Glucosylation of flavonoids and flavonoid glycosides by mutant dextransucrase from Lactobacillus reuteri TMW 1.106

Article abstract of DOI:10.1016/j.carres.2019.107741

Flavonoids are commonly abundant, plant-derived polyphenolic compounds which are responsible for color, taste, and antioxidant properties of certain plant based foods. Glucosylation by glucansucrases or other glycosyltransferases/glycoside hydrolases has

Full text of DOI:10.1016/j.carres.2019.107741

Accepted Manuscript
Glucosylation of flavonoids and flavonoid glycosides by mutant dextransucrase from
Lactobacillus reuteri TMW 1.106
Tizian Klingel, Martina Hadamjetz, Anja Fischer, Daniel Wefers
PII:
S0008-6215(19)30278-2
DOI:
https://doi.org/10.1016/j.carres.2019.107741
Article Number: 107741
Reference:
CAR 107741
To appear in: Carbohydrate Research
Received Date: 2 May 2019
Revised Date: 9 July 2019
Accepted Date: 9 July 2019
Please cite this article as: T. Klingel, M. Hadamjetz, A. Fischer, D. Wefers, Glucosylation of flavonoids
and flavonoid glycosides by mutant dextransucrase from Lactobacillus reuteri TMW 1.106, Carbohydrate
Research (2019), doi: https://doi.org/10.1016/j.carres.2019.107741.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Glc
α
O
OH
OH
OH
Sucrose
HO
O
O
HO
O
O
L. reuteri TMW 1.106
L242A dextransucrase
R = H,
O
R
O
R
Glcβ,
L-Rhaα1−6Glcβ
OH
OH
Glc
α
O
OH
OH
OH
OH
OH
HO
O
O
HO
O
HO
O
O
O
O
O
OH
OH
O
OH
Glcα1−3R
Glcα1−4R
Glcα1−3/4R
R = Glcβ
R = Glcβ
R = L-Rhaα1−6Glcβ

ACCEPTED MANUSCRIPT
Glucosylation of flavonoids and flavonoid glycosides by mutant
dextransucrase from Lactobacillus reuteri TMW 1.106
Tizian Klingel1, Martina Hadamjetz1, Anja Fischer1, Daniel Wefers1*
1
Department of Food Chemistry and Phytochemistry, Institute of Applied Biosciences,
Karlsruhe Institute of Technology (KIT), Adenauerring 20a, 76131 Karlsruhe, Germany
*Correspondence: Daniel Wefers, Department of Food Chemistry and Phytochemistry, Institute
of Applied Biosciences, Karlsruhe Institute of Technology (KIT), Adenauerring 20a, 76131
Karlsruhe, Germany. daniel.wefers@kit.edu, Phone: +49 721 608 42929, Fax: +49 721 608
47255
1

ACCEPTED MANUSCRIPT
1
ABSTRACT
2
3
Flavonoids are commonly abundant, plant-derived polyphenolic compounds which are
responsible for color, taste, and antioxidant properties of certain plant based food.
Glucosylation by glucansucrases or other glycosyltransferases/glycoside hydrolases has been
described to be a promising approach to modify stability, solubility, bioavailability, and taste
profile of flavonoids and related compounds. In this study, we modified and applied a
recombinant dextransucrase from Lactobacillus reuteri TMW 1.106 to glucosylate various
flavonoids and flavonoid glycosides. The glucoconjugates were subsequently isolated and
characterized by using two-dimensional NMR spectroscopy. Efficient glucosylation was
achieved for quercetin and its glycosides quercetin-3-O-β-glucoside and rutin. Significant
portions of α-glucose conjugates were also obtained for epigallocatechin gallate,
dihydromyricetin, and cyanidine-3-O-β-glucoside, whereas glucosylation efficiency was low
for naringin and neohesperidin dihydrochalcone. Most of the flavonoids with a catechol or
pyrogallol function at the B-ring were predominantly glucosylated at position O4’. However,
glycosyl substituents such as β-glucose, rutinose, or neohesperidose were glucosylated at
varying positions. Therefore, mutant dextransucrase from L. reuteri TMW 1.106 can be applied
for versatile structural modification of flavonoids.
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
KEYWORDS
19
20
glucansucrase, modification, NMR spectroscopy, acceptor reaction, structural characterization,
purification.
21
ABBREVIATIONS
22
23
24
L. = Lactobacillus, HSQC = Heteronuclear Single Quantum Coherence, COSY = H,H-
Correlated Spectroscopy, H2BC = Heteronuclear Two-bond Correlation, TOCSY = Total
Correlated Spectroscopy, HMBC = Heteronuclear Multiple Bond Correlation.
2

ACCEPTED MANUSCRIPT
25
1
INTRODUCTION
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
Flavonoids are a very heterogeneous group of polyphenolic secondary plant metabolites which
are widely distributed among different plants. Consumption of polyphenol-rich foods was
associated with beneficial health effects which were discussed to be related to the antioxidant
activity of flavonoids [1]. In addition, some flavonoids also affect the sensory properties of
plant based food. For example, anthocyanins are responsible for the color of many fruits,
whereas some flavonoids such as the flavanone naringin contribute to the bitter taste of some
citrus fruits [2]. Furthermore, the flavonoid derivative neohesperidin dihydrochalcone is added
to food products as a sweetener or flavor enhancer [3]. In nature, flavonoids are often
glycosylated at varying positions such as position O3 (if present) or position O7 and O5 [4]. It
has already been demonstrated that glycosylation significantly influences the solubility,
stability, bioavailability, and antioxidant properties [4-7]. Therefore, enzymatic in vitro
glycosylation has also been investigated as an approach to modify the described properties.
Besides cyclodextrin glucanotransferases, α-glucosidases, and maltogenic amylases [5, 7-13],
glucansucrases have already been applied to glucosylate selected flavonoids [14-20]. These
enzymes belong to the GH70 family and normally catalyze the conversion/incorporation of
sucrose into polymeric α-glucans, but by selecting appropriate reaction conditions,
oligosaccharides or other hydroxyl group containing compounds can also be conjugated with
α-glucose [21]. However, the above mentioned studies focused on glucosylation of non-
glycosylated flavonoids such as luteolin, quercetin, myricetin, dihydromyricetin, and
epigallocatechin gallate. In addition, only glucansucrases from Leuconostoc mesenteroides
have been used for flavonoid glucosylation. However, recombinant glucansucrases from L.
reuteri have been successfully applied for glucosylation of steviol glycosides or simple phenols
such as catechol [22-26]. Furthermore, site directed mutagenesis was applied to improve
glucosylation of these substrates and other non-carbohydrate molecules [23, 24, 26].
3

ACCEPTED MANUSCRIPT
50
51
52
53
54
55
56
57
58
In previous studies, it was demonstrated that dextransucrase from Lactobacillus (L.) reuteri
TMW 1.106 synthesizes O4-branched dextrans and thus produces a rather rare dextran type
[27, 28]. Furthermore, we observed a high expression level and thus a comparably high enzyme
activity after cloning of the encoding gene into a pET-based expression vector and heterologous
expression of the recombinant sucrase in E. coli. Consequently, this recombinant
dextransucrase is a promising candidate for the modification of polyphenolic compounds.
Therefore, the aim of this study was to analyze the structures of α-glucose conjugates formed
by the incubation of different flavonoids, flavonoid glycosides, and flavonoid derivatives with
dextransucrase from L. reuteri TMW 1.106.
59
60
2
RESULTS AND DISCUSSION
61
62
63
64
65
66
67
68
69
70
71
72
73
74
Several flavonoid substrates which showed a high structural heterogeneity of both the flavonoid
residues as well as the attached sugars were used for the glucosylation experiments (Fig. S1).
Besides the unglucosylated flavonole quercetin, quercetin-3-O-β-glucoside and rutin (which
corresponds to quercetin-3-O-rutinoside) were also used as substrates. Furthermore, the plant
pigment cyanidin-3-O-β-glucoside was isolated from blackberry and used for the glucosylation
experiments. With regards to taste-active flavonoids/flavonoid derivatives, naringin and
neohesperidin dihydrochalcone were used as substrates. These compounds carry a
neohesperidose residue at position O7, thus they show a different type and location of the sugar
side chain than the previously mentioned substrates. In addition, dihydromyricetin
(ampelopsin) and epigallocatechin gallate were used. In contrast to all other substrates, these
non-glycosylated flavonoids both carry a pyrogallol group at the B-ring (and, in case of
epigallocatechin gallate, at position O3 in form of gallic acid). Therefore, the substrates used in
this study exhibit different flavonoid types, different substitution patterns of the flavonoid B-
ring, and differences in the type and position of the flavonoid-bound sugar units.
4

ACCEPTED MANUSCRIPT
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
In preliminary experiments, wildtype dextransucrase from L. reuteri TMW 1.106 was
successfully applied for glucosylation of some flavonoid substrates, but a rather low
glucosylation efficiency was achieved. Devlamynck, te Poele, Meng, van Leeuwen and
Dijkhuizen [26] demonstrated that modification of certain amino acids in the +1 and +2
acceptor binding site of glucansucrase Gtf180 from L. reuteri 180 eventually led to an increased
glucosylation of non-carbohydrate molecules by suppressing synthesis of polymeric α-glucans.
Therefore, we aligned the amino acid sequence of L. reuteri TMW 1.106 dextransucrase and
Gtf180 to identify the leucine residue equivalent to L981 in Gtf180 (Fig. S2). Subsequently, we
changed this leucine residue to alanine by using site-directed mutagenesis (yielding L. reuteri
TMW 1.106 L242A). The mutant L. reuteri TMW 1.106 dextransucrase showed a clearly
increased glucosylation activity and was thus used for all further experiments. To evaluate
incubation conditions which allow for the isolation of conjugate amounts suitable for structure
elucidation by NMR spectroscopy, different amounts of substrate, sucrose, and enzyme were
tested in a small-scale. It was concluded that the highest amount of conjugates can be obtained
by using the highest acceptor concentration possible (with regards to solubility), 500 mM
sucrose, and 100 mU mutant L. reuteri TMW 1.106 dextransucrase / mL substrate solution.
Therefore, these conditions were used for the glucosylation of all substrates.
92
93
94
95
96
97
98
99
To obtain information on the portion and the molecular masses of the conjugates formed, the
incubation mixtures were analyzed by HPLC-DAD/MS. As a first approximation, conjugation
efficiency can be derived from the UV/Vis absorption signal intensity. However, changes in the
UV/Vis response due to conjugation need to be considered. The conjugates were named after
the substrate (prefix) and the number of added α-glucosyl moieties (suffix, derived from the
m/z of the individual conjugates). For unambiguous structural elucidation, conjugates were
isolated by preparative HPLC and characterized by using several one- and two-dimensional
NMR experiments. Although the proton spectra yield valuable information on the flavonoid
5

ACCEPTED MANUSCRIPT
100
101
102
103
104
105
106
107
108
109
110
derived protons, sugar-derived protons can often not be resolved due to signal overlap.
However, a Heteronuclear Single Quantum Coherence (HSQC) experiment mostly allows for
the resolution of all signals, because they are spread into the 13C dimension. To identify
downfield shifted HSQC signals due to α-glucose substitution, spectra of the untreated
compounds were also recorded. Identification and assignment of the individual HSQC signals
was achieved by using H,H-Correlated Spectroscopy (COSY), Heteronuclear Two-bond
Correlation (H2BC), Total Correlated Spectroscopy (TOCSY), and HSQC-TOCSY
experiments. Eventually, unambiguous information on interglycosidic linkages and on
substitution of the flavonoid residues was obtained from the Heteronuclear Multiple Bond
1
Correlation (HMBC) spectra. The H and 13C chemical shifts of all flavonoids and their
conjugates are provided in the supplementary data (Tab. S1-S8).
111
6

ACCEPTED MANUSCRIPT
112
3.1
Quercetin and quercetin glycosides
113
114
115
116
117
118
Fig. 1: HPLC-DAD/MS chromatograms of incubations of quercetin (Q, final concentration: 3
mM, subfigure A), quercetin-3-O-β-glucoside (Q3G, final concentration: 5 mM, subfigure B),
and rutin (R, final concentration: 3 mM, subfigure C) with 100 mU / mL mutant L. reuteri
TMW 1.106 dextransucrase and 500 mM sucrose. The structures of the conjugates (subfigure
D) were determined by NMR spectroscopy.
7

ACCEPTED MANUSCRIPT
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
One major monoglucosylated conjugate (Q-G1) was obtained after incubation of quercetin with
mutant L. reuteri TMW 1.106 dextransucrase (Fig. 1A). According to the peak areas, this
conjugate was present in about equal amounts than quercetin which indicates a high conjugation
efficiency (assuming an at least roughly comparable UV response). In accordance with its
molecular mass, signals of one α-glucose were detected in the HSQC spectrum of the isolated
1
compound. This sugar unit was identified by its characteristic H and 13C chemical shifts and
3
the JHH coupling constant of the anomeric proton. With regards to the quercetin unit, a
downfield shift of C5’H5’ and C6’H6’ was observed compared to unmodified quercetin.
Furthermore, H1 of glucose as well as H2’, H5’, and H6’ of quercetin showed an HMBC
13
correlation at 146.5 ppm. Because of this C chemical shift and because of the absence of
another HMBC signal between 145 and 150 ppm, these results show that the B-ring is
substituted at position O4’ (Fig. 1D). This conclusion was based on previous NMR data on
quercetin α-glucosides [17]. As reported by Moon, Lee, Jhon, Jun, Kang, Sim, Choi, Moon and
Kim [17], substitution at position O4’ leads to almost identical 13C chemical shifts of C3’ and
13
C4’ (146.4 and 146.9 ppm), whereas substitution at position O3’ results in two distinct C
chemical shifts for C3’ (144.8 ppm) and C4’ (149.8 ppm). This is also in good agreement with
NMR data from our laboratory for glucosylated phenolic acids with a catechol function.
Notably, the HPLC chromatogram also contained very low portions of another
monoglucosylated product (retention time 43 min) which most likely corresponds to the O3’
substituted conjugate. However, from the very low abundance of this conjugate, it can
concluded that position O4’ of quercetin is highly preferred. In previous reports, glucansucrases
from Leuconostoc mesenteroides B-1299CB and Leuconostoc mesenteroides NRRL B-512F
were also reported to prefer position O4’ over position O3’, but compared to our results, higher
portions of O3’ glucosylated quercetin were detected [14, 17]. Furthermore, Malbert, Moulis,
8

ACCEPTED MANUSCRIPT
144
145
146
147
148
149
150
151
152
153
Brison, Morel, Andre and Remaud-Simeon [20] obtained several mono- and oligoglucosylated
products by using a branching sucrase from Leuconostoc mesenteroides B-1299. However, the
selective glucosylation of quercetin at position O4’ observed for mutant L. reuteri TMW 1.106
dextransucrase may be advantageous for some applications. The glucosylation efficiencies
described for the previously used wildtype glucansucrases showed broad variation from 4 %
[14] to 25 % [17] and 50 % [20]. In addition, some mutants of the branching sucrase yielded
glucosylation efficiencies of even about 90 % [20]. Therefore, mutant L. reuteri TMW 1.106
dextransucrase provides a high glucosylation efficiency (about 50 % based on the peak areas)
compared to some other enzymes, but a higher efficiency may be achieved by creating and
screening more mutants.
154
155
156
157
Conjugation of quercetin-3-O-β-glucoside also resulted in one major product (Q3G-G1a in Fig
1B) which was isolated from the reaction mixture. In addition, it was possible to purify one
diglucosylated product (Q3G-G2) and two other monoglucosylated conjugates (Q3G-G1b and
Q3G-G1c).
158
159
160
161
162
163
164
Just as for Q-G1, comparison of the main product Q3G-G1a with unmodified quercetin-3-O-β-
glucoside showed that the HSQC signals derived from the B-ring of quercetin showed
significantly different chemical shifts. Furthermore, the chemical shifts of the additional α-
glucose unit were in good agreement with the chemical shifts of the α-glucose unit in Q-G1,
suggesting a similar chemical environment. Eventually, the attachment of α-glucose to position
O4’ of the B-ring was confirmed by an HMBC correlation between H1 of glucose and C4’ of
quercetin at 5.41 / 147.0 ppm.
165
166
167
168
For the two monoglucosylated conjugates Q3G-G1b and Q3G-G1c, quercetin derived HSQC
signals were not significantly altered, suggesting substitution of the β-glucose unit. After
assignment of the HSQC signals by using COSY/TOCSY spectra, characteristic chemical shifts
of the β-glucose derived signals were used to assign the substitution position of the β-glucose
9

ACCEPTED MANUSCRIPT
169
170
171
172
unit (Fig. S3). Q3G-G1b showed a significantly downfield shifted C3H3 signal (3.44 / 84.8
ppm), whereas Q3G-G1c showed a downfield shifted C4H4 signal (3.38 / 79.1 ppm). Therefore,
these two monoglucosylated conjugates result from substitution of β-glucose at position O3 and
O4, respectively.
173
174
175
176
177
178
179
180
181
182
183
As expected, the diglucosylated conjugate Q3G-G2 showed two sets of α-glucose derived
signals (Fig. 2). From the characteristically shifted HSQC signals of quercetin and the chemical
shifts of one α-glucose unit, it was concluded that quercetin is substituted at position O4’ of the
B-ring. The attachment of the corresponding glucose unit to position O4’ of quercetin was also
confirmed by an HMBC correlation between H1 of glucose and C4’ of quercetin. The chemical
shifts of the β-glucose unit and the remaining α-glucose unit were in good agreement with the
chemical shifts observed for α/β-glucose in Q3G-G1b. Only the C1H1 and C2H2 signals of β-
glucose were slightly shifted which is most likely a result of the B-ring substitution. However,
the attachment of the second α-glucose to position O3 of β-glucose was clearly demonstrated
by the diagnostic C3H3 signal of β-glucose (Fig. 2) and an HMBC correlation between H1 of
α-glucose and C3 of β-glucose.
184
10

ACCEPTED MANUSCRIPT
185
186
187
188
Fig. 2: HSQC spectrum and structure of the diglucosylated conjugate of quercetin-3-O-β-
glucoside (Q3G-G2). Signals belonging to the two α-glucose residues are colored in red and
blue. The diagnostic, downfield shifted C3H3 signal of the β-glucose unit is encircled in orange.
189
190
191
192
193
194
195
196
197
The predominant abundance of Q3G-G1a which is glucosylated at position O4’ of the B-ring
demonstrates that the preference for this position is not altered by the β-glucosyl residue at
position O3 of quercetin. Furthermore, the other three monoglucosylated conjugates show that
glucosylation can appear at different positions of the β-glucosyl residue. In addition, already
glucosylated compounds can be further substituted with α-glucose which results in the
formation of Q3G-G2. From the peak intensities in the chromatogram, a comparably high
glucosylation efficiency can be derived, because more than half of quercetin-3-O-β-glucoside
is glucosylated.
11

ACCEPTED MANUSCRIPT
198
199
200
201
202
203
When rutin was incubated with mutant L. reuteri TMW 1.106 dextransucrase, only two
monoglucosylated conjugates were detected. For R-G1a, the rutinose derived HSQC signals
1
13
were in good agreement with unsubstituted rutin. Furthermore, H and C chemical shifts of
quercetin as well as the additional α-glucose unit were in good agreement with the chemical
shifts obtained for Q3G-G1a. Therefore, R-G1a is glucosylated at position O4’ which was also
confirmed by an HMBC correlation between H1 of glucose and C4’ of quercetin.
204
205
206
207
208
209
210
For R-G1b, the quercetin signals were unmodified compared to the standard. In addition,
assignment of the rutinose HSQC signals demonstrated that the chemical shifts of β-glucose
were not significantly altered. However, α-rhamnose showed a clearly downfield shifted signal
at 3.22 / 81.8 ppm. This signal was identified as C4H4 by using the COSY spectrum as well as
an HMBC correlation between C4 and the characteristically located H6 of α-rhamnose (Fig. 3).
From an HMBC correlation between C4 of the α-rhamnose unit and H1 of the α-glucose unit,
it was also unambiguously demonstrated that the α-glucose unit is attached to this position.
211
212
213
214
215
These results demonstrate that the addition of a disaccharide to position O3 of quercetin does
not impede glucosylation at position O4’, but prevents glucosylation of the β-glucose unit.
Furthermore, the glucosylation efficiency seems to be lowered, as the major part of rutin
remained unmodified under the conditions used. However, glucosylation of the rhamnosyl
residue at position O4 is an interesting novel feature of glucansucrases.
216
12

ACCEPTED MANUSCRIPT
217
218
219
220
Fig. 3: HMBC spectrum and structure of the monoglucosylated conjugate of rutin (R-G1b).
Diagnostic signals which helped to identify C4 of α-rhamnose and demonstrate the attachment
of α-glucose to position O4 of α-rhamnose are encircled in red and blue.
221
13

ACCEPTED MANUSCRIPT
222
3.2
Cyanidin-3-glucoside
223
224
The chromatogram obtained for the cyanidin-3-O-β-glucoside incubation mixture contained
one additional, earlier eluting peak (Fig. 4).
225
226
227
228
229
230
Fig. 4: HPLC-DAD/MS chromatogram of the incubation of cyanidin-3-O-β-glucoside (Cy3G,
final concentration: 8 mM) with 100 mU / mL mutant L. reuteri TMW 1.106 dextransucrase
and 500 mM sucrose. The structure of the monoglucosylated conjugate Cy3G-G1 is shown in
Fig. 5.
231
232
233
234
235
236
237
238
The compound (Cy3G-G1) showed a quasimolecular ion at m/z 611 which demonstrates
monoglucosylation of cyanidin-3-O-β-glucoside. For structural elucidation, the conjugate was
isolated by preparative HPLC and NMR spectra were recorded in methanol-d4 with 0.1 % DCl.
This solvent was used because it provides complete solubilization and well resolved spectra of
the flavylium cation form. As expected, the HSQC spectrum contained signals typical for an α-
glucose in addition to the signals of cyanidin-3-O-β-glucoside. Besides the slightly shifted
C4H4 signal, the HSQC signals of cyanidin showed identical chemical shifts than for
14

ACCEPTED MANUSCRIPT
239
240
241
242
unmodified cyanidin-3-O-β-glucoside. Thus it can be concluded that the β-glucose residue is
glucosylated. After assigning the HSQC signals of the β-glucose unit by using COSY/TOCSY
experiments, it was observed that the C6H6 signals showed a characteristic downfield shift (Fig
5).
243
244
245
246
247
Fig. 5: HSQC spectrum and structure of the monoglucosylated conjugate of cyanidin-3-O-β-
glucoside (Cy3G-G1). The diagnostic, downfield shifted C6H6 signals of the β-glucose unit are
encircled in blue.
248
249
250
251
252
These downfield shifted signals strongly suggest substitution at position O6 of the β-glucose
unit, which was eventually confirmed by an HMBC correlation between C6 of the β-glucose
and H1 of the α-glucose. Therefore, cyanidin-3-O-β-glucoside shows a significantly different
product spectrum than the structurally related quercetin-3-O-β-glucoside which was
15

ACCEPTED MANUSCRIPT
253
254
255
glucosylated at position O4’ of the B-ring and at positions O3 and O4 of β-glucose. Most likely,
the structural differences between the two compounds result in a different orientation of
cyanidin-3-O-β-glucoside in the active site of the enzyme.
256
257
3.3
Naringin and neohesperidin dihydrochalcone
258
259
260
261
262
263
264
Fig. 6: HPLC-DAD/MS chromatograms of incubations of naringin (Na, final concentration: 30
mM, subfigure A) and neohesperidin dihydrochalcone (Ne, final concentration: 50 mM,
subfigure B) with 100 mU / mL mutant L. reuteri TMW 1.106 dextransucrase and 500 mM
sucrose. The peaks marked with an asterisk in subfigure B were also present in the untreated
neohesperidin dihydrochalcone preparation (see text). The structures of the conjugates
(subfigure C) were determined by NMR spectroscopy.
16

ACCEPTED MANUSCRIPT
265
266
267
268
269
270
271
272
273
274
275
Two monoglucosylated conjugates were detected in the incubation mixture of naringin with
mutant L. reuteri TMW 1.106 dextransucrase (Fig. 6A). However, the peak intensity suggested
that the conjugates are lowly abundant and that glucosylation efficiency is rather low.
Nevertheless, it was possible to isolate and characterize the two conjugates. The HSQC
spectrum of Na-G1a yielded unmodified neohesperidose signals and the anomeric signal of the
1
α-glucose showed a comparably high H chemical shift. This was already observed for
quercetin (glycoside) conjugates which are glucosylated at position O4’ of the B-ring. Because
the C2’H2’, C3’H3’, C5’H5’, C6’H6’ of naringin were also shifted downfield, it can be
concluded that naringin is glucosylated at the hydroxyl group of the B-ring. The attachment of
α-glucose to the B-ring was also confirmed by an HMBC correlation between H1 of α-glucose
and C4’ of naringin (Fig. S4).
276
277
278
279
280
281
The second monoglucosylated conjugate Na-G1b did not show any signal drift for the HSQC
signals derived from the aglycone and the α-rhamnose unit. However, just as for the conjugate
of cyanidin-3-O-β-glucoside, C6H6 signals of the naringin-bound β-glucose unit showed a
13
characteristic C downfield shift to 66 ppm (Fig. S5). Therefore it can be concluded that the
additional α-glucose unit is bound to position O6 of β-glucose which was also confirmed by a
corresponding HMBC correlation.
282
283
284
285
286
287
288
289
HPLC-DAD/MS and NMR spectroscopic investigations revealed that the neohesperidin
dihydrochalcone preparation contained some residual amounts of two flavonoids which were
identified as neohesperidin and diosmin-7-neohesperidoside (peaks marked with an asterisk in
Fig. 6B). The monoglucosylated neohesperidin dihydrochalcone conjugate formed by mutant
L. reuteri TMW 1.106 dextransucrase eluted near the two flavonoids and was therefore only
obtained in a mixture with neohesperidin. However, it was possible to assign all HSQC signals
corresponding to the aglycone (hesperetin) and the unsubstituted neohesperidose by using
COSY and TOCSY experiments. With regards to the monoglucosylated neohesperidin
17

ACCEPTED MANUSCRIPT
290
291
292
293
294
295
296
297
298
299
300
dihydrochalcone conjugate, the chemical shifts of the HSQC signals derived from the aglycone
were comparable to the chemical shifts obtained for unmodified neohesperidin
dihydrochalcone. In the anomeric region of the HSQC spectrum, two anomeric signals were
detected at 5.10/100.1 ppm and 5.14/97.0 ppm besides the additional signals of a carbohydrate-
bound α-glucose unit. The carbon shifts of these two signals were similar to unmodified
neohesperidose, therefore the signals can be assigned to β-glucose and α-rhamnose. The
assignment of the remaining HSQC signals of these two sugar units showed that the C4H4
signal of the α-rhamnose unit was clearly shifted downfield in the 13C dimension. Furthermore,
an HMBC correlation between H1 of α-glucose and C4 of α-rhamnose was found. Therefore,
it was possible to conclude that the α-glucose unit is bound to position O4 of the α-rhamnose
unit.
301
302
303
304
305
306
307
308
309
310
311
312
313
The structures of the naringin and neohesperidin dihydrochalcone conjugates demonstrate that
mutant L. reuteri TMW 1.106 dextransucrase is also able to glucosylate flavonoid-bound
neohesperidose units. However, naringin was glucosylated at the β-glucose unit and
neohesperidin dihydrochalcone was glucosylated at the α-rhamnose unit. These results suggest
that the aglycone has a significant influence on the glucosylation position. This may be caused
by different orientations of the acceptor substrate in the active site of the dextransucrases.
However, just as rutin (R-G1b), neohesperidin dihydrochalcone was glucosylated at position
O4 of the α-rhamnose unit. Thus, the position to which α-rhamnose is attached seems to be of
secondary importance and conjugation of α-rhamnose preferably occurs at position O4.
However, glucosylation of α-rhamnose was only observed for substrates which lack substitution
of the β-glucose unit. Therefore, it is likely that the corresponding conjugates are only formed
if the β-glucose units are not in a suitable position in the active site. Furthermore, Na-G1a,
which is substituted at position O4’ of naringin, is a rather uncommon reaction product because
18

ACCEPTED MANUSCRIPT
314
315
previous studies suggested that glucansucrases prefer vicinal phenolic hydroxyl groups for
glucosylation [14, 29].
316
317
3.4
Epigallocatechin gallate and dihydromyricetin
318
319
320
321
322
323
Fig. 7: HPLC-DAD/MS chromatograms of incubations of dihydromyricetin (DHM, final
concentration: 30 mM, subfigure A) and epigallocatechin gallate (EGCG, final concentration:
20 mM, subfigure B) with 100 mU / mL mutant L. reuteri TMW 1.106 dextransucrase and 500
mM sucrose. The structures of the conjugates (subfigure C) were determined by NMR
spectroscopy.
324
19

ACCEPTED MANUSCRIPT
325
326
327
328
329
In the incubation mixture of dihydromyricetin, a monoglucosylated conjugate was largely
predominant besides low amounts of a diglucosylated conjugate (Fig. 7A). Preparative HPLC
allowed for the isolation of both compounds, but both fractions still contained residual
unmodified dihydromyricetin. However, signals derived from the additional α-glucose units
were clearly assignable because of their different chemical shift.
330
331
332
333
334
335
The HSQC spectrum of the monoglucosylated conjugate DHM-G1 contained the characteristic
α-glucose signals as well as one downfield shifted signal in the aromatic region which was
identified as the C2’H2’ and C6’H6’ signals of the conjugated dihydromyricetin. Because
glucosylation at position O3’ / O5’ would result in two distinct signals, these results suggest
that dihydromyricetin is glucosylated at position O4’. This was also confirmed by an HMBC
correlation between C4’ of the conjugated dihydromyricetin and H1 of α-glucose.
336
337
338
339
340
341
342
343
344
345
For DHM-G2, the chemical shifts of conjugated dihydromyricetin were identical to DHM-G1,
thus this conjugate is also substituted at position O4’. As expected for a diglucosylated
conjugate, two anomeric signals were detected (Fig. S6). The anomeric signal of the
dihydromyricetin-bound glucosyl unit was identified based on its chemical shift and an HMBC
correlation between H1 and C4’ of dihydromyricetin. From the assignment of the remaining
HSQC signals, it was demonstrated that the C4H4 signal of the dihydromyricetin-bound α-
13
glucose unit was significantly downfield shifted in the C dimension (69.1 to 78.6 ppm). In
addition, the chemical shifts of the second glucosyl unit were in good agreement with glycoside-
bound, terminal α-glucose. Therefore, DHM-G2 contains a maltosyl unit bound to position O4’
of the B-ring.
346
347
348
349
Incubation of epigallocatechin gallate with mutant L. reuteri TMW 1.106 dextransucrase
yielded two monoglucosylated conjugates which eluted near to each other and also near to
unmodified epigallocatechin gallate (Fig. 7B). As a consequence, the two conjugates were only
isolated in a mixture with each other and unconjugated epigallocatechin gallate. Thus, the
20

ACCEPTED MANUSCRIPT
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
HSQC spectrum contained two close-by anomeric α-glucose signals (Fig. 8A). However, the
five non-anomeric HSQC signals of the two α-glucose units were identical. The aromatic region
of the spectrum was dominated by two signals (6.82/108.4 ppm and 6.40/105.2 ppm) which can
be assigned to the unconjugated B-ring (C2’H2’ / C6’H6’) and unconjugated gallic acid
(C2’’H2’’ / C6’’H6’’) of epigallocatechin gallate. Besides these signals (and the C6H6 / C8H8
signals), one additional signal at 6.48/105.2 ppm was detected. This signal showed a
comparable 13C chemical shift than C2’H2’/C6’H6’ of the unconjugated B-ring which suggests
that one conjugate is substituted at position O4’. This was confirmed by an HMBC correlation
between the H1 of the glucose unit at 4.83 ppm and the carbon with a chemical shift of 133.3
ppm (red bar in Fig 9B, assigned as C4’). The other α-glucose showed an HMBC correlation
between the anomeric proton (5.06 ppm) and a carbon at 137.7 ppm (blue bar in Fig. 8B) which
corresponds to C4’’ of the ester-bound gallic acid. Therefore, an α-glucose unit is bound to the
gallic acid residue, but glucosylation does not impact the proton shift of the substituted gallic
acid. This phenomenon was also previously observed by Moon, Kim, Lee, Jin, Kim and Kim
[16].
365
21

ACCEPTED MANUSCRIPT
366
367
368
369
370
371
Fig. 8. A: HSQC spectrum and structures of the two monoglucosylated conjugates of
epigallocatechin gallate (EGCG-G1a and EGCG-G1b) which were only obtained in a mixture.
Signals belonging to the two different α-glucose residues are colored in red and blue. B: HMBC
spectrum of the conjugate mixture. Diagnostic correlations between H1 of the α-glucose units
and C4’ of the B-ring and C4’’ of gallic acid are marked by red and blue bars, respectively.
372
373
374
375
376
377
378
379
380
The incubation of dihydromyricetin and epigallocatechin gallate with mutant L. reuteri TMW
1.106 dextransucrase demonstrates that flavonoids with a pyrogallol function (located at the B-
ring or ester-bound gallic acid) are also efficiently glucosylated by this enzyme. From the
structures of the conjugates it can also be derived that pyrogallol functions are preferably
glucosylated at position O4 as it was also observed for flavonoids with catechol functions.
Maltosyl-substituted conjugates were not detected for all substrates but dihydromyricetin,
therefore, the formation of this conjugate type is most likely also dependent on the structure of
the aglycone. Previous studies on the glucosylation of dihydromyricetin and epigallocatechin
22

ACCEPTED MANUSCRIPT
381
382
383
384
385
386
387
gallate by glucansucrase from Leuconostoc mesenteroides B-1299CB also showed preferred
glucosylation at position O4 of the B-ring and gallic acid [16, 18]. In contrast to the previous
studies, oligoglucosylated conjugates or conjugates which are additionally glucosylated at
position O7 of the flavonoid were not observed in our study. This may be a result of the
dextransucrase specificity and or the mutations introduced. However, these results once again
demonstrate that mutant L. reuteri TMW 1.106 dextransucrase is suitable for effective and
specific monoglucosylation without creating high amounts of other products.
388
389
3
CONCLUSION
390
391
392
393
394
395
396
397
398
399
400
401
402
The results of our study demonstrate that mutant L. reuteri TMW 1.106 dextransucrase is highly
suitable for the synthesis of different flavonoid glucosides. For substrates such as quercetin and
its glycosides, a high glucosylation efficiency was achieved, whereas the incubation of e.g.
neohesperidin dihydrochalcone and naringin only yielded low portions of glucoconjugates. The
different substrates were not only glucosylated with varying efficiencies but also at different
positions. While glucosylation of the phenolic hydroxyl groups of the flavonoid B-ring occurred
almost exclusively at position O4’, flavonoid-bound β-glucosyl units were conjugated at
position O3, O4, and O6 depending on the substrate. Furthermore, it was possible to
demonstrate that flavonoids with a rutinose and a neohesperidose residue can be glucosylated
either at position O6 of the glucose unit or at position O4 of the α-rhamnose unit. Overall,
glucosylation by mutant L. reuteri TMW 1.106 dextransucrase is highly dependent both on the
flavonoid residue and the glycosyl residue. Future studies could evaluate how stability,
antioxidant properties, bioavailability, and taste of the conjugates are affected by glucosylation.
403
23

ACCEPTED MANUSCRIPT
404
405
4
EXPERIMENTAL
Materials
4.1
406
407
408
409
410
411
412
413
414
Quercetin (≥ 98 %), quercetin-3-O-β-glucoside (≥ 98 %), rutin (≥ 97 %), naringin (≥ 98 %),
neohesperidin dihydrochalcone (≥ 98 %), epigallocatechin gallate (≥ 98 %), and (if not stated
otherwise) all other chemicals were purchased from VWR (Radnor, PA, USA), Sigma Aldrich
(Schnelldorf, Germany), Carl Roth (Karlsruhe, Germany), or Carbosynth (Newbury, UK).
Cyanidin-3-O-β-glucoside was extracted from freeze-dried blackberry powder by using acidic
methanol. Anthocyanins were concentrated by solid phase extraction as described previously
[30] and purified by preparative HPLC (see below). Dihydromyricetin was obtained from a
commercial dietary supplement. pLIC-SGC1 was a gift from Nicola Burgess-Brown (Addgene
plasmid # 39187).
415
4.2
Molecular cloning and site-directed mutagenesis
416
417
418
419
420
421
422
423
424
The gene encoding the previously described dextransucrase from L. reuteri TMW 1.106 [27]
was amplified from genomic DNA by using a Phusion High-Fidelity PCR kit (Thermo Fisher
Scientific, Waltham, MA, USA) and the primers listed in Table S9. The amplified and purified
gene was subsequently cloned into the pLIC-SGC1 vector by using a ligation independent
cloning approach [31-33]. Site directed mutagenesis of L. reuteri dextransucrase was conducted
by using a QuikChange Lightning Multi Site-Directed Mutagenesis Kit (Agilent Technologies,
Santa Clara, CA, USA) and the mutagenic primer listed in Table S9. After purification, the
sequence of the mutagenic plasmid was confirmed by Sanger sequencing (Eurofins GATC
Biotech, Konstanz, Germany).
425
4.3
Heterologous expression
426
427
For heterologous expression of the mutated dextransucrase gene, the plasmid was transformed
into OneShot BL21 Star (DE3) cells by heat shock. After transferring one colony into 5 mL of
24

ACCEPTED MANUSCRIPT
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
LB medium (100 µg ampicillin / mL), cells were grown at 37 °C and 225 rpm for 6 h. The
culture was used to inoculate 200 mL of fresh LB medium (100 µg ampicillin / mL) which was
further incubated at 37 °C and 225 rpm for 3 h. After induction of protein expression with
isopropyl-β-D-thiogalactopyranoside (final concentration: 0.1 mM), the culture was incubated
overnight at room temperature and 225 rpm. Cells were harvested by centrifugation and re-
suspended in binding buffer (50 mM sodium phosphate, 300 mM NaCl, 10 mM imidazole, pH
7.5). Lysis of the re-suspended cells was carried out by sonication (amplitude 50 %, 3x 20 s
pulse, 59.9 s pause) with a Sonifier W-250 D (Branson, Dabury, CT, USA) and cell debris were
removed by centrifugation (30 min, 4 °C, 20000 x g). For purification of the recombinant
dextransucrase, the clear supernatant was transferred to a HisPur Ni-NTA resin (Thermo Fisher
Scientific; pre-equilibrated with binding buffer) and incubated at 4 °C for 1 h. After washing
with 4 volumes of binding buffer, recombinant proteins were eluted with elution buffer (50 mM
sodium phosphate, 300 mM NaCl, 100 mM imidazole, pH 7.5) which resulted in a typical yield
of 100 mg dextransucrase / L culture. Hydrolytic activity of the recombinant dextransucrase
was determined by analyzing the release of para-nitrophenol from 35 mM para-nitrophenyl-α-
glucoside (25 mM sodium acetate buffer, supplemented with 1 mM CaCl2, pH 5.5) in a
microplate reader (Infinite M200 PRO, Tecan, Männedorf, Switzerland) at 400 nm. A standard
curve was used to calculate the volume activity (1 µmol of released para-nitrophenol / min) per
µL.
447
4.4
Acceptor reactions
448
449
450
451
452
For glucosylation, quercetin, quercetin-3-O-β-glucoside, rutin, cyanidin-3-O-β-glucoside,
naringin, neohesperidin dihydrochalcone, epigallocatechin gallate, and dihydromyricetin were
dissolved in DMSO and four volumes of 25 mM sodium acetate buffer (1 mM CaCl2, pH 5.5)
were added. This procedure allowed for solubilization of comparably high amounts of these
substrates (quercetin: 3 mM, quercetin-3-O-β-glucoside: 5 mM, rutin: 3 mM, cyanidin-3-O-β-
25

ACCEPTED MANUSCRIPT
453
454
455
456
457
glucoside: 8 mM, naringin: 30 mM, neohesperidin dihydrochalcone: 50 mM, epigallocatechin
gallate: 20 mM, dihydromyricetin: 30 mM). After the addition of sucrose (final concentration:
500 mM), 100 mU of L. reuteri TMW 1.106 L242A dextransucrase / mL substrate solution
were added and the mixture was incubated for 24 h at 30 °C. The enzyme was inactivated by
heating to 95 °C for 5 min and the mixture was used for further analysis and fractionation.
458
4.5
Glucoconjugate analysis and purification
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
For chromatographic analysis of the glucoconjugates, the glucosylation mixtures were diluted
and analyzed on a Surveyor HPLC-DAD system coupled to an LXQ linear ion trap MSn system
(Thermo Fisher Scientific). A Kinetex C18 column (100 mm x 4.6 mm i.d., 2.6 µm particle
size; Phenomenex, Torrance, CA, USA) and the following gradient composed of water with 0.1
% formic acid (eluent A) and acetonitrile with 0.1 % formic acid (eluent B) were used at 0.5
mL/min and 30 °C: 0-5 min, isocratic 3 % B; 5-25 min, linear to 8 % B; 25-45 min, linear to
28 % B; 45-55 min, linear to 60 % B; 55-56 min, linear to 100 % B; 56-62 min, isocratic 100
% B; 62-64 min, linear to 3 % B; 64-70 min, isocratic 3 % B. For the analysis of cyanidin-3-O-
β-glucoside, 0.5 % formic acid was added to both eluents. Flavonoids were detected at 280 nm
(neohesperidin dihydrochalcone, epigallocatechin gallate, and dihydromyricetin), 320 nm
(quercetin, quercetin-3-O-β-glucoside, rutin, and naringin), and 510 nm (cyanidin-3-O-β-
glucoside). For mass spectrometric analysis of the unmodified and conjugated flavonoids,
electrospray ionization was used in positive (cyanidin-3-O-β-glucoside) and negative (all other
compounds) mode. Prior to purification with preparative HPLC, conjugated flavonoids were
enriched by using solid phase extraction on a Chromabond C18 (Macherey Nagel, Düren,
Germany; preconditioned with methanol and water) stationary phase. The glucosylation
mixtures were loaded onto the column, the column was washed with water, and conjugates were
eluted by using methanol. Fractionation of the conjugates was conducted on a preparative
HPLC-UV system (LC-8A pumps, SPD-20A UV detector, Shimadzu, Kyoto, Japan) equipped
26

ACCEPTED MANUSCRIPT
478
479
480
481
482
483
484
with a Luna C18 column (250 mm × 10 mm i.d., 5 μm particle size, Phenomenex). The same
conditions as described above were used at 8 mL/min and room temperature. In some cases, the
gradient was truncated to: 0-5 min, linear from 3 % B to 8 % B; 5-25 min linear to 28 % B; 25-
35 min, linear to 40 % B; 35-36 min, linear to 100 % B; 36-41 min, isocratic 100 % B; 41-42
min, linear to 3 % B, 42-47 min, isocratic 3 % B. Detection of the conjugates was carried out
at 510 nm (cyanidin-3-O-β-glucoside) and 280 nm (all other compounds). The resulting
fractions were evaporated and freeze-dried.
485
4.6
NMR spectroscopy
486
487
488
489
490
491
492
NMR spectroscopic analysis of the standard compounds and conjugates was carried out on an
Ascend 500 MHz spectrometer (Bruker, Rheinstetten, Germany) equipped with a Prodigy
cryoprobe. Cyanidin-3-O-β-glucoside was dissolved in methanol-d4 whereas all other
conjugates were dissolved in DMSO-d6. Chemical shifts were referenced to the solvent residual
peak (methanol-d4: 3.31 / 49.00 ppm; DMSO-d6: 2.50 / 39.52 ppm) according to Gottlieb,
Kotlyar and Nudelman [34]. Standard Bruker parameter sets were used to record proton, HSQC,
H2BC, COSY, (HSQC-)TOCSY, and HMBC spectra.
493
ACKNOWLEDGEMENT
494
495
The authors are very grateful to Frank Jakob and Rudi Vogel from the Chair of Technical
Microbiology (TU Munich) for providing L. reuteri TMW 1.106 genomic DNA.
496
497
498
FUNDING
This work was supported by the Carl-Zeiss Foundation.
REFERENCES
499
500
501
502
[1] P.C.H. Hollman, A. Cassidy, B. Comte, M. Heinonen, M. Richelle, E. Richling, M. Serafini,
A. Scalbert, H. Sies, S. Vidry, The biological relevance of direct antioxidant effects of
polyphenols for cardiovascular health in humans is not established. J. Nutr. 141 (2011) 989S-
1009S. https://doi.org/10.3945/jn.110.131490.
27

ACCEPTED MANUSCRIPT
503
504
505
[2] E. Tripoli, M. La Guardia, S. Giammanco, D. Di Majo, M. Giammanco, Citrus flavonoids:
Molecular structure, biological activity and nutritional properties: A review. Food Chem. 104
(2007) 466-479. https://doi.org/10.1016/j.foodchem.2006.11.054.
506
507
508
[3] G.E. Inglett, L. Krbechek, B. Dowling, R. Wagner, Dihydrochalcone sweeteners - Sensory
and stability evaluation. J. Food Sci. 34 (1969) 101-103. https://doi.org/10.1111/j.1365-
2621.1969.tb14371.x.
509
510
511
[4] M. Plaza, T. Pozzo, J.Y. Liu, K.Z.G. Ara, C. Turner, E.N. Karlsson, Substituent effects on
in vitro antioxidizing properties, stability, and solubility in flavonoids. J. Agric. Food Chem.
62 (2014) 3321-3333. https://doi.org/10.1021/jf405570u.
512
513
514
515
[5] K. Murota, N. Matsuda, Y. Kashino, Y. Fujikura, T. Nakamura, Y. Kato, R. Shimizu, S.
Okuyama, H. Tanaka, T. Koda, K. Sekido, J. Terao, α-Oligoglucosylation of a sugar moiety
enhances the bioavailability of quercetin glucosides in humans. Arch. Biochem. Biophys. 501
(2010) 91-97. https://doi.org/10.1016/j.abb.2010.06.036.
516
517
518
[6] X. Zhang, J. Wang, J.M. Hu, Y.W. Huang, X.Y. Wu, C.T. Zi, X.J. Wang, J. Sheng, 2016.
Synthesis and biological testing of novel glucosylated epigallocatechin gallate (EGCG)
derivatives. Molecules. 21, 620. https://doi.org/10.3390/molecules21050620.
519
520
521
[7] Y.S. Lee, J.B. Woo, S.I. Ryu, S.K. Moon, N.S. Han, S.B. Lee, Glucosylation of flavonol
and flavanones by Bacillus cyclodextrin glucosyltransferase to enhance their solubility and
stability. Food Chem. 229 (2017) 75-83. https://doi.org/10.1016/j.foodchem.2017.02.057.
522
523
524
525
[8] C. Aoki, Y. Takeuchi, K. Higashi, Y. Okamoto, A. Nakanishi, M. Tandia, J. Uzawa, K.
Ueda, K. Moribe, Structural elucidation of a novel transglycosylated compound α-glucosyl
rhoifolin and of α-glucosyl rutin by NMR spectroscopy. Carbohydr. Res. 443 (2017) 37-41.
https://doi.org/10.1016/j.carres.2017.03.011.
526
527
528
[9] J.S. Cho, S.S. Yoo, T.K. Cheong, M.J. Kim, Y. Kim, K.H. Park, Transglycosylation of
neohesperidin dihydrochalcone by Bacillus stearothermophilus maltogenic amylase. J. Agric.
Food Chem. 48 (2000) 152-154. https://doi.org/10.1021/jf991044y.
529
530
531
532
[10] J.L. Gonzalez-Alfonso, L. Leemans, A. Poveda, J. Jimenez-Barbero, A.O. Ballesteros, F.J.
Plou, Efficient α-glucosylation of epigallocatechin gallate catalyzed by cyclodextrin
glucanotransferase from Thermoanaerobacter species. J. Agric. Food Chem. 66 (2018) 7402-
7408. https://doi.org/10.1021/acs.jafc.8b02143.
533
534
535
536
[11] T. Kometani, T. Nishimura, T. Nakae, H. Takii, S. Okada, Synthesis of neohesperidin
glycosides and naringin glycosides by cyclodextrin glucanotransferase from an alkalophilic
Bacillus
species.
Biosci.
Biotechnol.
Biochem.
60
(1996)
645-649.
https://doi.org/10.1271/bbb.60.645.
537
538
539
540
[12] S.J. Lee, J.C. Kim, M.J. Kim, M. Kitaoka, C.S. Park, S.Y. Lee, M.J. Ra, T.W. Moon, J.F.
Robyt, K.H. Park, Transglycosylation of naringin by Bacillus stearothermophilus maltogenic
amylase to give glycosylated naringin. J. Agric. Food Chem. 47 (1999) 3669-3674.
https://doi.org/10.1021/jf990034u.
541
542
[13] Y. Suzuki, K. Suzuki, Enzymatic formation of 4G-α-D-glucopyranosyl-rutin. Agric. Biol.
Chem. 55 (1991) 181-187. https://doi.org/10.1080/00021369.1991.10870539.
543
544
545
546
[14] A. Bertrand, S. Morel, F. Lefoulon, Y. Rolland, P. Monsan, M. Remaud-Simeon,
Leuconostoc mesenteroides glucansucrase synthesis of flavonoid glucosides by acceptor
reactions in aqueous-organic solvents. Carbohydr. Res. 341 (2006) 855-863.
https://doi.org/10.1016/j.carres.2006.02.008.
28