 in the presence of small amounts of water (see
Section 4).
enhanced acidity of the benzylic
a-CH proton for the respective
activated ester, such as 11a (Fig. 1), formed in situ. Alternatively,
Table 1
Degree of racemization for amide formation involving 7, 8, and NMM
Entry
Coupling Reagent
Solvent
Temperature, ꢀC
%Racemization
1
2
3
4
5
IBCFa
EtOAc
EtOAc
toluene
toluene
EtOAc
25
25
25
0
ꢁ20
16
16
30
13
8
CDMTb
DMTMMc
DMTMM
DMTMM
Amide 9 obtained from DMTMM-activated coupling of 7 and 8a
was carried forward to next step without purification or isolation.
Treating amide 9 with BH3-THF selectively reduced the amide and
a
IBCF: isobutyl chloroformate.
CDMT: 2-chloro-4,6-dimethoxytriazine.
DMTMM: 4-(4,6-dimethoxy-1,3,5-triazine-2-yl)-4-methyl-morpholinium chloride.
b
c

E.B. Villhauer et al. / Tetrahedron 65 (2009) 9067–9074
9069
CH3
N+
Cl-
MeO
N
N
N
O
OAc
MeO
O
N
OMe
DMTMM
9
7
O
O
O
N
N
S
NMM-HCl
HN
+
CH3
CF3
OMe
NMM-HCl
11a
OMe
8a
OAc
O
HCl
HN
O
S
O
N
OMe
CH3
O
N
N
NMM
+
CF3
OMe
OMe
11b
8
Figure 1. Cause and control of racemization.
suggested that these two reaction protocols must follow distinct
pathways. This prompted a more detailed analysis of these reactions.
As revealed by our 1H NMR studies in toluene-d8, a detectable
intermediate formed rapidly in 0.5 h when 10 was treated with
MsCl at 0 ꢀC (Fig. 4, NMR-B). The identity of this intermediate was
postulated to be aziridinium salt 14 based on a reported instability
and short lifetime of a benzylic mesylate.12 Over a period of 24 h at
rt, aziridinium salt 12 converted to another intermediate exclu-
sively (Fig. 4, NMR-D), which was isolated and fully characterized as
OAc
O
8a (free base)
8
(HCl salt)
DMTMM, NMM, -20 °C
8% racemization
OH
9
9
DMTMM, -20 °C
1.5% racemization
CF3
7
8a made in situ with CsCO3
DMTMM, -5 °C
2% racemization
9
b-chloro amine 15 (NMR, and elemental analyses). Chloride attack
of aziridinium salt 12 was selective for the benzylic over the
unsubstituted position, leading to benzylic chloride 15, which was
supported by similarities in 13C and 1H NMR chemical shifts of
C6H5CHCl between pure 15 (59.0 and 4.91 ppm, respectively, in
CDCl3) and an analogous molecule reported by Sharpless.11b Based
on these outcomes, a plausible mechanism to explain the kinetic
difference is shown in Scheme 3. If aziridinium salt 14 is exposed to
aniline 3 shortly after its formation, diamine 12 can be generated by
Figure 2. Effect of the amine form on degree of racemization.
acetate groups without reducing the sulfone or causing racemization.7
The purity of alcohol 10 obtained after aqueous sodium hydroxide
work up was acceptable to be used for the next step without further
purification.
a SN2 reaction via pathway 1. In the absence of 3, b-chloro amine 15
2.2. Regio-and stereoselective opening of aziridinium ion
is obtained via pathway 2, and is relatively unreactive to the sub-
sequent addition of 3. Regeneration of aziridinium ion 14 from 15
took place slowly over a period of 5 days via intra-molecular SN2
reaction. Ions 14 reacted with 3 to furnish diamine 12 exclusively,
which also supported the absolute configuration assignment of 15
since nucleophilic substitutions with double inversion of the con-
figuration can only contribute to the stereochemistry outcome of
Syntheses of vicinal diamines (such as 12, Scheme 3) through
aziridinium ions (such as 14) have been well-studied.8 Typically, an
aziridinium ion can be prepared by treating a b-amino alcohol with
mesyl chloride in the presence of Et3N. Regio-selectivity for the
opening of an aziridinium ion to produce a vicinal diamine is de-
pendent on the type of nucleophile and the nature of the aziridinium
ion.9 It has been reported that nucleophilic attacks on phenyl azir-
idinium ions at the more-hindered benzylic site by aliphatic amines
can be accomplished in high yields.9a,9c,10 Substitutions of azir-
idinium ions with aniline nucleophiles have also been reported.11
Herein, we disclose our own observations and insights into this type
of reaction. Although stable aziridinium salts have been isolated,9a
12. It is worth mentioning that (1) b-chloro amine 15 was generated
through the opening of aziridinium ion 14 with chloride, instead of
through the replacement of the mesylate;11,13 (2) 15 (a benzyl
chloride) was inert toward nucleophilic attack by aniline 3; and (3)
a regio-selective and stereo-specific nucleophilic opening of azir-
idinium ion 14 by 3 is accountable for the chiral vicinal diamine 12
synthesis.
our attempts to isolate aziridinium salt 14 after treating b-amino
alcohol 10 with mesyl chloride (MsCl) in the presence of Et3N were
unsuccessful. Therefore, a one-pot approach to synthesize vicinal
diamine 12 directly was investigated. We observed a temperature
effect on the rate of 12 formation (Fig. 3). Treating alcohol 10 with
MsCl at rt for 16 h and adding aniline 3 to the resulting intermediate,
vicinal diamine 12 was generated with high regio- and stereo-
selectivity (98:2 er) in 80% after 5 days. Under these conditions,
diamine 12 was the only regio-isomer that we observed according to
1H NMR analyses of the crude product. Interestingly, when the same
mesylation was carried out at 0 ꢀC for 30 min, followed by the ad-
dition of 3, the desired 12 was produced in >95% within 16 h. The
facts that formation of 12 had been faster at a lower temperature
2.3. Completion of the total synthesis
Our initial approach to remove p-methoxybenzyl group from 12
(Scheme 4) with neat TFA resulted in an quantitative conversion to
diamine 16 as TFA salt. From a large-scale preparation prospect,
employing large excess of volatile TFA (bp 72 ꢀC) is unsafe and
ecologically-unfriendly. An alternative and greener process was
developed for the deprotection utilizing stoichiometric amounts of
methanesulfonic acid (b. p. 167 ꢀC/10 mm) at elevated temperature
(100 ꢀC) in toluene. Diamine 16, obtained after base work up, was
treated with p-nitrobenzoic acid and furnished a crystalline salt 17

9070
E.B. Villhauer et al. / Tetrahedron 65 (2009) 9067–9074
H
OH
(S)
H
OMs
N
O
MsCl
Et3N
O
O
O
S
N
CH3
S
CH3
CF3
-
Et3N+Cl
CF3
OMe
OMe
10
13
SN2
-
OMs
pathway 1
MeO
O
H
SN2
N
X = OMs, Cl
H
O
Cl
O
O
-
NH2
O
O
X
(S)
S
Cl
N+
N
3
CH3
S
H
CH3
CF3
12
OMe
CF3
14a, X = OMs
14b, X = Cl
desired enantiomer
pathway 2
slow
SN2
-
Cl
H
O
O
Cl
H
N
NH2
H
O
O
Cl
Cl
O
O
3
S
(R)
N
CH3
S
N
CH3
CF3
CF3
inactive
OMe
OMe
15
12b
undesired enantiomer
Scheme 3. Plausible pathway for the formation of chiral vicinal diamine 12.
O
OH
Cl
NH
O
O
S
O
O
1) MsCl, Et3N, 25 °C, 16 h
1) MsCl, Et3N, 0 °C, 30 min
N
CH3
12
S
N
CH3
2)
2)
O
CF3
O
NH2
NH2
CF3
Cl
Cl
3
3
OMe
5 days, 80% conversion
10
16 h, >95% conversion
OMe
12
Figure 3. Impact of temperature on the rate of vicinal diamine 12 formation.
with high chemical purity (HPLC assay 98.1%). It is worth men-
tioning that salt 17 is the only crystalline intermediate of this
synthetic route starting from mandelic acid 5. The rest of the in-
termediates once generated were used for the next step without
further purification or isolation. As a result, the varieties and
amounts of solvents used were significantly minimized, which
follows the principles of green chemistry. This telescoped approach
achieved an overall yield of 34% from 5 to 17 (6 steps), which
demonstrated the effectiveness of our synthetic strategy. To
assemble the final trisubstituted imidazolidinone 1, salt 17 was first
converted to free base 16 with aqueous NaOH and then reacted
with triphosgene. The condensation from 16 took 15 min at rt and
provided 1 in quantitative yield. However, triphosgene is not an
acceptable choice for manufacturing purpose due to its toxicity.
Therefore, a safer alternative using carbonyl diimidazole as the
‘carbonyl’ surrogate was developed, affording 1 as a crystalline solid
in good yield (67%) with high purity (HPLC assay 99%). The enan-
tiomeric ratio of 1 was determined to be 96:4 (S:R) according to

E.B. Villhauer et al. / Tetrahedron 65 (2009) 9067–9074
9071
Figure 4. Conversion of b-amino alcohol 10 to b
-chloro amine 15 in toluene-d8 as monitored by 1H NMR.
a chiral HPLC analysis. The undesired (R)-enantiomer of 1 was
generated during mandelic acid 5 and amide 9 syntheses, with
about 2% from each individual steps. The possible conglomerate
nature14 of compound 1 contributed to the isolation of high en-
antiomeric purity 1 (HPLC 99.8:0.2 er) by a simple recrystallization
from ethanol.
reported synthesis furnished pure 1 without any chromatography
purification throughout the entire synthesis. As a result, conven-
tional approaches to purify or isolate intermediates using solvent
systems were significantly circumvented. This green and novel
synthesis appeared to be highly effective for the manufacturing of
multi-kilograms of a chemically- and optically-pure pharmaceu-
tical ingredient for clinical trials.
3. Conclusion
4. Experimental section
4.1. General
In conclusion, starting from a commercially available mandelic
acid 5, a scalable synthetic strategy leading to a chiral tri-
substituted imidazolidinone (1), a novel cannabinoid-1 antagonist,
is described. The key step involves a regio- and stereoselective
ring-opening of an aziridinium ion by an aniline nucleophile
affording vicinal diamine 12. A mechanistic study suggested two
plausible pathways that are accountable for the kinetic difference
of vicinal diamine formation under two different conditions. Al-
though most intermediates are not isolable by crystallization due
to their intrinsic physical properties (oil or foamy solid), the
All solvents and reagents were purchased from commercial
sources and used without further purification. NMR spectra were
recorded on either 400 or 500 MHz spectrometers. 1H and 13C NMR
chemical shifts are reported as d using residual solvent as an internal
standard. HPLC analyses were performed on a Waters, Varian, or
Rainin HPLC System connected to a PDA UV detector over 210–
400 nm or a UV detector at 254 nm.

9072
E.B. Villhauer et al. / Tetrahedron 65 (2009) 9067–9074
O
O
Cl
NH
Cl
NH
O
O
O
O
S
N
CH3
S
N
H
CH3
CF3
12
CF3
16
OMe
O
O
O
O
S
O
Cl
NH
CH3
[
16
]
Cl
N
N
O
O
N+
S
CH3
H
H
O
CF3
O
CF3
NO2
17
1
Scheme 4. End-game synthesis.
4.2. (S)-2-Acetoxy-2-(3-(trifluoromethyl)phenyl)acetic acid (7)
dioxane was added over a period of 45 min and stirred for an ad-
ditional 3 h. To the mixture, THF (1 L) was added. Any solids were
collected by filtration, rinsed with THF (2ꢂ500 mL), and dried at
50 ꢀC under vacuum for 16 h to obtain 8 (79.3 g, 73%) as a white
To a 4-L, three-necked, round bottom flask was charged (S)-3-
(trifluoromethyl)-mandelic acid (5) (120 g, 0.55 mol, 98:2 er), tol-
uene (960 mL), and acetyl chloride (66 g, 0.84 mol). The mixture
was stirred at rt for 16 h. The mixture was distilled at 40–45 ꢀC
under vacuum until the acetyl chloride content was determined to
be <1 wt % as measured by titration with 0.1 N AgNO3 solution. The
residue was diluted with toluene to a final volume of 400 mL to
obtain 7 as a solution, which was used for the next step without
solid: HPLC assay 99%; 1H NMR (500 MHz, DMSO-d6)
d 9.92 (br s,
1H), 7.53 (d, J¼8.5 Hz, 2H), 6.98 (d, J¼8.5 Hz, 2H), 4.11 (t, J¼5.0 Hz,
2H), 3.77 (s, 3H), 3.68 (m, 2H), 3.29 (m, 2H), 3.11 (s, 3H); 13C NMR
(125 MHz, DMSO-d6)
d 159.6, 131.6, 123.5, 113.9, 55.1, 49.5, 49.2,
40.7, 39.4. HRMS calcd for C11H17NO3S [MþH]þ 244.1007, found
244.0988.
further purification: 1H NMR (500 MHz, CDCl3)
(m, 1H), 7.40 (m, 1H), 7.12 (m, 1H), 7.03 (m, 1H), 5.89 (s, 1H), 2.10 (s,
3H); 13C NMR (125 MHz, CDCl3)
173.4, 170.2, 134.1, 131.4, 130.9,
d 7.62 (s, 1H), 7.54
4.5. (S)-2-((4-Methoxybenzyl)(2-(methylsulfonyl)-
ethyl)amino)-2-oxo-1-(3-(trifluoromethyl)phenyl)
ethyl acetate (9)
d
129.4, 126.2, 124.3, 123.9, 73.3, 20.5. HRMS calcd for C11H9F3O4
[MꢁH]ꢁ 261.0375, found 261.0367.
To a 4-L, three-necked, round bottom flask was charged 8
(100 g, 0.36 mol), THF (1.2 L), water (63 g), and cesium carbonate
(70 g, 0.22 mol). The mixture was stirred at rt for 1.5 h. To the
resulting mixture, a solution of 7 (0.38 mol) in toluene (400 mL),
prepared from the previously step, was added. The mixture was
stirred at rt for 30 min. In portions, 4-(4,6-dimethoxy-1,3,5-tri-
azin-2-yl)-4-methyl-morpholinium chloride (DMTMM, 109 g,
0.39 mol) was added. The mixture was stirred at ꢁ5.0 ꢀC for 16 h.
To the mixture was added 8% aqueous citric acid (800 mL) and THF
was evaporated at 45 ꢀC under vacuum. Toluene (730 mL) was
added to the residue and the aqueous layer was separated. The
organic layer was washed with 8% aqueous citric acid (1ꢂ500 mL),
5% aqueous NaHCO3 (1ꢂ500 mL), water (1ꢂ500 mL), and 15%
aqueous NaCl (1ꢂ500 mL). The organic phase was evaporated at
45 ꢀC under vacuum until the water content was determined to be
<0.15% as measured by Karl Fisher titration. The resulting solution
of 9 in toluene was used for the next step without further puri-
fication: chiral HPLC assay 96:4 er (2% from racemization, 2% from
4.3. (2-Methanesulfonyl)-N-(4-methoxy-benzyl)-acetamide
To a 100-L, three-necked reactor was charged sulfonyl acid 6
(1.22 kg, 8.83 mol), THF (36 L), and 4-methoxybenzyl amine (1.21 kg,
8.83 mol). To the slurry, was added 2-chloro-4,6-dimethoxy-
[1,3,5]triazine (1.7 kg, 9.71 mol), followed by N-methylmorpholine
(1.1 kg, 9.71 mol) over a period of 30 min. The mixture was stirred at
rt for an additional 2 h. To the cloudy solution,1 N NaOH (36.6 L) was
added over a period of 1 h and stirred for an additional 1 h. The
mixture was evaporated under vacuum to remove most of THF until
a suspension was obtained. Any solids were collected by filtration,
rinsed with water (4ꢂ5 L), and dried at 50 ꢀC under vacuum for 16 h
to give the title compound as a white solid (2.09 kg, 92%): 1H NMR
(500 MHz, DMSO-d6)
d
8.72 (t, J¼5.7 Hz, 1H), 7.21 (d, J¼8.5 Hz, 2H),
6.89 (d, J¼8.5 Hz, 2H), 4.26 (d, J¼5.7 Hz, 2H), 4.10 (s, 2H), 3.73 (s, 3H),
3.12 (s, 3H); 13C NMR (125 MHz, DMSO-d6)
d 161.8,158.3,130.3,128.6,
113.7, 59.4, 54.9, 41.8, 41.4. HRMS calcd for C11H15NO4S [MþH]þ
5); 1H NMR (500 MHz, CDCl3)
d
7.59 (d, J¼7.9 Hz, 1H), 7.53 (m, 2H),
258.0800, found 258.0811.
7.45 (t, J¼7.9 Hz, 1H), 6.98 (d, J¼8.8 Hz, 2H), 6.77 (d, J¼8.5 Hz, 2H),
6.18 (s, 1H), 4.55 (d, J¼16.4 Hz, 1H), 4.37 (d, J¼16.1 Hz, 1H), 3.72 (s,
3H), 3.68 (m, 1H), 3.66 (m, 1H), 3.19 (m, 1H), 3.06 (m, 1H), 2.82 (s,
4.4. (2-Methanesulfonyl-ethyl)-(4-methoxy-benzyl)-amine (8)
To a 5-L, three-necked, round bottom flask was charged
3H); 13C NMR (125 MHz, CDCl3)
d 170.6, 168.3, 159.6, 134.5, 131.8,
(2-methanesulfonyl)-N-(4-methoxy-benzyl)-acetamide
(100 g,
131.4, 129.6, 128.3, 126.6, 126.3, 125.3, 123.6, 114.5, 72.7, 55.2, 51.8,
51.5, 41.4, 41.1, 20.6. HRMS calcd for C22H24F3NO6S [MþH]þ
487.1355, found 487.1362. Chiral Technologies ChiralpakÔ AD,
388 mmol) and THF (330 mL). A solution of 1 N BH3-THF (777 mL,
777 mmol) was added over a period of 1 h. After the addition, the
mixture was allowed to warm to 50 ꢀC and stirred for an additional
7 h. The mixture was cooled to rt and stirred for 16 h. Methanol
(95 mL) was added and stirred for 5 min. A solution of 4 N HCl in
5
mm
250 mmꢂ4.6 mm, flow rate¼1 mL/min, 20 ꢀC, isocratic,
hexane–isopropanol¼75:25, UV 230 nm: (S)-enantiomer, tR¼
9.6 min; (R)-enantiomer, tR¼12.1 min.

E.B. Villhauer et al. / Tetrahedron 65 (2009) 9067–9074
9073
4.6. (S)-2-((4-Methoxybenzyl)(2-(methylsulfonyl)ethyl)-
4.8. [(S)-2-Chloro-2-(3-trifluoromethyl-phenyl)-ethyl]-(2-
amino)-1-(3-(trifluoromethyl)phenyl)ethanol (10)
methanesulfonyl-ethyl)-(4-methoxy-benzyl)-amine (15)
To a 1-L, three-necked, round bottom flask was charged a so-
lution of 9 (42 g, 86.4 mmol) in toluene (234 mL). A solution of 1 N
BH3-THF (216 mL, 216 mmol) was added over a period of 1 h,
maintaining the batch temperature below 22 ꢀC. After the addi-
tion, the mixture was stirred at ambient temperature for an ad-
ditional 20 h. The mixture was cooled to 10 ꢀC and treated with
3 N NaOH (130 mL) over a period of 1 h (Note: the first 8 mL of 3 N
NaOH addition caused significant foaming and should be added
very slowly). The mixture was allowed to warm to 50 ꢀC over
a period of 40 min and stirred for an additional 1 h. The mixture
was cooled to rt and extracted with toluene (35 mL). The organic
phase was separated, washed with 25% NaCl (3ꢂ130 mL), and
evaporated at 35–38 ꢀC under vacuum until a final volume of
230 mL was reached. The resulting solution of 10 in toluene was
used for the next step without further purification: HPLC assay
A 100-mL, four-necked, round bottom flask equipped with
a mechanical stirrer, and a thermocouple was charged 10
(4.31 g, 10.0 mmol), isopropyl acetate (27 mL), and triethyl-
amine (2.53 g, 25.0 mmol). The solution was cooled to ꢁ5 ꢀC
and methanesulfonyl chloride (1.37 g, 12.0 mmol) was added,
maintaining the batch temperature below ꢁ5 ꢀC. After the ad-
dition, the mixture was stirred at 0 ꢀC for an additional 30 min,
allowed to warm to 22 ꢀC, and stirred for an additional 16 h.
The mixture was quenched with water (9.0 mL). The organic
phase was separated, washed with water (9 mL), and evapo-
rated at 38 ꢀC under vacuum to dryness to obtain crude 15
(4.08 g) as a light yellow oil. Crude 15 (1.5 g) was purified by
column chromatography (silica gel, 40% ethyl acetate in hep-
tanes) to afford 15 (1.16 g) as a colorless oil: 1H NMR (400 MHz,
CDCl3)
d
7.62–7.47 (m, 4H), 7.10 (d, J¼8.6 Hz, 2H), 6.86 (d,
84%; 1H NMR (400 MHz, CDCl3)
d
7.58 (s, 1H), 7.55–7.41 (m, 3H),
J¼8.6 Hz, 2H), 4.91 (t, J¼7.3 Hz, 1H), 3.84 (s, 3H), 3.70 (d,
J¼13.1 Hz, 1H), 3.58 (d, J¼13.1 Hz, 1H), 3.17–3.08 (m, 6H), 2.82
7.23 (d, J¼9.6 Hz, 2H), 6.88 (d, J¼9.6 Hz, 2H), 4.75 (dd, J¼10.0,
3.2 Hz, 1H), 3.89 (br s, 1H), 3.82 (s, 3H), 3.81 (d, J¼12.9 Hz, 1H),
3.58 (d, J¼13.0 Hz, 1H), 3.25 (m, 1H), 3.20 (m, 1H), 3.10 (m, 1H),
3.05 (m, 1H), 2.85 (s, 3H), 2.73 (dd, J¼12.8, 3.2 Hz, 1H), 2.60 (dd,
(s, 3H); 13C NMR (125 MHz, CDCl3)
d 159.3, 141.0, 131.0, 130.9,
130.2, 129.2, 129.1, 125.3, 124.4, 123.8, 114.0, 62.8, 59.6, 59.0,
55.3, 52.7, 48.8, 41.8. Anal. Calcd for C20H23ClF3NO3S: C, 53.31;
H, 5.15; Cl, 7.88; N, 3.11; S, 7.13. Found: C, 53.40; H, 5.28; Cl,
7.93; N, 2.95; S, 7.10. HPLC for 15 (tR¼9.01 min): Waters Sym-
J¼12.8, 9.6 Hz, 1H); 13C NMR (125 MHz, CDCl3)
d 159.3, 143.0,
130.7, 130.4, 129.3, 129.2, 128.8, 124.3, 124.2, 122.7, 114.1, 69.9, 62.7,
58.6, 55.3, 52.3, 47.6, 41.7. HRMS calcd for C20H24F3NO4S [MþH]þ
metry C18 5
m
m 150 mmꢂ3.9 mm, flow rate¼1 mL/min, 22 ꢀC,
432.1456, found 432.1442. HPLC for
9
(tR¼8.7 min), 10
UV 230 nm, gradient elution from 10:90 A:B to 80:20 A:B over
7 min, held for 3 min; A¼0.1% TFA in acetonitrile; B¼0.1% TFA in
water.
(tR¼6.8 min): Waters Symmetry C18
5
mm 150 mmꢂ3.9 mm, flow
rate¼1.0 mL/min, 25 ꢀC, gradient elution from 90:10 A:B to 20:80
A:B over 7 min then held for an additional 3 min; A¼0.1% TFA in
water; B¼0.1% TFA in CH3CN.
4.9. (S)-N1-(4-(4-Chlorophenoxy)phenyl)-N2-(2-(methane-
sulfonyl)ethyl)-1-(3-(trifluoromethyl) phenyl)ethane-
1,2-diamine p-nitrobenzoate (17)
4.7. (S)-N1-(4-(4-Chlorophenoxy)phenyl)-N2-(4-methoxy-
benzyl)-N2-(2-(methylsulfonyl)ethyl)-1-(3-(trifluoro-
methyl)phenyl)ethane-1,2-diamine (12)
To a 2-L, three-necked, round bottom flask was charged a so-
lution of 12 (68 g, 108 mmol) in toluene (600 mL) and meth-
anesulfonic acid (125 g, 1.3 mol). The mixture was allowed to
warm to 100 ꢀC and stirred for an additional 2 h. The mixture was
cooled to 0 ꢀC and 7 N NaOH (200 mL) was added followed by 25%
NaCl (200 mL). The organic layer was separated, washed with 25%
NaCl (2ꢂ250 mL), and evaporated until the final volume of
210 mL was reached to obtain a solution of 16 in toluene. A so-
lution of p-nitrobenzoic acid (18 g, 108 mmol) in isopropanol
(180 mL) was then added. The mixture was allowed to warm to
75 ꢀC and stirred for an additional 30 min. Heptanes (975 mL) was
added over a period of 20 min, maintaining the batch tempera-
ture at 65 ꢀC. The mixture was cooled to 15 ꢀC over a period of 1 h
and stirred for an additional 1 h. The solids were collected by
filtration and rinsed with 7% ethanol in heptanes (250 mL). The
solids were dried under vacuum at 50 ꢀC for 16 h to give 17
(36.6 g, 34% from 5) as a yellow solid: mp 145–146 ꢀC; HPLC assay
To a 2-L, three-necked, round bottom flask was charged a so-
lution of 10 (72 g, 108 mmol) in toluene (83 mL) and toluene
(280 mL). The solution was cooled to 0 ꢀC and Et3N (27.4 g,
271 mmol) was added. Methanesulfonyl chloride (14.9 g, 130 mmol)
was charged over a period of 20 min. After the addition, the mix-
ture was stirred at 0 ꢀC for an additional 30 min. A solution of 3
(26 g, 118 mol) in toluene (230 mL) was added. The mixture was
allowed to warm to ambient temperature over a period of 1 h and
stirred for an additional 16 h. Water (75 mL) and saturated NaCl
(100 mL) were added. The organic layer was separated and
washed sequentially with a mixture of 20% citric acid (100 mL)
and saturated NaCl (100 mL), a mixture of water (100 mL) and
saturated NaCl (75 mL), saturated NaHCO3 (150 mL), and a mix-
ture of water (100 mL) and saturated NaCl (75 mL). The resulting
solution of 12 in toluene was used for the next step without
further purification: HPLC assay 82%; 1H NMR (500 MHz, CDCl3)
98.1%; 1H NMR (500 MHz, DMSO-d6)
d
8.32 (d, J¼9.0 Hz, 2H), 8.18
d
7.60 (s, 1H), 7.57 (d, J¼7.6 Hz, 1H), 7.52 (d, J¼7.6 Hz, 1H), 7.45 (t,
(d, J¼9.0 Hz, 2H), 7.79 (s, 1H), 7.74 (d, J¼7.0 Hz, 1H), 7.52–7.64 (m,
2H), 7.32 (d, J¼9.0 Hz, 2H), 6.83 (d, J¼9.0 Hz, 2H), 6.79 (d,
J¼9.0 Hz, 2H), 6.59 (d, J¼9.0 Hz, 2H), 6.23 (br, 1H), 4.62 (t,
J¼6.3 Hz, 1H), 3.28 (t, J¼6.7 Hz, 2H), 3.04 (t, J¼6.7 Hz, 2H), 2.97 (s,
J¼7.8 Hz, 1H), 7.37–7.15 (m, 4H), 6.87–6.80 (m, 4H), 7.77 (d,
J¼8.8 Hz, 2H), 6.43 (d, J¼9.1 Hz, 2H), 5.09 (br s, 1H), 4.27 (d, J¼4.1,
Hz, 1H), 4.25 (d, J¼4.4, Hz, 1H), 3.79 (s, 3H), 3.23 (m, 1H), 3.11 (m,
2H), 3.04 (m, 1H), 2.80 (s, 3H), 2.79–2.67 (m, 2H); 13C NMR
3H), 2.82–2.92 (m, 2H); 13C NMR (125 MHz, DMSO-d6)
d 166.0,
(125 MHz, CDCl3)
d
160.9, 156.5, 150.7, 139.0, 138.1, 132.4, 131.1,
157.6, 149.8, 145.7, 144.8, 144.7, 137.2, 131.0, 130.6, 129.5, 129.3,
128.9, 125.5, 125.4, 123.7, 123.6, 123.2, 120.9, 118.1, 113.9, 56.3,
55.0, 53.4, 42.2, 41.3. HRMS calcd for C24H24ClF3 N2O3S [MþH]þ:
513.1227, found 513.1235. HPLC for 16 (tR¼7.2 min), 12
(tR¼11.4 min), p-nitrobenzoic acid (tR¼6.2 min): Waters Symme-
130.8, 129.8, 129.5, 127.7, 125.7, 124.4, 123.7, 121.8, 120.4, 119.1,
117.6, 114.8, 58.5, 57.4, 57.2, 55.3, 49.5, 47.2, 41.7. HRMS calcd for
C32H32ClF3N2O4S [MþH]þ 633.1802, found 633.1793. HPLC for 10
(tR¼6.2 min), 3 (tR¼6.0 min), 12 (tR¼11.4 min): Waters Symmetry
C18
5
m
m 150 mmꢂ4.6 mm, flow rate¼1.0 mL/min, 25 ꢀC, gradient
try C18
5
m
m 150 mmꢂ4.6 mm, flow rate¼1.0 mL/min, 25 ꢀC,
elution from 90:10 A:B to 5:95 A:B over 10 min then held for an
additional 6 min; A¼0.1% formic acid in water; B¼0.1% formic acid
in CH3CN.
gradient elution from 90:10 A:B to 5:95 A:B over 10 min then
held for an additional 6 min; A¼0.1% formic acid in water; B¼0.1%
formic acid in CH3CN.

9074
E.B. Villhauer et al. / Tetrahedron 65 (2009) 9067–9074
4.10. (S)-3-[4-(4-Chloro-phenoxy)-phenyl]-4-(3-trifluoro-
methyl-phenyl)-1-(2-methanesulfonyl-ethyl)-
imidazolidin-2-one (1)
ChiralpakÔ IA, 5
m
m 250 mmꢂ4.6 mm, flow rate¼0.5 mL/min,
25 ꢀC, isocratic, CH3CN–water¼80:20. HPLC for 1 (tR¼10.1 min), 16
(tR¼7.1 min): Waters Symmetry C18
5
mm 150 mmꢂ4.6 mm, flow
rate¼1.0 mL/min, 25 ꢀC, gradient elution from 90:10 A:B to 5:95 A:B
over 10 min then held for an additional 6 min; A¼0.1% formic acid in
water; B¼0.1% formic acid in CH3CN.
To a 2-L, three-necked, round bottom flask was charged a solu-
tion of 17 (65 g, 95.4 mmol), isopropyl acetate (1 L), and 0.5 N NaOH
(400 mL). The mixture was stirred for 15 min. The organic layer was
separated, washed with water (2ꢂ700 mL), and evaporated under
vacuum until the final volume of 180 mL was reached. Toluene
(600 mL) was added and evaporated under vacuum until the final
volume of 180 mL was reached to obtain a solution of 16 in toluene.
1,10-Carbonyldiimidazole (23.2 g, 143 mmol) was added and the
mixture was allowed to warm to 100 ꢀC. The mixture was stirred for
an additional 4 h, cooled to rt and diluted with toluene (400 mL).
The organic phase was washed sequentially with a mixture of 1 N
HCl (400 mL) and 25% NaCl (400 mL), 25% NaCl (400 mL), a mixture
of 9% NaHCO3 (300 mL) and 25% NaCl (300 mL), and water (400 mL).
The organic phase was evaporated under vacuum until the final
volume of 300 mL was reached, treated with active charcoal (Pica
Pure HP100, 3 g) for 1 h, and filtered through a pad of Celite (16 g).
The organic solution was evaporated until the final volume of
200 mL was reached. The resulting solution was allowed to warm to
60 ꢀC, diluted with heptanes (135 mL), cooled to 20 ꢀC, and stirred
for an additional 3 h. To the resulting suspensions, ethanol (30 mL)
and heptanes (195 mL) were added, and stirred for an additional
12 h. The solids were collected by filtration, rinsed with a solution of
33% toluene in heptanes (150 mL), a solution of 7% ethanol in hep-
tanes (100 mL), and dried at 50 ꢀC under vacuum for 16 h to obtain 1
(33.3 g, 65%) as an off-white solid: chiral HPLC assay 96:4 er. The
following recrystallization was used to improve the optical purity.
Crude 1 (33.3 g) was dissolve into EtOH (125 mL) at 50 ꢀC, filtered,
cooled to 25 ꢀC and seeded. The batch was stirred at rt for 20 h. The
solids were collected by filtration, rinsed with EtOH (33 mL), dried
at 50 ꢀC under vacuum for 16 h to afford 1 (23.3 g, 70% yield) as
a white solid: mp 89–90 ꢀC; chiral HPLC assay 99.8:0.2 er; HPLC
Acknowledgements
We thank Dr. Jessica Liang, Mr. Anthony DiJulio, Dr. Kurt
Konisgberger and Mr. James Vivelo for engineering support during
the pilot plant campaign. We are indebted to Mr. Robert DuVal and
Dr. Tangqing Li for calorimetric studies and discussions on plant
safety. We appreciate analytical assistance from Dr. Donald Drink-
water, Ms. Sara Kessler, Ms. Mary Kolor, Ms. Tanise Ahattock, Ms.
Ping Zhu, Mr. Rudy Kircher, and Dr. Steven Fazio. We thank Dr.
Xiaohui He, Dr. Xuefeng Zhu, Dr. Dean Phillips, and Dr. Hong Liu for
their discovery of a cannabinoid-1 antagonist.
References and notes
1. (a) Liu, H.; He, X.; Phillips, D.; Zhu, X.; Yang, K.; Lau, T.; Wu, B.; Xie, Y.; Nguyen,
T.; Wang, X. WO 2008076754; Chem. Abstr. 2008, 771087. (b) Donohue, S. R.;
Halldin, C.; Pike, V. W. Tetrahedron Lett. 2008, 49, 2789–2791; (c) Kim, M.-a.;
Kim, J. Y.; Song, K.-S.; Kim, J.; Lee, J. Tetrahedron 2007, 63, 12845–12852.
2. Basavaish, D.; Krishna, P. R. Tetrahedron 1995, 51, 2403–2416.
3. Truce, W. E.; Norell, J. R. J. Am. Chem. Soc. 1963, 85, 3236–3239.
4. Shieh, W.-C.; Chen, Z.; Xue, S.; McKenna, J.; Wang, R.-M.; Prasad, K.; Repic, O.
Tetrahedron Lett. 2008, 49, 5359–5362.
5. (a) Fei, Z.; Wu, Q.; Zhang, F.; Cao, Y.; Liu, C.; Shieh, W.-C.; Xue, S.; McKenna, J.;
Prasad, K.; Prashad, M.; Baeschlin, D.; Namoto, K. J. Org. Chem. 2008, 73, 9016–
9021; (b) Garrett, C. E.; Jiang, X.; Prasad, K.; Repic, O. Tetrahedron Lett. 2002, 43,
4161–4165.
6. Shieh, W.-C.; Carlson, J.; Shore, M. Tetrahedron Lett. 1999, 40, 7167–7170.
7. Banfi, L.; Riva, R. Org. React. 2005, 65, 1–140.
8. For recent reviews on the synthesis of vicinal diamines through aziridium ions,
see: (a) Kurteva, V.; Lyapova, M. Central Eur. J. Chem. 2004, 2, 686–695; (b)
Lucet, D.; Le Gall, T. L.; Mioslowski, C. Angew. Chem., Int. Ed. 1998, 37, 2580–2627.
9. (a) Song, H. A.; Dadwal, M.; Lee, Y.; Mick, E.; Chong, H.-S. Angew. Chem., Int. Ed.
2009, 48, 1–4; (b) Couturier, C.; Blanchet, J.; Schlama, T.; Zhu, J. Org. Lett. 2006,
8, 2183–2186; (c) O’Brien, P.; Towers, T. D. J. Org. Chem. 2002, 67, 304–307.
10. (a) Curthbertson, E. O.; Brien, P.; Towers, T. D. Synthesis 2001, 693–695; (b)
Frizzle, M. J.; Caille, S.; Marshall, T. L.; McRae, K.; Nadeau, K.; Guo, G.; Wu, S.;
Martinelli, M. J.; Moniz, G. A. Org. Process Res. Dev. 2007,11, 215–222; (c) Saravanan,
P.; Bisai, A.; Baktharaman, S.; Chandrasekhar, M.; Singh, V. K. Tetrahedron 2002, 58,
4693–4706.
11. (a) Chuang, T.-H.; Sharpless, K. B. Org. Lett. 2000, 2, 3555–3557; (b) Chuang, T.-H.;
Sharpless, K. B. Org. Lett. 1999, 1, 1435–1437.
12. Bentley, T. W.; Christl, M.; Kemmer, R.; Llewellyn, G.; Oakley, J. E. J. Chem. Soc.,
Perkin Trans. 2 1994, 2531–2538.
13. Gmeiner, P.; Junge, D.; Ka¨rtner, A. J. Org. Chem. 1994, 59, 6766–6776.
14. Shieh, W.-C.; Carlson, J. A. J. Org. Chem. 1994, 59, 5463–5465.
assay 99.5%; 1H NMR (500 MHz, DMSO-d6)
d 7.77 (s,1H), 7.66 (d,1H,
J¼5.0 Hz, 1H), 7.63 (d, 1H, J¼5.0 Hz, 1H), 7.43 (m, 2H), 7.36 (m, 2H),
6.92 (m, 4H), 5.60 (dd, J¼10.0, 5.0 Hz, 1H), 3.99 (t, J¼10.0 Hz, 2H),
3.75 (m,1H), 3.67 (m,1H), 3.44 (m,1H), 3.32 (m,1H), 3.03 (s, 3H); 13C
NMR (125 MHz, DMSO-d6)
d 157.4, 156.0, 151.4, 142.0, 134.9, 130.6,
130.0, 129.7, 129.4, 126.8, 124.7, 124.1, 123.5, 121.8, 119.6, 119.2, 55.9,
50.9, 50.3, 40.6, 37.6. HRMS calcd for C25H22ClF3N2O4S [MþH]þ
539.1019, found 539.1011. Anal. Calcd for C25H22ClF3N2O4S: C, 55.71;
H, 4.11; Cl, 6.58; F,10.57; N, 5.20; S, 5.95. Found: C, 55.76; H, 4.32; Cl,
6.72; F, 10.67; N, 5.00; S, 6.09. Chiral HPLC for (S)-enantiomer
(tR¼9.4 min), (R)-enantiomer (tR¼12.3 min): Chiral Technologies