Synthesis of a novel farnesyl transferase inhibitor, ABT-100; Selective preparation of a stereogenic tertiary carbinol

Article abstract of DOI:10.1016/j.tet.2005.02.072

A stereoselective synthesis of ABT-100 1, a novel farnesyl transferase inhibitor, is described. The key step involves a stereoselective addition of the heterocyclic zinc reagent 10 to chiral α-keto ester 9 in >10:1 diastereoselectivity using menthol as th

Full text of DOI:10.1016/j.tet.2005.02.072

Tetrahedron 61 (2005) 4419–4425
Synthesis of a novel farnesyl transferase inhibitor, ABT-100;
selective preparation of a stereogenic tertiary carbinol
*
Michael J. Rozema, Albert W. Kruger, Bridget D. Rohde, Bhadra Shelat, Lakshmi Bhagavatula,
James J. Tien,† Weijiang Zhang‡ and Rodger F. Henry
GPRD Process Research and Development, 1401 Sheridan Rd., Dept. R450, Bldg. R8-1, Abbott Laboratories, North Chicago,
IL 60064-6285, USA
Received 23 December 2004; revised 21 February 2005; accepted 24 February 2005
Available online 16 March 2005
Abstract—A stereoselective synthesis of ABT-100 1, a novel farnesyl transferase inhibitor, is described. The key step involves a
stereoselective addition of the heterocyclic zinc reagent 10 to chiral a-keto ester 9 in O10:1 diastereoselectivity using menthol as the chiral
auxiliary. Crystallization of the product as the dimeric zinc complex facilitates isolation of product in O99:1 dr. The biaryl linkage is formed
by the use of a Suzuki coupling employing only 0.06 mol% of the catalyst. Coupling of the two fragments is accomplished using a SNAr
reaction between diol 5 and arylfluoride 4. The overall yield for the five step sequence is 37% on kilogram scale.
q 2005 Elsevier Ltd. All rights reserved.
1. Introduction
further biological evaluation. Key to our success was the
development of a method for the generation of the chiral
quaternary alcohol bearing a heterocyclic substituent.
Mutation of the ras-oncogene regulating cell growth and
proliferation is implicated in up to 25% of human cancers.1
After transcription of the protein and further activation by
normal ras-protein activation processes (cysteine-farnesyla-
tion, cleavage of a tripeptide and C-terminal methylation),
the mutated ras-protein drives uncontrolled cell growth and
proliferation.2 One strategy for disruption of this process is
by inhibition of the farnesylation step which is mediated by
farnesyl transferase (FT). ABT-100 13 has been identified as
an FT inhibitor possessing excellent potency, bioavailability
and pharmacokinetics. Herein, we disclose our development
of a kilogram-scale process to prepare ABT-100 to support
2. Results and discussion
Racemic ABT-1003 is readily generated by the non-
selective addition of imidazolyl Grignard reagent 2 to
ketone 3 (Eq. 1). While the enantiomers are readily
separated by chromatography on small scale (up to ca.
5 g),4 the limited solubility of rac-ABT-100 in typical
mobile phases severely limits the scalability of this method.
In addition, experiments conducted to effect a classical
resolution of the enantiomers by the formation of chiral salts
found no promising leads. Therefore, we desired to develop
an efficient and robust means to generate the single
enantiomer of ABT-100.
Keywords: Farnesyl-transferase; FTI; Inhibitor; Chiral keto-ester; Asym-
metric addition; Chiral tertiary alcohol; Chiral quaternary alcohol;
Organozinc reagent; Lewis acid; Suzuki coupling; Palladium coupling.
*
Corresponding author. Tel.: C1 847 935 3490; fax: C1 847 938 2258;
e-mail: michael.rozema@abbott.com
† Present address: ARYx Therapeutics 2255 Martin Ave. Suite F, Santa
Clara, CA 95050, USA.
(1)
‡ Present address: Theravance, Inc. 901 Gateway Boulevard, South San
Francisco, CA 94080, USA.
After a brief examination of enantioselective additions of 2
to ketone 3 without success, we shifted to the disconnection
0040–4020/$ - see front matter q 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2005.02.072

4420
M. J. Rozema et al. / Tetrahedron 61 (2005) 4419–4425
strategy shown in Eq. 2. We reasoned biaryl 4 could be
assembled from commercially available starting materials
using a Suzuki protocol. Diol 5 could come from the
diastereoselective addition of an imidazolyl organometallic
to a ketone of the type 6 or 7 with an appropriate chiral
auxiliary,5,6 and the two coupled together using an SNAr
reaction (Eq. 2).7
by Seebach.9 Thus, treatment of 8 with (K)-menthol
(1.5 equiv) in the presence of a catalytic amount of titanium
(IV) ethoxide (15 mol%) in xylene at 80 8C provided
(K)-menthyl ester 9 in 79% yield. The reaction was pushed
to completion by distillation of the ethanol generated by
passing a stream of nitrogen through the reaction vessel.
Using fewer than 1.5 equiv of menthol led to incomplete
conversion, presumably because the ethoxide ligands on the
catalyst were exchanged for menthol during the reaction.
While menthyl ester 9 is crystalline, it was difficult to
separate the product from the excess (K)-menthol without
resorting to a chromatographic purification. To render the
(K)-menthol inert in the next step, the hydroxy function
was protected in situ as an acetate. This conversion was best
accomplished by the addition of acetic anhydride directly to
the reaction mixture after cooling and then reheating to
80 8C. After an aqueous work-up, the mixture of menthyl
ester 9 and menthyl acetate is carried into the next reaction
as a concentrated solution in toluene.
(2)
Initial studies into the diastereoselective addition to
(K)-menthyl a-keto ester 9 employed Grignard reagent 2,
prepared by magnesium-iodide exchange using ethyl-
magnesium chloride,10 and resulted in 11 in a 2.3:1
diastereomeric ratio (dr). Pretreatment of 9 with magnesium
bromide etherate at low temperature (K40 8C) afforded
slightly better selectivity (dr 4:1). In general, literature
examples of the addition of Grignard reagents to
(K)-menthyl a-keto esters give the best results when the
Grignard reagent is used in the presence of zinc chloride.5a,b
In our hands, the addition of zinc chloride resulted in
precipitation of the Grignard reagent and low conversions,
although the stereoselectivity improved to O10:1. Reason-
ing that in most of these systems the Grignard reagent is first
transmetallated to the corresponding zinc reagent, direct
preparation of the imidazolyl-zinc reagent was pursued.
The diastereoselective addition of organometallics to a-keto
esters5 and to a lesser extent a-keto ethers6 bearing a chiral
auxiliary constitutes an efficient method for the stereo-
selective synthesis of tertiary alcohols. Auxiliaries such as
menthol,5a,b 8-phenyl-menthol,5c trans-2-substituted cyclo-
hexanols,5d,e carbohydrate5f and amino-indanol5g based
derivatives have found many useful applications. Use of
menthol as the chiral auxiliary provides many advantages:
low cost, availability, good selectivity (vide infra) and a
tendency for intermediates to be crystalline.
Starting from commercially available ethyl (4-cyano-
phenyl)-oxo-acetate 8 (Scheme 1), the (K)-menthyl ester
was prepared using a modification8 of the titanium (IV)
alkoxide-catalyzed transesterification conditions developed
Using the zinc activation procedure of Knochel,11 imida-
zolyl-zinc reagent 10 was conveniently prepared through
the direct insertion of zinc metal on laboratory scale (Eq. 3).
Scheme 1.

M. J. Rozema et al. / Tetrahedron 61 (2005) 4419–4425
4421
On larger scales, it was important to maintain adequate
stirring and to control the addition rate of imidazolyl iodide
12 to the activated zinc. If the concentration of 12 becomes
too high during the insertion reaction, a precipitate coats the
surface of the zinc metal and conversion to the imidazolzinc
reagent ceases. It was shown in laboratory experiments that
in THF the addition of 12 to 10 forms a stringy insoluble
substance, the exact nature of which was not determined,
but which is assumed to be a coordination complex between
the iodoimidazole nitrogen and imidazolylzinc reagent.12
Not surprisingly, it was also found that the stirring rate is
important for this heterogeneous reaction. Slow stir rates
increase the probability of the reaction stalling. Therefore,
controlling the addition rate of 12 and ensuring adequate
stirring of the heterogeneous reaction mixture led to an
efficient and reproducible reaction.
obtained in 65–75% yield and with a diastereomeric ratio of
O99.8:0.2.
Ester 11 was selectively reduced to diol 5 by reaction with
NaBH4 in methanolic THF at 50 8C (Scheme 1). Under
these conditions, concomitant reduction of the nitrile was
not observed. The only detected impurity was a trace
amount (!0.5%) of the nitrile methanolysis product.13
Nitrile reduction to the amine becomes more competitive
when stronger reducing agents are employed (NaBH4/
HOAc or LiBH4). Using the NaBH4 procedure, yields of
90–95% of pure 5 were routinely obtained.
The biphenyl subunit of ABT-100 was assembled from
boronic ester 13 and bromofluorobenzonitrile 14 (Eq. 4)
using a Suzuki protocol.14 While as little as 0.025 mol%
(Ph3P)2PdCl2 effects complete conversion in 15 h, to obtain
reasonable reaction times (!6 h), a larger amount of
catalyst (0.06 mol%) was typically used. Sodium bicarbon-
ate was used as the base in toluene/water mixtures under an
inert atmosphere. Biphenyl 415 was consistently produced in
98% yield and excellent purity.
(3)
In the absence of a Lewis acid, the imidazolylzinc reagent
10 does not react with the a-keto ester 9. Of the handful of
Lewis acids screened (ZnCl2, Zn(OTf)2, BF3$OEt2,
Ti(OEt)4, MgBr2$OEt2), only MgBr2$OEt2 effected the
addition of the imidazolyl moiety to the ketone. Thus, in the
presence of MgBr2$OEt2 in THF at 0 8C, the tertiary alcohol
11 was generated with a diastereomeric ratio of 10–11:1 and
in 75–85% yield. The selectivity of the addition under these
conditions is not greatly affected by temperature. Even at
elevated temperatures (50 8C), the selectivity decreases only
slightly to 7.5:1. At lower temperatures (K40 8C), the
reaction becomes gelatinous and difficult to stir and any
increase in selectivity is negated by lower conversions.
(4)
With the two coupling partners in hand, the aryl ether
formation was examined. The SNAr reaction could be
accomplished using a variety of bases (LiHMDS,
NaHMDS, KHMDS, KOtBu, KOH) in polar aprotic
solvents (DMF, DMSO). The reaction is conveniently
run with milled KOH in THF/DMSO at low temperature
(!15 8C) (Scheme 1). The choice of base and stoichiometry
influence not only the reaction rate but also the impurity
profile. As shown in Eq. 5, the alkylated diol 1 can react
with the excess base to form epoxide 15 and phenol 16.16
This degradation is more prevalent with the potassium
bases. However, using lithium bases results in slow
reactions and incomplete conversions. A balance of a
reasonable conversion with acceptable levels of decompo-
sition can be achieved by running the reaction at
temperatures below 15 8C. With KOH, trace amounts of
Due to the propensity of the imidazolyl moiety to bind to
zinc,12 the product is isolated as a 2:1 complex with zinc
chloride after washing the reaction mixture with saturated
aqueous NH4Cl and crystallizing from toluene. Using this
procedure, the diastereomeric ratio of the isolated ester was
increased to greater than 99:1. The structure of the zinc
complex has been established by single crystal X-ray
analysis of a sample crystallized from acetonitrile (Fig. 1).
In order to proceed to the reduction of the ester, the product
was liberated from the zinc complex. This was most
effectively accomplished by treatment of the complex with
the disodium salt of EDTA. Tertiary alcohol 11 is routinely
Figure 1.

4422
M. J. Rozema et al. / Tetrahedron 61 (2005) 4419–4425
nitrile hydrolysis products as well as a dialkylated product
are also seen.13 These impurities are rejected in the
precipitation of the freebase by the addition of methanol
and by final salt formation.
mixture was stirred for not less than 8 h. The reaction
progress was monitored by GC and was considered
complete when !1% ethyl ester remains. Upon completion,
the reaction mixture was cooled to ambient temperature.
Acetic anhydride (2.22 kg, 21.7 mol, 1.5 equiv) was added,
the reaction mixture was heated to 80 8C and stirred for not
less than 12 h. Upon completion, the reaction mixture was
cooled to ambient temperature and diluted with ethyl acetate
(44.2 kg). The ethyl acetate solution was washed twice with
a 10% HCl solution (2!19.6 L). The organic layer was
washed with a solution of NaHCO3 (0.98 kg) in distilled
water (19.6 L). Caution. CO2 evolution was observed
during neutralization. The organic layer was washed with
a solution of NaCl (3.92 kg) in distilled water (19.6 L). The
organic layer was filtered through a silica gel pad (1.96 kg)
and concentrated to an oil under vacuum. Toluene (8.65 kg)
was added and the resulting solution was concentrated under
vacuum to an oil. A second portion of toluene (8.65 kg) was
added and the resulting solution was concentrated under
vacuum to an oil to be used as is in the next reaction. A
portion of the oil was analyzed by HPLC and found to be ca.
50% 9 by weight (79% yield) Karl Fischer analysis of the oil
indicated the presence of 0.01% water. A sample of the
crude toluene solution was purified for characterization by
silica gel chromatography (95/5 to 90/10, hexanes/EtOAc).
The fractions that contained product were concentrated to
an oil that solidified. Mp 41.5–42.5 8C. 1H NMR (400 MHz,
CDCl3) d 8.10 (ddd, JZ8.4, 1.7, 1.5 Hz, 2H), 7.80 (ddd, JZ
8.4, 1.7, 1.5 Hz, 2H), 5.00 (td, JZ10.9, 4.5 Hz, 1H), 2.15
(m, 1H), 1.91 (m, 1H), 1.74 (m, 2H), 1.15 (m, 1H), 1.56 (m,
4H), 0.97 (d, JZ6.6 Hz, 3H), 0.91 (d, JZ7.0 Hz, 3H), 0.83
(d, JZ7.0 Hz, 3H) ppm. 13C NMR (100 MHz, CDCl3) d
184.2, 161.9, 135.3, 132.3, 129.9, 117.6, 117.3, 77.7, 46.9,
40.7, 34.2, 31.8, 26.5, 23.6, 22.2, 20.9, 16.5. IR (MIC) 2232,
1714, 1699 cmK1. [a]DZK84.4 (cZ0.963, MeOH). Anal.
Calcd for C19H23NO3, C 72.82, H 7.40, N 4.47; found C
72.66, H 7.53, N 4.29.
(5)
The final isolation consists of dissolving the unpurified
freebase in hot isopropanol, filtering and converting to the
HCl salt by the addition of aqueous HCl. The final product
can be recrystallized from EtOH to produce larger particles
with better handling properties.
In summary, we have developed a stereoselective and
scaleable synthesis of ABT-100 that produced material of
O99% ee and in an overall yield of 37% in five steps on
kilogram scale. The process is highlighted by a dia-
stereoselective addition of an imidazolylzinc reagent to an
a-ketoester to produce the stereogenic tertiary alcohol and
the formation of the biaryl ether through an SNAr reaction.
In addition, an efficient Suzuki reaction to produce the
biaryl moiety and a selective ester reduction to produce the
requisite diol were also developed.
3. Experimental
3.1. General
1H and 13C NMR spectra were taken in CDCl3 unless
otherwise indicated with CHCl3 (7.26 ppm) used as an
internal standard. All reactions were performed in appro-
priate glassware equipped with an overhead stirrer,
thermocouple, nitrogen inlet/outlet, and if necessary a
reflux condenser or distillation apparatus. All reactions
were conducted under positive nitrogen pressure. Commer-
cial grade anhydrous solvents and reagents were used
without any further purification. Analytical HPLC con-
ditions were Zorbax Eclipse XDB-C8 column with 5 mM
K2HPO4/5 mM KH2PO4 buffer (pH 7) and acetonitrile
mobile phase and detection at 205 nm unless otherwise
noted. Analytical GC was done on Alltech DB-1 column
and helium make-up gas. Infrared analyses were performed
on neat samples on microscope stage (MIC).
3.1.2. Imidazolylzinc iodide reagent (10). A suitable
reaction vessel was charged with zinc dust (1.51 kg,
23.1 mol, 4.5 equiv), THF (3.2 L) and 1,2-dibromoethane
(40 mL, 0.46 mol, 0.1 equiv). The resulting slurry was
heated to reflux and stirred for not less than 30 min with
effervescence observed due to the evolution of ethylene.
The reaction mixture was then cooled to !50 8C, the
chlorotrimethylsilane (56 mL, 0.44 mol) was added and
stirring was continued for not less than 5 min with a slight
exotherm of 0.5 8C observed. After heating the reaction
mixture back to reflux, a solution of 5-iodo-1-methyl-1H-
imidazole 12 (1.6 kg, 7.7 mol, 1.5 equiv) in THF (9.6 L)
was then added dropwise over 6 h using a metering pump.
The reaction mixture was sampled periodically to ensure
that reaction is proceeding, monitoring by HPLC the
conversion of 5-iodo-1-methyl-1H-imidazole 12 to
methyl-1H-imidazole (hydrolyzed zinc reagent 10)). After
an additional 15 min, the reaction was cooled to ambient
temperature, and the stirring was stopped to allow the excess
zinc dust to settle. The solution of 10 was decanted from the
excess zinc for use in the next step.
3.1.1. (4-Cyanophenyl)-2-oxo-acetic acid (1R,2S,5R)-
menthyl ester (9). A suitably equipped reaction vessel
was charged with ethyl 4-cyanobenzoylformate 8 (2.94 kg,
14.5 mol), xylene (2.94 kg), (1R,2S,5R)-(K)-menthol
(3.39 kg, 21.7 mol, 1.5 equiv), and Ti(OEt)4 (494.9 g,
2.17 mol, 0.15 equiv). Stirring was initiated under a
nitrogen atmosphere and the mixture was then heated
80 8C. The ethanol produced was distilled off with the aid of
a constant N2 flow through the reaction mixture. The
3.1.3. S-(4-Cyanophenyl)-hydroxy-(3-methyl-3H-imida-
zol-4-yl)-acetic acid (1R,2S,5R)-menthyl ester (11).
Addition of zinc reagent 10 to ketone 9. A suitable reaction

M. J. Rozema et al. / Tetrahedron 61 (2005) 4419–4425
4423
vessel was charged with THF (12 L) and magnesium
bromide diethyl etherate (1.32 kg, 5.1 mol, 1.0 equiv).
Note. The dissolution of magnesium bromide etherate is
exothermic and best performed by adding the solid to the
solvent. An external cooling bath is used to keep the
temperature below 35 8C. After cooling the resulting slurry
to below 25 8C, the (4-cyanophenyl)-oxo-acetic acid
menthyl ester 9 (3.12 kg, 51% potent; 1.59 kg, 5.08 mol)
was added. After stirring the resulting suspension for 30 min
at less than 25 8C, the solution of zinc reagent 10 was added
using a metering pump at such a rate to keep the reaction
temperature below 30 8C. The reaction mixture was stirred
at 25 8C and was considered complete when the ratio of
product to starting material remained unchanged by HPLC.
The reaction was quenched with a solution of NH4Cl
(3.6 kg) in water (12 L) and filtered through celite (350 g,
washing the pad with THF, 4 L). The layers were separated
and the organic layer washed with a solution of NH4Cl
(3.6 kg) and water (12 L). The THF layer was distilled down
to approximately 16 L then toluene (32 L) was added. The
volume was concentrated down to approximately 16 L by
distillation. The resulting suspension was heated to 90 8C
for not less than 8 h, and allowed to cool slowly over 16 h to
ambient temperature. The zinc complex of 11 was then
collected by filtration and washed twice with toluene (8 and
4 L). The complex was dried by a N2 flow. A sample
suitable for X-ray analysis was obtained by slow
evaporation from acetonitrile.
6.77 mol, 2.0 equiv) followed by THF (13.3 L, 10 vol). To
the resulting slurry was added MeOH (2.69 L, 2 mL/g) over
not less than 30 min in 3 portions. Caution. A large amount
of headspace should be allowed due to the large amount of
gas evolution in the quench step. A slow addition of MeOH
is used to minimize a small exotherm (6–7 8C) that is
otherwise experienced. The reaction mixture was then
warmed to 50G10 8C over 20–30 min. After less than 0.1%
of ester remains by HPLC (Zorbax Extend-C18 column,
mobile phase 10 mM NH4OH/MeOH), the mixture was
cooled to less than 30 8C and slowly quenched with aqueous
citric acid (40% w/w, 13.3 L) maintaining the temperature
below 40 8C. Caution. Care should be exercised in the rate
of addition of the quenching solution because the large
amount of heat and gas evolution can cause frothing and
foaming. After stirring overnight (12–18 h), the reaction
mixture was extracted with heptane (2!13.3 L), mixing the
contents for not less than 10 min. After the addition of THF
(13.3 L), a 50% w/w aqueous KOH solution was used to
adjust the pH to between 8 and 9. A water bath was used
to maintain temperature below 40 8C. The layers were
separated after the temperature is below 30 8C. To the
organic layer was added a solution of 40% w/w aqueous
citric acid (6.7 L) and the THF was removed by distillation.
The distillation was complete when the amount of THF
remaining was less than 1.6% (v/v) by GC analysis.
Crystallization of 5 was accomplished by adjusting the pH
to 8.5G0.5 using an aqueous KOH solution (50% w/w,
4.3 L). Note. A cooling bath was used to keep the internal
temperature below 35 8C. Diol 5 is collected by filtration,
washed with water (2!1.3 L) and dried in a vacuum oven at
60 8C for not less than 12 h to afford 0.74 kg of a white solid
(91%). The enantiomeric excess of the product was
determined to be 99.2% by HPLC analysis using a
Chiralpak AD column (4.6!250 mm; flow 1.0 mL/min;
mobile phase 80:20 hexane/ethanol; 23 8C; 210 nm;
9.09 min (major enantiomer) and 12.66 min (minor enan-
Decomplexation of zinc complex. The zinc complex of 11
was taken up in ethyl acetate (32 L) and THF (16 L). The
organic layer was washed three times with one-third of a
Na2EDTA solution prepared from Na2EDTA (5 kg) and
distilled water (95 L). The layers were allowed to stir for
20 min and then separated. The organic layer was then
filtered through celite (300 g), washing with ethyl acetate
(8 L), and then distilled down to 4 L. The resulting slurry
was heated to reflux to dissolve most of the solids. Heptane
(20 L) was added over 30 min maintaining a gentle reflux.
The resulting slurry was cooled to ambient temperature and
stirred for 10 h. Heptane (8 L) was added and stirred for 1 h.
The product was collected by filtration, washed with
heptane (4 L), and dried under N2 flow for 1 h. Product
was dried in vacuo at 50 8C for not less than 12 h to afford a
white solid 11 (1.38 kg, 68%). The diasteromeric ratio of the
product was determined to be 99.8:0.2 by HPLC analysis.
Mp 166.5–167.5 8C. 1H NMR (400 MHz, DMSO-d6) d 7.85
(dt, JZ8.6, 1.8 Hz, 2H), 7.64 (s, 1H), 7.49 (dt, JZ8.6,
1.8 Hz, 2H), 6.69 (d, JZ1.1 Hz, 1H), 7.07 (s, 1H), 4.61 (td,
JZ10.9, 4.3 Hz, 1H), 3.29 (s, 3H), 1.96 (m, 1H), 1.60 (m,
1H), 1.53 (m, 1H), 1.45 (m, 1H), 1.18 (m, 1H), 0.90 (m, 4H),
0.89 (d, JZ6.6 Hz, 3H), 0.60 (d, JZ6.9 Hz, 3H), 0.44 (d,
JZ6.9 Hz, 3H) ppm. 13C NMR (100 MHz, CDCl3) d 170.3,
144.7, 140.1, 131.6, 129.1, 127.4, 118.1, 112.1, 78.9, 75.4,
47.1, 40.6, 34.1, 33.3, 31.7, 25.7, 23.0, 22.2, 20.7, 15.7 ppm.
IR (KBr) 2960, 2240, 1740, 1460, 1200, 1090 cmK1. MS
(ESI) (MC1) 396. [a]DZK113.4 (cZ0.991, MeOH).
Anal. Calcd for C23H29N3O3, C 69.85, H 7.39, N 10.62;
found C 69.65, H 7.51, N 10.53. ICP ZnZ12 ppm.
1
tiomer). Mp 186.0–186.6 8C. H NMR (400 MHz, DMSO-
d6) d 7.70 (dt, JZ8.4, 1.8 Hz, 2H), 7.53 (dt, JZ8.4, 1.8 Hz,
2H), 7.53 (s, 1H), 7.2 (d, JZ1.0 Hz, 1H), 4.00 (d, JZ
11.3 Hz, 1H), 3.89 (d, JZ11.1 Hz, 1H), 3.30 (s, 3H) ppm.
13C NMR (100 MHz, DMSO-d6) d 149.1, 140.3, 134.4,
132.6, 128.1, 127.4, 119.3, 111.7, 75.1, 70.2, 33.5 ppm. IR
(MIC) 2220 cmK1. [a]DZK145.4 (cZ1.03, MeOH). Anal.
Calcd for C13H13N3O2, C 64.19, H 5.39, N 17.27; found C
63.93, H 5.70, N 17.10. MS (ACPI) (MC1) 244.0.
3.1.5. 6-Fluoro-40-(trifluoromethoxy)-[1,10-biphenyl]-3-
carbonitrile (4).16 A suitable reaction vessel was charged
with 3-bromo-4-fluorobenzonitrile 14 (2.63 kg, 12.9 mol),
4-trifluoromethoxyphenyl boronic acid 13 (3.06 kg,
14.7 mol, 1.1 equiv), bis(triphenylphosphine)-palladium(II)
chloride (6.0 g, 8.4!10K3 mol, 0.06 mol%), NaHCO3
(1.66 kg, 19.8 mol, 1.5 equiv), nitrogen-presparged toluene
(6.6 L), and presparged water (6.8 L). The reaction mixture
was heated to 75–85 8C until the aryl bromide was
consumed (!1%), as determined by HPLC analysis
(Zorbax Eclipse XDB-C8, 0.1% H3PO4/acetonitrile mobile
phase) (about 2 h). Note. Carbon dioxide gas evolution is
prominent as the reaction mixture is heated. Upon reaction
completion, the reaction mixture was cooled to 60–65 8C,
and the layers were separated above 45 8C. The organic
layer was filtered through a pad of silica gel (2.6 kg pre-wet
3.1.4. S-4-[1,2-Dihydroxy-1-(3-methyl-3H-imidazol-4-
yl)-ethyl]-benzonitrile (5). A suitable reaction vessel was
charged with 11 (1.34 kg, 3.39 mol) and NaBH4 (256 g,

4424
M. J. Rozema et al. / Tetrahedron 61 (2005) 4419–4425
with 5 L toluene), the silica gel pad was rinsed with toluene
(12.8 L) and the filtrates were concentrated. Note: Small
amounts of toluene are used to quantitate the transfers.
Residual toluene was removed from the resulting oil by
azeotropic distillation with ethanol–methanol (95:5;
3!6.4 L) until GC analysis showed that the percentage of
toluene in alcohol solvent was under 0.5%. The resultant
solid and/or solution was dissolved in ethanol–methanol
(95:5; 6.4 L) and heated to about 40 8C. The homogeneous
solution was removed from the heat, and water (6.4 L) was
added dropwise while the solution was stirred and allowed
to cool to room temperature. After cooling to room
temperature, the slurry was then filtered. The solid was
rinsed with 1:1 EtOH/water solution (12.8 L) at room
temperature and then rinsed with water (6.4 L). The white
solid was dried under nitrogen flow for approximately 1 h
and then dried in a vacuum oven with N2 stream at
maximum of 50 8C (25–30 mm Hg) for approximately 90 h
until the residual ethanol was minimized to afford 3.56 kg 4
(98% yield). Mp 62.2–62.9 8C. 1H NMR (400 MHz, CDCl3)
d 7.76 (dd, JZ7.0, 2.1 Hz, 1H), 7.68 (ddd, JZ8.4, 4.5,
2.2 Hz, 1H), 7.57 (m, 2H), 7.35 (m, 2H), 7.30 (dd, JZ9.7,
8.5 Hz, 1H) ppm. 13C NMR (100 MHz, CDCl3) d 161.9,
149.5, 134.9, 133.5, 131.9, 130.4, 129.5, 121.2 (2), 120.4,
117.8, 117.7, 109.1 ppm. IR (KBr) 2230, 1516, 1490, 1265,
1256, 1222, 1161 cmK1. HRMS (FAB): calcd for
C14H7F4NO [MC]: 281.0464, found 281.0466. Anal.
Calcd for C14H7F4NO, C 59.80, H 2.51, N 4.98; found C
59.74, H 2.56, N 4.90. ICP analysis: Pd !3 ppm, B !1 ppm,
Na 2 ppm.
internal temperature of the solution was adjusted to 60 8C
and 1 M HCl (37% HCl, 0.29 kg) in isopropanol (1.83 kg)
was added through the 0.2 mm in-line filter keeping the
temperature at not more than 65 8C. The internal tempera-
ture of the reaction mixture was adjusted to 0 8C and the
reaction mixture was stirred at 0 8C for 2 h. The precipitated
product was collected by filtration and washed with
isopropanol (6 kg, filtered through the in-line filter). The
solid was allowed to dry under a N2 flow for not less than
1 h. To achieve suitable particle size properties, the salt was
recrystallized from ethanol. The wet solid from the filter was
transferred into a clean reactor. Ethanol (35 kg) was charged
through a 0.2 mm in-line filter and the contents were heated
to reflux to dissolve the solids. The contents of the reactor
were distilled to a volume of approximately 20 L at
atmospheric pressure and the reaction mixture was stirred
with slow agitation at reflux for 3 h to help increase particle
size. The temperature of the slurry was adjusted slowly to
0 8C at a rate of approximately 10 8C/h because slow
cooling reduces the smaller size particle formation. The
slurry was stirred at 0 8C for 3 h. The solid product was
collected by filtration and dried under a N2 flow for 1 h. The
solid was dried in a dryer at 60G10 8C for 16 h under
nitrogen to afford 1.13 kg of a white flocculent powder
(75%) with 99.8% ee (HPLC, Chiralcel OD-RH column
(150!4.6 mm, 5 mm; flow 1.0 mL/min; mobile phase 50:50
20 mM KH2PO4/acetonitrile; ambient temperature; 234 nm;
6 min (minor enantiomer) and 9.5 min (major enantiomer)).
Mp 249.1–250.1 8C. 1H NMR (500 MHz, DMSO-d6) d
15.00 (s, 1H), 8.00 (d, JZ1.5 Hz, 1H), 7.86 (dd, JZ8.7,
2.3 Hz, 1H), 7.75 (d, JZ2.3 Hz, 1H), 7.67 (m, 2H), 7.47 (d,
JZ8.9 Hz, 1H), 7.42 (m, 2H), 7.37 (m, 4H), 7.18 (s, 1H),
4.76 (d, JZ10.1 Hz, 1H), 4.61 (d, JZ10.1 Hz, 1H), 3.37 (s,
3H) ppm; 13C NMR (125 MHz, DMSO-d6) d 158.1, 147.7,
145.5, 137.4, 134.8, 134.2, 134.0, 133.9, 131.9, 131.3,
129.5, 127.3, 123.2 (CF3), 121.1 (CF3), 120.4, 119.1 (CF3),
118.7, 118.4, 118.4, 117.0 (CF3), 113.8, 110.6, 103.7, 74.1,
3.1.6. 6-[2-(4-Cyano-phenyl)-2-hydroxy-2-(3-methyl-3H-
imidazol-4-yl)-ethoxy]-40-trifluoromethoxy-biphenyl-3-
carbonitrile (1). A suitable reaction vessel submersed in a
cooling bath was charged with diol 5 (0.71 kg, 2.92 mol),
biaryl fluoride 4 (1.07 kg, 3.81 mol, 1.3 equiv), and milled
KOH (0.20 kg, 3.20 mol, 1.1 equiv). The bath temperature
was adjusted to not more than 10 8C and THF (1.90 kg) was
charged. The slurry was stirred and the internal temperature
was adjusted to not more than K5 8C. Dimethylsulfoxide
(2.48 kg) was charged slowly to the reaction mixture. Note:
The reaction was slightly exothermic with the addition of
DMSO. The milky solution was stirred for not less than 5 h
at K2 8C. The reaction mixture was stirred for not less than
36 h at 14 8C. The reaction was judged to be complete if diol
5 peak area% was less than 3% by HPLC (Zorbax-C8
column 20 mM NH4OAc/acetonitrile mobile phase). The
internal temperature of the reaction mixture was lowered to
not more than 0 8C and a 20% methanol in water solution
(19 kg) was slowly charged keeping the temperature at not
more than 20 8C. Note. The quench of reaction mixture was
initially exothermic. The solid was collected by filtration
and washed with methanol (6 kg) and distilled water (6 kg).
The wet cake was allowed to dry under a N2 flow for 1 h.
The solid was washed with a 20% ethyl acetate in heptane
solution (2!5 kg) and then heptane (10 kg). The solid as
the crude free-base was allowed to dry under a N2 flow for
not less than 1 h. A suitable reaction vessel was charged
with the crude solid and isopropanol (40 kg). The
temperature was adjusted to 70 8C and the reaction mixture
was stirred until most of the solids dissolved. The resulting
solution was filtered into a suitable vessel through a 0.2 mm
in-line filter, rinsing the reactor with isopropanol (3 kg). The
72.0, 34.8 ppm. IR (MIC) 3274, 3160, 2232, 1261 cmK1
.
[a]DZK57.8 (cZ1.02, MeOH). Anal. Calcd for C27H20-
ClF3N4O3, C 59.95, H 3.73, N 10.36; found C 59.74, H 3.60,
N 10.29. ICP !20 ppm for all metals.
Acknowledgements
We thank Cathie Linton, Eric Kristensen and Mike Moore
for analytical support. We also thank Mike Fitzgerald for the
preparative separation of the enantiomers of the ABT-100
and Howard Morton for his helpful discussions.
References and notes
1. Bos, J. L. In Molecular Genetics in Cancer Diagnosis;
Cossman, J., Ed.; Elsevier: USA, 1990; p 273.
2. Prendergast, G. C.; Rane, N. Expert Opin. Investig. Drugs
2001, 10, 2105.
3. Claiborne, A. K.; Gwaltney, S. L., II; Hasvold, L. A.; Li, Q.;
Li, T.; Lin, N.- H.; Mantei, R. A.; Rockway, T. W.; Sham, H.
L.; Sullivan, G. M.; Tong, Y.; Wang, G.; Wang, L.; Wang, X.;
Wang, W.-B. Patent Appl. US2003/0087940 A1, May 8, 2003.

M. J. Rozema et al. / Tetrahedron 61 (2005) 4419–4425
4425
4. Conditions for the semi-preparative HPLC: Chiracel OD,
20 mDP (300 g), hexane–ethanol (70:30), 100 mL/min,
ambient temperature, 700 mg racemic material in 40 mL
warm ethanol injections.
Curr. Chem. 1975, 59, 33–64. (f) Burnnet, J. F.; Zahler, R. E.
Chem. Rev. 1951, 49, 273.
8. Krasik, P. Tetrahedron Lett. 1998, 39, 4223.
9. Seebach, D.; Hungerbuehler, E.; Naef, R.; Shnurrenberger, P.;
Weidman, B.; Zueger, M. Synthesis 1982, 138.
¨
5. (a) Boireau, G.; Deberly, A.; Abenhaım, D. Tetrahedron Lett.
¨
1988, 29, 2175. (b) Boireau, G.; Deberly, A.; Abenhaım, D.
10. (a) Holden, K. G.; Mattson, M. N.; Cha, K. H.; Rapoport, H.
J. Org. Chem. 2002, 67, 5913. (b) Panosyan, F. B.; Still,
I. W. J. Can. J. Chem. 2001, 79, 1110. (c) Burm, B. E. A.;
Blokker, P.; Jongmans, E.; van Kampen, E.; Wanner, M. J.;
Koomen, G.-J. Heterocycles 2001, 55, 495.
Tetrahedron 1989, 45, 5837. (c) Whitesell, J. K.; Deyo, D.;
Bhattacharya, A. J. Chem. Soc., Chem. Commun. 1983, 802.
(d) Basavaiah, D.; Bharathi, T. K. Tetrahedron Lett. 1991, 32,
3417. (e) Basavaiah, D.; Krishna, P. R. Tetrahedron 1995, 51,
12169. (f) Loupy, A.; Monteux, D. A. Tetrahedron 2002, 58,
1541. (g) Senanayake, C. H.; Fang, K.; Grover, P.; Bakale,
R. P.; Vandenbossche, C. P.; Wald, S. A. Tetrahedron Lett.
1999, 40, 819.
11. (a) Yeh, M. C. P.; Knochel, P. Tetrahedron Lett. 1988, 29,
2395. (b) Knochel, P.; Yeh, M. C. P.; Berk, S. C.; Talbert, J.
J. Org. Chem. 1988, 53, 2392.
12. Wilkinson, G., Ed.; Comprehensive Coordination Chemistry;
Pergamon: Oxford, 1987; Vol. 5, p 925.
6. Charette, A. B.; Benslimane, A. F.; Mellon, C. Tetrahedron
Lett. 1995, 36, 8557.
13. Evidence for this impurity comes from LC–MS data and by
independent synthesis.
7. For recent examples of aryl-ether synthesis using the SNAr
reaction on electron deficient fluoro-aromatics see: (a) Temal-
Laib, T.; Chastanet, J.; Zhu, J. J. Am. Chem. Soc. 2002, 124,
583. (b) Stark, H.; Sadek, B.; Krause, M.; Hu¨ls, A.; Ligneau,
X.; Ganellin, C. R.; Arrang, J.-M.; Schwartz, J.-C.; Schunack,
14. For reviews, see: (a) Miyaura, N.; Suzuki, A. Chem. Rev. 1995,
95, 2457. (b) Suzuki, A. J. Organomet. Chem. 1999, 576, 147.
15. Biphenyl 4 is referenced in a Japanese patent with no
procedure on its synthesis or characterization. Ikeda, A.;
Ozaki, M.; Honami, R.; Yumita, T.; Yano, H.; Nakano, Y;
Kurihara, Y.; Hirano, T. Preparation of triazole derivatives as
insecticides and acaricides. PCT Int. Appl, 1994.
´
W. J. Med. Chem. 2000, 43, 3987–3994. (c) East, S. P.; Joullie,
M. M. Tetrahedron Lett. 1998, 39, 7211. (d) For general
reviews on the SNAr reaction, see Wells, K. M.; Shi, Y.-J.;
Lynch, J. E.; Humphrey, G. R.; Volante, R. P.; Reider, P. J.
Tetradedron Lett. 1996, 37, 6439. (e) Zoltewicz, J. A. Top.
16. The labile epoxide 15 was never isolated. Evidence for its
existence comes from LC–MS data.