Quaternary chiral center via diastereoselective enolate amination enables the synthesis of an anti-inflammatory agent

Article abstract of DOI:10.1021/op900255k

The D-leucine amino acid residue necessary for the synthesis of BMS-561392,1, was employed as a chiral directing group for a diastereoselective enolate animation to establish the quaternary chiral center. Enhanced diastereomeric ratios were observed while conducting the enolate amination with 1-chloro-1-nitrosocyclo- pentane 6 in the presence of LiCl. Analogies are drawn between known tertiary amide amino alcohol chiral auxiliaries which have been used to effect diastereoselective enolate alkylations and aminations. Once the stereochemical features of 1 were established, an efficient reaction sequence was devised to complete its synthesis. During the course of this research, accelerated reaction calorimetry (ARC) data substantiated that the aminating agent 1-chloro-1- nitrosocyclopentane 6 was not safe to use as a neat compound. Consequently, a preparation and use of 6 as a stock solution in methyl tert-butyl ether (MTBE) was developed that rendered it safe for use.

Full text of DOI:10.1021/op900255k

Organic Process Research & Development 2010, 14, 159–167
Quaternary Chiral Center via Diastereoselective Enolate Amination Enables the
Synthesis of an Anti-inflammatory Agent†
Nicholas A. Magnus,*,¶ Silvio Campagna,‡ Pat N. Confalone,§ Scott Savage,⊥ David J. Meloni,⊥ Robert E. Waltermire,⊥
Robert G. Wethman,⊥ and Mathew Yates¶
Bristol-Myers Squibb Company, Process Research and DeVelopment, One Squibb DriVe, New Brunswick,
New Jersey 08903, U.S.A.
Abstract:
The -leucine amino acid residue necessary for the synthesis of
the perspective of ease of administration, reduced cost of
therapy, and patient compliance. The clinical development
candidate, BMS-561392, 1, is a TACE inhibitor, an inhibitor
of the enzyme which processes proTNFR to yield the soluble
TNFR cytokine product.2 The previously reported syntheses of
1 utilized racemic methods to establish the tetra-substituted
carbon stereocenter and relied on a subsequent chromatographic
separation of diastereomers2d or an enzymatic resolution2e to
establish the absolute stereochemistry. This article describes a
diastereoselective synthesis that affords the desired single
diastereomer of 1.
D
BMS-561392, 1, was employed as a chiral directing group for a
diastereoselective enolate amination to establish the quaternary
chiral center. Enhanced diastereomeric ratios were observed while
conducting the enolate amination with 1-chloro-1-nitrosocyclo-
pentane 6 in the presence of LiCl. Analogies are drawn between
known tertiary amide amino alcohol chiral auxiliaries which have
been used to effect diastereoselective enolate alkylations and
aminations. Once the stereochemical features of 1 were established,
an efficient reaction sequence was devised to complete its synthesis.
During the course of this research, accelerated reaction calorimetry
(ARC) data substantiated that the aminating agent 1-chloro-1-
nitrosocyclopentane 6 was not safe to use as a neat compound.
Consequently, a preparation and use of 6 as a stock solution in
methyl tert-butyl ether (MTBE) was developed that rendered it
safe for use.
Introduction
Results and Discussion
Tumor necrosis factor alpha (TNFR) is a cytokine produced
primarily by activated monocytes, macrophages, and T-cells,
which exhibits a wide range of biological actions relevant to
acute and chronic inflammation. The administration of neutral-
izing anti-TNFR antibodies or fusion proteins has been dem-
onstrated to be effective in the treatment of inflammatory
diseases in a clinical setting including rheumatoid arthritis,
psoriasis, Crohn’s diseases, and ankylosing spondylitis.1 The
clinical candidate synthesized in this article arose from a
discovery program targeted to identify an orally administered,
selective, small molecule capable of reducing the levels of
circulating TNFR. An oral agent was judged desirable from
The objective of the research described herein was to devise
an efficient asymmetric synthesis for potential commercializa-
tion of 1 with a focus on the key tetra-substituted carbon atom.
The preparation outlined in Scheme 1 gives access to the
γ-lactam 5. The methyl ester 2 in Scheme 1 is a known
compound.3 Allylation and ozonolysis of similar compounds
is described in the literature.4 The ozonolysis was run in EtOAc
to permit facile purification of the aldehyde 4 Via crystallization
from EtOAc/heptane, and the intermediate ozonide was reduced
by the combination of zinc and acetic acid to enable removal
of the byproduct salts by filtration. Next, an efficient reductive
† This manuscript is dedicated to Professor Philip D. Magnus, FRS, in honor
of his 66th birthday.
(2) (a) Lee, D. C.; Sunnarborg, S. W.; Hinkle, C. L.; Myers, T. J.;
Stevenson, M.; Russell, W. E.; Castner, B. J.; Gerhart, M. J.; Paxton,
R. J.; Black, R. A.; Chang, A.; Jackson, L. F. Ann. N.Y. Acad. Sci.
2003, 995, 22. (b) Newton, R. C.; Decicco, C. P. J. Med. Chem. 1999,
42 (13), 2295. (c) DasGupta, S.; Murumkar, P. R.; Giridhar, R.; Yadav,
M. R. Bioorg. Med. Chem. 2009, 17 (2), 444. (d) Maduskuie, T. P.,
Jr.; Duan, J.; Mercer, S. E. Novel Lactam Metalloprotease Inhibitors.
(Dupont Pharmaceuticals Company, U.S.A.). PCT Int. Appl. WO
2002004416 A2 20020117, 2002. (e) Savage, S. A.; Waltermire, R. E.;
Campagna, S.; Bordawekar, S.; Dalla Riva Toma, J. Org. Process
Res. DeV. 2009, 13, 510.
* Author to whom correspondence should be addressed. Phone: 317-651-
1470. Fax: 317-276-4507. E-mail: magnus_nicholas@lilly.com.
‡ Current address: Eisai Co. Ltd., 4 Corporate Drive, Andover, Massachusetts
01810-2447, U.S.A.
§ Current address: DuPont Crop Protection, Stine Haskell Research Center,
P.O. Box 30, Newark, Delaware 19714, U.S.A.
⊥ Current address: Bristol-Myers Squibb Company, Process Research and
Development, One Squibb Drive, New Brunswick, New Jersey 08903, U.S.A.
¶ Current address: Lilly Corporate Center, Indianapolis, Indiana 46285, U.S.A.
(1) (a) Klareskog, L.; Gaubitz, M.; Rodriguez-Valverde, V.; Malaise, M.;
Dougados, M.; Wajdula, J. Ann. Rheum. Dis. 2006, 65 (12), 1578.
(b) Tutrone, W. D.; Weinberg, J. M. Expert ReV. Dermatol. 2006, 1
(2), 207. (c) Bhatia, J. K.; Korelitz, B. I.; Panagopoulos, G.; Lobel,
E.; Mirsky, F.; Sultan, K.; DiSanti, W.; Chun, A.; Keenan, G.; Mamun,
K. J. Clin. Gastroenterol. 2007, 41 (7), 677. (d) Gorman, J. D.; Sack,
K. E.; Davis, J. C., Jr. N. Engl. J. Med. 2002, 346 (18), 1349.
(3) (a) Williams, D. R.; Sing-Yeun, S. J. Org. Chem. 1982, 47, 2846. (b)
Weiss, P. J. Am. Chem. Soc. 1948, 70, 4263.
(4) (a) Edgar, M. T.; Pettit, G. R.; Smith, T. H. J. Org. Chem. 1978, 43,
4115. (b) Pour, M.; Spula´k, M.; Buchta, V.; Kubanova´, P.; Voprsalova´,
M.; Wso´l, V.; Fa´kova´, H.; Koudelka, P.; Pourova´, H.; Schiller, R.
J. Med. Chem. 2001, 44, 2701.
10.1021/op900255k  2010 American Chemical Society
Published on Web 12/03/2009
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
159

Scheme 1
Scheme 2
Scheme 3
amination/cyclization5/hydrolysis sequence was conducted with-
out isolation of the intermediates. Aldehyde 4 was dissolved in
diethoxymethane and treated with D-leucine methyl ester
hydrochloride, followed by Hunig’s base. This mixture was
exposed to sodium triacetoxyborohydride to afford the
corresponding secondary amine. Lactamization was in-
duced thermally and the ester hydrolyzed by the action
of aqueous lithium hydroxide to afford γ-lactam 5 in 86%
overall yield from 4 to 5.
The proposal for the asymmetric synthesis of 1 was to
generate the dianion (carboxylate, enolate) of γ-lactam 5 and
trap the enolate on carbon with an electrophilic source of
nitrogen (i.e., electrophilic enolate amination).6 The hypothesis,
based on models, was that the existing chiral center of the amino
acid would influence the stereochemistry of the newly formed
quaternary chiral center (Vide infra). To probe the question of
diastereoselection and dianion formation, a deuteration experi-
ment was performed. γ-Lactam 5 was dissolved in THF and 2
equiv of n-butyllithium was added at -70 °C to generate the
dianion (Scheme 2). The dianion mixture was warmed to 0 °C
briefly, and cooled back to -70 °C in an attempt to effect
complete double deprotonation. CD3OD was subsequently
added to the reaction mixture.
Oppolzer employed chloronitroso reagents to efficiently
aminate the enolate of a sultam tertiary amide (Scheme 3).7
This methodology was expected to be compatible with the
proposed chemistry in Scheme 2, and therefore considered
worth investigation.
The dianion of γ-lactam 5 was generated as before, and the
resulting reaction mixture was treated with 1-chloro-1-nitroso-
cyclopentane 68 as a MTBE stock solution (Scheme 4). The
amination of the dianion of γ-lactam 5 proceeded at the rate of
addition of 6 to give greater than 95% conversion to the desired
product as indicated by HPLC. After an aqueous acidic workup,
(5) (a) Trybulski, E. J.; Zhang, J.; Kramss, R. H.; Mangano, R. M. J. Med.
Chem. 1993, 36, 3533. (b) Moss, N.; Beaulieu, P.; Duceppe, J. S.;
Ferland, J. M.; Gauthier, J.; Ghiro, E.; Goulet, S.; Guse, I.; Llina`s-
Brunet, M.; Plante, R.; Plamondon, L.; Wernic, D.; De´ziel, R. J. Med.
Chem. 1996, 39, 2178. (c) Shiniji, F.; Shuichi, N.; Yoshikazu, S.;
Shigeo, N. Heterocyles 1981, 15, 819.
(6) Erdik, E. Tetrahedron 2004, 60, 8747, and references cited therein.
(7) (a) Oppolzer, W.; Tamura, O. Tetrahedron Lett. 1990, 31, 991. (b)
Oppolzer, W.; Tamura, O.; Deerberg, J. HelV. Chim. Acta 1992, 75,
1965.
(8) Mu¨ller, E.; Metzeger, H.; Fries, D. Chem. Ber. 1954, 87, 1449.
(9) Hashimoto, N.; Aoyama, T.; Shioi, T. Chem. Pharm. Bull. 1981, 29,
1475.
(10) (a) Myers, A. G.; Schnider, P.; Kwon, S.; Kung, D. W. J. Org. Chem.
1999, 64, 3322. (b) Myers, A. G.; Yang, B. H.; Chen, H.; McKinstry,
L.; Kopecky, D. J.; Gleason, J. L. J. Am. Chem. Soc. 1997, 119, 6496.
(c) Myers, A. G.; Gleason, J. L.; Yoon, T.; Kung, D. W. J. Am. Chem.
Soc. 1997, 119, 656. (d) Myers, A. G.; McKinstry, L. J. Org. Chem.
1996, 61, 2428. (e) Myers, A. G.; Yoon, T. Tetrahedron Lett. 1995,
36, 9429. (f) Myers, A. G.; Gleason, J. L.; Yoon, T. J. Am. Chem.
Soc. 1995, 117, 8488. (g) Myers, A. G.; Yang, B. H.; Chen, H.;
Gleason, J. L. J. Am. Chem. Soc. 1994, 116, 9361. (h) Larcheveque,
M.; Ignatova, E.; Cuvigny, T. Tetrahedron Lett. 1978, 3961. (i)
Larcheveque, M.; Ignatova, E.; Cuvigny, T. J. Organomet. Chem.
1979, 177, 5.
The deuterated product was examined by 1H NMR integra-
tion, which indicated a 5:1 diastereomer ratio (dr). This result
demonstrated that the dianion had formed and that the enolate
quench had been influenced stereochemically by the amino acid
residue. With the diastereoselection concept validated, the
research was next directed towards selecting an appropriate
electrophilic aminating agent.6
160
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development

Scheme 4
the hydroxylamine 8 was isolated. A 7:3 mixture of diastere-
omers was apparent by NMR and HPLC. In order to determine
which diastereomer was favored in the amination it was
necessary to convert hydroxylamine 8 into the known inter-
mediate 9 from the original synthesis of 1.2d To this end,
hydroxylamine 8 was esterified by reaction with trimethylsi-
lyldiazomethane in the presence of a mixture of MeOH and
toluene,9 and the N-O bond was cleaved reductively by the
action of zinc in acidic media.7 The relative configuration of
ester 9 was confirmed by comparison of its NMR spectra with
those of an authentic sample.
As indicated above, 6 (Scheme 4) was selected as the
aminating agent. We selected 6 after comparing its performance
with that of 1-chloro-1-nitrosocyclohexane in the amination
reaction.8 It was found that 6 gave moderately cleaner reaction
profiles, perhaps due to less hindrance (more reactive) and
corresponding attenuation of competing processes. Accelerated
reaction calorimetry (ARC) analysis revealed that neat 6 was
already decomposing at 22 °C when the analysis began, and a
maximum “self heat” rate of 400 °C min-1 was measured
(Figure 1a). To enable the safe use of 6, solutions of 6 were
evaluated to determine their relative safety. As a 33 weight
percent (wt %) solution in MTBE, 6 had an onset of decom-
position at 70 °C with a maximum “self heat” rate of 4 °C
min-1 (Figure 1b). As MTBE has a boiling point of 55-56
°C, this gave the additional safety that the solvent would boil
before decomposition of the reagent could occur. As a result,
all further studies and scale-up were performed with 6 as a 33
wt % solution in MTBE.
1-Chloro-1-nitrosocyclopentane 6 was prepared by reacting
cyclopentanone oxime with chlorine in MTBE with a ReactIR
1000 Dicomp probe submerged into the reaction mixture. The
React IR data indicated a complete conversion of the starting
oxime to 6 with no side reactions detected. An unidentified,
transient, thick precipitate was observed during the course of
the reaction that was at maximum concentration at the point
when the oxime was completely consumed. The ReactIR plots
in Figures 2 and 3 show complete consumption of the oxime
OH and production of the NdO double bond. The component
profiles (Figure 4) illustrate complete consumption of the oxime
Figure 1. ARC data comparing neat 6 to 33 wt % 6 in MTBE.
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
161

Figure 2. Profile of peak consumption at 3363 cm-1 of the oxime OH of cyclopentanone oxime in MTBE.
Figure 3. Profile of peak production at 1578 cm-1 of the nitroso NdO of 6 in MTBE.
The work of A. G. Myers10 and D. A. Evans11 (Scheme 5,
eqs 1 and 2; concerning the diastereoselective alkylation of a
tertiary amide enolate dianion) and the enolate amination work
of Badia12 appeared to have similarities to the chemistry in
Scheme 4. The specific similarities are a tertiary amide bearing
a chiral directing group, a lithium alkoxide of an amino alcohol
as part of the directing group which is similar in arrangement
to the lithium carboxylate 5a (Scheme 5), and a tertiary amide
dianion nucleophile that is trapped on the enolate carbon by an
electrophile.
The origin of the diastereoselectivity in the alkylations
presented in eqs 1 and 2 (Scheme 5) are proposed to be a result
of the Z-enolate geometry (alkyl group on the same side of the
enolate as oxygen). In the Z-enolate configuration, the alkoxide-
bearing side chain can adopt a staggered confirmation in which
Figure 4. Component profiles from the ReactIR experiment.
(red line), an intermediate buildup which coincided with the
transient precipitate (blue line), and clean production of 6 (black
line).
The ReactIR data gave us the confidence to assume a
quantitative yield of 6 after consumption of the starting oxime
was complete.
(11) (a) Evans, D. A.; Takacs, J. M. Tetrahedron Lett. 1980, 21, 4233. (b)
Sonnet, P.; Heath, R. R. J. Org. Chem. 1980, 45, 3137.
(12) (a) Anakabe, E.; Vicario, J. L.; Bad´ıa, D.; Carrillo, L.; Yoldi, V. Eur.
J. Org. Chem. 2001, 4343. (b) Vicario, J. L.; Bad´ıa, D.; Dom´ınguez,
E.; Crespo, A.; Carrillo, L.; Anakabe, E. Tetrahedron Lett. 1999, 40,
7123.
162
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development

Scheme 5
Scheme 6
the C-H bond R to nitrogen lies in the plane of the enolate
oxygen, in keeping with allylic strain arguments. This places
the alkoxide (perhaps solvent coordinated or aggregated) side
chain in a position to block one of the faces of the enolate and
enhance the diastereoselectivity of the ensuing enolate trapping.
The possible intermediates 5a and 5b illustrate that only the
Z-enolate can form due to ring constraints (Scheme 5, eq 3). In
addition, if the enolate carboxylate of γ-lactam 5 is configured
with the C-H bond R to nitrogen lying in the plane of the
enolate oxygen 5a, the carboxylate does occupy the face
opposite to the newly introduced amino hydroxyl group of
compound 8 (Scheme 6). In Scheme 5, a cyclic transition state
5b is also drawn as an alternative predictive tool13 with the
proposition that the electrophile is introduced to the opposite
face with respect to the alkyl group of the amino acid.
at -70 °C (Scheme 6). Conducting the amination at 0 °C with
LiCl present gave a dr of hydroxylamine 8 of 9:1, with increased
levels of impurities when compared to the reaction at -70 °C.
It was observed that the impurity profile of the amination
reaction worsened when warmed prior to quenching for the
workup. To neutralize the carboxylate nitrone intermediate 7
(Scheme 6), an equivalent of methanesulfonic acid (MSA) as
a solution in THF was added to the reaction mixture at -70
°C, which improved the purity of the reaction mixture as it was
brought to 0 °C. In addition, it was found that unreacted 6,
which was apparent by its deep-blue color, also contributed to
increased impurity levels during the workup/warm-up. Trim-
ethylphosphite was added to the reaction mixture at -70 °C as
a titration, removing the blue color of 6 on contact, and
improving the purity of the reaction mixture. This workup
sequence resulted in a crude reaction mixture containing the
nitrone carboxylic acid of intermediate 7 (Scheme 6). The crude
mixture was treated with aqueous lithium hydroxide to reform
the lithium carboxylate nitrone intermediate 7 (Scheme 6) and
extract it into the aqueous phase. The aqueous phase was
washed twice with heptane to remove impurities and subse-
Noted in the Meyers’ work is the use of LiCl to increase
the reaction rate and in some cases the diastereoselectivity.10
The work of Seebach indicates that LiCl may modify the
aggregation state of enolates, and thereby the reactivity of an
enolate in solution.14 The addition of LiCl to the amination
reaction mixture increased the dr of hydroxylamine 8 to 92:8
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
163

quently treated with aqueous citric acid and EtOAc to establish
conditions amenable to hydrolysis of the nitrone 7 (Scheme 6).
The crude hydroxylamine 8 was ultimately extracted into the
EtOAc phase. The EtOAc was replaced with IPA Via azeotropic
distillation, followed by a water addition to induce crystallization
of the major diastereomer of hydroxylamine 8 (65% overall
yield and 99:1 dr as determined by HPLC). Colorless “irregular
block” tetragonal crystals of the major diastereomer of hy-
droxylamine 8 were found to crystallize from CDCl3, and its
absolute configuration was subsequently confirmed by single-
crystal X-ray analysis (Scheme 6).15
Conclusions
An asymmetric synthesis was discovered18 and developed
to enable the preparation of 1. The synthesis consists of an
efficient sequence of fifteen chemical transformations with only
seven isolations of crystalline intermediates in an overall yield
of 27%. The quaternary chiral center of 1 was established by
diastereoselective enolate amination with the requisite D-leucine
serving as the chiral directing group. To enable this chemistry,
a safe and “scalable” stock solution form of the electrophilic
aminating reagent 6 was developed.18 This synthesis was
selected for the preparation of 1 required to support late-stage
clinical development studies.19
The hydrogenation of 8 with palladium in methanol first
removed the benzyl group, followed by reduction of the N-O
bond (Scheme 6).16 After the double reduction was complete,
the catalyst was removed by filtration, and MSA was added to
induce Fischer esterification which afforded methyl ester 10 in
91% overall yield from hydroxylamine 8. Direct alkylation of
the phenol moiety of 10 with 4-chloromethyl-2-methylquinoline
11 with cesium carbonate to give quinoline adduct 12 was not
completely selective, affording a minor byproduct due to
alkylation of the primary amine. To counter this problem, a
Schiff base (imine) was formed in situ Via condensation of 10
with p-tolualdehyde in refluxing toluene. The imine was
concentrated to an oil and redissolved in acetonitrile, followed
by the addition of potassium carbonate, catalytic tetrabutylam-
monium iodide (TBAI), and 11. The resulting reaction mixture
was heated to 70 °C for an hour to give complete consumption
of the intermediate imine, and after imine hydrolysis, the free
base of 12 was obtained as a yellow oil. The free base of 12
was dissolved in a mixture of MeOH and IPA, and 2 equiv of
MSA was slowly added to the mixture to effect crystallization
of the bis-MSA salt 12 in 84% yield from 10. The last chemical
transformation required to synthesize 1 was conversion of the
methyl ester 12 into the requisite hydroxamic acid. The
hydroxamic acid 1 was prepared in 84-87% yield by exposing
the methyl ester 12 to hydroxylamine hydrochloride in a mixture
of sodium methoxide in methanol.2e,17,18
Experimental Section
General. Melting points were taken by differential scanning
calorimetry (DSC). Accurate mass spectra were recorded on a
Micromass LCT mass spectrometer (ESI-TOF-MS). Reagents
and solvents were purchased and used as received.
(4-Benzyloxyphenyl)acetic Acid Methyl Ester (2). Under a
nitrogen atmosphere, methyl-4-hydroxyphenylacetate (1.0 kg,
6.0 mol), potassium carbonate (325 mesh powder, 1.25 kg, 9.0
mol), and acetonitrile (8 L) were charged to a reactor. Benzyl
bromide (1.08 kg, 6.3 mol) was charged to the reaction, and
the resulting mixture was warmed to 75 °C. After 3 h, HPLC
indicated >95A% conversion of the methyl-4-hydroxypheny-
lacetate to 2. The reaction was cooled to 20 °C, and the solids
were filtered off. The solids were washed with acetonitrile (3
L), and the wash was combined with the original filtrate. The
combined solution was heated to reflux and 9-10 L of distillate
was removed. Heptane (5 L) was added to the mixture, and
the distillation was continued to remove a further 5 L of
distillate. The distillation temperature reached 97 °C, indicating
that the reaction solvent had been switched from acetonitrile to
heptane. Heptane (10 L) was added to the mixture while holding
the temperature above 60 °C. The mixture was cooled to 50
°C and seeded with 10 g of (4-benzyloxyphenyl)acetic acid
methyl ester 2. The mixture was held at 50 °C for 30 min before
cooling to 3 °C over 3 h. The mixture was held for 30 min,
then filtered to obtain the product. The product was washed
with cold heptane (3 L) and dried at 45 °C under vacuum to
constant weight. The (4-benzyloxyphenyl)acetic acid methyl
ester 2 (1.46 kg; 95% isolated yield) had a purity of >99.5A%
by HPLC analysis (see reference3 for spectroscopic data on
(4-benzyloxyphenyl)acetic acid methyl ester 2).
In an attempt to shorten the synthesis, the possibility of
replacing the benzyl protection group with the required 2,4-
dimethylquinoline of 1 was explored. After investigating this
possibility, it was found that the quinoline moiety was labile to
the required N-O cleavage reaction conditions and therefore
not compatible.
2-(4-Benzyloxyphenyl)-4-oxobutyric Acid Methyl Ester (4).
Under a nitrogen atmosphere, diisopropylamine (0.35 kg, 3.5
mol) and THF (3.75 L) were charged to a reactor. The resulting
mixture was cooled to -70 °C, and 2.5 M n-butyllithium in
hexanes (1.29 L, 3.2 mol) was added to the mixture, maintaining
a reaction temperature below -20 °C. The reaction mixture
was warmed briefly to -5 °C to ensure the formation of lithium
diisopropylamide (LDA) and cooled back to -70 °C. In a
separate vessel under a nitrogen atmosphere, (4-benzyloxyphe-
nyl)acetic acid methyl ester 2 (0.75 kg, 2.9 mol) was dissolved
(13) (a) Smitrovich, J. H.; DiMichele, L.; Qu, C.; Boice, G. N.; Nelson,
T. D.; Huffman, M. A.; Murry, J. J. Org. Chem. 2004, 69, 1903. (b)
Smitrovich, J. H.; Boice, G. N.; Qu, C.; DiMichele, L.; Nelson, T. D.;
Huffman, M. A.; Murry, J.; McNamara, J.; Reider, P. J. Org. Lett.
2002, 4, 1963.
(14) (a) Seebach, D.; Bossler, H.; Gru¨ndler, H.; Shoda, S.-I. HelV. Chim.
Acta 1991, 74, 197. (b) Miller, S. A.; Griffiths, S. L.; Seebach, D.
HelV. Chim. Acta 1993, 76, 563. (c) Bossler, H. G.; Seebach, D. HelV.
Chim. Acta 1994, 77, 1124.
(15) The following crystal structure has been deposited at the Cambridge
Crystallographic Data Centre and allocated the deposition number
CCDC 237521. http://www.ccdc.cam.ac.uk.
(16) Basheerudddin, K.; Siddiqui, A. A.; Khan, N. H.; Saleha, S. Synth.
Commun. 1979, 9, 705.
(17) Thouin, E.; Lubell, W. D. Tetrahedron Lett. 2000, 41, 457.
(18) Waltermire, R. E.; Savage, S. A.; Campagna, S.; Magnus, N. A.;
Confalone, P. N.; Yates, M.; Meloni, D. J. Asymmetric synthesis of
aminopyrrolidinones. (Bristol-Myers Squibb Company, U.S.A.). PCT
Int. Appl. WO 2003104220 A1 20031218, 2003, and U.S. Patent
6770763 B2 20040803, 2004.
(19) After further optimization, a large-scale campaign was initiated to
support full development. The initial portion of the campaign delivered
450 kg of lactam 5. Two pilot-scale batches, 20 kg input of lactam 5
per batch, of the key diastereoselective amination were conducted to
afford hydroxylamine 8 with >99:1 dr in 56% yield.
164
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development

in 2.2 L of THF. The THF solution of 2 was slowly charged to
the LDA solution while holding the temperature at -70 °C.
The resulting mixture was stirred at -70 °C for 1 h while a
light-green enolate slurry formed. Allyl bromide (0.46 kg, 3.8
mol) was added to the reactor, and the resulting mixture slowly
warmed to -40 °C and stirred for 1 h. A sample was taken for
HPLC analysis which indicated >97% conversion of 2. In a
separate vessel, an aqueous solution of dihydrogen sodium
phosphate monohydrate (0.75 kg, 5.4 mol) and water (8 L) was
prepared. The aqueous phosphate solution was charged to the
reaction mixture, allowing the reaction temperature to increase
to 20 °C. EtOAc (2.2 L) was charged to the reaction mixture,
and after mixing, the layers were allowed to separate. The lower
aqueous phase was removed, and the organic layer was washed
two times with water (2 × 6 L). Using vacuum distillation at
<40 °C, the solution was concentrated until the distillation rate
slowed substantially. EtOAc (2.2 L) was charged to the mixture,
and the vacuum distillation continued. When the distillation rate
slowed substantially, distillation was ceased and a sample of
the intermediate 2-(4-benzyloxyphenyl)-pent-4-enoic acid meth-
yl ester 3 was submitted for HPLC analysis. (The area percent
of 3 was >90%. Note: if the area percent of 3 is <90%,
crystallization of 4 becomes difficult.) The crude 2-(4-benzy-
loxyphenyl)-pent-4-enoic acid methyl ester 3 was diluted with
EtOAc (7.5 L) and the resulting mixture cooled to -50 to -55
°C. Ozone was introduced to the reaction mixture below the
surface of the liquid with rapid agitation. (Caution: the
combination of oxygen and EtOAc can present a fire hazard!
To mitigate risk, the outgas stream was diluted with nitrogen
gas.) When the solution turned light-gray/blue, a sample was
submitted for HPLC analysis which indicated >99A% conver-
sion of 3. The reaction mixture was warmed to -30 °C while
purging with nitrogen to remove excess ozone from solution.
Acetic acid (0.75 L) and water (0.75 L) were added to the
reaction mixture, followed by zinc (the zinc used was ∼325
mesh, 99.9% pure). The temperature was kept between -10
and -20 °C during the addition of four portions of zinc (total
of 0.37 kg, 5.6 mol of zinc added). Note: we waited a minimum
of 20 min between additions of zinc to avoid large exotherms).
After the zinc additions were complete, the mixture was stirred
at 0 to -10 °C for 30 min. A sample was submitted for HPLC
analysis which indicated >99.8A% conversion of the ozonide.
At 0 to -10 °C, the reaction mixture was filtered through a
3-in. bed of Celite. The Celite/zinc cake was washed with
EtOAc (3.0 L), the resulting filtrates were combined and washed
two times with water (2 × 4.5 L). Aqueous sodium bicarbonate
(4 wt %) (5 L) was combined with the organic phase and the
resulting mixture stirred for 15 min to give a pH of >7. The
aqueous layer was removed, and vacuum distillation at <30 °C
was employed to reduce the volume of the organic phase by
approximately 4 L. Heptane (6 L) was charged to the mixture,
and the vacuum distillation was continued to give an ending
volume of approximately 8 L. A sample was removed for GC
analysis, which indicated <20% EtOAc by volume when
compared to a reference solvent mixture. The mixture was
cooled to 0-10 °C and stirred for 1 h; the product was filtered
off, washed with heptane (3 L), and dried under vacuum to
constant weight at 45 °C. The title aldehyde 4 (698 g; 80%
isolated yield) was afforded as a tan solid. Mp (DSC) (10 °C/
min) onset 70.68 °C, peak 84.58 °C. IR (KBr pellet) 3441, 2952,
1
2833, 2727, 1729 cm-1. H NMR (400 MHz, CDCl3) δ 9.79
(1H, s), 7.46-7.34 (5H, m), 7.22 (2H, d, J ) 8.8 Hz), 6.96
(2H, d, J ) 8.8 Hz), 5.07 (2H, s), 4.11 (1H, dd, J ) 5.0, 9.6
Hz), 3.69 (3H, s), 3.39 (1H, dd, J ) 9.6, 18.7 Hz), 2.81 (1H,
dd, J ) 5.0, 18.7 Hz). 13C NMR (100 MHz, CDCl3) δ 199.6,
173.4, 158.1, 136.7, 129.8, 128.7, 128.5, 127.9, 127.3, 115.0,
69.8, 52.2, 47.2, 43.8. HRMS (ESI) calcd for C18 H19 O4 (M+)
299.128, found 299.128.
1-((R)-2-Amino-4-methylpentanoic acid)-3-(4-benzyloxyphe-
nyl)pyrrolidin-2-one (5). Under a nitrogen atmosphere, aldehyde
4 (628 g; 2.1 mol) was dissolved in diethoxymethane (DEM)
(6 L), and D-leucine methyl ester hydrochloride (419 g; 2.31
mol) was added, followed by diisopropylethylamine (320 g;
2.48 mol). The resulting mixture was stirred at 20 °C for 30
min and treated with sodium triacetoxyborohydride (540 g; 2.55
mol), giving a temperature rise to 31 °C. After 1 h, HPLC
indicated complete consumption of the starting material. The
resulting mixture was washed with water (2 × 4 L) and heated
to reflux. Distillate (2 L) was removed, and at a pot temperature
of 82 °C the reaction was held for 14 h giving complete
lactamization by HPLC. The reaction mixture was cooled to 5
°C, and aqueous LiOH (60 g; 2.5 mol in 1 L water) and MeOH
(1.5 L) were charged to the reaction. The resulting mixture was
warmed to 20 °C, and held for 4 h. HPLC indicated >95A%
conversion to γ-lactam 5. The pH was adjusted to between 2
to 3 with 1 N HCl and EtOAc (2 L) was charged. The aqueous
layer was removed, and the organic layer was washed with
water (2 × 3 L), concentrated Via distillation to 3.5 L, and held
at 65 °C. Heptane (3 L) was slowly added while holding at 65
°C. After 1 h, the resulting slurry was slowly cooled to 20 °C
and held for 4 h. The solids were filtered and washed with 30%
EtOAc/heptane (1 L) and finally with heptane (2 × 1 L). The
product was dried to constant weight at 50 °C giving γ-lactam
5 (692 g: 86% isolated yield) with a purity of >99.5A% by
HPLC analysis as a tan powder. Mp (DSC) (10 °C/min) onset
129.52 °C, peak 131.55 °C. IR (KBr pellet) 3449, 2958, 2873,
1
2583, 1735, 1632 cm-1. Mixture of diastereomers: H NMR
(400 MHz, CDCl3) δ 9.40 (1H, bs), 7.46-7.33 (5H, m), 7.22
(1H, d, J ) 8.6 Hz), 7.17 (1H, d, J ) 8.6 Hz), 6.96 (2H, d, J
) 8.6 Hz), 5.07 (1H, s), 5.05 (1H, s), 4.92 (1H, dd, J ) 8.0,
15.6 Hz), 3.75 (1H, m), 3.62 (1H, m), 3.42, (1H, m), 2.54 (1H,
m), 2.12 (1H, m), 1.81 (2H, m), 1.55 (1H, m), 0.99 (6H, m).
13C NMR (100 MHz, CDCl3) δ 176.8, 174.7, 174.4, 157.7,
136.8, 131.9, 131.5, 129.1, 128.8, 128.4, 127.7, 127.3, 114.9,
69.8, 52.4, 47.4, 42.3, 42.0, 37.3, 36.8, 28.5, 28.3, 24.9, 23.0,
21.1, 20.9. HRMS (ESI) calcd for C23 H28 NO4 (M+) 382.201,
found 382.202.
1-Chloro-1-nitrosocyclopentane (6). Under a nitrogen at-
mosphere, cyclopentanone oxime (600 g, 6.06 mols), and
MTBE (4.0 L) were combined. The reaction vessel was marked
at the solvent line for later reference, and the agitation was
started. After obtaining a solution, the mixture was held at
20-25 °C. Safety Note: The following addition was exothermic,
and the reactor was vented to an aqueous NaOH scrubber. The
NaOH charge to the scrubber was 5 times the excess of chlorine
gas (1 equiv of the chlorine was consumed by the cyclopen-
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
165

tanone oxime which produced 1 equiv of HCl, and the chlorine
gas had a maximum excess of 0.3 equiv). Chlorine (roughly
600 g, 8.46 mols) was charged to the reaction mixture through
a glass frit. (Note: lighting was kept as subdued as possible
due to the light-sensitive product). The chlorine addition
produced a deep-blue color and was charged at a rate to keep
the temperature between 20-35 °C. Good agitation was needed
because a thick slurry was produced during the addition. The
reaction was complete when the mixture returned to homoge-
neous and a halo of green (chlorine) was observed on the
reaction surface. The reaction was sparged with nitrogen to
remove any excess chlorine and HCl. The mixture was cooled
to 0-10 °C, and 2 N NaOH (4 to 5 L) was added at a rate that
kept the temperature between 0 to 25 °C until the pH was
constant at 14 (Note: The 2 N NaOH addition produced an
exotherm). The phases were separated, and the bottom aqueous
phase was removed. The organic phase was washed twice with
water (2 × 5.0 L) at 0-25 °C. The reactor was set up for
vacuum distillation with a Dean-Stark trap in place and an
addition funnel to add MTBE. (Note: the reaction volume was
not reduced below the starting volume of roughly 4 L marked
on the vessel after the initial oxime/MTBE charge due to the
enhanced stability of the product when diluted.) An initial
volume of MTBE ((needs to be <150 ppm water) or azeotropic
distillation was futile)) was charged to the mixture equal to the
volume of the Dean-Stark trap. During the vacuum distillation,
MTBE was charged to the vessel at roughly the same rate as it
was being distilled off, between 40 to 50 °C (the distillate had
a light-blue color). The Dean-Stark trap was emptied periodi-
cally, and after 8 L was removed, the reaction was checked by
Karl Fischer titration for water (criteria for completion was <150
ppm water). If the criteria were not met, the analysis was
repeated after every 4 L of distillate was collected until the
criteria were met. It was important that the weight and volume
of the reactor contents were measured, so that a weight percent
strength of the product could be estimated. The final volume
was roughly 4.0 L and was charged to amber glass bottles,
leaving as little head space as possible. Nitrogen was blown
into the head space, and the bottles were capped and parafilmed
(stored in a refrigerator at <7 °C in the dark). A quantitative
yield was assumed.
(R)-3-Aminohydroxy-1-((R)-2-amino-4-methylpentanoic acid)-
3-(4-benzyloxyphenyl)pyrrolidin-2-one (8). Under a nitrogen
atmosphere, LiCl (30 g; 708 mmol) was charged to a vessel,
followed by THF (1.05 L). The resulting mixture was stirred
at 20 °C for 30 min at which point the mixture was almost
homogeneous. γ-Lactam 5 (45 g; 118 mmol) was added, and
the mixture cooled to -50 °C. n-Butyllithium (2.5 N; 97.2 mL;
242 mmol) in hexanes was added over 30 min, and the resulting
mixture cooled to -70 °C. To the reaction mixture an estimated
23.5 wt % solution of 6 (113.2 mL; 153 mmol) in MTBE was
added over 30 min (HPLC indicated >95A% conversion and a
13:1 diastereomer ratio), followed by a THF solution of MSA
(8 mL; 124 mmol; in 90 mL of THF) over 15 min. Trimeth-
ylphosphite (12.52 mL; 106.17 mmol) was added to the reaction
mixture, followed by warming to 0 °C over 30 min, and holding
for an additional 30 min. An aqueous lithium hydroxide solution
(4.3 g; 177 mmol; in 1.35 L of water) was added to the reaction
mixture (aqueous pH >12), followed by heptane (400 mL). After
vigorous mixing, the phases were separated, and the aqueous
phase was washed with a further 300 mL of heptane. EtOAc
(450 mL) was added to the aqueous phase, and the resulting
mixture was warmed to 30 °C with vigorous mixing. Aqueous
citric acid (10%; 250 mL) was added to the mixture until pH
3.6 was reached. After 30 min of vigorous mixing to effect
hydrolysis of the nitrone intermediate 7, the organic phase was
separated and washed with brine (500 mL water + 100 mL
saturated aqueous brine), and the solvent was exchanged for
IPA Via azeotropic distillation (final volume of 400 mL). The
reaction mass was held at 60 °C, and water (600 mL) was added
to the mixture over 2.0 h, inducing precipitation. The resulting
slurry was cooled to 20 °C over 2.0 h, filtered, and washed
with 40% IPA/water (2 × 100 mL). After drying at 70 °C to
constant weight, the product hydroxylamine 8 (32.0 g; 65%
isolated yield) was afforded as a light-yellow solid, as a single
diastereomer as determined by HPLC and NMR. The absolute
stereochemistry of hydroxylamine 8 was confirmed by single-
crystal X-ray analysis.15 Mp (DSC) (10 °C/min) onset 140.63
°C, peak 147.46 °C. IR (KBr pellet) 3378, 2968, 2878, 1692
cm-1. Single diastereomer: 1H NMR (400 MHz, CDCl3) δ 7.61
(1H, s), 7.45-7.37 (5H, m), 7.33 (2H, d, J ) 8.8 Hz), 6.94
(2H, d, J ) 8.8 Hz), 5.09 (2H, s), 4.71 (1H, dd, J ) 4.0, 11.6
Hz), 3.34-3.20 (4H, m), 2.63-2.56 (1H, m), 2.13-2.09 (1H,
m), 1.81-1.74 (1H, m), 1.62-1.49 (2H, m), 0.92 (3H, d, J )
6.6 Hz), 0.87 (3H, d, J ) 6.6 Hz). 13C NMR (100 MHz, CDCl3)
δ 174.3, 172.6, 157.6, 137.1, 132.3, 128.4, 127.9, 127.7, 127.5,
114.3, 70.5, 69.1, 51.8, 39.0, 36.3, 29.7, 24.1, 23.1, 21.0. HRMS
(ESI) calcd for C23 H29 N2O5 (M+) 413.209, found 413.209.
[R]D (25 °C) -29.0 (c 1.064, MeOH).
(R)-3-Amino-1-((R)-2-amino-4-methylpentanoic acid methyl
ester)-3-(4-hydroxyphenyl)pyrrolidin-2-one (10). Under a ni-
trogen atmosphere, hydroxylamine 8 (10 g; 24.27 mmol) was
suspended in MeOH (250 mL) and heated to 40 °C for 30 min
giving a homogeneous mixture. Palladium on carbon (6.0 g;
54% wet; 8% palladium based on dry weight; 220 mg Pd;
Degussa) was charged to the reaction, and the nitrogen was
replaced with hydrogen. The reactor pressure was increased with
hydrogen to 60 psi and returned to atmospheric pressure three
times, and finally held at 60 psi. After 3.5 h, HPLC indicated
complete conversion to the corresponding phenol/primary amine
intermediate. The reaction mixture was cooled to 25 °C, and
filtered through a bed of Celite on filter paper. The clarified
mixture was held at 25 °C, and treated with MSA (3.15 mL;
48.54 mmol). After 18 h, HPLC indicated complete conversion
to the title ester 10. The resulting mixture was neutralized with
saturated aqueous sodium bicarbonate and extracted with
dichloromethane (2 × 500 mL). After drying the organic phase
with sodium sulfate and removing the solvent in Vacuo, a white
solid was obtained. Drying the white solid at 50 °C and 25
inHg to constant weight afforded ester 10 (7.0 g; 91% isolated
yield). The product was recrystallized from EtOAc/heptane to
give colorless needles as an EtOAc solvate. Mp (DSC) (10 °C/
min) onset 62.23 °C, peak 73.61 °C. IR (KBr pellet) 3406, 3358,
3298, 3144, 2964, 2874, 1746, 1688 cm-1. 1H NMR (400 MHz,
CDCl3) δ 7.23 (2H, d, J ) 8.8 Hz), 6.62 (2H, d, J ) 8.8 Hz),
5.00-4.96 (1H, m, apparent J ) 7.6 Hz), 3.98 (3H, bs), 3.70
166
•
Vol. 14, No. 1, 2010 / Organic Process Research & Development

(3H, s), 3.38-3.29 (2H, m), 2.51-2.46 (1H, m), 2.18-2.10
(1H, m), 1.81-1.77 (2H, m, apparent J ) 7.6 Hz), 1.56-1.49
(1H, m), 0.99 (3H, d, J ) 4.5 Hz), 0.97 (3H, d, J ) 4.5 Hz).
13C NMR (100 MHz, CDCl3) δ 178.0, 171.3, 156.0, 132.5,
126.7, 115.5, 62.8, 52.4, 52.1, 39.6, 36.8, 36.6, 24.6, 23.0, 21.1.
HRMS (ESI) calcd for C17 H25 N2O4 (M+) 321.181, found
321.181. [R]D (25 °C) -6.4 (c 0.962, MeOH).
salt of quinoline adduct 12 (1.25 g; 84% isolated yield) was
isolated as a white crystalline solid. Mp (DSC) (10 °C/min)
onset 251.32 °C, peak 252.82 °C. IR (KBr pellet) 3449, 3008,
2959, 2733, 2515, 2186, 1747, 1704 cm-1. 1H NMR (400 MHz,
CDCl3) δ 9.11-9.02 (2H, m), 8.46 (1H, d, J ) 8.3 Hz), 8.25
(1H, d, J ) 8.3 Hz), 8.13 (1H, dd, J ) 7.6 Hz), 8.09 (1H, s),
7.94 (1H, dd, J ) 7.6 Hz), 7.57 (2H, d, J ) 8.6 Hz), 7.37 (2H,
d, J ) 8.6 Hz), 5.93 (2H, s), 4.78 (1H, dd, J ) 4.5, 11.6 Hz),
3.60 (3H, s), 3.52-3.47 (1H, m), 3.23-3.17 (1H, m), 2.97 (3H,
s), 2.72-2.68 (1H, m), 2.50-2.43 (1H, m), 2.35 (6 H, s),
1.88-1.80 (1H, m), 1.69-1.59 (1H, m), 1.58-1.53 (1H, m),
0.95 (3H, d, J ) 6.6 Hz), 0.92 (3H, d, J ) 6.6 Hz). 13C NMR
(100 MHz, CDCl3) δ 170.8, 170.5, 158.0, 153.1, 137.3, 134.0,
128.9, 128.4, 127.9, 125.0, 123.6, 121.0, 120.2, 115.3, 66.0,
62.1, 52.3, 52.0, 40.1, 39.8, 38.8, 36.0, 32.1, 23.7, 23.1, 21.0,
20.9. HRMS (ESI) calcd for C28 H34 N3O4 (M+) 476.254, found
476.256. [R]D (25 °C) -16.3 (c 1.228, MeOH).
(R)-3-Amino-1-((R)-2-amino-4-methylpentanoic acid methyl
ester)-3-[4-(4-methanol-2-methylquinoline)phenyl]pyrrolidin-2-
one-bis-mesylate (12). Under a nitrogen atmosphere, ester 10
(1.0 g; 3.13 mmol) was dissolved in toluene (10 mL), and
p-tolualdehyde (380 µL, 3.13 mmol) was added to the mixture.
A Dean-Stark trap (prefilled with dry toluene) was fitted to
the reactor, and the mixture heated to reflux for 1 h to remove
water. The trap was emptied, and the reaction mixture cooled
to 40 °C. The pressure was reduced to 60 mbar to remove
toluene by distillation to afford the intermediate imine (con-
firmed by HPLC and MS) as a colorless oil. At atmospheric
pressure and 20 °C, the oil was taken up in acetonitrile (10
mL), and potassium carbonate (864 mg; 6.25 mmol) was added,
followed by TBAI (115 mg; 0.31 mmol). 4-Chloromethyl-2-
methyl-quinoline 11 (605 mg; 3.15 mmol) was added to the
reaction mixture, and the temperature increased to 70 °C. After
1 h, HPLC indicated a complete conversion to quinoline adduct
12. The reaction mixture was cooled to 20 °C, and 1N HCl
(20 mL) was added (aqueous pH 1). The resulting mixture was
washed with EtOAc (2 × 10 mL), and subsequently neutralized
with saturated aqueous sodium bicarbonate. Extraction with
EtOAc (2 × 40 mL), drying with sodium sulfate, and removing
the solvent in Vacuo gave a pale-yellow oil. The oil was
dissolved in 10% MeOH/IPA (20 mL) and heated to 50 °C.
Slow addition of MSA (405 µL; 6.25 mmol) at 50 °C caused
a precipitate to form, and subsequently, the reaction was cooled
to 20 °C over 1 h. The solids were filtered, rinsed with IPA (2
× 10), and dried at 50 °C and 25 inHg for 18 h. The bis-MSA
Note: for conversions of the methyl ester 12 into hydroxamic
acid 1 see references 2e, 17, and 18.
Acknowledgment
We thank John A. Castoro for the mass spectrometry.
Charles W. Ray and Robert G. Wethman are thanked for hazard
evaluations and developing the 1-chloro-1-nitrosocyclopentane
6 stock solution preparation. William J. Marshall is thanked
for X-ray determinations. We thank Thomas G. Neiss for NMR
spectrometry. We thank our Discovery colleagues, especially
Carl Decicco, Jim Duan, and Tom Maduskuie for countless
discussions and suggestions. We also thank David Kronenthal
for his careful review and editing of the manuscript.
Received for review September 30, 2009.
OP900255K
Vol. 14, No. 1, 2010 / Organic Process Research & Development
•
167