5-[4-(1-Hydroxyethyl)phenyl]-10,15,20-triphenylporphyrin as a probe of the transition-state conformation in hydrolase-catalyzed enantioselective transesterifications

Article abstract of DOI:10.1021/jo0109063

5-[4-(1-Hydroxyethyl)phenyl]-10,15,20-triphenylporphyrin (1a) and zinc porphyrin 1b were designed and synthesized to experimentally examine the validity of the transition-state model previously proposed for the lipase-catalyzed kinetic resolution of secondary alcohols. The lipases from Pseudomonas cepacia (lipase PS), Candida antarctica (CHIRAZYME L-2), Rhizomucor miehei (CHIRAZYME L-9), and Pseudomonas aeruginosa (lipase LIP) exhibited excellent enantioselectivity (E > 100 at 30 °C). Subtilisin Carlsberg from Bacillus licheniformis (ChiroCLEC-BL) also showed high enantioselectivity for 1a (E = 140 at 30 °C), and the thermodynamic parameters were determined: ΔΔH^? = -6.8 ± 0.8 kcal mol^-1, ΔΔS^? = -13 ± 3 cal mol^-1 K^-1. Lipases and subtilisin showed R- and S-preference for 1, respectively. The mechanisms underlying the experimental observations are explained in terms of the transition-state models. The large secondary alcohol 1 is a powerful tool for investigating the conformation of the transition state of the enzyme-catalyzed reactions. The fact that 1 was resolved with high enantioselectivity strongly suggests that the gauche conformation, but not the anti conformation, is taken in the transition state, in agreement with the transition-state models involving the stereoelectronic effect.

Full text of DOI:10.1021/jo0109063

2144
J . Org. Chem. 2002, 67, 2144-2151
5-[4-(1-Hyd r oxyeth yl)p h en yl]-10,15,20-tr ip h en ylp or p h yr in a s a
P r obe of th e Tr a n sition -Sta te Con for m a tion in
Hyd r ola se-Ca ta lyzed En a n tioselective Tr a n sester ifica tion s
Tadashi Ema,* Masahito J ittani, Kenji Furuie, Masanori Utaka, and Takashi Sakai*
Department of Applied Chemistry, Faculty of Engineering, Okayama University,
Tsushima, Okayama 700-8530, J apan
ema@cc.okayama-u.ac.jp
Received September 6, 2001
5-[4-(1-Hydroxyethyl)phenyl]-10,15,20-triphenylporphyrin (1a ) and zinc porphyrin 1b were designed
and synthesized to experimentally examine the validity of the transition-state model previously
proposed for the lipase-catalyzed kinetic resolution of secondary alcohols. The lipases from
Pseudomonas cepacia (lipase PS), Candida antarctica (CHIRAZYME L-2), Rhizomucor miehei
(CHIRAZYME L-9), and Pseudomonas aeruginosa (lipase LIP) exhibited excellent enantioselectivity
(E >100 at 30 °C). Subtilisin Carlsberg from Bacillus licheniformis (ChiroCLEC-BL) also showed
high enantioselectivity for 1a (E ) 140 at 30 °C), and the thermodynamic parameters were
determined: ∆∆Hq ) -6.8 ( 0.8 kcal mol-1, ∆∆Sq ) -13 ( 3 cal mol-1 K-1. Lipases and subtilisin
showed R- and S-preference for 1, respectively. The mechanisms underlying the experimental
observations are explained in terms of the transition-state models. The large secondary alcohol 1
is a powerful tool for investigating the conformation of the transition state of the enzyme-catalyzed
reactions. The fact that 1 was resolved with high enantioselectivity strongly suggests that the gauche
conformation, but not the anti conformation, is taken in the transition state, in agreement with
the transition-state models involving the stereoelectronic effect.
In tr od u ction
number of the substrates successfully resolved by a single
lipase increases enormously, mechanisms depending on
binding interactions are becoming inconsistent even if
drastic conformational change and different binding
modes are assumed. Moreover, it should also be noted
that in organic solvents, where hydrophobic interactions
are unlikely to work,4 binding interactions should be
weak.
We have postulated that any specific binding interac-
tion between an enzyme and variable regions (substitu-
ents) of substrates obstructs the simultaneous achieve-
ment of high enantioselectivity and broad substrate
specificity and have demonstrated that the enantiose-
lectivity of lipases for secondary alcohols can be accounted
for even without any binding interaction. The essence of
the stereo-sensing mechanism is represented in Figure
1b.5 In this model, enantioselectivity is explained only
by the conformational requirements and repulsive inter-
actions in the transition state, and any binding interac-
tion between enzyme’s pockets and substrate’s substit-
uents is not involved. In this sense, lipases are considered
to be “chemical reagent-like”.6 Further study has revealed
that such a concept can be applied to subtilisins (Figure
1c).7 These transition-state models can rationalize the
opposite empirical rules for lipases (typically, R-prefer-
ence)8 and subtilisins (typically, S-preference)7,9 (Figure
1a).
The most unusual but attractive feature of lipases is
that, unlike other enzymes, lipases simultaneously show
high enantioselectivity and broad substrate specificity.1
Uncovering the origin of this unique feature of lipases is
an intriguing subject. Traditionally, binding interactions
between enzyme’s pockets and substrate’s substituents
are believed to be essential for enantiomer discrimina-
tion. Many binding models and active-site models have
been proposed by substrate mapping,2 and most of the
computational calculations have been performed on the
basis of the concept of binding models.3 However, multi-
site binding interactions should necessarily restrict or
narrow substrate specificity, which is opposed to the
broad substrate specificity actually observed. As the
(1) For example: (a) Wong, C.-H.; Whitesides, G. M. Enzymes in
Synthetic Organic Chemistry; Elsevier: Oxford, 1994. (b) Enzyme
Catalysis in Organic Synthesis; Drauz, K., Waldmann, H., Eds.; VCH:
New York, 1995; Vol. 1. (c) Faber, K. Biotransformations in Organic
Chemistry; Springer-Verlag: Berlin, 1995. (d) Enzymatic Reactions in
Organic Media; Koskinen, A. M. P., Klibanov, A. M., Eds.; Blackie
Academic: Glasgow, 1996. (e) Bornscheuer, U. T.; Kazlauskas, R. J .
Hydrolases in Organic Synthesis; Wiley-VCH: Weinheim, 1999.
(2) (a) Itoh, T.; Kuroda, K.; Tomosada, M.; Takagi, Y. J . Org. Chem.
1991, 56, 797. (b) Hultin, P. G.; J ones, J . B. Tetrahedron Lett. 1992,
33, 1399. (c) Naemura, K.; Fukuda, R.; Murata, M.; Konishi, M.; Hirose,
K.; Tobe, Y. Tetrahedron: Asymmetry 1995, 6, 2385. (d) Nakamura,
K.; Kawasaki, M.; Ohno, A. Bull. Chem. Soc. J pn. 1996, 69, 1079. (e)
Nishizawa, K.; Ohgami, Y.; Matsuo, N.; Kisida, H.; Hirohara, H. J .
Chem. Soc., Perkin Trans. 2 1997, 1293.
(3) (a) Cygler, M.; Grochulski, P.; Kazlauskas, R. J .; Schrag, J . D.;
Bouthillier, F.; Rubin, B.; Serreqi, A. N.; Gupta, A. K. J . Am. Chem.
Soc. 1994, 116, 3180. (b) Lemke, K.; Lemke, M.; Theil, F. J . Org. Chem.
1997, 62, 6268. (c) Grabuleda, X.; J aime, C.; Guerrero, A. Tetrahe-
dron: Asymmetry 1997, 8, 3675. (d) Faeffner, F.; Norin, T.; Hult, K.
Biophys. J . 1998, 74, 1251. (e) Orrenius, C.; Faeffner, F.; Rotticci, D.;
O¨ hrner, N.; Norin, T.; Hult, K. Biocat. Biotrans. 1998, 16, 1. (f) Cuiper,
A. D.; Kouwijzer, M. L. C. E.; Grootenhuis, P. D. J .; Kellogg, R. M.;
Feringa, B. L. J . Org. Chem. 1999, 64, 9529.
(4) Israelachvili, J . N. Intermolecular and Surface Forces; Academic
Press: London, 1985.
(5) Ema, T.; Kobayashi, J .; Maeno, S.; Sakai, T.; Utaka, M. Bull.
Chem. Soc. J pn. 1998, 71, 443.
(6) In the field of synthetic organic chemistry, many transition-state
models, which involve no binding interaction, have been proposed for
various stereoselective chemical reactions. Gawley, R. E.; Aube´, J .
Principles of Asymmetric Synthesis; Elsevier: Oxford, 1996.
10.1021/jo0109063 CCC: $22.00 © 2002 American Chemical Society
Published on Web 03/01/2002

Modeling Hydrolase-Catalyzed Transesterifications
J . Org. Chem., Vol. 67, No. 7, 2002 2145
F igu r e 1. (a) Empirical rules for the lipase- and subtilisin-catalyzed kinetic resolutions of secondary alcohols. L and M represent
the larger and smaller substituents, respectively. Typically, (R)- and (S)-enantiomers react faster in the lipase- and subtilisin-
catalyzed kinetic resolutions, respectively. (b) (c) Transition-state models for the (b) lipase- and (c) subtilisin-catalyzed kinetic
resolutions of secondary alcohols. In both models, (i) the C-O bond of a substrate takes the gauche conformation with respect to
the breaking C- - -O bond, which is due to the stereoelectronic effect, and (ii) the H atom attached to the asymmetric C atom of
the substrate is syn-oriented toward the carbonyl O atom of the acetyl group. When such a locally favorable conformation is
taken, the faster-reacting enantiomer can direct the larger substituent (R1 in (b) and R2 in (c)) toward external solvent without
severe steric hindrance, whereas the slower-reacting enantiomer directs the larger substituent (R2 in (b) and R1 in (c)) toward the
protein wall, causing a severe steric repulsion. Even if any other conformation is taken, the slower-reacting enantiomer necessarily
becomes less stable than the antipodal enantiomer. For details, see ref 5.
Ch a r t 1
transition state. The transition-state models predict that
even very large secondary alcohols whose larger sub-
stituent cannot be accommodated in any pocket of the
enzymes can be efficiently resolved if they are ap-
propriately designed, which would be difficult to predict
on the basis of traditional binding models. Figure 1b and
c suggests that one enantiomer of 1 ((R)-1 for lipases and
(S)-1 for subtilisins) is susceptible to the acylation
because the tetraphenylporphyrin moiety can be directed
toward external solvent and that the other enantiomer
((S)-1 for lipases and (R)-1 for subtilisins) is highly
disfavored because of the severe steric repulsion between
the tetraphenylporphyrin moiety and the protein. In this
paper, we report the lipase- and subtilisin-catalyzed
kinetic resolutions of 1 and discuss the transition-state
conformation in detail.
The transition-state models have been derived on the
basis of the molecular orbital (MO) calculations and
molecular modeling5 and have been supported by the
kinetic5,7 and thermodynamic10 studies. Furthermore, as
reported in our preliminary papers, the extremely large
secondary alcohol, 5-[4-(1-hydroxyethyl)phenyl]-10,15,-
20-triphenylporphyrin (1) (Chart 1), undergoes lipase-
and subtilisin-catalyzed transesterifications.7,11 We de-
signed 1 in order to investigate the conformation of the
Resu lts
Lip a se-Ca ta lyzed Kin etic Resolu tion . Porphyrin
1a was synthesized by the condensation of benzaldehyde,
4-(1-hydroxyethyl)benzaldehyde (3),12 and pyrrole accord-
ing to the method of Lindsey.13 Zinc porphyrin 1b was
prepared by the standard method with Zn(OAc)2.14 The
lipase-catalyzed kinetic resolutions of 1 were performed
with vinyl acetate in dry diisopropyl ether at 30 °C
(Scheme 1). Several commercially available lipases were
examined. The enantiomeric purities (% ee) of 1 and 2
(7) Ema, T.; Okada, R.; Fukumoto, M.; J ittani, M.; Ishida, M.;
Furuie, K.; Yamaguchi, K.; Sakai, T.; Utaka, M. Tetrahedron Lett.
1999, 40, 4367.
(8) (a) Kazlauskas, R. J .; Weissfloch, A. N. E.; Rappaport, A. T.;
Cuccia, L. A. J . Org. Chem. 1991, 56, 2656. (b) Burgess, K.; J ennings,
L. D. J . Am. Chem. Soc. 1991, 113, 6129. (c) J anssen, A. J . M.; Klunder,
A. J . H.; Zwanenburg, B. Tetrahedron 1991, 47, 7645. (d) Xie, Z.-F.
Tetrahedron: Asymmetry 1991, 2, 733.
(9) (a) Fitzpatrick, P. A.; Klibanov, A. M. J . Am. Chem. Soc. 1991,
113, 3166. (b) Wang, Y.-F.; Yakovlevsky, K.; Zhang, B.; Margolin, A.
L. J . Org. Chem. 1997, 62, 3488. (c) Kazlauskas, R. J .; Weissfloch, A.
N. E. J . Mol. Catal. B: Enz. 1997, 3, 65. (d) Lloyd, R. C.; Dickman,
M.; J ones, J . B. Tetrahedron: Asymmetry 1998, 9, 551.
(10) Ema, T.; Yamaguchi, K.; Wakasa, Y.; Tanaka, N.; Utaka, M.;
Sakai, T. Chem. Lett. 2000, 782.
(11) Ema, T.; J ittani, M.; Sakai, T.; Utaka, M. Tetrahedron Lett.
1998, 39, 6311.
(12) Einhorn, J .; Einhorn, C.; Ratajczak, F.; Pierre, J .-L. J . Org.
Chem. 1996, 61, 7452.
(13) Lindsey, J . S.; Schreiman, I. C.; Hsu, H. C.; Kearney, P. C.;
Marguerettaz, A. M. J . Org. Chem. 1987, 52, 827.
(14) Smith, K. M. Porphyrins and Metalloporphyrins; Elsevier:
Amsterdam, 1975.

2146 J . Org. Chem., Vol. 67, No. 7, 2002
Ema et al.
Sch em e 1
Ta ble 1. Lip a se-Ca ta lyzed En a n tioselective Tr a n sester ifica tion s of 1a
% yieldc (% eed)
entry
lipase
lipase PSf
alcohol
time (h)
c (%)b
(R)-2
(S)-1
E valuee
1
2
3
4
5
6
7
8
1a
1b
1a
1b
1a
1b
1a
1b
23
23
9
48
48
48
5
48
14
50
20
5
8
50
18
44 (>98)
12 (>98)
45 (>98)
24 (>98)
8 (>98)
7 (>98)
46 (>98)
24 (>98)
47 (89)
82 (16)
47 (>98)
74 (25)
90 (5)
78 (8)
36 (>98)
72 (21)
>298
>116
>458
>126
>104
>107
>458
>122
lipase PS
CHIRAZYME L-2g
CHIRAZYME L-2
CHIRAZYME L-9h
CHIRAZYME L-9
lipase LIPi
lipase LIP
5
a
b
Conditions: lipase (900 mg), 1 (42 µmol), vinyl acetate (1.3 mmol), dry i-Pr2O (30 mL), 30 °C. Conversion calculated from ee(1)/
(ee(1) + ee(2)) according to ref 15. c Isolated yield. Determined by HPLC (Chiralcel OD-H) after conversion to 1a . e Calculated from E
d
) ln [1 - c(1 + ee(2))]/ln [1 - c(1 - ee(2))] according to ref 15. f Pseudomonas cepacia lipase (Amano Pharmaceutical). g Candida antarctica
h
i
lipase (Boehringer Mannheim). Rhizomucor miehei lipase (Boehringer Mannheim). Pseudomonas aeruginosa lipase (Toyobo).
Sch em e 2
were determined by means of HPLC after they were
Ta ble 2. Su btilisin -Ca ta lyzed En a n tioselective
converted to 1a , and the E values were calculated
according to the literature.15 The absolute configuration
of ester 2 obtained in the lipase-catalyzed transesterifi-
cations of 1 was determined to be (R) according to the
method described in the Experimental Section. It was
also confirmed by a control experiment without lipase
that zinc porphyrin 1b itself cannot function as a catalyst
for the transesterification. The results are listed in Table
1. Excellent enantioselectivities (E > 100) were observed
in all cases, and complete kinetic resolutions were
achieved in entries 3 and 7, where the enantiomeric
purities of both alcohol 1a and ester 2a were >98% ee.16
Su btilisin -Ca ta lyzed Kin etic Resolu tion . Cross-
linked enzyme crystals of subtilisin Carlsberg (Chiro-
Tr a n sester ifica tion s of 1a a
d
% yieldc (% ee
)
temp (°C)
time (h)
c (%)b
(S)-2a
(R)-1a
E valuee
20
25
30
35
40
45
3.5
4
3
3
2
45
38
42
42
37
37
41 (98)
34 (98)
40 (97)
39 (97)
38 (97)
32 (96)
54 (81)
61 (60)
57 (71)
59 (70)
59 (58)
68 (57)
249
183
140
138
118
87
2
a
Conditions: ChiroCLEC-BL (40 mg), 1a (42 µmol), vinyl
b
acetate (1.3 mmol), dry i-Pr2O (30 mL). Conversion calculated
from ee(1a )/(ee(1a ) + ee(2a )) according to ref 15. c Isolated yield.
Determined by HPLC (Chiralpak AD). e Calculated from E ) ln
d
[1 - c(1 + ee(2a ))]/ln [1 - c(1 - ee(2a ))] according to ref 15.
(15) Chen, C.-S.; Fujimoto, Y.; Girdaukas, G.; Sih, C. J . J . Am. Chem.
Soc. 1982, 104, 7294.
CLEC-BL) were used as a biocatalyst. The subtilisin-
catalyzed kinetic resolutions of 1a were performed with
vinyl acetate in dry diisopropyl ether (Scheme 2). The
(S)-enantiomer of 1a was acylated faster, and a high E
value was obtained at 30 °C (Table 2). The enantioselec-
tivity was high but incomplete, and therefore we were
able to determine the thermodynamic parameters, ∆∆Hq
(16) When both alcohol and ester obtained at 50% conversion are
enantiomerically pure, we call the reaction “complete kinetic resolu-
tion”. Many complete kinetic resolutions can be seen in the literature.
See, for example: (a) Tanaka, K.; Ogasawara, K. Synthesis 1995, 1237.
(b) Taniguchi, T.; Takeuchi, M.; Kadota, K.; ElAzab, A. S.; Ogasawara,
K. Synthesis 1999, 1325. (c) Nakashima, H.; Sato, M.; Taniguchi, T.;
Ogasawara, K. Synlett 1999, 1754. See also ref 22. Complete desym-
metrization of meso diols has also been reported and can be explained
in the same way. (d) J ohnson, C. R.; Golebiowski, A.; McGill, T. K.;
Steensma, D. H. Tetrahedron Lett. 1991, 32, 2597.
and ∆∆Sq, according to eq 1, where ∆∆Hq ) ∆Hq
-
fast

Modeling Hydrolase-Catalyzed Transesterifications
J . Org. Chem., Vol. 67, No. 7, 2002 2147
F igu r e 2. Plots of ln E vs 1/T for the subtilisin-catalyzed
kinetic resolutions of 1a .
F igu r e 3. A linear relationship between ∆∆Hq and ∆∆Sq
values in the subtilisin-catalyzed kinetic resolutions of a
variety of secondary alcohols. Except for 1a , all the data are
taken from ref 10. For comparison, the data for 1-phenyletha-
nol is indicated.
17
∆Hq
and ∆∆Sq ) ∆Sq - ∆Sq
.
slow
slow
fast
-∆∆Hq ∆∆Sq
ln E )
+
(1)
RT
R
stopped completely at 50% conversion and never pro-
ceeded further even after a prolonged reaction time.16
This observation indicates that the high enantioselectiv-
ity is achieved by the highly suppressed reactivity of the
(S)-enantiomer. Figure 1b suggests that a severe steric
repulsion occurs between the tetraphenylporphyrin moi-
ety of the slower-reacting enantiomer, (S)-1, and the
protein. Thus, the transition-state model proposed by us
contrasts with traditional binding models, in the latter
of which binding interactions between pockets of an
enzyme and substituents of a faster-reacting enantiomer
are the principal factor in reaction acceleration and high
enantioselectivity.
The E values obtained at several temperatures are
listed in Table 2. Plot of the ln E values against 1/T
(Figure 2) afforded the following values: ∆∆Hq) -6.8 (
0.8 kcal mol-1, ∆∆Sq ) -13 ( 3 cal mol-1 K-1. Previously,
we have found a linear relationship between the ∆∆Hq
and ∆∆Sq values for the subtilisin-catalyzed kinetic
resolutions of a variety of secondary alcohols.10 Interest-
ingly, the one set of the data for 1a was found to fall into
this linear relationship (Figure 3).18
Discu ssion
Four kinds of lipases exhibited excellent enantioselec-
tivity and R-preference for 1 (Table 1), as predicted by
the transition-state model. To our knowledge, 1 is the
largest secondary alcohol that has been successfully
resolved by the lipase-catalyzed reactions. The X-ray
crystal structures of lipases from Pseudomonas cepacia,19
Candida antarctica,20 and Rhizomucor miehei,21 which
are used in the present study, are available from the
Protein Data Bank. Obviously, the tetraphenylporphyrin
moiety of 1 is so large (edge-to-edge distance of ca. 17 Å)
that it cannot be accommodated by any pocket of these
enzymes. Therefore, the results in Table 1 strongly
suggest that the accommodation of the larger substituent
of substrates in a pocket of the enzymes is not necessarily
required for attaining high enantioselectivity. In entries
3 and 7 (Table 1), the transesterification reactions
Interestingly, zinc porphyrin 1b was acylated more
slowly than free base porphyrin 1a in most cases (Table
1). We suppose that 1b binds to polar amino acid residues
on the protein surface. This coordination may block the
active site or may decrease the protein flexibility es-
sential to catalytic activity. Because it is well-known that
a porphyrin ring is rigidified by complexation with a
metal atom, the increased rigidity of 1b might also be
related to the reduced reactivity. Even free base porphy-
rin 1a reacted relatively slowly in most cases, e.g., the
total turnover number (TTN) of the enzyme at 50%
conversion in entry 1 was calculated to be ca. 77, which
is much lower than ca. 5000 reported for 2-[(N,N-
dimethylcarbamoyl)methyl]-3-cyclopenten-1-ol under the
similar reaction conditions.22 Computer modeling sug-
gested that even (R)-1a can cause unfavorable steric
interactions with some parts of the protein such as the
lid (not shown). The reactivity of 1 may also be lowered
apparently by the low concentration of 1 (1.4 mM) or
inherently by the large size of the molecule.
(17) (a) Phillips, R. S. Trends Biotechnol. 1996, 14, 13. (b) Sakai,
T.; Kawabata, I.; Kishimoto, T.; Ema, T.; Utaka, M. J . Org. Chem. 1997,
62, 4906. (c) Sakai, T.; Kishimoto, T.; Tanaka, Y.; Ema, T.; Utaka, M.
Tetrahedron Lett. 1998, 39, 7881.
(18) The reaction conditions for 1a are slightly different from those
for secondary alcohols in ref 10. When molecular sieves 4Å were added,
less reliable data with larger standard deviations were obtained: ∆∆Hq
Subtilisin Carlsberg showed S-preference for 1a (Table
2), which is opposite to the R-preference of lipases. It is
known that an approximate mirror-image relationship
exists between the catalytic residues (the catalytic triad
and the oxyanion hole) of the serine proteases and those
) -6.6 ( 1.2 kcal mol-1, ∆∆Sq ) -12 ( 4 cal mol-1 K-1
.
(19) Kim, K. K.; Song, H. K.; Shin, D. H.; Hwang, K. Y.; Suh, S. W.
Structure 1997, 5, 173. The name of Pseudomonas cepacia has been
changed to Burkholderia cepacia.
(20) Uppenberg, J .; Hansen, M. T.; Patkar, S.; J ones, T. A. Structure
1994, 2, 293.
(21) Derewenda, U.; Brzozowski, A. M.; Lawson, D. M.; Derewenda,
Z. S. Biochemistry 1992, 31, 1532.
(22) Ema, T.; Maeno, S.; Takaya, Y.; Sakai, T.; Utaka, M. J . Org.
Chem. 1996, 61, 8610.

2148 J . Org. Chem., Vol. 67, No. 7, 2002
Ema et al.
of lipases.23 The opposite enantiopreferences of subtilisin
(S-preference) and lipases (R-preference) are reasonable
because the enantioselectivity results predominantly
from the local conformational requirements and repulsive
interactions in the transition state (Figure 1b and c).
Previously, we have observed a clear tendency that
subtilisin shows an enantioselectivity for various second-
ary alcohols lower than that of lipases.7 We have ascribed
the reduced enantioselectivity of subtilisin to the lack of
a protein wall corresponding to the “triangular wall” of
lipases, which we had proposed to be important for high
enantioselectivity of lipases.5 Instead of such a wall,
subtilisin has a shallow depression (S1′),24 which is made
up of the â-strand, the R-helical turn, and the histidine
imidazole of the catalytic triad and which is open to
external solvent (Figure 1c).25 Owing to this depression
(S1′), the steric repulsion point in subtilisin is distant
from the stereocenter of a secondary alcohol in the
transition state as compared to the case of lipases.
Besides, the active site of subtilisin is less hindered than
that of lipases having a lid. As a result of these structural
differences, subtilisin allowed the unfavorable enanti-
omer, (R)-1a , to react very slowly, resulting in the high
but incomplete enantioselectivity (Table 2), whereas
lipases completely blocked the reaction of (S)-1, leading
to the enantiospecific kinetic resolutions (Table 1).
We have previously observed a partial compensation
effect between the ∆∆Hq and ∆∆Sq values for subtilisin-
catalyzed kinetic resolutions of a variety of secondary
alcohols.10 As shown in Figure 3, the ∆∆Sq value de-
creases with a decrease in the ∆∆Hq value. Equation 1
indicates that a negative ∆∆Hq value contributes to an
increase in the E value, whereas a negative ∆∆Sq value
contributes to a decrease in the E value. We have
explained this phenomenon in terms of the transition-
state model (for details, see ref 10). Interestingly, the
present thermodynamic data for 1a also falls into this
linear relationship (Figure 3). This suggests that the
extremely large alcohol 1a is enantiodiscriminated basi-
cally in the same way as for other secondary alcohols of
smaller sizes. The E value for 1a (E ) 140) is much larger
than, for example, that for 1-phenylethanol (E ) 10).7
Figure 3 indicates that this enhanced E value for 1a is
due to the large, negative ∆∆Hq value, which in turn
suggests a large difference in the degree of the steric
repulsion and the strain caused in the transition state
between the enantiomers of 1a .
F igu r e 4. Conformations of the transition state liberating an
ester product in the (a) lipase- and (b) subtilisin-catalyzed
acylations of a secondary alcohol (ROH). In both cases, the
protein core is on the rear side of the paper, and on the front
side of the paper is external solvent. The gauche and anti
conformations are defined by the dihedral angle between the
R-O bond and the breaking C- - -O bond. The gauche confor-
mation is more favorable than the anti conformation because
of the stereoelectronic effect (see text or ref 5).
conformation. As can be seen in the transition-state
models (Figure 1b and c), this conformational require-
ment due to the stereoelectronic effect plays a key role
in chiral discrimination.5 On the other hand, studies
based on an X-ray crystal structure of a lipase complexed
with a chiral transition-state analogue, menthyl phos-
phonate,3a or based on molecular mechanics (MM) cal-
culations on tetrahedral intermediates3d-f have failed to
find the significance of the stereoelectronic effect, and the
researchers seem to believe (i) the anti conformation is
the reactive conformation and (ii) binding pockets are
essential for chiral discrimination. Truly, the tetrahedral
geometry at the P atom (phosphonate inhibitor) or at the
central C atom (tetrahedral intermediate) and the elec-
tronegative O atom locally mimic the transition-state
structure of the esterification/hydrolysis. However, it is
doubtful whether even the conformation mimics the real
transition state. The stereoelectronic effect in the P-
containing complex must be small, because the long P-O
bond distance and the differences in orbital size and
energy level between the P and O atoms weaken the
n-σ* orbital mixing,28 and because the complex is in the
ground state.29 The stereoelectronic effect in the tetra-
hedral intermediate is also small as reported previ-
ously.5,29 The large alcohol 1 is a powerful tool to address
this conformational issue. In particular, subtilisin Carls-
Because the stereoelectronic effect originates from the
intrinsic nature of chemical bonds,26 it should be retained
not only in chemical reactions but also in enzymatic
reactions.27 Figure 4 illustrates that the gauche confor-
mation, with the lone-pair orbital (n) mixing with the
antibonding orbital (σ*), facilitates the C-O bond cleav-
age more efficiently than the anti conformation and that
the protein is flanked by the substrate part (R) more
closely in the gauche conformation than in the anti
(28) The X-ray crystal structures of lipase-inhibitor complexes in
the Protein Data Bank show that the dihedral angle in question (C-
O-P-Oγ) varies widely, although there is a weak tendency to fall into
the gauche conformation: 91.4° (1EX9), 124.9° (1HQD), 140.0° (5LIP),
75.4° (1LPB), 130.3° (1LPM), 70.8° (1LPS), 94.1° (4TGL), and 70.5°
(5TGL).
(29) To simulate the enantioselectivity, many researchers have
employed the tetrahedral intermediate, instead of the real transition
state, in the MM calculations. However, it has been revealed by the
MO calculations that the stereoelectronic effect operates effectively in
the bond-breaking/forming transition state rather than in the tetra-
hedral intermediate.5 The energetic benefit due to the stereoelectronic
effect is inherently small in the tetrahedral intermediate and probably
in the enzyme-phosphonate complex because they have no transiently
breaking/forming chemical bond. Moreover, the MM calculations
cannot properly evaluate the stereoelectronic effect that explicitly
involves the interactions of molecular orbitals. For these reasons, the
real transition state should be treated by MO methods. As an
alternative method, Tafi et al. have involved the concept of the
stereoelectronic theory in the MM calculations on the tetrahedral
intermediates to reproduce the enantiopreference of a lipase. Tafi, A.;
van Almsick, A.; Corelli, F.; Crusco, M.; Laumen, K. E.; Schneider, M.
P.; Botta, M. J . Org. Chem. 2000, 65, 3659.
(23) Derewenda, Z. S.; Wei, Y. J . Am. Chem. Soc. 1995, 117, 2104.
(24) The S1′ subsite was originally defined as a binding region for
the leaving group of peptide substrates. Schechter, I.; Berger, A.
Biochem. Biophys. Res. Commun. 1967, 27, 157.
(25) (a) Neidhart, D. J .; Petsko, G. A. Protein Eng. 1988, 2, 271. (b)
McPhalen, C. A.; J ames, M. N. G. Biochemistry 1988, 27, 6582.
(26) (a) Deslongchamps, P. Stereoelectronic Effects in Organic
Chemistry; Pergamon: Oxford, 1983. (b) Kirby, A. J . Stereoelectronic
Effects; Oxford University Press: Oxford, 1996.
(27) Dugas, H. Bioorganic Chemistry: A Chemical Approach to
Enzyme Action, 3rd ed.; Springer-Verlag: New York, 1996.

Modeling Hydrolase-Catalyzed Transesterifications
J . Org. Chem., Vol. 67, No. 7, 2002 2149
Con clu sion
Despite their extreme simplicity, the transition-state
models are indeed consistent with several experimental
observations for lipases and subtilisins. They can clearly
explain (i) the simultaneous achievement of high enan-
tioselectivity and broad substrate specificity,5 (ii) the
opposite enantiopreferences of lipases and subtilisins for
secondary alcohols,5,7 (iii) low activity for secondary
alcohols having bulky substituents on both sides,5 (iv)
little or no activity for tertiary alcohols,5 and (v) the
reason the large alcohol 1 can be resolved with high
enantioselectivity.7,11 In this paper, the experimental
observations for the enzyme-catalyzed kinetic resolutions
of 1 have been explained in detail in terms of the
transition-state models. By using 1 as a substrate, it has
been experimentally supported that the chemical reagent-
like feature of the enzymes (conformational requirements
and repulsive interactions in the transition state) is the
determinant of the enantioselectivity and that the gauche
conformation is taken in the transition state. It can
therefore be predicted that artificial catalysts mimicking
the catalytic machinery of lipases and subtilisins will
show R- and S-preference, respectively, for a variety of
secondary alcohols. The transition-state models are so
useful that they will enable us to plan efficient chemo-
enzymatic syntheses and rational alteration of the
enzymes.
F igu r e 5. Comparison of the gauche and anti conformations
in the subtilisin-catalyzed transesterification of 1a . (a) The anti
conformation for (S)-1a . (b) The anti conformation for (R)-1a .
(c) The gauche conformation for (S)-1a . (d) The gauche
conformation for (R)-1a . The enzyme is shown in green, (S)-
1a is in red, and (R)-1a is in magenda. The orientation of the
protein is similar to that in Figure 1c. Close-up representations
of (c) and (d) are drawn in (e) and (f), respectively, where the
tetraphenylporphyrin moiety is shown by an ellipse. In (a),
(b), (c), and (e), the H atom attached to the asymmetric C atom
of 1a is syn-oriented toward the oxyanion, which is conforma-
tionally favorable. In the case of (d) and (f), the tetraphe-
nylporphyrin moiety is forced to point outside, with the methyl
group syn-oriented toward the oxyanion, which is conforma-
tionally unfavorable due to the torsional strain and the steric
repulsion.5 These structures are shown to help the readers
understand the discussion, and it is only for easy manipula-
tions that we employed the tetrahedral intermediate. Because
the stereoelectronic effect in the tetrahedral intermediate is
small, the tetrahedral intermediate cannot be a substitute for
the real transition state. See refs 5 and 29.
Exp er im en ta l Section
Ma ter ia ls a n d Meth od s. Silica gel and basic alumina
column chromatography was performed using Fuji Silysia BW-
127 ZH (100-270 mesh) and Merck aluminum oxide 90 active
basic (0.063-0.200 mm), respectively. Thin-layer chromatog-
raphy (TLC) was performed on Merck silica gel 60 F254. Lipase
PS (1% (w/w) powder) was provided by Amano Pharmaceutical
Co. CHIRAZYME L-2 (carrier-fixed C2, isoform B) and L-9
(lyo) were provided by Boehringer Mannheim GmbH. Lipase
LIP and ChiroCLEC-BL were purchased from Toyobo Co. and
Altus Biologics Inc, respectively. 4-(1-Hydroxyethyl)benzalde-
hyde (3) was prepared according to the literature.12 Dry
diisopropyl ether and dry Et2O were distilled from sodium, and
dry CH2Cl2 was distilled from CaH2. The enantiomeric purities
of 1a , 2a , 3, 5, and 6 were determined by HPLC using Chiralcel
OD-H (1a ), Chiralpak AD (1a , 2a , 5, 6), and Chiralcel OB (3)
columns (Daicel Chemical Industries); that of 7 was deter-
mined by capillary gas chromatography using a CP-cyclodex-
trin-â-2,3,6-M-19 column (Chrompack, φ 0.25 mm × 25 m).1H
and 13C NMR spectra measured in CDCl3 at 200 and 50 MHz,
respectively, are reported.
Com p u ter Mod elin g. Molecular modeling was performed
with SYBYL 6.4 (Tripos, Inc.) on an O2 workstation (Silicon
Graphics, Inc.). The X-ray crystal structures of lipases from
P. cepacia (1OIL),19 C. antarctica (1LBS),20 and R. miehei
(4TGL)21 and that of subtilisin Carlsberg (2SEC)25b available
from the Protein Data Bank were used. After an inhibitor and
water molecules were removed, the tetrahedral intermediate
for (R)- or (S)-1a was constructed, and the hydrogen atoms
berg, whose active site is much less crowded than that
of lipases, provides a clear argument. If the anti confor-
mation is allowed, the tetraphenylporphyrin moiety of
both enantiomers makes no contact with the protein in
such a conformation that the H atom attached to the
asymmetric C atom of 1a is syn-oriented toward the
oxyanion (Figure 5a and b). This will result in no
enantioselectivity. In contrast, when the gauche confor-
mation is taken, the tetraphenylporphyrin moiety of (S)-
1a can be favorably directed toward external solvent
(Figure 5c and e), whereas (R)-1a undergoes a steric
repulsion and/or a strain considerably (Figure 5d and f).
This can lead to differential activation energy. Clearly,
the facts that the high E value (E ) 140 at 30 °C) and
the large, negative ∆∆Hq value (-6.8 kcal mol-1) were
observed for 1a strongly support the gauche conforma-
tion.30
(30) We also examined three kinds of lipases (1OIL, 1LBS, 4TGL).
Because the active site of the lipases is very crowded, it is surprising
indeed that 1 was allowed to react. Visual inspection of the tetrahedral
intermediates indicated that (R)-1a can direct the tetraphenylporphy-
rin moiety toward external solvent in a locally favorable conformation
shown in Figure 1b, although even this conformation obviously suffered
to some degree from a steric repulsion with the peripheral protein
parts, which can explain the low TTN (see text). The anti conformation
of (R)-1a and both conformations of (S)-1a had the tetraphenylpor-
phyrin moiety unacceptably overlapped with the protein or underwent
a severe strain. Accordingly, (R)-1a most likely takes the gauche
conformation in the transition state, and (S)-1a seems to have no way
to overcome the high energy barrier.

2150 J . Org. Chem., Vol. 67, No. 7, 2002
Ema et al.
Sch em e 3a
were generated. The conformation was manually adjusted
appropriately. The active site of the lipases was too crowded
to perform a reliable calculation. Since the active site of
subtilisin Carlsberg is not so crowded, the MM calculations
were performed. The whole protein was treated as a rigid body
using Aggregates, and the remaining parts were allowed to
move using the Tripos force field and the Gasteiger-Hu¨ckel
charges. The obtained structures were then compared.
5-[4-(1-H yd r oxyet h yl)p h en yl]-10,15,20-t r ip h en ylp or -
p h yr in (1a ). To a solution of 3 (602 mg, 4.01 mmol), benzal-
dehyde (1.27 g, 12.0 mmol), and pyrrole (1.10 g, 16.4 mmol) in
dry CH2Cl2 (1.5 L) was added BF3‚OEt2 (224 mg, 1.58 mmol)
under N2. After the solution was stirred at room temperature
for 3.5 h in the dark, o-chloranil (2.73 g, 11.1 mmol) was added.
After being stirred for 1 h, the solution was concentrated to
ca. 100 mL and washed with saturated aqueous NaHCO3 (100
mL × 2). The organic layer was separated and dried over Na2-
SO4. After the solution was concentrated, the product was
purified by repeated silica gel and basic alumina column
chromatography (hexane/CHCl3/EtOAc (20:20:3)) to afford 1a
as a purple solid (175 mg, 7%): mp > 280 °C; 1H NMR δ -2.61
(br s, 2H), 1.79 (d, J ) 6.4 Hz, 3H), 2.13 (br s, 1H), 5.19 (q, J
) 6.4 Hz, 1H), 7.70-7.82 (m, 11H), 8.23-8.34 (m, 8H), 8.96
(s, 8H); 13C NMR δ 25.4, 70.3, 119.9, 120.2, 123.7, 126.7, 127.7,
131.1, 134.5, 134.7, 141.2, 142.1, 145.0; IR (KBr) 3428, 3318
cm-1; HR FAB-MS calcd for C46H35N4O 659.2811, found
659.2775; Anal. Calcd for C46H34N4O: C, 83.87; H, 5.20; N,
8.50. Found: C, 83.61; H, 5.10; N, 8.39.
a
Reagents and conditions: (a) CHIRAZYME L-2, vinyl acetate,
i-Pr2O, -10 °C; (b) NaBH4, EtOH, 0 °C to rt; (c) CBr4, PPh3,
CH2Cl2, 0 °C to rt; (d) LiAlH4, Et2O, 0 °C to rt; (e) pyrrole, PhCHO,
BF3‚OEt2, CH2Cl2, rt, then o-chloranil.
monitored by TLC. The mixture was filtered and concentrated.
Alcohol (R)-1a and ester (S)-2a were separated by silica gel
column chromatography as shown above. HPLC for 1a: Chiral-
pak AD, hexane/2-PrOH ) 92:8, flow rate 0.3 mL/min, detec-
tion 420 nm, (R) 54.0 min, (S) 66.4 min. HPLC for 2a :
Chiralpak AD, hexane/2-PrOH ) 98:2, flow rate 0.3 mL/min,
detection 420 nm, (R) 67.1 min, (S) 81.9 min.
Deter m in a tion of Absolu te Con figu r a tion . The absolute
stereochemistry of 1 and 2 was established according to
Scheme 3. The absolute configuration of ester 4 obtained in
the lipase-catalyzed kinetic resolution of 3 was determined to
be (R) by confirming that 4 was converted to (R)-7, whose
specific rotation has been reported.31 This result agrees with
the prediction by the transition-state model that (R)-4 will be
obtained. The authentic (R)-4 was converted to (R)-2a . By
comparing the HPLC retention times, the absolute configu-
ration of 2a obtained in the lipase-catalyzed kinetic resolution
of 1a was determined to be (R). Detailed procedures are
described below.
{5-[4-(1-Hyd r oxyeth yl)p h en yl]-10,15,20-tr ip h en ylp or -
p h yr in a to}zin c(II) (1b). To a solution of 1a (122 mg, 185
µmol) in CHCl3 (50 mL) was added a solution of Zn(OAc)2‚
2H2O (433 mg, 1.97 mmol) in MeOH (10 mL). The solution
was heated at reflux for 30 min. The progress of the reaction
was monitored by UV-vis spectroscopy. After brine (100 mL)
was added to the solution, the organic layer was separated,
dried over Na2SO4, and concentrated. Recrystallization from
CHCl3/hexane afforded 1b as purple crystals (99 mg, 74%):
1
mp > 280 °C; H NMR δ 1.69 (d, J ) 6.3 Hz, 3H), 1.89 (br s
1H), 5.04-5.14 (m, 1H), 7.63-7.77 (m, 11H), 8.16-8.24 (m,
8H), 8.95 (s, 8H); 13C NMR δ 25.3, 70.4, 120.8, 121.1, 123.6,
126.5, 127.5, 132.0, 134.4, 134.6, 142.0, 142.8, 144.7, 150.2;
IR (KBr) 3413 cm-1; HR FAB-MS calcd for C46H33N4OZn
721.1946, found 721.1902.
(R)-4-(1-Acetoxyeth yl)ben za ld eh yd e ((R)-4).32 After a
mixture of 3 (301 mg, 2.00 mmol) and CHIRAZYME L-2 (667
mg) in dry i-Pr2O (6.7 mL) was stirred at -10 °C for 10 min,
vinyl acetate (345 mg, 4.01 mmol) was added. After being
stirred at -10 °C for 3 h, the mixture was filtered and
concentrated. Alcohol (S)-3 and ester (R)-4 were separated by
silica gel column chromatography (hexane/EtOAc (4:1) to (3:
Lip a se-Ca ta lyzed Kin etic Resolu tion s of 1. To a mixture
of 1a (28 mg, 42 µmol) or 1b (30 mg, 42 µmol) and lipase (900
mg) in dry i-Pr2O (30 mL) was added vinyl acetate (112 mg,
1.30 mmol) at 30 °C. The mixture was stirred at ca. 450 rpm
in the dark. The progress of the reaction was monitored by
TLC. The mixture was filtered and concentrated. Alcohol (S)-1
and ester (R)-2 were separated by silica gel column chroma-
tography (hexane/CHC13/EtOAc (28:16:3) to (20:20:3)). (R)-
2a : mp 270-273 °C; 1H NMR δ -2.76 (br s, 2H), 1.85 (d, J )
6.6 Hz, 3H), 2.28 (s, 3H), 6.27 (q, J ) 6.6 Hz, 1H), 7.71-7.78
(m, 11H), 8.21-8.26 (m, 8H), 8.87 (s, 8H); 13C NMR δ 21.6,
22.5, 72.3, 119.6, 120.2, 124.4, 126.7, 127.7, 131.1, 134.6, 134.7,
141.1, 141.7, 142.1, 170.6; IR (KBr) 3319, 1727 cm-1; HR FAB-
MS calcd for C48H37N4O2 701.2917, found 701.2871. (R)-2b: mp
273-275 °C; 1H NMR δ 1.83 (d, J ) 6.6 Hz, 3H), 2.23 (s, 3H),
6.22 (q, J ) 6.6 Hz, 1H), 7.70-7.79 (m, 11H), 8.19-8.25 (m,
8H), 8.96 (s, 8H); 13C NMR δ 21.5, 22.5, 72.3, 120.6, 121.1,
124.3, 126.5, 127.5, 132.0, 134.4, 134.5, 140.8, 142.3, 142.8,
150.1, 150.2, 170.6; IR (KBr) 1736 cm-1; HR FAB-MS calcd
for C48H35N4O2Zn 763.2051, found 763.2087. HPLC for 1a :
Chiralcel OD-H, hexane/2-PrOH ) 9:1, flow rate 0.5 mL/min,
detection 420 nm, (R) 27.2 min, (S) 32.2 min. The enantiomeric
purities of 2a , 2b, and 1b were determined after conversion
to 1a .
1)). (R)-4: 40% yield, >98% ee; [R]25 ) +119 (c 1.04, CHCl3);
D
1H NMR δ 1.54 (d, J ) 6.6 Hz, 3H), 2.10 (s, 3H), 5.90 (q, J )
6.6 Hz, 1H), 7.50 (d, J ) 8.2 Hz, 2H), 7.87 (d, J ) 8.2 Hz, 2H),
10.0 (s, 1H); 13C NMR δ 21.1, 22.2, 71.7, 126.4, 130.0, 135.8,
148.4, 170.1, 191.7; IR (neat) 1738, 1704 cm-1. (S)-3: 53% yield,
71% ee; [R]25 ) -36.5 (c 1.25, CHCl3), lit.33 [R]24 ) +50.2 (c
D
D
3.11, CHCl3) for (R)-3 with 88% ee (the absolute configuration
reported as a tentative assignment in ref 33 was found to be
correct). HPLC for 3: Chiralcel OB, hexane/2-PrOH ) 95:5,
flow rate 0.5 mL/min, detection 254 nm, (S) 35.9 min, (R) 41.3
min. The enantiomeric purity of 4 was determined after
conversion to 3.
(R)-4-(1-Acetoxyeth yl)ben zyl Alcoh ol ((R)-5). To a solu-
tion of (R)-4 (480 mg, 2.50 mmol) in EtOH (10 mL) and
saturated aqueous NaHCO3 (0.4 mL) was added portionwise
NaBH4 (47 mg, 1.24 mmol) at 0 °C over 5 min. After the
mixture was stirred at room temperature for 1.5 h, brine (3
mL) was added. After being stirred for 20 min, the solution
was concentrated and extracted with Et2O (6 mL × 6). The
organic layer was dried over Na2SO4 and concentrated. The
Su btilisin -Ca ta lyzed Kin etic Resolu tion s of 1a . General
procedure for the subtilisin-catalyzed kinetic resolutions of 1a
is similar to that for the lipase-catalyzed kinetic resolutions.
After a mixture of 1a (28 mg, 42 µmol) and ChiroCLEC-BL
(40 mg) in dry i-Pr2O (30 mL) was stirred for 30 min in a water
bath thermostated at the reaction temperature, vinyl acetate
(112 mg, 1.30 mmol) was added. The mixture was stirred at
ca. 450 rpm in the dark. The progress of the reaction was
(31) Kitayama, K.; Uozumi, Y.; Hayashi, T. J . Chem. Soc., Chem.
Commun. 1995, 1533.
(32) Salomon, R. G.; Reuter, J . M. J . Am. Chem. Soc. 1977, 99, 4372.
(33) Soai, K.; Hori, H.; Kawahara, M. Tetrahedron: Asymmetry
1990, 1, 769.

Modeling Hydrolase-Catalyzed Transesterifications
J . Org. Chem., Vol. 67, No. 7, 2002 2151
product was purified by silica gel column chromatography
(Et2O) to afford (R)-5 as a colorless oil (303 mg, 62%, >98%
ee): bp 76 °C (4 mmHg); [R]26D ) +124 (c 1.02, EtOH); 1H NMR
δ 1.52 (d, J ) 6.6 Hz, 3H), 2.04 (s, 3H), 2.37 (br s, 1H), 4.63 (s,
2H), 5.85 (q, J ) 6.6 Hz, 1H), 7.33 (s, 4H); 13C NMR δ 21.2,
22.1, 64.7, 72.1, 126.2, 127.0, 140.6, 140.9, 170.4; IR (neat)
3418, 1737 cm-1; MS (EI) m/z 194 (M+). HPLC: Chiralpak AD,
hexane/2-PrOH ) 95:5, flow rate 0.5 mL/min, detection 254
nm, (R) 23.4 min, (S) 27.2 min.
ee; 1H NMR δ 1.48 (d, J ) 6.5 Hz, 3H), 2.11 (br s, 1H), 2.36 (s,
3H), 4.85 (q, J ) 6.5 Hz, 1H), 7.17 (d, J ) 8.1 Hz, 2H), 7.27 (d,
J ) 8.1 Hz, 2H); 13C NMR δ 21.0, 25.0, 70.1, 125.3, 129.1,
137.0, 142.8; IR (neat) 3358 cm-1; MS (EI) m/z 136 (M+). GC:
CP-cyclodextrin-â-2,3,6-M-19, injection temperature 280 °C,
column temperature 120 °C, detection temperature 220 °C,
carrier gas He, (R) 13.1 min, (S) 14.1 min.
(R)-5-[4-(1-Acetoxyeth yl)ph en yl]-10,15,20-tr iph en ylpor -
p h yr in ((R)-2a ). To a solution of (R)-4 (98 mg, 0.51 mmol),
benzaldehyde (161 mg, 1.52 mmol), and pyrrole (135 mg, 2.01
mmol) in dry CH2Cl2 (200 mL) was added BF3‚OEt2 (28 mg,
0.20 mmol). After the mixture was stirred at room temperature
for 3 h in the dark, o-chloranil (339 mg, 1.38 mmol) was added.
After being stirred for 1 h, the solution was concentrated to
ca. 50 mL and washed with saturated aqueous NaHCO3 (15
mL). The organic layer was separated, dried over MgSO4, and
concentrated. The product was purified by repeated silica gel
column chromatography (hexane/CHCl3/EtOAc (2:1:0) to (10:
5:2)) to afford (R)-2a as a purple solid (43 mg, 12%, >98% ee).
(R)-4-(1-Acetoxyeth yl)ben zyl Br om id e ((R)-6). To a
solution of (R)-5 (254 mg, 1.31 mmol) and CBr4 (650 mg, 1.96
mmol) in dry CH2Cl2 (6.5 mL) was added portionwise PPh3
(516 mg, 1.97 mmol) at 0 °C over 5 min. The solution was
stirred at room temperature for 3.5 h. The solvent was removed
under reduced pressure. The products were dissolved in Et2O
and filtered (3 mL × 5). The filtrate was concentrated under
reduced pressure. Purification by silica gel column chroma-
tography (hexane/Et2O (5:1)) afforded (R)-6 as a colorless oil
(313 mg, 93%, >98% ee): bp 50 °C (4 mmHg); [R]26 ) +99.1
D
1
(c 1.10, EtOH); H NMR δ 1.52 (d, J ) 6.6 Hz, 3H), 2.07 (s,
3H), 4.48 (s, 2H), 5.87 (q, J ) 6.6 Hz, 1H), 7.32 (d, J ) 8.6 Hz,
2H), 7.38 (d, J ) 8.6 Hz, 2H); 13C NMR δ 21.3, 22.1, 33.1, 71.8,
126.5, 129.2, 137.3, 141.9, 170.2; IR (neat) 1732 cm-1; MS (EI)
m/z 256, 258 (M+). Anal. Calcd for C11H13O2Br: C, 51.38; H,
5.10. Found: C, 51.54; H, 5.23. HPLC: Chiralpak AD, hexane/
2-PrOH ) 95:5, flow rate 0.3 mL/min, detection 254 nm, (R)
17.1 min, (S) 20.7 min.
Ack n ow led gm en t. We are grateful to Amano Phar-
maceutical Co. and Boehringer Mannheim GmbH for
providing us with lipase PS and CHIRAZYME L-2, L-9,
respectively. We are grateful to the SC-NMR Laboratory
of Okayama University and the MS Laboratory of
Faculty of Agriculture, Okayama University for the
measurements of NMR and mass spectra, respectively.
We also thank Mr. T. Nitoda and Mr. K. Murai for the
measurement of HR FAB-MS and EI-MS, respectively.
This work was partially supported by a Grant-in-Aid
for Scientific Research from the Ministry of Education,
Science, Sports and Culture of J apan.
(R)-1-(p-Tolyl)eth a n ol ((R)-7). To a suspension of LiAlH4
(28 mg, 0.74 mmol) in dry Et2O (3 mL) was added dropwise a
solution of (R)-6 (128 mg, 0.50 mmol) in dry Et2O (2 mL) under
N2 at 0 °C over 5 min. The mixture was stirred at room
temperature for 2 h. EtOAc (0.5 mL) was added dropwise at 0
°C over 5 min, and the mixture was stirred for 20 min. After
brine (1 mL) was added to the mixture, the organic layer was
separated. The aqueous layer was extracted with Et2O (3 mL
× 5). The combined organic layer was dried over Na2SO4 and
concentrated. Purification by silica gel column chromatography
(hexane/Et2O (4:1)) afforded (R)-7 as a colorless oil (48 mg,
Su p p or t in g In for m a t ion Ava ila b le: 1H and 13C NMR
spectra for 1a , 1b, 2a , 2b, and 3-7. This material is available
free of charge via the Internet at http://pubs.acs.org.
70%, >99% ee): bp 47 °C (5 mmHg); [R]29 ) +57.8 (c 1.10,
D
CHCl3), lit.31 [R]25D ) +44.7 (c 1.27, CHCl3) for (R)-7 with 89%
J O0109063