Experimental and computation studies on Candida antarctica lipase B-catalyzed enantioselective alcoholysis of 4-bromomethyl-β-lactone leading to enantiopure 4-bromo-3-hydroxybutanoate

Article abstract of DOI:10.1002/adsc.201200901

Both enantiomers of optically pure 4-bromo-3-hydroxybutanoate, which is an important chiral building block in the syntheses of various biologically active compounds including statins, were synthesized from rac-4-bromomethyl-β- lactone through kinetic resolution. Candida antarctica lipase B (CAL-B) enantioselectively catalyzes the ring opening of the β-lactone with ethanol to yield ethyl (R)-4-bromo-3-hydroxybutanoate with high enantioselectivity (E>200). The unreacted (S)-4-bromomethyl-β-lactone was converted to ethyl (S)-4-bromo-3-hydroxybutanoate (>99% ee), which can be further transformed to ethyl (R)-4-cyano-3-hydroxybutanoate, through an acid-catalyzed ring opening in ethanol. Molecular modeling revealed that the stereocenter of the fast-reacting enantiomer, (R)-bromomethyl-β-lactone, is ～2 A from the reacting carbonyl carbon. In addition, the slow-reacting enantiomer, (S)-4-bromomethyl-β-lactone, encounters steric hindrance between the bromo substituent and the side chain of the Leu278 residue, while the fast-reacting enantiomer does not have any steric clash. Copyright

Full text of DOI:10.1002/adsc.201200901

FULL PAPERS
DOI: 10.1002/adsc.201200901
Experimental and Computation Studies on Candida antarctica
Lipase B-Catalyzed Enantioselective Alcoholysis of
4-Bromomethyl-b-lactone Leading to Enantiopure
4-Bromo-3-hydroxybutanoate
Jung Yun Lim,a Nan Young Jeon,a A-Reum Park,a,b Bora Min,a Bum Tae Kim,a
Seongsoon Park,c,* and Hyuk Leea,b,*
a
Medicinal Chemistry Research Center, Korea research Institute of Chemical Technology, 141 Gajeong-ro, Yuseong-gu,
Daejeon 305-600, Republic of Korea
Fax : (+82)-42-861-0307; phone: (+82)-42-860-7019; e-mail: leeh@krict.re.kr
Medicinal and Pharmaceutical Chemistry, University of Science & Technology, 176 Gajeong-ro, Yuseong-gu, Daejeon
b
305-350, Republic of Korea
c
Department of Chemistry, Center for NanoBio Applied Technology, Institute of Basic Sciences, Sungshin Womenꢀs
University, Seoul 142-732, Republic of Korea
Fax : (+82)-2-920-2047; phone: (+82)-2-920-7646; e-mail: spark@sungshin.ac.kr
Received: October 5, 2012; Revised: December 16, 2012; Published online: February 28, 2013
Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/adsc.201200901.
Abstract: Both enantiomers of optically pure 4- droxybutanoate, through an acid-catalyzed ring
bromo-3-hydroxybutanoate, which is an important opening in ethanol. Molecular modeling revealed
chiral building block in the syntheses of various bio- that the stereocenter of the fast-reacting enantiomer,
logically active compounds including statins, were (R)-bromomethyl-b-lactone, is ~2 ꢁ from the react-
synthesized
from
rac-4-bromomethyl-b-lactone ing carbonyl carbon. In addition, the slow-reacting
through kinetic resolution. Candida antarctica lipase enantiomer, (S)-4-bromomethyl-b-lactone, encoun-
B (CAL-B) enantioselectively catalyzes the ring ters steric hindrance between the bromo substituent
opening of the b-lactone with ethanol to yield ethyl and the side chain of the Leu278 residue, while the
(R)-4-bromo-3-hydroxybutanoate with high enantio- fast-reacting enantiomer does not have any steric
selectivity (E>200). The unreacted (S)-4-bromo- clash.
methyl-b-lactone was converted to ethyl (S)-4-
bromo-3-hydroxybutanoate (>99% ee), which can Keywords: kinetic resolution; b-lactones; lipase; mo-
be further transformed to ethyl (R)-4-cyano-3-hy- lecular modeling; statins
Introduction
ylglutaryl-coenzyme A (HMG-CoA) in cholesterol
biosynthesis. Since mevastatin was discovered in
Both enantiomers of 4-halo-3-hydroxybutanoates are 1971,[3] numerous statins, such as atorvastatin, fluvas-
valuable building blocks in the syntheses of biologi- tatin, lovastatin, pitavastatin, pravastatin, rosuvasta-
cally active compounds or pharmaceutical drugs. The tin, and simvastatin, have been derived from fungal
R-enantiomer has been used as a starting material in metabolites or synthesized for clinical use. Structural-
the synthesis of l-carnitine, which is a carrier mole- ly, the stereochemistry of a 3,5-dihydroxy acid side
cule for the transport of fatty acids into the mitochon- chain[4] is important for the drug activity of all statins
drial matrix.[1] Similarly, the S-enantiomer can be con- except fluvastatin, which has been used as a racemic
verted to (R)-4-cyano-3-hydroxybutanoate, a synthetic mixture.[5] Thus, high optical purity of the statin side
intermediate for various statins, which are some of chain is critically important. However, obtaining
the best-selling pharmaceuticals with sales of $10 bil- enantiopure statin side chains is synthetically chal-
lion a year.[2]
lenging, and several chemoenzymatic processes have
Statins are cholesterol-lowering drugs that inhibit therefore been developed to prepare building blocks
the production of mevalonate from 3-hydroxy-3-meth- of high stereochemical purity.[6]
1808
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2013, 355, 1808 – 1816

Experimental and Computation Studies on Candida antarctica Lipase B
Three types of biocatalysts have been used for the use computer modeling to propose a molecular basis
direct preparation of (S)-4-chloro-3-hydroxybutanoate for the high enantioselectivity of the lipase-catalyzed
or (R)-4-cyano-3-hydroxybutanoate: reductase or al- ring-opening reaction.
cohol dehydrogenase, nitrilase, and halodehydroge-
nase. Simizu and co-workers developed the enantiose-
lective reduction of 4-chloro-3-ketobutanoate ester by Results and Discussion
Candida magnoliae ketone reductase to provide (S)-4-
chloro-3-hydroxybutanoate. Bacillus megaterium glu- Screening Lipases and Esterases toward the
cose dehydrogenase was used to regenerate the Enantioselective Ring Opening of rac-4-
NADPH consumed in the reaction.[7] A similar ap- Bromomethyl-b-lactone (rac-1a) with n-Butanol
proach has been developed using alcohol dehydrogen-
ase combined with halohydrin dehalogenase.[8] Alco- The starting racemic 4-bromomethyl-b-lactone (rac-
hol dehydrogenase reduces alkyl 4-chloro-3-ketobut- 1a) was synthesized from the reaction of vinyl acetic
ACHTUNGTRENNUNGanoate to alkyl (S)-4-chloro-3-hydroxybutanoate and acid with bromine according to the method in the lit-
halohydrin dehalogenase catalyzes the conversion of erature.[13] The ring-opening alcoholysis of rac-1a with
alkyl (S)-4-chloro-3-hydroxybutanoate to (R)-3-hy- n-butanol produces n-butyl 4-bromo-3-hydroxybutan-
droxy-4-cyanobutanoate by replacing the chlorine
ACHTUNGTRENoNUNG ate (2a). Twenty commercial lipases or esterases
atom with cyano group. An alternative approach has were screened toward the ring opening of the b-lac-
been developed based on hydrolysis by nitrilase. Burk tone (rac-1a) with n-butanol (3 equiv.) (Scheme 1,
and co-workers demonstrated hydrolytic desymmetri-
zation of 3-hydroxyglutaronitrile to provide (R)-4-
cyano-3-hydroxybutanoic acid.[9]
Although direct syntheses utilizing reductases are
valuable methods, the high cost of cofactor (i.e.,
NADPH) necessitates a cofactor regeneration system.
However, cofactor regeneration often leads to diffi-
cult product recovery due to high concentrations of
side products. Thus, a synthetic route using a cofactor-
free enzyme is an attractive alternative, and the
lipase-catalyzed kinetic resolution is one of the most
promising examples of such routes to obtain both
Scheme 1. Kinetic resolution through ring opening of rac-1a.
enantiomers.
Lipases do not require cofactors and are relatively Supporting Information, Table S1). The reactions
cost-effective enzymes due to their ease of produc- were performed at room temperature for 24 h. Ten li-
tion. In addition, lipases are very stable in organic sol- pases catalyzed the alcoholysis and five enzymes
vents and at high temperature. Although lipase-cata- showed high enantioselectivity (E>200) (Table 1).
lyzed kinetic resolution of racemic 4-halo-3-hydroxy- However, most lipases showed low conversion (<
butanoate or 4-cyano-3-hydroxybutanoate esters is 10%) but Candida antarctica lipase B (CAL-B, Novo-
the simplest approach to obtaining the enantiopure zym 435) showed near 50% conversion. Hence, CAL-
acids, it is very challenging because the stereocenter is B (Novozym 435) was selected for further optimiza-
far from the reacting carbonyl carbon. Moreover, li- tion of the enantioselective ring opening of the b-lac-
pases do not efficiently resolve racemic mixtures pos- tone. Among the rest of the lipases with lower enan-
sessing the stereocenter on the acyl moiety[10]; indeed, tioselectivity (E<15), Burkholderia cepacia lipase
a lipase-catalyzed kinetic resolution has shown only (Amano Lipase PS-CI) has high reactivity with 53%
a moderate enantiomeric ratio (E= ~36).[11] An alter- conversion in 24 h (Table 1).
native starting compound for lipase-catalyzed kinetic
resolution is 4-halomethyl-b-lactone, which can afford
alkyl 4-halo-3-hydroxybutanoate through the ring- Optimization of CAL-B-Catalyzed Enantioselective
opening reaction with an alcohol. Compared to the Ring Opening of rac-4-Bromomethyl-b-lactone (rac-
esters, the stereocenter of the b-lactone is closer to 1a)
the reacting carbonyl carbon, suggesting that the ring
opening of the b-lactone could provide higher enan- In general, lipase-catalyzed reactions are affected by
tioselectivity.[12] Herein, we present a novel approach particular reaction conditions, such as solvents, reac-
based on the lipase-catalyzed ring-opening reaction of tion temperature, and substrates. Thus, we explored
the b-lactone to produce both enantiomers of alkyl 4- such reaction conditions to enhance the reaction rate
bromo-3-hydroxybutanoate with over 99% ee and an without loss of enantioselectivity in the CAL-B-cata-
enantiomeric ratio greater than 200. In addition, we lyzed ring-opening alcoholysis (Table 2).
Adv. Synth. Catal. 2013, 355, 1808 – 1816
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1809

Jung Yun Lim et al.
FULL PAPERS
Table 1. The list of enzymes catalyzing the enantioselective alcoholysis of rac-1a.[a]
Enzyme
% ee
Conversion [%]
E[b]
S-1a
R-2a
Candida antarctica lipase B (Novozym 435)
Candida rugosa lipase
Hog pancreas lipase
Thermomyces lanuginosus lipase
Streptomyces diastatochromogenes esterase
Burkholderia cepacia lipase (Amano Lipase PS-C I)
>99
3.4
8.5
3.8
1.5
63
93
52
>200
>200
>200
>200
>200
7
>99
>99
>99
>99
56
3.3
7.8
3.6
1.5
53
[a]
rac-1a (0.08 mmol), n-BuOH (0.24 mmol) and enzyme (2 mg) in 1 mL of hexane/diethyl ether (9:1) were shaken for 24 h
at 258C.
[b]
R-selective. E=enantiomeric ratio as defined in ref.[22]
Table 2. Rate and enantioselectivity of the CAL-B-catalyzed ring-opening reaction of rac-1a.[a]
Entry Alcohol (equiv.) Solvent (1 mL)
Temp. [8C] Time [h] % ee (1) % ee (2) Conversion [%][b] E[c]
1
2
3
4
5
6
7
8
n-butanol (3)
n-butanol (3)
n-butanol (3)
n-butanol (3)
n-butanol (3)
n-butanol (3)
n-butanol (3)
n-butanol (3)
n-butanol (3)
n-butanol (3)
n-butanol (1)
n-butanol (3)
n-butanol (5)
n-butanol (5)
hexane/diethyl ether (9:1) 25
5
5
5
5
5
5
3
2
1
1
0.5
0.5
0.5
1
1
0.5
1
0.5
1
0.5
1
>99
>99
28
>99
n.d.[d]
>99
>99
>99
>99
55
>99
77
71
>99
14
88
>99 (2a) 50
>99 (2a) 50
>99 (2a) 22
>99 (2a) 50
>200
>200
>200
>200
n.d.
diethyl ether
THF
25
25
25
25
30
40
50
60
t-butyl methyl ether
methylene chloride
diethyl ether
diethyl ether
diethyl ether
diethyl ether
n.d.
n.d.
>99 (2a) 50
>99 (2a) 50
>99 (2a) 50
>99 (2a) 50
>99 (2a) 35
>99 (2a) 50
>99 (2a) 43
>99 (2a) 41
>99 (2a) 50
>99 (2a) 12
>99 (2b) 47
>99 (2b) 50
>99 (2c) 44
>99 (2c) 50
>99 (2d) 46
>99 (2d) 50
>200
>200
>200
>200
>200
>200
>200
>200
>200
>200
>200
>200
>200
>200
>200
>200
9
10
11
12
13
14
15
16
17
18
19
20
21
hexane/diethyl ether (9:1) 60
diethyl ether
diethyl ether
diethyl ether
diethyl ether
60
60
60
60
60
60
60
60
60
60
60
n-butanol (100) not added
methanol (1)
methanol (1)
ethanol (1)
ethanol (1)
n-propanol (1)
n-propanol (1)
diethyl ether
diethyl ether
diethyl ether
diethyl ether
diethyl ether
diethyl ether
>99
79
>99
84
>99
[a]
[b]
rac-1a (0.08 mmol), alcohol (0.24 mmol) and enzyme (2 mg) in 1 mL of solvent were shaken.
The reaction was monitored by GC until the progress proceeded to 50% conversion. For a highly enantioselective reac-
tion, the maximum conversion is 50%.
R-selective.
n.d.=not detected.
[c]
[d]
First, we investigated various solvents, such as creasing reaction temperature, the rate of alcoholysis
hexane, diethyl ether, tert-butyl methyl ether, THF, increased and the reaction reached 50% conversion at
and methylene chloride. The starting lactone rac-1a is 608C in 1 h (entry 9 in Table 2). We also varied the
poorly miscible in hexane and thus a mixed solution amount of n-butanol and examined its influence on
with diethyl ether was examined instead. The reac- reaction conversion (entries 11–15 in Table 2); while
tions in a hexane/diethyl ether mixture (9:1), diethyl three equiv. of n-butanol were used in the initial stud-
ether, and tert-butyl methyl ether reached 50% con- ies, the reaction with one equivalent of n-butanol pro-
version in 5 h without loss of enantioselectivity while vided the highest conversion in the shortest time
the reactions in THF or methylene chloride were (entry 11 in Table 2). In contrast, 5 equiv. of n-butanol
slower or did not occur, respectively (entries 1–5 in slightly decreased the conversion (entry 13 in
Table 2). The temperature effect was also investigated Table 2), and a large excess (100 equiv.) of n-butanol
on the ring-opening reaction (entries 6–10 in Table 2), was significantly detrimental (12% conversion in 1 h),
although it was completed at 258C in 5 h. With in- although high enantioselectivity (E>200) was main-
1810
asc.wiley-vch.de
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2013, 355, 1808 – 1816

Experimental and Computation Studies on Candida antarctica Lipase B
tained (entry 15 in Table 2). Other alcohols, such as tiomeric excess of (R)-4-cyano-3-hydroxybutanoate
methanol, ethanol, and n-propanol, were also tested was higher than 99% ([a]2D5: À288 (c 1.0, EtOH); lit.
for the alcoholysis of rac-1a (entries 16–21 in Table 2). [a]2D5: À318 (c 1.0, EtOH)).[14]
The reactions with the equivalent amount of such al-
cohols also provided high enantioselectivity (E>200)
and the reactions were completed in 1 h (entries 17, Enantioselective Ring Opening of Racemic
19, and 21). However, the reactions with those alco- Chloromethyl- and Azidomethyl-b-lactones (rac-1)
hols did not achieve 50% conversion in 0.5 h while n- with Ethanol by CAL-B
butanol provides 50% conversion in the same time
(entries 11, 16, 18, and 20 in Table 2). Although n-bu- We expanded the substrate range to chloromethyl-
tanol afforded the highest conversion within 0.5 h, (rac-1b) and azidomethyl-b-lactone (rac-1c). The rac-
ethanol was used for the larger-scale enantioselective emic 4-chloromethyl-b-lactone (rac-1b) was prepared
ring opening of the lactone (rac-1) because the re- from the reaction of vinylacetic acid and ZnCl2 ac-
maining ethanol can be conveniently removed by cording to the method in the literature,[15] and rac-4-
vacuum evaporation after the completion of the reac- azidomethyl-b-lactone (rac-1c) was obtained from the
tion.
reaction of rac-1a and NaN3 in DMSO at 08C. The
ring openings of rac-1b and rac-1c were carried out
with ethanol at 608C in tert-butyl methyl ether. The
reactions showed rather different enantioselectivities
(Table 3). In the reaction with the azido compound
(rac-1c), the enantiomeric ratio is as high as that in
Synthesis of Optically Pure Ethyl (S)-4-Bromo-3-
hydroxybutanoate (S-2c)
The more useful building block of the statin side the ring opening of 4-bromomethyl-b-lactone (rac-
chain is (S)-4-bromo-3-hydroxybutanoate, which can 1a), while the enantiomeric ratio of the reaction to-
be converted to (R)-4-cyano-3-hydroxybutanoate. The wards the chloro compound (rac-1b) decreased to
CAL-B-catalyzed ring opening of rac-4-bromomethyl- 161. Presumably, a smaller substituent (i.e., the chloro
b-lactone (rac-1a), meanwhile, produces the opposite group) than the bromo group causes lower enantiose-
enantiomer [R-2c: [a]D25: +10.98 (c 1.0, EtOH)] and lectivity in the b-lactone ring-opening reactions. How-
thus it is necessary to convert the unreacted (S)-4- ever, when the reaction with the chloro compound
bromomethyl-b-lactone [S-1a: [a]2D5: À40.88 (c 0.5, was performed at lower temperature (308C), the
CHCl3)] to ethyl (S)-4-bromo-3-hydroxybutanoate (S- enantiomeric ratio was greater than 200, although the
2c). The reaction was achieved by an acid-catalyzed reaction proceeded slowly to reach 50% conversion.
ring opening in ethanol that maintained the enantio-
meric excess (>99% ee) and exhibited an optical ro-
tation value (S-2c: [a]2D5: À9.98 (c 1.0, EtOH)) similar Molecular Basis of the Ring Opening of 4-
to that in the literature [lit. [a]2D5: À118 (c 1.0, Bromomethyl-b-lactone
EtOH)][14] (Scheme 2). The further conversion of (S)-
4-bromo-3-hydroxybutanoate to (R)-4-cyano-3-hy- Computer modeling was used to rationalize the high
droxybutanoate was obtained by a reaction with enantioselectivity of CAL-B toward the b-lactone.
NaCN in an aqueous ethanol solution (80% v/v) ac- The mechanism of the lipase-catalyzed ring opening
cording to the method in the literature.[14] The enan- of the b-lactone is presumably composed of two steps.
In the first step, the side chain of the catalytic serine
residue (Ser105) attacks the carbonyl carbon atom
and an acyl enzyme intermediate is produced through
a tetrahedral intermediate. Then, the alcohol sub-
strate attacks the carbonyl carbon atom of the acyl
enzyme and the consecutive deacylation affords the
product ester in the second step. We hypothesized
that the first step primarily determines the enantiose-
lectivity of the reaction because the reverse ring for-
mation of the first step barely occurs and the second
step (acyl enzyme alcoholysis) is identical to what
happens in the kinetic resolution of esters, which has
been demonstrated to be of low enantioselectivity.[11]
Hence, molecular modeling was focused on the ring-
opening step.
In the ring-opening step of the reaction, a tetrahe-
Scheme 2. Synthesis of S-2c.
dral intermediate is formed before the acyl enzyme is
Adv. Synth. Catal. 2013, 355, 1808 – 1816
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1811

Jung Yun Lim et al.
FULL PAPERS
Table 3. CAL-B-catalyzed ring opening of rac-b-lactones (rac-1) with EtOH.[a]
R
Temp. [8C]
Time [h]
% ee
Conversion [%]
E
S-1
R-2
Cl (1b)
Cl (1b)
N3 (1c)
60
30
60
2
6
1
92
>99
>99
96 (2e)
>99 (2e)
>99 (2f)
49
50
50
161
>200
>200
[a]
rac-1 (0.08 mmol), EtOH (0.24 mmol) and CAL-B (Novozym 435) (2 mg) in 1 mL of tert-butyl methyl ether were shaken.
produced. The tetrahedral intermediate is constructed
by the attack of the catalytic serine to the carbonyl
carbon atom of the b-lactone. For catalytically pro-
ductive binding, the substrate positions its larger sub-
stituent towards the surface of CAL-B.[16] The catalyt-
ically productive tetrahedral intermediate should pos-
sess the key hydrogen bonds with the catalytic resi-
dues, such as Thr40, Ser105, Gln106, and His224 [(a)–
(e) in Figure 1].[17] The catalytic histidine (His224)
forms hydrogen bonds with both the serine Og and O
of the lactone ring, and the oxyanion is stabilized
through hydrogen bonding with Thr40 and Gln106.
The conformations of the tetrahedral intermediate
of both enantiomers can be defined according to the
location of the bromo substituent. The bromo group
freely rotates and may choose one of the three con-
Figure 1. Line diagrams of a possible catalytically productive
conformation of the tetrahedral intermediate for the ring
opening of the (R)-b-lactone (R-1a, the fast-reacting enan-
tiomer) catalyzed by CAL-B.
formations for each enantiomer as defined in The hydrogen bond between the nitrogen atom of the
Figure 2. In one of the conformations for both enan- catalytic histidine and the oxygen atom of the lactone
tiomers, the bromine atom is oriented toward the oxy- was weak or missing due to steric strain of the bromo
À
anion hole [(c) and (f) in Figure 2]. The large bromo substituent with Ile189. The N O distance is 3.65 ꢁ,
group probably interferes with the formation of hy- which is not acceptable for a hydrogen bond. Thus,
drogen bonds between the oxyanion and oxyanion conformation (b) cannot be catalytically productive.
hole residues, Thr40 and Gln106. Thus, these two con-
For the slow enantiomer, none of the conforma-
formations [(c) and (f) in Figure 2] are not reactive tions maintained all of the key hydrogen bonds re-
structures and were therefore discarded from further quired for catalysis. Conformation (d) of the slow
consideration in modeling studies.
enantiomer is unlikely to be catalytically productive
We minimized all models of the remaining confor- because the hydrogen bond between the His224 Ne
mations [(a), (b), (d), and (e) in Figure 2] of the tetra- and b-lactone oxygen is weak or missing (3.88 ꢁ).
hedral intermediate and found only one conformation The orientation of the bromo substituent in confor-
that maintained all the key hydrogen bonds (Table 4). mation (d) generates a steric clash with Leu278. Simi-
For the conformation (a) of the fast enantiomer, (R)- larly, conformation (e) is catalytically unproductive
1a, the bromo group was positioned toward Leu278. due to the absence of a hydrogen bond between
The bromo substituent did not have any steric clash His224 Ne and the b-lactone oxygen; this distance of
with enzyme residues and all key hydrogen bonds 3.87 ꢁ is beyond the limit for hydrogen bonding.
were maintained. However, the conformation (b)
Rationalizing the high enantioselectivity of the
(bromo pointing toward Ile189) of the fast enantio- CAL-B-catalyzed ring-opening reaction of the b-lac-
mer did not possess all of the key hydrogen bonds. tone was straightforward since both conformations for
1812
asc.wiley-vch.de
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2013, 355, 1808 – 1816

Experimental and Computation Studies on Candida antarctica Lipase B
Figure 2. Proposed orientation of the b-lactone in the substrate-binding site of CAL-B. The conformations (c) and (f) are un-
likely for catalysis due to the high crowding.
Table 4. Key hydrogen bonds in possible transition state models of the CAL-B-catalyzed ring opening of the b-lactone.
Hydrogen H-bond dis- Comments
bond[a]
a
tance,
ꢁ (angle)[b])
conformation (a)[c] of the
3.11 (1328) all key H-bonds exist, a catalytically productive structure
R-enantiomer
b
c
3.17 (1568)
3.03 (1688)
d
e
a
2.49 (1758)
2.67 (1708)
conformation (b) of the R-
2.96 (1488) H-bond ꢃbꢀ is missing, it shows steric strain between Ile189 and the
enantiomer
bromo substituent of the b-lactone
b
c
3.65 (1528)
3.01 (1688) unlikely to be a catalytically productive structure
d
e
a
2.49 (1758)
2.69 (1688)
conformation (d) of the S-
3.17 (1438) H-bond ꢃbꢀ is missing, it shows steric strain between Leu278 and the
enantiomer
bromo substituent of the b-lactone
b
c
3.88 (1578)
2.85 (1678) unlikely to be a catalytically productive structure.
d
e
a
2.51 (1748)
2.70 (1688)
conformation (e) of the S-
2.96 (1448) H-bond ꢃbꢀ is missing, it shows steric strain between Leu278 and the
enantiomer
bromo substituent of the b-lactone
b
c
3.87 (1608)
2.89 (1688) unlikely to be a catalytically productive structure.
d
e
2.49 (1768)
2.71 (1688)
[a]
Scheme 1 defines the hydrogen bonds.
Distance between non-hydrogen atoms (N N, N O or O O). Distances of 2.7–3.2 ꢁ are considered as a range for a hy-
[b]
À
À
À
À À
drogen bond. Angles refer to the N H O or similar angle. For an ideal hydrogen bond, this angle is 1808, but angles of
>1208 are consistent with a hydrogen bond.
[c]
The conformations were defined as in Figure 2.
the slow-reacting enantiomer lack one of the key hy- Conclusions
drogen bonds and encounter steric clashes with
Leu278, whereas one of the conformations for the Lipase-catalyzed kinetic resolution is one of the most
fast-reacting enantiomer maintains all the key hydro- useful methods to obtain an enantiopure chiral com-
gen bonds and does not have any steric clashes pound. In general, lipases successfully resolve chiral
(Figure 3).
alcohols as the ester forms with a stereocenter near
Adv. Synth. Catal. 2013, 355, 1808 – 1816
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1813

Jung Yun Lim et al.
FULL PAPERS
used statin intermediate, ethyl (R)-4-cyano-3-hydroxy-
butanoate, without loss of optical purity. Molecular
modeling revealed that the slow-reacting enantiomer,
(S)-4-bromomethyl-b-lactone, encounters steric hin-
drance between the bromomethylene substituent and
the side chain of the the Leu278 residue, while one of
the conformations of the tetrahedral intermediate for
the fast-reacting enantiomer shows a productive con-
formation without any steric clash.
Experimental Section
General remarks, characterization data of rac-1c including
spectroscopic and analytical data, experimental details (2c,
2e, and 2f), and Table S1 are available as Supporting Infor-
mation.
Figure 3. A proposed tetrahedral intermediate of the fast-
reacting enantiomer of the b-lactone. CAL-B is shown in
ribbon representation, except for Ile189, Leu278, His224,
Ser105, Gln106, and Thr42, which are represented by sticks.
The b-lactone is also shown as a stick model. The dashed
lines indicate hydrogen bonds.
Enzyme-Catalyzed Enantioselective Ring Opening of
rac-1a with Alcohol
As a representative procedure, rac-1a (13 mg, 0.08 mmol),
Novozym 435 (Candida antarctica lipase B, 2 mg) and n-
BuOH (18 mg, 0.24 mmol) were dissolved in hexane/diethyl
ether (9:1). The mixture was shaken at 258C for 24 h, and
filtered to remove Novozym 435. The residue was analyzed
by GC. GC analyses were conducted with an Agilent 6890N
equipped with a flame ionization detector using lipodex-A
(30 mꢄ0.25 mm I.D., MN). The injector and detector tem-
peratures were 2008C and 2208C, respectively. The GC
column temperature was initially programmed at 608C for
30 min, then gradient to 1008C at 18C·minÀ1, and then held
for 10 min. S-1a and R-1a were detected at 43.9 min and
45.1 min, respectively. S-2a and R-2a were detected at
85.9 min and 86.4 min, respectively.
The progress of the optimization reactions and the ee
values were followed by taking samples at intervals and ana-
lyzing them by GC until the reaction proceeded to ~50%
conversion. When the CAL-B-catalyzed ring opening of rac-
1a with methanol (Table 2, entries 16–17), the resulting ring-
opened enantiomers, S-2b and R-2b, were detected at
52.8 min and 53.4 min, respectively. When ethanol was used
instead of methanol in the reaction (Table 2, entries 18–19),
S-2c and R-2c were detected at 63.5 min and 65.3 min, re-
spectively. In the reaction with n-propanol (Table 2, en-
tries 20–21), S-2d and R-2d were detected at 73.9 min and
74.6 min, respectively.
the reacting carbonyl carbon (i.e., secondary alco-
hols), whereas they do not show high enantioselectivi-
ty toward chiral acids or alcohol esters possessing the
stereocenter far away from the reacting carbonyl
carbon. CAL-B has shown only moderate enantiose-
lectivity toward ethyl 4-chloro-3-hydroxybutanoate in
ammonolysis,[11] presumably because the stereocenter
is located on the acyl moiety and is far from the react-
ing carbonyl carbon (~2.6 ꢁ). Thus, we reasoned that
a substrate, such as the b-lactone, with the chiral
atom near the reacting carbonyl carbon might in-
crease the enantioselectivity of CAL-B. Indeed, it has
been reported that CAL-B can resolve a b-lactam,
which is structurally similar to the b-lactone, with
high enantioselectivity.[18] In the current investigation,
we chose 4-bromomethyl-b-lactone as a substrate in
a lipase-catalyzed resolution to obtain enantiopure 4-
bromo-3-hydroxybutanoate; the stereocenter of this
substrate is much closer (~2 ꢁ) to the reacting car-
bonyl carbon than that of the ester form (see
Scheme 1). In addition, the chiral atom in a substrate-
docking model exists in the alcohol binding pocket
like a chiral alcohol, which is a substrate with high
enantioselectivity in CAL-B-catalyzed kinetic resolu-
tions. CAL-B-catalyzed ring opening of the 4-bromo-
methyl-b-lactone with ethanol shows high enantiose-
lectivity (E>200) and produces optically pure ethyl
(R)-4-bromo-3-hydroxybutanoate. The unreacted (S)-
4-bromomethyl-b-lactone can be converted to ethyl
(S)-4-bromo-3-hydroxybutanoate through an acid-cat-
alyzed ring opening in ethanol and the consecutive re-
Synthesis of rac-4-Azidomethyl-b-lactone (rac-1c)
NaN3 (60 mg, 1.4 mmol) was added portionwise into a solu-
tion of rac-1a (1.2 mmol, 200 mg) in DMSO (5.0 mL) at
08C. After stirring for 4 h at room temperature, the reaction
mixture was quenched by H2O and extracted with CH2Cl2
(3ꢄ30 mL). The organic layer was washed with brine
(20 mL), dried over anhydrous MgSO4, and then evaporated
under reduced pressure. The crude product was purified by
silica gel flash column chromatography (EtOAc/hexane 1:4)
to afford 57 mg of rac-1c with a 50% yield. 1H NMR
(300 MHz, CDCl3): d=4.60 (m, 1H), 3.70 (q, J=3.5 Hz,
action with sodium cyanide produces a commonly 1H), 3.53 (dd, J=4.6, 3.6 Hz, 1H), 3.47 (q, J=5.9 Hz, 1H),
1814
asc.wiley-vch.de
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2013, 355, 1808 – 1816

Experimental and Computation Studies on Candida antarctica Lipase B
3.32 (q, J=4.2 Hz, 1H); 13C NMR (125 MHz, CDCl3): d=
Molecular Modeling
166.5, 68.3, 52.5, 40.6; IR (neat): nmax =3880, 3658, 2924,
2348, 2116, 1832 cmÀ1; MS (relative intensity): m/z=127
(M+1, 128), 71 (91), 54 (22), 43 (100), 42 (65), 41 (20), 40
(10); HR-MS (EI+): m/z=127.0380, calcd. for C4H5N3O2:
127.0382.
Computer modeling was performed using SYBYL 8.1
(Tripos). The X-ray crystal structure of CAL-B (pdb en-
try 1TCA) was obtained from the Protein Data Bank. The
enzyme structure containing the tetrahedral intermediate
form of the substrate was manually created in accordance
with the previous literature.[19] The Kollman All Atom force
field[20] was used with a distant dependent dielectric constant
and a non-bonded cut-off of 8 ꢁ for the protein. The param-
eters for the core structure of the tetrahedral intermediate
were obtained from the literature[16] and some missing pa-
rameters of the remaining atoms of the intermediate were
adapted from a general AMBER force field.[21] Partial
charges of the atoms of the enzyme and the tetrahedral
form of the substrate were calculated by the Pullman
method and the Mulliken method, respectively, with
a formal charge of À1 for the substrate oxyanion. Energy
minimization calculations were performed by the Powell
method in SYBYL and preceded until the rms derivatives
reached less than 0.005 kcalmolÀ1 ꢁÀ1.
Gram-Scale Syntheses of R-2c and S-2c by Enantio-
selective Ring Opening of rac-1a with EtOH by
CAL-B
Novozym 435 (CAL-B, 300 mg), rac-1a (2.0 g, 12 mmol),
and ethanol (553 mg, 12 mmol) were dissolved in tert-butyl
methyl ether (100 mL). The reaction mixture was shaken at
608C for 1 h. The mixture was filtered to remove Novozym
435. The filtrate was concentrated in vacuum, and then puri-
fied on a silica gel column chromatography (hexane/ethyl
acetate 1:1) to afford an (R)-ethyl 4-bromo-3-hydroxybut-
ACHTUNGTRENNUNG
anoate {R-2c, yield: 1.10 g, >99% ee, [a]2D5: +10.98 (c 1.0,
EtOH)} and (S)-4-bromomethyl-b-lactone (S-1a, yield:
0.94 g, >99% ee), i.e., 45% and 47% yields, respectively.
The unreacted S-1a (0.94 g, 5.7 mmol) was dissolved in
ethanol (20 mL) with p-toluenesulfonic acid·H2O (0.11 g,
0.57 mmol). The reaction mixture was stirred at 608C for
1 h, and then concentrated in vacuum. The residues were
purified on a silica gel column chromatography (hexane/
ethyl acetate 1:1) to afford an S-2c {yield: 1.3 g, >99% ee,
[a]2D5: À9.98 (c 1.0, EtOH)}, i.e., 99% yield.
Acknowledgements
This research was supported by the Korea Research Institute
of Chemical Technology (SI-1206) and by Basic Science Re-
search Program through the National Research Foundation
of Korea (NRF) funded by the Ministry of Education, Sci-
ence and Technology (2010-0022069). The authors thank
Prof. Geoffrey Horsman (Wilfrid Laurier University,
Canada) for useful comments.
CAL-B-Catalyzed Enantioselective Ring Opening of
rac-1b and rac-1c with Ethanol
4-Chloromethyl-b-lactone (rac-1b) (9.6 mg, 0.08 mmol), No-
vozym 435 (CAL-B, 2 mg) and EtOH (11 mg, 0.24 mmol)
were added in tert-butyl methyl ether (1 mL). The mixture References
was shaken at 30 and 608C, and the progress of the reaction
and the ee values were monitored through analyzing the
samples at particular intervals by GC until the reaction pro-
ceeded to 50% conversion. Then the mixture was filtered to
remove Novozym 435. The residue was analyzed by GC. GC
analyses were conducted with an Agilent 6890N equipped
with a flame ionization detector using lipodex-A (30 mꢄ
0.25 mm I.D., MN). The injector and detector temperature
were 2008C and 2208C, respectively. The GC column tem-
perature was initially programmed at 508C for 40 min, then
gradient to 808C at 18C·minÀ1, and then held for 30 min. S-
1b and R-1b were detected at 47.6 min and 49.1 min, respec-
tively. S-2e and R-2e were detected at 71.4 min and
72.3 min, respectively.
The ring opening of 4-azidomethyl-b-lactone (rac-1c,
10 mg, 0.08 mmol) with ethanol (11 mg, 0.24 mmol) by No-
vozym 435 (2 mg) at 608C for 1 h was conducted and ana-
lyzed in the same manner as for rac-1b. GC analyses were
conducted with an Agilent 6890N equipped with a flame
ionization detector using Gamma Cyclodextrin trifluoroace-
tyl (30 mꢄ0.25 mm I.D., Astec). The injector and detector
temperature were 2008C and 2208C, respectively. The GC
column temperature was initially programmed at 608C for
2 min, then gradient to 1008C at 108C·minÀ1, and then held
for 15 min. S-1c and R-1c were detected at 41.3 min and
43.8 min, respectively. S-2f and R-2f were detected at
37.1 min and 38.3 min, respectively.
[1] C. H. Wong, D. G. Drueckhammer, H. M. Sweers, J.
Am. Chem. Soc. 1985, 107, 4208–4231.
[2] S. K. Ma, J. Gruber, C. Davis, L. Newman, D. Gray, A.
Wang, J. Grate, G. W. Huisman, R. A. Sheldon, Green
Chem. 2010, 12, 81–86.
[3] A. Endo, J. Lipid Res. 1992, 33, 1569–1582.
[4] A. Lilkeblad, A. Kallinen, L. T. Kanerva, Curr. Org.
Synth. 2009, 6, 362–379.
[5] M. Mꢅller, Angew. Chem. 2005, 117, 366–369; Angew.
Chem. Int. Ed. 2005, 44, 362–365.
[6] S. Panke, M. Wubbolts, Curr. Opin. Chem. Biol. 2005,
9, 188–194.
[7] a) Y. Yasohara, N. Kizaki, J. Hasegawa, M. Wada, M.
Kataoka, S. Shimizu, Tetrahedron: Asymmetry 2001, 12,
1713–1718; b) N. Kizaki, Y. Yasohara, J. Hasegawa, M.
Wada, M. Kataoka, S. Shimizu, Appl. Microbiol. Bio-
technol. 2001, 55, 590–595; c) M. Kataoka, L. P. S.
Rohani, K. Yamamoto, M. Wada, H. Kawabata, K.
Kita, H. Yanase, S. Shimizu, Appl. Microbiol. Biotech-
nol. 1997, 48, 699–703.
[8] S. C. Davis, J. H. Grate, D. R. Gray, J. M. Gruber, G. W.
Huisman, S. K. Ma, L. M. Newman, R. Sheldon, L. A.
Wang, (Codexis Inc.), Patent WO 2004/015132, 2004.
[9] a) G. DeSantis, K. Wong, B. Farwell, K. Chatman, Z.
Zhu, G. Tomlinson, H. Huang, X. Tan, L. Bibbs, P.
Chen, K. Kretz, M. J. Burk, J. Am. Chem. Soc. 2003,
Adv. Synth. Catal. 2013, 355, 1808 – 1816
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1815

Jung Yun Lim et al.
FULL PAPERS
125, 11476–11477; b) G. DeSantis, Z. Zhu, W. A.
Greenberg, K. Wong, J. Chaplin, S. R. Hanson, B. Far-
well, L. W. Nicholson, C. L. Rand, D. P. Weiner, D. E.
Robertson, J. Burk, J. Am. Chem. Soc. 2002, 124, 9024–
9025; c) see also: S. J. Maddrell, N. J. Turner, A. Ker-
ridge, A. J. Willetts, J. Crosby, Tetrahedron Lett. 1996,
37, 6001–6004.
[15] J. A. Lꢇpez-Lꢇpez, F. M. Guerra, F. J. Moreno-Dorado,
Z. D. Jorge, G. M. Massanet, Tetrahedron Lett. 2007,
48, 1749–1752.
[16] F. Hæffner, T. Norin, K. Hult, Biophys. J. 1998, 74,
1251–1262.
[17] W. V. Tuomi, R. J. Kazlauskas, J. Org. Chem. 1999, 64,
2638–2647.
[10] U. T. Bornscheuer, R. J. Kazlauskas, Hydrolases in Or-
ganic Synthesis, 2nd edn. Wiley-VCH, Weinheim, 2006.
[11] E. Garcꢆa-Urdiales, F. Rebolledo, V. Gotor, Tetrahe-
dron: Asymmetry 1999, 10, 721–726.
[18] S. Park, E. Forrꢇ, H. Grewal, F. Fꢅlçp, R. J. Kazlauskas,
Adv. Synth. Catal. 2003, 345, 986–995.
[19] S. Raza, L. Fransson, K. Hult, Protein Sci. 2001, 10,
329–338.
[12] a) Y. Koichi, K. Suginaka, Y. Yamamoto, J. Chem. Soc.
Perkin Trans. 1 1995, 1645–1646; b) W. Adam, P. Groer,
C. R. Saha-Mçller, Tetrahedron: Asymmetry 1997, 8,
833–883; c) S. G. Nelson, K. L. Spencer, J. Org. Chem.
2000, 65, 1227–1230; d) D. Klomp, J. A. Peters, U. Ha-
nefeld, Tetrahedron: Asymmetry 2005, 16, 3892–3896.
[13] W. E. Barnett, J. C. Mckenna, Tetrahedron Lett. 1971,
12, 2595–2598.
[20] a) S. J. Weiner, P. A. Kollman, D. A. Case, U. C. Singh,
C. Ghio, G. Alagona, S. Profeta, P. Weiner, J. Am.
Chem. Soc. 1984, 106, 765–784; b) S. J. Weiner, P. A.
Kollman, D. T. Nguyen, D. A. Case, J. Comput. Chem.
1986, 7, 230–252.
[21] J. Wang, W. Wang, P. A. Kollman, J. Mol. Graphics
Modell. 2006, 25, 247–260.
[22] C. S. Chen, Y. Fujimoto, G. Girdaukas, C. J. Sih, J. Am.
Chem. Soc. 1982, 104, 7294–7299.
[14] P. Kumar, A. N. Deshmukh, R. K. Upadhyay, M. K.
Gurjar, Tetrahedron: Asymmetry 2005, 16, 2717–2721.
1816
asc.wiley-vch.de
ꢂ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2013, 355, 1808 – 1816