Tuning lipase enantioselectivity in organic media using solid-state buffers

Article abstract of DOI:10.1021/jo0101104

The enantioselectivity exhibited by Candida antarctica lipase B (CALB) in predominantly organic media has been studied for different enzyme protonation states. Alcoholysis of (±)-2-phenyl-4-benzyloxazol-5(4H)-one (1) using butan-1-ol as the nucleophile in low-water organic solvents was used as a model reaction. Using either organo-soluble bases or the newly introduced solid-state buffers of known pK_a, the protonation state of the lipase was altered. By choice of the appropriate solid-state buffer or organic base, the enantioselectivity could be selectively tuned. Both Et₃N and the solid-state buffer pair CAPSO/CAPSO.Na were found to increase the enantioselectivity of the reaction catalyzed by CALB and that of another lipase (Mucor miehei). Significant differences to both the enantioselectivity and catalytic rate were observed, especially under hydrated conditions where byproduct acid was formed.

Full text of DOI:10.1021/jo0101104

5074
J . Org. Chem. 2001, 66, 5074-5079
Tu n in g Lip a se En a n tioselectivity in Or ga n ic Med ia Usin g
Solid -Sta te Bu ffer s
Margarita Quiro´s,† Marie-Claire Parker,*,† and Nicholas J . Turner‡
Department of Chemistry, University of Glasgow, Glasgow, G12 8QQ,U.K., and Edinburgh Centre for
Protein Technology, Department of Chemistry, Kings Buildings, University of Edinburgh,
Edinburgh, EH9 3J J , U.K.
mariec@chem.gla.ac.uk
Received J anuary 29, 2001
The enantioselectivity exhibited by Candida antarctica lipase B (CALB) in predominantly organic
media has been studied for different enzyme protonation states. Alcoholysis of (()-2-phenyl-4-
benzyloxazol-5(4H)-one (1) using butan-1-ol as the nucleophile in low-water organic solvents was
used as a model reaction. Using either organo-soluble bases or the newly introduced solid-state
buffers of known pKa, the protonation state of the lipase was altered. By choice of the appropriate
solid-state buffer or organic base, the enantioselectivity could be selectively tuned. Both Et3N and
the solid-state buffer pair CAPSO/CAPSO.Na were found to increase the enantioselectivity of the
reaction catalyzed by CALB and that of another lipase (Mucor miehei). Significant differences to
both the enantioselectivity and catalytic rate were observed, especially under hydrated conditions
where byproduct acid was formed.
In tr od u ction
control of ionization-state in organic media.5 It is well-
known that lyophilized powders of enzymes suspended
in organic media show a pH “memory effect”, i.e., the
activity of the enzyme can be correlated with the last
aqueous pH to which the enzyme was exposed.6 However,
in many enzymatic reactions there may be unrecognized
changes to the acid-base conditions due to, for example,
the production of acidic side products. The result of such
changes to acid/base conditions is erasure of pH memory.7
This erasure is reflected in an altered protonation-state
of either ionizable groups directly involved in catalysis
or those residues elsewhere that play a key role in the
overall retention of the active enzyme conformation. It
has recently been reported that a convenient and practi-
cal way to avoid such problems is to use solid-state
buffers5a,e in the reaction mixture. These buffers are able
to reset and then fix the enzyme ionization-state. So far
these buffers have only been extended to enhancing the
catalytic activity of serine proteases in a model system.5e,8
Recently, we have shown that the CALB-catalyzed
dynamic resolution of 4-substituted oxazol-5(4H)-ones
was critically dependent on reaction conditions9 and could
be improved markedly on addition of a catalytic amount
of a nonreactive organic base such as Et3N to the reaction
Much effort, over the past decade has focused on the
study of the physicochemical behavior of isolated enzymes
in organic solvent environments and their application to
the efficient synthesis of biologically active pharmaceuti-
cal intermediates. In this sense, lipases have attracted
much interest since they have proved to be powerful tools
for the resolution of a wide variety of chiral compounds
with potential industrial applications.1 To this end, a
range of strategies which aim to engineer the micro-
environment of the lipase are currently employed to
optimize lipase activity and stereoespecificity in organic
solvents. Relevant factors include the degree of protein
hydration,2 nature of the organic solvent,3 modification
of substrates,4 and ionization-state of the enzyme.5 With
the exception of ionization-state studies, most of these
parameters have been studied and reported at length.
There are, however, few reports concerned with the
* To whom correspondence should be addressed. Fax: (+44) 141-
330-4888.
† University of Glasgow. E-mail for M. Quiro´s: margarq@
chem.gla.ac.uk.
‡ University of Edinburgh. Fax: (+44) 131-650-4717/-4743. E-
mail: n.j.turner@ed.ac.uk.
(1) (a) Rubin, B.; Dennis, E. A. Lipases. In Methods in Enzymology.
Part A: Biotechnology, Vol. 284, 1997, and Part B: Enzyme Charac-
terisation and Utilisation, Vol. 286, 1997. (b) Ortaggi, G.; J aeger, K.-
E., Eds. J . Mol. Catal. B: Enzym. (Special issue: Microbial Lipases in
Biocatalysis) 1997, 3, 1. (c) Santaniello, E.; Ferraboschi, P.; Grisenti,
P.; Manzocchi, A. Chem. Rev. 1992, 92, 1071-1140. (d) Schmid, R. D.;
Verger, R. Angew. Chem., Int. Ed. 1998, 37, 1609-1633.
(2) (a) Halling, P. J . Enzyme Microb. Technol. 1994, 16, 178-206.
(b) Parker, M. C.; Blacker, A. J .; Moore, B. D. Biotechnol. Bioeng. 1995,
46, 452-458. (c) Halling, P. J . Enzyme Microb. Technol. 1984, 6, 513-
516.
(3) (a) Sakurai, T.; Margolin, A. L.; Russell, A. J .; Klibanov, A. M.
J . Am. Chem. Soc. 1988, 110, 7236-7237. (b) Parida, S.; Dordick, J .
S. J . Am. Chem. Soc. 1991, 113, 2253-2259. (c) Zaks, A.; Klibanov, A.
M. Science 1984, 224, 1249-1251. (d) Dordick, J . Enzyme Microb.
Technol. 1989, 11, 194-211.
(4) (a) Santaniello, E.; Chiari, M.; Ferraboschi, P.; Trave, S. J . Org.
Chem. 1988, 53, 1567-1569. (b) Itoh, T.; Takagi, Y.; Nishiyama, S. J .
Org. Chem. 1991, 56, 1521-1524.
(5) (a) Zacharis, E.; Moore, B. D.; Halling, P. J . J . Am. Chem. Soc.
1997, 119, 12396-12397. (b) Blackwood, A. D.; Curran, L. J .; Moore,
B. D.; Halling, P. J . Biochim. Biophys. Acta 1994, 1206, 161-165. (c)
Halling, P. J .; Blackwood, A. D.; Moore, B. D. Ann. N. Y. Acad. Sci.
1996, 799, 251-256. (d) Xu, K.; Klibanov, A. M. J . Am. Chem. Soc.
1996, 118, 9815-9819. (e) Harper, N.; Dolman, M.; Moore, B. D.;
Halling, P. J . Chem.: A Eur. J ., in press.
(6) (a) Zaks, A.; Klibanov, A. M. Proc. Natl. Acad. Sci. U.S.A. 1985,
82, 3192-3196. (b) Zaks, A.; Klibanov, A. M. J . Biol. Chem. 1988, 263,
8017-8021. (c) Yang, Z.; Zacherl, D.; Russel, A. J . J . Am. Chem. Soc.
1993, 115, 12251-12257.
(7) Valivety, R. H.; Rakels, J . L. L.; Blanco, R. M.; J ohnston, G. A.;
Brown, L.; Suckling, C. J .; Halling, P. J . Biotechnol. Lett. 1990, 12,
475-480.
(8) Patridge J .; Halling, P. J .; Moore, B. D. J . Chem. Soc., Perkin
Trans. 2 2000, 465-471.
(9) Parker, M. C.; Brown, S. A.; Robertson, L.; Turner, N. J . Chem.
Commun. 1998, 2247-2248.
10.1021/jo0101104 CCC: $20.00 © 2001 American Chemical Society
Published on Web 06/22/2001

Tuning Lipase Enantioselectivity in Organic Media
J . Org. Chem., Vol. 66, No. 15, 2001 5075
Sch em e 1
medium. 5(4H)-Oxazalones undergo spontaneous hy-
drolysis and racemize with ease;10 the role of the base is
not, however, to facilitate this racemization. CALB is able
to accept a wide substrate range, has proven to be very
stereoselective for the resolution of secondary alcohols,11
and thus is an ideal candidate for our studies.
In this paper, we have shown that it is possible to
selectively tune both the catalytic activity and more
importantly the enantioselectivity by changing the pro-
tonation-state of an enzyme in organic media. Using
lipase B from Candida antarctica (CALB, Chirazyme L-2)
and another lipase (Mucor miehei (Lipozyme, RM IM)),
the enantioselectivity in organic media could be selec-
tively altered and reset using solid-state organic buffers
that modify the ionization state of proteins in situ.
Resu lts a n d Discu ssion
The CALB-catalyzed alcoholysis of (()-2-phenyl-4-
benzyl-5(4H)-oxazolone (1) is shown in Scheme 1; the (S)-
N-Benzoylphenylalanine butyl ester is obtained as the
major enantiomer. This reaction is an example of a
dynamic resolution (potential of 100% yield and 100% ee)
and ultimately yields optically active R-amino esters that
are appropriate precursors of R-amino acids.12 Thus, a
wide variety of side chains can be used in order to obtain
a range of unnatural synthetic amino acids.13
To investigate our initial observation9 of enhanced
enantioselectivity in the presence of Et3N in more detail,
we studied systematically for the first time a range of
organo-soluble bases of varying pKa. In parallel with this
we also decided to investigate the use of solid-state
buffers that could buffer the ionization state of CALB in
situ within a given pKa range.
F igu r e 1. Effect of adding acids on (a) the catalytic activity
and (b) enantioselectivity of CALB-butanolysis of (()-2-phenyl-
4-benzyl-5(4H)-oxazolone (1). The figures show the results
obtained in the presence of N-benzoylphenylalanine (3) (9),
acetic acid (0), Et3N (b), Et3N, and (3) (O), Et3N, and acetic
acid (1) and additive free (3). The amount of Et3N added was
14 mol % and that for both acids was 24 mol %. The reactions
were carried out in dry toluene (aw ≈ 0.0) at 37 °C.
from an initial value of E ) 24 (92% ee) to E ) 4 (59%
ee). Since the presence of acetic acid attenuates the
activity and E-value in a similar way to that of N-
benzoylphenylalanine, chiral inhibition can be ruled out
as a possible mechanism.
We had previously shown that Et3N increases the
enantioselectivity of CALB in organic solvents by ion-
pairing byproduct acid, thereby removing acid from the
enzyme microenvironment. Other researchers, too, have
also shown for a number of different lipases that poor E
can be overcome by the addition of an organic base.14
However, there has been no systematic investigation of
the effect of the base pKa on the enantioselective behavior
of lipases. We, therefore, decided to use a range of bases
However, it was first important to establish if the
inactivation and reduction to the enantiomeric ratio (E)
that we observed during the course of a reaction was due
to chiral inhibition by the acid generated. To investigate
this a dry preparation of CALB in toluene (aw ∼ 0) that
contained either trace byproduct acid (3) (N-benzoyl-
phenylalanine) or a nonspecific acid, such as acetic acid,
was assayed. From Figure 1a,b it can be seen that the
presence of acetic acid significantly reduces both the
catalytic activity (∼50%) and enantioselectivity of CALB
15
of differing pKa to investigate this. The bases we chose
for study were N,N-dimethylaniline (DMA, pKa ) 4.38),
N,N-4-(dimethylamino)pyridine (DMAP; pKa 6.09), tri-
ethylamine (Et3N, pKa ) 10.93), and N,N-diisopropyl-
amine (DIPA, pKa ) 11.05). The reactions were all carried
out in dry toluene (aw ∼ 0.0), and the organic base was
added at the start of each reaction. The results obtained
are shown in Figure 2a,b. Et3N, DMAP,16 and DIPA
(10) (a) Goodman, M.; Levine, L. J . Am. Chem. Soc. 1964, 86, 2918-
2922. (b) Bevinakatti, H. S.; Banerji, A. A.; Newadkar, R. V.; Mokashi,
A. A. Tetrahedron: Asymmetry 1992, 3, 1505-1508. (c) Bevinakatti,
H. S.; Newadkar, R. V.; Banerji, A. A. J . Chem. Soc., Chem. Commun.
1990, 1091-1092. (d) Gu, R.-L.; Lee, I.-S.; Sih, C. J . Tetrahedron Lett.
1992, 33, 1953-1956. (e) Benoiton, N. L. The Peptides 1983, 5, 246.
(11) (a) Uppenberg, J .; O¨ hrner, N.; Norin, M.; Hult, K.; Kleywegt,
G. J .; Patkar, S.; Waagen, V.; Anthonsen, T.; J ones, T. A. Biochemistry
1995, 34, 4, 16838-16851. (b) Rottici, D.; Hæffner, F.; Orrenius, C.;
Norin, T.; Hult, K. J . Mol. Catal. B: Enzymol. 1998, 5, 267-272. (c)
Frykman, H.; O¨ hrner, N.; Norin, T.; Hult, K. Tetrahedron Lett. 1993,
34, 1367-1370.
(14) (a) Theil, F. Tetrahedron 2000, 56, 2905-2919. (b) Berger, B.;
Rabiller, C. G.; Ko¨nigsberger, K.; Faber, K.; Griengl, H. Tetrahedron:
Asymmetry 1990, 1, 541-546. (c) Rakels, J . L. L.; Straathof, A. J . J .;
Heijen, J . J . Tetrahedron: Asymmetry 1994, 5, 93-100. (d) Stead, P.;
Marley, H.; Mahmoudian, M.; Webb, G.; Noble, D.; Ip, Y. T.; Piga, E.;
Rossi, S.; Roberts, S. M.; Dawson, M. J . Tetrahedron: Asymmetry 1996,
7, 2247-2250.
(12) Turner, N. J .; Winterman, R.; McCague, J . S.; Parratt, J . S.;
Taylor, S. J . C. Tetrahedron Lett. 1995, 36, 1113-1116.
(13) Brown S. A.; Parker M. C.; Fisher A.; Turner N. J . Tetrahe-
dron: Asymmetry 2000, 11, 1687-1690.
(15) Lide, D. R., Ed. CRC Handbook of Chemistry and Physics;
CRC: London (U.K.), 1998; pp 7-9, 8-46.

5076 J . Org. Chem., Vol. 66, No. 15, 2001
Quiro´s et al.
maintain a constant E-value; those of low pKa are unable
to maintain/control CALB enantioselectivity.
It is, however, difficult to quantify the effect of base
pKa on the enantioselectivity of CALB for a number of
reasons. First, base solubility is solvent dependent, and
this will result in differential binding of base to the
enzyme surface and at or near the active-site. Also the
base strength will depend on solvent polarity and protic/
nonprotic nature.17
Potentially anomalous results associated with using
organic bases can be ruled out if the enzyme protonation-
state is altered in situ by a solid-state organic buffer. The
solid-state buffers used in this study are zwitterionic
biological buffers of the kind commonly used for control-
ling aqueous pH. As shown in eq 1, the acid and sodium
salt together control the ionization-state of the enzyme’s
carboxylic groups. The buffer cannot bind to the protein
surface/active site but can only exchange ions (Na+/H+),
and therefore any changes to the E-value can be at-
tributed to an altered enzyme protonation-state. Solid-
state buffers can therefore be used in systems where
reactants or products with acid-base properties are
produced. These buffers have a significant advantage
over the organo-soluble bases: they are present in the
solvent as crystalline solids and will generate a single
fixed value of exchange potential at equilibrium regard-
less of the solvent since they have a fixed thermodynamic
activity.
F igu r e 2. Effect of different organo-soluble bases on the
catalytic activity (a) and enantioselectivity (b) of CALB-
butanolysis of (()-2-phenyl-4-benzyl-5(4H)-oxazolone. The bases
added were Et3N (O), iPr2NH (1), DMAP (3), N,N-dimethyl-
aniline (9), and in the absence of base (b). A catalytic amount
(14 mol %) of the given base was added at the beginning of
the reaction. The reactions were carried out in dry toluene (aw
≈ 0.0) at 37 °C.
With this aim in mind, we screened a range of biologi-
cal buffers of differing acid-base strength: CAPSO (9.7);
AMPSO (9.3); HEPPSO (7.8); HEPES (7.5); MOPS (7.2),
and PIPES (6.8). The pKa values of these buffers in
aqueous solution are given in parentheses.15 Prior to
addition of the substrates, the solid-state buffer and the
corresponding sodium salt were preequilibrated with the
enzyme. The results obtained are shown in Figures 3a,b.
We observed (Figure 3b) that as the aqueous pKa of
the buffer increases, so too does the E-value, with the
exception of HEPES/HEPES.Na and HEPPSO/HEPP-
SO.Na. Harper et al. have recently used an organo-
soluble indicator5e to screen for appropriate buffer pairs
that would be within the correct pKa for controlling the
catalytic activity of the serine protease, subtilisin Carls-
berg. In the presence of these buffers, they found an
anomalous indicator response using either HEPES/
HEPES.Na or HEPPSO/HEPPSO.Na that may also
provide an explanation for our observations.
We can also re-set a low E-value to a higher one, using
these solid-state buffers. An initially low E-value, as a
result of the presence of DMAP (pKa 6.09) (E ) 7, 75%
ee) can be reset in situ to a higher level (E ) 10, 81% ee)
by the addition of a solid-state buffer of higher pKa
(CAPSO, pKa 9.7). Full recovery of CALB enantioselec-
tivity to the level observed with CALB and CAPSO/
CAPSO.Na is not found, which may be as a result of
residual DMAP bound to the protein.
accelerate the reaction rate by approximately 10-fold over
that observed for the CALB-catalyzed reaction alone. The
low pKa base, N,N-dimethylaniline marginally reduces
the enzyme-catalyzed reaction rate compared to additive-
free CALB.
Figure 2b shows the effect of the base pKa on the
E-value for the product (S) butyl ester obtained from the
CALB-catalyzed reaction. For a reaction containing no
organic base, we observe an initially high E-value that
steadily decreases with time. Remarkably, this decrease
can be arrested immediately when a base of high pKa
(DIPA (11.05) or Et3N (10.93)) is added. For these
systems, control of the E-value is obtained and remains
constant and high throughout the course of the reaction
(24 h). Switching to a base of low pKa fails to control the
enantioselectivity, as observed for a reaction containing
DMA (pKa 4.38).
The highest E value was observed with Et3N (E ) 32,
94% ee). Addition of DIPA also results in enhancement
and stabilization to the E-ratio (E ) 24, 92% ee), as found
for Et3N. Addition of DMAP leads to the lowest E-value
of all the bases tested, E ) 7 (75% ee). In this case, the
E-value is attenuated to a low but constant value. DMA
reduces the E-value over time, to a value comparable with
that found for CALB alone. It can be seen from these data
that there is a correlation between CALB enantioselec-
tivity and base pKa: bases of high pKa control and
As discussed previously, solid-state buffers will produce
a fixed ionization-state regardless of the solvent, and
indeed when we use t-BuOMe as the reaction solvent,
(16) Maugard, T.; Remaud-Simeon, M.; Monsan, P. Biochim. Bio-
phys. Acta 1998, 1387, 177-183.
(17) Davis, M. M. Acid Behavior in Aprotic Organic Solvents; Nat.
Bur. Stand. Monograph No. 105, Washington, D. C., 1968.

Tuning Lipase Enantioselectivity in Organic Media
J . Org. Chem., Vol. 66, No. 15, 2001 5077
F igu r e 4. Catalytic performance of CALB in t-BuOMe.
Catalytic activity (a) and enantioselctivity (b) in the presence
of Et3N (O), DMAP (1), the biological buffer CAPSO/sodium
salt (3), and additive free (b). The reactions were carried out
at 37 °C.
F igu r e 3. Effect of solid-state buffers on catalytic activity (a)
and enantioselectivity (b) of CALB-butanolysis of (1). Figures
show the results obtained after 4 h of reaction. Solid-state
buffers/sodium salts were preequilibrated with the enzyme
prior to the addition of substrates. The reactions were carried
out in dry toluene (aw ≈ 0.0) at 37 °C. The solid-state buffer
pairs were PIPES/PIPES.Na ([), MOPS/MOPS.Na (4), HEPES/
HEPES.Na (2), HEPPSO/HEPPSO.Na (0), AMPSO/AMP-
SO.Na (9), CAPSO/CAPSO.Na (O). Also, the results obtained
in the presence of Et3N (b) are shown. Reaction catalyzed by
CALB without any additive gave 19% product and an E-value
of 14.3. Standard deviations average over three measurements
for each one were between (1.00% to (2.50% for the % product
values, and (0.00% to (1.15% for the E-values.
Ta ble 1. Effect of Ad d itives on Lip a se-Ca ta lyzed
Dyn a m ic Resolu tion of 1
CALB
CALB-CLEC
M. miehei
additive
% Pa
% eeb
% Pa
% eeb
% Pa
% eeb
none
Et3N
CAPSO
19
62
43
87c
94
91
21
72
54
87d
95
93
43
47
32
48e
61
57
a
Percentage of N-benzoyphenylalanine butyl ester after 4 h of
reaction. ee of (S)-butyl ester (2) after 4 h of reaction. c Initial
b
we find a very similar trend to that observed in toluene,
thereby ruling out any specific solvent effects of toluene.
Figure 4a,b shows the effect of Et3N, DMAP, and CAPSO/
CAPSO.Na on the enantioselectivity of the CALB-
catalyzed (S)-butyl ester in t-BuOMe. Without the addi-
tion of any of these additives, there is a decrease to the
E-value when monitored over 24 h from E ) 7 (75% ee)
to E ) 2 (33% ee).
Together, with the possibility of a specific solvent effect,
we also considered the posibility of interference in
binding of base/buffer to the support material used in the
preparation of CALB. Therefore, we used a pure prepara-
tion of CALB supplied as cross-linked enzyme crystals
(CLEC-CALB). Similarly, we probed the response to
changes in E-value of a different lipase (from Mucor
miehei) using either Et3N or CAPSO/CAPSO.Na, Lipase
from Mucor miehei has been reported by several groups14a
to be activated in the presence of Et3N. From Table 1 it
can be seen that the support has no affect on CALB
enantioselectivity. Although the intrinsic selectivity of the
Mucor miehei lipase is lower than that found for CALB,
both Et3N and CAPSO/CAPSO.Na enhance both the
activity and selectivity.
value 92%. Initial value 91%. e Initial value 59%.
d
Since most enzyme-catalyzed reactions in organic
solvents are not practicably carried out under anhydrous
conditions, we tested the catalytic performance of CALB
as a function of the thermodynamic water activity (aw).18
Figure 5a,b shows the effect of either CAPSO/CAPSO.Na
or Et3N on the catalytic activity and E-value, respectively.
We can see that in all cases the catalytic rate decreases
with increasing hydration level. This effect is offset by
Et3N, and only under dry or low hydration conditions (aw
∼ 0, 0.11) by CAPSO/CAPSO.Na. In Figure 5b the
E-value (t ) 24 h) is plotted as a function of aw. Whereas,
the E-value decreases significantly as a function of aw,
remarkably both Et3N and CAPSO maintain higher E
values over the water activity range. Reactions were not
carried out at aw > 0.86 as a small amount of CAPSO.Na
is likely to dissolve in the organic solvent, and this will
(18) Hutcheon, G. A., Halling, P. J ., Moore, B. D. Measurement and
control of hydration in nonaqueous biocatalysis. Methods in Enzymol-
ogy; Rubin, B.; Dennis, E. A., Eds.; Academic Press: New York, 1997;
Lipases Part B, Vol. 286, pp 465-472.

5078 J . Org. Chem., Vol. 66, No. 15, 2001
Quiro´s et al.
indicates that both the solid-state buffers and the organo-
soluble bases (of sufficiently high pKa) are acting via the
same mechanism. A hydrated system, which results in
poor enantioselectivity, leads to an increase in the rate
of the R-enantiomer and a decrease in the catalytic rate
of the S-enantiomer relative to that of CALB alone. When
the medium is acidifed, the same effect is found.
There are a number of possible reasons for the observed
effects. The ability of bases and solid-state buffers to
selectively modulate the activity and enantioselectivity
may be attributable to control of the protonation of the
catalytic His-224. Protonation of this residue by acidic
species may result in modulation of the activity and
enantioselectivity. Kazlauskas and co-workers19 have
suggested a structural basis for the chiral preference of
lipases. High resolution 3D structures derived from X-ray
data show that the interactions between the menthyl ring
of the slow-reacting enantiomer (1S)-menthyl hexylphos-
phonate and the histidine of the catalytic triad disrupt
the hydrogen bond between the Nꢀ2 of the imidazole ring
and the menthol oxygen atom, most likely accounting for
the slower reaction of the (1S)-enantiomer.
Uppenberg et al.11a have determined the three-dimen-
sional structure of CALB cocrystallized with the deter-
gent Tween 80 and in another structure in complex with
a racemic mixture of a covalently bound phosphonate
inhibitor. They observed that as the pH was raised, the
short R5 helix (a potential lid forming R-helix) became
disordered, as manifested by a disappearance of the
electron density for this helix. Therefore, this raises the
possibility that the disordered state of the helix under
conditions where deprotonation occurs in organic media
also results in an increase of disorder of the helix (i.e.,
an ionization dependence) thereby adopting a more open
conformation for substrate access.
F igu r e 5. Catalytic performance of CALB as a function of
thermodynamic water activity (aw). Catalytic activity (a) and
enantioselectivity (b) obtained after 24 h of reaction in the
presence of Et3N (9) and CAPSO/sodium salt (3) and additive
free (b) at different aw. The reactions were carried out in dry
toluene at 37 °C.
therefore adversely affect the acid-base behavior of the
buffer at aw > 0.86.5e
Con clu sion
Despite the large number of studies14 that have re-
ported a qualitatively beneficial effect of Et3N on a range
lypolytic enantioselectivity, there have been no system-
atic studies aimed at quantifying and establishing the
molecular basis for this effect. We have therefore sought
to address this issue via controlled protonation-state
changes to CALB. This enzyme has proven to be a potent
biocatalyst for the stereoselective kinetic resolution of
secondary alcohols and is also able to accept a broad
substrate range.11
We have shown that a simple strategy for controlling
the protonation state of a lipase, Candida antarctica
lipase B, in organic solvents is to use solid-state buffers
suspended in the reaction media or soluble organic bases
of high aqueous pKa. The enantioselectivity and the
catalytic activity of CALB can be substantially improved
by either addition of a base or a solid-state buffer of high
pKa (Et3N/CAPSO.CAPSO.Na).
The method may therefore provide a reliable strategy
for the prediction and control of enantioselectivity of
enzymes that employ general acid-base chemistry for
their catalytic processes.
The propensity for organic bases to increase the
catalytic activity of lipases has in some cases14c,16 been
attributed to the formation of an ion-pair between the
base and acid byproduct. The rationale and evidence for
this is that the base removes acid from the enzyme’s
microenvironment and restores catalytic function.9 Even
minor amounts of an acidic component present in the
solvent will result in protonation and reversible loss to
catalytic activity as demonstrated using a chromoiono-
phoric indicator that is sensitive to protonation changes.5e
To gain an insight into the mechanism of enhancement
or attenuation to the enantioselectivity of CALB, we have
compared the initial rate for each of the enantiomers
under a given set of conditions. We observed that when
the additive to the system is either organic base or a
solid-state buffer (of high pKa, e.g., CAPSO) both the
catalytic rate of the R- and favored S-butyl ester increase
relative to that of CALB with no additive present. This
Exp er im en ta l Section
Rea gen ts a n d Solven ts. The lipases, Chirazyme L-2, c-f.
C2, lyo (Candida antarctica lipase B) and Lipozyme, RM IM
(Mucor miehei lipase) were received as an immobilized prepa-
ration (Boehringer Mannheim, Germany and Novo Nordisk,
Denmark respectively) and were dehydrated over P2O5 (at
room temperature) for 4 days before use. Cross-linked enzyme
crystals of Candida antarctica lipase, CALB-CLEC were
obtained from Altus Biologics Inc. (Cambridge, MA) as an
aqueous suspension. The substrate, (()-2-phenyl-4-benzyl-
oxazol-5(4H)-one (1) was prepared from dl-phenylalanine as
previously described.12 nBuOH was purchased from Aldrich
(19) 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.

Tuning Lipase Enantioselectivity in Organic Media
J . Org. Chem., Vol. 66, No. 15, 2001 5079
(HPLC grade). Inorganic salts (analytical grade) were obtained
from Aldrich U.K. and used without further purification.
Biological buffers: AMPSO (3-[(1,1-dimethyl-2-hydroxyethyl)-
amino]-2-hydroxypropanesulfonic acid), CAPSO (3-[cyclohexyl-
amino]-2-hydroxy-1-propanesulfonic acid), HEPES (N-[2-
hydroxyethyl]piperazine-N′-[2-ethanesulfonic acid]), HEPPSO
(N-[2-hydroxyethyl]piperazine-N′-[2-hydroxypropanesulfonic
acid]), MOPS (3-[N-morpholino]propanesulfonic acid), PIPES
(piperazine-N,N′-bis[2-ethanesulfonic acid), and their sodium
salts were purchased from Sigma. All other reagents were
obtained from Aldrich, Sigma (U.K.), Fisher (U.K), or Fluka
(U.K) and were of analytical grade. Solvents used in the
enzyme-catalyzed reactions were of HPLC grade, dried, and
stored over freshly reactivated 4 Å molecular sieves. The water
content of the dry solvents was determined using a 684 Karl
Fischer Coulometer coulometric titrator20 and found to be
<0.001 wt %. Hydrated solvents were also analyzed by Karl-
Fischer titration.
No product was detected when the reactions were carried
out in the absence of the biocatalyst.
En zym e P r ep a r a tion of CALB-CLEC for Use in Or -
ga n ic Solven t. The enzyme crystals were supplied and stored
in an aqueous solution. Prior to use, the aqueous CLEC
suspension (150 µL; 7.5 mg) was washed with propan-1-ol (4
× 1.3 mL) and then with butan-1-ol (2 × 0.6 mL). After each
wash, the enzyme suspension was centrifuged and the solvent
removed. Care was taken to ensure that the enzyme pellet
remained wet at all times. The enzyme was then used directly
in a reaction by resuspension in the appropriate organic
solvent.
En zym a tic Rea ction in th e P r esen ce of Or ga n ic Ba ses.
The substrate, (()-2-phenyl-4-benzyloxazol-5(4H)-one (1) (0.16
mmol; 40 mg), was placed in a screw top vial together with
the solvent (4 mL; either anhydrous or hydrated), butan-1-ol
(0.24 mmol; 21.9 µL), enzyme (CALB (40 mg), Mucor miehei
(5 mg) or CALB-CLEC (7.5 mg)) and the organic base (14 mol
%; Et3N, DMAP, N,N-dimethylaniline or N,N-diisopropyl-
amine). The reaction and subsequent analyses were performed
using the procedure described above.
In str u m en ts. Chiral HPLC analyses (Chiralcel-OD column
(Daicel 250 × 4.6 mm) were carried out on a Waters 2690
Separations module (Waters U.K.)
En zym e a n d Solven t Hyd r a tion . Vials containing chira-
zyme and solvent (toluene) were placed in separate sealed jars
containing a saturated salt solution for control of the thermo-
dynamic water activity (aw).18 Typically, samples of enzyme
were equilibrated for 4 days and solvent for 24 h, both at room
temperature.
En zym a tic Rea ction in th e P r esen ce of Or ga n ic Acid s.
The substrate, (()-2-phenyl-4-benzyloxazol-5(4H)-one (1) (0.16
mmol; 40 mg), was placed in a screw top vial together with
the solvent (4 mL; either anhydrous or hydrated), butan-1-ol
(0.24 mmol; 21.9 µL), enzyme (CALB (40 mg), Mucor miehei
(5 mg) or CALB-CLEC (7.5 mg)) and the acid (24 mol %;
N-benzoylphenylalanine (3) or acetic acid). The reactions were
carried out both in the presence and absence of Et3N (14 mol
%). The reactions and subsequent analyses were performed
using the procedure described above.
En zym a tic Rea ction Usin g Solid -Sta te Bu ffer s. The
solid-state buffers/sodium salt (25 mg of each form) were
equilibrated together with the catalyst (CALB (40 mg), Mucor
miehei (5 mg) or CALB-CLEC (7.5 mg)) for 3 h in the case of
CALB and CALB-CLEC and for 5 h in that of Mucor miehei
in dry toluene prior to the addition of the substrates and
initiation of the reaction. The substrates 1 (0.16 mmol; 40 mg)
and BuOH (0.24 mmol; 21.9 µL) were added to the reaction
vial and shaken at 250 rpm on a rotary shaker at 37 °C. The
progress of the reaction and ee of the product obtained were
monitored by HPLC (Chiralcel-OD) as previously described.
Typ ica l En zym a tic Rea ction . The substrate, (()-2-phen-
yl-4-benzyloxazol-5(4H)-one (1) (0.16 mmol; 40 mg), was placed
in a screw top vial together with the solvent (4 mL; either
anhydrous or hydrated), butan-1-ol (0.24 mmol; 21.9 µL), and
the enzyme (CALB (40 mg), Mucor miehei (5 mg) or CALB-
CLEC (7.5 mg)). The reaction vial was shaken at 250 rpm on
a rotatory shaker at 37 °C. The progress of the reaction and
the ee of the product (N-benzoylphenylalanine butyl ester) (2)
were monitored by HPLC using chiral chromatography (Chiral-
cel-OD), and the products were monitored by UV at λ ) 254
nm. Aliquots (50 µL) were periodically withdrawn and dis-
solved with hexane/propan-2-ol (90:10 v/v) to 1 mL final
volume, filtered, and then analyzed by HPLC as described. The
mobile phase composition was hexane/propan-2-ol (90:10 v/v).
The percentage product obtained was calculated using an
external standard method. A linear calibration curve was
constructed for the racemic butyl ester by preparing standard
solutions containing the pure racemic butyl ester at five
different known concentrations, covering that obtained for a
typical reaction. Chiral HPLC analyses: (S)- N-benzoylphenyl-
alanine butyl ester, tR 8.95 min; (()-N-benzoylphenylalanine
butyl ester, tR 5.67 and 8.69 min; RS ) 4.25 (hexane:propan-
2-ol, 90:10 (v/v); 0.8 mL/min). The enantioselectivity values
were calculated according to the literature.21
Ack n ow led gm en t. We are grateful to the BBSRC
for funding (M.Q.) and for a David Phillips Fellowship
(M.C.P.). We would also like to thank Dr. Barry D.
Moore, Prof. Peter J . Halling, and Prof. Karl Hult for
their advice and invaluable discussions.
J O0101104
(20) Laitennen, H. A.; Harris, W. E. Chemical Analysis, 2nd edn,
McGraw-Hill, 1975, 3, 61.
(21) Tan, D. S.; Gu¨nter, M. M.; Drueckhammer, D. G. J . Am. Chem.
Soc. 1995, 117, 9093-9094.