Optimized catalytic enantioselective aryl transfer process gives access to mGlu2 receptor potentiators

Article abstract of DOI:10.1021/op0602268

An asymmetric enantioselective aryl transfer reaction was developed to give access to the diarylmethanol 7 and ultimately acetate 2 which is useful for the preparation of mGlu2 receptor potentiators (Scheme 3). The aryl transfer chemistry involved the pre

Full text of DOI:10.1021/op0602268

Organic Process Research & Development 2007, 11, 560−567
Optimized Catalytic Enantioselective Aryl Transfer Process Gives Access to
mGlu2 Receptor Potentiators
Nicholas A. Magnus,* Peter B. Anzeveno, D. Scott Coffey, David A. Hay, Michael E. Laurila, Jeffrey M. Schkeryantz,
Bruce W. Shaw, and Michael A. Staszak
Eli Lilly and Company, Chemical Product Research and DeVelopment DiVision, Indianapolis, Indiana 46285, U.S.A.
Abstract:
An asymmetric enantioselective aryl transfer reaction was
developed to give access to the diarylmethanol 7 and ultimately
acetate 2 which is useful for the preparation of mGlu2 receptor
potentiators (Scheme 3). The aryl transfer chemistry involved
the preparation of a proposed arylalkylzinc species 14 from
boroxine 16 and diethylzinc (DEZ), and reacting this mixture
with aldehyde 5 in the presence of chiral ligand 15. During the
course of optimizing the preparations of proposed intermediate
Results and Discussion
In order to fund toxicological evaluations, 500 g of
diarylmethanol 1 was synthesized using the chemistry
outlined in Scheme 1.
1
4 and diarylmethanol 7, an understanding of optimal stoichi-
The Scheme 1 synthesis began with the protection of
ometry and reaction times was gained through empirical
observation, the use of solution IR, and analyzing off-gases via
real time gas analysis/mass spectroscopy. The preparation of
diarylmethanol 7 and subsequent conversion into acetate 2
required carefully selected workups, selective extractions, and
azeotropic distillations to generate a series of stock solutions
to accommodate oil intermediates that finally gave acetate 2 as
a crystalline solid with >99% ee.
4
-bromobenzyl alcohol 3 as tert-butyldimethylsilyl (TBS)
ether 4 by exposure to tert-butyldimethylsilyl chloride (TBS-
Cl) in the presence of triethylamine (TEA) and 4-dimethy-
laminopyridine (DMAP) in CH
Grignard reagent of ether 4 was prepared by reaction of 4
4
2
Cl
2
. The corresponding
5
with magnesium turnings in refluxing THF, which was
subsequently reacted with 3-cyanobenzaldehyde 5 to afford
an 80% yield of diarylmethanol 6 as a racemic mixture. The
racemic diarylmethanol 6 was subjected to an arduous chiral
chromatography to give a 47% yield of the desired diaryl-
methanol enantiomer 7. In order to differentiate the alcohols
of 7, the secondary alcohol was protected as a tetrahydro-
pyran (THP) acetal by reaction with 3,4-dihydro-2H-pyran
Introduction
The excitatory amino acid L-glutamate mediates most of
the excitatory neurotransmission within the central nervous
system (CNS). Glutamate receptors are classified into two
main types, ionotropic (iGlu), which are glutamate-mediated
ion channels, and metabotropic (mGlu), which are a class
2
and catalytic pyridinium p-toluene sulfonate (PPTS) in CH -
6
2
Cl to afford 8. Reaction of 8 with tetrabutylammonium
fluoride (TBAF) in THF afforded good yields of primary
alcohol 9. Reaction of 9 with methanesulfonyl chloride
1
of G-protein-coupled receptors. The mGlu receptors have
been divided into three main groups (I-III) with the group
2 2
(MsCl) and TEA in CH Cl gave mixtures of mesylate 10a
II (mGlu2 and -3) mGlu receptors being largely presynaptic
and chloride 10b. The chloride 10b is a major product formed
in the mesylation, because the TEA-hydrochloride that is
generated has good solubility in the reaction media and can
supply chloride ion to displace the mesylate from 10a. Due
to the sluggish reactivity of chloride 10b, the 10a/10b
mixture was reacted with sodium iodide under Finkelstein
halogen, halogen exchange conditions to give the intermedi-
ate iodide 10c which reacted with the phenolate of 11 in a
Williamson ether synthesis to afford ether 12. Standard THP
cleavage conditions (PPTS in wet EtOH) gave the chiral
2
and generally inhibiting neurotransmission. To access mul-
tikilogram quantities of mGlu2 receptor potentiators to fund
Lilly’s research directed towards potential therapies for the
acute treatment of migraine headaches, processes needed to
be developed to synthesize the ether-linked diarylmethanols
3
1
and 2.
*
To whom correspondence should be addressed. E-mail: magnus_nicholas@
lilly.com.
(1) For reviews on the pharmacology of Metabotropic Glutamate Receptors,
7
see: (a) Monn, J. A.; Schoepp, D. D. Annu. Rep. Med. Chem. 1997, 32, 1.
diarylmethanol 1 in 80% overall yield from 9.
(b) Schoepp, D. D.; Jane, D. E.; Monn, J. A. Neuropharmacology 1999,
In order to rapidly prepare multikilogram amounts of 1
in “fixed” pilot-plant reactors, the plan was to develop the
3
8, 1431.
(
2) (a) Lam, A. G. M.; Soriano, M. A.; Monn, J. A.; Schoepp, D. D.; Lodge,
D.; McCulloch, J. Neurosci. Lett. 1998, 254, 121. (b) Kingston, A. E.;
O’Neill, M. J.; Lam, A.; Bales, K. R.; Monn, J. A.; Schoepp, D. D. Eur. J.
Pharmacol. 1999, 377, 155. (c) Helton, D. R.; Tizzano, J. P.; Monn, J. A.;
Schoepp, D. D.; Kallman, M. J. J. Pharmacol. Exp. Ther. 1998, 284, 651.
3) Aicher, T. D.; Cortez, G. S.; Groendyke, T. M.; Khilevich, A.; Knobelsdorf,
J. A.; Magnus, N. A.; Marmsater, F. P.; Schkeryantz, J. M.; Tang, T. P.
(4) Boaz, N. W.; Fox, K. M. J. Org. Chem. 1993, 58, 3042.
(5) Guthikonda, R. N.; Cama, L. D.; Quesada, M.; Woods, M. F.; Salzmann,
T. N.; Christensen, B. G. J. Med. Chem. 1987, 30, 871.
(6) Parham, W. E.; Anderson, E. L. J. Am. Chem. Soc. 1948, 70, 4187.
(7) (a) Corey, E. J.; Niwa, H.; Knolle, J. J. Am. Chem. Soc. 1978, 100, 1942.
(b) Gala, D.; Steinman, M.; Jaret, R. S. J. Org. Chem. 1986, 51, 4488. (c)
Miyashita, M.; Yoshikoshi, A.; Grieco, P. A. J. Org. Chem. 1977, 42, 3772.
(
(Eli Lilly and Company, U.S.A.). PCT Int. Appl. 2006, WO 2006057870
A1 20060601.
5
60
•
Vol. 11, No. 3, 2007 / Organic Process Research & Development
10.1021/op0602268 CCC: $37.00 © 2007 American Chemical Society
Published on Web 01/17/2007

Scheme 1
Scheme 1 chemistry for scale-up and replace the chiral
chromatography with an asymmetric synthesis method. The
initial goal was an asymmetric preparation of diarylmethanol
efficient asymmetric aryl transfer process.10 The Peric a` s
group employed solution IR to monitor the rate of phenyl
transfer from diphenylzinc (DPZ), relative to the phenyl
transfer rate from mixtures of DPZ and diethylzinc (DEZ),
with and without an amino alcohol catalyst present, using
p-tolualdehyde as the electrophile. The kinetic study showed
that DPZ transferred phenyl to the aldehyde at 0 °C with no
ligand present, mixtures of DPZ and DEZ did not transfer
phenyl or ethyl to the aldehyde at 0 °C until an amino alcohol
catalyst was present (i.e., no background reaction), and in
the mixed zinc species experiments the phenyl group
transferred preferentially. The optimal ratio of DPZ to DEZ
to ensure high conversion of the aldehyde and selective
phenyl transfer was determined to be 1.32 equiv of DEZ to
7
(Scheme 1). To achieve this goal, asymmetric reduction
of the appropriate ketone and resolution of 6 via enzymatic
methods were pursued as the highest probability solutions
for preparations of 7. In addition to these approaches, the
recent advances in asymmetric aryl transfer chemistry of aryl
zinc species to aromatic aldehydes was difficult to discount
8
as a possible solution. The work of Bolm’s group involving
arylboronic acids as precursors to aryl zinc species was
particularly attractive since it permitted the use of the Scheme
1
aryl bromide 4 as the precursor to the requisite, known
9
boronic acid 13 (Scheme 2). The Bolm methodology of
0
.64 equiv of DPZ (i.e., 2 to 1), which equates to about a 28
converting arylboronic acids to aryl zinc species uses about
mol % excess of phenyl groups. In order to answer the
question of whether this more desirable aryl group stoichi-
ometry from the Peric a` s work could extrapolate to the Bolm
arylboronic acid technique of generating aryl zinc species,
the following research was conducted.
1
30 mol % excess of arylboronic acid; therefore, reduction
8
of this amount would be a focus for process development.
The recent work of the Peric a` s group also appeared to have
important information relevant to the development of an
The arylboronic acid 13 was prepared according to the
literature procedure (Scheme 2). The reports from Bolm and
(
8) Schmidt, F.; Stemmler, R. T.; Rudolph, J.; Bolm, C. Chem. Soc. ReV. 2006,
5, 454 and references therein.
9
3
(9) Zheng, N.; Armstrong, J. D., III; Kan K. E.; Keller, J.; Liu, T.; Purick, R.;
Lynch, J.; Hartner, F. W.; Volante, R. P. Tetrahedron Asymmetry 2003,
(10) Fontes, M.; Verdaguer, X.; Sol a` , L.; Peric a` s, M. A.; Riera, A. J. Org. Chem.
2004, 69, 2532.
1
4, 3435.
Vol. 11, No. 3, 2007 / Organic Process Research & Development
•
561

Scheme 2
Scheme 3
Peric a` s led us to the design of this initial reaction: a toluene
mixture of arylboronic acid 13, DEZ, the polymer additive
dimethoxypolyethyleneglycol-m2000 (DiMPEG), and the
relatively simple amino alcohol ligand (R)-(-)-2-piperdino-
1,1,2-triphenylethanol, 15,8,11 was reacted with 3-cyanoben-
zaldehyde 5 (Scheme 2). The result of this first attempt was
(11) Sol a` , L.; Reddy, K. S.; Vidal-Ferran, A.; Moyano, A.; Peric a` s, M. A.; Riera,
A.; Alvarez-Larena, A.; Piniella, J. F. J. Org. Chem. 1998, 63, 7078.
562
•
Vol. 11, No. 3, 2007 / Organic Process Research & Development

to produce the diarylmethanol 7 in about 95% yield with
>
94% ee.
While the Scheme 2 chemistry had given a desirable yield
and enantioselectivity in the production of 7, there were some
potential areas for development to make it more useful for
the preparation of kilogram amounts: (1) improve efficiency
by reducing or removing the poorly soluble DiMPEG
polymer additive; (2) reduce the equivalents of ligand 15;
(3) reduce the equivalents of arylboronic acid 13; (4) reduce
the equivalents of DEZ; (5) reduce the reaction volumes
which were 300 mL/g of aldehyde 5, and reduce the workup
volume which was 700 mL/g of aldehyde 5; (6) remove the
chromatography that had been required to purify 7 due to
the excesses of reagents present (crystallization of 7 was not
an option due to its being an oil).
To address these development needs, a multigram prepa-
ration of arylboronic acid 13 was conducted with the slight
modification of using n-hexyllithium instead of n-butyl-
lithium to avoid the dangers of butane release during
processing. During the proton NMR characterization of this
multigram batch of arylboronic acid 13, the integration of
the -OH group protons did not add up to the required
amount of 2 when run in deuterated DMSO as had been
9
reported in the literature characterization. The proton NMR
spectrum of 13 was next collected in deuterated chloroform,
and the -OH group protons were completely absent. Since
toluene was the reaction solvent that 13 was to be reacted
in, the proton NMR was also gathered in deuterated toluene
which also indicated the absence of -OH protons. In
addition, the literature preparation of 13 reported an IR
Figure 1. Liquid IR absorbance changes and ethane gas
evolution observed during the addition of DEZ to boroxine 16
in toluene.
with caustic, followed by filtration, afforded an unoptimized
70% recovery of the ligand 15. Subsequent chiral HPLC
analysis and subjection of the recovered ligand 15 to the
aryl transfer reaction protocol conclusively proved it to be
recyclable.
Optimization of the equivalents of boroxine 16 and DEZ
began with an attempt to analyze the reaction between these
components in toluene. Initially, the boron-zinc exchange
reactions between boroxine 16 and DEZ were performed per
the literature references, with a 15-20 h stir period at 60
-
1
spectrum with no bands beyond 3000 cm which was not
1
2
in keeping with a typical IR spectrum for a boronic acid.
After reviewing the spectroscopic data associated with 13,
it became apparent that the structure of arylboronic acid 13
was not correct and that the actual structure was the boroxine
16 (Scheme 3). This structural understanding was critical
for the pending stoichiometry optimization due to the
significant molecular weight differences between 13 and 16.
The DiMPEG polymer was removed from the Scheme 2
aryl transfer reaction, and at the original ligand 15 loading
of 15 mol % the enantioselectivity dropped from >94% to
°
C with no method of analysis. In an attempt to ensure
reproducibility, solution IR (ReactIR 4000 instrument with
DiComp probe) and a real-time gas analyzer-mass spec-
trometer (RTGA-MS) were used to better understand the
reaction between boroxine 16 and DEZ in toluene. During
the DEZ addition to the boroxine 16 in toluene, the
temperature rose from 20 to 26 °C, and the IR scans showed
91%. However, if the ligand 15 loading was increased to
20%, the enantioselectivity of the aryl transfer returned to
>
94%. A ligand 15 loading of 30% gave 96% enantiose-
lectivity, and a ligand 15 loading of 40% gave a modest gain
in enantioselectivity to 96.8% which indicated that increases
in ligand 15 loading did not have a linear relationship with
respect to increases in enantioselectivity. Removal of the
DiMPEG polymer led to a substantial reduction in reaction
volume, and afforded clean layer separations during the
workup which had proven problematic previously. In addi-
tion, the piperidine-based ligand 15 could now be removed
easily from the crude reaction mixtures with aqueous HCl
washes, an event that had been complicated by the presence
of the DiMPEG polymer. Those acidic extracts were es-
sentially pure ligand 15-hydrochloride, and neutralization
-
1
consumption of a band at 1408 cm and production of a
-
1
band at 1289 cm with little or no change within minutes
after the addition was complete. In addition, the RTGA-
MS data indicated a rapid decay in the evolution of ethane
from the system after the DEZ addition was complete. The
top half of Figure 1 illustrates the absorbance changes
observed by IR as a result of the DEZ addition. The bottom
half of Figure 1 illustrates the monitoring of ethane evolution
and decay as a result of the DEZ addition, and mass spectra
shown in Figure 2 confirm that the gas monitored was ethane
and not ethylene. Subsequent heating of the reaction mixture
(
12) Siverstein, R. M.; Bassler, G. C.; Morrill, T. C. Spectrometric Identification
of Organic Compounds, 4th Ed.; John Wiley & Sons: New York, 1981;
pp 112-120.
-
1
to 60 °C caused the new band at 1289 cm to be consumed
within 0.5 h (Figure 3), and the resulting mixture was found
Vol. 11, No. 3, 2007 / Organic Process Research & Development
•
563

Figure 2. Mass spectrum of gas evolved during the addition of DEZ to boroxine 16 in toluene, and comparison to standards of
ethane and ethylene to confirm the identity.
to be proficient in the aryl transfer chemistry. Based on the
results of these experiments, the current 15-20 h period at
had been consumed, additional small aliquots of aldehyde 5
were added to the reactions over time, allowing each to be
consumed. The HPLC analysis after each aldehyde 5 addition
revealed that the boroxine 16 equivalents could be reduced
to 0.45 (35 mol % excess of aryl equivalents) with good
production of diarylmethanol 7 and enantioselectivties
maintained at >94%. These results were comparable to the
optimal DPZ stoichiometry described by the Peric a` s group.10
Attempts to reduce the boroxine 16 equivalents to lower than
0.45 led to undesired ethyl transfer to the aldehyde 5. In
addition, it was found that the amount of DEZ could not be
60 °C after the DEZ addition could be reduced to a 2 h period
at 60 °C to affect the boron-zinc exchange with no concerns
of reproducibility due to an incomplete reaction.
In order to optimize the amount of boroxine 16 used for
the preparation of diarylmethanol 7, reactions were set up
essentially as depicted in Scheme 2 with 0.78 equiv of
boroxine 16 (130 mol % excess of aryl equivalents) relative
to aldehyde 5, but without the DiMPEG polymer. After
HPLC analysis confirmed that all of the starting aldehyde 5
564
•
Vol. 11, No. 3, 2007 / Organic Process Research & Development

Figure 3. IR waterfall plots correspond to the addition of diethyl zinc solution to the boroxine 16 in toluene between 20 °C to 26
-
1
°
C, and the subsequent heatup to 60 °C. The band at 1289 cm that is produced during the DEZ addition is consumed at 60 °C.
reduced to much lower than 2.8 equiv (relative to the
boroxine 16), from the original 7.7 equiv, without deleterious
affects to the conversion of aldehyde 5. The effect of
temperature was also examined in the production of diaryl-
methanol 7, and it was found that no improvements in
conversion or enantioselectivity were observed below -10
boroxine 16 with results comparable to those observed in
the laboratory.
To simplify the stereochemistry and improve functional
group compatibility, acetate was chosen as a replacement
3
for the previous THP protecting group for 7. The CH CN
stock solution of 7 was reacted with acetic anhydride in the
presence of TEA and catalytic DMAP (DMAP was necessary
for good conversion) to afford the acetate 18 (Scheme 3). It
was found that the reaction mixture of 18 could be treated
with aqueous HCl directly to effect removal of the TBS
group and give the oil primary alcohol 19 (Scheme 3). Once
again, taking advantage of the immiscibility of heptane with
°C, and that temperatures of 10 °C or above led to undesired
ethyl transfer to the aldehyde 5 and slightly lower enanti-
oselectivities.
With the reagent stoichiometry optimized, the workup of
the enantioselective aryl transfer reaction to produce a stock
solution of the oil diarylmethanol 7 was pursued. To cope
with the zinc waste produced in this chemistry, the reaction
mixture was treated with wet acetic acid to directly precipitate
CH
of 19 was extracted with heptane. After the extraction, the
CH CN was replaced via azeotropic distillation for toluene,
3
CN, the TBS-OH that was generated in the production
Zn(OAc)
granular solid (dry acetic acid as the quench produced slimy
Zn(OAc) which was difficult to remove). After removal of
2 2
‚2H O (mp 237 °C) which is a readily filtered
3
and ammonium salts generated from the prior chemistry were
removed by water washing. The toluene solution of 19 was
reacted with MsCl in the presence of TEA to form the
mesylate 20 and precipitate TEA-hydrochloride. Due to the
insolubility of TEA-hydrochloride in the primarily toluene
reaction media, the chloride derivative of 20 was negligible,
and this obviated the need for the Finkelstein halogen,
halogen exchange that had been used in the Scheme 1
chemistry. The TEA-hydrochloride was washed away with
water to give a toluene solution of 20 which was diluted
with acetone and reacted with the phenolic compound 11 in
the presence of potassium carbonate to give acetate 2
2
the zinc waste, the ligand 15 could be extracted into aqueous
HCl for subsequent recycling, and this operation further
removed salts. After salt and ligand 15 removal, the
remaining impurity in the primarily toluene organic phase
was the benzyloxy-tert-butyl-dimethyl-silane “proteo-deriva-
tive” 17 (Scheme 3). The silane 17 was not tolerated well in
the subsequent chemistry; thus, it was necessary to remove
it from the toluene solution containing 7. This was achieved
by employing azeotropic distillation to replace the toluene
with CH
3
CN followed by selectively extracting 17 from the
CN and heptane are essentially
CH CN with heptane (CH
3
3
immiscible). The result of these workup operations was to
produce a solution of 7 in 88% yield with >94% ee. As a
result of the optimizations to the aryl transfer reaction and
the workup, the maximum reaction volumes which were 300
mL/g of aldehyde 5 went down to 37 mL/g, and the
maximum workup volume which was 700 mL/g of alde-
hyde 5 went down to 60 mL/g. This reaction was run in a
pilot plant using 17.6 kg of aldehyde 5 and 47.6 kg of
(Scheme 3). The acetate 2 was afforded in 73% yield over
four steps after crystallization from EtOH in >96% ee.
Recrystallization of acetate 2 from a mixture of MTBE and
heptane increased the ee to >99% with an 83% yield. With
procedures developed to prepare kilograms of acetate 2 a
more complete assessment of potential mGlu2 receptor
potentiator therapeutics can now be realized.
Vol. 11, No. 3, 2007 / Organic Process Research & Development
•
565

Conclusions
removed, and the mixture was cooled to -10 °C using an
ice/acetone bath. (R)-(-)-2-Piperidino-1,1,2-triphenylethanol
15 (6.7 g, 18.7 mmol) was dissolved in toluene (70 mL)
and added to the reaction mixture via syringe. Stirring was
continued for 30 min. 3-Cyanobenzaldehyde 5 (12.2 g, 93.3
mmol) was dissolved in toluene (40 mL) and transferred to
the addition funnel. This solution was added dropwise to
the reaction mixture while maintaining the pot temperature
below -5 °C. After 4 h at -10 to -5 °C, the reaction
mixture was quenched by the dropwise addition of a mixture
In short, the mGlu2 receptor potentiator precursor mol-
ecule 1 was prepared via a racemic synthesis combined with
a chiral chromatography as outlined in Scheme 1. In addi-
tion, an asymmetric enantioselective aryl transfer reaction
was developed to give ready access to kilograms of the chi-
ral diarylmethanol 7 and ultimately acetate 2 which is also
useful for the preparation of mGlu2 receptor potentiators
(Scheme 3).
Experimental Section
HPLC Method for Analyzing the Scheme 3 Synthesis
of Acetate 2. Column: Zorbax Rapid Resolution SB-C8,
2
of AcOH (59.0 mL, 1003 mmol) and H O (14 mL). The
resulting slurry was filtered, and the solids (zinc acetate
dihydrate) were rinsed with toluene (50 mL). The filtrate
was transferred to a separatory funnel and washed sequen-
4
.6 mm × 75 mm, 3.5 µm. Flow rate: 2 mL/min. Column
temperature: 30 °C. Wavelength: 220 nm. A ) 0.1% H
PO in Milli-Q water. B ) acetonitrile. Gradient: 80% A at
min to 10% A at 7 min. Hold at 10% A for 1 min. Return
3
-
tially with 0.5 M HCl (2 × 200 mL), H
2
O (2 × 100 mL),
4
0
.5 M NaOH (100 mL), and H O (100 mL). The organic
2
0
portion was concentrated in vacuo using a 45 °C bath. Crude
chiral diarylmethanol 7 (40.4 g) was recovered (theory )
to 80% A over 0.5 min and hold at 80% A for 0.5 min prior
to next injection. HPLC method for chiral analysis: Col-
umn: Chiralpak AD 250 mm × 4.6 mm, 10 µm particle
size. Flow rate: 1 mL/min. Column temperature: ambient.
Wavelength: 280 nm. Mobile phase: heptane/2-propanol/
TFA (60:40:0.01 v/v/v). Gradient: isocratic.
3
3.0 g). Crude 7 (40.4 g gross, 33.0 g theory, 93.3 mmol)
was dissolved in CH CN (330 mL) and transferred to a 1-L
3
separatory funnel. Heptane (66 mL) was added, the mixture
was shaken, and the layers were allowed to separate. This
extractive process was repeated with 5 × 33 mL of heptane,
2,4,6-Tris-[4-(tert-butyldimethylsilanyloxymethyl)-phen-
allowing ∼15 min separation time per extract. The CH
3
CN
9
yl]-cyclotriboroxine (16). Under a nitrogen atmosphere at
3 °C, (4-bromobenzyloxy)-tert-butyldimethylsilane 4 (390.0
g, 1.29 mol), anhydrous THF (3.90 L), anhydrous toluene
0.83 L), and trisopropylborate (117.7 g, 0.625 mol) were
(
lower) layer was weighed (314.3 g) and held for assay and
2
direct carry-thru to the next step. HPLC analysis employing
a standard curve based upon purified oil 7 indicated that the
(
CH
88.3% yield). Chiral HPLC assay results: 95.5% ee. H
NMR (DMSO-d , 500.0 MHz): δ 0.04 (s, 6H), 0.87 (s, 9H),
.64 (s, 2H), 5.75 (d, 1H, J ) 3.8 Hz), 6.08 (d, 1H J ) 3.8
Hz), 7.22 (d, 2H, J ) 7.2 Hz), 7.33 (d, 2H, J ) 7.2 Hz),
.48 (t, 1H, J ) 7.2 Hz), 7.65 (m, 2H), 7.78 (s, 2H).
S)-Acetic Acid [4-(tert-butyl-dimethyl-silanyloxy-
methyl)-phenyl]-(3-cyano-phenyl)-methyl Ester (18). Un-
der a nitrogen atmosphere, a flask was charged with a CH
CN solution of 7 (28.35 g, 80.2 mmol). Acetic anhydride
10.6 g, 104.2 mmol), triethylamine (11.4 g, 112.4 mmol),
3
CN solution contained 29.1 g of the desired product
combined. The resulting mixture was stirred at ambient
temperature for 30 min, then cooled to -78 °C using a dry
ice/acetone bath. A 2.3 M hexane solution of n-hexyllithium
1
(
6
4
(195.2 g, 0.634 mol) was transferred to an addition funnel,
then added dropwise to the above mixture over 2.5 h. The
reaction mixture was allowed to warm to -20 °C and
quenched by the addition of 2 M HCl (1.32 L). After stirring
at 0 °C for 0.5 h, EtOAc (2.1 L) was added, and the mixture
was stirred for an additional 0.5 h. Stirring was stopped, and
the layers were separated. The organic portion was washed
7
(
3
-
(
with 5% aqueous NaHCO
to a volume of 3 L. This solution was treated with CH
3 L) and was reconcentrated to 3 L volume. This process
3
and then concentrated in vacuo
and DMAP (0.23 g, 1.8 mmol) were added to the reaction
mixture, and stirring was continued for 1 h. This reaction
3
CN
(
mixture of 18 was used directly in the next step without
was repeated 2 more times, resulting in a white suspension.
The mixture was filtered, and the resulting solids were
washed with CH
1
workup. H NMR (CDCl
3
, 500.0 MHz) δ 0.10 (s, 6H), 0.94
(s, 9H), 2.18 (s, 3H), 4.73 (s, 2H), 6.85 (s, 1H), 7.27 (d, 2H,
3
CN (0.5 L). After vacuum drying the solids
J ) 8 Hz), 7.32 (d, 2H, J ) 8 Hz), 7.44 (t, 1H, J ) 8 Hz),
7.56 (m, 2H), 7.64 (s, 1H).
(
(
(
45 °C, 48 h), 245 g of 16 was recovered as a white powder
76.0%). H NMR (CDCl
1
3
, 500 MHz) δ 0.13 (s, 18H), 0.98
(S)-Acetic Acid (3-cyano-phenyl)-(4-hydroxymethyl-
s, 27H), 4.85 (s, 6H), 7.46 (d, 6H, J ) 8 Hz), 8.20 (d, 6H,
1
3
phenyl)-methyl Ester (19). To a flask containing a CH CN
3
J ) 8 Hz). C NMR (100 MHz, CDCl
2
3
) δ -5.23, 18.43,
5.96, 64.95, 125.42, 128.76, 135.67, 146.27. IR (KBr) 1611,
solution of 18 (31.7 g, 80.2 mmol) was added 5 M HCl (28
mL, 140 mmol). The resulting mixture was stirred at 23 °C
for 2.75 h, transferred to a separatory funnel, and extracted
with heptane (3 × 320 mL). The tert-butyl-dimethyl-silanol-
containing heptane extracts were discarded. Toluene (476
mL) and D.I. water (320 mL) were added to the remaining
(CH CN) layer, and after shaking, the layers were allowed
3
to separate. The aqueous layer was back extracted with
toluene (320 mL). The organic phases were combined and
-
1
1
410, 1369, 1347, 1300, 1250, 1087, and 836 cm .
S)-3-{[4-(tert-Butyl-dimethyl-silanyloxymethyl)-phen-
yl]-hydroxy-methyl}-benzonitrile (7). The boroxine 16
30.1 g, 40.4 mmol) was charged to a reactor and subjected
(
(
to five alternating vacuum/nitrogen purges via a Firestone
apparatus and was finally purged under a nitrogen flow.
Diethylzinc (DEZ) (1.1 M) in toluene (312.3 mL, 343.5
mmol) was transferred to an addition funnel via cannula and
was added to 16. After 5 min, heating was initiated with a
set point of 60 °C. After 17 h at 60 °C, the mantle was
washed with saturated NaHCO
3
(320 mL) followed by D.I.
water (320 mL). The solution was concentrated by vacuum
566
•
Vol. 11, No. 3, 2007 / Organic Process Research & Development

distillation to a volume of 40 mL and then was diluted with
The solids were washed with cold absolute EtOH (0 °C, 34
mL) and further dried in a vacuum oven at 45 °C. Compound
2 (24.5 g) was isolated as an off-white solid (72.9% from
7). Chiral assay: 96.7% ee.
toluene (286 mL). The resulting mixture of 19 was held for
1
transfer to the next step. H NMR (CDCl
3
, 400.0 MHz) δ
2
.17 (s, 3H), 4.68 (s, 2H), 6.85 (s, 1H), 7.30 (d, 2H, J ) 8
Hz), 7.37 (d, 2H, J ) 8 Hz), 7.46 (t, 1H, J ) 8 Hz), 7.56
m, 2H), 7.63 (s, 1H).
S)-Acetic Acid (3-Cyano-phenyl)-(4-methanesulfo-
Procedure for ee Upgrade of 2. Under a nitrogen
atmosphere, a flask was charged with crude 7 (24.5 g, 53.6
mmol) and MTBE (123 mL). The suspension was heated to
reflux, held for 10 min, and then allowed to cool to 27 °C.
Heptane (50 mL) was charged to an addition funnel and then
added to the above mixture over 20 min. The resulting
mixture was stirred at 23 °C for 2 h and filtered. The crystals
were washed with 1:1 MTBE/heptane (50 mL), and vacuum-
dried at 45 °C to afford 20.4 g (83.1%) of off-white crystals
of 2. Chiral assay: 99.4% ee. mp (DSC) (10 °C/min) onset
(
(
nyloxymethyl-phenyl)-methyl Ester (20). Under a nitrogen
atmosphere, a flask was charged with a toluene solution of
alcohol 19 (20.7 g, 73.6 mmol) followed by triethylamine
(12.3 mL, 88.4 mmol), and the resulting mixture was cooled
to 0 °C. Methanesulfonyl chloride (6.6 mL, 84.6 mmol) was
charged to an addition funnel and added dropwise to the
above mixture over 30 min. The reaction mixture was stirred
at 0 °C for 1 h and transferred to a separatory funnel. After
washing with D.I. water (2 × 50 mL), the toluene solution
2
1
96.18 °C, peak 100.05 °C; [R]
D
20.1 (c ) 1.0 DMSO); IR
(KBr pellet) 3466, 3085, 3065, 2967, 2933, 2867, 2225,
1
-1
1
of mesylate 19 was held for transfer to the next step. H
6
1740, 1620, 1497, and 1417 cm ; H NMR (DMSO-d ,
NMR (CDCl
s, 2H), 6.85 (s, 1H), 7.36 (d, 2H, J ) 8 Hz), 7.42 (d, 2H,
J ) 8 Hz), 7.46 (t, 1H, J ) 8 Hz), 7.55 (d, 1H, J ) 8 Hz),
.59 (d, 2H, J ) 8 Hz), 7.63 (s, 1H).
S)-Acetic Acid [4-(4-Acetyl-3-hydroxy-2-propyl-
phenoxymethyl)-phenyl]-(3-cyano-phenyl)-methyl Ester
2). Under a nitrogen atmosphere, a flask was charged with
3
, 500.0 MHz) δ 2.19 (s, 3H), 2.97 (s, 3H), 5.22
500.0 MHz): δ 0.84 (t, 3H, J ) 7 Hz), 1.42-1.49 (m, 2H),
2.15 (s, 3H), 2.54 (s, 3H), 2.55-2.58 (m, 2H), 5.21 (s, 2H),
6.68 (d, 1H, J ) 8 Hz), 6.83 (s, 1H), 7.42 (d, 2H, J ) 8
Hz), 7.45 (d, 2H, J ) 8 Hz), 7.55 (t, 1H, J ) 7 Hz), 7.72-
(
7
13
(
7.77 (m, 3H), 7.91 (s, 1H), 12.82 (s, 1H); C NMR (DMSO-
6
d , 100 MHz) δ 14.4, 21.3, 21.9, 24.4, 26.8, 69.5, 75.6, 104.4,
(
112.1, 114.3, 117.3, 118.9, 127.3, 127.8, 130.3, 130.4, 131.6,
a toluene solution of mesylate 19 (26.4 g, 73.6 mmol).
Acetone (344 mL), 2′4′-dihydroxy-3′-propyl-acetophenone
131.7, 132.1, 137.2, 139.8, 142.6, 161.5, 162.5, 169.9, 204.4;
+
HRMS (AP ; accurate mass) calcd for C28
H
27NO
5
457.1889,
5
: C, 73.51; H,
1
1 (12.9 g, 66.5 mmol), and K
2
CO
3
(10.2 g, 73.9 mmol)
found 457.1892. Anal. Calcd for C28
H27NO
were charged to the reaction flask, and the resulting mixture
was heated to 60 °C and stirred for 6.5 h. An additional
charge of 2′4′-dihydroxy-3′-propyl-acetophenone (0.67 g, 3.4
mmol) was made to the reaction mixture, and stirring was
continued for 5.5 h. Heating was stopped, and the flask
contents were allowed to cool to 23 °C and then were filtered.
The solids were rinsed with toluene (75 mL), and the
resulting filtrate was concentrated in vacuo to a total volume
of 132 mL. The solution was transferred to a separatory
funnel and washed with D.I. water (2 × 132 mL). The
organic portion was further concentrated in vacuo to 60 mL.
To the still warm solution (70 °C), absolute EtOH (240 mL)
was added and the mixture again concentrated to 234 mL.
To the still warm solution (70 °C), absolute EtOH (66 mL)
was added and the resulting solution concentrated to 100
mL. To the still warm solution (70 °C), absolute EtOH (190
mL) was added and the mixture allowed to cool slowly to
5.95; N, 3.06. Found: C, 73.39; H, 5.95; N, 3.00.
Acknowledgment
Special thanks to the process chemistry group at Carbo-
Gen AG, Aarau, Switzerland, and in particular to Dr. Marc
Lanz and Dr. Roland Keilitz, for further developing and
demonstrating the Scheme 3 chemistry in their scale-up
facility. In addition, Mrs. Amanda L. McDaniel and Dr. Mark
A. LaPack are thanked for the solution IR and RTGA-MS
experiments and interpretation.
Supporting Information Available
Experimental procedures and spectral data for the Scheme
1
chemistry to prepare 1. This material is available free of
charge via the Internet at http://pubs.acs.org.
Received for review October 30, 2006.
OP0602268
23 °C. The resulting suspension was stirred for 15 h and
filtered.
Vol. 11, No. 3, 2007 / Organic Process Research & Development
•
567