The formation of a crystalline oxazolidin-5-one from (L)-alanine and its use as a chiral template in the practical synthesis of α-substituted alanine esters

Article abstract of DOI:10.2533/chimia.2006.566

Three different protocols to synthesize oxazolidin-5-ones have been studied with the goal to develop a method to synthesize a diastereomerically pure oxazolidin-5-one. A novel method is reported that uses a dynamic crystallization-induced asymmetric transformation to isolate a single diastereomer of an oxazolidin-5-one in 92% yield on kilogram scale. Alkylation of the oxazolidin-5-one template leads to good-to-excellent yields of N-protected α-substituted alanine esters in >98-99% ee. Schweizerische Chemische Gesellschaft.

Full text of DOI:10.2533/chimia.2006.566

Process chemistry
566
doi:10.2533/chimia.2006.566
chimiA 2006, 60, No. 9
Chimia 60 (2006) 566–573
© Schweizerische Chemische Gesellschaft
ISSN 0009–4293
The Formation of a Crystalline
Oxazolidin-5-one from (l)-Alanine and its
Use as a Chiral Template in the Practical
Synthesis of α-Substituted Alanine Esters
magnus eriksson*, elio Napolitano, Jinghua Xu, suresh Kapadia, Denis Byrne, Laurence Nummy,
Nelu Grinberg, sherri shen, heewon Lee, and Vittorio Farina
Abstract: three different protocols to synthesize oxazolidin-5-ones have been studied with the goal to develop a
method to synthesize a diastereomerically pure oxazolidin-5-one. A novel method is reported that uses a dynamic
crystallization-induced asymmetric transformation to isolate a single diastereomer of an oxazolidin-5-one in 92%
yield on kilogram scale. Alkylation of the oxazolidin-5-one template leads to good-to-excellent yields of N-pro-
tected α-substituted alanine esters in >98–99% ee.
Keywords: crystallization-induced asymmetric transformation · Diastereoselective alkylation ·
α,α’-Disubstituted amino acid · oxazolidin-5-one
Introduction
methods have been reviewed [4]. Of par-
ticular relevance to this work are methods
The subject of nonproteinogenic α-amino based on alkylation of chiral templates, and
acids has been an area of growing interest examples can be found in the work by the
in recent years. Substituted α-amino acids groups of Schöllkopf [5], Williams [6], and
have been used in research in several areas, Seebach [7] (Scheme 2).
notably peptidomimetics, protein synthe-
sis, natural products synthesis and in the
synthesis of pharmaceutically interesting
compounds [1].
The synthesis of 2 was envisioned via
the procedure oftentimes referred to as
Self-Reproduction of Chirality (SROC)
which was initially developed by Seebach
et al. [8]. Its key feature is the diastereose-
lective alkylation of stereochemically de-
fined imidazolidinones or oxazolidinones.
Thus, deprotonation at the α-position fol-
Our interest in the synthesis of α,α’-
.
disubstituted amino acids stems from a
research program on protein–protein inter-
actions [2] and more specifically on small-
molecule inhibitors of cell–cell interac-
tions between leukocyte function-assisted
antigen (LFA-1) and intercellular adhesion
molecule (ICAM-1). This research led to
the discovery of a potent LFA-1 inhibitor,
BIRT 377 (1) shown in Scheme 1 [3].
lowed by alkylation gives α,α’-disubsti-
.
tuted imidazolidinones or oxazolidinones.
Since the alkylation is highly stereospe-
cific, one is able to isolate chiral α,α’-di-
.
*Correspondence: Dr. m eriksson
Department of chemical Development
Boehringer ingelheim Pharmaceuticals inc.
P.o. Box 368
ridgefield, ct 06877, UsA
tel.: +1 203 798 4056
substituted amino acids of high enantio-
meric purity after hydrolysis/deprotection
of the imidazolidinone/oxazolidinone ring.
The key to the success of this protocol is
the ability to prepare 2-substituted imidaz-
olidinones or oxazolidinones from chiral
Retrosynthetically,1canbederivedfrom
the chiral α,α’-disubstituted amino acid 2.
.
There are a number of protocols available
for synthesis of these compounds and these
Fax: +1 203 837 4056
e-mail: eriksson@rdg.boehringer-ingelheim.com

Process chemistry
567
chimiA 2006, 60, No. 9
a 2:1 ratio of cis and trans oxazolidinones,
lower than the 5:1 ratio reported by Seebach
and Fadel [7a]. On the other hand, acylation
of 3b (R = Ph) gave 4b in 1:7.3 ratio of cis
and trans diastereomers, similar to the 1:7.5
reported by Fadel and Salaün [14].
During our studies we have found that
N-acyloxazolidinones cis-4a and trans-4b
easily undergo epimerization at C(2) when
exposed to acidic reagents such as ZnCl2 in
dichloromethane to produce a 60:40 mix-
ture of cis and trans-4a (22 h) or 4b (2 h)
when stirred at rt. The same ratios can be
obtained starting from trans-4a and cis-4b
diastereomers.Theseresultsclearlyindicate
the thermodynamically preferred composi-
tion for 4a and 4b and the ratios reported
by Seebach and Fadel [7a] and Fadel and
Salaün [14] therefore represent mixtures
obtained under non-thermodynamic condi-
tions (Scheme 3).
scheme 1. retrosynthetic analysis of Birt-377
scheme 2. chiral templates used for the synthesis of α,α’-disubstituted
amino acids
Unfortunately, this protocol turned out
to be less suitable on larger scale due to op-
erational difficulties. We found the Na-salt
of (l)-alanine to be very hygroscopic and
prone to forming a thick oil. Its conversion
to the Schiff base under heterogeneous con-
ditions in pentane or higher hydrocarbons
was plagued by long reaction times and er-
ratic results and the acylation of the imines
gave products with variable cis:trans ratios.
We therefore decided to search for alterna-
tive methods to prepare diastereomerically
pure oxazolidinones.
Results and Discussion
amino acids and to obtain these templates
as pure diastereomers in high yield. Using
the imidazolidinone approach, two routes
to BIRT 377 starting from Boc-protected
(d)-alanine were developed [9–11]. How-
ever, a route via a diastereomerically pure
oxazolidinone seemed more attractive due
to a potentially easier hydrolysis of the in-
termediate alkylated template and also the
possibility of using cheap (l)-alanine as
starting material for the synthesis of 1, as
opposed to its (d)-enantiomer, which is re-
quired in our previous approach [9]. To this
end, we have communicated two protocols
using oxazolidinones as templates for α-al-
kylation to make chiral quaternary amino
acids [12][13]. In this paper, we report fur-
ther details of our work and, more precisely,
we describe our attempts at gaining some
insight into the mechanisms of the different
methods to make oxazolidinones.
Acylation of Imines
We were initially interested in applying
the Seebach protocol for the generation of
the oxazolidinone template. The generic
Seebach protocol is shown in Table 1 and
consists of acylation of an imine salt, such
as 3, made from an aromatic or aliphatic al-
dehyde and a sodium salt of an amino acid.
Table 1 shows some examples from the lit-
erature based on (l)-alanine.
Evident from these results is the wide
variation in cis:trans ratios and the some-
times low yield obtained. Particularly in-
triguing is the switch in diastereoselectivity
of 4 in going from the alkylimine (4a, R =
t-Bu) to the aryl imine (4b, R = Ph). It is
also unclear from these results whether the
reported cis:trans ratios reflect thermody-
namic or kinetic compositions.
Cyclization Promoted by SOCl2 and
ZnCl2
In our search of alternative condensa-
tion methods to make oxazolidinones, we
attempted to use a one-step procedure by
Karady et al. [18] In this protocol, N-Cbz-
(l)-phenylalanine was condensed with 2
equiv. benzaldehyde in the presence of 1
equiv. p-TsOH to obtain a 9:1 cis:trans mix-
ture of oxazolidinones in 40% yield. This
procedure was later modified by Cheng
et al. [19] and their approach was to react
N-Cbz-(l)-phenylalanine (5) with benzal-
dehyde dimethyl acetal (6) in the presence
of BF3•Et2O to afford oxazolidinone 7
(Scheme 4). A similar protocol was used by
Schrader and Marlowe [20] using N-Cbz-
(l)-alanine to obtain the cis-oxazolidinone
in 75% yield after crystallization.
In our hands, treatment of 3a (R = t-Bu)
with benzoyl chloride at reflux gave 4a in
table 1. Data collected from the literature for cis:trans ratios of oxazolidinones
We sought alternative reaction condi-
tions for this protocol because we eventu-
ally wanted to implement this on a larger
scale and therefore wanted to avoid the use
of corrosive BF3•Et2O and flammable di-
ethyl ether as solvent. The low temperature
condition was also a concern on large scale.
Several Lewis acids were evaluated but nei-
ther gave any cyclized product. Only ZnCl2,
ZnBr2 or TiCl4 gave oxazolidinone 9 in ap-
preciable yield (Table 2).
Entry
R
t-Bu
X
Cis:Trans
5:1
Yield of 4 [%]
Ref.
[7a]
[14]
[15]
[16]
[17]
a
b
c
d
e
Ph
92
94
31
50
95
Ph
Ph
1:7.5
Ph
Bno
Ph
1:2.5
Ph
trans only
cis only
Ferrocenyl
t-Bu
During this work, we became aware of

Process chemistry
568
chimiA 2006, 60, No. 9
scheme 3. equilibration of cis- and trans-N-benzoyl oxazolidinone
scheme 5. Activation of 6 with thionyl chloride
scheme 4. Protocol used by cheng et al. [19]
a report by Micheel and Meckstroth [21]
where catalytic amounts of SOCl2 were
used to affect the cyclization of glycine and
aldehydes to oxazolidinones. In our case,
addition of 1 equiv. SOCl2 to the reaction
increased the yield to 66% (entry 7). A
limited solvent study of these conditions
showed that THF gives the highest cis:trans
scheme 6. oxazolidinone via α-chloroether/Zncl2
ratio. The reaction proceeds also in EtOAc
or CH3CN. Furthermore, oxalyl chloride
can be used in place of thionyl chloride.
However, at least equiv. each of ZnCl2 and
SOCl2 are necessary. With lower amounts,
the reaction remained incomplete. The opti-
mized reaction conditions give 9 in ca. 75%
crude yield with 15:1 to 20:1 ratio. Recrys-
tallization gives 9 in better than 50:1 cis:
trans ratio and ca. 65% yield [12].
believe that the acid chloride of N-Cbz-(l)- the appearance of any oxazolidinone (9) as
alanine could be an intermediate. However, determined by 1H-NMR. Thus, in the pres-
1H-NMR experiments in d8-THF showed a
ence of ZnCl2, 10 reacts with 8 to afford the
fast reaction, <20 min at rt, between SOCl2
and 6 in the presence of ZnCl2 to yield α-
chloro ether 10 (Scheme 5).
cyclized product 9 (presumably through 11
and 12) as shown in Scheme 6.
From our early studies of this reaction
The corresponding formation of 10 in we noted that a cis:trans ratio of 5:1 was
the absence of ZnCl2 is very slow, ca. 40% common. This ratio was also obtained if
after 18 h at room temperature. Further-
more, when ZnCl2 is added to a mixture of
N-Cbz-(l)-alanine, SOCl2 and benzalde-
hyde dimethyl acetal, 10 is observed before
pure cis-9 was treated with ZnCl2 in CH2Cl2
suggesting that 5:1 is the thermodynamic
composition. Interestingly, we found that a
3:1 cis:trans mixture of 9 could be enriched
to 8:1 by stirring with ZnCl2 in THF for 20
h at rt.
The use of at least 1 equiv. of SOCl2
with N-Cbz-(l)-alanine initially led us to
table 2. cyclization experiments with different Lewis acids
To clarify this behavior, we followed
the reaction by 1H-NMR. Surprisingly, the
trans-9 was formed selectively in the very
early stages of the reaction as identified by
the chemical shift of H(2). The selective
formation of the trans-diastereomer in the
early stages of the reaction is also observed
by HPLC sampling of the reaction mixture.
With time, however, the cis-diasteromer be-
comes the predominant species and by the
end of the reaction the ratio of cis:trans is
ca. 9:1. This indicates that trans is the ki-
netic diastereomer and the thermodynamic
composition in THF is ca. 9:1 (cis:trans).
After work-up, 9 is isolated as a 15:1 to
20:1 mixture of cis and trans diastereomers.
Apparently, there is some kind of enrich-
ment taking place after the cyclization is
completed. The exact ratio varies somewhat
between runs and seems to depend also on
the scale of the reaction and the reaction
temperature during aqueous quench. We
Entry
Lewis acid
Zncl2
Solvent
et2o
Yield [%]
Cis:Trans
1
2
3
4
5
6
7
8
30
–
5:1
Zncl2
mtBe
et2o
–
5:1
ticl4
25
35
20
30
66
76
ZnBr2(2 equiv.)
ZnBr2(1 equiv.)
Zncl2(2 equiv.)
Zncl2/socl2
ZnCl2/SOCl2
Bu2o
Bu2o
Bu2o
Bu2o
THF
5:1
5:1
5:1
5.5:1
15:1 to 20:1

Process chemistry
569
chimiA 2006, 60, No. 9
reasoned that perhaps the minor diastereo-
mer may be selectively hydrolyzed by acid
released during the aqueous quench and
that this could explain the improved ratios
observed in the crude product. Therefore, a
5:1 cis:trans mixture of 9 was dissolved in
THF and to this solution, a small amount of
ZnCl2 and SOCl2 was added to mimic actual
reaction conditions because these reagents
are used in slight excess. This mixture
was quenched with water slowly keeping
the temperature below 10 °C. Samples for
HPLC taken during the quench confirmed
that the trans diastereomer is hydrolyzed
more rapidly than the cis.
scheme 7. Novel synthesis of oxazolidinones
of pure cis-15d in ca. 82% yield.Analysis of
the mother liquors showed minor amounts
of cis and trans product in a 85:15 ratio,
along with some unreacted carbamate and
α,α-dichloromethyl biphenyl. The pres-
ence of an 85:15 ratio of cis- and trans-15d
in the mother liquor is clear evidence that
a crystallization-induced asymmetric trans-
formation operates to provide pure cis-dia-
stereomer.
nones from N-methoxy and N-i-butyloxy
carbamates and different aldehydes. From
our work on the corresponding N-amides,
we knew that epimerization at the C(2) po-
sition readily takes place upon exposure to
a variety of acidic reagents such as ZnCl2 in
CH2Cl2 or p-toluenesulfonic acid in reflux-
ing benzene. Thus, formation of a variety
of oxazolidinones with ZnCl2 in CH2Cl2
established the equilibrium composition of
cis and trans isomers and this is shown in
Table 3.
Although this protocol to make 9 was
amenable to scale-up, it suffered from the
use of more than stoichiometric amounts of
ZnCl2 and SOCl2 and also the fact that the
crude material needed recrystallization to
reach high diastereomeric purity.
A small survey of Lewis acids showed
that SnCl4 gave the highest yield of 15d
Cyclization Catalyzed by ZnCl2 or
SnCl4
The preference for the cis-diastereomer (Table 4). Using the optimized conditions
in 15 is thought to arise from A1,3-strain
[26] with the N-acyl substituent pointing
in the opposite direction from the substitu-
the oxazolidinone template was isolated in
92% yield on kilogram scale.
Our previous work on imidazolidinones
[9–11] had demonstrated the possibility of a
crystallization-driven asymmetric transfor-
mation [22] process to provide a pure dia-
stereomer from a thermodynamic mixture
of cis:trans isomers for use as a template for
subsequent alkylation. Our intention was to
take advantage of this reaction feature in the
oxazolidinone synthesis as well. We also
wanted to explore the use of acid chlorides
as substrates for the oxazolidinone forma-
tion, based on the literature precedent that
acetyl chloride reacts with aromatic alde-
hydes to form α-chlorobenzyl acetates in
the presence of Lewis acids (e.g. ZnCl2)
[23]. Our hypothesis was that the in situ
generation of an acid chloride from a car-
bamate-protected amino acid followed by
reaction with an aromatic aldehyde would
afford an oxazolidinone directly.
The structure of 15d was confirmed by
ents in the 2- and 4-positions, thus slightly single crystal X-ray analysis and is shown in
lowering the energy of the cis-isomer. In Fig. 1. The cis relationship of the 4-methyl
entry 5, the ratio of cis and trans is almost and the 2-biphenyl substituents is therefore
equal and this is likely due to the steric bulk confirmed.
of the 9-anthranyl substituent leading to a
very small energy difference between the (l)-alanine and (COCl)2 is rapid at room
two isomers.
The attempts to obtain a crystalline
oxazolidinone were successful with N-i-
The reaction of N-i-butyloxycarbonyl-
temperature and forms acid chloride 14 in
quantitative yield as determined by 1H-
NMR in CD2Cl2. No changes occur when
13C-labeled benzaldehyde (δ = 192.6 ppm
butyloxy carbamate and 4’-biphenylcar-
boxaldehyde to give 15d, Table 3, entry for 13C=O) is added to the acid chloride
4. In solution, the 85:15 ratio of cis- and (Fig. 2, spectrum a). When 10 mol% SnCl4
is added to this mixture, a major unknown
peak is observed at 83.4 ppm as well as new
peaks at 89.3 and 89.5 ppm corresponding
trans-15d was observed. However, when
the solvent was partially removed and the
residue suspended in MTBE, an off-white
solid could be filtered off which consisted to C(2) in cis-15c and trans-15c, respective-
However, acid chlorides from N-acyl
amino acids are difficult to prepare and are
known to be unstable. Buckley et al. have
shown that acid chlorides from N-amides
of amino acids undergo racemization via
azlactone formation [24]. This reaction is
much slower with N-carbamates [25] and
it is known [24] that N-ethoxycarbonyl ala-
nine can be converted to the corresponding
acid chloride using oxalyl chloride/DMF
without racemization.
Thus, treatment of N-i-butyloxycarbon-
yl-(l)-alanine (13) in CH2Cl2 with catalytic
DMF and 1 equiv. oxalyl chloride leads to
the acid chloride 14 in quantitative yield.
It can be isolated as a yellow oil or used
directly in solution. Addition of benzalde-
hyde followed by a catalytic amount of solid
ZnCl2 leads to the formation of oxazolidi-
none 15c in good yield as a thermodynamic
mixture of cis and trans diastereomers in
85:15 ratio as outlined in Scheme 7.
table 3. equilibration ratios of different oxazolidinones
Entry
R
Aldehyde
Ph
Cis:Trans
85:15
Product
15a
1
2
3
4
5
me
me
4`-biphenyl
Ph
85:15
15b
i-Bu
i-Bu
i-Bu
85:15
15c
4`-biphenyl
9-anthranyl
85:15
15d
55:45
15e
table 4. influence of catalyst on the conversion to oxazolidinone 15d
Catalyst
Zncl2
Conversion
87%
Cis:Trans
5:1
5:1
3:1
–
sncl4
94%
ticl4
<10%
trace
ti(i-Pro)4
We studied the formation of oxazolidi-

Process chemistry
570
chimiA 2006, 60, No. 9
scheme 8. synthesis of 13c-labeled α-chlorobenzyl esters
Fig. 1. single crystal X-ray structure (orteP
representation) for (l)-oxazolidinone (15d)
ly (Fig. 2, spectrum b). The peak at 83.4
ppm diminishes whereas the peaks at 89.3
and 89.5 from 15c increase over time (Fig.
2, spectra c and d).
scheme 9. Proposed mechanism for oxazolidinone formation
We propose that the unknown species
observed at 83.4 ppm corresponds to an
intermediate α-chloro benzyl ester 16 as
shown in Scheme 8 [27].
The main support for intermediate 16
comes from correlation to the 13C-NMR
study of the analogous reaction with acetyl
chloride (18) and 13C-labeled benzaldehyde
in the presence of catalytic SnCl4 to give
α-chlorobenzyl acetate (19), as laid out in
Scheme 8. The signal for the benzylic car-
bon in 19 is readily observed at 82.9 ppm.
a CH2Cl2 solution of the acid chloride in
the presence of catalytic SnCl4, a new car-
bonyl resonance at 1800 cm–1 is observed
which is attributed to the oxazolidinone
product. The benzaldehyde signal at 1702
cm–1 disappears at ca. 3 times faster rate
than the appearance of the oxazolidinone
signal suggesting the formation of an inter-
mediate which is transformed into the oxa-
zolidinone product at a slightly slower rate.
Scheme 9 shows the proposed mechanism
for oxazolidinone formation.
This is very close to the signal at 83.4 ppm
observed for the unknown intermediate
(vide supra).
Unfortunately, the isolation of 16 is
not possible since it undergoes rapid cy-
clization to the oxazolidinone. Additional
qualitative evidence for an intermediate
in the cyclization comes from ReactIR ex-
periments. When benzaldehyde is added to
Alkylation of Oxazolidinones
Alkylation of 15d is stereospecific and
takes place from the face opposite the bi-
phenyl moiety, leading via 21 to deriva-
tives of 22 in >98% ee as determined by
chiral-phase HPLC (Scheme 10). In the
procedure, LiHMDS is added at –25 °C to
a mixture of electrophile and 15d. Alter-
natively, the enolate can be generated at
–78 °C followed by addition of the elec-
trophile. The results were poorer when the
enolate was first generated at –25 °C fol-
lowed by addition of electrophile, indicat-
ing that the enolate is stable at –78 °C but
not at –25 °C.
Crude oxazolidinone 21 is then treated
with MeOLi in MeOH to make the methyl
ester 22. This liberates the 4’-biphenylcar-
boxaldehyde [28] which is isolated during
workup as its sodium bisulfite adduct via
filtration and can subsequently be regener-
ated [29] for use in the synthesis of 15d.
The scope of the alkylation is shown in
Table 5. Activated allylic-type electrophiles
generally give excellent yield of derivatives
of 22. A non-activated alkyl electrophile
Fig. 2. Formation of oxazolidinone 15a using 13c-labelled Phcho. 13c-spectra; a) before 10 mol%
sncl4. b) 2 min after 10 mol% sncl4. c) 30 min d) 60 min.

Process chemistry
571
chimiA 2006, 60, No. 9
NJ. Enantiomeric ratios were determined
on a Chiralcel OD column using hexanes/
isopropanol as eluent. Authentic samples
of the opposite enantiomer were prepared
in all cases. All chemicals were obtained
from commercial sources and used as re-
ceived unless otherwise noted. Anhydrous
solvents from Aldrich were used directly.
scheme 10. Diastereoselective synthesis of α,α’-aminoesters
table 5. Alkylation/transesterification of 15d
(S)-2-Isobutoxycarbonylamino-
propionic Acid (13c) [31]
(l)-Alanine (411 g, 4.61 mol, 1 equiv.)
was added to a 12 l four-neck round bottom
flask, fitted with mechanical stirrer, addi-
tion funnel and temperature probe. Water
(3.42 l) was added and the resulting solu-
tion cooled to 10–15 °C with an ice/water
bath. Solid NaOH (184.4 g, 4.61 mol, 1
equiv.) was added in portions to give a so-
lution with pH 11–12. The temperature rose
to ca. 20 °C with each portion of NaOH.
Dimethylaminopyridine (DMAP, 23 g,
0.189 mol, 0.041 equiv.) was added and the
reaction cooled back down to ca. 15 °C. To
this solution was added i-butylchlorofor-
mate (688 ml, 5.30 mol, 1.15 equiv.) in one
portion. The reaction was stirred vigorously
and 30% aq. NaOH was added via the ad-
dition funnel at a rate to maintain the pH
around 10–11. During the aq. NaOH ad-
dition the reaction temperature rose to 42
°C and was cooled down by addition of ice
directly into the reaction mixture. A total of
ca. 900 ml of 30% aq. NaOH was added.
Once the pH had stabilized at pH 10–11,
the reaction was cooled to room tempera-
ture and stirred overnight.
Yield of 22a
[%]
Enantiomeric
purity [% ee]
Entry
Electrophile, R-X
a
86%
>99
b
80%
93%
98.9
>99
c
d
70%
>99
e
97%
70%
>99
>99
f
The reaction solution was cooled to 2–3
°C by addition of ice directly to the flask
and the pH was then adjusted to 1–2 by ad-
dition of conc. HCl. The resulting turbid re-
action mixture was extracted with CH2Cl2
(3 × 1.5 l) and the combined organic layer
was concentrated on a rotavap and dried
further under high vacuum to afford 843 g
of a white, waxy solid (96% yield).
g
67%
90%
>99
>99
h
Anal. calc. for C8H15NO4, C; 50.78, H;
7.99; N; 7.40: Found: C; 50.92, H; 8.29, N;
7.19.
1H-NMR and 13C-NMR data correlate
well with published data [31].
a isolated yields over two steps. b DmPU (1 equiv.) was added.
(e.g. butyl bromide) gave lower yield as can tion has been demonstrated on kilogram
be expected. However, addition of 1 equiv. scale. Alkylation of the oxazolidinone fol-
DMPU to the enolate before addition of the lowed by hydrolysis leads to α-substituted
BuBr increased the isolated yield to 70% esters of (l)-alanine in >98–99 % ee.
over two steps [30].
(2S,4S)-2-Biphenyl-4-yl-4-methyl-5-
oxo-oxazolidine-3-carboxylic Acid
Isobutyl Ester (15d)
A 12 l four-neck round bottom flask,
fitted with mechanical stirrer, addition fun-
nel, reflux condenser, argon inlet and tem-
perature probe, was purged with argon for
30 min. N-i-Butyloxycarbonyl-(l)-alanine
(641 g, 3.39 mol, 1 equiv.) was added to
the flask and dissolved in CH2Cl2 (3.3 l).
Dimethylformamide (13.2 ml, 0.17 mol,
0.05 equiv.) was added, the mixture stirred
at 200 rpm and cooled to 7–8 °C in an
ice/water bath. Once at that temperature,
Experimental
Summary
General
1
In summary, three methods to synthe-
All melting points are uncorrected. H
size enantiomerically pure oxazolidinones and 13C NMR spectra were obtained on
from (l)-alanine have been evaluated and Bruker spectrometers operating at 500/400
a novel synthesis of a chiral oxazolidinone and 125/100 MHz, respectively. Elemental
using a dynamic asymmetric transforma- analysis were performed by QTI Analyses,

Process chemistry
572
chimiA 2006, 60, No. 9
oxalyl chloride (301.7 ml, 3.46 mol, 1.02 off. The filtrate layers were separated and
equiv.) was added in one portion via the the aqueous phase extracted with CH2Cl2
addition funnel. A 1–2 °C temperature rise (30 ml). The combined organic layer was
was observed along with the release of gas washed with water (50 ml), brine (50 ml)
(HCl, CO2, CO) from the reaction. The and dried (MgSO4) to afford the methyl
ice/water bath was removed and the solu- ester as a yellow oil after evaporation of
tion stirred at ambient temperature for 5 h the solvent. The methyl esters were puri-
until all gas bubbling had ceased. Biphe- fied by flash chromatography on silica gel
nylcarboxaldehyde (661 g, 3.46 mol, 1.02 using EtOAc/hexanes.
(R)-2-Isobutoxycarbonylamino-3-(3-
methoxy-phenyl)-2-methyl-propi-
onic Acid Methyl Ester (22d)
Flash chromatography on silica gel
using 10% EtOAc/hexanes gave a color-
1
less oil (70% over two steps). H-NMR
(CDCl3): 7.14–7.26 (m, 1H), 6.75–7.79
(m, 1H), 6.59–6.65 (m, 2H), 5.41 (br, 1H),
3.84–3.89 (m, 2H), 3.75 (s, 3H), 3.36 (bd,
1H), 3.16 (d, J = 14 Hz, 1H), 1.88–1.94 (m,
J = 7 Hz, 1H), 1.62 (s, 3H), 0.93 (d, J = 7
Hz, 6H). 13C-NMR (CDCl3): 174.2, 159.4,
155.1, 137.7, 129.2, 122.2, 115.7, 112.3,
70.86, 60.57, 55.06, 52.62, 41.76, 27.98,
23.71, 19.03. LCMS: m/z = 324.3 (M+H)+.
Anal. calc. for C17H25NO5, C; 63.14, H;
7.79; N; 4.33: Found, C; 63.12, H; 7.76, N;
4.29. Enantiomeric purity; >99% ee.
equiv.) was added as a solid in one por-
tion to afford an orange solution. Within 5 (R)-2-Isobutoxycarbonylamino-2-
min, a solution of SnCl4 (170 ml, 1.0 M in methyl-pent-4-enoic Acid Methyl
CH2Cl2, 0.05 equiv.) was added, leading Ester (22a)
to a red solution and a rise in temperature
Flash chromatography on silica gel us-
from 13 to 26 °C over 5–10 min. The tem- ing 15% EtOAc/hexanes gave a light yel-
1
perature stabilized and the solution was low oil (86% over two steps). H-NMR
stirred at 25 °C for 20 h with a slow argon (DMSO-d6): 7.48 (br s, 1H), 5.65–5.74 (m,
purge to remove HCl gas formed during J = 7.5, 9.4, 18.0 Hz, 1H), 5.08 (dd, J =
the reaction. The reaction mixture was 9.4,18.0, 2H), 3.68–3.76 (m, 2H), 3.59 (s,
then concentrated under reduced pressure 3H), 2.59 (dd, J = 7.0, 13.5 Hz, 1H), 2.39
until ca. 1 l CH2Cl2 remained to afford a (dd, J = 7.0, 13.5 Hz, 1H), 1.80–1.84 (m,
suspension. MTBE (1.9 l) was added and 1H), 1.30 (s, 3H), 0.88 (d, J = 6.5 Hz, 6H).
the suspension was stirred at ambient tem- 13C-NMR (DMSO-d6): 173.04, 154.32,
perature for ca. 15 h. The thick suspen- 131.90, 117.87, 68.72, 57.27, 50.91, 40.04,
sion was cooled to 2–3 °C with ice/water 26.77, 21.36, 17.92. LCMS: m/z = 244.1
bath, stirred at that temperature for 2.5 h (M+H)+. Anal. calc. for C12H21NO4, C;
and then filtered on a medium frit Büch- 59.24, H; 8.70; N; 5.76: Found: C; 59.19,
ner funnel to afford an off-white solid. H; 8.73, N; 5.72. Enantiomeric purity;
The solid was washed with a 1 l portion >99 % ee.
(R)-3-(4-Bromo-phenyl)-2-isobu-
toxycarbonylamino-2-methyl-propi-
onic Acid Methyl Ester (22e)
Flashchromatographyonsilicagelusing
20% EtOAc/hexanes gave 1.44 g (97% over
1
two steps) slightly greenish oil. H-NMR
(CDCl3): 7.37 (d, J = 8.3 Hz, 2H), 6.92 (d,
J = 8.3 Hz, 2H), 5.35 (br, 1H), 3.89–3.92
(m, 1H), 3.82–3.86 (m, 1H), 3.75 (s, 3H),
3.40 (bd, J = 13 Hz, 1H), 3.15 (d, J = 13
Hz, 1H), 1.88–1.97 (m, 1H), 1.60 (s, 3H),
0.93 (d, J = 7 Hz, 6H). 13C-NMR (CDCl3):
171.84, 152.94, 133.24, 129.48, 129.23,
118.90, 68.85, 58.42, 50.57, 38.85, 25.93,
21.65, 16.92. LCMS: m/z = 372.8 (M+H)+.
Anal. calc. for C16H22BrNO4, C; 51.62, H;
5.96; N; 3.76: Found, C; 51.57, H; 5.88, N;
3.70. Enantiomeric purity; >99% ee.
of cold MTBE and dried on the frit under
vacuum for 3 h to give 1.10 kg (92%) oxa- (E)-(R)-2-Isobutoxycarbonylamino-
zolidinone product.
2-methyl-5-phenyl-pent-4-enoic
1H-NMR, 13C-NMR and combustion Acid Methyl Ester (22b)
analysis data correlate well with those pub-
lished [13].
Flash chromatography on silica gel us-
ing 15% EtOAc/hexanes gave a colorless oil
1
(80% over two steps). H-NMR (CDCl3):
General Procedure for Alkylation of
15d
7.22–7.34 (m, 5H), 6.45 (d, J = 16 Hz, 1H),
6.04 (dt, J = 7, 16 Hz, 1H), 3.84 (m, 2H),
3.77 (s, 3H), 2.95 (br, 1H), 2.72–2.78 (dd,
J = 7, 14 Hz, 1H), 1.90 (m, J = 7 Hz, 1H),
1.62 (s, 3H), 0.91 (d, J = 7 Hz, 6H). 13C-
NMR (CDCl3): 174.3, 136.9, 134.3, 128.5,
127.5, 126.2, 123.6, 70.84, 59.74, 52.69,
40.47, 28.00, 23.35, 19.02. LCMS: m/z =
320.3 (M+H)+. Anal. calc. for C18H25NO4,
C; 67.69, H; 7.89; N; 4.39: Found: C; 67.60,
H; 7.90, N; 4.28. Enantiomeric purity;
98.9 % ee.
(R)-2-Isobutoxycarbonylamino-2-
methyl-hexanoic Acid Methyl Ester
(22f)
The oxazolidinone template 15d (4
mmol) and the corresponding alkyl halide
(4.2 mmol) were suspended in THF (12
ml) under argon in a dry 50 ml flask. The
mixture was cooled to –25 °C with a dry
ice/acetone bath. LiHMDS (1.0 M in THF;
4.2 ml) was added slowly over 20 minutes
via syringe. The mixture was stirred at –25
°C for 30 min and at room temperature for
3 h to afford an orange solution. Saturated
NH4Cl (15 ml) was added and the mixture
was diluted with EtOAc (50 ml). The layers
were separated and the aqueous layer was
extracted with EtOAc (20 ml). The com-
bined organic layer was washed with brine
(30 ml) and dried over MgSO4. Evaporation
of the solvent afforded the crude alkylation
product as an oil.
The crude alkylation product (~4
mmol) was dissolved in dry MeOH (16
ml) and MeOLi (4.4 ml, 1.0 M in MeOH)
was added. The solution stirred at room
temperature until TLC indicated disap-
pearance of starting material (1–3 h). A
mixture of saturated aqueous NaHSO3
(30 ml) and CH2Cl2 (40 ml) was added.
A white precipitate formed and the mix-
ture was stirred at room temperature for
ca. 1 h after which the solid was filtered
The oxazolidinone template (4 mmol)
was suspended in THF (12 ml) under ar-
gon in a dry 50 ml flask. The mixture was
cooled to –35 °C and LiHMDS (1.0 M in
THF; 4.2 ml, 1.05 equiv.) was added slowly
over 5 min via syringe to afford a yellow
solution that was stirred for 30 min at –35
°C. DMPU (4 mmol, 1 equiv.) was added
and the solution stirred another 30 min. n-
BuBr (4.2 mmol, 1.05 equiv.) was added
and the solution was allowed to reach rt
over 4 h. See general experimental proce-
dure for work-up and hydrolysis/esterifica-
tion. Flash chromatography on silica gel us-
ing 10% EtOAc/hexanes gave 0.73 g light
yellow oil (70% over two steps). 1H-NMR
(CDCl3): 5.45 (br, 1H), 3.81(d, J = 6.6 Hz,
2H), 3.75 (s, 3H), 2.09 (br, 1H), 1.88 (m, J
= 6.7 Hz, 1H), 1.70–1.81 (m, 1H), 1.56 (s,
3H), 1.17–1.34 (m, 3H), 0.98–1.12 (m, 1H),
0.91 (d, J = 6.7 Hz, 6H), 0.87 (t, J = 6.9 Hz,
3H). 13C-NMR (CDCl3): 175.17, 155.25,
70.90, 60.05, 52.76, 37.02, 28.20, 26.40,
23.57, 22.76, 19.21, 14.09. LCMS: m/z =
260.2 (M+H)+. Anal. calc. for C13H25NO4,
C; 60.21, H; 9.72; N; 5.40: Found: C; 60.00,
(R)-2-Isobutoxycarbonylamino-2-
methyl-3-phenyl-propionic Acid
Methyl Ester (22c)
Flash chromatography on silica gel us-
ing 10% EtOAc/hexanes gave a light yel-
1
low oil (93% over two steps). H-NMR
(CDCl3): 7.22–7.26 (m, 3H), 7.03–7.06 (m,
2H), 5.36 (br, 1H), 3.84–3.93 (m, 2H), 3.73
(s, 3H), 3.93 (bd, 1H), 3.18 (d, J = 14 Hz,
1H), 1.89–1.96 (m, J = 7 Hz, 1H), 1.61 (s,
3H), 0.92–0.94 (d, J = 7 Hz, 6H). 13C-NMR
(CDCl3): 174.18, 155.12, 136.21, 129.92,
128.22, 126.92, 70.83, 60.60, 52.56, 41.74,
28.01, 23.63, 19.02. LCMS: m/z = 294.1
(M+H)+. Anal. calc. for C16H23NO4, C;
65.51, H; 7.70; N; 4.77: Found, C; 65.52,
H; 7.84, N; 4.72. Enantiomeric purity;
>99 % ee.

Process chemistry
573
chimiA 2006, 60, No. 9
[1] a) R.M. Williams, ‘Synthesis of Optically
Active α-Amino Acids’, Organic Chem-
istry Series, Vol. 7, Eds. J.E. Baldwin, P.D.
Magnus, Pergamon Press, Oxford 1989;
b) H. Heimgartner, Angew. Chem. Int. Ed.
Engl. 1991, 30, 238–264.
[2] K. Last-Barney, W. Davidson, M. Cardo-
zo, L.L. Frye, C.A. Grygon, J.L. Hopkins,
D.D. Jeanfavre, S. Pav, C. Qian, J.M. Ste-
venson, L. Tong, R. Zindell, T.A. Kelly, J.
Am. Chem. Soc. 2001, 123, 5643–5650.
[3] T.A. Kelly, D.D. Jeanfavre, D.W. McNeil,
J. Woska, P.L. Reilly, E.A. Mainolfi, K.M.
Kishimoto, G.H. Nabozny, R. Zinter, B.J.
Bormann, R.J. Rothlein, Immunol. 1999,
163, 5173.
Farina, Tetrahedron Asymmetry 2003, 14,
3495–3501.
[12] S.R. Kapadia, D.M. Spero, M. Eriksson,
J. Org. Chem. 2001, 66, 1903–1905.
[13] E. Napolitano, V. Farina, Tetrahedron Lett.
2001, 42, 3231–3234.
[14] A. Fadel, J. Salaün, Tetrahedron Lett.
1987, 28, 2243–2246.
[15] E. Altmann, K. Nebel, M. Mutter, Helv.
Chim. Acta 1991, 74, 800–806.
[16] K. Nebel, M. Mutter, Tetrahedron 1988,
44, 4793–4796.
[17] A. Alonso, S.G. Davis, A.S. Elend, J.L.
Haggit, J. Chem. Soc., Perkin Trans. 1
1998, 257–264.
[18] S. Karady, J.S. Amato, L.M. Weinstock,
Tetrahedron Lett. 1984, 25, 4337–4340.
[19] H. Cheng, P. Keitz, J.B. Jones, J. Org.
Chem. 1994, 59, 7671–7676.
[20] W.D. Shrader, C.K. Marlowe, Bioorg.&
Med. Chem. Lett. 1995, 5, 2207–2210.
[21] F. Micheel, W. Meckstroth, Chem. Ber.
1959, 92, 1675–1679.
H; 9.83, N; 5.17. Enantiomeric purity;
>99% ee.
(R)-2-Isobutoxycarbonylamino-2-
methyl-3-(4-pyrimidin-5-yl-phenyl)-
propionic Acid Methyl Ester (22g)
The alkylation reaction was stirred at
–25 °C for 16 h before warming to rt. Flash
chromatography on silica gel using 35%
EtOAc/hexanes gave a slightly greenish
solid (67% over two steps). 1H-NMR (CD-
Cl3): 9.11 (s, 1H), 8.86 (s, 2H), 7.42 (d, J =
8.3 Hz, 2H), 7.17 (d, J = 8.3 Hz, 2H), 5.55
(br, 1H), 3.87–3.91 (m, 1H), 3.80–3.84 (m,
1H), 3.75 (s, 3H), 3.47 (bd, J = 12 Hz, 1H),
3.25 (d, J = 13 Hz, 1H), 1.85–1.94 (m, 1H),
1.60 (s, 3H), 0.90 (d, J = 7 Hz, 6H). 13C-
NMR (CDCl3): 173.69, 157.04, 154.79,
154.36, 137.26, 133.55, 132.47, 130.70,
126.36, 70.58, 60.23, 52.33, 40.83, 27.70,
23.46, 18.67. LCMS: m/z = 372.8 (M+H)+.
Anal. calc. for C20H25N3O4, C; 64.67, H;
6.78; N; 11.31: Found: C; 64.49, H; 6.62,
N; 11.18. Enantiomeric purity; >99% ee.
[4] a) C. Cativiela, M.D. Diaz-de-Villegas,
Tetrahedron Asymmetry 1998, 9, 3517–
3599; b) R.O. Duthaler, Tetrahedron 1994,
50, 1539–1650; c) T. Wirth, Angew. Chem.
Int. Ed. Engl. 1997, 36, 225–227.
[5] a) U. Schöllkopf, W. Hartwig, U. Groth, K.-
O. Westphalen, Liebigs Ann. Chem. 1981,
696; b) U. Schöllkopf, U. Groth, K.-O.
Westphalen, C. Deng, Synthesis 1981, 969–
971. c) U. Schöllkopf, Tetrahedron 1983,
39, 2085–2091 and references cited therein.
See also J. Chen, S.P. Corbin, N.J. Holman,
Org. Process Res. Dev. 2005, 9, 185–187.
[6] a) R.M. Williams, M.-N. Im, J. Am. Chem.
Soc. 1991, 113, 9276–9286; b) R.M. Wil-
liams, M.-N. Im, Tetrahedron Lett. 1988,
29, 6075–6078; c) P.J. Sinclair, D. Zhai,
J. Reibenspies, R.M. Williams, J. Am.
Chem. Soc. 1986, 108, 1103–1104; d)
R.M. Williams, in ‘Advances inAsymmet-
ric Synthesis’, Ed. A. Hassner, JAI Press,
Greenwich, CT, 1995;Vol 1, pp 45–94 and
references cited therein.
[7] a) D. Seebach, A. Fadel, Helv. Chim.
Acta 1985, 68, 1243–1250; b) R. Naef,
D. Seebach, Helv. Chim. Acta 1985, 68,
135–143.
[8] D. Seebach, A.R. Sting, M. Hoffmann,
Angew. Chem. Int. Ed. Engl. 1996, 35,
2708–2748 and references cited therein.
[9] N. Yee, Org. Lett. 2000, 2, 2781–2783.
[10] R.P. Frutos, S. Stehle, L. Nummy, N. Yee,
Tetrahedron Asymmetry 2001, 12, 101–
104.
[22] N.G. Anderson, Org. Process Res. Dev.
2005, 9, 800–813 and references therein.
[23] M. Neuenschwander, P. Bigler, K. Chris-
ten, R. Isell, R. Kyburz, H. Mühle, Helv.
Chim. Acta 1978, 61, 2047–2058.
[24] T.F. Buckley III, H. Rapoport, J. Am.
Chem. Soc. 1981, 103, 6157–6160.
[25] J.H. Jones, M.J. Witty, J. Chem. Soc., Per-
kin Trans. 1 1979, 3203–3206.
(R)-2-Isobutoxycarbonylamino-2-
methyl-3-(4-trifluoromethoxy-phe-
nyl)-propionic Acid Methyl Ester
(22h)
Flash chromatography on silica gel us-
ing 10% EtOAc/hexanes gave 1.36 g (90%
1
over two steps) slightly greenish oil. H-
[26] R.W. Hoffmann, Chem. Rev. 1989, 89,
1841–1860.
NMR (DMSO-d6): 7.44 (s, 1H), 7.27 (d, J
= 7.6 Hz, 2H), 7.21 (d, J = 8.1 Hz, 2H),
3.82 (m, 1H), 3.74 (m, 1H), 3.61 (s, 3H),
3.31 (d, J = 13 Hz, 1H), 3.00 (d, J = 13.9
Hz, 1H), 1.86 (m, 1H), 1.21 (s, 3H), 0.90
(d, J = 6.7 Hz, 6H). 13C-NMR (DMSO-d6):
175.62, 156.93, 148.84, 137.45, 133.81,
121.96, 121.70 (JC-F = 254 Hz), 71.38,
60.29, 53.54, 41.54, 29.26, 23.92, 20.39.
LCMS: m/z = 378.3 (M+H)+. Anal. calc.
for C17H22F3NO5, C; 54.11, H; 5.88; N;
3.71: Found: C; 54.25, H; 6.12, N; 3.70.
Enantiomeric purity; >99% ee.
[27] Only one peak is observed although 16
should be present as two diastereomers.
It is possible that the two signals overlap
given the closeness of the signals from
C(2) in cis- and trans-15c.
[28] This material is commercially available
on metric ton scale for less than $50/kg.
In comparison, the price for research
amounts is about $10/g.
[29] D.P. Kjell, B.J. Slattery, M.J. Semo, J.
Org. Chem. 1999, 64, 5722–5724.
[30] Only 33% yield over two steps was ob-
tained in the absence of DMPU.
Acknowledgements
[31] R.A. Aitken, N. Karodia, T. Massil, R.J.
Young, J. Chem. Soc., Perkin Trans. 1
2002, 533–541.
Assistance with the single-crystal X-ray
determination from Dr. Andre White is much
appreciated.
[11] N.Yee, L.J. Nummy, R.P. Frutos, J.J. Song,
E. Napolitano, D.P. Byrne, P.-J. Jones, V.
Received: June 20, 2006