Process research and development for an efficient synthesis of the HIV protease inhibitor BMS-232632

Article abstract of DOI:10.1021/op025504r

Development of an efficient and scalable process for the human immunodeficiency virus (HIV) protease inhibitor BMS-232632 1-[4-(pyridin-2-yl)phenyl]-5(S)-2,5-bis{[N-(methoxycarbonyl)L-tert- leucinyl]-amino}-4(S)-hydroxy-6-phenyl-2-azahexane, is described. The key step in the synthesis of the intermediate N-1-(tert-butyloxycarbonyl)-N-2-[4-(pyridin-2-yl)benzylidene]hydrazone (11) was the Pd-mediated coupling of boronic acid 9 with 2-bromopyridine. An efficient procedure was developed for the chemoselective reduction of hydrazone 11 to hydrazine carbamate 4. The key intermediate N-(tert-butyloxycarbonyl)-2(S)-amino-1-phenyl-3(R)-3,4-epoxy-butane (6) was prepared stereoselectively from chiral diol 10. The subsequent union of 4 and 6 followed by coupling with N-methoxycarbonyl-L-tert-leucine provided the free base BMS-232632 in high yield. Evaluation of a variety of salts and identification of bisulfate salt 19 with enhanced bioavallability are also described.

Full text of DOI:10.1021/op025504r

Organic Process Research & Development 2002, 6, 323−328
Process Research and Development for an Efficient Synthesis of the HIV
Protease Inhibitor BMS-232632
Zhongmin Xu,*,† Janak Singh,† Mark D. Schwinden,† Bin Zheng,‡ Thomas P. Kissick,† Bharat Patel,†
Michael J. Humora,‡ Fernando Quiroz,‡ Lin Dong,‡ Dau-Ming Hsieh,‡ James E. Heikes,‡ Madhusudhan Pudipeddi,§
Mark D. Lindrud,‡ Sushil K. Srivastava,‡ David R. Kronenthal,‡ and Richard H. Mueller‡
The Bristol-Myers Squibb Pharmaceutical Research Institute,
Process Research and DeVelopment, P.O. Box 4000, Princeton, New Jersey 08543, U.S.A.,
Process Research and DeVelopment, P.O. Box 191, New Brunswick, New Jersey 08903, U.S.A. and
Biopharmaceutics Research and DeVelopment, P.O. Box 191, New Brunswick, New Jersey 08903, U.S.A.
Abstract:
studies.2 However, this route was suboptimal for the produc-
tion of the larger quantities required for further toxicology
studies, preclinical studies, and clinical trials. Significant
modifications in the synthesis of key intermediates as well
as the replacement of reagents in the subsequent steps were
required. Furthermore, the crystalline free base form of BMS-
232632 as a suspension in water or oil had poor oral
bioavailability in animals, presumably due to its extremely
low solubility in these vehicles. For development of phar-
maceutical formulations, particularly oral dosage forms, a
salt form of BMS-232632 with enhanced bioavailability had
to be identified. We herein report on a more efficient
synthesis of this compound and identification of a salt with
appropriate properties.
Development of an efficient and scalable process for the human
immunodeficiency virus (HIV) protease inhibitor BMS-232632
1-[4-(pyridin-2-yl)phenyl]-5(S)-2,5-bis{[N-(methoxycarbonyl)-
L
-tert-leucinyl]-amino}-4(S)-hydroxy-6-phenyl-2-azahexane, is
described. The key step in the synthesis of the intermediate N-1-
(tert-butyloxycarbonyl)-N-2-[4-(pyridin-2-yl)benzylidene]hydra-
zone (11) was the Pd-mediated coupling of boronic acid 9 with
2-bromopyridine. An efficient procedure was developed for the
chemoselective reduction of hydrazone 11 to hydrazine car-
bamate 4. The key intermediate N-(tert-butyloxycarbonyl)-2(S)-
amino-1-phenyl-3(R)-3,4-epoxy-butane (6) was prepared stere-
oselectively from chiral diol 10. The subsequent union of 4 and
6 followed by coupling with N-methoxycarbonyl-L-tert-leucine
Several limitations of the original process presented a
formidable challenge for the preparation of material in large
quantity. The pyridinyl benzaldehyde 3 was made in three
steps involving nickel-catalyzed coupling of Grignard reagent
2 with 2-bromopyridine. Although the overall yield from 1
was high (84%), the use of a hazardous reagent diisobutyl-
aluminum hydride and protection/deprotection of the alde-
hyde moiety were required.2 Epoxide 6 was prepared from
L-Boc-phenylalaninal (5) in two steps including a Wittig
reaction with methyltriphenylphosphonium bromide. Not
surprisingly, even at -78 °C, racemization of 5 occurred
during the Wittig reaction, which in turn resulted in epoxide
with only 80% ee. Furthermore, purification of both the
Wittig reaction product and epoxide 6 proved difficult, and
silica gel chromatographies were required. In the final step,
coupling with N-methoxycarbonyl-L-tert-leucine, the expen-
provided the free base BMS-232632 in high yield. Evaluation
of a variety of salts and identification of bisulfate salt 19 with
enhanced bioavailability are also described.
Introduction
BMS-232632 is an acyclic aza-peptidomimetic that is a
potent human immunodeficiency virus (HIV) protease in-
hibitor.1 Its bisulfate salt has better bioavailability than the
free base, with a half-life suitable for once-daily dosing. We
have developed a practical synthesis of this compound.
(1) (a) Fassler, A.; Bold, G.; Capraro, H.; Cozens, R.; Mestan, J.; Poncioni,
B.; Rosel, J.; Tintelnot-Blomley, M.; Lang, M. J. Med. Chem. 1996, 39,
3203-3216. (b) Bold, G.; Faessler, A.; Capraro, H.; Cozens, R.; Klimkait,
T.; Lazdins, J.; Mestan, J.; Poncioni, B.; Roesel, J.; Stover, D.; Tintelnot-
Blomley, M.; Acemoglu, F.; Beck, W.; Boss, E.; Eschbach, M.; Huerlimann,
T.; Masso, E.; Roussel, S.; Ucci-Stoll, K.; Wyss, D.; Lang, M. J. Med.
Chem. 1998, 41, 3387-3401. (c) Rusano, G. L.; Bilello, J. A.; Preston, S.
L.; O’Mara, E.; Kaul, S.; Schnittman, S.; Echols, R. J. Infect. Dis. 2001,
183, 1126-1129. (d) Gong, Y.; Robinson, B. S.; Rose, R. E.; Deminie,
C.; Spicer, T. P.; Stock, D.; Colonno, R. J.; Lin, P. Antimicrob. Agents
Chemother. 2000, 44, 2319-2326. (e) Robinson, B. S.; Riccardi, K. A.;
Gong, Y.; Guo, Q.; Stock, D. A.; Blair, W. S.; Terry, B. J.; Deminie, C.
A.; Djang, F.; Colonno, R. J.; Lin, P. Antimicrob. Agents Chemother. 2000,
44, 2093-2099. (f) Rabasseda, X.; Silvestre, J.; Castaner, J. Drugs Future
1999, 24, 375-380.
The original process for BMS-232632 (Scheme 1) was
satisfactory for preparation of material for early toxicology
† Process Research and Development, P.O. Box 4000, Princeton, New Jersey
08543.
‡ Process Research and Development, P.O. Box 191, New Brunswick, New
Jersey 08903.
(2) (a) Giordano, C.; Pozzoli, C.; Benedetti, F. W.O. Patent 012083, 2001. (b)
Fassler, A.; Bold, G.; Capraro, H.; Steiner, H. W.O. Patent 9746514, 1997.
(c) Reference 1a.
§ Biopharmaceutics Research and Development, P.O. Box 191, New Brun-
swick, New Jersey 08903.
10.1021/op025504r CCC: $22.00 © 2002 American Chemical Society
Published on Web 04/20/2002
Vol. 6, No. 3, 2002 / Organic Process Research & Development
•
323

Scheme 1. Original Synthesis of BMS-232632
Scheme 2. Retrosynthesis for BMS-232632
Scheme 3. Synthesis of hydrazino carbamate 4
sive O-(1,2-dihydro-2-oxo-1-pyridyl)-N,N,N′,N′-tetramethyl-
uronium-tetrafluoro-borate (TPTU) was used.
In addition, a mediocre yield was obtained due in part to
the low enantiomeric purity of the epoxide starting material.
Besides the above-mentioned issues, environmentally un-
friendly solvents, such as highly flammable diethyl ether and
suspected carcinogen 1,4-dioxane, were employed in this
process. Therefore, a strong need for the development of a
more efficient and environmentally suitable synthesis for
BMS-232632 existed.
philes, the Suzuki coupling is highly tolerant of functionality.3
The requisite aldehyde 3 (Scheme 3) was prepared through
Suzuki coupling of boronic acid 9 to 2-bromopyridine. The
reaction was reasonably straightforward. However, it proved
critical to use a 4:3 mixture of toluene-ethanol as the solvent
in order to dissolve all the reactants, thus affording a facile
reaction. Also, use of a minimal amount (0.1 mol %) of Pd-
(PPh3)4 catalyst was required to deliver product with low
amount (5-50 ppm) of palladium. Under the optimal
conditions, the Suzuki coupling afforded the desired aldehyde
3 after crystallization from toluene-n-heptane in 79.5% yield
and with HPLC area % (AP) 94-99. It was subsequently
found that the toluene solution of crude aldehyde 3 resulting
from an extractive workup could be used in the next step
without further purification. Curiously, wide variations in
residual Pd levels (5-400 ppm) in the product were observed
from batch to batch.
Synthesis Strategy
As depicted in the retrosynthetic analysis (Scheme 2), we
envisaged new methods for preparation of hydrazinocar-
bamate 4 and epoxide 6. Aldehyde 3 might be accessed
through a Suzuki coupling of commercially available 4-formyl-
benzeneboronic acid (9) with 2-bromopyridine, eliminating
the need to protect and deprotect the aldehyde functionality.
To obtain enantiomerically pure epoxide 6, we planned to
use the commercially available diol 10, possessing the
S-configuration at the carbinol center. A selective formation
of 6 was expected through an intramolecular SN2 displace-
ment reaction. The convergent formation of BMS-232632
from intermediates 4 and 6 is analogous to the original
synthesis.2
The subsequent condensation of aldehyde 3 with tert-butyl
carbazate was initially carried out by heating in ethanol and
then precipitating the product by the addition of water. This
procedure afforded a 93% yield of 11 with less than optimal
purity, ∼95 AP. We found that the use of a solvent mixture
Results and Discussion
Synthesis of Hydrazinocarbamate 4. Unlike the metal-
catalyzed couplings of Grignard reagents with aryl electro-
(3) For reviews, see: (a) Miyaura, N.; Suzuki, A. Chem. ReV. 1995, 95, 2457-
2483. (b) Suzuki, A. J. Organomet. Chem. 1999, 576 (1-2), 147-168.
324
•
Vol. 6, No. 3, 2002 / Organic Process Research & Development

Scheme 4. Synthesis of (2S,3R)-epoxide 6
of 4:3 toluene-2-propanol was ideal for both the condensa-
tion and subsequent crystallization. Heating the crude alde-
hyde 3 and tert-butyl carbazate in this mixture of solvents
for 2 h and subsequent cooling and filtration provided
compound 11 in 92% yield and with ∼100 AP. In addition,
the residual Pd level was reduced about 10-fold compared
to that in the aldehyde starting material. This procedure was
scaled up to afford 5 kg of 11 in a single batch.
Our studies showed that under the original conditions (Pd/
C, MeOH, and 1 atm H2),2 the conversion of 11 to 4 was
variable, ranging from 80 to 98%. Also problematic were
the side reactions involving hydrogenolysis and overreduc-
tion.4 In contrast to the hydrogenolysis and overreduction
byproducts, the removal of the unreacted hydrazone 11
through crystallization proved difficult due to its extremely
low solubility compared to product 4. Thus, a more efficient
conversion was desired. Replacement of Pd/C with Pd-
(OH)2/C provided improved conversion (>97%) while the
extent of hydrogenolysis was less than 2%. Hydrogenation
of 11 using Pd(OH)2 (MeOH, 1 atm H2, 6 h) followed by
crystallization provided 4 in 80-90% yields and with 94-
99 AP (1-5 AP 11). Despite these encouraging results, the
required catalyst loading of 10 wt % of Pd(OH)2 rendered
this reduction unattractive.
air, it readily underwent air oxidation in MeOH or MeCN
solution to afford hydrazone 11. Therefore, solutions of 4
were handled under inert gas.
Synthesis of (2S,3R)-Epoxide 6. Methods for the conver-
sion of a vicinal diol to an epoxide are well established.12
Through modification of the reported procedures,12d-f
a
practical process for this transformation was developed. As
illustrated in Scheme 4, a quantitative yield of silyl mesylate
12 was produced in one pot through selective silylation (2.2
equiv of TBSCl, DMAP (cat), 2.2 equiv of Et3N, 50 °C,
toluene) and subsequent mesylation (1.1 equiv of MsCl, 5
°C, toluene) of diol 10. This oily intermediate was carried
into the next step without further purification.
We found that desilylation of 12 could be effected using
the inexpensive reagent ammonium fluoride13 instead of the
more traditional tetrabutylammonium fluoride. The resulting
solid product 13 could be readily isolated and further purified
through crystallization from IPA-H2O in 80% yield with
AP 98. Several bases were screened for epoxide formation
from 13. KOtBu was found to be the base of choice, giving
enantiomerically pure epoxide product in 90% crystallized
yield. Interestingly, use of KOH in EtOH gave an ethanolysis
product 14 exclusively.14 The reaction with KOtBu in THF-
IPA was reproducible on scale-up. Following the above four-
step procedure, a total of 2.7 kg of epoxide 6 was prepared
in 70% average yield with 100 AP. The primary attraction
of this route is that product 6 with 100% ee is obtained
cleanly.
5
Chemical reductions of hydrazone 11 using NaBH4
(various solvents, with additives such as NiCl2,6 ZrCl4,7 and
HOAc), NaB(OAc)3H-HOAc,8 NaBH2S3,9 LiBH4, and
Al(iBu)2H were sluggish; generally <20% product was
observed. Reduction with Zn/HOAc10 in refluxing methanol
resulted in ∼40% conversion to hydrazinocarbamate 4.
In search for a more effective procedure, we examined
catalytic phase-transfer hydrogenation.11 Although reduction
with 2 equiv of sodium formate and 1 mol % of Pd/C
required an elevated temperature, complete reaction could
be obtained reliably in 2 h at 56 °C, yielding crude 4 with
∼98 AP. Crystallization from tert-butyl methyl ether and
n-heptane furnished the desired hydrazinocarbamate in 78%
isolated yield with 100 AP. This reduction procedure was
successfully scaled up to afford 3.7 kg of 4 (78% yield, AP
100) in a single batch. It is noteworthy to mention that
although solid hydrazinocarbamate 4 was perfectly stable in
(12) Via conversion of 1° alcohol to a leaving group: (a) Clayden, J.; McElroy,
A. B.; Warren, S. J. Chem. Soc., Perkin Trans. 1 1995, 15, 1913-1934.
(b) Shibuya, M.; Terauchi, H. Tetrahedron Lett. 1987, 28, 2619-2623. (c)
Masaki, Y.; Serizawa, Y.; Nagata, K.; Oda, H. Tetrahedron Lett. 1986,
27, 231-234. Via conversion of 2° alcohol to a leaving group: (d) Castejon,
P.; Pasto, M.; Moyano, A.; Pericas, M. A.; Riera, A. Tetrahedron Lett.
1995, 36, 3019-3022. (e) Dondoni, A.; Fantin, G.; Fogagnolo, M.; Medici,
A. Angew. Chem, Int. Ed. Engl. 1986, 25, 835-839. (f) Qlan, X.; Moris-
Varas, F.; Wong, C. H. Bioorg. Med. Chem. Lett. 1996, 6, 1117-1122.
(13) (a) Hassan, A. E.; Nishizono, N.; Minakawa, N.; Shuto, S.; Matsuda, A. J.
Org. Chem. 1996, 61, 6261-6267. (b) Seki, M.; Kondo, K.; Kuroda, T.;
Yamanaka, T.; Iwasaki, T. Synlett 1995, 609-611.
(14) It is possible that the desired epoxide 6 was initially formed and that it
rapidly underwent a rearrangement under the reaction conditions. This
possibility was supported by the finding that subjection of 6 to excess KOH
in EtOH resulted in the formation of 14. Compound 14 was isolated and
fully characterized. 1H NMR (400 MHz, CDCl3) δ 7.4-7.2 (m, 5H), 5.14
(d, J ) 9.6 Hz, 1H), 3.75 (m, 1 H), 3.45 (q, J ) 7.1 Hz, 2H), 3.36 (d, J )
7.0 Hz, 2H), 3.24 (s, 1H), 2.89 (m, 2H), 1.40 (s, 9H), 1.13 (t, J ) 7.1 Hz,
3H); 13C NMR (100 MHz, CDCl3) δ 14.9, 28.4, 38.3, 52.8, 66.5, 69.3,
72.4, 72.6, 76.7, 79.1, 126.1, 128.2, 129.2, 138.1, 155.6; IR (1% KBr pellet)
3440, 1713, 1695 cm-1; [R]D ) -31.3 (c ) 1, MeOH, 22 °C); Anal. Calcad
for C18H29NO4: C, 65.99; H, 8.80; N 4.53. Found: C, 65.88; H, 8.82; N
4.26.
(4) Byproducts i and ii were formed by the hydrogenolysis and iii was formed
by overreduction.
(5) (a) Karmakar, D.; Rajapati, D.; Sandhu, J. S. J. Chem. Res. Synop. 1996,
10, 464-465. (b) Wann, R. S.; Thorsen, T. P.; Kreevoy, M. M. J. Org.
Chem. 1981, 46, 2579-2581.
(6) Rao, H. P.; Reddy, K. S.; Turnbull, K.; Borchers, V. Synth. Commun. 1992,
22, 1339-1343.
(7) Chary, K. P.; Ram, S. R.; Salahuddin, S.; Iyengar, D. S. Synth. Commun.
2000, 30 (19), 3559-3563.
(8) (a) Singh, J.; Sharma, M.; Kaur, I.; Kad, L. C. Synth. Commun. 2000, 30,
1515-1519. (b) Baruah, B.; Dutta, P. M.; Boruah, A.; Prajapati, D.; Sandhu,
J. S. Synlett 1999, 4, 409-410.
(9) Ramanujam, V. S.; Trieff, M. N. J. Chem. Soc., Perkin. Trans. 2 1976, 15,
1811-1815.
(10) Hoffmann, R. W.; Sieber, W. Justus Liebigs Ann. Chem. 1967, 703, 96-
100.
(11) (a) Watanabe, T.; Nishiyama, S.; Yamamura, S.; Kato, K.; Nagai, M.;
Takita, T. Tetrahedron Lett. 1991, 32, 2399-2400. (b) Berlin, W. K.;
Zhang, W.-S.; Shen, T. Y. Tetrahedron 1991, 47, 1-20.
Vol. 6, No. 3, 2002 / Organic Process Research & Development
•
325

Scheme 5. Synthesis of BMS-232632 and its bisulfate salt
19
50 °C. Although a clean deprotection product 18 was
generated, the isolation of this trihydrochloride salt proved
difficult due to its hygroscopic nature. Attempted crystal-
lization or precipitation from several solvent systems afforded
material which was still extremely hygroscopic. The corre-
sponding MsOH and HI salts 18a and 18b were isolated by
deprotection using MsOH and TMSI, respectively. These
materials were foamy solids and also very hygroscopic. To
circumvent the problematic isolation of 18, a procedure for
the deprotection and in situ coupling with N-methoxycar-
bonyl-L-tert-leucine was developed. After removal of the Boc
groups from 15 with HCl, the wet THF layer was decanted.
The residual oily product was then washed with THF and
taken up in dichloromethane. The resulting solution was
subjected to the original coupling conditions (3 equiv of
N-methoxycarbonyl-L-tert-leucine, 1 equiv of TPTU, 6 equiv
of EtNiPr2, CH2Cl2),2 affording BMS-232632 in ∼80% yield
with high purity. To avoid use of the expensive reagent
TPTU, the coupling step was reevaluated. Replacement of
TPTU by a combination of water-soluble carbodiimide
[WSC, 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hy-
drochloride] and 1-hydroxybenzotrizaole (HOBT) formed the
coupling product in a similar yield. The solution of crude
trihydrochloride salt 18 was treated with diisopropylethyl-
amine and added to a mixture of N-methoxycarbonyl-L-tert-
leucine, WSC and HOBT in dichloromethane. A minimum
of 2.6 equiv each of WSC and HOBT, and 2.5 equiv of
N-methoxycarbonyl-L-tert-leucine were required for complete
reaction.15 Using this process we prepared 3.5 kg of BMS-
232632 (yield 82% from 15, AP 98.8).
Formation of the Bisulfate Salt 19.16 Free base BMS-
232632 is highly insoluble in water (<1 µg/mL at 24 °C)
and has poor oral bioavailability in animals. Therefore,
several acid salts were explored. A number of commonly
used acid salts such as the hydrochloride, benzenesulfonate,
methanesulfonate, p-toluenesulfonate, phosphate, nitrate, 1,2-
ethanedisulfonate, 2-hydroxyethanesulfonate, sulfate, and
bisulfate were evaluated. While these salts in their crystalline
form exhibited higher aqueous solubility (0.5-5 mg/mL)
than the free base, bisulfate salt 19 was chosen as the final
form for development due to its superior solubility (4-5 mg/
mL).
Synthesis of BMS-232632. As illustrated in Scheme 5,
the coupling of hydrazinocarbamate 4 and epoxide 6 was
performed in refluxing IPA for 24 h, followed by the addition
of water to precipitate the crude product. Subsequent
recrystallization from MeCN-H2O furnished 15 in 85% yield
with AP 98. In the coupling reaction, low levels (<5%) of
impurity 16,
Bisulfate salt 19 was obtained in two crystalline forms,
as determined by powder X-ray diffraction patterns. Type-1
crystals were formed in acetone, MeCN, or EtOH; the Type-
II crystals were obtained from IPA. Due to its reproducibility,
the Type-I form was selected for further development. The
Type-I form 19 was initially prepared in MeCN with 9 M
H2SO4 and in the presence of seed crystals. Although a very
easily filterable, dense, sandy solid was obtained, this
procedure was abandoned due to concerns of possible toxicity
associated with residual trace amounts of MeCN. Alterna-
tively, the formation of 19 was performed with concentrated
H2SO4 in EtOH at ambient temperature. Direct crystallization
a consequence of epoxide opening by the solvent, were
observed by in-process HPLC. Employment of tert-butyl
alcohol produced 15 in comparable yields; however, the
reaction required ∼48 h for completion. Replacement of IPA
with several aprotic solvents (toluene, acetonitrile) resulted
in very sluggish reactions. Impurity 16 as well as 17, resulting
from epoxide opening by the hydrogenolysis product Boc-
NHNH2 mentioned above,4 was purged during the initial
precipitation of the product.
(15) In-process HPLC analysis indicated that the amide bond on the amine was
formed prior to that on the hydrazine. Fewer equivalents of WSC, HOBT
and N-methoxycarbonyl-L-tert-leucine resulted in an incomplete reaction,
evidenced by a high ratio of monocoupling product to BMS-232632.
(16) Singh, J.; Pudipeddi, M.; Lindrud, D. M. W.O. Patent 9936404, 1999.
Subsequent removal of the two Boc groups in 15 was
performed with concentrated hydrochloric acid in THF at
326
•
Vol. 6, No. 3, 2002 / Organic Process Research & Development

by addition of n-heptane provided the bisulfate salt 19 as an
easily filterable solid in 85% yield with >99.6 AP on 3.6
kg input scale. H NMR analyses of the free base BMS-
232632 and bisulfate salt 19 indicated that the pyridine ring
in 19 was protonated, as evidenced by the downfield
chemical shifts of protons Ha-d compared to those of the
free base (Scheme 5, starred chemical shifts are for the
protons in the salt 19).
(113.4 g, 0.382 mol), and 10% palladium on activated carbon
(4.05 g) in 382 mL of ethanol, a solution of sodium formate
(46.7 g, 0.687 mol) in water (70 mL) was added. The
resulting mixture was heated to 57 °C, and slight gas
evolution was observed within 5 min. After 1.5 h, the reaction
was cooled to about 40 °C, and MTBE (380 mL) was added.
The solid was removed through filtration. The filtrate was
washed with 10% sodium chloride solution (550 mL), dried
over anhydrous sodium sulfate (100 g), filtered, and con-
centrated to provide a colorless gel which was dissolved in
MTBE (110 mL) under argon atmosphere at 50 °C. n-
Heptane (350 mL) was added dropwise to the warm solution,
and the resulting mixture was slowly cooled to 23 °C and
stirred for 16 h. Filtration and washing with heptane (2 ×
300 mL) afforded 4 as a white solid (89.1 g, 78% yield, AP
100).
[3-tert-Butyl-dimethylsilanyloxy-2(S)-[(methylsulfonyl)-
oxy]-1(S)-(phenylmethyl)propyl]-carbamic Acid, 1,1-Di-
methylethyl Ester (12). A solution of diol 10 (544 g, 1.034
mol) in 1.2 L of toluene was heated to 88 °C, and a clear
solution was obtained. The solution was then cooled to 50
°C. Dimethylamino pyridine (23.6 g, 0.195 mol) and
triethylamine (325 mL, 2.32 mol) were charged followed
by the slow addition of tert-butyl-dimethylsilyl chloride (350
g, 2.32 mol) while keeping the internal temperature around
50 °C. The reaction mixture was cooled to 0 °C over 3 h.
Triethylamine (417 mL) was added followed by the slow
addition of trifluoromethanesulfonyl chloride (198 mL),
keeping the internal temperature under 5 °C. The resulting
mixture was stirred at 0 °C for about 3 h. The solid was
filtered through Celite and washed with toluene (2 × 700
mL). The filtrate was washed with water (4 L), 1 N HCl (4
L), and brine (4 L), in that order, and then concentrated in
a vacuum to afford 1.04 kg of product 12 as a yellow oil.
This product was subjected to the next step without further
purification.
[3-Hydroxy-2(S)-[(methylsulfonyl)oxy]-1(S)-(phenyl-
methyl)propyl]-carbamic Acid, 1,1-Dimethylethyl Ester
(13). Into a reactor was charged ammonium fluoride (358
g, 9.67 mol), a solution of the crude mesylate 12 (1.04 kg,
1.034 mol) in methanol (5.6 L), and acetic acid (550 mL).
The mixture was stirred at ambient temperature for 11 h.
The reaction mixture was concentrated to dryness to afford
a solid, which was dissolved in 11 L of methyl tert-butyl
ether. The resulting solution was washed with water (5 L),
5% sodium bicarbonate (3 × 4 L), and brine (4 L) and then
dried over MgSO4 (300 g). Filtration and partial concentra-
tion afforded 5 L of solution. The concentrated solution was
then cooled to 4 °C and stirred at this temperature for 18 h
to give a slurry. The solid was filtered, washed with cold
MTBE (200 mL) and dried under partial pressure to afford
489.1 g of 13. The filtrate was concentrated to 1 L, cooled
to 4 °C, and stirred at this temperature for 18 h to give a
slurry. Another 61.7 g of solid was obtained after filtration
and drying. Thus, a total of 550.8 g of product 13 was
obtained as a white solid (80% yield, AP 98).
1
Summary
A short and efficient synthesis of hydrazinocarbamate 4
has been developed. A novel process for selective formation
of epoxide 6 from diol 10 has been established. The
convergent formation of BMS-232632 from 4 and 6 has been
significantly improved, providing the free base BMS-232632
in good yield. Finally, the bisulfate salt 19 with enhanced
bioavailability was identified and obtained in excellent purity.
Experimental Section
All new compounds described in the Experimental Section
were fully characterized. Analytical and spectral data for
these compounds are for lab batches. HPLC analysis results
are described as area % (AP).
4-Pyridin-2-yl-benzaldehyde (3). A mixture of 4-formyl-
phenyl-boronic acid (9) (200 g, 666.9 mmol) and 2-bro-
mopyridine (110.6 g, 700.3 mmol) in 700 mL of 4:3 toluene/
95% ethanol was purged with N2 for 15 min and then heated
under a N2 atmosphere, resulting in a clear solution. A slurry
of Pd(PPh3)4 (1 g, 0.86 mmol) in 10 mL of a 4:3 mixture of
toluene and 95% ethanol was added, followed by 1.16 L of
3 M aqueous Na2CO3. The resulting mixture was gently
refluxed at ∼76 °C for 13 h, at which time an additional
10-mL slurry of Pd(PPh3)4 (540 mg, 0.467 mmol) was added.
The reaction was continued for another 7 h and then cooled
to ambient temperature. The solid was removed by filtration,
and the filtrate was poured into a separatory funnel. The
layers were separated, and the aqueous layer was washed
with toluene (2 × 250 mL). The combined organic layers
were stirred over charcoal (10 g) for 5 h, filtered, and partially
concentrated in vacuo to provide a toluene solution of
aldehyde 3, which contained approximately 150 mL of
toluene. This unpurified product was submitted directly to
the next reaction. Alternatively, the crude 3 could be purified
through crystallization from a mixture of 1:1 toluene and
n-hexane. Typically, a 80% crystallized yield was obtained.
N-1-(tert-Butyloxycarbonyl)-N-2-[4-(pyridin-2yl)ben-
zylidene]-hydrazone (11). To a 100-L glass plant reactor
was added 4-pyridin-2-yl-benzaldehyde (3) (4.131 kg, 22.25
mol), tert-butyl carbazate (2.967 kg, 22.23 mol), 6.05 L of
2-propanol, and 8.01 L of toluene. The mixture was heated
to reflux (80-85 °C) under inert atmosphere for 2 h, cooled
to 22 °C over 4 h, and stirred at that temperature for about
14 h. The resulting mixture was filtered, and the filter cake
was washed with a cold mixture of toluene and heptane (1:
3, 0-5 °C, 4 × 3.56 L). The cake was air-dried under
vacuum to afford 5.644 kg of 11 (85.3% yield, 99.8 AP).
N-1-(tert-Butyloxycarbonyl)N-2-[4-(pyridin-2-yl)ben-
zylidene]-hydrazine (4). In to a mixture of hydrazone 11
N-(tert-butyloxycarbonyl)-2(S)-amino-1-phenyl-3(R)-
3,4-epoxy-butane (6). To a clear solution of hydroxy
Vol. 6, No. 3, 2002 / Organic Process Research & Development
•
327

mesylate 13 (629.9 g, 1.75 mol) in a mixture of IPA (6.3 L)
and THF (1.8 L) at 17 °C, was added KOtBu (207 g, 95%,
1.75 mol) over 20 min. The mixture was stirred for 1.5 h
followed by addition of 30 mL of acetic acid over 15 min.
The resulting solution was concentrated under vacuum to
dryness to afford a white solid. The solid was dissolved in
MTBE (9.0 L), and the resulting solution was washed with
water (4.5 L), saturated sodium bicarbonate solution (4.5 L),
and brine (4.5 L), dried over anhydrous Na2SO4, filtered,
and concentrated to give an oil (455.2 g). The oil was diluted
with hexane (1.3 L) followed by addition of water (200 mL).
The mixture was cooled to -4 °C, and solid was observed.
The solid was collected by filtration, washed with 700 mL
of cold hexane (0 °C), and dried under vacuum for 18 h to
give epoxide 6 as a white solid (400.5 g, 88% yield, AP
97).
1-[4-(Pyridin-2-yl)phenyl]-5(S)-2,5-bis[(tert-butyloxy-
carbonyl)-amino]-4(S)-hydroxy-6-phenyl-2-azahexane (15).
To a mixture of compounds 4 (1.68 kg, 5.58 mol) and 6
(1.735 kg, 6.56 mol) was added 23.83 L of 2-propanol. The
solution was refluxed for 24 h under an inert atmosphere
and then cooled to 22 °C. Water (28.9 L) was added, and
the resulting mixture was stirred at 22 °C for 16 h to afford
a slurry. The solid was collected by filtration and washed
with water (2 × 11 L) and cold MTBE (∼0 °C, 2 × 4 L).
This wet cake was dissolved in acetonitrile (44 L), and water
(44 L). The crystallized material was filtered and dried to
give 2.681 kg (85.4% yield, AP 99) of 15.
N-Methoxycarbonyl-L-tert-leucine. To a solution of
L-tert-leucine (750 g, 5.7 mol) was added 9.42 L of 2 N
aqueous sodium hydroxide and methylchloroformate (882
mL, 11.34 mol). The resulting mixture was heated at 60 °C
for 18 h and then cooled to 22 °C. The reaction mixture
was acidified by addition of 4 N HCl (4.8 L) to pH ∼2
followed by an extractive workup with ethyl acetate (3 × 6
L). The combined extracts were washed with brine and
concentrated under vacuum to afford an oil. To the oil 7.7
L hexane was added to produce a large solid mass. The solid
was collected by filtration and air-dried to afford 883 g of
N-methoxycarbonyl-L-tert-leucine (82% yield).
1-[4-(Pyridin-2-yl)phenyl]-5(S)-2,5-bis{[N-(methoxy-
carbonyl)-L-tert-leucinyl]amino}-4(S)-hydroxy-6-phenyl-
2-azahexane (BMS-232632). To a stirred solution of 15 (3.5
kg, 6.22 mol) in 17.5 L of THF, concentrated HCl (2.38 L,
12 N) was added dropwise at 22 °C to give an amber-colored
solution. The mixture was stirred for 3 h at 45-55 °C to
produce a biphasic mixture. The mixture was cooled to 22-
25 °C over 3 h, and the wet THF layer was decanted away
from a brown oil. The oil was rinsed with 7 L of THF, which
was decanted. The brown oil was dissolved in a mixture of
14 L of CH2Cl2 and 6.65 L of DIPEA. The resulting mixture
was slowly transferred by pump at 22 °C into a premixed
solution of N-methoxycarbonyl-L-tert-leucine (2.94 kg, 15.5
mol), HOBT (2.32 kg, 17.2 mol), and WSC (3.13 kg, 16.3
mol) in 4.4 L of CH2Cl2. The new reaction mixture was
stirred for 3-4 h at 25 °C. The reaction mixture was then
washed with water (1 × 13.1 L), NaHCO3 (1 × 13.1 L),
and brine (1 × 14 L). The organic layer was concentrated
to a viscous oil and treated with 42 L of a 98:2 isopropyl
ether-EtOH solution at 55-70 °C until a mild reflux was
observed. The slurry was cooled to 25-30 °C, and the solid
was collected by filtration, washed with 13.3 L of a 98:2
mixture of isopropyl ether-EtOH, and dried at 35-40 °C
under vacuum to give 3.86 kg of crude free base. The crude
product was crystallized from 36.4 L of a mixture of
ethanol-water (45/55), filtered, washed, and dried to afford
3.64 kg of BMS-232632 in 82.1% yield and with AP 98.8.
1-[4-(Pyridin-2-yl)phenyl]-5(S)-2,5-bis{[N-(methoxy-
carbonyl)-L-tert-leucinyl]amino}-4(S)-hydroxy-6-phenyl-
2-azahexane, Bisulfate Salt (19). To a solution of BMS-
232632 (3.6 kg, 5.1 mol) in 27 L of ethanol, concentrated
H2SO4 (309 mL, 5.65 mol) was added dropwise at 22 °C to
give a yellow-orange colored solution. After complete
dissolution, the solution was filtered and transferred into
another vessel. Then 18 L of n-heptane was added followed
by seeding with 10 g of pure 19. After seeding, an additional
12 L of n-heptane was added. The mixture was stirred for
15 h. The solid was collected through filtration and washed
with heptane-ethanol (1:1) solution and dried to afford 3.49
1
kg of 19 as a white solid (85.2% yield, AP 99.8). H NMR
(270 MHz DMSO-d6) δ 9.35 (s, 1H), 9.0 (d, J ) 4.1 Hz,
1H), 8.64 (t, J ) 8.2 Hz, 1 H), 8.43 (d, J ) 8.2 Hz, 1H),
8.07 (m, 2H), 8.32 (t, J ) 8.2 Hz, 1H), 8.16 (d, J ) 8.2 Hz,
1H), 8.73 (d, J ) 8.2 Hz, 1H), 7.67 (m, 1H), 7.32 (d, J )
4.7 Hz, 1H), 7.27 (m, 3H), 7.09 (d, J ) 9.4 Hz, 1H), 7.03
(d, J ) 9.3 Hz, 1H), 4.20 (m, 1H), 3.97 (m, 2H), 3.76 (d, J
) 9.4 Hz, 1H), 3.62 (m, 2H), 3.64 (s, 3H), 3.60 (s, 3H),
2.90-2.63 (m, 2H), 0.89 (s, 9H), 0.75 (s, 9H); 13C NMR
(68 MHz DMSO-d6) δ 26.3, 26.7, 33.4, 33.6, 37.7, 51.4,
51.6, 60.7, 61.0, 61.3, 63.2, 68.3, 124.2, 124.8, 125.9, 127.5,
128.0, 129.1, 129.3, 131.3, 139.0, 141.6, 144.2, 144.7, 152.3,
156.5, 170.1; IR (1% KBr pellet) 3426, 2959, 1701, 1676,
1653, 1514, 1370, 1244, 1065, 777; [R]D ) -46.1 (c ) 1,
1:1 MeOH/H2O, pH ) 2.6, 22 °C); Anal. Calcd for
C38H52N6O7‚H2SO4‚0.35 H2O: C, 56.40; H, 6.81; N 10.39;
S, 3.96; H2O, 0.78. Found: C, 56.54; H, 6.81; N 10.35; S,
3.98; H2O, 0.74.
Acknowledgment
We thank the Analytical R & D Department, Bristol-
Myers Squibb for their valuable support during the course
of this work.
Received for review January 18, 2002.
OP025504R
328
•
Vol. 6, No. 3, 2002 / Organic Process Research & Development