Efficient synthesis of a GABAA α2,3-selective allosteric modulator via a sequential Pd-catalyzed cross-coupling approach

Article abstract of DOI:10.1021/jo050741o

A practical synthesis of 2-[3-(4-fluoro-3-pyridin-3-yl-phenyl)-imidazo[1,2- a]pyrimidin-7-yl]-propan-2-ol (1), an oral GABA_A α _2/3-selective agonist, is described. The five-step process, which afforded 1 in 40% overall yield, included imidazopyrimidine 2 and pyridine boronic acid 4 as key fragments. The synthesis is highlighted by consecutive Pd-catalyzed coupling steps to assemble the final free base 1 in high yield and regioselectivity. A novel method for Pd removal in the final step is also described.

Full text of DOI:10.1021/jo050741o

Efficient Synthesis of a GABAA r2,3-Selective Allosteric Modulator
via a Sequential Pd-Catalyzed Cross-Coupling Approach
Mark S. Jensen,* R. Scott Hoerrner,* Wenjie Li, Dorian P. Nelson, Gary J. Javadi,
Peter G. Dormer, Dongwei Cai, and Robert D. Larsen1
Department of Process Research, Merck Research Laboratories, P.O. Box, 2000,
Rahway, New Jersey 07065
mark_jensen@merck.com
Received April 13, 2005
A practical synthesis of 2-[3-(4-fluoro-3-pyridin-3-yl-phenyl)-imidazo[1,2-a]pyrimidin-7-yl]-propan-
2-ol (1), an oral GABAA R2/3-selective agonist, is described. The five-step process, which afforded 1
in 40% overall yield, included imidazopyrimidine 2 and pyridine boronic acid 4 as key fragments.
The synthesis is highlighted by consecutive Pd-catalyzed coupling steps to assemble the final free
base 1 in high yield and regioselectivity. A novel method for Pd removal in the final step is also
described.
Introduction
a potential R2/R3-selective agonist, and a synthetic route
capable of supplying kilo-scale quantities was needed.5
GABA (γ-aminobutyric acid) is the major inhibitory
neurotransmitter in the central nervous system. One of
its sites of action is the GABAA receptor family. Within
this family there are various subtypes that, when selec-
tively allosterically modulated, elicit different responses.
For example, selective agonists of the R1 subtype elicit a
sedative/muscle relaxation effect, while agonists of the
R2 and R3 subtypes are believed to mediate anxiety.2
Benzodiazepines such as diazepam bind nonselectively
to the R1, R2, R3, and R5 subtypes with nanomolar affinity.
While the benzodiazepines are widely used in the clinic,
this nonselective binding leads to a variety of undesirable
side effects.3 Recently, workers at Merck identified 14 as
The approach taken for the assembly of this molecule
relies on Pd catalysis to form the two aryl-aryl bonds.
The key to the success of this approach is the availability
(1) Current address: Amgen, Inc., Thousand Oaks, CA 91320-1799.
(2) (a) McKernan, R. M.; Rosahl, T. W.; Reynolds, D. S.; Sur, C.;
Wafford, K. A.; Atack, J. R.; Farrar, S.; Myers, J.; Cook, G.; Ferris, P.;
Garrett, L.; Bristow, L.; Marshall, G.; Macaulay, A.; Brown, N.; Howell,
O.; Moore, K. W.; Carling, R. W.; Street, L. J.; Castro, J. L.; Ragan, C.
I.; Dawson, G. R.; Whiting, P. J. Nat. Neurosci. 2000, 3, 587-592. (b)
Sieghart, W. Trends Pharmacol. Sci. 2000, 21, 411-413. (c) Rudolph,
U.; Crestani, F.; Benke, D.; Brunig, I.; Benson, J. A.; Fritschy, J.-M.;
Martin, J. R.; Bluethmann, H.; Mohler, H. Nature 1999, 401, 796-
800.
(4) (a) Chambers, M. S.; Goodacre, S. C.; Hallett, D. J.; Jennings,
A.; Jones, P.; Lewis, R. T.; Moore, K. W.; Street, L. J.; Szekeres, H. J.
Imidazo-Pyrimidine Derivatives as Ligands for GABA Receptors. World
Patent WO 02/074772 A1, September 26, 2002. (b) Chambers, M. S.;
Goodacre, S. C.; Hallett, D. J.; Jennings, A.; Jones, P.; Lewis, R. T.;
Moore, K. W.; Russell, M. G. N.; Street, L. J.; Szekeres, H. J. Imidazo-
Pyrimidine Derivatives as Ligands for GABA Receptors. World Patent
WO 02/074773 A1, September 26, 2002.
(5) Li, W.; Cai, D.; Hoerrner, R. S.; Jensen, M. S.; Larsen, R. D.;
Petasis, N. A. Preparation of Substituted Imidazopyrimidines. World
Patent WO 03/080621 A1, October 2, 2003.
(3) Costa, E.; Guidotti, A. Trends Pharmacol. Sci. 1996, 17, 192-
200.
10.1021/jo050741o CCC: $30.25 © 2005 American Chemical Society
Published on Web 06/29/2005
6034
J. Org. Chem. 2005, 70, 6034-6039

Practical Synthesis of a GABAA R2/3-Selective Agonist
SCHEME 1a
of 4-bromo-2-chloro-1-fluorobenzene (3). With the proper
ligand choice, a Heck-type reaction of 3 with imidazopy-
rimidine 2 should be feasible to provide a coupled aryl
chloride product. In the final step, another proper ligand
choice would enable the coupling of the aryl chloride with
3-pyridine boronic acid (4) in a Suzuki reaction to provide
the target compound 1. The major challenges anticipated
for this approach were the lack of availability of 4,
difficulty in removing residual Pd from this nitrogen-rich
compound, and chemoselectivity in the coupling reac-
tions. Herein we report a scalable, high-yielding process
for the preparation of 1. This work highlights the
versatility of Pd-catalyzed coupling and also describes a
novel method of removing residual Pd from the process
stream.
a Reagents and conditions: (a) Ac2O, DMAP, Et3N, CH2Cl2
(96%); (b) HC(OEt)3, BFOEt2, CH2Cl2; i-Pr2-NEt (69%); (c) 2-ami-
noimidazole sulfate, EtOH, NaOMe (74%).
SCHEME 2
Results
The required imidazopyrimidine intermediate 2 was
prepared in three steps from commercially available
starting materials. In particular, the availability of
2-aminoimidazole sulfate in bulk made this strategy,
which is based on literature precedent,6 attractive. The
alcohol function of 3-hydroxy-3-methyl-2-butanone (5)
was protected as an acetate in 96% yield using standard
conditions. Attempts to use bulkier protecting groups
(trialkyl silyl or benzoate) were less successful. One-
carbon homologation of the resulting methyl ketone 6 to
provide the â-ketoacetal 7 was accomplished by treating
triethyl orthoformate with boron trifluoride etherate at
-65 °C. The resulting complex was warmed to 0 °C and
again cooled to -65 °C, and the methyl ketone 6 was then
introduced followed by the diisopropylethylamine.7 Fol-
lowing an aqueous workup, 7 was isolated as a crude oil
and used directly in the next step. To form the imida-
zopyrimidine ring system, 2-aminoimidazole sulfate was
treated with 2 equiv of NaOMe followed by the â-keto-
acetal 7. It has been previously reported that the regio-
selectivity of this type of ring formation is strongly
influenced by pH.6a High pH promotes formation of the
7-substituted isomer (2), while low pH promotes forma-
tion of the 5-substituted isomer (8). This trend held here,
and using the excess NaOMe gave a 96:4 mixture of 2
and 8. The isolation of 2 was key to success here, and a
crystallization method was developed. The crude reaction
mixture was adjusted to pH 6 with HCl/IPA, and the
solvent was switched from ethanol to ethyl acetate.
Under these conditions, imidazopyrimidine 2 crystallized
and was isolated as a free-flowing solid. The undesired
isomer 8 was completely rejected in the crystallization.
The overall yield of the three-step process just described
was 48-50% and provided 2 as a crystalline compound
with excellent purity.
SCHEME 3a
a Reagents and conditions: (a) Pd(OAc)2, Ph3P, Cs2CO3, 1,4-
dioxane, 90 °C, 12 h (88%); (b) 4, Pd(dba)2, t-Bu3P, K3PO4, H2O,
1,4-dioxane, 80 °C, 16 h (96%); (c) HCl, EtOH, ∆, then IPAC (88%).
a Suzuki coupling to form the carbon-carbon bond
(Scheme 2). To accomplish this, 2 was first brominated
with bromine in the presence of sodium acetate, potas-
sium bromide, and methanol to provide the 3-bromo
imidazopyrimidine 9 in excellent yield. When arylboronic
acid 10 was used as the partner in the Suzuki coupling
with 9, the target compound 1 was the product. While
the convergency of this approach was attractive, the
preparation of the requisite arylboronic acid 10 was
lengthy. In a new strategy (Scheme 3), the coupling
reaction between 2 and 4-bromo-2-chloro-1-fluorobenzene
(3) would directly install the aryl group.8 The chloride of
11 could then be used as a handle to install the 3-pyridyl
Heck-Type Coupling. The initial approach4 to couple
imidazopyrimidine 2 with the central aryl ring relied on
(6) For a similar approach to 7-substituted imidazo[1,2-a]pyrim-
idines, see: (a) Guerret, P.; Jacquier, R.; Maury, G. Bull. Chem. Soc.
France 1972, 3503-3511. (b) Ku¨nstlinger, M.; Breitmaier, E. Synthesis
1983, 161-162. The parent as well as 2- and 3-substituted imidazo-
[1,2-a]pyrimidines have been prepared by the reaction of a 2-aminopy-
rimidine with an R-halocarbonyl compound: (c) Rival, Y.; Grassy, G.;
Taudou, A.; Ecalle, R. Eur. J. Med. Chem. 1991, 26, 13-18. (d)
Tschitschibabin, A. E. Chem. Ber. 1925, 58, 1704-1706. (e) Lo¨ber, S.;
Hu¨bner, H.; Gmeiner, P. Bioorg. Med. Chem. Lett. 1999, 9, 97-102.
(7) Suzuki, M.; Yanagisawa, A.; Noyori, R. Tetrahedron Lett. 1982,
23, 3595-3596.
J. Org. Chem, Vol. 70, No. 15, 2005 6035

Jensen et al.
moiety. With this two-step approach, the bromination of
2 and the numerous functional group manipulations
needed to prepare boronic acid 10 would be eliminated.
In the event, treatment of 2 with 3, Pd(OAc)2, Ph3P,
and Cs2CO3 in 1,4-dioxane at 90 °C gave, after an
extractive workup, coupled product 11 as a crystalline
compound in 88% isolated yield. No other cross-coupled
products were observed. K2CO3 could also be used as
base in this reaction; however, other bases such as Na2-
CO3, phosphates, and amines failed. Ph3P proved to be
superior to other common ligands, and DMF was an
acceptable alternative solvent. The regiochemical out-
come of this reaction can be rationalized solely on the
electronic nature of 2.9 Nucleophiles typically attack the
imidazopyrimidine at the 5- and 7-positions, while elec-
trophiles react at the 3-position.
FIGURE 1. NOE studies of 1 and 12.
SCHEME 4
Suzuki Coupling. With 11 in hand, only installation
of the 3-pyridyl moiety remained. Recently reported
successes in the cross-coupling of aryl chlorides in the
Suzuki reaction10 led us to focus our efforts in that
direction.11 The main drawback to this approach was the
lack of access to the requisite 3-pyridyl boronic acid (4).
While 4 is commercially available, it is quite expensive
and supplies are limited. These limitations led to the
development of an improved method for preparing 4.12
When n-BuLi was added over 1 h to a -40 °C solution of
3-bromopyridine, triisopropylborate, toluene, and THF,
4 was the result. Following an extractive workup, 4 was
isolated in 91% yield.13
Our initial attempts to couple 11 and 4 utilized
anhydrous conditions14 [Pd2dba3, P(t-Bu)3, KF, THF], but
to our surprise, no reaction occurred under these condi-
tions. Anhydrous conditions with other typical solvents
(i.e., 1,4-dioxane) and additives (i.e., Cs2CO3, CsF, K3PO4)
also failed. The reaction between aryl chloride 11 and
phenyl boronic acid, however, worked well under these
conditions, so the 3-pyridyl boronic acid was viewed as
the culprit.
and 13C NMR spectra were similar. There was a distinct
difference in their UV spectra. The major product exhib-
ited λmax at 230 and 240 nm, while the minor component
had λmax at 201, 228, and 312 nm. NOE experiments15
unambiguously established that the two compounds
differed in substitution on the imidazole ring (Figure 1).
During the Suzuki reaction, a Dimroth rearrangement16
had occurred (Scheme 4). Studying the reaction profile
using HPLC revealed that the rearrangement occurred
both at aryl chloride 11 and the coupled product 1 to yield
aryl chloride 13 and coupled product 12, respectively. The
rearranged aryl chloride 13 subsequently underwent a
Suzuki coupling to provide 12. While the Dimroth re-
arrangement is typically a reversible process, in this
system, substitution at the 2-position with an aryl group
is thermodynamically favored to substitution at the
3-position. In compounds 13 and 12, the aryl ring can be
coplanar and thus conjugated with the imidazopyrimi-
dine ring system. This is evidenced by the UV data cited
above and also supported by molecular modeling calcula-
tions. This conjugation is not possible with compounds
11 and 1 due the steric interaction between the ortho
hydrogens of the aryl ring and the 5-hydrogen of the
pyrimidine ring. In fact, under basic aqueous conditions,
1 can be cleanly and completely converted to 12. No
rearrangement, however, was seen in the arylation
reaction of 2 and 3 to produce 11. This is due to the
anhydrous conditions of the reaction. A plausible mech-
anism for this Dimroth rearrangement is shown in Figure
2.
Water was essential for the reaction.11 In the first
successful reaction, the coupling partners were subjected
to Pd2dba3, t-Bu3P, Na2CO3, THF, and water at 70 °C for
48 h. While these conditions provided 1, a new challenge
was revealed. Two products were formed. Both had the
1
empirical formula of the desired product, and their H
(8) (a) Li, W.; Nelson, D. P.; Jensen, M. S.; Hoerrner, R. S.; Javadi,
G. J.; Cai, D.; Larsen, R. D. Org. Lett. 2003, 5, 4835-4837. (b) Recently,
other arylation reactions of imidazoles, oxazoles, and thiazoles have
been reported. For a leading reference, see: Sezen, B.; Sames, D. J.
Am. Chem. Soc. 2003, 125, 10580-10585.
(9) Sliskovic, D. R. In Comprehensive Heterocyclic Chemistry II;
Katritzky, A. R., Rees, C. W., Scriven, E. F., V., Eds; Pergamon:
Oxford, UK, 1996; Vol. 8, p 345.
(10) For a review of palladium-catalyzed coupling reactions with aryl
chlorides, see: Littke, A. F.; Fu, G. C. Angew. Chem., Int. Ed. 2002,
41, 4176-4211.
(11) For a discussion of Suzuki-Miyaura reactions using electron-
deficient arylboronic acids, including 4, see: Barder, T. E.; Walker, S.
D.; Martinelli, J. R.; Buchwald, S. L. J. Am. Chem. Soc. 2005, 127,
4685-4696.
(12) (a) Li, W.; Nelson, D. P.; Jensen, M. S.; Hoerrner, R. S.; Cai,
D.; Larsen, R. D.; Reider, P. J. J. Org. Chem. 2002, 67, 5394-5397.
(b) Cai, D.; Larsen, R. D.; Reider, P. J. Tetrahedron Lett. 2002, 43,
4285-4287.
(13) For a recently reported preparation of diethyl-3-pyridyl borane
from 3-lithio pyridine, see: Lipton, M. F.; Mauragis, M. A.; Maloney,
M. T.; Veley, M. F.; VanderBor, D. W.; Newby, J. J.; Appell, R. B.;
Daugs, E. D. Org. Process Res. Dev. 2003, 7, 385-392.
(14) Littke, A. F.; Dai, C.; Fu, G. C. J. Am. Chem. Soc. 2000, 122,
4020-4028.
(15) (a) Kessler, H.; Oschkinat, H.; Griesinger, C.; Bermel, W. J.
Magn. Reson. 1986, 70, 106-133. (b) Stonehouse, J.; Adell, P.; Keeler,
J.; Shaka, A. J. J. Am. Chem. Soc. 1994, 116, 6037-6038. (c) Stott,
K.; Stonehouse, J.; Keeler, J.; Hwang, T. L.; Shaka, A. J. J. Am. Chem.
Soc. 1995, 117, 4199-4200.
(16) Vaughan, K.; LaFrance, R. J.; Tang, Y. J. Heterocycl. Chem.
1991, 28, 1709-1713.
6036 J. Org. Chem., Vol. 70, No. 15, 2005

Practical Synthesis of a GABAA R2/3-Selective Agonist
During the workup of the Suzuki reaction, the acidic
aqueous solution of 1 was clear yellow, leading us to
suspect that the metal was Pd(II). With the hope that
reducing the metal to Pd(0) would facilitate its removal,
borane-trimethylamine complex was added to the acidic
aqueous solution. This was rewarded with the formation
of a black precipitate. Filtration through a nylon mem-
brane with a 1 µm pore size removed the precipitated
Pd. Addition of NaOH to the now colorless solution gave
crystalline free base 1 that contained about 40 ppm Pd.
This low-Pd free base was then taken up in HCl and
ethanol, polished with Darco-G60, and then crystallized
by addition of IPAC to provide 14 with less than 3 ppm
Pd.
FIGURE 2. Plausible rearrangement mechanism.
Since the Suzuki coupling is the last bond-forming step
of this synthesis, the appearance of this impurity pre-
sented a significant problem. Impurities introduced at
the late stage of a drug synthesis can be difficult to reject,
and the yield loss due to their formation should be
avoided. A careful study of solvents, bases, reaction
temperatures, and amounts of water was undertaken.
Stability studies where 1 was warmed in a mixture of
aqueous Cs2CO3 and an organic solvent indicated that
the rearrangement is rapid in dimethylacetamide (DMAC),
acetonitrile, and ethanol but much slower in 1,4-dioxane,
THF, and 2-Me THF. When the latter solvents were used
in the reaction, the product crystallized out of solution
as the reaction progressed, presumably protecting the
compound from rearrangement. Also, minimizing the
amount of water was essential. With 1,4-dioxane, 25 vol
% water was optimal. More water increased the re-
arrangement, and less water gave a sluggish reaction
that did not go to completion. Under the optimized
reaction conditions, a 102:1 mixture of 1 and 12 was
formed. Following an extractive workup, 1 was isolated
as a crystalline free base in 96% yield. Free base 1 was
then dissolved in a hot solution of HCl and ethanol.
Addition of isopropyl acetate (IPAC) and cooling provided
crystalline bis-HCl salt 14. During this sequence, the
Dimroth rearrangement product 12 was completely
rejected.
Pd Removal. While the current Pd-catalyzed cross-
coupling route to 1 is efficient in forming the carbon-
carbon bonds, it has a significant drawback. Residual Pd
can be challenging to remove late in a synthesis, par-
ticularly when the target compound is rich in hetero-
atoms, as 1 is. Without any treatment to remove Pd, free
base 1 was crystallized with ca. 4000 ppm Pd. The salt
formation using this material gave bis-HCl salt 14 with
ca. 2700 ppm Pd. The target Pd level was <10 ppm.
Treatment of an acidic ethanolic solution of 1 with
numerous commercially available Pd scavengers, includ-
ing polymer-bound phosphines, thiols, sulfonates, car-
boxylates, and isothiouroniums, all failed to significantly
lower the Pd levels. Treatment with n-Bu3P17 also failed.
Darco G-60, an activated charcoal, gave good removal of
Pd, but 100-200 wt %, based on 1, was required. While
this is a functional solution to the problem, we were
concerned with variable product loss to the charcoal.
Conclusion
A new route to the GABAA R2,3-selective agonist 1 has
been developed. A known route to the imidazopyrimidines
was refined to provide 2 in good yield and excellent
purity. A new two-step route from 2 to 1 that relies on
Pd catalysis to form both aryl-aryl bonds was identified
and developed. In this route, 4-bromo-2-chloro-1-fluo-
robenzene (3) was used, and the ligand choice in the Pd-
catalyzed reactions exquisitely controlled the regioselec-
tivity. This work also revealed an improved method for
preparing 3-pyridyl boronic acid. Impurity formation via
the Dimroth rearrangement was identified and sup-
pressed, and a new method of removing Pd from the
process stream was found. The overall yield from com-
mercially available 5 to free base 1 was 40.5% with the
longest linear sequence being five steps. The bis-HCl salt
14 was prepared in 35.6% overall yield from 5.
Experimental Section
Acetic Acid 1,1-Dimethyl-2-oxo-propyl Ester (6). A
mechanically stirred 3 L round-bottomed flask under an
atmosphere of nitrogen was charged with 3-hydroxy-3-methyl-
2-butanone (5, 100.0 g, 979 mmol) and CH2Cl2 (900 mL). A
solution of acetic anhydride (92.4 mL, 979 mmol) and CH2Cl2
(50 mL) was added over 10 min, and the resulting solution
was cooled to 10 °C. DMAP (5.98 g, 49.0 mmol) was added in
one portion followed by Et3N (205 mL, 1.47 mol) over 50 min
while maintaining an internal temperature below 16 °C. After
the addition, the mixture was aged for 16 h at room temper-
ature. The resulting solution was poured onto MeOH (200 mL),
and after 15 min, 2 N HCl (400 mL) was added. The phases
were separated, and the organic phase was washed with water
(500 mL) followed by saturated NaHCO3 (500 mL) and then
dried over Na2SO4. The solvent was removed in vacuo to
provide 6 as a golden oil (132.5 g, 96.9 wt % pure, 95.8%
isolated yield). The characterization data was identical to that
reported previously.18
Acetic Acid 4,4-Diethoxy-1,1-dimethyl-2-oxo-butyl Es-
ter (7). A 3 L round-bottomed flask fitted with a mechanical
stirrer under an atmosphere of N2 was charged with triethyl
orthoformate (304 mL, 1.83 mol) and CH2Cl2 (1.30 L). The
solution was cooled to -65 °C, and BF3‚OEt2 (255 mL, 2.01
mol) was added over 30 min to provide a yellow solution. After
an additional 15 min, the cold bath was removed, and the flask
contents were allowed to warm to 0 °C. After aging at 0 °C for
45 min, the mixture was cooled to -65 °C to provide a thin
white slurry. A solution of methyl ketone 6 (131.55 g, 0.912
mol) and CH2Cl2 (100 mL) was added over 17 min. After the
(17) Larsen, R. D.; King, A. O.; Chen, C. Y.; Corley, E. G.; Foster,
B. S.; Roberts, F. E.; Yang, C.; Lieberman, D. R.; Reamer, R. A.;
Tschaen, D. M.; Verhoeven, T. R.; Reider, P. J. J. Org. Chem. 1994,
59, 6391-6394.
(18) Mitsudo, T.; Hori, Y.; Yamakawa, Y.; Watanabe, Y. J. Org.
Chem. 1987, 52, 2230-2239.
J. Org. Chem, Vol. 70, No. 15, 2005 6037

Jensen et al.
mixture was aged for 40 min at -65 °C, diisopropylethylamine
(404 mL, 2.32 mol) was added over 30 min maintaining an
internal temperature below -63 °C. After an additional 45 min
at -65 °C, the mixture was warmed to room temperature over
45 min and aged at that temperature for 5 h. The reaction
mixture was poured onto saturated aqueous NaHCO3 (800 mL)
and stirred for 15 min. The layers were partitioned, and the
organic phase was washed with 1 M H2SO4 at 5 °C (2 × 800
mL) and water at 5 °C (2 × 800 mL) and then dried over Na2-
SO4. The solvent was removed in vacuo to provide 7 as an oil
(270 g, 57 wt % pure, 69% isolated yield). An analytical sample
was prepared by HPLC on a Zorbax C-8 preparative column
with gradient elution from 30% ACN, 70% 0.1% H3PO4/H2O
to 70% ACN, 30% 0.1% H3PO4/H2O over 20 min: 1H NMR
(CDCl3, 400.13 MHz) δ 4.95 (t, J ) 5.6 Hz, 1H), 3.69 (m, 2H),
3.53 (m, 2H), 2.79 (d, J ) 5.6 Hz, 2H), 2.07 (s, 3H), 1.47 (s,
6H), 1.18 (t, J ) 6.8 Hz, 6H); 13C NMR (CDCl3, 100.61 MHz)
δ 205.7, 170.4, 100.8, 83.7, 63.2, 41.4, 22.9, 21.1, 15.1.
to 90 °C and then cooled over several hours to 0 °C. The
crystalline solids were collected on a frit, rinsed with IPAC/
EtOAc (1:1), and air-dried to provide the coupled product 11
(16.19 g, 88%) as a cream-colored crystalline solid: mp 180-
1
182 °C; H NMR (CDCl3, 400.25 MHz) δ 8.53 (d, J ) 7.2 Hz,
1H), 7.86 (s, 1H), 7.59 (dd, J ) 6.8, 2.0 Hz, 1H), 7.42 (m, 1H),
7.34 (m, 1H), 7.14 (d, J ) 7.2 Hz, 1H), 1.63 (s, 6H); 13C NMR
(CDCl3, 100.64 MHz) δ 167.8, 158.1 (d, J ) 252.2 Hz), 148.0,
133.8, 131.4, 130.0, 127.8 (d, J ) 7.2 Hz), 125.5 (d, J ) 4.0
Hz), 122.4 (d, J ) 18.5 Hz), 122.1, 117.8 (d, J ) 20.9 Hz), 105.7,
72.5, 29.9. Anal. Calcd for C15H13ClFN3O: C, 58.93; H, 4.29;
N, 13.74. Found: C, 58.65; H, 4.11; N, 13.66.
2-[3-(4-Fluoro-3-pyridin-3-yl-phenyl)-imidazo[1,2-a]py-
rimidin-7-yl]-propan-2-ol (1). A 5 L round-bottomed flask
equipped with a mechanical stirrer, thermocouple, nitrogen/
vacuum inlet adapter, and septum was charged with aryl
chloride 11 (101.59 g, 333 mmol), 3-pyridine boronic acid (4,
42.98 g, 350 mmol), K3PO4 (111.69 g, 526 mmol), 1,4-dioxane
(1.20 L), and water (406 mL). Cycling vacuum and then
nitrogen three times degassed the stirred slurry. All solids
dissolved on warming to 70 °C.
2-Imidazo[1,2-a]pyrimidin-7-yl-propan-2-ol (2). A 3 L
round-bottomed flask equipped with a mechanical stirrer,
thermocouple and nitrogen inlet adaptor was charged with
aminoimidazole sulfate (84.6 g, 641 mmol) and EtOH (900 mL,
absolute). The resulting slurry was stirred at 24 °C, and solid
NaOMe (69.2 g, 1.28 mol) was added in two portions over 15
min. After the addition was complete, the reaction temperature
was 46 °C. The mixture was warmed to 60 °C and stirred for
45 min. After the aging, the mixture was cooled to 50 °C. A
solution of crude acetal 7 (253.5 g of 57 wt % material, 611
mmol) and EtOH (150 mL, absolute) was added over 1 h,
maintaining a reaction temperature of about 45 °C. After the
addition, the reaction was maintained at 60 °C for 3 h and
then cooled to 20 °C. The pH was adjusted to 5.9 by addition
of aqueous HCl (5 N, 130 mL) in 10 mL portions. Darco G-60
(9.0 g) was added, and the slurry was stirred for 30 min. The
solids were removed by filtration through a pad of Solka-floc,
and the cake was rinsed with EtOH (200 mL). The filtrate was
concentrated to ca. 200 mL via a rotovap; then, 300 mL of
EtOAc was added, and the mixture was concentrated to ca.
300 mL. This was repeated three times, with crystals forming
during the process. The final slurry was diluted to 600 mL
with EtOAc and then cooled to 5 °C with stirring. The solids
were collected on a frit, rinsed with cold EtOAc (50 mL), and
dried overnight under vacuum at 22 °C to provide 2 (79 g, 99
wt % pure, 74% isolated yield) as a crystalline solid: mp 150-
A 500 mL round-bottomed-flask equipped with a magnetic
stir bar, thermocouple, nitrogen/vacuum inlet adapter, and
septum was charged with Pd(dba)2 (9.57 g, 16.7 mmol) and
1,4-dioxane (130 mL). The resulting solution was degassed as
above, and then t-Bu3P (49.8 mL of a 10 wt % solution in
hexanes, 16.7 mmol) was added by syringe. The solution was
degassed again and then warmed to 70 °C. The warm catalyst
solution was cannulated to the 5 L flask, and the resulting
mixture was stirred at 70 °C for 16 h. At the end of reaction
most, of the product had crystallized out to provide a gray
slurry. The reaction mixture was partitioned between 2 N HCl
(1.8 L) and toluene (0.9 L). The clear yellow aqueous phase
was transferred to a stirred 4 L Erlenmeyer flask with a
nitrogen sweep, and borane trimethylamine complex (1.87 g)
was added. After 90 min, the resulting black solids were
removed by filtration through a 1.0 µm nylon filter. The filtrate
was transferred to a mechanically stirred 5 L flask equipped
with a pH probe. The pH was adjusted to 3.8 by slow addition
of 50 wt % NaOH (ca. 130 mL). The mixture self-nucleated to
provide a slurry. Additional 50 wt % NaOH (ca. 36 mL) was
added over 30 min to pH 7.1 to provide a cream-colored slurry.
The solids were collected on a frit, washed with water (250
mL), and air-dried to give the free base 1 as an off-white
crystalline solid (117.38 g, 94.5 wt % pure, 95.6% isolated yield,
41 ppm Pd): mp 234-236 °C; 1H NMR (DMSO-d6, 399.87
MHz) δ 9.00 (d, J ) 7.2 Hz, 1H), 8.85 (bs, 1H), 8.60 (dd, J )
4.8, 1.6 Hz, 1H), 8.06 (m, 1H), 7.92 (s, 1H), 7.87 (dd, J ) 7.2,
2.4 Hz, 1H), 7.74 (m, 1H), 7.76-7.68 (om, 2H), 7.37 (d, J )
7.2 Hz, 1H), 5.49 (s, 1H), 1.47 (s, 6H); 13C NMR (DMSO-d6,
100.55 MHz) δ 169.6, 159.3 (d, J ) 248.2 Hz), 149.9 (d, J )
3.2 Hz), 149.6, 148.5, 137.0 (d, J ) 3.2 Hz), 134.0, 133.5, 130.9,
130.3 (d, J ) 3.2 Hz), 130.0 (d, J ) 8.0 Hz), 126.6 (d, J ) 14.5
Hz), 126.1 (d, J ) 4.0 Hz), 124.1, 122.7, 117.7 (d, J ) 22.5
Hz), 106.3, 73.0, 30.3. Anal. Calcd for C20H17FN4O: C, 68.95;
H, 4.92; N, 16.08. Found: C, 68.55; H, 4.84; N, 15.90.
2-[3-(4-Fluoro-3-pyridin-3-yl-phenyl)-imidazo[1,2-a]py-
rimidin-7-yl]-propan-2-ol Dihydrogen Chloride Salt (14).
A magnetically stirred, 2 L, three-necked, round-bottomed
flask equipped with a condenser, thermocouple, and drying
tube was charged with free base 1 (25.00 g, 71.8 mmol) and
EtOH (950 mL). The suspension was warmed to 75 °C, then
HCl (47 mL of a 4.6 M solution in IPA) was added to provide
a clear solution. Powdered Darco G-60 (2.50 g) was added and
the mixture refluxed for 2 h. The suspension was cooled to 60
°C, and the solids were removed by filtration through a pad of
Solka-floc. The resulting clear solution was transferred to a
mechanically stirred 3 L round-bottomed flask equipped with
a reflux condenser. The solution was warmed to 75 °C, and
IPAC (1.00 L) was added. During the IPAC addition, the
solution self-nucleated to provide a white suspension, which
was cooled over several hours to room temperature. The solids
1
152 °C; H NMR (MeOH-d6, 400.13 MHz) δ 8.80 (d, J ) 7.2
Hz, 1H), 7.77 (d, J ) 1.6 Hz, 1H), 7.66 (d, J ) 1.6 Hz, 1H),
7.39 (d, J ) 7.2 Hz, 1H), 1.60 (s, 6H); 13C NMR (MeOH-d6,
100.61 MHz) δ 169.7, 147.8, 134.7, 132.8, 111.1, 105.4, 72.8,
28.4. Anal. Calcd for C9H11N3O: C, 61.00; H, 6.26; N, 23.71.
Found: C, 60.79; H, 6.30; N, 23.50.
2-[3-(3-Chloro-4-fluoro-phenyl)-imidazo[1,2-a]pyrimi-
din-7-yl]-propan-2-ol (11). A 500 mL round-bottomed flask
equipped with a mechanical stirrer, thermocouple and nitrogen/
vacuum inlet was charged with 2 (10.62 g, 60.0 mmol),
4-bromo-2-chloro-1-fluorobenzene (3, 8.00 mL, 66.0 mmol),
palladium(II) acetate (270 mg, 1.20 mmol), triphenylphosphine
(630 mg, 2.40 mmol), cesium carbonate (19.6 g, 60.2 mmol),
and anhydrous 1,4-dioxane (120 mL). The slurry was degassed
using five vacuum/nitrogen back-fill cycles, heated to 90 °C,
and aged for 12 h. The mixture was cooled to 60 °C; then, water
(150 mL) and EtOAc (150 mL) were added to the crude reaction
mixture, and the slurry was stirred at 45 °C until all of the
solids were dissolved. The layers were partitioned, and the
aqueous phase was extracted again with EtOAc (50 mL). The
organic phases were combined and extracted twice with 2 N
HCl (100 mL and then 30 mL). The acidic aqueous phases were
combined with EtOAc (250 mL), and NaOH (10.9 g, final pH
) 10) was added slowly with stirring; the mixture was warmed
to 65 °C to dissolve the solids. The phases were separated,
and the organic phase was concentrated to ca. 200 mL to
provide a slurry. IPAC (110 mL) was added, and the slurry
was concentrated again to ca. 200 mL. The slurry was warmed
6038 J. Org. Chem., Vol. 70, No. 15, 2005

Practical Synthesis of a GABAA R2/3-Selective Agonist
were collected on a frit, rinsed with 1:1 EtOH/IPAC (100 mL),
and then dried in a 60 °C vacuum oven to provide the bis-HCl
salt 14 as colorless crystals (26.76 g, 88%, <3 ppm Pd, <30
ppm P, <10 ppm B): mp 241-250 °C; 1H NMR (CD3OD,
400.13 MHz) δ 9.32 (m, 1H), 9.30 (d, J ) 7.2 Hz, 1H), 9.02 (m,
1H), 8.97 (m, 1H), 8.38 (s, 1H), 8.28 (m, 1H), 8.18 (dd, J ) 7.2,
2.0 Hz, 1H), 8.03 (d, J ) 7.2 Hz, 1H), 7.99 (m, 1H), 7.68 (dd,
J ) 10.4, 8.4 Hz, 1H), 1.67 (s, 6H); 13C NMR (CD3OD, 100.61
MHz) δ 178.4, 160.5 (d, J ) 254.0 Hz), 146.9 (d, J ) 3.4 Hz),
143.5, 141.4 (d, J ) 4.0 Hz), 140.7, 135.8, 134.4, 133.3 (d, J )
9.4 Hz), 131.9 (d, J ) 2.4 Hz), 127.5, 124.3, 123.1 (d, J ) 13.6
Hz), 122.4 (d, J ) 3.8 Hz), 120.8, 117.8 (d, J ) 23.2 Hz), 110.7,
73.2, 28.5.
MHz) δ 8.91 (d, J ) 7.2 Hz, 1H), 8.84 (m, 1H), 8.64 (dd, J )
4.8, 1.6 Hz, 1H), 8.43 (s, 1H), 8.18 (dd, J ) 7.6, 2.0 Hz, 1H),
8.09-8.05 (m, 2H), 7.55 (ddd, J ) 8.0, 4.8, 0.8 Hz, 1H), 7.46
(dd, J ) 10.8, 8.8 Hz, 1H), 7.35 (d, J ) 6.8 Hz, 1H), 1.49 (s,
6H); 13C NMR (DMSO-d6, 100.62 MHz) δ 170.0, 159.3 (d, J )
247.5 Hz), 149.6 (d, J ) 3.2 Hz), 149.4, 147.8, 144.1, 136.7 (d,
J ) 3.2 Hz), 135.4, 131.22 (d, J ) 3.2 Hz), 131.17 (d, J ) 1.6
Hz), 128.1 (d, J ) 3.2 Hz), 127.7 (d, J ) 8.0 Hz), 125.8 (d, J )
14.4 Hz), 124.1, 117.1 (d, J ) 23.2 Hz), 107.8, 105.9, 72.8, 26.6.
Acknowledgment. We gratefully acknowledge Prof.
N. Petasis and Dr. D. J. Hallett for helpful discussions.
2-[2-(4-Fluoro-3-pyridin-3-yl-phenyl)-imidazo[1,2-r]py-
rimidin-7-yl]-propan-2-ol (12). 1H NMR (DMSO-d6, 400.13
JO050741O
J. Org. Chem, Vol. 70, No. 15, 2005 6039