Diastereoselective synthesis of (R)-(alkyl)-β-d-galactopyranoside by using β-galactosidase (Aspergillus oryzae) in low-water media

Article abstract of DOI:10.1016/j.bmcl.2007.11.006

A β-galactosidase (from Aspergillus oryzae) preparation viz. EPRP (enzyme precipitated and rinsed with propanol), obtained by the removal of bulk water by precipitation with n-propanol, showed higher biological activity than the lyophilized powder. FT-IR study confirmed that EPRP had retained the α-helical content of the native structure better than the lyophilized form. Use of this formulation of β-galactosidase under low water conditions (temperature 55 °C, reaction time of 4 h) gave enantioselectivity, E > 1000 for the stereoselective synthesis of (R)-(1-phenylethyl)-β-d-galactopyranoside, starting from racemic 1-phenylethanol and d-galactose. For racemic 2-octanol also, EPRP worked better. Under similar conditions, (R)-(2-octyl)-β-d-galactopyranoside was formed with an enantioselectivity, E = 38.

Full text of DOI:10.1016/j.bmcl.2007.11.006

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 124–128
Diastereoselective synthesis of (R)-(alkyl)-b-D-galactopyranoside
by using b-galactosidase (Aspergillus oryzae) in low-water media
Abir B. Majumder, Bhupender Singh and Munishwar N. Gupta*
Chemistry Department, Indian Institute of Technology Delhi, Hauz Khas, New Delhi 110 016, India
Received 6 June 2007; revised 13 October 2007; accepted 1 November 2007
Available online 5 November 2007
Abstract—A b-galactosidase (from Aspergillus oryzae) preparation viz. EPRP (enzyme precipitated and rinsed with propanol),
obtained by the removal of bulk water by precipitation with n-propanol, showed higher biological activity than the lyophilized pow-
der. FT-IR study confirmed that EPRP had retained the a-helical content of the native structure better than the lyophilized form.
Use of this formulation of b-galactosidase under low water conditions (temperature 55 °C, reaction time of 4 h) gave enantioselec-
tivity, E > 1000 for the stereoselective synthesis of (R)-(1-phenylethyl)-b-D-galactopyranoside, starting from racemic 1-phenyletha-
nol and D-galactose. For racemic 2-octanol also, EPRP worked better. Under similar conditions, (R)-(2-octyl)-b-D-
galactopyranoside was formed with an enantioselectivity, E = 38.
Ó 2007 Elsevier Ltd. All rights reserved.
The products arising out of the combination of carbohy-
drates and other specific molecules are collectively
with a diastereomeric excess of 42% (reported R:S gluco-
6
side ratio was 0.4:1.0). The use of enzymes in low-water
7–9
1
,2
known as glycoconjugates. In view of their impor-
tance in many biological processes and as starting mate-
rials for many drugs, synthesis of glycoconjugates has
attracted considerable attention. The chemical synthesis
is generally a multi-step procedure with low yield and
contamination with unwanted enantiomer. The use of
glycosidases is an attractive alternative and can be car-
ried out in two ways. The frequently used way is to carry
out a transglycosylation where an activated glycosyl do-
media is now widely known. Unfortunately, as is now
also well established, the catalytic efficiency of the en-
zymes in such low-water media is quite low as compared
8
–10
to their performance in aqueous media.
improve their catalytic performance in non-aqueous
In order to
9,10
media many approaches have been described.
1
Recently, Bridiau et al. have looked at the optimization
of b-galactosidase catalyzed selective galactosylation of
a number of aromatic compounds. The % conversions
3
–5
nor is used. As pointed out by Vic and Crout, this is
4
both expensive as well as gives too many side products.
1
ranged from 9.5% to 96%. The worst conversion was
The second approach is to carry out reverse hydrolysis
in low-water media. Synthesis of alkyl-b-D-glycosides
from D-glucose or from D-galactose by reverse hydroly-
sis using b-glycosidases (glucosidase or galactosidase)
has been reported with alcohols which lack chiral cen-
obtained with 1-phenylethanol as the acceptor. Also
the resolution of (R,S)-1-phenylethanol has been exten-
sively investigated to evaluate and validate different en-
1
1,12
zyme based approaches for kinetic resolution.
This
is because enantiopure 1-phenylethanol is a chiral syn-
thon for synthesis of industrially important compounds
2
,4,5
ter.
However, the use of reverse hydrolysis in enzy-
1
3
matic enantioselective glycosylation with a racemic
mixture of a chiral alcohol giving a diastereomeric prod-
uct pair in low-water media has been reported with lim-
ited success. Vic et al. synthesized (S)-1-phenylethyl-b-D-
glucoside from D-glucose using almond-b-D-glucosidase
like primeveroside. Hence resolution of 1-phenyletha-
nol was taken up in the present work.
It is a well known practice to ‘pH tune’ the enzyme pow-
7
,9
der before using it in non-aqueous media. This con-
sists of dissolving the enzyme in a buffer at optimum
pH and lyophilizing the solution. Table 1 shows that
Keywords: Glycoconjugates; Galactoside synthesis; b-Galactosidase;
Reverse hydrolysis; Low-water media; Diastereoselective synthesis.
1
4
pH tuned lyophilized powder as a catalyst produced
1
5
2
3% conversion to (R)-(1-phenylethyl)-b-D-galactopy-
3,16
*
Corresponding author. Tel.: +91 11 26591503; fax: +91 11
6581073; e-mail: appliedbiocat@yahoo.co.in
ranoside (compound 3, Scheme 1) with 93% deP.
2
0
960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.11.006

A. B. Majumder et al. / Bioorg. Med. Chem. Lett. 18 (2008) 124–128
125
Table 1. Galactosylation of 1-phenylethanol with different formulations of b-galactosidase
À1 À1
a
C (%) (16 h)
b
b
S
c
Biocatalyst type
Initial rate (lmol mg
h
)
de
P
(%)
ee
(%)
E
Lyophilized
Colyophilized (with galactose)
EPRP
575
700
23
28
42
93
95
95
28
36
68
36
56
80
1050
14
Different formulations of b-galactosidase were used. The reaction was performed in acetonitrile, dimethyl formamide, water (in 9:0.5:0.5v/v/v ratio).
The minimum amount of water (5%, v/v) was required to dissolve the sugar under the reaction condition. Each reaction set was performed twice and
the deviations between the pair of readings were within 2%.
a
C (conversion) = ee /ee + de . Further incubation caused a drop in C value: from 16 h to 32 h, C dropped from 23% to 10%, 30% to 13%, and 42%
S S P
to 25%.
b
c
P S
de , ee stand for diastereomeric excess of the product (Scheme 1, 3 and 4) and enantiomeric excess of the unreacted substrate (Scheme 1, 2a,b),
respectively.
23
E = ln[1 À C(1 + ee
P
)]/ln[1 À C(1 À ee
P P P P
)], here de = eeP. Diastereomeic excess of the product (de ) can replace ee in Chen’s equation as there is
only a pair of diastereomeric products with b-stereochemistry in the sugar part and differing in the configuration of the aglycon portion due to
enantioselection from racemic 1-phenylethanol.
Scheme 1.
2
5
Another strategy which has been used for improving the
reaction rates in non-aqueous media is the use of sugars
Table 2. FT-IR analysis of amide I band of b-galactosidase enzyme
formulations
9
a
Peaks (cm )
À1
as lyoprotectants during lyophilization. Presence of
such additives during lyophilization minimizes the struc-
tural changes which proteins are known to undergo dur-
Type
FD (%)
EPRP (%)
1
1
1
610 ± 1
635 ± 1
654 ± 2
Side chain residues
b-Sheet
a-Helix
Contribution of turn
and b-segments
Turn
14
45
10
23
14
34
21
16
1
7
ing lyophilization. In the present instance, as the
substrate galactose itself is a sugar, it was a natural
choice as a lyoprotectant during lyophilization. More-
over, it has also been pointed out that the presence of
a substrate during lyophilization may also minimize
1670 ± 2
1688 ± 2
11
15
The freeze-dried (FD) or lyophilized and EPRP formulations of the
enzyme were analyzed in the solid form and the changes in the sec-
ondary structure were recorded (see Figs. 2a and b for the FT-IR
spectra in the Supplementary data).
1
8
structural damage during lyophilization. Thus, it was
hoped that for both reasons, lyophilization of the en-
zymes, with D-galactose, would improve reaction rates.
Table 1 shows that the strategy did work. In recent
years, the removal of bulk water of enzyme by rinsing/
precipitation with an organic solvent rather than lyoph-
ilization has emerged as an attractive option to obtain
more active biocatalyst preparation for use in low-water
a
Peaks were identified as mentioned in the earlier studies of b-
galactosidase.
2
1
45% for freeze-dried form) than the freeze-dried form.
This is in agreement with the known picture about the
structural changes undergone by protein during lyophi-
lization. It is believed that lyophilization damages the
protein structure by reducing the a-helical content of
this protein. The reported a-helical content of this en-
zyme in the native form is 46%. Thus FT-IR data indi-
cate that removal of bulk water by precipitation with n-
propanol is less harsh for protein structure as compared
to lyophilization. This could explain the higher activity
displayed by EPRP. Using this biocatalyst preparation
1
9,20
media.
zyme precipitated and rinsed with propanol) yielded
even better initial rates and consequently 42% conver-
Use of such a formulation called EPRP (en-
1
4
1
7
sion
(
to
compound 3, Scheme 1) in 16 h with an improved
(R)-(1-phenylethyl)-b-D-galactopyranoside
2
1
9
to evaluate the secondary structure of the proteins espe-
5% de (Table 1). FT-IR has emerged as a powerful tool
1
7
cially in the context of ‘drying’ the enzymes. Table 2
shows that precipitated enzyme (EPRP) did have higher
a-helix content (21% for EPRP and 10% for freeze-dried
form) and a lower b-sheet content (34% for EPRP and
1
4
(EPRP), different reaction media were tried (Table 3)

126
A. B. Majumder et al. / Bioorg. Med. Chem. Lett. 18 (2008) 124–128
Table 3. Effect of the change in the solvent system for galactosylation
of 1-phenylethanol with EPRP
a
C (%) de
a
P
a
S
a
Solvent
Initial rate
ee
E
À1 À1
(lmol mg
h
)
(4 h)
(%) (%)
Acetonitrile/DMF/
water (9:0.5:0.5, v/v)
1050
15
97
97
98
17
27
78
86
Acetonitrile/DMSO 1675
9:1, v/v)
DMSO
22
43
(
5953
73 220
Variation in the initial rates and enantioselectivity. EPRP of b-galac-
1
4
tosidase was tried with different solvent system. After 4 h the de and
ee values were noted in each case and conversion and E values were
calculated. Each reaction set was performed twice and the deviations
between the pair of readings were within 2%.
a
C, de , ee , E bear the same meanings as described in Table 1.
P S
with a reduced reaction time of 4 h. It was found that
DMSO could replace both DMF plus water and this im-
proved initial rates as well as conversion (%) without
2
2
loss of diastereoselectivity. As mentioned in Table 1,
E = 220 was obtained. Glycosidation reaction is accom-
panied by production of water as well. The water con-
tent in the reaction medium is known to influence the
7
reaction rates in non-aqueous enzymology. Hence the
reaction in DMSO was repeated in the presence of the
molecular sieves. The effect of the amount of the molec-
ular sieves present in the reaction medium on the initial
rate of product formation and deP (%) is shown in
Figure 1. There was some increase in initial rates as well
as in de (%) with optimum amount of molecular sieves.
P
Reducing water content in organic media is known to
3
2
improve enantioselectivity. This is presumed to be
due to the reduction in conformational flexibility which
2
4
favors enantioselectivity by the enzyme. The separa-
tion of the reactants from the products was done on
1
6
the HPLC chiral column. Thus, 45% conversion with
99% de under the best conditions was achieved. The
>
enantioselectivity, E, was based on the de of the product
and the ee of the remaining unreacted (S)-enriched 1-
phenylethanol. Corresponding to 45% conversion this
was >1000.
Figure 1. Change in the initial rates with EPRP and de (product) with
the varied amount of molecular sieves in reaction media: Activated
˚
molecular sieves (3 A) were added in varied amounts (10–60%, w/w
In order to evaluate possible versatility of this approach,
galactosylation of another alcohol was attempted. Mat-
sumura et al. have attempted to prepare enantiomeri-
cally pure (R)-2-octyl-b-D-galactopyranoside by
transgalactosylation. Only 17% yield with 63% ee could
be obtained. So 2-octanol was chosen as the second
alcohol for galactosylation by reverse hydrolysis. Table
reactants) in the reaction medium DMSO, incubated at 55 °C at
2
00 rpm in an orbital shaker, and initial rates were studied in each
case. (a) Synthesis of (R)-1-phenylethyl-b-D-galactopyranoside: Dia-
stereomeric excess (de) for the product was determined after 4 h. The
maximum de value corresponds to 45% conversion with 20% (w/w,
reactants) molecular sieves. The reactions were done in duplicates and
the deviations between the pair of readings were within 2%. (b)
Synthesis of (R)-2-octyl-b-D-galactopyranoside: Diastereomeric excess
4
shows the variation in the initial rates with enantiose-
lectivity of different formulations of b-galactosidase for
the reaction with racemic 2-octanol in 20% (v/v) DMSO
in acetonitrile. The medium was chosen from our earlier
experience which showed that DMSO as a cosolvent in
acetonitrile was a better option to start with. The same
trend was observed in this case also. The better values
were achieved with EPRP in DMSO (Table 4). Presence
of 10% molecular sieves (w/w, reactants) in the reaction
medium further improved the initial rate, conversion,
(
de) for the product was determined after 10 h. The maximum de value
corresponds to 33% conversion with 10% (w/w, reactants) molecular
sieves. The reactions were done in duplicates and the deviations
between the pair of readings were within 2%.
obtained within 10 h with a 92% deP (continuing the
reaction beyond 10 h gave a lower deP).
and the de (Fig. 1b). Under optimized conditions, thus,
P
À1 À1
an improved rate of 1320 lmol mg
h
corresponding
To conclude, for galactosylation of racemic 1-phenyleth-
anol, a conversion of 45% with a de > 99% (E > 1000)
to 33% conversion to 2-octyl-b-D-galactopyranoside was
P

A. B. Majumder et al. / Bioorg. Med. Chem. Lett. 18 (2008) 124–128
127
Table 4. Synthesis of (R)-2-octyl-b-D-galactopyranoside in 20%
DMSO (v/v) in acetonitrile
7. Gupta, M. N. Eur. J. Biochem. 1992, 203, 25.
8. Partridge, J.; Harper, N.; Moore, B. D.; Halling, P. J. In
Methods in Biotechnology; Vulfson, E. N., Halling, P. J.,
Holland, H. L., Eds.; Humana Press: New Jersey, 2002; p
a
C (%) de
a
a
S
a
Biocatalyst type Initial rate
P
ee
E
À1 À1
(
lmol mg
h
)
(10 h)
(%)
(%)
9
7.
. Lee, M. Y.; Dordick, J. S. Curr. Opin. Biotechnol. 2002,
3, 376.
Lyophiliized
Co-lyophilized
EPRP
534
735
936
17
21
27
30
72
78
88
90
14
20
32
38
7
9
10
21
28
1
1
1
1
0. Hudson, E. P.; Eppler, R. K.; Clark, D. S. Curr. Opin.
Biochem. 2005, 16, 637.
1. Shah, S.; Gupta, M. N. Bioorg. Med. Chem. Lett. 2007, 17,
b
EPRP
1206
Variation of initial rates and enantioselectivity of the different for-
mulations of b-galactosidase. Each reaction set was performed twice
and the deviations between the pair of readings were within 2%.
9
21.
2. Griebenow, K.; Laureano, Y. D.; Santos, A. M.; Cle-
mente, I. M.; Rodriguez, L.; Vidal, M. W.; Barletta, G.
J. Am. Chem. Soc. 1999, 121, 8157.
a
C, de
P
, ee
S
, E bear the sam meanings as described in Table 1.
b
The solvent used in this case for the reaction was 100% DMSO.
1
1
3. Hirata, T.; Shimoda, K.; Fujino, T.; Yamane, S-Y.; Ohta,
S. Bull. Chem. Soc. Jpn. 2001, 74, 539.
4. Preparation of different formulations of the biocatalyst
for low-water media: Lyophilized enzyme preparation:
b-Galactosidase powder (5 mg) was dissolved in
in 4 h is a considerable improvement as compared to
1
9
.5% conversion reported by Bridiau et al. and 39%
conversion with a de (reported as ee) of 98% (corre-
P
3
3
200 lL of 100 mM phosphate buffer (pH 7.3),
sponding E = 188) reported by Matsumura et al. For
frozen at À23 °C, and was lyophilized for 24 h. In
another case, the same amount of enzyme was
dissolved along with galactose (18.6 mg) and was
lyophilized for 24 h.
galactosylation of racemic 2-octanol, again EPRP was
found to be better than the other formulations of b-
galactosidase (Table 4). Under optimized conditions,
3
high 92% de which is a considerable improvement over
the earlier reported value of 63% de . The present work
P
outlines an optimization strategy which combines the
use of a highly active biocatalyst with a suitable reaction
medium for the diastereoselective galactosylation of sec-
ondary alcohols.
3% conversion was achieved within 10 h with a fairly
EPRP (enzyme precipitated and rinsed with propanol):
25 lL of enzyme solution (20 mg/100 lL of 100 mM
phosphate buffer, pH 7.3) was precipitated in 200 lL
of ice cold dry n-propanol, centrifuged, and the
recovered pellet was repeatedly rinsed with n-propa-
nol before using for the reaction.
P
3
1
5. b-Galactosylation with 1-phenylethanol: In screw-capped
vials, D-galactose (18.5 mg) and 25 lL of racemic 1-
phenylethanol (Merck, Germany) were dissolved in a
solution of acetonitrile, DMF, water (in 9:0.5:0.5, v/v
ratio) so that the final volume of the reaction volume was
Acknowledgments
1
.0 mL. After pre-incubation at 55 °C for 15 min, 5 mg
The authors are grateful to Dr. N. G. Ramesh (Chemis-
try Department, IIT Delhi) for helpful discussions. This
work was supported by ‘Core group grant for applied
biocatalysis’ by Department of Science and Technology,
Government of India (DST). The financial supports
provided by Council of Scientific and Industrial Re-
search (India) to ABM as Senior Research Fellowship
are also gratefully acknowledged.
(10%, w/w reactants) enzyme preparation was added. As
an alternative reaction medium, anhydrous DMSO (water
content <0.005%, v/v, Acros Organics, USA) was tried
instead of DMF and water. The reaction was carried out
at 55 °C at 200 rpm in an orbital shaker. b-Galactosylation
with 2-octanol: D-Galactose (18.5 mg) and 35 lL of race-
mic 2-octanol (Merck, Germany) were dissolved in a
solution of acetonitrile, DMSO (in 4: 1, v/v ratio) and the
reaction was performed as mentioned earlier.
1
6. HPLC analysis and determination of the diastereomeric
excess (de) of the product: The product analysis was done
with chiral column [Chiracel OD RH; eluents: (i) hexane/
isopropanol/ethanol = 75:23:2 (v/v) at a flow rate of
Supplementary data
0
.5 mL/min by DAD-UV (217 nm) for 1-phenylethanol
system, (ii) acetonitrile/water = 9:1 (v/v) at a flow rate of
.25 mL/min by RID for 2-octanol system] fitted in
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.bmcl.
0
2
007.11.006.
Agilent 1100 series HPLC system. The peaks (retention
time) were identified as (S)-1-phenylethyl-b-D-galactopy-
ranoside (4.43 min), (R)-1-phenylethyl-b-D-galactopyran-
oside (4.68 min), (R)-1-phenylethanol (4.90 min), (S)-1-
phenylethanol (5.06 min), and (S)-2-octyl-b-D-galactopy-
ranoside (8.75 min), (R)-2-octyl-b-D-galactopyranoside
References and notes
. Bridiau, N.; Taboubi, S.; Marzouki, N.; Legoy, M. D.;
Maugard, T. Biotechnol. Prog. 2006, 22, 326.
1
(
9.20 min), (R)-2-octanol (10.04 min), (S)-2-octanol
10.3 min). Finally, the enantioselectivity was calculated
2
3
. Oosterom, M. W-. V.; van Belle, H. J. A.; van
Rantwijk, F.; Sheldon, R. A. J. Mol. Catal. A: Chem.
(
by Chen’s equation.
2
3
1
998, 134, 267.
1
1
1
2
7. Anchordogny, T. J.; Carpenter, J. F. Arch. Biochem.
Biophys. 1996, 332, 231.
8. Mishra, P.; Griebenow, K.; Klibanov, A. M. Biotechnol.
Bioeng. 1996, 52, 609.
9. Partridge, J.; Halling, P. J.; Moore, B. D. J. Chem. Soc.,
Chem. Commun. 1998, 7, 841.
0. Roy, I.; Gupta, M. N. Bioorg. Med. Chem. Lett. 2004, 14,
. Matsumura, S.; Yamazaki, H.; Toshima, K. Biotechnol.
Lett. 1997, 19, 583.
4
5
. Vic, G.; Crout, D. H. G. Carbohydr. Res. 1995, 279, 315.
. Garcia-Garibay, M.; Lopez-Munguia, A.; Berzana, E.
Biotechnol. Bioeng. 2000, 69, 627.
6
. Vic, G.; Biton, J.; Le Beller, D.; Michael, J.-M.; Thomas,
D. Biotechnol. Bioeng. 1995, 46, 109.
2
191.

128
A. B. Majumder et al. / Bioorg. Med. Chem. Lett. 18 (2008) 124–128
TM
2
2
2
2
1. Degraeve, P.; Rubens, P.; Lemay, P.; Heremans, K.
Enzyme Microb. Technol. 2002, 31, 673.
2. Watanabe, K.; Ueji, S. J. Chem. Soc., Perkin Trans. 1
6700 spectrophotometer with Smart Miracle
ATR
accessories having zinc selenide crystal. Different formu-
lations of b-galactosidase, freeze-dried (FD), and EPRP,
were taken directly on the zinc selenide crystal plate and
an average of 200 scan was taken with a resolution of
2001, 386.
3. Chen, C.-S.; Fujimoto, Y.; Girdalukas, G.; Sih, C. J.
J. Am. Chem. Soc. 1982, 104, 7294.
4. Scheper, T.; Bornscheuer, U.; Kolisis, F. N.; Meyer, H.
M.; Capewell, A.; Wendel, V.; Kreye, L.; Herar, A.
Tetrahedron: Asymmetry 1993, 4, 1007.
À1
À1
2 cm . The spectra were analyzed over 1600–1700 cm
2
1
(amide-I band). The data collection and Fourier Self
Deconvoluation (FSD) were done by using OMNIC 2.1
software. The peaks were resolved and were fitted with
Gaussian curves. From the best fit, the peak areas were
noted.
2
5. FT-IR analysis of the different formulations of b-galac-
tosidase: The spectra were recorded on Nicolet FT-IR
TM