Process development of the synthetic route to R116301

Article abstract of DOI:10.1021/op700086d

We describe in this paper the synthesis of compound 1 (R116301), which was developed to prepare pilot scale quantities (20-50 kg) of drug substance. The synthesis involves the ^sBuLi deprotonation of Boc-protected piperidone acetal 2, followed b

Full text of DOI:10.1021/op700086d

Organic Process Research & Development 2007, 11, 1079–1086
Process Development of the Synthetic Route to R116301
Michel Guillaume,* Jef Cuypers, and Jul Dingenen
Chemical Process Research, Johnson & Johnson Pharmaceutical Research and DeVelopment,
Turnhoutseweg, 30, 2340 Beerse, Belgium
Abstract:
We describe in this paper the synthesis of compound 1 (R116301),
which was developed to prepare pilot scale quantities (20–50 kg)
of drug substance. The synthesis involves the sBuLi deprotonation
of Boc-protected piperidone acetal 2, followed by benzaldehyde
addition and ring closure to cyclic carbamate 4. Piperidine acetal
5 is resolved with Brown’s acid and acylated. The ketone obtained
after piperidine acetal deprotection undergoes reductive amination
with N-benzyl piperazine, the most critical step in the synthesis.
The original synthetic route to 1 is depicted in Scheme 1.
As evident from its inefficiency (<1% overall yield), this
synthesis was not suitable for pilot scale.
After debenzylation, final coupling and salt formation, compound
1 is obtained over 10 steps with 4% overall yield.
The most efficient way to supply 1 in sufficient quantity
(up to 20 kg) and within the required time scale was to adapt
the current synthesis procedures. We discuss our improvements
in the next section.
Introduction
Tachykinin peptides are central and peripheral neurotrans-
mitters and neuromodulators. Substance P (SP), neurokinin A
(NKA) and neurokinin B (NKB) interact with, respectively, the
NK1, NK2 and NK3 receptors.1 The interaction of SP with the
NK1 receptor activates neuronal pathways regulating mood,
anxiety, emesis and pain.
Several NK1 antagonists have been reported to show good
activity in preclinical assays relevant to pathological conditions.2
This was recently supported by clinical findings demonstrating
that Aprepitant, a CNS penetrating NK1 antagonist, has efficacy
as antidepressant3 and antiemetic.4
The aim of this discovery project was to identify nonpeptidic,
orally active, centrally penetrating and selective NK1 small
molecule receptor antagonists.5
The production of kilogram batches of compound 1 was
required to carry out several toxicological, formulation and
clinical studies.
Results and Discussion
The reaction of the anion derived from compound 2 (1.3
s
equiv of BuLi in diethyl ether (Et2O)6 in the presence of 1
equiv of TMEDA at -70 °C) with benzaldehyde (1.1 equiv),7
followed by quenching at -50 °C with NH4Cl and evaporating
the solvent, led to residue 3.
Originally, 3 was cyclized to 4 (45% yield from 2) by
adding toluene and NaOMe/MeOH (1 equiv), evaporating
MeOH/toluene and adding di-isopropyl ether (DIPE) to
the oily residue in order to precipitate the product. This
laborious procedure was circumvented by replacing tolu-
i
ene by PrOH (1 L/mol) and using catalytic quantities of
tBuOK (0.05 equiv): compound 4 precipitated as pure
crystalline product in 55% overall yield from 2. We also
tried the more straightforward carbanion attack on benzyl
bromide. In this case, however, exchange reactions
together with Wurtz coupling prevented us from obtaining
pure material and acceptable yields.
* To whom correspondence should be addressed. E-mail: mguillau@
prdbe.jnj.com.
(1) Pennefather, J. N.; Lecci, A.; Candenas, M. L.; Patak, E.; Pinto, F. M.;
Maggi, C. A. Life Sci. 2004, 74, 1445.
(2) (a) Rupniak, N. M. J.; Kramer, M. S. Trends Pharmacol. Sci. 1999,
20, 485. (b) Seward, E. M.; Swain, C. J. Expert Opin. Ther. Pat. 1999,
9, 571. (c) Diemunsch, P.; Grélot, L. Drugs 2000, 60, 533. (d) Rupniak,
N. M. J. Curr. Opin. InVest. Drugs 2002, 3, 257. (e) Humphrey, J. M.
Curr. Top. Med. Chem. 2003, 3, 1423.
In the next step, the benzylic C-O hydrogenolysis was
followed by carbamate fragmentation to amine 5 in almost
quantitative yield.8 At this stage, we had to solve the following
problem: build-up of CO2 within the vessel induced progressive
(3) Kramer, M. S.; Rupniak, N. M. J. Science 1998, 281, 1640. (b)
Argyropoulos, S. V.; Nutt, D. J Expert Opin. InVest. Drugs 2000, 9,
1871. (c) Sorbera, L. A.; Castaner, J.; Bayés, M.; Silvestre, J. Drugs
Future 2002, 27, 211.
(6) The use of this highly flammable solvent, associated with the low
temperature of the reaction, led us to out-source the reaction in an ad
hoc facility, where special care was taken for prior peroxide testing,
reactor inertization, and connection to earth. Interestingly, we per-
formed more recently at lab scale a couple of experiments using 2-Me-
THF as solvent and successfully obtained comparable results as with
Et2O. Aycock, D. F. Org. Process Res. DeV. 2007, 11, 156.
(4) (a) Patel, L.; Lindley, C. Expert Opin. Pharmacother. 2003, 4, 2279.
(b) Hesketh, P. J.; Grunberg, S. M.; Gralla, R. J.; Warr, D. G.; Roila,
F.; de Wit, R.; Chawla, S. P.; Carides, A. D.; Ianus, J.; Elmer, M. E.;
Evans, J. K.; Beck, K.; Reines, S.; Horgan, K. J. J. Clin. Oncol. 2003,
21, 4112.
(5) Megens, A. A. H. P.; Ashton, D.; Vermeire, J. C. A.; Vermote,
P. C. M.; Hens, K. A.; Hillen, L. C.; Fransen, J. F.; Mahieu, M.;
Heylen, L.; Leysen, J. E.; Jurzak, M. R.; Janssens, F. J. Parmacol.
Exp. Ther. 2002, 302, 696.
(7) (a) For reviews on this topic, see: Hoppe, D.; Hense, T. Angew. Chem.,
Int. Ed. Engl. 1997, 36, 2282. (b) Beak, P.; Basu, A.; Gallagher, D. J.;
Park, Y. S.; Thayumanavan, S. Acc. Chem. Res. 1996, 29, 552.
(8) Effenberger, F.; Jäger, J. Chem. Eur. J. 1997, 3, 1370.
10.1021/op700086d CCC: $37.00
Published on Web 11/16/2007
2007 American Chemical Society
Vol. 11, No. 6, 2007 / Organic Process Research & Development
•
1079

Scheme 1. Synthesis of compound 1a
a (i) Et2O, TMEDA, sBuLi, PhCHO; (ii) NaOMe (1 equiv), toluene, rfx (45% over 2 steps); (iii) H2, Pd/C, MeOH, 50 °C (85%); (iv) 3,5-bis-trifluoromethyl benzoyl
chloride, Et3N, toluene; (v) HCl, THF/H2O (38% over 2 steps); (vi) N-benzyl-piperazine, Ti(OiPr)4, NaBH3CN, CH2Cl2 (10%); (vii) chromatography (quant); (viii) H2,
Pd/C, MeOH, 50 °C (quant); (ix) N-(R-chloroacetyl)-2,6-dimethylanilide 11, K2CO3, MIK, 100°C (82%); (x) L-malic acid (1.1 equiv), EtOH (80%).
Scheme 2
slowing of the rate of hydrogenation and several reflushing
cycles with H2 were necessary in order to bring the reaction to
completion. Therefore, we added 1.05 equiv of NaOMe to take
up the CO2. That resulted in a much faster reaction (100%
conversion after 2 h, 91% yield) and led to protected piperidone
(()-5 as an oily residue after work-up.
Originally, enantiomers were separated later in the synthesis
by means of preparative chiral chromatography. However, we
decided to perform the resolution at this stage: the use of
Brown’s acid9 led to (R)-(–)-5 in 30% yield (max 50%!) and
90% ee after the release of the base. This ee, albeit modest, was
sufficient to obtain eventually final compound 1 with 98% ee.
After amine release, residue (R)-(–)-5 was reacted with 3,5-
bis(trifluoromethyl)-benzoyl chloride in toluene.
The reductive amination step to 8 proved to be by far the
most difficult in the synthesis (Scheme 3).
Originally, the ketone 7 was reacted with N-benzylpiperazine
in the presence of Ti(OiPr)4 (1 equiv) and then reduced with
NaBH3CN according to a classical procedure;10 the desired
trans diastereomer was obtained with only 25:75 trans/cis-
selectivity.
We improved this step by performing a catalytic hydrogena-
tion instead of the borohydride reduction.11 Through Pt-
catalyzed hydrogenation, we obtained a 70:30 trans/cis-
selectivity.
We performed more than 150 experiments to further improve
the trans-selectivity but did not obtain better results despite the
many variations we tried (Table 1).
In the original synthesis, the ketone underwent deprotection
in THF/H2O. We performed this step with aqueous 5 N HCl
(5 equiv) in EtOH at 50 °C. Compound (+)-7 was obtained in
85% yield (Scheme 2).
The experiments shown below are representative of the
parameters we varied in order to improve selectivity:
(i) metal catalysts: Pt, Ni, Pd, Rh and Ir
(ii) alkaloid-modified Pt/C catalyst12
(iii) catalyst support: C, Al2O3, acid-treated C, sulfur-treated
C, CaCO3
(9) (a) Brown, E.; Viot, F.; Le Floc’h, Y. Tetrahedron Lett. 1985, 26,
4451. (b) Brown, E.; Chevalier, C.; Huet, F.; Le Grumelec, C.; Lézé,
A.; Touet, J. Tetrahedron: Asymmetry 1994, 5, 1191. For lab scale
purposes, Brown’s acid was purchased at Acros Organics. Large scale
quantities of this reagent were made according to an internal,
unpublished experimental procedure. Other chiral acids were screened
as well but proved unsuccessful for this resolution: L-(+)-tartaric acid,
L-(–)-dibenzoyl-tartaric acid monohydrate, (–)-di-p-toluil-L-tartaric
acid, S-(+)-mandelic acid, L-(–)-malic acid, L-(+)-lactic acid,
L-pyroglutamic acid, and D-(+)-10-camphorsulfonic acid.
(10) (a) Mattson, R. J.; Pham, K. M.; Leuck, D. J.; Cowen, K. A. J. Org.
Chem. 1990, 55, 2552. (b) Kumpaty, H. J.; Bhattacharyya, S.; Rehr,
E. W.; Gonzalez, A. M. Synthesis 2003, 2206 and references therein.
(11) Previous trials with NaBH4, NaBH4/ZnCl2, and NaBH4/CaCl2 always
showed limited selectivity for the undesired cis-isomer.
(12) For a review on this topic, see: Kacprzac, K.; Gawroñski, J. Synthesis
2001, 961.
1080
•
Vol. 11, No. 6, 2007 / Organic Process Research & Development

Scheme 3
(iv) solvent: EtOH, MeOH, THF, toluene and EtOAc
(v) amination conditions: Ti(OiPr)4, Al(OiPr)3, Ti(OnBu)4,
Ti(OEt)4, MgSO4, Mg(OEt)2
The precipitate, obtained in 80% yield, was in turn recrystallized
from EtOH, leading to pure compound 1 as white crystals (80%
yield).
(vi) ketone substrate: N-bis(3,5-trifluoromethyl benzoyl)-
(compound 7), N-carbethoxy-, N-benzyloxycarbonyl-, N-BOC-,
N-acetyl-, N-benzoyl-, N-trifluoroacetyl-, N-benzyl-, N-1-naph-
toyl-, N-2-naphtoyl- and N-tosyl piperidones
(vii) amine substrates: N-benzyl-, N-Boc- and N-carbethoxy
piperazine
The results clearly indicate a strong influence of the
piperidone substitution and of the reducing system (Pt > Pd ≈
NaBH3CN) on the trans-selectivity.
Using the experimental procedure described here, we were
able to produce about 20–50 kg of drug substance.
In order to have a more efficient synthesis, the more
convergent reductive amination depicted in Scheme 4 was
tried.
Despite its apparent simplicity, a convenient process to
manufacture piperazine 12 was demanding to develop. We
described this topic in a preceding paper.13
Ketone 7 underwent reductive amination with secondary
amine 12 under the same conditions as described above
[Al(OiPr)3, neat, then iPrOH, H2, Pt/C]. However, a 50:50 cis/
trans mixture was obtained. Further optimization would be
necessary to improve selectivity if this convergent approach
were to be pursued further.
In conclusion, process development on R116301 resulted
in the manufacture of multikilogram quantities of drug sub-
stance. The overall yield of about 4% represents a 5- to 10-
fold increase compared with the former synthesis.
The most difficult step was the reductive amination,
to which we brought two modifications: the catalytic
hydrogenation conditions (Pt/C) for the reduction, which
led to an inversion of the stereochemical selectivity (from
trans/cis 25:75 to 70:30), and the use of Al(OiPr)3 instead
of Ti(OiPr)4 to avoid the formation of insoluble material.
A practical problem arose at this stage when Ti alkoxide
were used for the amination step: after the hydrogenation, work-
up with water and toluene generated a milky precipitate that
was almost impossible to filter. Use of acid to dissolve the
precipitate was incompatible with the amine, which would also
transfer into the aqueous phase. While magnesium sulfate and
magnesium ethoxide did not lead to completion of the reaction,
aluminum isopropoxide (Al(OiPr)3) gave similar conversion and
similar trans-selectivity as Ti(OiPr)4. A simple alkaline work-
up, followed by extraction with toluene, afforded the 70:30
trans/cis mixture, which after chromatography gave compound
trans-9.
For the separation of the cis- and trans-isomers, a
standard analytical screen on normal silica gel, polar
modified silica gel (diol-, aminopropyl-, cyanopropyl-) and
some reversed phase materials was performed. However,
these experiments did not result in an acceptable separation
for large-scale preparative chromatography. Therefore,
some experimental work was done on a number of chiral
stationary phases. On Chiralcel OD [Daicel (3,5-dimethyl
phenyl carbamate of cellulose)], the isomers were nicely
separated using EtOH denatured with 2% of MeOH as
the mobile phase.
Compound 9 was debenzylated to 10 (Pd/C, 50 °C, MeOH);
once the reaction was complete, the catalyst was filtered off,
the solvent was evaporated, and methyl isopropyl ketone (MIK)
was added for the next step. To this solution were added
compound 11 (1 equiv) and Na2CO3 (1.2 equiv). After reaction
(2 h at 100 °C) and work-up (water and toluene), the organic
phase was evaporated and iPrOH added for the following step.
This solution was filtered through Dicalite (in order to avoid
insoluble particles in the end product). (S)-Malic acid was added,
and the reaction mixture was heated to reflux and then cooled.
Experimental Section
General Procedures. All materials were purchased from
commercial suppliers and used without further purification.
All reactions were conducted under an atmosphere of
nitrogen unless noted otherwise. In the laboratory, only
glass vessels were used; in the pilot plant, either steel or
glass-lined vessels were used. For each reaction, a sample
of the reaction mixture was collected and analyzed by
means of GC or HPLC.
(()-Tetrahydro-1′-phenylspiro(1,3-dioxolan-2,7′(8′H)-3H-
oxazolo[3,4-a]pyridin)-3′-one 4. A mixture of 2 (24.3 g, 0.1
mol) and TMEDA (15.1 mL, 1 equiv) in diethyl ether (150
mL, 1.5 L/mol) was cooled down to -70 °C with acetone/dry
ice. sBuLi 1.3 M (100 mL, 1.3 equiv) was added dropwise so
(13) Guillaume, M.; Cuypers, J.; Vervest, I.; De Smaele, D.; Leurs, S. Org.
Process Res. DeV. 2003, 7, 939.
Vol. 11, No. 6, 2007 / Organic Process Research & Development
•
1081

Table 1. Representative experiments for the reductive amination stepa
1082
•
Vol. 11, No. 6, 2007 / Organic Process Research & Development

Table 1. Continued
a The yield (*) refers to the desired trans-isomer. In most cases, limited amounts (0–3%) of reduced ketone and starting material are found in the reaction mixture as well.
Scheme 4
cooled down, HCl 5 N (40 mL, 2 equiv) was added (CO2
evolution!), the catalyst was filtered off, and the product was
extracted with toluene (100 mL, 1 L/mol). After drying the
organic layer over Na2SO4, the solvent was distilled off. (()-5
was obtained as yellow oil. It was used directly in the next step.
Yield: 21.1 g (91%).
2. Pilot Plant Procedure. Note: The above-described labora-
tory procedure had not been yet implemented in the pilot plant
when the project was discontinued. We describe therefore the
process performed without NaOMe.
that the temperature remained under -60 °C. After the addition,
the mixture was stirred at -70 °C. Benzaldehyde (11.7 mL,
1.1 equiv) was added at such a rate that the temperature
remained under -60 °C. After stirring 2 h at -70 °C, NH4Cl
1 N (100 mL) was added and the temperature was allowed to
rise to room temperature. The organic layer was evaporated to
afford 3 as a yellow oil. Intermediate analysis (GC area %)
indicated 92% conversion.
iPrOH (100 mL, 1 L/mol) and tBuOK (600 mg, 5 mol%)
were added, and the reaction mixture was heated to reflux,
stirred at that temperature for 4 h, and then cooled down
to room temperature. After 1 h, a precipitate appeared
(seeding is sometimes necessary), and the mixture was
further cooled down to 0 °C. After stirring 30 min at that
temperature, the precipitate was filtered off, washed with
iPrOH (20 mL, 0.02 L/mol), and dried over 16 h at 40 °C
under vacuum. Yield: 15.2 g (55%). Mp: 129 °C. 1H NMR
(CDCl3, 400 MHz, major diastereomer): δ 1.59–1.82 (m,
3H), 1.98 (m, 1H), 3.10 (m, 1H), 3.76 (ddd, J ) 11.9,
6.6, 4.0 Hz, 1H), 3.80–4.04 (m, 5H), 5.03 (d, J ) 6.6 Hz,
1H), 7.25–7.44 (m, 5H). Anal. Calcd for C15H17NO4:
C, 65.44; H, 6.22; N, 5.09. Found: C, 65.56; H, 6.30; N,
4.89.
To an inertized 2000 L hydrogenation reactor was
added 4 (200 kg, 726.5 mol), followed by MeOH (1300
L, 1.8 L/mol) and Pd/C 5% wet (15.40 kg, 21.2 g/mol).
After starting the stirrer, the reaction mixture was heated
up to 50 °C and hydrogenated at that temperature. When
an uptake drop was observed, the vessel was degassed
(CO2 was removed) and reflushed with H2. After a couple
of degas/reflush operations, a process control (GC analy-
sis) indicated that the SM had completely disappeared.
The reaction mixture was flushed with nitrogen, and then
thiophene (73 g, ca. 1 g/200 g catalyst) was added. The
catalyst was filtered over Dicalite, rinsed with MeOH (145
L, 0.2 L/mol) and water (145 L, 0.2 L/mol); 1100 L
solvent was evaporated and the remaining solution was
collected and weighed. Yield: 493.2 kg solution. Base
titration: 36.6% w/w
174 kg 4 (quantitative yield).
7-(Phenylmethyl)-1,4-dioxa-8-azaspiro [4.5] decane (–)-
5·Brown’s Acid. 1. Laboratory Procedure. Compound 4 (233
g, 1 mol) was dissolved in MeOH (1900 mL, 1.8 L/mol).
Brown’s acid (209.0 g, 1 equiv) was added, and the mixture
was heated to reflux. After 15 min at reflux, the reaction mixture
was cooled down to 55 °C, seeded and stirred at that tempera-
ture for 1 h, and then further cooled down to 20 °C over 4 h;
the precipitate was filtered off, washed with MeOH (120 mL,
0.12 L/mol), and dried at 40 °C for 4 h under vacuum. Yield:
133 g (30%). Mp: 174 °C. HPLC: 100.6% (% w/w abs). ee )
90% (HPLC).
(()-7-(Phenylmethyl)-1,4-dioxa-8-azaspiro[4.5]decane (()-
5. 1. Laboratory Procedure. A mixture of 4 (27.5 g, 0.1 mol)
was dissolved in MeOH (100 mL, 1 L/mol). A 30% NaOMe
solution in MeOH (20 mL, 1.05 equiv) was added, followed
by Pd/C (5% wet, 2.5 g, 25 g/mol). The reaction mixture was
heated up to 50 °C and hydrogenated over 2 h. After it was
Vol. 11, No. 6, 2007 / Organic Process Research & Development
•
1083

As mentioned above, the enantiomeric purity is sufficient
to eventually obtain 1 of sufficient enantiomeric purity. How-
ever, for analytical purposes, we recrystallized 5 g of 5 from
MeOH (10 mL/g). Yield: 3 g (60% from raw 5). Mp: 181 °C.
HPLC: 100.5% (% w/w abs). ee >99% (chiral HPLC). Anal.
Calcd for C24H30N2O6: C, 65.14; H, 6.83; N, 6.33. Found: C,
65.08; H, 6.75; N, 6.26.
A sample of this salt (2.3 g) was in turn neutralized with
NaOH 1 N and extracted twice with toluene. After drying the
extract and evaporating the solvent, (–)-5 was obtained as a
transparent oil (1.1g, 92%) with the following characteristics:
HPLC:14 99.3% w/w abs. ee >99% (derivatization + HPLC15).
[R]D20 ) -16.6°. 1H NMR (CDCl3, 360 MHz): δ 1.45 (dd, J
) 11.4, 12.9 Hz, 1H), 1.57–1.71 (m, 3H), 1.77 (bdt, J ) 12.9
Hz, 1H), 2.59 (dd, J ) 13.3, 8.5 Hz, 1H), 2.66–2.79 (m, 2H),
2.92–3.03 (m, 2H), 3.89–3.98 (m,4H), 7.17–7.25 (m, 3H),
7.26–7.33 (m, 2H). Anal. Calcd for C14H19NO2: C, 72.07; H,
8.21; N, 6.00. Found: C, 73.33; H, 8.91; N, 6.01.
2. Pilot Plant Procedure. In a 1000 L reactor, 4 (24.2 kg,
104 mol), MeOH (187 L, 1.8 L/mol) and Brown’s acid (21.8
kg, 1 equiv) were mixed together. After refluxing the mixture
for 20 min, cooling down to 55 °C, seeding at that temperature,
cooling down further to 20 °C over 3 h, and stirring at that
temperature for 1 h, the white precipitate was centrifuged,
washed with MeOH (12.5 L, 0.12L/mol) and dried at 40 °C
under vacuum for 18 h to afford 15.9 kg of (–)-5·Brown’s acid
(35% yield), which was pure enough to use further in the next
step. ee ) 86.2%.
(+)-1-[3,5-Bis(trifluoromethyl)benzoyl]-2-(phenylmethyl)-
4-piperidinone (+)-7. 1. Laboratory Procedure. (–)-5·Brown’s
acid (140.0 g, 0.32 mol) was suspended in water (640 mL, 2
L/mol). NaOH 50% (18.0 mL, 1.1 equiv) and toluene (640 mL,
2 L/mol) were added, and the mixture was stirred vigorously
at room temperature. The water layer was extracted again with
toluene (64 mL, 0.2 L/mol). The combined organic fractions
were dried over Na2SO4 and put in a 3-necked flask, which
was cooled to 15 °C with water. The addition of triethylamine
(44.8 mL, 1.5 equiv) was followed by the dropwise addition
of 3,5-bis(trifluoromethyl)benzoyl chloride (88.3 g, 1 equiv) so
that the temperature remained under 30 °C. After stirring 30
min at 25 °C, water (320 mL, 1 L/mol) was added, and the
toluene extract was evaporated at 50 °C under vacuum. EtOH
(160 mL, 0.5 L/mol) was added, followed by water (160 mL,
0.5 L/mol) and HCl 5 N (160 mL, 2.5 equiv). The reaction
mixture was heated to 60 °C for 16 h and then cooled to 25
°C. Toluene (320 mL, 1 L/mol) was added, the water layer
was discarded, the organic extract was dried over Na2SO4, and
the solvent was distilled off at 50 °C under vacuum. Yield:
130.0 g (85% yield). An analytical sample of (–)-5 (1 g, see
above) was reacted as described above to afford after flash
chromatography 1 g (+)-7 (54%). HPLC:16 98.5%. [R]D
)
20
+7.3° (0.01% in MeOH). Anal. Calcd for C21H17F6NO2: C,
58.75; H, 3.99; N, 3.26. Found: C, 56.43; H, 4.02; N, 3.25.. 1H
NMR (DMSO-d6, 400 MHz) δ: 2.24 (bd, J ) 13.7 Hz, 1H),
2.40 (bd, J ) 13.3 Hz, 1H), 2.67 (m, 1H), 2.82 (m, 1H), 2.96
(m, 1H), 3.11 (bdd, J ) 14.2, 6.2 Hz, 1H), 3.57 (m, 1H), 4.10
(m, 1H), 4.75 (m, 1H), 6.99 (bd, J ) 6.2 Hz, 2H), 7.16–7.39
(m, 3H), 7.48 (bs, 2H), 8.13 (bs, 1H).
2. Pilot Plant Procedure. The following procedure describes
the process starting from racemic 5 (not as a salt). Compound
(()-5 (50.7 kg) was brought in the reactor together with toluene
(434 L) and triethylamine (22.2 kg), which formed a clear
solution. The acid chloride (63.4 kg) was added dropwise at a
temperature <50 °C. The reaction mixture was cooled to 25
°C and stirred at that temperature over 2 h. Water (87 L) was
added, and the mixture was stirred for 30 min. The water layer
was discarded, and the organic layer was evaporated at 80 °C
under vacuum. Water (107 L), EtOH (107 L), and HCl (12 N,
107 L) were added to the residue. The mixture was heated to
60 °C, stirred for 8 h at that temperature, and then cooled to
15–20 °C. The precipitate (()-7 was centrifuged, washed with
water, and dried (35 °C, vac, 52 h). Yield: 87%.
(+)-(2R-trans)-1-[3,5-Bis(trifluoromethyl)benzoyl]-2-(phen-
ylmethyl)-4-[4-(phenylmethyl)-1-piperazinyl]piperidine9.1. Labo-
ratory Procedure. Method A (with Ti(OiPr)4). Compound (+)-7
(171.7 g, 0.4 mol) was stirred at room temperature with
N-benzyl piperazine (206.1 g, 0.48 mol, 1.2 equiv) and
Ti(OiPr)4 (125.1 g, 0.44 mol, 1.1 equiv) at 40 °C for 3 h.
iPrOH (1.2 L, 3 L/mol) and Pt/C (5% dry, 16 g, 40 g/mol)
was added, and the reaction mixture was heated to 50 °C
and hydrogenated over 16 h at that temperature. After
purging the vessel with N2, the mixture was cooled to 20–25
°C, and the catalyst was filtered off and washed with toluene
(2 × 320 mL, 2 × 0.8L/mol). Toluene (800 mL, 2 L/mol)
was added to the filtrate, followed by Dicalite (40 g, 100
g/mol) and NaOH (800 mL, 2 equiv). After stirring 30 min
at room temperature, the milky precipitate was filtered over
Dicalite (difficult filtration!). The layers of the filtrate were
separated, and the organic one was filtered over Na2SO4 and
evaporated. The oily residue (cis/trans mixture) was chro-
matographed to afford pure trans-isomer as a pale yellow
oil, which was used further in the next step.Method B (with
Al(OiPr)3). N-Benzyl piperazine (17.6 g, 0.1 mol), Al(OiPr)3
i
(20.4 g, 0.1 mol), Pt/C 5% dry (2 g, 20 g/mol), and PrOH
(200 mL, 2 L/mol) were added to (+)-6 (42.9 g, 0.1 mol) in
toluene (200 mL, 2 L/mol). The mixture was heated to 60
°C, stirred at that temperature for 3 h, and then hydrogenated
over 16 h. After the reaction mixture was cooled to room
(14) HPLC Method. Hypersil BDS 100 × 4.0 mm i.d., 3 µm spherical
material; eluent A 0.2% (NH4)2CO3; eluent B CH3CN; temp 30 °C;
injection volume 6 mL; flow 1 mL/min. Gradient: 0 min 95% A, 5%
B; 7 min 75% A, 25% B; 14 min 25% A, 75% B; 16 min 5% A, 95%
B; 19 min 5% A, 95% B; 19.5 min 95% A, 5% B; 25 min 95% A;
5% B. UV detection at 210 nm. Retention time: 9.5 min.
(15) HPLC Method. the compound is derivatized with GITC (excess) in
DMF, followed by achiral HPLC. Hypersil BDS 100 × 4.0 mm i.d.,
3 µm spherical material; eluent A 0.5% (NH4)OAc; eluent B MeOH;
temp 25 °C; injection volume 6 mL; flow 1 mL/min. Gradient: 0 min
45% A, 55% B; 25 min 45% A, 55% B. UV detection at 225 nm.
Retention time: 20.4 min.
(16) HPLC Method. Hypersil BDS 100 × 4.0 mm i.d., 3 µm spherical
material; eluent A 1% H3PO4; eluent B CH3CN; temp 25 °C; injection
volume 6 mL; flow 1.2 mL/min. Gradient: 0 min 95% A, 5% B; 15
min 0% A, 100% B; 18 min 0% A, 100% B; 19 min 95% A, 5% B;
19 min 5% A, 95% B; 25 min 95% A; 5% B. UV detection at 210
nm. Retention time: 10.9 min.
1084
•
Vol. 11, No. 6, 2007 / Organic Process Research & Development

temperature and flushed with N2, NaOH 5 N (100 mL, 5
equiv) was added, and the catalyst was filtered off over
Dicalite and rinsed with toluene (20 mL, 0.2 L/mol). The
layers were separated, the organic layer waswas dried over
Na2SO4, and the solvent was distilled off at 50 °C under
vacuum. Yield: 70.4g (corresponds to 46% active yield of
trans-isomer). The mixture was chromatographed to afford
(+)-9 as a pale yellow oil that was directly dissolved in
MeOH (28 L) and Pd/C 5% wet (580 g) were introduced.
Stirring was started, and the mixture was heated to 50 °C. The
debenzylation through hydrogenolysis was performed at that
temperature. After 16 h, a process control indicated complete
conversion. The reaction mixture was cooled to 25 °C,
thiophene (3 g) was added, and the filtration was performed.
The cake was washed with MeOH (12.5 L), and the solution
was evaporated to dryness. For the next step, 4-methyl-2-
pentanone (38.8 L) was added and stirring was continued for
30 min The solution of (+)-10 so obtained was used further
after quantitative HPLC analysis. Yield: 92%.
1
MeOH, titrated, and used further in the next step. H NMR
(CDCl3, 400 MHz) δ: 1.49–1.70 (m, 2H), 1.97 (m, 1H), 2.14
(m, 1H), 2.38–2.72 (m, 8H), 2.75–2.90 (m, 2H), 3.09–3.21
(m, 2H), 3.46–3.58 (m, 2H), 3.91 (m, 1H), 4.76 (m, 1H),
6.87 (d, J ) 6.2 Hz, 2H), 7.11 (s, 2H), 7.18–7.37 (m, 8H),
7.78 (s, 1H).
(+)-(B-trans)-4-[1-[3,5-Bis(trifluoromethyl)benzoyl]-2-
(phenylmethyl)-4-piperidinyl]-N-(2,6-dimethylphenyl)-1-
piperazineacetamide·(L)-Malic Acid 1. 1. Laboratory Pro-
cedure. Compound (+)-10 (0.22 mol) was dissolved in MIK
(440 mL, 2 L/mol). N-(R-Chloroacetyl)-2,6-dimethylanilide 11
(54.4 g, 0.22 mol) and Na2CO3 (26.5 g, 1.2 equiv) were added,
and the reaction mixture was heated to 100 °C. After stirring
2 h at that temperature, the mixture was cooled to room
temperature, water (880 mL, 4 L/mol) was added, and the
2. Pilot Plant Procedure (with Al(OiPr)3). To a 300 L glass-
lined inertized vessel were added (+)-6 (55 kg, 128 mol) in
iPrOH (128 L, 1 L/mol), N-benzyl piperazine (22.5 kg, 1 equiv),
Al(OiPr)3 (26.1 kg, 1 equiv), and Pt/C 5% dry (2.56 kg, 20
g/mol). The mixture was heated to 60 °C, stirred at that
temperature for 3 h, and then hydrogenated over 16 h. The
reaction mixture was cooled to room temperature, flushed with
N2, and then transferred to another vessel. Toluene (512 L, 4
L/mol) and NaOH 5 N (256 L, 10 equiv) were added over 15
min, followed by Dicalite (3.8 kg, 30 g/mol). After filtration,
the layers were separated, and the organic layer was evaporated
under vacuum. EtOH denatured with 2% MeOH (141 L, 1.1
L/mol) was added, and the mixture was further chromato-
graphed as such. Yield: 43 kg (corresponds to 57% active yield
of trans-isomer).
Chromatographic Separation at Large Scale. (Not the
same batch as the one described in the preceding paragraph!)
Stationary phase: Chiralcel OD [Daicel (3,5-dimethyl phenyl
carbamate of cellulose)]. Mobile phase: EtOH denatured with
2% of MeOH. A solution of 100 g of product/L eluent was
prepared. A 20 cm i.d. dynamic axial compression column was
filled with 6 kg of the stationary phase and the total amount of
product (148 kg) was chromatographed by means of repetitive
injections of 1 L of sample solutions corresponding to an
injection amount of 100 g (a loading capacity of 16.7 mg/g of
packing material). After evaporation of the solvent, 86.6 kg of
the desired isomer was obtained, which corresponds to 59%
active yield (based on (+)-9.
(+)-(2R-trans)-1-[3,5-Bis(trifluoromethyl)benzoyl]-2-(phen-
ylmethyl)-4-(1-piperazinyl)piperidine 10. 1. Laboratory Pro-
cedure. Compound (+)-9 (137.5 g, 0.23 mol) was dissolved in
MeOH (700 mL, 3 L/mol) and Pd/C 5% wet (7.0 g, 30 g/mol)
was added. The reaction mixture was heated to 50 °C and
hydrogenated at that temperature for 5 h. The catalyst was
filtered off over Dicalite, rinsed with MeOH (35 mL, 0.15
L/mol). The solvent was distilled off at 50 °C under vacuum
to afford (+)-10 as a pale yellow oil. Yield: 114.5 g (95% yield).
1H NMR (CDCl3, 360 MHz) δ: 1.49–1.75 (m, 2H), 1.90 (bs,
1H, NH), 1.98 (m, 1H), 2.16 (m, 1H), 2.49–2.68 (m, 4H),
2.77–2.97 (m, 6H), 3.18 (m, 2H), 3.93 (m, 1H), 4.77 (m, 1H),
6.88 (d, J ) 6.3 Hz, 2H), 7.08 (bs, 2H), 7.20–7.37 (m, 3H),
7.78 (bs, 1H).
i
organic layer was evaporated at 50 °C under vacuum. PrOH
(440 mL, 2 L/mol) was added, and the solution was filtered
over Dicalite in order to remove the insoluble materials. Malic
acid (29.5 g, 1 equiv) was added, and the mixture was heated
to reflux, while a precipitate formed. After 30 min reflux, the
mixture was cooled to 20 °C and stirred at that temperature for
3 h. The precipitate was filtered off, washed with 44 mL (0.2
L/mol) iPrOH, and dried for 16 h at 50 °C under vacuum. Yield:
122.2 g (70% over 2 steps), off-white precipitate. The salt was
recrystallized from EtOH (8 mL solvent/g salt, 80% yield) to
obtain a purity which was always consistent with quality
requirements. HPLC:17 99.6% w/w. ee: 98% (CE18). Anal. Calcd
for C35H38F6N4O2.C4H6O5: C, 58.94; H, 5.58; N, 7.05. Found:
1
C, 58.76; H, 5.44; N, 7.01. H NMR (CDCl3, 400 MHz, free
base): δ 1.58 (m, 1H), 1.68 (m, 1H), 1.97 (m, 1H), 2.14 (m,
1H), 2.23 (s, 6H), 2.55–2.80 (m, 9H), 2.82–2.94 (m, 1H),
3.10–3.22 (m, 4H), 3.94 (m, 1H), 4.77 (m, 1H), 6.87 (d, J )
6.6 Hz, 2H), 7.03–7.10 (m, 3H), 7.12 (s, 2H), 7.20–7.36 (m,
3H), 7.79 (s, 1H), 8.62 (s, 1H, CONH).
2. Pilot Plant Procedure. In an inerted reactor, (+)-10 (146
kg solution in MIK) was introduced together with N-(R-
chloroacetyl)-2,6-dimethylanilide 11 (8.6 kg) and Na2CO3 (5.37
kg). The reaction mixture was heated to 100 °C and stirred at
that temperature for 2 h. Process control (HPLC) indicated a
complete conversion, the reaction mixture was then cooled to
25 °C and water (174 L) was added. After stirring for 30 min
the water layer was discarded. The organic layer was evaporated
at 70 °C under vacuum to dryness. iPrOH (84 L) and Dicalite
(17) HPLC Method. Hypersil BDS 100 × 4.0 mm i.d., 3 µm spherical
material; eluent A 10mM TBAHS; eluent B CH3CN; eluent C MeOH;
temp 25 °C; injection volume 6 mL; flow 1 mL/min. Gradient: 0 min
90% A, 5% B, 5% C; 15 min 0% A, 50% B, 50% C; 18 min 0% A,
50% B, 50% C; 19 min 90% A, 5% B, 5% C; 25 min 90% A, 5% B,
5% C. UV detection at 215 nm. Retention time: 10.1 min.
(18) CE Method. TSP capillary electrophoresis system; capillary uncoated
fused silica i.d. 75 µm, 25 cm total length, temp 20 °C. Background
electrolyte: 25 mM hydroxypropyl-γ-CD in 100 mM H3PO4/trietha-
nolamine, buffer pH ) 2.5; separate +100 mA; UV detection at 200
nm. Retention time: 15.8 min.
2. Pilot Plant Procedure. The hydrogenation vessel was
inerted with nitrogen, and then (+)-9 (40.2 kg in MeOH),
Vol. 11, No. 6, 2007 / Organic Process Research & Development
•
1085

(435 g) were added. The mixture was filtered and transferred
into another reactor. A solution of L-malic acid (6.52 kg) in
iPrOH (78 L) was added to the solution of the free amine. The
mixture was heated to reflux (ca. 80 °C). During this process,
crystallization occurred. The solution was refluxed for 30 min
The reaction mixture was cooled to 25 °C over 3 h and stirred
overnight at 22 °C. The precipitate was centrifuged, washed
with 10 L iPrOH and dried (50 °C, vac, 16 h). Yield: 86%.
Received for review April 17, 2007.
OP700086D
1086
•
Vol. 11, No. 6, 2007 / Organic Process Research & Development