Preparation of the enantiomers of 1-phenylethan-1,2-diol. Regio- and enantioselectivity of acylase I and Candida antarctica lipases A and B

Article abstract of DOI:10.1016/S0957-4166(01)00428-1

Acylase I and Candida antarctica lipases A (CAL-A) and B (CAL-B) were evaluated for the preparation of the enantiomers of 1-phenylethan-1,2-diol. In the presence of CAL-B, the sequential one-pot methanolysis of the diacetate in acetonitrile allowed the preparation of (S)-diol (e.e. 97%) and (R)-1-acetoxy-1-phenylethanol (e.e. 94%). Base-catalyzed methanolysis of the monoacetate resulted in the corresponding (R)-diol. When one of the diol enantiomers was subjected to Mitsunobu esterification, inversion of configuration occurred, allowing transformation of the initially racemic mixture to one enantiomer. Acylase I-catalysis led to the chemo- and enantioselective formation of (S)-1-acetoxy-1-phenylethanol (e.e. 97%) in the presence of the primary hydroxyl function through acetylation of the secondary hydroxyl group. The low chemical yield (ca. 25%) was due to the moderate enzymatic regioselectivity. CAL-A behaved in a similar way to acylase I.

Full text of DOI:10.1016/S0957-4166(01)00428-1

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 12 (2001) 2447–2455
Preparation of the enantiomers of 1-phenylethan-1,2-diol. Regio-
and enantioselectivity of acylase I and Candida antarctica lipases
A and B
Pauliina Virsu, Arto Liljeblad, Anu Kanerva and Liisa T. Kanerva*
Laboratory of Synthetic Drug Chemistry and Department of Chemistry, University of Turku, Lemminka¨isenkatu 2,
FIN-20520 Turku, Finland
Received 31 August 2001; accepted 2 October 2001
Abstract—Acylase I and Candida antarctica lipases A (CAL-A) and B (CAL-B) were evaluated for the preparation of the
enantiomers of 1-phenylethan-1,2-diol. In the presence of CAL-B, the sequential one-pot methanolysis of the diacetate in
acetonitrile allowed the preparation of (S)-diol (e.e. 97%) and (R)-1-acetoxy-1-phenylethanol (e.e. 94%). Base-catalyzed methanol-
ysis of the monoacetate resulted in the corresponding (R)-diol. When one of the diol enantiomers was subjected to Mitsunobu
esterification, inversion of configuration occurred, allowing transformation of the initially racemic mixture to one enantiomer.
Acylase I-catalysis led to the chemo- and enantioselective formation of (S)-1-acetoxy-1-phenylethanol (e.e. 97%) in the presence
of the primary hydroxyl function through acetylation of the secondary hydroxyl group. The low chemical yield (ca. 25%) was due
to the moderate enzymatic regioselectivity. CAL-A behaved in a similar way to acylase I. © 2001 Elsevier Science Ltd. All rights
reserved.
1. Introduction
very poor enantioselectivity although the substrate
structure and the nature of a lipase can be critical in
this respect. The role of a lipase is clearly seen for the
methanolysis of dibutanoylated 1-O-trityl glycerol.10
Thus, in the presence of Pseudomonas cepacia (lipase
PS) and fluorescens lipases (lipase AK) the highly regio-
and enantioselective (E>100; E is the enantiomer
ratio11) reaction in diisopropyl ether gave the (R)-2-
monoacylated product at 50% conversion and the reac-
1,2-Diols are important intermediates and building
blocks for various synthetic applications. Thus, the
enantiomers of 1-phenylethan-1,2-diol 1 are useful for
the synthesis of chiral catalysts,1,2 macrocyclic
polyether–diester ligands,3 pharmaceutically active
compounds4 and liquid crystals.5 For the preparation of
enantiomeric alcohols, the lipase (EC 3.1.1.3)-catalyzed
kinetic resolution through transesterification in an
appropriate organic solvent is among the most power-
ful procedures. Such resolution methods afford the two
enantiomers in 50% conversion at best.
tion stopped. When Candida antarctica lipase
B
(CAL-B) was used as the catalyst, the initially formed
2-monoacylated product was racemic and further reac-
tion resulted in (R)-1-O-tritylglycerol together with the
less reactive (S)-2-monobutanoylated counterpart. The
latter type of reaction was previously called one-pot
sequential resolution and was successfully used for the
lipase PS-catalyzed acylation of aryl-substituted 1,2-
diols in organic solvents.7–9,12
The biological function of lipases is to catalyze the
hydrolysis of triacylglycerols at water–lipid interfaces.
In these reactions, lipases are typically 1,3-regioselec-
tive. On this basis, it is hardly a surprise that the
primary hydroxyl function reacts favourably in the
lipase-catalyzed acylation of 1,2-diols in organic
media.6–9 Similarly, in the alcoholysis of the corre-
sponding diesters a primary alcohol is first liberated in
a highly regioselective manner.7,10 These reactions,
occurring far from the stereocentre, often proceed with
The use of separately monoprotected 1,2-diols as sub-
strates is a chemical means of forcing an enzymatic
reaction to occur at a stereocentre rather than else-
where in a molecule. Accordingly, highly enantioselec-
tive acylations of 1-phenoxy-2-alkanols with activated
butanoates were previously performed in the presence
of CAL-B.13 On the other hand, the use of protecting
* Corresponding author. E-mail: lkanerva@utu.fi
0957-4166/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(01)00428-1

2448
P. Virsu et al. / Tetrahedron: Asymmetry 12 (2001) 2447–2455
were screened using 1-phenylethanol and the corre-
sponding esters as model substrates.
groups increases the number of synthetic steps and for
that reason is usually avoided when other possibilities
are at hand.
2. Results and discussion
The third strategy for the resolution of 1,2-diols (or
their diesters) that has not been described in the litera-
ture is to use an enzyme that is 2- rather than 1,3-
regioselective towards triacylglycerols, leading to a
reaction at the stereocentre rather than remote from it.
Interestingly, C. antarctica lipase A (CAL-A) was previ-
ously shown to exert clear preference for the 2-position
when triacylglycerols were hydrolyzed.14
The enzymatic acetylation of 1 and alcoholysis of 4
were studied paying attention to regio- and enantiose-
lectivities. The acetylation of 1 to product 4 [or alcohol-
ysis of 4 to product 1] can proceed through two
different monoacetylated products 2 and 3 (Scheme 1).
The reaction is said to be from low to highly regioselec-
tive when it proceeds from equal to different propor-
tions of the monoacetylated products via steps A and C
(or B and D). The progression curves with time for the
various products 1–4 in the reaction mixture are shown
in Fig. 1(a–c) for the acetylation of 1 and in Fig. 1(d–f)
for the alcoholysis of 4. Enantiodiscriminations (which
enantiomer reacts faster) for the sequential steps A–D
by different enzymes are shown in Table 1. It is notable
that in the case of the present enzymes, the regioselec-
tivity steps A and C for acetylation and B and D for
alcoholysis proceed with opposite enantiodiscrimina-
tions except when one of the steps leads to a racemic
product. Accordingly, the enantiomer that reacts faster
depends on enzymatic regioselectivity.
In the work described herein, CAL-A, CAL-B and
acylase I were evaluated for the preparation of the
enantiomers of 1-phenylethan-1,2-diol without the use
of protecting groups. The purpose was to exploit
sequential resolution in one-pot through the acetylation
of 1 in vinyl acetate and through the alcoholysis of the
corresponding diacetate 4 in butanol or in acetonitrile
(or some other organic solvent) containing methanol
(Scheme 1). Using this strategy, the aim was to obtain
the two enantiomers of the diol simultaneously. The
possibility of transforming one of the enantiomers to its
antipode through the Mitsunobu reaction was also
studied.15 A different strategy was to study enzymes
that allow an enantioselective reaction at the free or
acetylated secondary hydroxyl group of 1 and 4, respec-
tively, in the presence of the primary one. Monoesters 2
and 3 were chemically prepared and separately sub-
jected under enzymatic reaction conditions in order to
study which of the enantiomers reacts faster to give 1
upon alcoholysis and 4 upon acetylation. Lipases were
chosen on the basis of their previous behaviors.10,14 On
the other hand, acylase I (N-acylamino acid amidohy-
drolase, EC 3.5.1.14) represents an unexplored possibil-
ity in catalyzing the resolution of 1,2-diols. Acylase I
was previously shown to be useful for the acylation of
primary and secondary alcohols as well as for the
alcoholysis of carboxylic acid esters as novel reac-
tions.16–21 Commercial enzymes from Aspergillus species
2.1. Acylase I catalysis
Acylase I is well characterized as a Zn(II)-containing
enzyme (1–3 zinc ions per subunit).22,23 The exact struc-
ture of the active site is not known and the reaction
mechanism is simply proposed on the basis of the
mechanisms of Zn(II)-containing carboxypeptidases.24
The reaction proceeds without the formation of a cova-
lent acyl-enzyme intermediate (the formation of which
is typical of lipases). The enzyme is relatively stable in
polar and hydrophilic solvents such as alcohols.20,21
For acylase I-catalyzed transesterifications, the previous
results indicated higher enantioselectivity for the
acylation of secondary rather than primary alcohols.16–19
On this basis, commercial acylase I enzymes were first
screened for the acetylation of 1-phenylethanol in vinyl
acetate and for the butanolysis of 1-phenylethyl acetate
and butanoate in the neat alcohol as model reactions.
The results are shown in Table 2. Acetylation proceeds
considerably faster than butanolysis, the (R)-enan-
tiomer always reacting faster than the (S)-isomer.
Somewhat higher enantioselectivity is observed for the
butanolysis of 1-phenylethyl acetate than for that of the
butanoate. The enzyme on Eupergit C is practically
ineffective. Clearly, acylase I from Aspergillus melleus is
the best for acetylation, while the enzyme from A. genus
is most usable for alcoholysis and the enzymes were,
respectively, used in the present work.
A
OH
OH
R1OH
Ph OH
1
Ph OCOMe
3
AcOCH=CH2
C
B
AcOCH=CH2
AcOCH=CH2
R1OH
R1OH
OCOMe
OCOMe
AcOCH=CH2
The acylase I (A. melleus)-catalyzed acetylation of 1
with vinyl acetate favors the formation of 2 (ꢀ)
through step C over the formation of 3 (ꢁ) through
step A (Scheme 1; Fig. 1(a)). Interestingly, step A is
highly enantioselective, allowing the preparation of
monoester (S)-3 in 97% e.e. and ca. 25% chemical yield.
Ph OH
Ph OCOMe
4
R1OH
D
2
Scheme 1.

P. Virsu et al. / Tetrahedron: Asymmetry 12 (2001) 2447–2455
2449
Figure 1. Progression curves with time for the acetylation of 1 in vinyl acetate in the presence of (a) acylase I (A. melleus), (b)
CAL-A and (c) CAL-B and for the alcoholysis of 4 in butanol in the presence of (d) acylase I (A. genus) and with methanol (0.8
M) in acetonitrile (e) CAL-A and (f) CAL-B. (ꢂ) for 1, (ꢀ) for 2, (ꢁ) for 3 and (") for 4; enantiomeric excess (e.e./%) for highly
enantioselective cases.
Table 1. Reactive enantiomer of steps A–D for the acetylation of 1 and for the alcoholysis of 4 by acylase I (A. melleus and
genus), CAL-A and CAL-B
Step
Acetylation
CAL-A
Alcoholysis
CAL-A
Acylase I
CAL-B
Acylase I
CAL-B
A
B
C
D
S
S
S
S (E=3)b
R
S (E=2)b
R/S
S (Eꢀ100)b
R
R (E=2)
(R/S)
R (E=20)a
R
S (E=2)a
R
R (E=7)a
R/S
R/S (E=1)b
R/S
R (E=6)b
R
R/S (E=1)a
R/S (E=1)a
S (E=35)
a E for the enzymatic acetylation of chemically prepared 2 and 3 in vinyl acetate.
b E for the enzymatic alcoholysis of chemically prepared 2 in 3 in butanol or in acetonitrile containing methanol (0.8 M).

2450
P. Virsu et al. / Tetrahedron: Asymmetry 12 (2001) 2447–2455
Table 2. Acylase I-catalyzed acetylation of 1-phenylethanol in vinyl acetate and butanolysis of 1-phenylethyl acetate/1-
phenylethyl butanoate in butanol at room temperature (25°C)a
Acetylation
Butanolysis
Time (h)
Conversion (%)
E
Time (h)
Conversion (%)
E
Aspergillus melleus
Aspergillus genus
Aspergillus on Eupergit C
26
26
26
50
51
5
107
61
3
53/115
53/115
20/115
13/42
27/52
0/18
72/51
146/129
–/15
a 75 mg/mL of the enzyme preparation.
The formation of the (S)-enantiomer is in accordance
with the identical steric demands of (S)-3 and (R)-1-
phenylethanol. Step C is slightly (R)-enantioselective.
The formation of (S)-4 (") as the result of sequential
reactions with e.e. between 20 and 50% (depending on
the stage of the reaction) needs more careful consider-
ation. Namely, the enzymatic acetylation of chemically
prepared ( )-2 proceeds slowly (10% conversion after
75 h) to ( )-4, while the acetylation of ( )-3 (50%
conversion at 24 h) leads to (R)-4 with E=20 (Table 1).
On this basis, the acetylation of 1 by acylase I is easily
expected to lead to the formation of (R)-4 rather than
to the observed product with the (S) absolute configu-
ration. As an explanation, the high concentration
excess of (S)-3 and the predominance of the sequential
step B over D together result in the observed stereo-
chemical outcome.
of (S)-3 with 80% e.e. in ca. 35% chemical yield. The
formation of (R)-2 in step C is less enantioselective,
proceeding at an e.e. of around 50% at every stage of
the reaction. In order to study the nature of the sequen-
tial steps B and D, chemically prepared monoesters 2
and 3 were separately subjected under the enzymatic
reaction conditions, resulting in 4 with E=1 (77%
conversion in 2 h) and E=2 (67% conversion in 2
hours), respectively. Thus, the formation of enan-
tiomerically enriched diacetate (S)-4 (") (e.e. 25% after
3 h reaction) is explained by the opposite enan-
tiodiscriminations of steps A and C together with low
enantioselectivities of the almost equally reactive
sequential steps B and D, respectively.
The CAL-A-catalyzed sequential one-pot methanolysis
of diacetate 4 to diol 1 in acetonitrile proceeds slowly
and with enantioselectivities which are too low to be of
practical value (Fig. 1(e)). It is interesting to note that
the reaction at the stereocentre in the formation of
monoester 2 (ꢀ; step D) is favored over the formation
of 3 (ꢁ; step B). The CAL-A-catalyzed methanolysis of
chemically prepared monoesters 2 and 3 reveal that diol
1 is mainly obtained when product 2 reacts further
through the slightly (R)-enantioselective step C (Table
1).
The acylase I (A. genus)-catalyzed sequential one-pot
butanolysis of diacetate 4 to diol 1 in neat butanol
proceeds too slowly to be of practical value (Fig. 1(d)).
Regioselectivity is moderate, favoring the formation of
3 (ꢁ) through step B in an (R)-enantioselective reac-
tion. The sequential step A was shown to be slightly (S)
enantioselective (Table 1). Similar to the acetylation
case above, the concentration effect of (R)-3 evidently
leads to the formation of slightly enantiomerically
enriched (R)-1 as the final product. For the butanolysis
of 4, the enzyme from A. melleus behaves in the same
way, except that the reactions are even slower as
expected according to the screening results in Table 2.
Noteworthy of the CAL-A-catalyzed reactions is that
the same enantiodiscrimination is observed in the
sequential steps A and B as well as C and D (or a
product is racemic), although the regiodiscrimination
steps A and C as well as B and D are of the opposite
enantiodiscrimination (Table 1).
2.2. CAL-A catalysis
CAL-A is a calcium dependent, thermostable lipase
that is reported to be highly active in a non-specific
manner.25,26 The enzyme is only rarely used in enan-
tioselective reactions. In the present work it was studied
because of its unique sn-2 preference towards triac-
ylglycerols and of the highly enantioselective acylation
of the amino groups in aliphatic b-substituted b-amino
esters.14,27
2.3. CAL-B catalysis
The properties and the wide application area of CAL-B
as a regio- and enantioselective catalyst in organic
synthesis are well described.25,26,28,29 The enzyme is
calcium independent and works without interfacial acti-
vation. It is also active in many hydrophilic solvents
and tolerates high concentrations of various alkan-1-ols
including methanol.10
The CAL-A-catalyzed acetylation of 1 in vinyl acetate
proceeds smoothly, leading to the disappearance of the
diol in 3–4 h (ꢂ, Fig. 1(b)). The enzyme is of low
regioselectivity as product 2 (ꢀ) is obtained in addition
to the expected monoester 3 (ꢁ),14 the progression
curves being practically one on the other. Interestingly,
the enzyme is highly (S) enantioselective for the reac-
tion through step A (Table 1), enabling the formation
The CAL-B-catalyzed acetylation of 1 in vinyl acetate
leads to the rapid consumption of the diol (ꢂ, step C)
and to the formation of practically racemic monoac-
etate 2 (ꢀ) in a highly regioselective manner through
step C (Fig. 1(c)). This reaction is followed by sequen-
tial acetylation through step D, allowing the formation
of (S)-4 (") at around 90% e.e. in over 40% yield. The

P. Virsu et al. / Tetrahedron: Asymmetry 12 (2001) 2447–2455
2451
Lipase-catalyzed reactions proceed through an acyl-
enzyme intermediate. In this intermediate, an ester bond
covalently binds the serine residue of the catalytic triad
to the carbonyl carbon of an acyl donor at the active site
of the enzyme.28 Typical of neutral and basic ester
hydrolysis mechanisms, an anionic tetrahedral intermedi-
ate precedes the formation and follows the decomposi-
tion of the ester bond. Substrate/active site models for
the anionic tetrahedral intermediates of the faster react-
ing enantiomers where the secondary (I) and primary (II)
hydroxyl groups of the present substrates are bound to
the carbonyl carbon during acetylation and methanolysis
are given in Scheme 2 (R=H or Ac). The opposite
enantiodiscriminations for the sequential steps A and B
as well as for C and D can now be attributed to steric
requirements for correct orientation of an alcohol (or an
ester) in the three-dimensional space of the active site of
CAL-B. Scheme 2 is a modification according to the
previous presentation for 1-O-alkyl-2-alkanols in the
active site cleft of CAL-B.13
enantioselectivity of step D is in accordance with E=35,
which was calculated for the CAL-B-catalyzed acetyla-
tion of chemically prepared 2 in vinyl acetate (Table 1).
In spite of the high regioselectivity, traces [5% relative
proportion for (S)-3 (e.e. 40%) after 15 min] of mono-
acetate 3 were detected in the very beginning of the
reaction before 1 was consumed in the fast step C.
Evidently, the traces smoothly disappear through the
sequential step B, explaining the initially moderate e.e.
values (71 and 86%) for (S)-4. This conclusion is sup-
ported by the fact that chemically prepared monoester
3 was totally acetylated with vinyl acetate and CAL-B
in an (R)-enantioselective reaction (E=7) in 1 day (Table
1).
The CAL-B-catalyzed methanolysis of 4 (") in acetoni-
trile first proceeds smoothly to ca. 60% conversion. The
reaction slows down when the (R)-enantiomer has more
or less reacted to monoacetate 3 (ꢁ) through step B (Fig.
1(f)). E=15 can be approximated for this reaction using
the GLC data that are collected before the first signs of
1 appear. The highly (S)-enantioselective methanolysis of
3 in step A finally drives the sequential resolution to the
end in the formation of (S)-1 (ꢂ) at 97% e.e. and (R)-3
(ꢁ) at 95% e.e. both close to 50% chemical yields (Table
3, row 1). The CAL-B-catalyzed methanolysis of chem-
ically prepared monoesters 3 and 2 in acetonitrile led to
(S)-1 with Eꢀ100 (conversion 15% after 2 h) through
step A and (R)-1 with E=2 (conversion 80% after 2 h)
through step C (Table 1), respectively. These results
further evidence the favor of highly regioselective forma-
tion of 3 over 2 in the methanolysis of 4.
As shown in Fig. 1(f) and Table 3, small amounts (ca.
2%) of 2 (ꢀ) are detected throughout the course of the
sequential methanolysis of 4 although the reaction is
highly regioselective and as stated before the methanol-
ysis of 2 produces diol 1 smoothly (80% conversion in
2 h). Acyl migration is proposed as an explanation.
Intramolecular acyl migration in monoacylated products
can take place while heating and in the presence of acids
and bases, the transformation from 3 to 2 being energet-
ically more favorable.30,31 For GLC analysis, the deriva-
tization of free hydroxyl groups in a sample with
Table 3. Effect of methanol concentration on the CAL-B-catalyzed methanolysis of 4 in acetonitrile at room temperature:
relative proportion/e.e./% for compounds 1–4
Row
MeOH (M)
Time (h)
1
2
3
4
1
2
3
4
5
0.8a
0.8b
0.8c
1.2a
1.2b
41
89
79
79
110
44/97
44/96
43/96
46/96
44/97
2/34
2/26
3/7
2/24
2/13
53/95
53/94
51/90
51/96
53/94
1/17
1/8
3/4
1/10
1/39
a Small-scale.
b Gram-scale.
c BuOH in the place of MeOH.
Large group
Ph
Large group
Ph
II
I
Stereospecificity
pocket
Stereospecificity
pocket
H
H
O
OR
OR
O
Trp 104
Trp 104
O
O-
O
O-
His 224
His 224
105 Ser
105 Ser
Oxyanion
hole
Oxyanion
hole
Scheme 2.

2452
P. Virsu et al. / Tetrahedron: Asymmetry 12 (2001) 2447–2455
trifluoroacetic anhydride (liberates trifluoroacetic acid)
can expose 3 to acyl migration. Accordingly, it was
possible to show that the sample of (R)-3 (0.1 M, 200
mL, e.e. 94%) was transformed to the 3/2 mixture of
(R)-3 (e.e. 45%)/(R)-2 (e.e. 18%) when treated with
trifluoroacetic acid (20 mL) for 5 h before the anhy-
dride (50 mL) was added for derivatization. This result
shows that considerable racemization is accompanied
by acyl migration. When trifluoroacetic anhydride was
added before the addition of trifluoroacetic acid, nor-
mal 2% of 2 was detected by GLC and the racemiza-
tion of derivatized (R)-3 was not observed after acid
treatment overnight. Separation on silica may also
expose 3 to slightly acidic conditions. Accordingly,
only 2% of 2 in the resolution mixture (1.2 M MeOH;
Table 3, row 5) increased to 5% after separation of the
products by column chromatography (as described in
Section 4.3).
Ester alcoholysis in organic solvents most probably stops
at an equilibrium conversion, leading to the gradual
racemization of the less reactive enantiomer when the
product enantiomer converts back to the starting mate-
rial. The reaction conditions for the methanolysis of 4
were chosen according to the previous results, where 0.8
M methanol in acetonitrile was enough to prevent the
effects of equilibria.10 It seems that the reaction of (S)-3
to (S)-1 does not quite go to completion with 0.8 M
alcohol solutions of the present work (Table 3, rows 1–3).
Thus, the methanolysis of 4 was repeated using 1.2 M
methanol in acetonitrile (rows 4 and 5). Although the
reaction takes more time with increased methanol concen-
tration and there is no clear difference in the results, the
gram-scale resolution of 4 was performed using both
methanol concentrations and the resolution in the pres-
ence of 1.2 M methanol is described in Section 4.3. For
unknown reasons, more time was needed for the comple-
tion of the reaction when a small-scale reaction (rows 1
and 4) was changed to a gram-scale volume (rows 2 and
5).
2.3.1. Optimization for gram-scale resolution. The
CAL-B-catalyzed methanolysis of 4 in acetonitrile
allows the preparation of both enantiomers of 1 simul-
taneously through sequential resolution in one-pot
(Schemes 1 and 3). In this reaction, the first step B is
moderately (R)-enantioselective (approximated E=15)
and thus the wrong enantiomer is initially in excess for
the reaction through step A. Thus, the time needed for
the transformation of 4 to the mixture of (S)-1 and
(R)-3 would be shorter if step B were (S)-selective or
non-enantioselective. Solvent effects were studied as a
means of decreasing the reaction time. Tetra-
hydrofuran or toluene in the place of acetonitrile
retarded the formation of diol (S)-1 through step A
(only 10% formed in 41 h) without affecting the reac-
tion through step B. In diisopropyl ether, the reaction
behaved almost identically to that in acetonitrile. The
use of butanol in the place of methanol as a nucle-
ophile in acetonitrile caused a drop in E from 15 to 10
for the reaction through step B without the desired
effect on reaction time (Table 3, row 3). On this basis,
methanolysis in acetonitrile was chosen as a method
for the present work.
2.4. Preparation of (R)-3 and (S)-3; inversion of
configuration via Mitsunobu esterification
The disadvantage of conventional resolution is the 50%
limit for the chemical yield of one enantiomer without
special arrangements. Resolution followed by Mitsunobu
esterification of a free hydroxyl group in one of the
enantiomers was previously shown to lead to inversion
at the stereocentre and accordingly to give yield enhance-
ment for the other enantiomer.15 Under Mitsunobu
conditions, 1,2-diols undergo predominant ester forma-
tion at the secondary centre due to the formation of a
phosphorane intermediate, which opens via protonation
of the least hindered oxygen.32 When commercial (R)-1
(e.e. 98%) was acetylated under Mitsunobu conditions the
product was the corresponding monoester (S)-3 with 98%
e.e., indicating that the resolution–inversion method is
excellent for the transformation of (S)-1 to (R)-3. To this
end, (S)-1 with 97% e.e. was subjected to Mitsunobu
conditions resulting in (R)-3 with 95% e.e. as was
expected.
Ph3P
AcOH
DEAD
MeOH in
OCOMe
OH
OH
OH
OCOMe
MeCN
+
Ph
(S)-1
Ph
Ph OCOMe
CAL-B
4
(R)-3
Amberlyst
IRA-401
MeOH
Amberlyst
IRA-401
MeOH
Ph3P
AcOH
DEAD
OH
OH
Ph
OCOMe
Ph
OH
(S)-3
(R)-1
Scheme 3.

P. Virsu et al. / Tetrahedron: Asymmetry 12 (2001) 2447–2455
2453
In order to exploit the Mitsunobu reaction to give the
(R)-enantiomer, (R)-3 must be first transformed to the
diol (R)-1 (Scheme 3). The base-catalyzed deacylation
of 1,2-diol esters has proved its applicability in various
connections.6–9 As described in Section 4.4, the method
was also successfully used in the present work.
many) and Tokyo Kasei (Tokyo, Japan), respectively.
Acylase I from Aspergillus on Eupergit C (388.4 U/g) was
a product from Fluka (Buchs, Switzerland). The origin
of the immobilized enzyme is not reported more accu-
rately by the producer. Immobilized C. antarctica lipase
B (CAL-B, Chirazyme L5) and C. antarctica lipase A
(CAL-A, Chirazyme L5) were the products of Boehringer
Mannheim. Before use, CAL-A was immobilized on
Celite in the presence of sucrose (20% lipase on Celite).33
1-Phenylethanol, 1-phenylethan-1,2-diol 1 (racemic and
the pure (R)-enantiomer), vinyl acetate and tri-
fluoroacetic anhydride (used as derivatization reagent)
were products of Aldrich. 1-Phenylethyl acetate and
butanoate were from the previous work.15 The solvents
were of the highest analytical grade and obtained from
Lab Scan Ltd., Riedel-de Hae¨n and J.T. Baker.
3. Conclusions
The regio- and enantioselectivities of acylase I, CAL-A
and CAL-B were evaluated in order to study their
applicability to the preparation of the enantiomers of
1-phenylethan-1,2-diol 1. To this end, 1 was subjected to
acetylation in vinyl acetate and the corresponding diac-
etate 4 for alcoholysis in butanol or in acetonitrile
containing methanol (Scheme 1). For the present
enzymes, the enantiomer that reacts faster was shown to
depend on enzymatic regioselectivity, i.e. the formation
of monoester 2 was shown to proceed with opposite
enantiodiscrimination to the formation of monoester 3
(Table 1). In the case of CAL-B catalysis, the substrate/
active site models for the anionic tetrahedral intermedi-
ates that precede the formation and follow the
decomposition of an acyl-enzyme intermediate were used
to explain the opposite enantiodiscriminations (if not one
of the steps is racemic) for the sequential steps.
4.1.1. Preparation of ( )-1,2-diacetoxy-1-phenylethane 4.
1-Phenylethan-1,2-diol 1 (1.43 g, 10.3 mmol) was dis-
solved in dichloromethane (25 mL) with triethylamine
(4.30 mL, 31.0 mmol), DMAP (4-N,N-dimethyl-
aminopyridine; 63.2 mg, 0.52 mmol) and acetic anhy-
dride (2.94 mL, 31.0 mmol). The mixture was stirred at
room temperature (25°C) for 24 h. The addition of
methanol (10 mL) stopped the reaction. After evapora-
tion in vacuo the product was purified by column
chromatography on silica using acetone:petroleum ether
(3:7) as an eluent, yielding 4 (2.21 g, 9.9 mmol). 1H NMR:
l (ppm) 2.03 (s, 3H, CH3CO2CH2), 2.09 (s, 3H,
CH3CO2CH), 4.29 (m, 2H, CH2), 5.99 (dd, 1H, CH),
7.27–7.38 (m, 5H, Ar-H). 13C NMR l (ppm) 20.7
(CH3CO2CH2), 21.0 (CH3CO2CH), 66.0 (CH2), 73.2
(CH), 126.6, 128.6 and 136.4 (6C, Ar-C), 170.0
(CO2CH2), 170.6 (CO2CH). Mass spectrum: M+=222.
Excellent regioselectivity towards the reaction at the
primary alcohol (acetylation, step C) or at its ester
(methanolysis, step B) functions is obvious for CAL-B
catalysis, the respective sequential steps D and A being
highly enantioselective (Scheme 1, Fig. 1(c and f)).
Accordingly, the sequential one-pot resolution of 4 with
methanol in acetonitrile allowed the preparation of (S)-1
(e.e. 97%) and (R)-3 (e.e. 94%) on gram-scale (Scheme
3, Table 3). (R)-3 was further transformed to (R)-1 via
base-catalyzed methanolysis. When only one of the
enantiomers of 1 is needed, the Mitsunobu acetylation at
the secondary hydroxyl group of (R)-1 or (S)-1 was
shown to allow the transformation of ( )-1 to one
enantiomer.
4.1.2. Preparation of ( )-1-acetoxy-1-phenylethanol 3.
Prepared by Mitsunobu reaction:15 diol 1 (2.16 g, 15.6
mmol) was dissolved in tetrahydrofuran (30 mL) and
triphenylphosphine (4.09 g, 15.6 mmol) and acetic acid
(0.89 mL, 15.6 mmol) were added followed by the
dropwise addition of DEAD (diethyl azodicarboxylate;
2.43 mL, 15.6 mmol). The mixture was stirred at room
temperature (25°C) for 21 h before the work-up as above,
yielding 3 (2.08 g, 11.5 mmol) containing 10% of ( )-2
Due to low regioselectivities, CAL-A and acylase I
catalyze the formation of both monoesters 2 and 3, the
formation of 3 proceeding in a highly enantioselective
manner (Fig. 1(a, b, d and e)). Thus, for the acetylation
of 1, the preparation of (S)-3 with 97% e.e. (25%
theoretical yield) in the case of acylase I and with 80%
e.e. (35% theoretical yield) in the case of CAL-A are
possible (Fig. 1(a and b)). The reactions are unique in
that the secondary alcohol function is enantioselectively
acetylated in the presence of the primary one. More work
is needed in order to increase the amounts of the fraction
of (S)-3 over that of 2.
1
according to the H NMR spectra. A small amount of
diacetate 4 was also produced in the reaction mixture in
accordance with the previous Mitsunobu acylation of 1
but it was easily separated during the workup.32 1H
NMR: l (ppm) 2.04 (s, 3H, CH3CO2), 3.75 (qd, 2H,
CH2), 5.75 (dd, 1H, CH), 7.14–7.31 (m, 5H, Ar-H). 13C
NMR l (ppm) 21.1 (CH3CO2), 65.8 (CH2OH), 76.8
(CH), 126.6, 128.6 and 137.0 (6C, Ar-C), 170.7 (CO2).
Mass spectrum: M+=180.
4.1.3. Preparation of ( )-2-acetoxy-1-phenylethanol 2.
Diol
1 (2.30 g, 16.6 mmol) was dissolved in
dichloromethane (20 mL) and triethylamine (3.46 mL,
25.0 mmol) and DMAP (0.10 g, 0.8 mmol) were added
in an ice/ethanol bath. Acetic anhydride (1.57 mL 16.6
mmol) was added dropwise to the reaction mixture. The
solution was stirred for 1.5 h. The addition of methanol
(15 mL) stopped the reaction, work-up yielding 2 (0.85
g, 4.7 mmol) containing 7% of ( )-3 according to GLC
4. Experimental
4.1. Materials
Acylase I from A. melleus (0.49 U/mg) and genus (35
U/mg) were obtained from Sigma (Deisenhofen, Ger-

2454
P. Virsu et al. / Tetrahedron: Asymmetry 12 (2001) 2447–2455
1
analysis. H NMR: l (ppm) 2.09 (s, 3H, CH3CO2), 2.80
(s, 1H, CHOH), 4.19 (qd, 2H, CH2), 4.93 (dd, 1H,
CH), 7.26–7.40 (m, 5H, Ar-H). 13C NMR l (ppm) 20.8
(CH3CO2), 69.2 (CHOH), 72.3 (CH2), 126.1, 128.5 and
139.8 (6C, Ar-C) 171.1 (CO2). Mass spectrum: M+=
180.
(m, 2H, CH2), 4.79, (m, 1H, CH), 7.21–7.37 (m, 5H,
Ar-H). 13C NMR l (ppm) 68.0 (CH2), 74.7 (CH),
126.0, 128.0, 128.5 and 140.4 (6C, Ar-C). Mass spec-
trum: M+=138. Melting point: 64°C; (lit.)34 67–69°C.
Elemental analysis: obs. C, 69.36, H; 7.51%. Calcd C,
69.54; H, 7.30% for the C8H10O2. The spectroscopic
data for (R)-3 are in accordance with those for ( )-3.
4.2. Methods
4.4. Inversion of configuration; transformation to the diol
The progress of the reactions was followed by taking
samples (200 mL for acetylation and 500 mL for
butanolysis) at intervals, filtering off the enzyme and
analyzing the sample by GLC on Astec Chiraldex G-
TA column. Free alcohol groups in the samples were
derivatized with 50 mL of trifluoroacetic anhydride
before the GLC analysis. For alcoholysis reactions, the
unreacted alcohol was evaporated and the residue dis-
solved in 200 mL of dichloromethane before the deriva-
tization. Enantiomeric excess values (e.e.) and relative
proportions of the components in the reaction mixture
at a given time were determined according to the peak
areas in the chromatograms and to the calibration
mixture of 1–4 being 0.1 M for each.
A solution of (S)-1 (0.150 g, 1.1 mmol) in tetra-
hydrofuran (5 mL) was treated with triphenylphosphine
(0.285 g, 1.1 mmol) and acetic acid (62 mL, 1.1 mmol),
followed by the addition of DEAD (169 mL, 1.1 mmol).
The work-up was as above. The product was isolated
yielding (R)-3 [82 mg, 0.46 mmol, e.e. 95%, [h]2D0 −80 (c
1, CHCl3)] and containing 6% of ( )-2 according to the
GLC method.
(R)-3 (0.266 g, 1.5 mmol, e.e. 94%) was dissolved in
methanol (10 mL) and basic ion-exchange resin Amber-
lite IRA-401 (ca 1 g) was added. Deacetylation was
complete in 1 h, as detected by TLC. The ion-exchange
resin was removed by filtration and the filtrate was
dried with Na2SO4. Evaporation of the solvent yielded
(R)-1 (0.184 g, 1.3 mmol, e.e. 94%).
1H and 13C NMR spectra were measured in CDCl3 on
a Jeol Lambda 400 or Bruker 200 Spectrometer with
tetramethylsilane as an internal standard. MS spectra
were recorded on a VG Analytical 7070E instrument
equipped with a ZAXstation 3100 M76 computer. Opti-
cal rotations were measured using a Jasco DIP-360
polarimeter. Elemental analyses were performed using a
Perkin–Elmer CHNS-240 Ser II Elemental Analyzer.
The melting point of compound 1 was measured using
4.5. Determination of absolute configurations
The absolute configurations for the enantiomers of 1
were determined by using commercial (R)-1. For that
purpose (R)-1 in CH2Cl2 (0.1 M, 100 mL) was added
into the solution of ( )-1 (0.1 M, 100 mL). The mixture
was derivatized with trifluoroacetic anhydride and ana-
lyzed by the GLC method. The absolute configurations
for the enantiomers of 4 were determined in the same
way by acetylating the mixture of commercial (R)-1
(0.1 M, 100 mL) and ( )-1 (0.1 M, 100 mL) with acetic
anhydride (10 mL) in the presence of DMAP (10 mL, 5%
solution in pyridine). In the case of 2, the absolute
configurations were obtained by preparing (R)-2 from
commercial (R)-1, as described above. In the case of 3,
absolute configurations were decided from the results
for the CAL-B-catalyzed alcoholysis of 4 leading to the
formation of (S)-1 and accordingly leaving (R)-3 unre-
acted. The sample was analyzed by GLC. Absolute
configurations were defined according to the enan-
tiomeric peak areas in the chromatogram.
8
BUCHI 510 Melting Point instrument.
4.3. Enzymatic resolution
The reactions were typically performed as small scale
experiments where one of the substrates 1–4, 1-
phenylethanol and 1-phenylethyl acetate or butanoate
(0.1 M) was dissolved in vinyl acetate, 1-butanol or
acetonitrile containing methanol or butanol (0.8 or 1.2
M). Diisopropyl ether, toluene and tetrahydrofuran
were also used as solvents for the CAL-B-catalyzed
methanolysis of 4. The enzyme preparation (75 mg/mL)
was added in order to start the reaction. The reaction
mixture was shaken at room temperature (25°C).
4.3.1. Gram-scale resolution of 1,2-diacetoxy-1-
phenylethane 4. 1,2-Diacetoxy-1-phenylethane 4 (2.00 g,
9.0 mmol) was dissolved in acetonitrile (90.0 mL).
Methanol (4.38 mL, 108 mmol) and CAL-B (6.7 g)
were added. After 110 h the enzyme was removed by
filtration and washed with acetonitrile. Purification by
column chromatography (acetone:petroleum ether, 3:7)
yielded (S)-1 [0.58 g, 4.2 mmol, e.e. 97%, [h]2D0 +63 (c 1,
CHCl3); [h]2D0 (lit.)34 +66 (c 1, CHCl3)] and (R)-3 [0.51
g, 2.8 mmol, e.e. 94%, [h]2D0 −80 (c 1, CHCl3)] contain-
ing 5% of 2 according to GLC analysis. It is possible
that the real amount of 2 is less than 5% in the
separated (R)-3 (see acyl migration during derivatiza-
References
1. Brown, J. M.; Murrer, B. A. J. Chem. Soc., Perkin Trans.
2 1982, 489–497.
2. King, R. B.; Bakos, J.; Hoff, C. D.; Marko´, L. J. Org.
Chem. 1979, 44, 1729–1931.
3. Bradshaw, J. S.; Jolley, S. T.; Izatt, R. M. J. Org. Chem.
1982, 47, 1229–1232.
4. Ma¨rki, H. P.; Crameri, Y.; Eigenmann, R.; Krasso, A.;
Ramuz, H.; Bernauer, K.; Goodman, M.; Melmon, K. L.
Helv. Chim. Acta 1988, 71, 320–336.
1
tion as described in Section 2.3). (S)-1: H NMR: l
(ppm) 2.43, (s, 1H, CH2OH), 2.86 (s, 1H, CHOH), 3.70

P. Virsu et al. / Tetrahedron: Asymmetry 12 (2001) 2447–2455
2455
20. Liljeblad, A.; Kanerva, L. T. Tetrahedron: Asymmetry 1999,
10, 4405–4415.
21. Liljeblad, A.; Lindborg, J.; Kanerva, L. T. Tetrahedron:
Asymmetry 2000, 11, 3957–3966.
22. Chenault, H. K.; Dahmer, J.; Whitesides, G. M. J. Am.
Chem. Soc. 1989, 111, 6354–6364.
23. Anders, M. W.; Dekant, W. Adv. Pharm. 1994, 59, 431–448.
24. Silverman, R. B. In The Organic Chemistry of Enzyme-Cat-
alyzed Reactions; Academic Press: San Diego, 2000; pp.
55–60.
25. Nielsen, T. B.; Ishii, M.; Kirk, O. Biotechnol. Appl.
Cold-Adapted Org. 1999, 49–61.
26. Anderson, E. M.; Larsson, K. M.; Kirk, O. Biocat. Biotrans.
1998, 16, 181–204.
27. Gedey, S.; Liljeblad, A.; La´za´r, L.; Fu¨lo¨p, F.; Kanerva, L.
T. Tetrahedron: Asymmetry 2001, 12, 105–110.
28. Bornscheuer, U. T.;Kazlauskas, R. J. HydrolasesinOrganic
Synthesis. Regio- and Stereoselective Biotransformations;
Wiley: Weinheim, 1999.
5. Pauluth, D.; Wa¨chtler, E. F. In Chirality in Industry II;
Collins, A. N.; Sheldrake, G. N.; Crosby, J., Eds.; John
Wiley & Sons: Chichester, 1997; pp. 280–281.
6. Janssen, A. J. M.; Klunder, A. J. H.; Zwanenburg, B.
Tetrahedron 1991, 47, 7409–7416.
7. Bosetti, A.; Bianchi, D.; Cesti, P.; Golini, P.; Spezia, S. J.
Chem. Soc., Perkin Trans. 1 1992, 239–242.
8. Lemke, K.; Theil, F.; Kunath, A.; Schick, H. Tetrahedron:
Asymmetry 1996, 7, 971–974.
9. Theil, F.; Weidner, J.; Ballschuh, S.; Kunath, A.; Schick,
H. J. Org. Chem. 1994, 59, 388–393.
10. Liljeblad, A.; Va¨nttinen, E.; Kanerva, L. T. Chirality 1999,
11, 432–439.
11. Chen, C.-S.; Fujimoto, Y.; Girdaukas, G.; Sih, C. J. J. Am.
Chem. Soc. 1982, 104, 7294–7298.
12. Lundell, K.; Kanerva, L. T. Tetrahedron: Asymmetry 1995,
6, 2281–2286.
13. Engholm Jacobsen, E.; Hoff, B. H.; Anthonsen, T. Chirality
2000, 12, 654–659.
8
14. Rogalska, E.; Cudrey, C.; Ferrato, F.; Verger, R. Chirality
1993, 5, 24–30.
15. Va¨nttinen, E.; Kanerva, L. T. Tetrahedron: Asymmetry
1995, 6, 1779–1786.
29. Orrenius, C.; Hæffner, F.; Rotticci, D.; Ohrner, N.; Norin,
T.; Hult, K. Biocat. Biotrans. 1998, 16, 1–15.
30. van Lohuizen, O. E.; Verkade, P. E. Recl. Trav. Chim.
Pays-Bas 1960, 79, 133–159.
16. Herrado´n, B.; Valverde, S. Synlett 1995, 599–602.
17. Ors, M.; Morcuende, A.; Jime´nez-Vacas, M. I.; Valverde,
S.; Herrado´n, B. Synlett 1996, 449–451.
31. Lok, C. M.; Ward, J. P. Chem. Phys. Lipids 1986, 39, 19–29.
32. Pautard, A. M.; Evans, S. A. J. Org. Chem. 1988, 53,
2300–2303.
18. Faraldos, J.; Arroyo, E.; Herrado´n, B. Synlett 1997,
367–370.
33. Kanerva, L. T.; Sundholm, O. J. Chem. Soc., Perkin Trans.
1 1993, 2407–2410.
19. Bakker, M.; Spruijt, A. S.; van Rantwijk, F.; Sheldon, R.
A. Tetrahedron: Asymmetry 2000, 11, 1801–1808.
34. Handbook of Fine Chemicals and Laboratory Equipment,
Aldrich, 2000–2001, p. 1346.