Synthesis of the CETP inhibitor torcetrapib: The resolution route and origin of stereoselectivity in the iminium ion cyclization

Article abstract of DOI:10.1021/op060014a

A practical, efficient synthesis of (-)-(2R,4S)-4-[(3,5-bis- trifluoromethyl-benzyl)methoxycarbonylamino]-2-ethyl-6-trifluoromethyl-3, 4-dihydro-2H-quinoline-1-carboxylic acid ethyl ester (1), a cholesteryl ester transfer protein (CETP) inhibitor, is desc

Full text of DOI:10.1021/op060014a

Organic Process Research & Development 2006, 10, 464−471
Synthesis of the CETP Inhibitor Torcetrapib: The Resolution Route and Origin
of Stereoselectivity in the Iminium Ion Cyclization
David B. Damon,† Robert W. Dugger,*,†, George Magnus-Aryitey,‡ Roger B. Ruggeri,‡ Ronald T. Wester,‡
Meihua Tu,§ and Yuriy Abramov*,§,|
Chemical Research and DeVelopment, Department of Medicinal Chemistry, Exploratory Medicinal Chemistry,
Pfizer Global Research and DeVelopment, Groton Labs, Eastern Point Road, Groton, Connecticut 06340, U.S.A.
Abstract:
A practical, efficient synthesis of (-)-(2R,4S)-4-[(3,5-bis-tri-
fluoromethyl-benzyl)methoxycarbonylamino]-2-ethyl-6-tri-
fluoromethyl-3,4-dihydro-2H-quinoline-1-carboxylic acid ethyl
ester (1), a cholesteryl ester transfer protein (CETP) inhibitor,
Figure 1. Structure of torcetrapib (CP-529,414).
is described. The key reaction in the synthesis, addition of an
N-vinylcarbamate to an iminium ion rapidly followed by an
iminium ion cyclization onto the aryl ring, sets up the cis
relationship of the two subsituents of the tetrahydoquinoline
ring of 6. The origin of the high cis stereoselectivity in the
cyclization was explored using high-level quantum chemistry
calculations.
Scheme 1
Results and Discussion
Synthesis. Key to the synthesis of 1 is the construction
of the tetrahydroquinoline nucleus with the cis stereochem-
istry of the substituents. The aza Diels-Alder reaction of
N-arylimines and N-vinyl compounds (Scheme 1) seemed
ideal since it is often associated with a high degree of cis
selectivity.3
Imine 3 (Scheme 2) was generated in situ in CH2Cl2 from
p-trifluoromethylaniline4 and n-propanal using TiCl4 as a
water scavenger. Addition of vinyl carbamate 55 and catalytic
amounts of BF3‚OEt2 produced 6 with no detectable amount
of the trans isomer (by NMR); however, the yield was low
(40-60%), and the purity of the product was modest
(∼70%). After several attempts at improving the yield by
varying conditions it became clear to us that imine 3 was
not being formed cleanly; significant amounts of aminal 4
were also formed, thus leading to the low yields and low
purity.
Since imine 3 seemed to be unstable towards further
additions, we sought to trap 3 with a nucleophile that could
subsequently regenerate the imine in situ in the next
step. The variation reported by Katritzky6 in which the imine
is trapped as its benzotriazole adduct seemed ideal. In
Katritzky’s work, the cis isomer was also the only stereo-
isomer observed.
Introduction
Despite continuing advances in their understanding and
treatment, cardiovascular diseases remain the leading cause
of morbidity and mortality in the industrialized world. While
most of the existing therapies in this area address the risk
factors of hypertension and elevated low-density lipoprotein
cholesterol (LDL-C), there remains a paucity of treatments
for raising levels of high-density lipoprotein cholesterol
(HDL-C). Pronounced elevation of HDL-C levels observed
in subjects lacking a functional copy of the gene for
cholesteryl ester transfer protein (CETP) offered strong
evidence that an inhibitor of CETP might provide a therapy
for raising low levels of HDL-C and the treatment of
cardiovascular diseases. Torcetrapib (1, CP-529,414) is an
inhibitor of CETP that is currently being developed for the
treatment of atherosclerosis and coronary heart disease. The
success of this program required a practical method for
preparing torcetrapib on a large scale in a reliable manner.
Synthetic routes used to prepare materials to support the
clinical development of torcetrapib are described in this
article1 and in the following article.2
* To whom correspondence should be addressed. E-mail: robert.w.dugger@
pfizer.com, yuriy.a.abramov@pfizer.com.
In our hands, stirring aniline 2 with n-propanal and
benzotriazole in toluene followed by addition of n-heptane
Correspondence regarding the synthetic aspects of this paper should be
directed to R. Dugger (robert.w.wdugger@pfizer.com).
| Correspondence regarding the computational aspects of this paper should
be directed to Y. Abramov (yuriy.a.abramov@pfizer.com).
† Chemical Research and Development.
(3) (a) Crousse, B.; Be´gue´, J.; Bonnet-Delpon, D. J. Org. Chem. 2000, 65,
5009. (b) Stevenson, G.; Leeson, P.; Rowley, M.; Sanderson, I.; Stansfield,
I. Bioorg. Med. Chem. Lett. 1992, 2, 371.
(4) Hazards associated with the storage of p-trifluoromethylaniline have been
described. Users of this compound are urged to take the appropriate
precautions. See Tickner, D.; Kasthurikrishnan, N. Org. Process Res. DeV.
2001, 5, 270.
(5) am Ende, D. J.; DeVries, K. M.; Clifford, P. J.; Brenek, S. J. Org. Process
Res. DeV. 1998, 2, 382.
(6) Katrizky, A. R.; Rachwal, B.; Rachwal, S. J. Org. Chem. 1995, 60, 3993.
‡ Department of Medicinal Chemistry.
§ Exploratory Medicinal Chemistry.
(1) See Damon, D. B.; Dugger, R. W. Eur. Pat. Appl. EP1125929, Aug. 22,
2001 and U.S. Patent 6,313,142, Nov. 6, 2001 for the original disclosure of
this work.
(2) Damon, D. B.; Dugger, R. W.; Hubbs, S. E.; Scott, J. M.; Scott, R. W.
Org. Process Res. DeV. 2006, 10, 472 (the following article in this issue).
464
•
Vol. 10, No. 3, 2006 / Organic Process Research & Development
10.1021/op060014a CCC: $33.50 © 2006 American Chemical Society
Published on Web 03/16/2006

Scheme 2
Scheme 3. Partial cyclization of 7 and 5
genolysis conditions using ammonium formate as the reduc-
ing agent.7 We screened numerous chiral acids to find a
suitable salt for resolution of 11. Dibenzoyl-L-tartrate was
the best lead identified from our initial screen. Aqueous
ethanol was identified as the optimal solvent system. Using
3% water (w/w) gave the best results, a salt of 98% de in
39% yield (78% of theory). NMR integration of the salt
indicated that it was a hemi-tartrate salt (2 equiv of amine
per tartrate). The structure and absolute stereochemistry of
12 were confirmed by an X-ray crystallographic structure
determination.8
Breaking the salt and isolation of the free amine were
often accompanied by the formation of byproducts. Fortu-
nately, we found that isolation of the free amine was
unnecessary. Slurrying salt 12 in dichloroethane and treating
with aqueous NaOH produced a dichloroethane solution of
the free base that could be dried and carried directly into
the reductive amination reaction. Adding aldehyde 13 and
sodium triacetoxyborohydride9 to the dichloroethane solution
resulted in conversion to desired benzylamine 14. We found
it convenient to isolate the product of the reductive amination
as the toluenesulfonate salt by dissolving the crude free base
in a mixture of methanol and diisopropyl ether and adding
toluenesulfonic acid. This process provided 14 in 86% yield
from 12. Installing the final carbamate by treating 14 with
methylchloroformate in THF with Na2CO3 as the base to
neutralize the toluenesulfonic acid and HCl produced 1 in
high yield. The final product is very soluble in most organic
solvents but can be crystallized from concentrated ethanol
solutions as the monoethanolate in 74% yield.
resulted in 7 crystallizing from the reaction mixture in 91%
yield (two crops). With a crystalline imine surrogate in hand,
we next investigated its reaction with vinyl carbamate 5.
Treating a toluene solution of 7 and 5 with 1 mol %
toluenesulfonic acid resulted in a 78% yield of 6. Again, no
trans isomer was detectable. This method of producing 6
also has the advantage of not requiring the use of the
moisture-sensitive Lewis acids, TiCl4 and BF3‚OEt2.
Interestingly, in one run using a low-quality batch of 5
we observed the formation of a large amount of a byproduct
(Scheme 3). The byproduct was isolated and shown to be 8
(a 1:1 mixture of diastereomers). Resubjecting 8 to the
standard cyclization conditions provided 6 in excellent yield
(92%). Again, only the cis isomer was observed. This
observation is consistent with the stepwise process reported
by Katritzky.6
Acylation of 6 (Scheme 4) is difficult due to steric
hindrance of the adjacent ethyl group and the low nucleo-
philicity of the aniline nitrogen due to the electron-withdraw-
ing effect of the p-trifluoromethyl group. A variety of
reaction conditions were investigated to find the optimum
conditions. Pyridine was found to be the best base. The best
yield of 10 was obtained when an excess of ethyl chloro-
formate (5 equiv) was slowly added to the reaction mixture
over a period of 4 h. Aqueous workup followed by
crystallization provided 10 in 83% yield. Although multigram
quantities of 10 could be separated into its enantiomers via
chiral chromatography, for the quantities we needed for
development of this candidate we felt that a classical
resolution would be more expedient. Amine 11 was prepared
in excellent yield (93%) by hydrogenolysis of the benzyl-
carbamate protecting group of 10 under transfer hydro-
Origin of the Stereoselectivity. Although high cis
selectivity is precedented in the literature,3,4 it seemed unusual
that we could detect no trans isomer at all, and so it was
decided to utilize molecular modeling to gain a better
understanding of this process. High-level quantum chemistry
calculations were adopted to rationalize the cis selectivity
in the cyclization of the activated iminium ion 9. For
(7) Johnstone, R. A. W.; Wilby, A. H.; Entwistle, I. D. Chem. ReV. 1985, 85,
129.
(8) Details can be found in the Supporting Information.
(9) Abdel-Magid, A. F.; Maryanoff, C. A.; Carson, K. G. Tetrahedron. Lett.
1990, 31, 5595.
Vol. 10, No. 3, 2006 / Organic Process Research & Development
•
465

Scheme 4
Scheme 5. Cyclization pathways to cis and trans products
simplicity, the methyl carbamate 15 was modeled instead
of the benzyl carbamate (Scheme 5).
present a Curtin-Hammet kinetic system, in which a reactant
exists in two interconverting forms, each of which give
different products.10 To evaluate a product ratio for such a
system, a consideration of both reaction kinetics and ki-
netics of the conformational interconversion should be
performed.
Stereoselectivity, as a general phenomenon, arises from
free energy differences associated with diastereomeric transi-
tion states (kinetic control) or products (thermodynamic
control). There are four possible diastereomers of the
intermediate cyclization products of 15, two cis and two
trans. Only two of them, 16 and 18, were found to have the
lowest energies and became the focus of further modeling.
The calculated relative energy difference for those dia-
stereomers, as well as for the corresponding final products,
17 and 19, appeared to be -0.8 and -0.3 kcal/mol,
respectively. Since these results do not account for the
observed high stereoselectivity, the kinetic consideration was
pursued. For this, the formation of the intermediate cycliza-
tion products, 16 and 18, was assumed to be a rate-limiting
step.
The cyclization and conformational pathways are pre-
sented in Figure 4a,b. According to MP2 calculations, the
RC1 conformer is the global minimum of the potential
surface area and is separated from RC3 by 3.0 kcal/mol. The
energy difference increases up to 4.4 kcal/mol for the cis
and trans cyclization transition states. At the same time, the
absolute values of the activation energies appear to be very
low, especially for the cis cyclization. This result is a likely
consequence of the implicit solvation model (using the
nonionic solvent, DMSO), which underestimates the elec-
trostatic stabilization of the cation, 15, by a counterion. The
stabilization is less favorable for the reaction transition states
than for the RC1 and RC3 conformers, since the cationic
charge becomes more diffused as a result of a partial charge
transfer (of about 0.16 e- for both TScis and TStrans) be-
The three lowest-energy conformations of 15 (reactant
conformers, RC) are shown in Figure 2. Among those, RC1
and RC3 are precursors for the respective cis and trans
cyclization. The corresponding transition state (TS) geom-
etries for cyclization, TScis and TStrans, are shown in Figure
3. Conformations RC1 and RC3 and products 16 and 18
(10) Seeman, J. I. Chem. ReV. 1983, 83, 83.
466
•
Vol. 10, No. 3, 2006 / Organic Process Research & Development

Figure 2. Lowest-energy conformers of 15. RC1 and RC3 are precursors for respectively cis and trans cyclization. Distances
between atoms closing the ring are indicated in Å.
Figure 3. Transition-state geometries for the cis and trans cyclization. Distances between atoms closing the ring are indicated in
Å.
tween nucleophilic (trifluoromethylaniline) and electrophilic
(iminium side chain) fragments of the molecule. As a result,
the reaction activation barriers should increase by ap-
proximately an equal amount. To verify this consideration,
a simplified model study was performed. Each of the RC1,
RC3, TScis, and TStrans structures was explicitly solvated by
one Cl- counterion, displaying a close contact with the H(N)
hydrogen of the iminium cation. For this purpose, only
counterion positions were optimized, while keeping the
molecule rigid. This was followed by a single-point implicit
solvation of the neutral complex in toluene. The resulting
energy differences between the “ground” and “transition”
states were compared with the corresponding values obtained
at the same level of theory and adopting implicit DMSO
solvation with no counterion consideration. The estimated
increase of the activation energies for the trans and cis
cyclization, due to the counterion stabilization, was respec-
tively 2.5 and 3.4 kcal/mol (3.6 and 4.4 kcal/mol in gas
phase). The estimated change of the energy difference
between the reaction transition states was 0.2 kcal/mol.
TS31 displays the lowest barrier along the conformational
pathway between RC1 and RC3 (Figure 4b). The resulting
activation energies for the direct and reverse interconversion
between the conformers are 5.2 kcal/mol and 2.2 kcal/mol,
respectively. The latter energy is comparable with the trans
cyclization activation barrier of 1.9 kcal/mol, without any
adjustment on the counterion effect. However, the atomic
charge analysis of the TS31 reflects its higher polarity relative
to that of the RC1 and RC3 conformers: the charge
separation between nucleophilic and electrophilic fragments
is about 0.12 e- higher. As a result, an additional stabilization
of this transition state and a lowering of the interconversion
Vol. 10, No. 3, 2006 / Organic Process Research & Development
•
467

HOMO-LUMO gap from 3.20 eV, for the RC1, to 3.75
eV, for TScis.
In terms of the FMO theory, the observed stereoselectivity
may be rationalized as a result of the more favorable
secondary interaction between LUMO, at the N-CR bond
of the carbamate, and HOMO, at the C4-C3 bond of the
phenyl ring: the difference between N-C4 distances of the
TStrans and TScis transition states is 0.8 Å.
Conclusion
The route we developed to prepare 1 proceeds in 13.5%
overall yield over seven steps from p-trifluoromethylaniline.
It has been scaled several times to produce multikilogram
quantities needed for the development of this exciting drug
candidate. Although this process was suitable for early
development purposes, a better route was needed for larger-
scale manufacture. Investigations towards an improved route
are described in the following paper. The origin of the high
cis stereoselectivity in the cyclization was rationalized by
high-level quantum chemistry calculations. It was demon-
strated that the selectivity is kinetically controlled, originating
from both the ground-state conformational preference and
from the lower activation energy of the cis cyclization.
Computational Section
The origin of cis stereoselectivity of the cyclization prod-
uct of 15 (Scheme 5) was investigated by means of quantum
chemical calculations performed with the Gaussian03 pro-
gram.12 Initial conformational search for 15 and cyclization
products 16, 17, 18, and 19 was performed using the Monte
Carlo Multiple Minimum (MCMM) method, adopting
MMF94s force field with GBSA solvation as implemented
in Macromodel.13 It was followed by gas-phase geometry
optimization at the B3LYP/6-31G(d) level of theory.14
Transition states were calculated at the same level of theory
using a quadratic synchronous transit algorithm (STQN).15
The nature of all stationary points was confirmed by the
frequency calculations. GaussView 3.0916 was used to
animate the imaginary frequencies of the TS structures; this
was done to confirm that the displacement vectors are placed
Figure 4. Potential energy profile of 15: (a) cyclization
pathway; (b) conformational pathway.
barrier between RC3 and RC1 should take place due to the
electrostatic interaction with the counterion in solution.
On the basis of the above considerations, we assume that
the interconversion between the reactant conformers, RC1
and RC3, takes place faster (a lower barrier) than cyclization
towards the cis and trans products (higher barriers). In that
case, the cis-to-trans product ratio can be described by the
Arrenhius equation: cis/trans ) e-(∆G(TScis)-∆G(TStrans))/RT. The
predicted energy difference of 4.4 kcal/mol accounts for the
cis stereoselectivity of more than 99.9%. Therefore, the origin
of the stereoselectivity takes place from a noticeably more
stable reactant conformer and the lower activation energy
for the cis cyclization.
The cyclization reaction can be also qualitatively rational-
ized in terms of frontier molecular orbital (FMO) theory.11
Within this approach, the reaction is driven by the favorable
interaction between the highest occupied (HOMO) and the
lowest unoccupied (LUMO) molecular orbitals of the
respectively nucleophilic and electrophilic species. The
frontier orbitals of 15 are predominately located at the phenyl
(HOMO) and carbamate fragments (LUMO), displaying
maximum constructive overlap along the reaction path
(Figure 5). An intramolecular mixing of these orbitals during
the cyclization leads to electronic stabilization of the transi-
tion state; this is due to lowering the HOMO and raising the
LUMO energies. The net effect, at the PCM-B3LYP/6-
31+G(d,p) level of theory in DMSO, is an increase of the
(12) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M.
A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.;
Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.;
Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.;
Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.;
Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li,
X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C.;
Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.;
Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.;
Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich,
S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.;
Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.;
Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz,
P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.;
Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson,
B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian 03,
Revision B.04; Gaussian, Inc.: Wallingford CT, 2004.
(13) MacroModel; Schrodinger Inc.: Portland, OR
(14) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648. (b) Lee, C.; Yang, W.;
Parr, R. G. Phys. ReV. B 1988, 37, 785. (c) Hariharan, P. C., Pople J. A.
Theor. Chim. Acta 1973, 28, 213.
(15) (a) Peng, C.; Ayala, P. Y.; Schlegel, H. B.; Frisch, M. J. J. Comput. Chem.
1996, 17, 49. (b) Peng, C.; Schlegel, H. B. Isr. J. Chem. 1994, 33, 449.
(16) GaussView, rev. 3.0, Semichem, Inc. for Gaussian, Inc. 1998.
(11) Fukui, K. Acc. Chem. Res. 1981, 14, 363.
468
•
Vol. 10, No. 3, 2006 / Organic Process Research & Development

Figure 5. Frontier orbital analysis of the cis cyclization. The (0.05 au isosurfaces of the orbitals are shown.
along the reaction coordinates. Finally, the reported free
energies at the zero temperature were calculated at the
optimized geometries as: ∆G ) Eel + ZPE + ∆Gsol . The
electronic energies, Eel, were obtained by the single-point
MP2/6-31+G(d,p)17 calculations in gas phase. Zero-point
energy (ZPE) corrections at the B3LYP/6-31G(d) level were
scaled by a factor of 0.9806.18 The solvation free energies,
∆Gsol, were calculated in DMSO and in toluene, using the
PCM model19 at the B3LYP/6-31+G(d) level of theory,
adopting UAKS atomic radii.20 Modeling of the salt bridge
with Cl- anion was performed at the B3LYP/6-31+G(d)//
B3LYP/3-21G(d) level of theory. The atomic charges were
estimated according to the Mulliken population.
2.25 (m, 2H), 6.49 (m, 1H), 6.80 (d, 2H, J ) 8.7 Hz), 7.35
(m, 3H), 7.50 (m, 1H), 7.88 (d, 1H, J ) 8.3 Hz), 7.99 (m,
1H), 8.09 (d, 1H, J ) 8.5 Hz), none of the N-2 benzotriazole
isomer was observed;6 13C NMR (DMSO-d6, 100 MHz) δ
149.3, 146.1, 131.4, 127.7, 126.8, 125.3 (q, J ) 270 Hz),
124.4, 119.8, 118.2 (q, J ) 31.7 Hz), 112.9, 111.5, 71.0,
28.0, 10.2; DEPT spectrum: quaternary carbons δ 149.3,
146.1, 131.4, 125.3, 118.2; CH carbons δ 127.7, 126.8, 124.4,
119.8, 112.9, 111.5, 71.0; CH2 carbon δ 28.0; CH3 carbon
δ 10.2; IR (drifts) 3292 (s), 3038 (m), 2975 (m), 1621 (s),
1331 (s), 1320 (s), 1114 (vs); Anal. Calcd for C16H15N4F3:
C, 59.99; H, 4.72; N, 17.49. Found (first crop): C, 60.16;
H, 4.74; N, 17.86. Found (second crop): C, 59.97; H, 4.66;
N, 17.63.
Experimental Section
cis-(2-Ethyl-6-trifluoromethyl-1,2,3,4-tetrahydro-quin-
olin-4-yl)carbamic Acid Benzyl Ester (6). To a 1-L flask
under nitrogen atmosphere charged with 5 (27.66 g, 156
mmol, 1.0 equiv) were added dry toluene (500 mL), 7 (50.0
g, 156 mmol, 1.0 equiv), and p-toluenesulfonic acid mono-
hydrate (297 mg, 1.56 mmol, 0.01 equiv). The mixture was
heated to 70 °C. After 2 h, the mixture was cooled to room
temperature and transferred to a separatory funnel. Ethyl
acetate (500 mL) was added. The mixture was washed 1 ×
200 mL of 1 N NaOH, 1 × 200 mL of H2O, 1 × 200 mL of
brine, and dried (MgSO4). The mixture was filtered, and the
solids were washed 1 × 50 mL of ethyl acetate. The filtrate
was concentrated to ∼250 mL; 500 mL of toluene was added,
and the mixture was concentrated to ∼500 mL. Five hundred
milliliters of n-heptane was added, the slurry was stirred 1
h and filtered through a Buchner funnel and dried. 6 was
isolated as a white powder (45.04 g, 76%): mp 155-157
All materials were purchased from commercial suppliers
and used without further purification. All reactions were
conducted under an atmosphere of nitrogen unless noted
1
otherwise. H and 13C-spectroscopy was performed at 300
and 400 MHz on Bruker-Spectrospin Avance instruments.
(1-Benzotriazol-1-yl-propyl)-(4-trifluoromethyl-phenyl)-
amine (7). A 2-L flask under nitrogen atmosphere was
charged with benzotriazole (36.96 g, 310 mmol, 1.0 equiv)
and dry toluene (400 mL). A room-temperature solution of
p-trifluoromethylaniline4 (39.1 mL, 310 mmol, 1.0 equiv)
and 50 mL of toluene was added over 1 min. A room-
temperature solution of n-propanal (24.6 mL, 341 mmol, 1.1
equiv) and 50 mL of toluene was then added over 20 min.
There was an exotherm from 23 to 30 °C during this addition.
After stirring 24 h, n-heptane (500 mL) was added, and the
slurry stirred an additional 1 h. The suspension was filtered,
and the solids were washed with n-heptane (1 × 100 mL
and then 1 × 200 mL) and were dried. 7 was isolated as
shiny white needles (81.3 g, 82%). After 24 h, a second crop
was isolated from the filtrate (8.7 g, 9%). mp 130-132 °C;
1H NMR (DMSO-d6, 400 MHz) δ 0.82 (t, 3H, J ) 7.5 Hz),
1
°C; H NMR (DMSO-d6, 400 MHz) δ 0.92 (t, 3H, J ) 7.5
Hz), 1.5 (m, 3H), 2.00 (m, 1H), 3.35 (m, 1H), 4.77 (m, 1H),
5.07 (d, 1H, J ) 12.5 Hz), 5.15 (d, 1H, J ) 12.5 Hz), 6.35
(s, 1H), 6.61 (d, 1H, J ) 8.5 Hz), 7.12 (s, 1H), 7.18 (dd,
1H, J ) 1.9, 8.5 Hz), 7.4 (m, 5H), 7.70 (d, 1H, J ) 9.1 Hz);
13C NMR (DMSO-d6, 100 MHz) δ 157.03, 149.02, 137.79,
128.82, 128.23, 128.03, 125.9 (q, J ) 270 Hz), 125.0, 123.5,
121.7, 115.2 (q, J ) 31.7 Hz), 113.3, 65.8, 52.0, 47.8, 34.0,
28.6, 9.9; DEPT spectrum: quaternary carbons δ 157.0,
149.0, 137.7, 125.9, 121.7, 115.2; CH carbons δ 128.8, 128.2,
(17) (a) Møller, C.; Plesset, M. S. Phys. ReV. 1934, 46, 618. (b) Frisch, M. J.;
Head-Gordon, M.; Pople, J. A. Chem. Phys. Lett. 1990, 166, 275.
(18) Scott, A. P.; Radom, L. J. Phys. Chem. 1996, 100, 16502.
(19) (a) Cossi, M.; Scalmani, G.; Rega, N.; Barone, V. J. Chem. Phys. 2002,
117, 43. (b) Cances, E.; Mennucci, B.; Tomasi, J. J. Chem. Phys. 1997,
107, 3032.
(20) Barone, V.; Cossi, M.; Tomasi, J. J. Chem. Phys. 1997, 107, 3210.
Vol. 10, No. 3, 2006 / Organic Process Research & Development
•
469

128.0, 125.0, 123.5, 113.3, 52.0, 47.8; CH2 carbons δ 65.8,
34.0, 28.6; CH3 carbon δ 9.9; IR (drifts) 3430 (m), 3303
(s), 2951 (m), 1686 (vs), 1542 (vs), 1088 (vs); MS (APCI+)
m/z (rel intens) 379 (M + H+, 53), 228 (100); Anal. Calcd
for C20H21N2O2F3: C, 63.48; H, 5.59; N, 7.40; Found: C,
63.69; H, 6.06, N, 7.36.
was washed 1 × 125 mL of saturated aq. NaHCO3 and 1 ×
100 mL of brine, and dried (Na2SO4). The mixture was
filtered and the filtrate concentrated to a clear oil. The oil
was crystallized from 100 mL of n-heptane to give 11 as a
white crystalline solid (26.05 g, 93%): mp 61.5-63.5 °C;
1H NMR (CDCl3, 400 MHz) δ 0.79 (t, 3H, J ) 7.5 Hz),
1.24 (m, 4H), 1.42 (m, 1H), 1.51 (br s, 2H), 1.62 (m, 1H),
2.46 (m, 1H), 3.73 (m, 1H), 4.17 (m, 2H), 4.36 (m, 1H),
7.44 (m, 2H), 7.66 (m, 1H); 13C NMR (CDCl3, 100 MHz) δ
154.6, 139.3, 138.9, 126.3 (q, J ) 32 Hz), 125.7, 124.3 (q,
J ) 271 Hz), 123.5, 119.8, 61.6, 54.6, 46.1, 41.0, 28.5, 14.8,
9.0; DEPT spectrum: quaternary carbons δ 154.6, 139.3,
138.9, 126.3, 124.3; CH carbons δ 125.7, 123.5, 119.8, 54.6,
46.1; CH2 carbons δ 61.6, 41.0, 28.5; CH3 carbons δ 14.8,
9.0; IR (drifts) 3350 (s), 3293 (m), 2972 (s), 1697 (vs); MS
(ES+) m/z (rel intens) 358 (M + H + CH3CN+, 55), 317
(M + H+, 7), 300 (100); Anal. Calcd for C15H19N2O2F3: C,
56.96; H, 6.06; N, 8.86. Found: C, 56.86; H, 6.28; N, 8.82.
(-)-(2R,4S)-4-Amino-2-ethyl-6-trifluoromethyl-3,4-di-
hydro-2H-quinoline-1-carboxylic Acid Ethyl Ester Hemi-
(-)-dibenzoyl-L-tartrate Salt (12). A 1-L flask under
nitrogen atmosphere was charged with 11 (24.0 g, 75.9
mmol, 1.0 equiv) and (-) dibenzoyl-L-tartaric acid (anhy-
drous) (27.19 g, 75.9 mmol, 1.0 equiv). Ethanol (300 mL of
∼97%, prepared by adding 10.5 mL of H2O to 500 mL of
absolute ethanol, mixing, and measuring out 300 mL) was
added. The mixture was stirred at room temperature for 18
h and then filtered. The solids were washed 1 × 48 mL of
∼97% ethanol and dried to give 12 as a white crystalline
cis-4-Benzyloxycarbonylamino-2-ethyl-6-trifluoromethyl-
3,4-dihydro-2H-quinoline-1-carboxylic Acid Ethyl Ester
(10). A 3-L flask under nitrogen atmosphere was charged
with 6 (96.0 g, 254 mmol, 1.0 equiv), dry dichloromethane
(720 mL), and dry pyridine (103 mL, 1.27 mol, 5.0 equiv).
A solution of ethyl chloroformate (121 mL, 1.27 mol, 5.0
equiv), in dry dichloromethane (240 mL), was added slowly
over 4 h. The addition was exothermic and required a reflux
condenser. Once the chloroformate addition was complete,
the reaction was cooled in an ice bath and 1350 mL of 1 N
NaOH was added. The mixture was stirred 15 min and then
transferred to a separatory funnel. The layers were separated,
and the aqueous was extracted 1 × 1 L of dichloromethane.
The combined dichloromethane layers were washed 1 ×
1350 mL of 1 N HCl, 1 × 1 L of saturated aq NaHCO3, 1
× 1 L of brine, and dried (Na2SO4). The mixture was filtered,
and the filtrate concentrated to an orange oil. Five hundred
and seventy milliliters of abs. ethanol was added, and the
solution was concentrated. The solids were dissolved in 1370
mL of abs. ethanol, and 570 mL of H2O were added dropwise
over 45 min. The resultant thick slurry stirred 18 h and
filtered. The solids were washed with cold 7:3 abs. ethanol/
water (1 × 250 mL, then 1 × 100 mL) and dried (vac oven,
45 °C) to give 10 as a white, crystalline solid (94.54 g,
1
solid (14.77 g, 39%): mp 189.5-191.5 °C dec; H NMR
1
83%): mp 92-96 °C; H NMR (CDCl3, 400 MHz) δ 0.84
(DMSO-d6, 400 MHz) δ 0.62 (t, 3H, J ) 7.3 Hz), 1.16 (t,
3H, J ) 7.1 Hz), 1.3 (m, 3H), 2.5 (m, 1H), 4.1 (m, 4H),
5.63 (s, 1H, methine proton in DBTA), 7.47 (m, 2H, DBTA
aromatic H’s), 7.6 (m, 3H, DBTA aromatic H’s), 7.68 (s,
(t, 3H, J ) 7.4 Hz), 1.28 (t, 3H, J ) 7.0 Hz), 1.4 (m, 2H),
1.62 (m, 1H), 2.53 (m, 1H), 4.23 (m, 2H), 4.47 (m, 1H),
4.79 (m, 1H), 5.01 (d, 1H, J ) 9.2 Hz), 5.18 (m, 2H), 7.4
(m, 5H), 7.5 (m, 2H), 7.57 (m, 1H); 13C NMR (CDCl3, 100
MHz) δ 155.9, 154.4, 139.4, 136.2, 134.3, 128.6, 128.3,
128.2, 126.3 (q, J ) 31.7 Hz), 126.1, 124.2, 124.19, 124.12
(q, J ) 273 Hz), 120.74, 120.70, 67.2, 62.2, 53.4, 46.7, 37.7,
28.2, 14.3, 9.7; DEPT spectrum: quaternary carbons δ 155.9,
154.4, 139.4, 136.2, 134.3, 126.3, 124.12; CH carbons δ
128.6, 128.3, 128.2, 126.1, 124.22, 124.19, 120.74, 120.70,
53.4, 46.7; CH2 carbons δ 67.2, 62.2, 37.7, 28.2; CH3 carbons
δ 14.3, 9.7; IR (drifts) 3304 (s), 3067 (m), 3033 (m), 2982
(m), 2932 (m), 1723 (s), 1693 (s), 1545 (s); MS (APCI+)
m/z (rel intens) 451 (M + H+, 2), 300 (100); Anal. Calcd
for C23H25N2O4F3: C, 61.33; H, 5.60; N, 6.22. Found: C,
61.07; H, 5.69; N, 6.22.
cis-4-Amino-2-ethyl-6-trifluoromethyl-3,4-dihydro-2H-
quinoline-1-carboxylic Acid Ethyl Ester (11). A 1-L, four-
neck flask under nitrogen atmosphere was charged with 10
(40.1 g, 89 mmol, 1.0 equiv), methanol (400 mL), and
ammonium formate (14.0 g, 223 mmol, 2.5 equiv). To the
slurry was added 10% Pd/C, 50% water wet (4.0 g), and the
slurry was heated to 40 °C over 1 h. After 1.5 h, the mixture
was cooled to room temperature and filtered through Celite.
The cake was washed 2 × 100 mL of methanol. The filtrate
was concentrated to ∼75 mL, transferred to a separatory
funnel, and diluted with 400 mL of ethyl acetate. The mixture
+
1H), 7.95 (m, 2H), 8.2 (br s, NH3 , did not integrate); 13C
NMR (DMSO-d6, 100 MHz) δ 169.8, 165.5, 154.1, 140.1,
134.5, 133.5, 130.7, 129.6, 128.9, 126.7, 124.8 (q, J ) 31.7
Hz), 124.6 (q, J ) 271 Hz), 124.5, 120.9, 74.4, 62.1, 53.5,
45.9, 38.8, 28.2, 14.6, 9.5; DEPT spectrum: quaternary
carbons δ 169.8, 165.5, 154.1, 140.1, 134.5, 130.7, 124.8,
124.6; CH carbons δ 133.5, 129.6, 128.9, 126.7, 124.5, 120.9,
74.4, 53.5, 45.9; CH2 carbons δ 62.1, 38.8, 28.2; CH3 carbons
δ 14.6, 9.5; IR (drifts) 3278 (m), 2400-3100 (broad), 1703
(vs); MS (ES+) m/z (rel intens) 358 (M + H + CH3CN+,
55), 317 (M + H+, 7), 300 (100); Anal. Calcd for
C15H19N2O2F3.C9H7O4: C, 58.18; H, 5.29; N, 5.65. Found:
C, 57.99; H, 5.15; N, 5.64; Chiral HPLC: mobile phase 950:
50:2 n-hexane/2-propanol/HOAc, flow rate 1.50 mL/min,
column temp 40 °C, Chiralpak AD 4.6 mm × 250 mm,
sample concentration ∼0.5 mg/mL in ∼1:1 n-hexane/2-
propanol. Retention times: 7.5 min 99% (2R,4S) and 10.0
min 1% (2S,4R). [R]D ) -153 (c ) 1.07, CH3OH).
(-)-(2R,4S)-4-(3,5-Bis-trifluoromethyl-benzylamino)-
2-ethyl-6-trifluoromethyl-3,4-dihydro-2H-quinoline-1-car-
boxylic Acid Ethyl Ester Tosylate Salt (14). 12 (13.0 g,
26.2 mmol, 1.0 equiv) was suspended in 1,2-dichloroethane
(260 mL) in a 500-mL separatory funnel. The mixture was
washed 1 × 65 mL of 1 N NaOH, 1 × 65 mL of brine, and
470
•
Vol. 10, No. 3, 2006 / Organic Process Research & Development

dried (MgSO4). The mixture was filtered, concentrated to
∼80 mL, and transferred to a 250-mL three-neck flask. 13
(4.53 mL, 27.5 mmol, 1.05 equiv) was added, and the
mixture stirred 1 h at room temperature under nitrogen
atmosphere. Sodium triacetoxyborohydride (11.1 g, 52.4
mmol, 2.0 equiv) was added in one portion, and the white
slurry stirred 18 h. To the slurry were added 50 mL of 1,2-
dichloroethane and 50 mL of 2 N NaOH, and the aqueous
layer was extracted 2 × 50 mL of 1,2-dichloroethane. The
combined organic extracts were washed 1 × 31 mL of 1 N
HCl, 1 × 50 mL of saturated aq NaHCO3, 1 × 50 mL of
brine, and dried (Na2SO4). The mixture was filtered and
concentrated to a clear oil. The oil was dissolved in methanol
(71 mL). p-Toluenesulfonic acid monohydrate (5.23 g, 27.5
mmol, 1.05 equiv) was added. After 5 min, 284 mL of
isopropyl ether was added.21 The solution was concentrated
to ∼35 mL, transferred to a 500 mL three neck flask (mech.
stirrer), and diluted with 284 mL isopropyl ether. A thick
white slurry formed in 10 min. After stirring 3 h, the slurry
was filtered and the cake washed 2 × 70 mL of isopropyl
ether. After drying, 14 was isolated as a white powder (16.18
g, 86% overall): mp 191-192 °C; 1H NMR (DMSO-d6, 400
MHz) δ 0.78 (t, 3H, J ) 7.5 Hz), 1.21 (t, 3H, J ) 7.0 Hz),
1.5 (m, 3H), 2.24 (s, 3H), 3.08 (m, 1H), 4.17 (m, 2H), 4.41
(m, 1H), 4.50 (m, 2H), 4.79 (m, 1H), 7.04 (d, 2H, J ) 7.9
Hz), 7.42 (d, 2H, J ) 7.9 Hz), 7.7 (m, 2H), 7.81 (s, 1H),
8.21 (s, 1H), 8.35 (s, 2H), 9.58 (br s, 1H), 9.83 (br s, 1H);
13C NMR (DMSO-d6, 100 MHz) δ 154.0, 145.4, 140.2,
138.3, 135.3, 132.5, 131.6, 130.7 (q, J ) 33.2 Hz), 128.4,
127.4, 125.8, 125.3, 124.9 (q, J ) 31.7 Hz), 124.5 (q, J )
271 Hz), 123.6 (q, J ) 273 Hz), 123.4, 120.3, 62.3, 53.9,
53.7, 47.9, 33.3, 28.6, 21.1, 14.6, 9.5; DEPT spectrum:
quaternary carbons δ 154.0, 145.4, 140.2, 138.3, 135.3,
130.7, 124.9, 124.5, 123.6; CH carbons δ 132.5, 131.6, 128.4,
127.4, 125.8, 125.3, 123.4, 120.3, 53.9, 53.7; CH2 carbons
δ 62.3, 47.98, 33.3, 28.6; CH3 carbons δ 21.1, 14.6, 9.5; IR
(drifts) 2300-3100 (br), 2974 (m), 2731 (m), 2620 (m), 2455
(m), 1714 (s), 1621 (m), 1283 (vs), 1169 (vs), 1126 (vs);
MS (ES+) m/z (rel intens) 584 (M + H + CH3CN+, 100),
543 (M + H+, 80); Anal. Calcd for C24H23N2O2F9.
C7H8O3S: C, 52.11; H, 4.37; N, 3.92. Found: C, 52.15; H,
4.22; N, 3.69; [R]D ) -77.9 (c ) 1.05, CH3OH).
concentrated to 65 mL, diluted with 260 mL of ethyl acetate,
and transferred to a separatory funnel. The mixture was
washed 1 × 90 mL of 1 N HCl (CO2 evolution), 1 × 90 mL
of saturated aq NaHCO3, 1 × 90 mL of brine, and dried
(MgSO4). Filtration and concentration of filtrate afforded a
clear oil, which was costripped 3 × 33 mL of 2B ethanol.
The oil was dissolved in 33 mL of 2B ethanol and seeded
with a few milligrams of 1. After stirring 18 h at room
temperature, the slurry was filtered and dried to give 1
monoethanolate as a white crystalline powder (8.66 g,
1
74%): mp 54-58 °C; H NMR (CDCl3, 400 MHz, 55 °C)
δ 0.73 (t, 3H, J ) 7.0 Hz), 1.20 (t, EtOH), 1.27 (t, 3H, J )
7.1 Hz), 1.42 (m, 2H), 1.66 (m, 1H), 2.25 (br s, 1H), 3.67
(q, EtOH), 3.79 (s, 3H), 4.2 (m, 3H), 4.33 (m, 1H), 5.2 (br
s, 2H), 7.12 (s, 1H), 7.49 (d, 1H, J ) 8.3 Hz), 7.57 (d, 1H,
J ) 8.5 Hz), 7.73 (s, 2H), 7.78 (s, 1H); 13C NMR (CDCl3,
400 MHz) δ 157.7, 154.3, 141.7, 140.0, 133.8, 132.1 (q, J
) 33 Hz), 126.9, 124.4, 123.9 (q, J ) 273 Hz), 123.1 (q, J
) 273 Hz), 121.3, 119.1, 62.2, 58.2, 54.4, 53.71, 53.0, 46.6,
37.0, 29.0, 18.2, 14.3, 9.2, (note: the fourth quartet appears
to be buried under the δ 126.9 peak, with J ≈ 32 Hz); DEPT
spectrum: quaternary carbons δ 157.7, 154.3, 141.7, 140.0,
133.8, 132.1, 123.9, 123.1; CH carbons δ 126.9, 124.4, 121.3,
119.1, 54.4, 53.0; CH2 carbons δ 62.2, 58.2, 46.6, 37.0, 29.0;
CH3 carbons δ 53.7, 18.2, 14.3, 9.2; IR (drifts) 3489 (s),
2974 (s), 2884 (m), 1701 (vs), 1280 (vs), 1131 (vs); MS
(ES+) m/z (rel intens) 601 (M + H+, 100); Anal. Calcd for
C26H25N2O4F9.C2H6O: C, 52.01; H, 4.83; N, 4.33. Found:
C, 51.84; H, 4.54; N, 4.33; chiral HPLC: mobile phase 950:
50:2 n-hexane/2-propanol/HOAc, flow rate 1.0 mL/min, 254
nm, Chiralpak AD 4.6 mm × 250 mm, column temp 40 °C,
sample concentration ∼0.5 mg/mL in 90:10 n-hexane/2-
propanol, authentic racemate retention times 3.6 and 4.6 min.
(-)-(2R,4S)-4-[(3,5-Bis-trifluoromethyl-benzyl)methoxycar-
bonylamino]-2-ethyl-6-trifluoromethyl-3,4-dihydro-2H-quin-
oline-1-carboxylic acid ethyl ester monoethanolate shows 4.6
min, 99.1% and 3.6 min, not detected; [R]D ) -93.3 (c )
1.08, CH3OH).
Acknowledgment
We thank Jon Bordner for the X-ray crystallographic
structure determination. We also thank Keith Devries, Jerry
Murry, and Steve Ley for helpful chemistry discussions.
(-)-(2R,4S)-4-[(3,5-Bis-trifluoromethyl-benzyl)meth-
oxycarbonylamino]-2-ethyl-6-trifluoromethyl-3,4-dihydro-
2H-quinoline-1-carboxylic Acid Ethyl Ester Mono-
ethanolate (1). Na2CO3 (s) (6.75 g, 63.7 mmol, 3.5 equiv)
was added to a room-temperature solution of 14 (13.0 g,
18.2 mmol, 1.0 equiv) in dry THF (130 mL). Methyl-
chloroformate (3.51 mL, 45.5 mmol, 2.5 equiv) was added
neat, dropwise over 2 min. After 24 h, the mixture was
Supporting Information Available
Computational data and X-ray crystallographic data. This
material is available free of charge via the Internet at http://
pubs.acs.org.
Received for review January 13, 2006.
OP060014A
(21) Any use of isopropyl ether can be hazardous. Please use with caution. See
Laird, T. Org. Process Res. DeV. 2006, 8, 815.
Vol. 10, No. 3, 2006 / Organic Process Research & Development
•
471