Enantioselective acylation of alcohols with fluorinated β-phenyl-β-lactams in the presence of Burkholderia cepacia lipase

Article abstract of DOI:10.1016/j.tetasy.2007.06.033

This paper concentrates on studies of the acylation of alcohols with 3,3-difluoro-4-phenylazetidin-2-one rac-1, trans-3-fluoro-4-phenylazetidin-2-one rac-2 and 4-phenylazetidin-2-one rac-3 in the presence of immobilized lipase PS from Burkholderia cepacia in dry tert-butyl methyl ether (TBME). Fluorine activation in the compounds studied was essential in order to split the β-lactam ring with lipase PS. The highly enantioselective ring opening of rac-1 and rac-2 with methanol (1-butanol was also studied) allowed the preparation of the (R/(3R,4R))-β-lactams as the unreacted enantiomers and (S/(2S,3S))-β-amino esters as the product enantiomers with an ee >99%. Under the same conditions, rac-3 was totally unreactive. The possibility for a competing hydrolysis caused by water in the enzyme preparations is also discussed.

Full text of DOI:10.1016/j.tetasy.2007.06.033

Tetrahedron: Asymmetry 18 (2007) 1567–1573
Enantioselective acylation of alcohols with fluorinated b-phenyl-
b-lactams in the presence of Burkholderia cepacia lipase
Xiang-Guo Li, Maria Lahitie, Mari Paivio and Liisa T. Kanerva*
¨
¨
¨
Department of Pharmacology, Drug Development and Therapeutics/Laboratory of Synthetic Drug Chemistry
and Department of Chemistry, University of Turku, Lemminkaisenkatu 5C, FIN-20520 Turku, Finland
Received 29 May 2007; accepted 28 June 2007
Abstract—This paper concentrates on studies of the acylation of alcohols with 3,3-difluoro-4-phenylazetidin-2-one rac-1, trans-3-fluoro-
4-phenylazetidin-2-one rac-2 and 4-phenylazetidin-2-one rac-3 in the presence of immobilized lipase PS from Burkholderia cepacia in dry
tert-butyl methyl ether (TBME). Fluorine activation in the compounds studied was essential in order to split the b-lactam ring with lipase
PS. The highly enantioselective ring opening of rac-1 and rac-2 with methanol (1-butanol was also studied) allowed the preparation of the
(R/(3R,4R))-b-lactams as the unreacted enantiomers and (S/(2S,3S))-b-amino esters as the product enantiomers with an ee >99%. Under
the same conditions, rac-3 was totally unreactive. The possibility for a competing hydrolysis caused by water in the enzyme preparations
is also discussed.
ꢀ 2007 Elsevier Ltd. All rights reserved.
1. Introduction
and their precursors have previously been reviewed.4 The
results clearly indicate that 2,2,2-trifluoroethyl butanoate
is a typical achiral acyl donor and Candida antarctica lipase
A (CAL-A), a catalyst for the asymmetric N-acylation of
b-amino esters. On the other hand, 2,2,2-trifluoroethyl
esters of b-amino acids were previously prepared by
making use of C. antarctica lipase B (CAL-B)-catalyzed
interesterification (two ester substrates give products with
exchanged alkyl counterparts) between a b-amino ester
and a 2,2,2-trifluoroethyl ester.5 Accordingly, lipases will
accept alkyl-fluorinated carboxylic esters as substrates.
Enzymatic kinetic resolutions of b-amino acid derivatives
and precursors containing fluorine in the acid moiety are
rarely investigated. The studies seem to be restricted to
the hydrolysis of N-phenylacetyl derivatives of b-fluoro-
alkyl- and b-fluoroaryl-b-amino acids by penicilline acyl-
ase6–8 and to the hydrolysis of b-fluorophenyl-b-alanine
ethyl ester by Burkholderia cepacia lipase (lipase PS).9
Natural organic compounds relatively rarely contain fluo-
rine. On the other hand, fluorine has gained an important
position as an isostere for hydrogen, especially in lead
modifications to potent medicinal agents. This is based
on various desirable effects caused by the presence of fluo-
rine on the physiological activity and metabolism of com-
pounds. Fluorine can also affect the binding of the
substrate at the active site of the enzyme, for instance, by
accepting hydrogen bonds from the protein or from the
active site water. As a consequence, half of the top ten
drugs sold in 2005 were fluorinated organic molecules,
while 20% of the pharmaceuticals on the market were
estimated to contain fluorine.1 The a,a-difluoro-b-amino
acid residue of (S)-4 (Scheme 1) in the side chain of
dodetaxel analogues is a good example of an important
fluorine-containing b-amino acid.2
The enantiomers of various b-amino acids and their pre-
cursors and intermediates have widespread applications
as building blocks and intermediates in the design of
medicinally important compounds.3 Results for the enzy-
matic kinetic resolution of nonfluorinated b-amino esters
b-Lactams (2-azetidinones) are interesting precursors of
b-amino acids, while the lipases are highly usable and
economical catalysts in organic chemistry. The enantio-
selective hydrolyses of many b-lactams with CAL-B
(Novozym 435 preparation) were previously reported with
1 equiv of water in diisopropyl ether (DIPE) at an elevated
temperature.10,11 Enantioselective alcoholysis of rac-4-
phenyl-2-azetidinone with CAL-B and 2-octanol was also
reported, although the ester product was not detected.12
*
Corresponding author. Tel.: +358 2 333 6773; fax: +358 2 333 7955;
e-mail: lkanerva@utu.fi
0957-4166/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetasy.2007.06.033

1568
X.-G. Li et al. / Tetrahedron: Asymmetry 18 (2007) 1567–1573
R
R1
R
R1
R
R
R1
CO2Me
NH2
O
O
O
R1
3
4
2
OMe
iii
ii
NH
N
HN
+
Ph
rac-1, R=R1=F
rac-2, R=H, R1=F
rac-3, R=R1=H
rac-1a, R=R1=F
(S)-4, R=R1=F
(R)-1, R=R1=F
rac-2a, R=H, R1=F
(2S,3S)-5, R=H, R1=F
(3R,4R)-2, R=H, R1=F
i
iv
R
R
OMe
O
R1
CO2H
NH2
R1
HN
+
N
H
(S)-6, R=R1=F
(R)-1, R=R1=F
(2S,3S)-7, R=H, R1=F
(R)-8, R=R1=H
(3R,4R)-2, R=H, R1=F
(S)-3, R=R1=H
Scheme 1. Chemoenzymatic routes to the observed enantiomeric products of the work. Reagents: (i) BrCF2CO2Et/BrCHFCO2Et, Zn, THF; (ii)
Ce(NH4)2(NO3)6 (CAN), CH3CN, H2O; (iii) CH3OH, lipase PS; (iv) H2O, lipase PS-D or CAL-B.10
In our previous studies of the reaction of 4-benzyl-2-azetid-
inone in the presence of methanol and CAL-B in dry
organic solvents, two important observations were evident:
(1) the added methanol and the water in the enzyme prep-
aration served as competitive nucleophiles for ring open-
ing; (2) and the methyl ester produced by the action of
methanol was further enzymatically hydrolyzed, the enan-
tiopreference being opposite to that observed for the
hydrolysis of the b-lactam ring.13 Moreover, there was
experimental evidence that the lipase PS preparation (lipase
adsorbed on Celite in the presence of sucrose)14 in dry
organic solvents exposed the b-lactam to hydrolysis less
than it did in the CAL-B preparation. This is in accordance
with the observation that the lipophilic polymethacrylate
carrier of Novozym 435 easily releases the attached water,
giving hydrolysis of hydrolyzable compounds.15 Sih et al.
were the first able to open a b-lactam ring with methanol
in tert-butyl methyl ether (TBME) in the presence of lipase
PS.16 In their study, the N-benzoyl activation of the b-lac-
tam was supposed to make the ring susceptible to nucleo-
philic attack.
bon of the lactam ring (Scheme 2).4 During the following
second addition–elimination process, the obtained ester
intermediate reacts with a nucleophile, and leads to the for-
mation of the corresponding final product (an acid with
water and an ester with an alcohol).
F
F
F
O
a
F
H2N
O
Enz
+ HO-Enz
NH
O
1
(S)-
acyl-enzyme
intermediate
O
b
c
Enz
+
ROH
PrCO2R + HO-Enz
Pr
O
acyl-enzyme
intermediate
+H2O
-H2O
4
(S)- + HO-Enz
PrCO2H + HO-Enz
Herein, the ability of lipases to enantioselectively acylate
alcohols to afford b-amino esters with fluorinated 4-phen-
yl-2-azetidinones rac-1 and rac-2 as well as nonfluorinated
rac-3 under mild conditions is considered (Scheme 1, route
iii). The enzymatic preparation of b-amino esters directly
from b-lactams offers an attractive alternative to the ring
opening with alcohols by chemical means.3b More specifi-
cally, we have studied the importance of fluorine activation
on the ring opening and kinetic resolution of our substrates
by methanol and 1-butanol and the possibility to suppress
the competing hydrolysis route iv by using mainly lipase PS
adsorbed on Celite as a catalyst in TBME. The critical step
for the ring opening by lipase catalysis is the formation of
the so called acyl-enzyme intermediate when the active site
serine hydroxyl reacts as a nucleophile to the carbonyl car-
Scheme 2. Ring opening by the alcohol generated in situ from an PrCO2R
(R = Me or Bu, Enz = lipase). The water is from the enzyme preparation.
The synthesis of fluorinated b-amino acids has been
reviewed.3,17 Compound (S)-6 was previously prepared,
based on various Reformatsky-type reactions in the pres-
ence of chiral auxiliaries.18,19 On the other hand, the key
step to (2S,3S)-7 was the tandem conjugate addition of
lithium (S)-N-benzyl-N-a-methylbenzylamide to tert-butyl
cinnamate.20 Based on these articles, the absolute configu-
rations of the present resolution products are known (see
Section 4.5) and are shown in Scheme 1. The main benefit
of the lipase-catalyzed kinetic resolution pathways over the
published synthetic routes to enantiopure compounds18–20

X.-G. Li et al. / Tetrahedron: Asymmetry 18 (2007) 1567–1573
1569
is that chiral induction now is catalytic rather than stoichio-
metric, meaning there is no need to prepare enantiopure
auxiliaries, and mild reaction conditions are possible.
2.2. Enantioselective ring opening of rac-1–3 by alcoholysis
For optimization, rac-1 was used as a model substrate.
Free lipase PS and various immobilized lipase PS, CAL-
B and CAL-A preparations were screened for the alcohol-
ysis of rac-1 with methanol (5 equiv) in dry TBME (Table
1; Scheme 1, route iii). CAL-A did not catalyze the reaction
(entry 6) while all lipase PS preparations and CAL-B gave
highly enantioselective reactions, which were not contami-
nated by hydrolysis (entries 1–5) on the basis of stoichio-
metry [the sum of the amounts of (R)-1 and (S)-4 equals
to the initial amount of rac-1]. CAL-B and lipase PS
preparations, other than unimmobilized lipase PS powder
(entry 3), gave good reactivities as measured with the con-
version reached after 7 h. Due to the well-known tendency
of CAL-B to hydrolyze b-lactams,10–13 lipase PS adsorbed
on Celite in the presence of sucrose (called lipase PS prep-
aration from this point on)14 was the choice for further
optimizations. Later, when commercial lipase PS-D (lipase
PS also on Celite) from Amano was available, we decided
to replace our own preparation as the results were very sim-
ilar. In practise, lipase PS-D was used for the studies with
rac-2 and rac-3 and in all preparative-scale reactions.
2. Results and discussion
2.1. Synthesis of racemic b-lactams
N-Protected rac-1a and rac-2a were prepared with reason-
able chemical yields (81% and 49%, respectively) via a
Reformatsky addition, where the Schiff base between p-
methoxylbenzylimine and benzaldehyde reacted with ethyl
bromodifluoroacetate and bromofluoroacetate (2 equiv),
respectively, in the presence of Zn dust as previously
described for rac-1a (Scheme 1, i).18 The following oxidative
deprotection furnished rac-1 with 80% and rac-2 as the
trans-diastereomer with 57% yield (route ii). In addition
to the trans-product, the synthesis of rac-2a produced
two unidentified products on the basis of TLC. The
trans-stereochemistry in rac-2 is based upon the lack of
NOESY correlation between the protons on carbon atoms
3 and 4. As further evidence for the trans-assignment, the
specific rotation and the coupling constants of (2S,3S)-7
(obtained through hydrolysis of enzymatic acylation prod-
uct (2S,3S)-5, see Section 4.5) were closely identical with
the published data.20 rac-3 was obtained by the 1,2-dipolar
cycloaddition of chlorosulfonyl isocyanate to styrene as
previously described.10
The effect of methanol concentration (0.1–0.5 M) was stud-
ied for the alcoholysis of rac-1 (0.05 M) in TBME in the
presence of lipase PS preparation (Table 2; Scheme 1, route
iii). Enantioselectivity (measured by the enantiomer ratio,
E) was excellent in each case, and accordingly the reaction
was at (entries 1, 2 and 4) or was approaching (entry 7)
50% conversion after 24 h, producing enantiopure resolu-
tion products (R)-1 and (S)-4. Enzymatic reactivity clearly
decreased with increasing methanol concentration, as can
be seen from the conversions between 41% and 28% after
the reaction of 7 h (entries 1, 2, 4 and 7). Elevated temper-
atures (entry 3 compared to 2) and higher enzyme contents
(entry 5 compared to 4) enhanced reactivity. The ring open-
ing was somewhat faster in DIPE (entry 6) than in TBME
(entry 4). For further studies TBME was preferred because
it does not form peroxides. 1-Butanol (entry 8), when used
as a nucleophile, was practically as effective as methanol
(entry 4). Although hydrolysis was not a side reaction in
the alcoholysis of rac-1 with methanol or 1-butanol, the
lipase PS preparation gave surprisingly fast and highly
enantioselective hydrolysis when no alcohol was added
Table 1. Lipase screening for the alcoholysis of rac-1 (0.05 M) by lipase
preparations (30 mg/mL) with methanol (0.25 M) in TBME at room
temperature
Entry Lipase
Time Conversion ee(R)-1 ee(S)-4
E
(h)
(%)
(%)
(%)
1
2
3
4
5
6
Lipase PSa
7
7
7
7
7
33
43
1
46
33
0
49
76
1
86
48
—
>99
>99
>99
>99
>99
—
>200
>200
>200
>200
>200
—
Lipase PS-D
Lipase PSb
Lipase PS-C II
CAL-B
CAL-Aa
24
a Lipase preparation on Celite containing 20% (w/w) of the commercial
enzyme powder.
b Free lipase PS powder.
Table 2. Alcoholysis of rac-1 (0.05 M) catalyzed by lipase PS preparation (30 mg/mL) in TBME at room temperature
Entry
Alcohol (M)
Time (h)
Conversion (%)
ee(R)-1 (%)
ee(S)-4 (%)
E
1
2
3
4
5
6
7
8
9
CH3OH (0.1)
7/24
7/24
5
7/24
5/24
7/24
7/24
7/24
24
41/50
40/50
49
33/50
33/50
46/50
28/47
36/50
47
70/>99
66/>99
95
49/>99
50/>99
86/>99
38/90
57/>99
90
>99
>99
>99
>99
>99
>99
>99
>99d
>99e
>200
>200
>200
>200
>200
>200
>200
>200
>200
CH3OH (0.15)
CH3OH (0.15)a
CH3OH (0.25)
CH3OH (0.25)b
CH3OH (0.25)c
CH3OH (0.5)
CH3(CH2)3OH (0.25)
No alcohol
a Temperature 47 ꢁC.
b Lipase PS preparation (50 mg/mL).
c DIPE in the place of TBME as a solvent.
d ee for the (S)-butyl ester.
e ee(S)ꢀ6
.

1570
X.-G. Li et al. / Tetrahedron: Asymmetry 18 (2007) 1567–1573
(entry 9; route vi). No reaction was observed in the absence
of the enzyme.
rac-1 (entry 1), 9% of the observed total conversion
(46%) proceeded by hydrolysis in the case of rac-2 under
the same conditions (entry 3). In spite of the competitive
hydrolysis, the formation of (2S,3S)-5 was highly enantio-
selective (eeester >99%). The proportion of the hydrolysis
was reduced by using higher enzyme or methanol contents
(entries 4–6). When competitive alcoholysis and hydrolysis
were observed, enantioselectivity was not expressed using
E-values. In the case of rac-3, there was no reaction at all
in the presence or absence of methanol or added water (en-
tries 8–11). When CAL-B replaced lipase PS-D, hydrolysis
of rac-3 with 1 equiv of water in TBME proceeded
smoothly (entry 12). This is in accordance with the
published data for the hydrolysis of rac-3 in DIPE at
60 ꢁC (entry 13).10 The low enantioselectivity in our case
at room temperature (entry 12) can be explained by the
previous observation that elevated temperature is a
common prerequisite for highly enantioselective hydrolysis
of b-lactams in organic solvents.10–12
The b-amino ester (S)-4, which was produced when rac-1
reacts with methanol in the presence of the lipase PS prep-
aration (Scheme 1, route iii), can be further enzymatically
hydrolyzed. In order to study this, rac-4 (prepared by open-
ing the b-lactam ring of rac-1 by sodium methoxide)21 was
incubated in TBME in the presence of the lipase PS prep-
aration. As a result, the hydrolysis of (S)-4 turned to be
highly unlikely since rac-4 was hydrolyzed with excellent
(R)-enantiopreference (25% after 24 h, E > 200; Scheme
3). The observed opposite enantiopreference for the ring
opening of the b-lactam and for the hydrolysis of the cor-
responding b-amino ester in the case of lipase PS-catalysis
is in agreement with that which was observed earlier for 4-
benzyl-2-azetidinone and CAL-B catalysis.13 Interestingly,
when rac-4 was incubated in the presence of methanol
(5 equiv) and the lipase PS preparation, its hydrolysis in
TBME was totally suppressed.
2.3. Enantioselective ring opening of rac-1 by an inter-
esterification-type reaction
In the next step, we chose 0.25 M methanol and subjected
rac-1–3 to alcoholysis in the presence of lipase PS-D in dry
TBME at room temperature (Table 3). According to the
results, the importance of fluorine activation for the open-
ing of the lactam ring with methanol and/or water in the
presence of lipase PS-D is evident. While the ring opening
exclusively took place through alcoholysis in the case of
In a highly enantioselective alcoholysis of rac-1 with meth-
anol above, (S)-1 as an acyl donor forms an acyl-enzyme
intermediate with lipase PS which in the next mechanistic
step reacts with methanol (ROH) giving the product (S)-
4 (Scheme 2). This gave us an idea to make use of our expe-
F
F
F
COOH
NH2
COOMe
MeOOC
H2N
lipase PS
H2O
F
F
F
(
+
NH2
4
-
R)-6
4
S)-
rac
(
Scheme 3. Hydrolysis of rac-4 by the water in the enzyme preparation.
Table 3. Enantioselective ring opening of rac-(1–3) (0.05 M) with methanol in the presence of lipase PS-D (30 mg/mL) in TBME at room temperature
Entry
Compound
Nucleophile (M)
Time (h)
Conversion/(a)a (%)
ee1–3 (%)
eeester (%)
E
1
2
3
4
5
6
7
8
9
10
11
12
13
rac-1
rac-1
rac-2
rac-2c
rac-2
rac-2c
rac-2
rac-3
rac-3
rac-3e
rac-3
rac-3f
rac-3g
MeOH (0.25)
No alcohol
24
24
24
24
24
24
24
24
24
24
48
24
24
50/(0)
—/(47)
46/(9)
47/(3)
24/(4)
40/(2)
—/(35)
0
>99
>99
>99b
>99
>99
>99
>99
—
—
—
—
—
>200
>200
90
59
83
41
60
46
—
—
—
—
66
99
d
MeOH (0.25)
MeOH (0.25)
MeOH (0.5)
MeOH (0.5)
No alcohol
MeOH (0.25)
No alcohol
No alcohol
—
—
—
—
d
d
d
26
—
—
—
—
0
0
0
Water (1 equiv)
Water (1 equiv)
Water (1 equiv)
—/(46)
—/(50)
—
99
15
>200
a The proportion of hydrolysis (a).
(S) ꢀ 6
b ee
.
c Lipase PS preparation (50 mg/mL).
d E not given due to competitive alcoholysis and hydrolysis reactions.
e Temperature 47 ꢁC.
f CAL-B in the place of lipase PS-D.
g Ref. 10; CAL-B in DIPE at 60 ꢁC, eeacid is given.

X.-G. Li et al. / Tetrahedron: Asymmetry 18 (2007) 1567–1573
1571
rience from previous interesterification work with lipase
3. Conclusions
catalysis5,22 where the alcohol ROH was generated in situ
from an achiral ester PrCO2R through route b and reacted
further with a chiral ester. When rac-1 was dissolved in
TBME/PrCO2R (1:1; R = Me or Bu) and lipase PS-D
was added, the formation of (S)-4 (R = Me) or the corre-
sponding butyl ester (R = Bu) took place with reasonable
reactivity, especially in the case of butyl butanoate as the
source of an alcohol (Scheme 2, Table 4). Surprisingly,
the alcohol (methanol and 1-butanol) obtained from the
butanoate ester in situ was able to suppress the competing
hydrolysis of rac-1 completely. This result supports our
earlier explanation5,22 that the alcohol ROH does not nec-
essarily leave the active site as is expected according to the
classic ping-pong bi-bi mechanism.23 Instead the alcohol
stays bound at the active site, replacing the active site water
or is inserted into a possible water network in the vicinity
of the active site. This process enables the equivalent
amounts of ROH to react with the acyl-enzyme intermedi-
ate. This means that the water in the enzyme preparation
first leads to the hydrolysis of the butanoate ester (route
b), leaving an alcohol product in the active site. When
(S)-1 in rac-1forms an acyl-enzyme intermediate (route a)
at the active site bearing the bound alcohol, the formation
of (S)-4 or its butyl ester analogue becomes possible
through route c.
Herein, we have reported a Reformatsky type reaction
leading to fluorinated 4-phenyl-2-azetidinones, rac-1 and
rac-2 (Scheme 1). The racemates and the corresponding
nonfluorinated analogue rac-3 were subjected to enantio-
selective alcoholysis in dry TBME in the presence of lipase
PS-D (or lipase PS adsorbed on Celite in the presence of
sucrose) from B. cepacia. The results indicate the impor-
tance of fluorine activation for the lipase PS-catalyzed ring
opening in general and for the opening with alcohols in
particular. As a result, fluorine activation was necessary
for the acylation of methanol and 1-butanol with lipase
PS-D, with rac-3 being totally unreactive. Interestingly,
hydrolysis by water in the seemingly dry enzyme prepara-
tion was totally suppressed in the case of rac-1 and could
be minimized in the case of rac-2 allowing the kinetic reso-
lution to produce the respective enantiopure resolution
products (R)-1 and (S)-4 and (3R,4R)-2 and (2S,3S)-5 at
50% conversion. On the other hand, the relatively fast ring
opening of rac-1 and rac-2 through hydrolysis was evident
in the absence of an added alcohol while there was no
hydrolysis with rac-3 as the substrate.
The results with rac-1 and achiral alkyl butanoates also
showed highly enantioselective ring opening of the b-lac-
tam, indicating that an added alcohol can be replaced by
an alcohol produced in situ in the active site of the lipase.
Table 4. Enantioselective ring opening of rac-1 (0.05 M) in the mixture of
TBME/PrCO2R (1:1) in the presence of lipase PS-D (20 mg/mL) at room
temperature
4. Experimental
Entry
R
Time
(h)
Conversion
(%)
ee(R)ꢀ1
(%)
ee(S)ꢀ4
(%)
E
4.1. Materials and methods
1
2
Me
Bu
24
24
29
40
40
68
>99
>200
>200
>99a
Zinc, ethyl bromodifluoroacetate, ethyl bromofluoroace-
tate, cerium(IV) ammonium nitrate (CAN) and solvents
were products of Aldrich or Fluka. rac-118 and rac-310 were
prepared as previously described. All solvents were of the
highest analytical grade and were dried by standard meth-
ods when necessary. Lipase PS from B. cepacia (previously
Pseudomonas cepacia) as a free enzyme powder and when
immobilized on Celite (lipase PS-D) and on ceramic parti-
cles (lipase PS-C II) was purchased from Amano Europe,
England. C. antarctica lipase A (CAL-A) was the product
of Roche. Before use, lipase PS and CAL-A powders were
adsorbed on Celite by dissolving the enzyme (5 g) and
sucrose (3 g) in Tris–HCl buffer (250 mL, 20 mM, pH
7.9) followed by the addition of Celite (17 g) as previously
described.14 The lipase preparation containing 20% (w/w)
of the original enzyme powder was thus obtained. C. ant-
arctica lipase B (CAL-B, Novozym 435) was a generous
gift from Novo Nordisk. Preparative chromatographic sep-
arations were performed by column chromatography on
Merck Kieselgel 60 (0.063–0.200 lm). TLC was carried
out with Merck Kieselgel 60F254 sheets. Melting points
were measured on a Sanyo instrument at a heating rate
of 2 ꢁC/min. Optical rotations were determined with a Per-
kin–Elmer polarimeter, and [a]D values are given in units of
10ꢀ1 deg cm2 gꢀ1. The determination of E was based on the
equation E = ln[(1 ꢀ c)(1 ꢀ eeS)]/ln[(1 ꢀ c)(1 + eeS)] with
the use of linear regression, E being the slope of the line
ln[(1 ꢀ c)(1 ꢀ eeS)] versus ln[(1 ꢀ c)(1 + eeS)].24
a ee for the (S)-butyl ester.
The results in Table 4 (reaction in TBME/PrCO2R) are
unoptimized and cannot be compared to those in Table 2
(reaction in TBME) except to the extent that reactions in
both systems proceed with the same high enantioselectivity.
As the most important conclusion, it is clear that an alco-
hol added can be replaced by the respective alcohol pro-
duced in situ at the active site of the lipase.
2.4. Preparative-scale kinetic resolutions
The preparative-scale kinetic resolution of rac-1 and rac-2
with methanol and lipase PS-D was performed as described
in Section 4. Unreacted lactams (R)-1, (3R,4R)-2 and ami-
no esters (S)-4 and (2S,3S)-5 were obtained in enantiopure
forms at 50% conversion (Scheme 1). The specific rotation
22
½aꢁD ¼ ꢀ76:6 (c 1.0, CHCl3) for the prepared (R)-1dis-
25
agreed with the literature value18 ½aꢁD ¼ þ38:3 (c 1.01,
CHCl3) for (S)-1. In order to show that our compound
was correct, (R)-1 was transformed into the hydrochloride
20
of (R)-6, giving ½aꢁD ¼ ꢀ3:4 (c 0.84, MeOH) in agreement
20
with the literature value18 ½aꢁD ¼ þ3:3 (c 0.84, MeOH)
for the corresponding (S)-enantiomer. As rac-3 and lipase
PS-D were unable to acylate methanol, the gram-scale
kinetic resolution was impossible.

1572
X.-G. Li et al. / Tetrahedron: Asymmetry 18 (2007) 1567–1573
1
The H and 13C NMR spectra were recorded on a Bruker
(145 mg, 0.88 mmol, 57%). Mp 107–108 ꢁC. 1H NMR
(500 MHz, CDCl3) d 4.82–4.84 (dd, J = 11.0 Hz, 1.5 Hz,
1H, CH), 5.15–5.26 (dt, J = 53.5 Hz, 2.0 Hz, 1H, CHF),
6.61 (s, 1H, NH), 7.35–7.45 (m, 5arom. H); 13C NMR
(126 MHz, CDCl3) d 59.70–59.89 (d, J = 23.9 Hz, CH),
98.10–99.90 (d, J = 226.8 Hz, CHF), 125.85, 129.14,
129.16, 136.41 (arom. C), 163.81 (CO); HRMS: M+ found
(M+ calculated for C9H8FNO) 165.05970 (165.05899); MS:
m/z (relative intensity) 165 (2), 136 (4), 122 (100), 109 (3),
104 (10), 101 (5), 96 (11), 77 (7).
500 spectrometer with tetramethylsilane (TMS) as an inter-
nal standard. H–1H COSY, H–13C HQSC and H–13C
HMBC spectra were used for the assignment of the chemi-
cal shifts when necessary. 2,2,2-Trifluoroacetic acid was
used as the external standard for measuring 19F NMRs.
Mass spectra were taken on a VG 7070E mass spectrometer.
1
1
1
In a typical small-scale experiment, one of the lipase prep-
arations was added to one of the substrates rac-1–3
(0.05 M) and a nucleophile (an alcohol, 0.05–0.5 M) or
alkyl butanoate in TBME (1 mL). Unless stated otherwise,
the enzymatic reactions were performed at room tempera-
ture (23 ꢁC). The progress of the reaction was followed by
taking samples from the reaction mixture at intervals and
analyzing them by HPLC on a CHIRACEL-OD column
(0.46 · 25 cm) and GC on a Chrompack CP-Chirasil-
DEX CB column or a Chrompack CP-Chirasil-L-Valine
column. 4-Fluorophenyl acetonitrile was used as an inter-
nal standard for quantitative analyses of the resolution
mixture.
4.3. Kinetic resolution of rac-1 with methanol
rac-1 (200 mg, 1.09 mmol) was dissolved in TBME (22 mL)
after which methanol (174 mg, 5.45 mmol) and lipase PS-D
(660 mg) were added. The reaction was stopped after 17.5 h
by filtering off the enzyme at 50% conversion. The solvent
was evaporated off and the residue purified on a silica gel
column to afford (R)-1 as a solid {98 mg, yield 98%, mp
22
62–63 ꢁC, ee(R)-1 >99%, ½aꢁD ¼ ꢀ76:6 (c 1.0, CHCl3); known
25
data18 for (S)-1 ½aꢁD ¼ þ38:3 (c 1.01, CHCl3)}. (S)-4 was
obtained as an oily product {92 mg, yield 78%, ee(S)-4
22
4.2. Preparation of trans-3-fluoro-4-phenylazetidin-2-one
rac-2
>99%, ½aꢁD ¼ ꢀ5:8 (c 1.0, CHCl3)} with 1H NMR
(500 MHz, CDCl3) d 3.80 (s, 3H, OCH3), 4.44–4.49 (dd,
J = 15.5 Hz, 10.5 Hz, 1H, CH), 7.34–7.37 (m, 5arom. H);
13C NMR (126 MHz, CDCl3) d 53.25 (CH3), 58.25–58.64
(t, J = 23.9 Hz, CH), 113.33–117.39 (t, J = 255.8 Hz,
CF2), 127.91, 128.61, 128.94, 136.22 (arom. C), 164.07–
164.59 (t, J = 32.8 Hz, CO); 19F NMR (471 MHz, CDCl3)
d ꢀ99.25–(ꢀ98.67) (dd, J = 253.8 Hz, 14.1 Hz), ꢀ95.35–
(ꢀ94.78) (dd, J = 253.8 Hz, 14.1 Hz); HRMS: M+ found
(M+ calculated for C10H11F2NO2) 215.07470 (215.07579);
MS: m/z (relative intensity) 215 (0.01), 200 (0.03), 195
(0.23), 180 (0.14), 164 (0.15), 140 (4), 106 (100), 79 (15).
rac-2a was prepared following the published method for
rac-1a.18 A solution of the imine (1.52 g, 6.77 mmol) and
ethyl bromofluoroacetate (2.50 g, 13.54 mmol) in THF
(2 mL) was added to a refluxing suspension of Zn dust
(834 mg, 12.83 mmol) in THF (3 mL). After 2.5 h, the reac-
tion mixture was cooled down and quenched by adding sat-
urated NH4Cl (10 mL). The aqueous layer was extracted
with CH2Cl2, and the organic layer dried over anhydrous
Na2SO4 and filtered. After concentration, the residue was
purified on a silica gel column eluting with petroleum
ether/ethyl acetate (8:1) to afford rac-2a as an oil
4.4. Kinetic resolution of rac-2 with methanol
1
(944 mg, 3.31 mmol, 49%). H NMR (500 MHz, CDCl3)
d 3.74–3.77 (d, J = 14.9 Hz, 1H, CH2), 3.82 (s, 3H,
OCH3), 4.46–4.48 (d, J = 9.9 Hz, 1H, CH), 4.82–4.85 (d,
J = 14.8 Hz, 1H, CH2), 5.19–5.30 (d, J = 54.1 Hz, 1H,
CHF), 6.81–6.84 (m, 2arom. H), 7.05–7.07 (m, 2arom.
H), 7.21–7.23 (m, 2.7 Hz, 2arom. H), 7.42–7.44 (m, 3arom.
H); 13C NMR (126 MHz, CDCl3) d 43.90 (CH2), 55.30
(OCH3), 62.22–62.41 (d, J = 25.2 Hz, CH), 96.91–98.70
(d, J = 226.4 Hz, CHF), 114.27, 126.30, 126.80, 129.24,
129.28, 130.01, 134.45, 159.42, 163.45–163.63 (d,
rac-2 (200 mg, 1.21 mmol) was dissolved in TBME (24 mL)
after which methanol (194 mg, 6.05 mmol) and lipase PS-D
(1.20 g) were added. The reaction was stopped after 48 h by
filtering off the enzyme at 50% conversion. The solvent was
evaporated off and the residue purified on a silica gel col-
umn to afford (3R,4R)-2 as a solid (96 mg, 0.58 mmol, yield
22
96%): mp 107–108 ꢁC, ee(3R,4R)-2 >99%, ½aꢁD ¼ ꢀ19:1 (c
1.0, CHCl3)}. Compound (2S,3S)-5 was obtained as an oily
product {108 mg, 0.55 mmol, yield 91%, ee(2S,3S)-5 >99%,
22
1
J = 22.9 Hz, CO). 19F NMR (471 MHz, CDCl3)
d
½aꢁD ¼ þ10:1 (c 1.0, CHCl3)} with H NMR (500 MHz,
CDCl3) d 3.68 (s, 3H, OCH3), 4.39–4.44 (dd, J = 20.6 Hz,
4.7 Hz, 1H, CH), 5.03–5.13 (dd, J = 48.6 Hz, 4.8 Hz, 1H,
CHF), 7.29–7.37 (m, 5arom. H); 13C NMR (126 MHz,
CDCl3) d 52.21 (CH3), 57.22, 57.39 (d, J = 20.9 Hz, CH),
91.58, 93.09 (d, J = 189.9 Hz, CHF), 127.11, 128.15,
128.58, 139.50 (arom. C), 168.26–168.45 (d, J = 23.2 Hz,
CO); MS: m/z (relative intensity) 191 (10), 123 (6), 106
(100), 91(6), 79 (15).
ꢀ203.08–(ꢀ203.22) (dd, J = 56.5 Hz, 14.1 Hz.). HRMS:
M+ found (M+ calculated for C17H16FNO2) 285.11600
(285.11651); MS: m/z (relative intensity) 285 (18), 226 (2),
163 (69), 131 (3), 121 (100), 107 (6), 91 (5), 77 (10).
CAN (2.55 g, 4.65 mmol) was added in small portions to a
solution of rac-2a (441 mg, 1.55 mmol) in CH3CN/H2O
(9:1, v/v, 80 mL). After 6 h at room temperature, the mix-
ture was poured into water (100 mL). The aqueous layer
was extracted with EtOAc. The organic layer was washed
with 5% NaHCO3 (50 mL) and 10% Na2SO3 (50 mL) and
dried over anhydrous Na2SO4. The mixture was filtered
and concentrated under vacuum. The residue was purified
on silica gel column with petroleum ether/ethyl acetate
(6:1) being the eluent, affording rac-2 as a solid product
4.5. Determination of absolute configurations
The absolute configurations were determined by correlat-
ing the specific rotation data of (2S,3S)-7 and the HCl salt
of (R)-6 to the reported data.18,20 Accordingly, a solution
of (R)-1 (40 mg, 0.22 mmol) in aqueous HCl (6 M, 4 mL)

X.-G. Li et al. / Tetrahedron: Asymmetry 18 (2007) 1567–1573
1573
4. Liljeblad, A.; Kanerva, L. T. Tetrahedron 2006, 62, 5831–
5854.
was refluxed for 2 h before the solvent was evaporated,
affording the HCl salt of (R)-6 (mp 160–161 ꢁC) in a quan-
20
5. Li, X.-G.; Kanerva, L. T. Org. Lett. 2006, 8, 5593–5596.
6. Soloshonok, V. A.; Kirilenko, A. G.; Fokina, N. A.;
Shishkina, I. P.; Galushko, S. V.; Kukhar, V. P.; Svedas, V.
K.; Kozlova, E. V. Tetrahedron: Asymmetry 1994, 5, 1119–
1126.
7. Soloshonok, V. A.; Kirilenko, A. G.; Fokina, N. A.; Kukhar,
V. P.; Galushko, S. V.; Svedas, V. K.; Resnati, G. Tetra-
hedron: Asymmetry 1994, 5, 1225–1228.
8. Soloshonok, V. A.; Fokina, A. F.; Rybakova, A. V.;
Shishkina, I. P.; Galushko, S. V.; Sorochinsky, A. E.;
Kukhar, V. P.; Savchenko, M. V.; Svedas, V. K. Tetrahedron:
Asymmetry 1995, 6, 1601–1610.
9. Faulconbridge, S. J.; Holt, K. E.; Sevellano, L. G.; Lock, C.
J.; Tiffin, P. D.; Tremayne, N.; Winter, S. Tetrahedron Lett.
2000, 41, 2679–2681.
titative yield with ½aꢁD ¼ ꢀ3:4 (c 0.84, MeOH) {the litera-
20
ture value18 for the HCl salt of (S)-6 ½aꢁD ¼ þ3:3 (c 0.84,
1
MeOH)}. H NMR (500 MHz, CD3OD) d 5.06–5.11 (dd,
J = 6.0 Hz, 19.0 Hz, 1H, CH), 7.52–7.56 (m, 5arom. H);
13C NMR (126 MHz, CD3OD)
d
56.72–57.10 (t,
J = 23.9 Hz, CH), 112.15–116.27 (t, J = 257.8 Hz, CF2),
128.24, 128.51, 129.31, 130.49 (arom. C), 166.6 (t,
J = 28.9 Hz, CO).
A solution of (2S,3S)-5 (30 mg, 0.15 mmol, ee >99%) in
aqueous HCl (6 M, 2 mL) was refluxed for 4 h followed
by evaporation of the solvent. The resulting solid was trea-
ted with iPrOH (2 mL) and propylene oxide (35 mg,
0.60 mmol) for 3 h. The precipitate was collected and
´
´
´
10. Forro, E.; Paal, T.; Tasnadi, G.; Fulo¨p, F. Adv. Synth. Catal.
¨
i
washed with PrOH (4 mL) to afford amino acid (2S,3S)-
2006, 348, 917–923.
7 as a solid (mp 178–180 ꢁC) in a quantitative yield with
´
11. Forro, E.; Fulo¨p, F. Mini-Rev. Org. Chem. 2004, 1, 93–
¨
22
½aꢁD ¼ ꢀ42:8 (c 0.4, CH3OH) {the literature value20 for
102.
22
(2S,3S)-7 ½aꢁD ¼ ꢀ40 (c 0.4, CH3OH)}. 1H NMR
´
12. Park, S.; Forro, E.; Grewal, H.; Fulo¨p, F.; Kazlauskas, R. J.
¨
Adv. Synth. Catal. 2003, 345, 986–995.
13. Li, X.-G.; Kanerva, L. T. Adv. Synth. Catal. 2006, 348, 197–
205.
14. Sundholm, O.; Kanerva, L. T. J. Chem. Soc., Perkin Trans. 1
1993, 2407–2410.
15. Veum, L.; Kanerva, L. T.; Halling, P. J.; Maschmeyer, T.;
Hanefeld, U. Adv. Synth. Catal. 2005, 347, 1015–1021.
16. Brieva, R.; Crich, J. Z.; Sih, C. J. J. Org. Chem. 1993, 58,
1068–1075.
(500 MHz, CD3OD) d 4.68–4.73 (dd, J = 24.2 Hz, 3.8 Hz,
1H, CH), 4.98–5.09 (dd, J = 50.9, 3.9 Hz, 1H, CHF),
7.32–7.43 (m, 5 arom. H); 13C NMR (126 MHz, CD3OD)
d 57.51, 57.68 (d, J = 21.4 Hz, CH), 89.99, 91.52 (d,
J = 192.3 Hz, CHF), 129.73, 129.93, 130.52, 133.91 (arom.
C), 172.02 (CO).
17. Qiu, X.-L.; Meng, W.-D.; Qing, F.-L. Tetrahedron 2004, 60,
6711–6745.
Acknowledgement
18. Marcotte, S.; Pannecoucke, X.; Feason, C.; Quirion, J.-C. J.
Org. Chem. 1999, 64, 8461–8464.
This work was supported by the Academy of Finland
(Grant 109743 to L.K.).
19. Sorochinsky, A.; Voloshin, N.; Markovsky, A.; Belik, M.;
Yasuda, N.; Uekusa, H.; Ono, T.; Berbasov, D. O.;
Soloshonok, V. A. J. Org. Chem. 2003, 68, 7448–7454.
20. Andrews, P. C.; Bhaskar, V.; Bromfield, K. M.; Dodd, A. M.;
Duggan, P. J.; Duggan, S. A. M.; McCarthy, T. D. Synlett
2004, 791–794.
References
1. Thayer, A. M. Chem. Eng. News 2006, 84, 15–24.
2. Uoto, K.; Ohsuki, S.; Takenoshita, H.; Ishiyama, T.; Iimura,
S.; Hirota, Y.; Mitsui, I.; Hirofumi, T.; Soga, T. Chem.
Pharm. Bull. 1997, 45, 1793–1804.
3. (a) Enantioselective Synthesis of b-Amino Acids, 2nd ed.;
Juaristi, E., Soloshonok, V., Eds.; Wiley-VHC: New York,
2005; (b) Palomo, C.; Aizpurua, J. M.; Ganboa, I.; Oiarbide,
M. In Enantioselective Synthesis of b-Amino Acids, 2nd ed.;
Wiley-VHC: New York, 2005; pp 477–495.
´
21. Gyonfalvi, S.; Szakonyi, Z.; Fulo¨p, F. Tetrahedron: Asym-
¨
metry 2003, 14, 3965–3972.
22. Liljeblad, A.; Kavenius, H.-M.; Ta¨htinen, P.; Kanerva, L. T.
Tetrahedron: Asymmetry 2007, 18, 181–191.
23. Straathof, A. J. J.; Rakels, J. L. L.; Heijnen, J. J. Biocatalysis
1992, 7, 13–27.
24. Chen, C.-S.; Fujimoto, Y.; Girdaukas, G.; Sih, C. J. J. Am.
Chem. Soc. 1982, 104, 7294–7299.