Regioselective glucosylation of aromatic compounds: Screening of a recombinant glycosyltransferase library to identify biocatalysts

Article abstract of DOI:10.1002/anie.200504505

(Chemical Equation Presented) A novel whole-cell screen for the identification of new glucosyltransferase (GT) biocatalysts within a recombinant enzyme library was developed. Following biotransformation, levels of D-glucose were used as a measure of biocatalysis. Twenty five enzymes that transfer D-glucose to trans-resveratrol were identified that have the regioselectivity shown in the picture.

Full text of DOI:10.1002/anie.200504505

Communications
Biocatalyst Identification
Herin, we wanted to apply the whole-cell biocatalysis
system in a format that would enable us to screen simulta-
neously a library consisting of multiple GTs. Thus, single
colonies of E.coli , each expressing an individual GT, were
cultured in 96-well titer plates. The screen for catalytic activity
needed to be independent of aglycone if the method was to be
generic. Therefore, we used a colorimetric detection system
for d-glucose,[10] which is experimentally released from
glucosides formed during the biocatalysis. The principle aim
of the study has been to provide a rapid assessment of GTs to
detect those with a high potential for development into
whole-cell biocatalysts. This provides the foundation for the
subsequent detailed analysis and choice of enzyme to use or
improve for the synthesis of aromatic glucosides.
DOI: 10.1002/anie.200504505
Regioselective Glucosylation of Aromatic
Compounds: Screening of a Recombinant
Glycosyltransferase Library to Identify
Biocatalysts**
Markus Weis, Eng-Kiat Lim, Neil Bruce, and
Dianna Bowles*
The chemical synthesis of glycosides requires glycosyl activa-
tion and involves multiple steps of protection/deprotection to
control regioselectivity. This can often reduce the yield of the
final product.[1] Glycosyltransferases (GTs) offer
a potential solution to this problem[2] as the
enzymes use unprotected aglycones in aqueous
solution and their catalytic activity is chemo-,
regio- and enantioselective.[3] However, to date,
the availability of characterized enzymes has been
limited and their use as biocatalysts is constrained
by the need to supply activated sugars for the
synthesis of the glycosides. Recently, a large
multigene family of GTs has been identified in
Arabidopsis thaliana[4] and expressed as recombi-
nant enzymes in Escherichia coli. This strategy has
enabled new catalytic activities towards small
aromatic molecules to be identified in vitro.[5,6]
Studies have shown that plant GTs are capable
of recognizing compounds beyond those found in
plants.[7] The need to add activated sugars has
been successfully overcome by the use of
recombinant GTs in a whole-cell biocatalysis
system.[8] Based on this system, we have devel-
oped an assay method in which an entire library of
GTs can be screened against potential substrates
without the requirement of recombinant protein
purification and supplementary activated sugars,
thereby allowing rapid identification of relevant
biocatalysts for future use in chemical synthesis or
fermentation. The method is in contrast to those
previously reported including liquid chromatog-
raphy (LC), LC–MS, or radiometric detection,[9]
The method, illustrated in Scheme 1, was established and
optimized for a 96-well plate format by using the conversion
Scheme 1. Design of the rapid-screening method. This method consists of three
stages: aglycone biotransformation (stage 1), cleavage of the glucoside (stage 2),
and detection of the released d-glucose in a coupled enzymatic assay (stage 3).
which were developed for analyzing the activity of
purified recombinant GTs.
[*] M. Weis, Dr. E. K. Lim, Prof. N. Bruce, Prof. D. Bowles
CNAP, Department of Biology
University of York
of the hydroxycoumarin, scopoletin (1), into scopolin (2) as a
model system. In vitro catalysis had already demonstrated
that the substrate was recognized by multiple recombinant
arabidopsis GTs.[6] Cells were cultured in standard media
before transferring to d-glucose-minus medium in which l-
arabinose was the carbon source. Following induction,
addition of substrate, and incubation, cells were separated
and the media from each well were collected and samples
either analyzed directly by using reverse-phase (RP) HPLC
or treated with b-glucosidase, filtered through polyvinyl-
York YO10 5DD (UK)
Fax: (+44)1904-328-772
E-mail: djb32@york.ac.uk
[**] This work was supported by The Garfield Weston Foundation. We
thank Dr. David Ashford, Dr. Gideon Grogan, and members of the
group for helpful discussions.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
3534
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 3534 –3538

Angewandte
Chemie
polypyrrolidone (PVPP) to remove remaining aglycone, and
the levels of d-glucose then detected in an enzymatic assay.
Figure 1 illustrates the GT activities towards scopoletin and
demonstrates a linear relationship between the amount of
scopolin formed in each reaction and d-glucose detection.
The whole-cell biocatalysis and screen identified 45 GTs with
activity towards scopoletin, confirming and extending the
earlier data from in vitro catalysis. Invariably, a negative
result in the d-glucose detection assay correlated with a
negative result in the RP-HPLC analysis.
Figure 1. Screening of a GT library against the aglycone scopoletin.
a) The readings at A405nm for d-glucose detection are presented in a
color-code format. b) The correlation of the colorimetric detection at
A405nm and the HPLC analysis. HPLC quantification of glucosides are
normalized on the strongest peak and annotated in percentage.
c) Examples of RP-HPLC chromatographs of active and non-active GTs
in whole-cell biocatalysis are illustrated.
Figure 2. Screening of a GT library against the aglycone daidzein.
a) The readings at A405nm for d-glucose detection are presented in a
color-code format. b) Examples of RP-HPLC traces of active and non-
active GTs in whole-cell biocatalysis are illustrated. c) The regioselec-
tivity of the active GTs towards daidzein, defined by the percentage of
a regiospecific glucoside in the total amount of monoglucosides
formed.
Angew. Chem. Int. Ed. 2006, 45, 3534 –3538ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3535

Communications
The utility of the method to discover novel biocatalysts
correlation to d-glucose detection for both substrates (see
the Supporting Information). The mono- and diglucosides of
daidzein (4–6) and trans-resveratrols (8–10), which eluted
earlier than the two aglycones under the RP-HPLC con-
ditions used (Figure 2b and Figure 3b), were identified by
using external standards when available, or by electrospray
was investigated by using the isoflavone, daidzein (3), and the
stilbene, trans-resveratrol (7). Both compounds exist as
glucosides, have attracted considerable pharmaceutical inter-
est,[11] and chemical synthesis of their different glycosides has
been attempted but resulted in poor yields and lack of
regioselective discrimination.[12] Daidzein, as well as other
isoflavones, occurs naturally in legumes as the 7- and 4’-b-O-
glucosides (4 and 5, respectively).[13] trans-Resveratrol, a
naturally occurring hydroxystilbene, is found as glucosides[14]
and methoxides.[15] Piceid (3-b-O-glucoside; 8) and resvera-
troloside (4’-b-O-glucoside 9) are the most abundant con-
jugates. Bioactivity of these compounds has been reported in
relation to cancer prevention,[16] coronary heart disease,[17]
antioxidant activity,[18] and estrogenic activity[19] As neither
daidzein nor trans-resveratrol is reported to occur in arabid-
posis, they represent non-natural substrates for the GT
screen.
1
LC–MS. H NMR analysis was used to confirm the structure
of the monoglucosides (see the Supporting Information).
From the 13 GTs that recognized daidzein, three (GTs 84A1,
73B2, and 73B1) were found to be 100% regioselective for
the 7-OH; the remaining enzymes glucosylated the 4’-OH and
7-OH positions to varying degrees, and one GT, 73C4,
produced the diglucoside in addition to the monoglucosides
(Figure 2b). Similarly, regioselective glucosylation of trans-
resveratrol was observed. From the 25 enzymes that recog-
nized the substrate, five GTs were specific for the 3-OH
position (GTs 71D1, 71C2, 88A1, 72D1 and 71C4) and one
GT 74B1 was specific for the 4’-OH position (Figure 3b).
Only trace levels of a diglucoside were observed under the
reaction conditions used. As before, for the biotransformation
of both daidzein and trans-resveratrol, the d-glucose-based
detection system did not miss any positive enzyme activities;
however, in these assays, two false positives in screens of each
compound were observed in which an intense absorption was
The utility of the screening method and regioselective
biocatalysis by the GTs are illustrated in Figure 2and
Figure 3. 13 GTs recognized daidzein and 25 GTs were
identified that glucosylated trans-resveratrol. As previously
described for scopoletin, RP-HPLC quantification of the
glucosides formed in the biocatalysis revealed a linear
Figure 3. Screening of a GT library against the aglycone trans-resveratrol. a) The readings at A405nm for d-glucose detection are presented in a
color-code format. b) Examples of RP–HPLC traces of active and non-active GTs in whole-cell biocatalysis are illustrated. c) The regioselectivity of
the active GTs towards trans-resveratrol (defined by the percentage of a regiospecific glucoside in the total amount of monoglucosides formed).
3536
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 3534 –3538

Angewandte
Chemie
Other experimental details and analytical methods as well as
physical data were included as Supporting Information.
not associated with any product formation. To confirm that
the GTactivity detected in the whole-cell system remain after
purification, the activity of a number of individual recombi-
nant GTs were also analyzed in vitro. The GT activities
detected in the whole-cell screen were confirmed in vitro and
their specific activities determined (see the Supporting
Information).
The potential of GTs for use in preparative synthesis of
daidzein and trans-resveratrol glucosides was explored on a
fermenter scale by using bacterial cells expressing GT 73B1
and GT 73C6, respectively. This confirmed that GTs offer
major opportunities for preparative-scale biotransformations
as the glucosides could readily be obtained in significant
quantities (see the Supporting Information).
Received: December 19, 2005
Published online: April 24, 2006
Keywords: enzymes · glycosides · glycosylation · regioselectivity ·
screening
.
[1] a) H. Pellissier, Tetrahedron 2005, 61, 2947 – 2993; b) K. C.
Nicolaou, H. J. Mitchell, Angew.Chem. 2001, 113, 1624 – 1672;
Angew.Chem.Int.Ed.
2001, 40, 1576 – 1624; c) S. Hanessian,
B. L. Lou, Chem.Rev. 2000, 100, 4443 – 4463.
[2] a) S. L. Flitsch, Curr.Opin.Chem.Biol.
2000, 4, 619 – 625;
b) B. G. Davis, V. Boyer, Nat.Prod.Rep. 2001, 18, 618 – 640; c) P.
Sears, C. H. Wong, Science 2001, 291, 2344 – 2350; d) Z. Nahalka,
Y. Liu, X. Chen, P. G. Wang, Chem.Eur.J. 2003, 9, 373 – 377;
e) E. K. Lim, D. J. Bowles, EMBO J. 2004, 23, 2915 – 2922;
f) K. M. Koeller, C. H. Wong, Chem.Rev. 2000, 100, 4465 – 4493.
[3] a) E. Garcia-Junceda, J. F. Garcia-Garcia, A. Bastida, A. Fer-
nandez-Mayoralas, Bioorg.Med.Chem. 2004, 12, 1817 – 1834;
b) E. K. Lim, Chem.Eur.J. 2005, 11, 5486 – 5494.
The use of this screening method in whole-cell biocatalysis
can be readily extended for the detection of other glycosides.
This can be achieved by combining, for example, xylosidase
and pyranose oxidase[20] to detect xylosides, or galactosides
through the use of galactosidase and galactose oxidase.[21]
In conclusion, we have successfully developed a generic
screen to determine the activity of recombinant GT libraries
towards aromatic compounds in whole-cell biocatalysis. We
have demonstrated that the method provides the means to
rapidly identify GTs of high utility that can be further
developed for use in biotransformations or chemo-enzymatic
synthesis of small-molecule glycosides. The regio- and enan-
tioselectivity of GT biocatalysts offers a useful complement to
classical chemical approaches.
[4] Y. Li, S. Baldauf, E. K. Lim, D. J. Bowles, J.Biol.Chem. 2001,
276, 4338 – 4343.
[5] a) B. Hou, E. K. Lim, G. S. Higgins, D. J. Bowles, J.Biol.Chem.
2004, 279, 47822 – 47832; b) E. K. Lim, C. J. Doucet, Y. Li, L.
Elias, D. Worrall, S. P. Spencer, J. Ross, D. J. Bowles, J.Biol.
Chem. 2002, 277, 586 – 592; c) E. K. Lim, C. J. Doucet, B. Hou,
R. G. Jackson, S. R. Abrams, D. J. Bowles, Tetrahedron: Asym-
metry 2005, 16, 143 – 147.
[6] E. K. Lim, S. Baldauf, Y. Li, L. Elias, D. Worrall, S. P. Spencer,
R. G. Jackson, G. Taguchi, J. Ross, D. J. Bowles, Glycobiology
2003, 13, 139 – 145.
[7] a) C. Loutre, D. P. Dixon, M. Brazier, M. Slater, D. J. Cole, R.
Edwards, Plant J. 2003, 34, 485 – 493; b) B. Poppenberger, F.
Berthiller, D. Lucyshyn, T. Sieberer, R. Schuhmacher, R. Krska,
K. Kuchler, J. Glossl, C. Luschnig, G. Adam, J.Biol.Chem. 2003,
278, 47905 – 47914; c) T. Hefner, J. Arend, H. Warzecha, K.
Siems, J. Stockigt, Bioorg.Med.Chem. 2002, 10, 1731 – 1741;
d) B. Messner, O. Thulke, A. R. Schaffner, Planta 2003, 217,
138 – 146.
[8] a) V. Kren, J. Thiem, Chem.Soc.Rev. 1997, 26, 463 – 473; b) S.
Koizumi, T. Endo, K. Tabata, A. Ozaki, Nat.Biotechnol. 1998,
16, 847 – 850; c) E. K. Lim, D. A. Ashford, B. Hou, R. G.
Jackson, D. J. Bowles, Biotechnol.Bioeng. 2004, 87, 623 – 631;
d) S. Koizumi, Trends Glycosci.Glycotechnol. 2003, 15, 65 – 74;
e) J. Arend, H. Warzecha, T. Hefner, J. Stockigt, Biotechnol.
Bioeng. 2001, 76, 126 – 131.
[9] a) M. Yang, M. Brazier, R. Edwards, B. G. Davis, ChemBioChem
2005, 6, 346 – 357; b) A. Di Marco, M. DꢀAntoni, S. Attaccalite, P.
Carotenuto, R. Laufer, Drug Metab.Dispos. 2005, 33, 812– 819;
c) M. M. Palcic, K. Sujino, Trends Glycosci.Glycotechnol. 2001,
13, 361 – 370; d) J. Wongkongkatep, Y. Miyahara, A. Ojida, I.
Hamachi, Angew.Chem. 2006, 118, 681 – 684; Angew.Chem.Int.
Ed. 2006, 45, 665 – 668; B. Schwartz, J. A. Markwalder, S. P. Seitz,
Y. Wang, R. L. Stein, Biochemistry 2002, 41, 12552; e) C. Deng,
R, Chen, Anal.Biochem. 2004, 330, 219 – 226; f) H. C. Hang, C.
Yu, M. R. Pratt, C. R. Bertozzi, J.Am.Chem.Soc. 2004, 126, 6 –
7; g) H. Liu, T. K. Ritter, R. Sadamoto, P. S. Sears, M. Wu, C.-H.
Wong, ChemBioChem 2003, 4, 603 – 609.
Experimental Section
Stage 1, biotransformation: Single colonies of the GT library grown
on Lauria Bertani (LB)–agar plates overnight were transferred to
individual wells in a 96-well bacterial culture plate containing 2” YT
medium (400 mL; bacto tryptone (16 gLÀ1), yeast extract (10 gLÀ1),
NaCl (5 gLÀ1)) and ampicillin (50 mgmLÀ1). The plate was covered
with an adhesive plate seal (ABgene, U.K.) and incubated at 378C
(250 rpm). The bacterial growth was monitored at 595 nm by a plate
reader (Bio-Tec Instruments Inc., U.S.A). After 4 h, the cultures had
reached the exponential phase. The plate was centrifuged (4000 g,
10 min), the supernatants discarded, and cell pellets were resus-
pended in isopropyl-b-d-thiogalactopyranoside (IPTG, 1 mm), 2-(N-
morpholino) ethanesulfonic acid (MES, 50 mm, pH 7.0), ampicillin
(50 mgmLÀ1), l-arabinose (10 gLÀ1), and 500 mm of aglycone to a total
whole-cell reaction volume of 400 mL per well. The 96-well plate was
closed with
a gas-permeable adhesive plate seal, wrapped in
aluminium foil for light protection, and incubated at 258C
(250 rpm). After 44 h, the cultures were centrifuged (4000 g,
15 min) and the supernatants analyzed.
Stage 2, cleavage: Supernatants (100 mL) were transferred to a
microtiter plate, b-glucosidase (1 mL, 1 UmLÀ1) was added and the
plate incubated for 90 min at 378C.
Stage 3, detection: The reaction mix (50 mL) was transferred to a
96-well filtration plate (ABgene, U.K.), mixed with an equal volume
of PVPP aqueous suspension (25 gLÀ1), and shaken for 1 h at 258C
before centrifugation (1000 g, 5 min). To each filtrate, MES (50 mm;
pH 7.0), 2,2ꢀ-azino-bis(3-ethylbenzthiazoline-6-sulfonic acid) (ABTS;
4 mm), peroxidase (2UmL À1), and glucose oxidase (2UmL À1) were
added to a final volume of 125 mL. The formation of the green dye was
monitored at 405 nm for 30 min by using a plate reader.
[10] a) D. C. Williams, G. F. Huff, W. R. Seitz, Clin.Chem. 1976, 22,
372– 374; b) N. P. Groome, J.Clin.Chem.Clin.Biochem.
18, 345 – 349.
1980,
[11] a) G. Galati, P. J. OꢀBrien, Free Radical Biol.Med. 2004, 37, 287 –
303; b) S. Barnes, J. Prasain, Curr.Opin.Plant.Biol. 2005, 8,
324 – 328; c) L. Fremont, Life Sci. 2000, 66, 663 – 673; d) P.
Angew. Chem. Int. Ed. 2006, 45, 3534 –3538ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3537

Communications
Signorelli, R. Ghidoni, J.Nutr.Biochem. 2005, 16, 449 – 466; e) J.
Reynaud, D. Guilet, R. Terreux, M. Lussignol, N. Walchshofer,
Nat.Prod.Rep. 2005, 22, 504 – 515; f) K. D. R. Setchell, A.
Cassidy, J.Nutr. 1999, 129, 758S-767S.
[12] a) P. W. Needs, G. Williamson, Carbohydr.Res. 2001, 330, 511 –
515; b) F. Orsini, F. Pelizzoni, B. Bellini, G. Miglierini, Carbo-
hydr.Res. 1997, 301, 95 – 109; c) P. T. Lewis, K. Wahala,
Tetrahedron Lett. 1998, 39, 9559 – 9562.
[13] Y. Shibuya, S. Tahara, Y. Kimura, J. Miyzutani, Z.Naturforsch.C
1991, 46, 513 – 518.
[14] D. M. Goldberg, E. Ng, A. Karumanchiri, E. P. Diamandis, G. J.
Soleas, Am.J.Enol.Vitic. 1996, 47, 415 – 420.
[15] P. Langcake, C. A. Cornford, R. J. Pryce, Phytochemistry 1979,
18, 1025 – 1027.
[16] a) H. Adlercreutz, M. Yaghoob, K. Hoeckerstedt, Acta Oncol.
1992, 350, 115 – 181; b) M. Jang, L. Cai, G. O. Udeani, K. V.
Slowing, C. F. Thomas, C. W. Beecher, H. H. Fong, N. R. Farns-
worth, A. D. Kinghorn, R. G. Mehta, R. C. Moon, J. M. Pezzuto,
Science 1997, 275, 218 – 220; c) D. Ingram, K. Sanders, M.
Kolybaba, D. Lopez, Lancet 1998, 350, 990 – 994.
[17] a) S. Samman, P. Lyons-Wall, N. Cook, Antioxid.Health Dis.
1998, 7, 469 – 481; b) P. Nestel, T. Yamashita, T. Sasahara, S.
Pomeroy, A. Dart, P. Komesaroff, A. Owen, A. Abbey,
Arterioscler.Thromb.Vasc.Biol. 1997, 17, 3392– 3398.
[18] a) M. Begona-Ruiz-Larrera, A. Moham, G. Paganga, N. Miller,
G. Bolwell, C. Rice-Evans, Free Radical Res. 1997, 26, 63 – 70;
b) M. J. Tikkanen, K. Wahala, S. Ojala, V. Vihma, H. Adler-
creutz, Proc.Natl.Acad.Sci.USA 1998, 95, 3106 – 3110.
[19] a) S. D. Garrett, H. A. Lee, M. R. A. Morgan, Nat.Biotechnol.
1999, 17, 1219 – 1222; b) B. D. Gehm, J. M. McAndrews, P. Y.
Chien, J. L. Jameson, Proc.Natl.Acad.Sci.USA
1997, 94,
14138 – 14143.
[20] K. Wagschal, D. Franqui-Espiet, C. C. Lee, G. H. Robertson,
D. W. S. Wong, Appl.Environ.Microbiol. 2005, 71, 5318 – 5323.
[21] a) K. Schmid, R. Schmitt, Eur.J.Biochem. 1976, 67, 95 – 104;
b) G. Avigad, D. Amaral, C. Asensio, B. L. Horecker, J.Biol.
Chem. 1962, 237, 2736 – 2743; c) C. S. Frings, H. L. Pardue, Anal.
Chem. 1964, 36, 2477 – 2478.
3538
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 3534 –3538