Glycosyltransferase Co-Immobilization for Natural Product Glycosylation: Cascade Biosynthesis of the C-Glucoside Nothofagin with Efficient Reuse of Enzymes

Article abstract of DOI:10.1002/adsc.202001549

Sugar nucleotide-dependent (Leloir) glycosyltransferases are synthetically important for oligosaccharides and small molecule glycosides. Their practical use involves one-pot cascade reactions to regenerate the sugar nucleotide substrate. Glycosyltransferase co-immobilization is vital to advance multi-enzyme glycosylation systems on solid support. Here, we show glycosyltransferase chimeras with the cationic binding module Z_basic2 for efficient and well-controllable two-enzyme co-immobilization on anionic (ReliSorb SP400) carrier material. We use the C-glycosyltransferase from rice (Oryza sativa; OsCGT) and the sucrose synthase from soybean (Glycine max; GmSuSy) to synthesize nothofagin, the natural 3’-C-β-d-glucoside of the dihydrochalcone phloretin, with regeneration of uridine 5’-diphosphate (UDP) glucose from sucrose and UDP. Exploiting enzyme surface tethering via Z_basic2, we achieve programmable loading of the glycosyltransferases (～18 mg/g carrier; 60%–70% yield; ～80% effectiveness) in an activity ratio (OsCGT:GmSuSy=～1.2) optimal for the overall reaction rate (～0.2 mmol h^?1 g^?1 catalyst; 30 °C, pH 7.5). Using phloretin solubilized at 120 mM as inclusion complex with 2-hydroxypropyl-β-cyclodextrin, we demonstrate complete substrate conversion into nothofagin (～52 g/L; 21.8 mg product h^?1 g^?1 catalyst) at 4% mass loading of the catalyst. The UDP-glucose was recycled 240 times. The solid catalyst showed excellent reusability, retaining ～40% of initial activity after 15 cycles of phloretin conversion (60 mM) with a catalyst turnover number of ～273 g nothofagin/g protein used. Our study presents important progress towards applied bio-catalysis with immobilized glycosyltransferase cascades. (Figure presented.).

Full text of DOI:10.1002/adsc.202001549

FULL PAPER
DOI: 10.1002/adsc.202001549
Glycosyltransferase Co-Immobilization for Natural Product
Glycosylation: Cascade Biosynthesis of the C-Glucoside
Nothofagin with Efficient Reuse of Enzymes
Hui Liu,a Gregor Tegl,a and Bernd Nidetzkya, b,*
a
Institute of Biotechnology and Biochemical Engineering,
Graz University of Technology, NAWI Graz,
Petersgasse 12, 8010 Graz, Austria
Austrian Centre of Industrial Biotechnology (acib),
b
Petersgasse 14, 8010 Graz, Austria
Fax: (+43)-316-873-8434
Phone: (+43)-316-873-8400
E-mail: bernd.nidetzky@tugraz.at
Manuscript received: December 10, 2020; Revised manuscript received: February 9, 2021;
Version of record online: March 2, 2021
Supporting information for this article is available on the WWW under https://doi.org/10.1002/adsc.202001549
© 2021 The Authors. Advanced Synthesis & Catalysis published by Wiley-VCH GmbH. This is an open access article under the
terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided
the original work is properly cited.
Abstract: Sugar nucleotide-dependent (Leloir) glycosyltransferases are synthetically important for oligosac-
charides and small molecule glycosides. Their practical use involves one-pot cascade reactions to regenerate
the sugar nucleotide substrate. Glycosyltransferase co-immobilization is vital to advance multi-enzyme
glycosylation systems on solid support. Here, we show glycosyltransferase chimeras with the cationic binding
module Zbasic2 for efficient and well-controllable two-enzyme co-immobilization on anionic (ReliSorb SP400)
carrier material. We use the C-glycosyltransferase from rice (Oryza sativa; OsCGT) and the sucrose synthase
from soybean (Glycine max; GmSuSy) to synthesize nothofagin, the natural 3’-C-β-d-glucoside of the
dihydrochalcone phloretin, with regeneration of uridine 5’-diphosphate (UDP) glucose from sucrose and UDP.
Exploiting enzyme surface tethering via Zbasic2, we achieve programmable loading of the glycosyltransferases
(~18 mg/g carrier; 60%–70% yield; ~80% effectiveness) in an activity ratio (OsCGT:GmSuSy= ~1.2) optimal
for the overall reaction rate (~0.2 mmolhÀ 1 gÀ 1 catalyst; 30 C, pH 7.5). Using phloretin solubilized at 120 mM
°
as inclusion complex with 2-hydroxypropyl-β-cyclodextrin, we demonstrate complete substrate conversion
into nothofagin (~52 g/L; 21.8 mg product hÀ 1 gÀ 1 catalyst) at 4% mass loading of the catalyst. The UDP-
glucose was recycled 240 times. The solid catalyst showed excellent reusability, retaining ~40% of initial
activity after 15 cycles of phloretin conversion (60 mM) with a catalyst turnover number of ~273 g
nothofagin/g protein used. Our study presents important progress towards applied bio-catalysis with
immobilized glycosyltransferase cascades.
Keywords: Leloir glycosyltransferase; sugar nucleotide regeneration; co-immobilization; cascade bio-catalysis;
C-glycosylation; nothofagin
Introduction
often represents the key central step of synthetic routes
towards
oligosaccharides,[1b,2]
and glycoconjugates,
natural
product
including
Sugar
nucleotide-dependent
(Leloir) glycosides[3]
glycosyltransferases are generally regarded as efficient glycoproteins.[4] These are important product classes
and practically useful catalysts of glycosylation with promising applications in different commercial
reactions.[1] In carbohydrate chemistry, glycosylation sectors, including medical healthcare in particular. The
Adv. Synth. Catal. 2021, 363, 2157–2169
2157
© 2021 The Authors. Advanced Synthesis & Catalysis
published by Wiley-VCH GmbH

FULL PAPER
asc.wiley-vch.de
use of glycosyltransferases in applied bio-catalysis a recyclable catalyst. Thus, the practical use of
must involve an integrated concept of reaction glycosyltransferases in glycoside synthesis could be-
development.[1a] Like other enzymatic reactions that come substantially more operative. There has been
require (co)-substrates (e.g., nicotinamide coenzymes high interest recently in the identification, character-
for oxidoreductions;[5] nucleoside triphosphates for ization
and
engineering
of
Leloir
C-
phosphorylation[6] or nucleotide transfer;[7] S- adenosyl- glycosyltransferases[1a,3d,12] as well as in the use of these
methionine for methylation[8]) which for reasons of enzymes as catalysts of synthetically important C-
atom economy and costs cannot be used stoichiometri- glycosylation reaction.[13] The idea of this study, to
cally, the overall glycosyltransferase reaction typically develop a generic process technology for the applica-
consists of a multistep cascade transformation in which tion of immobilized glycosyltransferase cascades, is
the immediate enzymatic glycosylation is supplied thus strongly supported. Advances made in the (semi)
in situ with the corresponding sugar nucleotide donor continuous synthesis of nothofagin, according to
substrate.[1,9] Besides synthesis from free monosacchar- Scheme 1 and using co-immobilized enzymes, would
ide via phosphorylation and nucleotide transfer seem to be largely transferrable to other C-glycosyl-
cascades,[9a] sugar nucleotides are often prepared transferase-catalyzed glycosylation reactions from
through a second glycosyltransferase reaction, run in UDP-glucose of synthetic importance.[1a,13]
reverse direction and operated in parallel to the
Immobilization is an important method from the
synthetic reaction.[1a,3,9b] A suitably activated glycoside general applied bio-catalysis toolbox for reaction
serves as the donor for in situ sugar nucleotide syn- development at the interface with process engineering.
[14]
thesis. A synthetically important example is the supply
Besides occasional reports scattered across glyco-
of uridine 5’-diphosphate (UDP) glucose (UDP-glu- syltransferase
types (e.g.,
sialyltransferase,[15a]
cose) from sucrose and UDP, catalyzed by sucrose galactosyltransferase,[15b] antibiotic glycosyltrans-
synthase (SuSy).[9b,10] Scheme 1 shows the glycosyl- ferase,[15c] sucrose synthase[15d,e]), enzyme immobili-
transferase cascade reaction of this study, involving the zation for (semi)continuous synthesis with reuse of
3’-C-β-d-glycosylation of the dihydrochalcone phlor- solid catalyst is not well developed for the Leloir
etin from UDP-glucose that is continuously supplied glycosyltransferases. In particular, broadly applicable
and regenerated from UDP and sucrose. Nothofagin technology for glycosyltransferase co-immobilization
production according to Scheme 1 was used in our to establish glycosylation cascades on an insoluble
earlier studies[11] to demonstrate a systematic approach support is lacking. Generally, co-immobilization (the
to process intensification for a glycosyltransferase colocalizing immobilization of multiple enzymes on
cascade reaction, performed in batch conversion and in the same solid carrier) can benefit the efficiency of
homogeneous solution. The current inquiry was set out heterogeneously catalyzed enzymatic cascade reactions
to implement the same glycosyltransferase process on by exploiting the effects of spatial proximity.[14c,16]
the surface of a solid support, with the aim of Glycosyltransferase co-immobilization can be achieved
fundamental improvement in reaction efficiency due to in principle via immobilizing each enzyme according
Scheme 1. Coupled glycosyltransferase reaction for nothofagin synthesis via 3’-C-β-glycosylation of phloretin from UDP-glucose
which in turn is prepared from sucrose and UDP. The Zbasic2 fusions of the rice (Oryza sativa) C-glycosyltransferase (OsCGT) and
the soybean (Glycine max) sucrose synthase (GmSuSy) are used as co-immobilized enzyme preparation on ReliSorb SP400.
Catalytic amounts of UDP are used.
Adv. Synth. Catal. 2021, 363, 2157–2169
2158
© 2021 The Authors. Advanced Synthesis & Catalysis
published by Wiley-VCH GmbH

FULL PAPER
asc.wiley-vch.de
to its own, enzyme-specific strategy. Guisan, Rocha- conversion of phloretin and sucrose into nothofagin.
Martin and coworkers have recently presented an The solid catalyst was easily reusable, retaining ~40%
important example of such “individualized” glycosyl- of initial activity after 15 cycles of phloretin conver-
transferase co-immobilization, applied to a natural sion (60 mM) with a catalyst turnover number of
product glycosyltransferase and a sucrose synthase, as ~273 g nothofagin/g protein used. Our study presents
discussed later in context of the evidence from the important progress towards applied bio-catalysis with
current study.[17] Alternatively, glycosyltransferase co- immobilized glycosyltransferase cascades.
immobilization can be uniform in the strategy used for
tethering the enzymes to the solid carrier surface. It is
this latter strategy, realized in the form of glycosyl- Results and Discussion
transferase fusions with the cationic binding module
Glycosyltransferase Fusions with Zbasic2
Zbasic2, that we have pursued here with the idea of a
facile and efficient enzyme co-immobilization usable N-terminal fusions of OsCGT and GmSuSy with the
for nothofagin synthesis. Zbasic2 module were constructed and the resulting
Zbasic2 is a small protein of 58 amino acids and chimeric proteins obtained by expression in E. coli.
~7 kilodalton size.[18] It is an engineered binding The purified enzymes showed the expected mass for
module, originally developed from the Z-domain of the respective full-length protein fused to Zbasic2 (Fig-
staphylococcal protein A. Placed within fusion pro- ure S1). The specific activity of Z-OsCGT (2.06 U/mg)
teins, the Zbasic2 module can fold autonomously to adopt and Z-GmSuSy (0.77 U/mg) was decreased to, respec-
its stable three α-helical bundle structure.[18] Driven by tively, 67% and 37% of the corresponding N-termi-
electrostatic interactions from multiple arginine resi- nally Strep-tagged enzyme. Comparison of specific
dues exposed on one of its sides,[18] the functional activities in purified enzyme and in cell lysate
Zbasic2 will bind with considerable affinity to negatively (Table S1) suggested that expression of Zbasic2 protein
charged surfaces, such as those of cation exchange was to a level of ~20–30% of total intracellular
resins[19a] and silica materials.[19b] We have demon- protein.
strated fusion to Zbasic2 for enzyme immobilization,[19]
The Z-glycosyltransferases were conveniently re-
and recently also for multiple enzyme co- covered from the cell lysate (Figure S1). Moreover, as
immobilization,[20] with excellent control, and modu- shown later, the Z-glycosyltransferases could be
larity. Important elements of control are enzyme immobilized directly from cell lysate, without require-
orientation on the solid surface, especially for co- ment for further enrichment and purification of the
immobilization, and relative amount of enzyme loaded enzymes. The specific activity of OsCGT and GmSuSy
onto the solid carrier. Moreover, enzymes immobilized was decreased by ~40–60% in the presence of
via Zbasic2 are homogeneously distributed in porous 250 mM NaCl. The effect was distinct for the Z-
carriers,[19a,e] which is important to enable the co- fusions of the glycosyltransferases. The Strep-tagged
localization of multiple enzymes. Cello-oligosacchar- enzymes, by contrast, were unaffected by 250 mM
ide synthesis from sucrose and glucose by three NaCl (Table S1). The addition of 250 mM NaCl was
glycoside phosphorylases shows practical use of Zbasic2
based enzyme co-immobilization applied to a hetero- immobilization[19] have shown that this can enhance
geneously catalyzed cascade transformation.[19a]
the selectivity of the interaction of the Zbasic2 module
-
considered because earlier studies of Z-enzyme
Here, we present an efficient and flexible strategy with the used ReliSorb SP400 carrier.
for the co-immobilization of the rice (Oryza sativa) C-
glycosyltransferase (OsCGT) and the soybean (Glycine
max) sucrose synthase (GmSuSy). Building on surface
Single Enzyme Immobilization
tethering via the Zbasic2 module, this strategy is novel to The Z-glycosyltransferases were immobilized individ-
the class of Leloir glycosyltransferases; and it opens ually on ReliSorb SP400 resin (Figure S2 and Ta-
up new important opportunities for development of a ble S2). ReliSorb SP400 are polymethacrylate particles
glycosyltransferase-based process technology for gly- of spherical shape (75–200 μm diameter; 120 μm mean
coside synthesis. Fusion to the Zbasic2 module was diameter) and involving a network of macro-pores
reasonably tolerated as regards the enzyme specific with a size of ~100 nm. The ReliSorb SP400 material
activity (�40% of the native glycosyltransferase). offers sulfonate surface groups as negatively charged
Single-enzyme immobilization as well as two-enzyme interaction sites for oriented binding of the fusion
co-immobilization were demonstrated on ReliSorb protein via the cationic Zbasic2 module.
SP400 carrier in good yield, 50–100% depending on
Previous studies have shown excellent combination
the enzyme loading. The immobilized enzymes showed of efficiency and selectivity of Z-enzyme immobiliza-
high catalytic effectiveness (�50%). Programmable tion on ReliSorb SP400,[19a,20] encouraging our efforts
loading of the two glycosyltransferases gave an here to directly immobilize the Z-glycosyltransferases
activity ratio (OsCGT:GmSuSy= ~1.2) optimal for from the cell lysate. Results in Figure S2 show
Adv. Synth. Catal. 2021, 363, 2157–2169
2159
© 2021 The Authors. Advanced Synthesis & Catalysis
published by Wiley-VCH GmbH

FULL PAPER
asc.wiley-vch.de
selective adsorption of both Z-OsCGT and Z-GmSuSy expected for the immobilized enzyme based on the
to the ReliSorb SP400 carrier, leading to near-complete activity yield. For each enzyme, the observable activity
depletion of the target protein from cell lysate. increased dependent on the activity loaded, enabling an
Although the immobilized Z-glycosyltransferases were approximate range of 8–15 U/g for the active solid
tethered non-covalently, their binding appeared to be catalyst. The yield and the catalytic effectiveness both
strong: washing the carrier (~60 mg) two times with decreased with increasing enzyme loading, from a near
1 mL of 50 mM HEPES buffer (pH 7.5; 250 mM 100% at the lowest loading to around 50–60% at the
NaCl) did not cause elution of enzyme activity.
highest loading. The catalytic effectiveness of ~100%
As shown in Figure 1, we analyzed the individual implies that no activity was lost as result of the Z-
immobilization of Z-OsCGT and Z-GmSuSy on enzyme attachment to the solid carrier. The finding is
ReliSorb SP400 by monitoring, at different loadings of as expected for a perfectly oriented immobilization of
enzyme activity on the unit mass of carrier, the activity the Z-glycosyltransferase where the chimeric enzyme
yield (% of the offered activity bound to the carrier) is tethered solely by its Zbasic2 module and the function
and the catalytic effectiveness (% activity retention) of of the “catalytic module” is unaffected by the solid
the immobilized enzyme. The catalytic effectiveness is surface (for general discussion, see reference 14c). The
the ratio (×100, %) between the observable activity of decrease in catalytic effectiveness at higher enzyme
the immobilized enzyme (U/g) and the activity loading might arise due to effects of enzyme crowding
or aggregation on the solid surface. Pore clogging by
immobilized enzyme might be another reason. It seems
unlikely that at the enzyme loading used, the carrier
was internally saturated with protein.[19a] In terms of
protein immobilized, up to ~30 mg/g and ~42 mg/g
were bound for Z-OsCGT and Z-GmSuSy, respec-
tively. Differences in protein binding may be due to
the oligomer structure of the enzymes which differs for
Z-OsCGT (monomer[11a]) and Z-GmSuSy (homo-
tetramer[10a]). Immobilization of Z-GmSuSy might thus
benefit from multivalent binding via two or more of its
Zbasic2 modules. Results of co-immobilized enzyme
reuse experiments to be reported later show, however,
that the Z-GmSuSy was bound on the carrier not much
more strongly than the monomeric Z-OsCGT. The
structure of Arabidopsis thaliana SuSy[21] shows that
the N-termini of neighboring protein subunits point to
opposite sides of the enzyme tetramer. Structural
constraints would therefore seem to limit the degree of
multivalent binding in GmSuSy immobilization to
maximally two Zbasic2 modules being involved simulta-
neously in tethering the enzyme tetramer to the solid
surface. We presume at this stage, therefore, that multi-
valency effects in Z-GmSuSy immobilization were
small. The relatively similar behavior of the two Z-
glycosyltransferases in terms of catalytic effectiveness,
and dependence thereof on enzyme loading, is worth
noting.
Glycosyltransferase Co-Immobilization
Recent study suggests that immobilization of O-
glycosyltransferase (from apple) and sucrose synthase
Figure 1. Immobilization of (A) Z-OsCGT and (B) Z-GmSuSy
on ReliSorb SP400. Enzymes were immobilized directly from
(from Acidithiobacillus caldus) to co-localize the two
enzyme activities on the same porous carrier benefits
the synthetic efficiency of the bi-enzymatic glycosyla-
tion cascade.[17] The rate of 5-O-β-glycosylation of
piceid (3-O-β-d-glucoside of resveratrol) was en-
hanced ~2.4-fold compared to reaction of the two
enzymes immobilized individually on separate
the E. coli cell lysates. A 50 mM HEPES buffer (pH 7.5;
250 mM NaCl) was used for enzyme loading and washing. The
carrier loading was 60 mg/ml. The buffer pH was unchanged
during incubation. Immobilized yield (squares, black), observ-
able activity (circles, red) and catalytic effectiveness (triangles,
blue) are shown. Activities of soluble and immobilized enzymes
were determined as described in the Methods.
Adv. Synth. Catal. 2021, 363, 2157–2169
2160
© 2021 The Authors. Advanced Synthesis & Catalysis
published by Wiley-VCH GmbH

FULL PAPER
asc.wiley-vch.de
carriers.[17] Therefore, to perform the 3-C-β-glycosyla-
tion of phloretin efficiently, we considered co-immobi-
lization of OsCGT and GmSuSy to place the enzymes
closely together under the confinement of the solid
surface. This would set particle radius of the carrier as
the upper limit for the maximum diffusional distance
in the reaction. Proximity effects could thus be
exploited to maximize the overall flux through the
two-step cascade reaction on solid support (for general
discussion, see reference 16). We immobilized Z-
OsCGT and Z-GmSuSy in an activity ratio for the
enzyme loaded that varied between 10.8 and 0.1. The
maximum activity loaded was 43.2 U/g for Z-OsCGT
(activity ratio 10.8) and 36 U/g for Z-GmSuSy (activity
ratio 0.1). For each activity ratio used, we measured
the observable activity (rate of nothofagin synthesis) of
the solid catalyst and the product yield from the
coupled reaction. Figure 2A shows that both parame-
ters passed through a maximum at an activity ratio of
1.2 and dropped sharply at lower and higher activity
ratios. The optimum synthesis rate was ~13 U/g.
Analysis with SDS PAGE (Figure 2B) reveals, for each
activity ratio used, the protein distribution for Z-
OsCGT and Z-GmSuSy on the solid carrier. The
semiquantitative picture from the polyacrylamide gel
(Figure 2B) suggests that the activity ratio as loaded
was largely retained in the solid catalyst. Using the
optimum ratio of 1.2, we determined the observable
activity for the individual enzymes (Z-OsCGT: 10.9 U/
g; Z-GmSuSy: 10.4 U/g) in the co-immobilized prepa-
ration. The Z-OsCGT and the Z-GmSuSy were
immobilized in a yield of ~60–70% and show an
effectiveness of ~80%, both in good accordance with
the results of single enzyme immobilization studies.
These findings suggest that in Z-enzyme co-immobili-
zation, the immobilization of one enzyme did not
affect negatively the immobilization of the other.
Generally, this is important as it enables programmable
enzyme co-immobilization based on evidence from
single enzyme immobilization experiments.
Figure 2. Co-immobilization of Z-OsCGT and Z-GmSuSy on
ReliSorb SP400. (A) Observable activity of nothofagin syn-
thesis (squares, black) and nothofagin yield from the coupled
reaction (circles, red) dependent on the activity ratio of Z-
OsCGT and Z-GmSuSy loaded. The co-immobilization was
from E. coli lysates mixed in 50 mM HEPES buffer (pH 7.5;
250 mM NaCl). The activity measurements used the following
conditions in 50 mM HEPES buffer (pH 7.5): 1 mM phloretin,
500 mM sucrose, 0.5 mM UDP, 13 mM MgCl2, 50 mM KCl,
°
20% (by volume) DMSO, 1.3 mg/ml BSA; 30 C and 1000 rpm
agitation rate. (B) SDS polyacrylamide gel showing the proteins
bound on ReliSorb SP400 dependent on the activity ratios
loaded. The protein band corresponding to Z-OsCGT (red
arrow) and Z-GmSuSy (green arrow) is indicated. The activity
ratio (Z-OsCGT/Z-GmSuSy) was 10.8 (a), 2.8 (b), 1.2 (c), 0.5
(d), and 0.1 (e).
Characterization of the Co-Immobilized Glycosyl-
transferases
The pH dependence of activity for soluble and
immobilized enzyme preparations was determined at
°
30 C. As shown in Figure 3A, the activity of the
soluble Z-OsCGT dropped sharply at high pH whereas
the immobilized enzyme retained its maximum activity
of pH 7.5 up to pH 10.5. For Z-GmSuSy, the pH activity at high pH was attenuated compared to the
dependence of activity was relatively narrow, with soluble enzymes. Considering the behavior of the
optimum activity at pH of 6.5. Loss of activity at high single enzyme immobilized preparations (Fig-
pH was less pronounced in the immobilized enzyme ure 3A,B), the decrease in activity at high pH for the
(Figure 3B). Interestingly, an opposite trend was found co-immobilized preparation appears to be dominated
at low pH where immobilized Z-GmSuSy lost activity by the pH dependence of the Z-GmSuSy activity. The
more dramatically than the soluble enzyme. In the co- optimum pH for the co-immobilized enzyme prepara-
immobilized enzyme preparation (Figure 3C), loss of tion was ~7.5.
Adv. Synth. Catal. 2021, 363, 2157–2169
2161
© 2021 The Authors. Advanced Synthesis & Catalysis
published by Wiley-VCH GmbH

FULL PAPER
asc.wiley-vch.de
Figure 3. Effect of pH on the activity of the soluble (full squares, black) and the (co)-immobilized enzyme preparations (full circles,
red). (A) Z-OsCGT, (B) Z-GmSuSy, and (C) both enzymes used together in cascade reaction. The reaction buffers used were
25 mM sodium citrate/phosphate (pH 3.5–6.5), 50 mM HEPES (pH 7.5–8.5), and 50 mM CHES (pH 9.5.–10.5). The pH shown is
the final pH of the suspension containing the carrier particles. Enzyme activities were measured as described in the Methods. The
observable activity at the optimum pH (100%) was 14 U/g for immobilized Z-OsCGT, 13.5 U/g for immobilized Z-GmSuSy and
13 U/g for the co-immobilized enzymes.
The temperature dependence of enzyme stability enzyme as result of the immobilization. An operational
°
°
(30 min incubation) was analyzed at pH 7.5 by temperature range between 30 C and 40 C for the bi-
comparing soluble and immobilized enzyme prepara- enzymatic synthesis of nothofagin was suggested from
tions similarly as done for the pH-activity profiles. the data in Figure 4.
Results in Figure 4 (% activity loss dependent on
temperature) reveal that except for Z-GmSuSy that was
Synthetic Activity and Recyclability of the Co-
Immobilized Glycosyltransferases
°
more stable at 50 C due to immobilization, the soluble
and (co)-immobilized enzymes exhibited comparable
stabilities. The idea of oriented immobilization via the The co-immobilized enzyme preparation (activity ratio
Zbasic2 module is that the characteristics of the enzyme 1.2) was used to examine its synthetic applicability and
module are largely unaffected by tethering of the Z- recyclability. Initially, to provide a proof of concept, a
enzyme to the solid surface. One would, therefore, not relatively low concentration of the phloretin substrate
expect a large stabilization of the activity of the Z- (1 mM) was used. The time course of nothofagin
Figure 4. Thermostability of soluble (dark gray) and immobilized enzyme preparations (black). (A) Z-OsCGT, (B) Z-GmSuSy, and
(C) both enzymes used together in cascade reaction. Enzyme activities were measured as described in the Methods. Enzyme
activities were measured as described in the Methods. The observable activity at the optimum pH (100%) was 14 U/g for
immobilized Z-OsCGT, 13.5 U/g for immobilized Z-GmSuSy and 13 U/g for the co-immobilized enzymes.
Adv. Synth. Catal. 2021, 363, 2157–2169
2162
© 2021 The Authors. Advanced Synthesis & Catalysis
published by Wiley-VCH GmbH

FULL PAPER
asc.wiley-vch.de
production is shown in Figure 5. The used phloretin min) was even higher than in reactions using the non-
was converted fully within ~15 min. The solid catalyst complexed phloretin (~0.1 mM/min; Figure 5B, Fig-
was recovered and the reaction was repeated 15 times. ure S3). A space-time yield of ~5 mM/h (2.2 gLÀ 1 hÀ 1)
Figure 5B shows the nothofagin yield in each reaction was calculated for the reaction run to completion. We
of 15 min. The yield decreased gradually to ~60% in increased the phloretin concentration to 120 mM and
the last reaction. The solid catalyst was analyzed by show its full conversion into nothofagin (Figure 6B).
SDS PAGE at reaction start and after the last reaction The time required for production was longer (60 h),
(Figure 5A). Comparison of the two samples reveals due to a decreased initial conversion rate (~0.1 mM/
clear decrease in protein band intensity for both min) and a pronounced slowing down of the reaction
enzymes after repeated use of the catalyst. Despite the already at low degrees of conversion (~50%). The
semiquantitative nature of the image, the loss in band space-time yield of ~2 mM/h (0.87 gLÀ 1 hÀ 1) was also
intensity (Figure 5A), indicating partial elution of the decreased compared to the 60 mM reaction. Fluid
non-covalently bound enzyme, might well account for viscosity which is increased strongly at high concen-
the observed decrease in synthetic activity (Figure 5C). tration of the phloretin inclusion complex with 2-
hydroxypropyl-β-cyclodextrin[11c] could be a relevant
factor. In a comparison at equivalent volumetric
activity of the enzymes used, the solid catalyst was
only about half as efficient in terms of space-time yield
Nothofagin Synthesis with Re-Use of the Co-Immo-
bilized Glycosyltransferases
Due to the excellent recyclability of the co-immobi- as the soluble enzymes. Viscosity effects are likely to
lized enzyme preparation, we were encouraged to be stronger in the heterogeneously catalyzed reaction
explore strategies of reaction intensification applicable and thus provide a possible explanation.
to the solid catalyst. In particular, phloretin solubility
To examine re-use of the solid catalyst, we
should be enhanced to increase the product output. performed a total of 15 consecutive reactions with
Using 20% DMSO co-solvent, the phloretin can be 60 mM phloretin, whereby each reaction lasted 12 h
dissolved to ~10 mM and we show its full conversion and involved recovery of the immobilized enzyme for
from 5 mM or 10 mM initial concentration in Fig- the next reaction. As shown in Figure 7, the nothofagin
ure S3. To further increase the soluble phloretin yield decreased from a near 100% in the first reaction
concentration, we used inclusion complexation with 2- to ~40% in the last reaction. As in the re-use experi-
hydroxypropyl-β-cyclodextrin, as used by us previ- ment in Figure 5, partial elution of enzyme from the
ously for nothofagin production with the soluble solid catalysts (data not shown) seemed to be respon-
enzymes.[11c] We show in Figure 6A that phloretin sible for the gradual loss of catalyst efficiency over the
solubilized at 60 mM was converted very efficiently number of reactions performed. We can calculate from
(�98%) within ~12 h. The conversion rate (~0.2 mM/ the total nothofagin produced in the 15 reactions
Figure 5. Synthetic activity and recyclability of co-immobilized Z-OsCGT and Z-GmSuSy. (A) SDS polyacrylamide gel showing
the protein on the ReliSorb SP400 carrier at reaction start (Int) and after the last reaction (Fin). (B) Reaction time course. (C)
Repeated reaction as in panel B with reuse of the solid catalyst. Each reaction lasted 15 min. Conditions: 1 mM phloretin, 500 mM
°
sucrose, 0.5 mM UDP, 50 mM KCl, 13 mM MgCl2, 1.3 mg/ml BSA, 20% DMSO, 30 C, 50 mM HEPES buffer (pH 7.5). The
starting activity of immobilized Z-OsCGT and Z-GmSuSy was 0.23 U/ml and 0.21 U/ml, respectively. The solid catalyst was used
at 20 mg/ml. The loaded activity ratio was 1.2.
Adv. Synth. Catal. 2021, 363, 2157–2169
2163
© 2021 The Authors. Advanced Synthesis & Catalysis
published by Wiley-VCH GmbH

FULL PAPER
asc.wiley-vch.de
Figure 7. Re-use of the co-immobilized enzyme preparation for
nothofagin production. Reaction conditions: 60 mM phloretin as
inclusion complex with 2-hydroxypropyl-β-cyclodextrin,
500 mM sucrose, 0.5 mM UDP, 50 mM KCl, 13 mM MgCl2,
°
1.3 mg/ml BSA, 50 mM HEPES buffer (pH 7.5) and 30 C. The
starting activity of immobilized Z-OsCGT and Z-GmSuSy was
0.47 U/ml and 0.42 U/ml, respectively. The solid catalyst was
used at 40 mg/ml. The loaded activity ratio was 1.2. Each
reaction cycle lasted 12 h.
strated at 100 g scale in an earlier study from this
laboratory.[11c] The isolated yield was �65% and the
purity was �95%. Considering the product down-
stream processing to have been established in prior
research, we did not pursue isolation of nothofagin in
the current study.
Figure 6. Nothofagin production (circles, red) from phloretin
(squares, black) solubilized as inclusion complex with 2-
hydroxypropyl-β-cyclodextrin (A, 60 mM; B, 120 mM) using
co-immobilized Z-OsCGT and Z-GmSuSy. Reaction conditions:
500 mM sucrose, 0.5 mM UDP, 50 mM KCl, 13 mM MgCl2,
°
1.3 mg/ml BSA, 50 mM HEPES buffer (pH 7.5) and 30 C. The
The Advantage of Enzyme Co-Immobilization for
Efficient Glycosylation
starting activity of immobilized Z- OsCGT and Z-GmSuSy was
0.47 U/ml and 0.42 U/ml, respectively. The solid catalyst was
used at 40 mg/ml. The loaded activity ratio was 1.2.
Fusion to Zbasic2 can introduce modular design to
enzyme immobilization.[14c,19a,e] Unlike unstructured
peptide tags, Zbasic2 represents a structurally independ-
(278 mg, 15 ml) and the amount of immobilized Z- ent and functionally distinct unit (module) of the fusion
OsCGT present on the solid catalyst (0.30 mg, protein. It can fold separately and provides a program-
recycled) that the C-glycosyltransferase reached a mable function according to a uniform mechanistic
mass-based turnover number of 927 (= 278/0.30) g/g. principle of surface binding.[18] Due to its independent
With a Mr of 436 for nothofagin and 57,800 for Z- function in the modularly organized fusion protein,
OsCGT, the corresponding mol-based turnover number Zbasic2 hardly interferes with the activity of its partner
is 1.2×105 mol/mol. Considering the Z-GmSuSy module,
as
shown
for
different
enzymes
(0.72 mg, recycled) that is additionally required, the previously[19,20,22] and here for the first time for
mass-based turnover number becomes 273 (=278/ representatives of the Leloir glycosyltransferase class.
1.02). These are excellent turnover numbers for a Thus, immobilization via Zbasic2 is advantageous in
Leloir glycosyltransferase[1a,11a] and highlight the effi- particular when multiple enzymes in cascade bio-
ciency of the synthetic reaction when the solid catalyst catalysis should be tethered as a protein ensemble on a
is reused. It may be noted that nothofagin recovery solid surface, to establish multistep reaction sequences
from an enzymatic reaction mixture comparable to the on a single insoluble carrier. Proof of principle was
ones obtained in Figure 6 and Figure 7 was demon- shown from the co-immobilization of the Z-fusions of
Adv. Synth. Catal. 2021, 363, 2157–2169
2164
© 2021 The Authors. Advanced Synthesis & Catalysis
published by Wiley-VCH GmbH

FULL PAPER
asc.wiley-vch.de
three glycoside phosphorylases and the application of
the solid catalyst thus obtained for cello-oligosacchar-
ide synthesis.[20a] However, decision about enzyme co-
immobilization on the same carrier, as opposed to
single enzyme immobilization on different carriers,
must balance gain in catalyst efficiency with increased
complexity of catalyst preparation when co-immobi-
lized enzymes are used.[16] We therefore set out experi-
ments to compare co-immobilized preparation of Z-
OsCGT and Z-GmSuSy with a mixture of the individ-
ually immobilized enzymes for nothofagin synthesis
under otherwise exactly comparable conditions. Time
courses of conversion of 60 mM phloretin inclusion
complex with 2-hydroxypropyl-β-cyclodextrin are
shown in Figure 8. In both reactions, the phloretin was
converted fully. In terms of initial reaction rate,
reaction of the co-immobilized enzyme (34 mM/h) was
~2.5-fold faster than reaction of the individually
immobilized enzymes. The space-time yield for full
conversion (�98%) was ~4-fold higher for the co-
immobilized enzyme than for individually immobilized
Z-OsCGT and Z-GmSuSy. These results quantify the
intensification of nothofagin production due to prox-
imity effects of enzyme co-immobilization. The rate
enhancement for the co-immobilized enzyme prepara-
tion can probably be ascribed to shortened diffusion
paths when enzymes are co-localized on the porous
surface of the solid carrier. Our results are consistent
with study of Rocha-Martin and colleagues who
showed 2.4-fold benefit of glycosyltransferase co-
immobilization, compared to individual enzyme immo-
bilization on separate carrier particles, on piceid Figure 8. Synthesis of nothofagin using co-immobilized (A) or
glycosylation from sucrose-derived UDP-glucose.[17]
The effect was dedicated to an enhanced efficiency of
the UDP/UDP-glucose shuttle when enzymes are co-
localized on the solid carrier.
individually immobilized (B) enzyme preparations. Phloretin
(squares, black) and nothofagin (circles, red) are shown.
Reaction conditions: 60 mM phloretin as inclusion complex
with 2-hydroxypropyl-β-cyclodextrin, 500 mM sucrose, 0.5 mM
UDP, 50 mM KCl, 13 mM MgCl2, 1.3 mg/ml BSA, 50 mM
°
HEPES buffer (pH 7.5) and 30 C. The starting activity of
Co-Immobilization of Z-OsCGT and Z-GmSuSy in
Context
immobilized Z-OsCGT and Z-GmSuSy was 1.17 U/ml and
1.05 U/ml, respectively. The solid catalyst was used at 100 mg/
ml. The loaded activity ratio was 1.2.
Leloir glycosyltransferases are widely recognized for
their synthetic importance.[1–4] They are considered
“precision catalysts” of glycosylation with high poten-
tial for applied bio-catalysis. There have been limited enzyme for up to 10 cycles was shown. Important
efforts, however, to develop immobilized glycosyl- advance made with the Zbasic2 fusion approach reported
transferases for heterogeneous catalysis applications here is fourfold. First, the immobilization no longer
(for reviews, see references 1a, 2a). Recently, Guisan, requires purified enzyme and is conveniently done
Rocha-Martin and their co-workers have shown from the cell lysate. Second, it builds on modular
immobilization of bacterial sucrose synthases on design. A proven engineering principle of enzyme
suitably activated agarose carriers.[15d] They also tethering to the solid surface is flexibly applied to the
demonstrated glycosyltransferase co-immobilization on (co)-immobilization of different glycosyltransferases.
agarose,[17] as already discussed. Their studies showed Efficiency parameters of the (co)-immobilization are
glycosyltransferase immobilization in good loading good to excellent when applying standard immobiliza-
(several mg protein/g carrier), high yield (�50%) and tion conditions without enzyme-specific optimization.
typically good effectiveness (�44%). Stabilization of Third, a standard (anionic) carrier material can be used
tetrameric sucrose synthase required post-immobiliza- for immobilization without requirement for chemical
tion crosslinking.[15d] Re-use of the immobilized derivatization/activation of the solid surface. Fourth,
Adv. Synth. Catal. 2021, 363, 2157–2169
2165
© 2021 The Authors. Advanced Synthesis & Catalysis
published by Wiley-VCH GmbH

FULL PAPER
asc.wiley-vch.de
materials used are described in the Supporting Information.
Authentic standard of nothofagin (phloretin 3’-C-β-D-gluco-
side; �99%) was obtained from recent study using enzymatic
synthesis.[11c]
enzyme co-localization on the solid carrier does not
require specific adjustment and finely tuned control of
the conditions for single-enzyme immobilization. Im-
portantly, the co-immobilization study of Trobo-
Maseda et al.[17] shows that unless proper control for
enzyme co-localization is applied to the immobiliza-
tion of the single enzymes, the effect of enzyme co-
immobilization can be lost completely (~50-fold).
The lifetime of our co-immobilized glycosyltrans-
ferase catalyst is sufficient for multiple cycles of
reaction. However, further stabilization of Z-enzyme
binding to the solid surface will be important to
enhance the turnover number of this catalyst. Post-
immobilization techniques of protein crosslinking, as
applied to sucrose synthase in the study of Trobo-
Z-Enzymes
OsCGT (GenBank id: FM179712) and GmSuSy (GenBank id:
AF030231) were used. Enzyme chimeras harboring the Zbasic2
module fused to the N-terminus were constructed from the
plasmid vector pT7_Z_P450 BM3[20c] using overlap extension
PCR. The oligonucleotide primers used (Table S1) and the PCR
conditions applied are found in the Supporting Information. The
used expression vectors are referred to pT7_Z_OsCGT and
pT7_Z_GmSuSy. The relevant protein sequences and the
methods used for protein expression in E. coli BL21(DE3) and
Maseda et al.,[15d] may be useful. However, the enzyme purification by cation exchange chromatography are detailed in
the Supporting Information. Protein purification was monitored
by SDS PAGE and measurement of specific enzyme activity.
Enzymes harboring N-terminal Strep-tag II have previously
been reported[11a] and are used as reference. Protein concen-
trations were measured with ROTI Quant assay (Carl Roth,
Karlsruhe, Germany) referenced against BSA. Z-OsCGT (~
15 mg/ml) and Z-GmSuSy (~20 mg/ml) were stored in 50 mM
turnover number from the current work (~273 g/g)
shows Leloir glycosyltransferases to be already on par
with other enzymes (e.g., glycoside hydrolases,[23]
glycoside phosphorylases[20a,24] perceived as more
robust for application in glycosylation.
°
HEPES buffer (pH 7.5) at À 20 C. Enzyme preparations
Conclusion
retained full activity for at least 2 weeks.
Modular bioengineering is a powerful approach for
accelerated development in biotechnology.[25] Fusion to
the Zbasic2 module is shown here for glycosyltransferase
co-immobilization to implement sugar nucleotide-
dependent glycosylation cascade reactions on solid
support. The example of Z-OsCGT and Z-GmSuSy
demonstrates convenient production of the Z-glycosyl-
transferases in E. coli and programmable enzyme co-
immobilization in a facile, high-yielding procedure
directly from the cell lysates. The solid catalyst with
Z-OsCGT and Z-GmSuSy loaded in optimum ratio
shows excellent overall activity for nothofagin syn-
thesis from sucrose and phloretin, and it can be
recycled conveniently from the reaction. Output
parameters, such as yield, product concentration,
space-time yield, and recycling of UDP as catalytic
reagent, are similar for batch syntheses by soluble and
co-immobilized enzymes. Recycling of the co-immobi-
lized enzymes enhances the enzyme turnover number
by another �10-fold. Collectively, this study empha-
sizes enzyme co-immobilization as an important
element of a comprehensive process intensification
strategy for glycosylation reactions catalyzed by
coupled Leloir glycosyltransferases.
The cell lysates used for immobilization were prepared by ultra-
sonication of E. coli cells suspended in a 2-fold volume of 50
mM HEPES buffer (pH 7.5). A Sonic Dismembrator Model 505
(Fisher Scientific, Vienna, Austria) was used and the sonication
protocol (6 min in total) involved alternating 3 s pulse on and
9 s pulse off at 60% amplitude and cooling on ice. Cell lysate
°
was recovered by centrifugation (15,000 rpm, 4 C, 40 min;
Centrifuge 5424R, Eppendorf, Vienna, Austria).
Enzyme Assays
°
Reactions were performed in 1 ml total volume at 30 C in 2 ml
plastic tubes using agitation at 400 rpm (soluble enzymes) or
1000 rpm (immobilized enzymes) in an Eppendorf Thermo-
Mixer C instrument (Vienna, Austria). Samples (20 μl) were
withdrawn after 5, 10 and 20 min and reactions quenched with
the same volume of acetonitrile. Solid material (precipitated
protein, immobilization carrier) was centrifuged off
°
(13200 rpm, 4 C, 20 min; Centrifuge 5424R, Eppendorf) and
the supernatant analyzed by ion-pairing reversed-phase
HPLC.[26] A Shimadzu model UFLC HPLC system equipped
with a Kinetex® 5 μm EVO C18 LC column (100 Å, 150×
4.6 mm; Phenomenex, Merck, Vienna, Austria) was used. The
°
column was equilibrated at 25 C in 20 mM potassium
phosphate buffer (pH 5.9) containing 40 mM tetrabutylammo-
nium bromide. Elution was with acetonitrile. UV detection was
used.
Experimental Section
Z-OsCGT. The reaction mixture contained in 50 mM HEPES
buffer (pH 7.5), 1 mM phloretin, 2 mM UDP-glucose, 13 mM
MgCl2, 50 mM KCl, 4% DMSO, and 1.3 mg/ml BSA. NaCl
(250 mM) was added optionally. HPLC analysis used an
acetonitrile gradient (20%–50%) at 1 ml/min and detection at
282 nm. Rates were determined from [nothofagin]/time. One
Materials
ReliSorb SP400 carrier was from Resindion (Binasco, Italy). 2-
Hydroxypropyl-β-cyclodextrin (>98%), phloretin (>98%),
UDP (97%) and UDP-glucose (>98%) were from Carbosynth
(Compton, UK). Other chemicals were of reagent grade. Further
Adv. Synth. Catal. 2021, 363, 2157–2169
2166
© 2021 The Authors. Advanced Synthesis & Catalysis
published by Wiley-VCH GmbH

FULL PAPER
asc.wiley-vch.de
unit (U) of activity is the enzyme amount releasing 1 μmol
nothofagin/min under the conditions used.
Immobilization yield. This is expressed as ΔA/A0 (×100, %).
Observable activity. This is the activity of the solid catalyst
Z-GmSuSy. The reaction mixture contained in 50 mM BisTris
buffer (pH 6.5), 500 mM sucrose, 2 mM UDP, 13 mM MgCl2,
50 mM KCl and 1.3 mg/ml BSA. NaCl (250 mM) was added
optionally. HPLC analysis used isocratic acetonitrile conditions
(12.5%) at 2 ml/min and detection at 262 nm. Rates were
determined from [UDP-glucose]/time. One U is the enzyme
amount releasing 1 μmol UDP-glucose/min under the conditions
used.
directly measured in an activity assay. It is expressed as U/g.
Catalytic effectiveness. This is the ratio of observable
activity and bound activity (×100, %). Immobilizations were
performed in triplicates and the mean value with standard error
is reported.
Immobilized Enzyme Characterization
Coupled reaction. The reaction mixture contained in 50 mM
HEPES buffer (pH 7.5), 1 mM phloretin, 500 mM sucrose,
0.5 mM UDP, 13 mM MgCl2, 50 mM KCl, 20% DMSO and
1.3 mg/ml BSA. HPLC analysis of the Z-OsCGT assay was
used. Rates were determined from [nothofagin]/time. One U is
the enzyme amount releasing 1 μmol nothofagin/min under the
conditions used.
The pH-activity profile of soluble and (co)-immobilized
°
enzymes was recorded at 30 C in the pH range 3.5–10.5. The
buffers used were 25 mM sodium citrate/phosphate (pH 3.5–
6.5), 50 mM HEPES (pH 7.5–8.5), and 50 mM CHES (pH 9.5–
10.5). The other reaction conditions and the analytical
procedures used were the ones of the enzyme assays described
above. The pH values given are from the suspension of carrier
°
particles. Enzyme stability at different temperatures in 10 C
Immobilization
°
°
interval between 20 C and 60 C were recorded in 50 mM
HEPES buffer (pH 7.5) and using incubation for 30 min.
ReliSorb SP400 carrier (60 mg dry material) was weighed into
2 ml Eppendorf tubes, washed three times with water and
afterwards two times with HEPES buffer (50 mM, 250 mM
NaCl, pH 7.5). The pH of the carrier suspension was verified as
7.5. For single enzyme immobilization, E. coli cell lysate (1 ml;
2–15 mg total protein/ml) was added and incubated at room
temperature at 40 rpm on an end-over-end rotator for 2 h.
Supernatant was removed and carrier washed twice with 50 mM
HEPES (pH 7.5; Z-OsCGT) or 50 mM Bistris (pH 6.5; Z-
GmSuSy) buffer, each containing 250 mM NaCl. Supernatant
and washing solutions were collected for protein and activity
measurements. The pH was always checked and was constant at
~7.5. The carrier was also recovered and used for activity
measurement. For enzyme co-immobilization, E. coli cell
lysates containing Z-OsCGT and Z-GmSuSy were mixed in
variable volume ratio (10.8–0.1) to give 1 ml of total lysate to
be added to 60 mg dry carrier. The mixture was incubated
(50 mM HEPES buffer, pH 7.5; 250 mM NaCl) and processed
as described above. The individual activities of Z-OsCGT and
Z-GmSuSy as well as the combined (coupled) activity of the
two enzymes were measured using the assays already stated.
Unless mentioned, in the co-immobilization experiment activity
refers to the coupled reaction.
°
Residual activity was then recorded at 30 C using the standard
activity assays. In pH and temperature studies, purified
preparations of the soluble enzymes were used. The soluble
enzyme concentration used was typically ~0.1 mg/ml, equiv-
alent to ~0.2 U/ml. The volumetric concentration of the
immobilized enzyme, as prepared from loading cell lysate on
ReliSorb SP400, was ~0.25 U/ml. The co-immobilized enzyme
preparation involved a ~1.2 activity ratio of Z-OsCGT and Z-
GmSuSy loaded on the carrier.
Enzymatic Synthesis of Nothofagin
Formation of the phloretin inclusion complex. 2-Hydrox-
ypropyl-β-cyclodextrin (4.2 g; 3 mmol) was dissolved to
~300 mM in ~4 ml of deionized water in a 50 ml tube.
Microwave heating at 750 W was used. The total heating time
was ~20 s and several steps of heating and mixing by inversion
were used. Phloretin (0.66 g; 2.4 mmol; ~240 mM) was added
and dissolved as before. Not all phloretin could be dissolved
with the procedure used. The final volume was set to 10 ml.
The mixture was equilibrated in a drying chamber at 70 C for
1 h and inverted every 15 min. If required, the solution was
stored overnight at 4 C and microwave-heated shortly prior to
use. Insoluble phloretin was centrifuged off (5000 rpm, 5 min,
room temperature). In each conversion experiment, the actual
phloretin concentration in solution was determined by HPLC.
°
°
Parameters used to evaluate the immobilization were the
following.
Loaded activity. This is the enzyme activity loaded on the
carrier and is expressed as U/g. Cell lysates used for
immobilization had a known volumetric activity (U/ml) and
protein concentration (mg/ml). The activity loaded was calcu-
lated from the protein loaded in the experiment.
Enzymatic conversion. Reactions were performed using non-
complexed phloretin at low concentrations (1.0–10 mM; 20%
DMSO cosolvent) or phloretin inclusion complex (60 mM,
120 mM; no cosolvent). For 1 mM phloretin (20% DMSO
cosolvent), the reaction was performed in 50 mM HEPES buffer
(pH 7.5) containing 20 mg/ml solid catalyst, 500 mM sucrose,
0.5 mM UDP, 50 mM KCl, 13 mM MgCl2 and 1.3 mg/ml BSA.
A total liquid volume of 500 μl was used. For 60 mM phloretin
inclusion complex, the reaction was performed in 50 mM
HEPES buffer (pH 7.5) containing 40 mg/ml solid catalyst,
500 mM sucrose, 0.5 mM UDP, 50 mM KCl, 13 mM MgCl2
and 1.3 mg/ml BSA. A total liquid volume of 1 ml was used.
Bound activity. This is the activity bound to the solid carrier. It
was calculated as difference in the soluble enzyme activity (U)
before (A0) and after the immobilization (A). The activity A is
the total activity present in the supernatant of the immobiliza-
tion and the washing solutions. The bound activity is calculated
as ΔA (=A0À A) divided by the carrier mass (g). The bound
activity is expressed as U/g.
Adv. Synth. Catal. 2021, 363, 2157–2169
2167
© 2021 The Authors. Advanced Synthesis & Catalysis
published by Wiley-VCH GmbH

FULL PAPER
asc.wiley-vch.de
Unless mentioned, a 1.2 activity ratio of Z-OsCGT and Z-
GmSuSy loaded on the carrier was used. Especially note, the
term of solid catalyst refers to the enzymes co-immobilized on
Relisorb SP400 throughout the manuscript. The volumetric
activities varied as indicated in the text. The temperature was
2015, 44, 8350—8374; c) W. Schwab, T. C. Fischer, A.
Giri, M. Wüst, Appl. Microbiol. Biotechnol. 2015, 99,
165–174; d) G. Tegl, B. Nidetzky, Biochem. Soc. Trans.
2020, 48, 1583–1598.
[4] a) C. Li, L. X. Wang, Chem. Rev. 2018, 118, 8359–8413;
b) P. Bojarová, R. R. Rosencrantz, L. Elling, V. Křen,
Chem. Soc. Rev. 2013, 42, 4774–4797.
°
30 C and agitation was at 1000 rpm in an Eppendorf
ThermoMixer C. Samples (30 μl) were taken at certain times
and processed as described above for the assay of the coupled
reaction. All reactions were performed in triplicates and the
mean value with standard error is reported.
[5] a) Y. Wang, H. Ren, H. Zhao, Crit. Rev. Biochem. Mol.
2018, 53, 115–129; b) J. J. Dong, E. Fernández-Fueyo, F.
Hollmann, C. E. Paul, S. Younes, W. Zhang, Angew.
Chem. Int. Ed. 2018, 57, 9238–9261; Angew. Chem.
2018, 130, 9380–9404; c) S. Wu, R. Snajdrova, J. C.
Moore, K. Baldenius, U. T. Bornscheuer, Angew. Chem.
Int. Ed. 2020, 59, 2–34; d) R. Yuryev, S. Strompen, A.
Liese, Beilstein J. Org. Chem. 2011, 7, 1449–1467.
[6] a) S. Mordhorst, J. Singh, M. K. F. Mohr, R. Hinkel-
mann, H. J. Jessen, J. N. Andexer, ChemBioChem 2019,
20, 1019–1022; b) J. N. Andexer, M. Richter, ChemBio-
Chem 2015, 16, 380–386.
Re-use of co-immobilized enzymes. At the end of the
enzymatic reaction, the solid carrier was centrifuged off
°
(4000 rpm, 20 min, 4 C; Centrifuge 5424R, Eppendorf) and the
supernatant withdrawn for analysis. The carrier washed twice
with 50 mM HEPES buffer (pH 7.5) and then resuspended in
500 μl (1 mM) or 1 ml (60 mM) fresh reaction mixture for a
consecutive round of conversion. Incubation was done for a set
time (15 min, 12 h) depending on the phloretin concentration
(1 mM, 60 mM) used. The carrier re-use was done 15 times.
[7] S. C. Timmons, R. H. Mosher, S. A. Knowles, D. L.
Jakeman, Org. Lett. 2007, 9, 857–860.
Enzyme turnover determination. We define the turnover
number as the mass nothofagin produced in 15 reaction cycles/
mass enzyme used. To determine the total enzyme mass applied
on solid carrier in the reaction, we used the observable activity
of each Z-OsCGT and Z-GmSuSy (U/g carrier) as the basis.
The observable activity was corrected for the known catalytic
effectiveness of the immobilized enzyme (Z-OsCGT: 77%; Z-
GmSuSy: 76%), yielding the bound activity of the immobilized
enzyme if it were fully active. This activity was divided by the
specific activity of the purified enzyme, thus giving the mg of
each Z-OsCGT and Z-GmSuSy used on carrier in the reaction.
The mol-based turnover number was calculated with the Mr of
nothofagin (436) and Z-OsCGT (57800).
[8] C. Liao, F. P. Seebeck, Nat. Catal. 2019, 2, 696–701.
[9] a) W. Li, J. B. McArthur, X. Chen, Carbohydr. Res.
2019, 472, 86–97; b) B. Sauerzapfe, L. Elling, In: Multi-
step Enzyme Catalysis: Biotransformations and Chemo-
enzymatic Synthesis, 1st ed., Eds.: E. Garcia-Junceda,
Wiley-VCH, Weinheim, 2008, pp. 83–102.
[10] a) K. Schmölzer, A. Gutmann, M. Diricks, T. Desmet, B.
Nidetzky, Biotechnol. Adv. 2016, 34, 88–111; b) A.
Zervosen, L. Elling, In: Carbohydrate Biotechnology
Protocols: Methods in Biotechnology, 1st ed., Vol. 10,
Eds: C. Bucke, Springer, 1999, pp. 235–254.
[11] a) L. Bungaruang, A. Gutmann, B. Nidetzky, Adv. Synth.
Catal. 2013, 355, 2757–2763; b) L. Bungaruang, A.
Gutmann, B. Nidetzky, Adv. Synth. Catal. 2016,
358,486–493; c) K. Schmölzer, M. Lemmerer, B. Ni-
detzky, Biotechnol. Bioeng. 2018, 115, 545–556.
[12] a) N. Putkaradze, D. Teze, F. Fredslund, D. H. Welner,
Nat. Prod. Rep. 2021, doi.org/10.1039/D0NP00040 J;
b) J. B. He, P. Zhao, Z. M. Hu, S. Liu, X. Qiao, M. Ye,
Angew. Chem. Int. Ed. 2019, 58, 11513–11520; c) M.
Zhang, F. D. Li, K. Li, X. Qiao, M. Ye, J. Am. Chem.
Soc. 2020, 142, 3506–3512; d) F. Mendoza, G. A. Jaña,
J. Chem. Inf. Model. 2020, 60, 933–943; e) Z. L. Wang,
H. M. Wang, S. Zhang, X. Qiao, M. Ye, Proc. Mont.
Acad. Sci. 2020, 117, 30816–30823.
[13] a) D. Chen, R. Chen, R. Wang, X. Yu, J. Dai, Angew.
Chem. Int. Ed. 2015, 127, 12869–12873; b) D. Chen, S.
Fan, R. Chen, Z. Yang, J. Dai, ACS Catal. 2018, 8,
4917–4927; c) D. Chen, R. Chen, K. Xie, F. Ye, J. Dai,
Org. Lett. 2018, 20, 1634–1637; d) K. Xie, X. Zhang, S.
Sui, F. Ye, J. Dai, Nat. Commun. 2020, 11, 1–12.
[14] a) U. Hanefeld, L. Gardossi, E. Magner, Chem. Soc. Rev.
2009, 38, 453–468; b) R. C. Rodrigues, C. Ortiz, Á.
Berenguer-Murcia, R. Torres, R. Fernández-Lafuente,
Chem. Soc. Rev. 2013, 42, 6290–6307; c) J. Boudrant,
J. M. Woodley, R. Fernandez-Lafuente, Process Bio-
chem. 2020, 90, 66–80; d) C. Mateo, J. M. Palomo, G.
Acknowledgements
Martin Pfeiffer, Peter Wied and Chao Zhong (Institute of
Biotechnology and Biochemical Engineering, Graz University
of Technology) are acknowledged for their support. H.L. is
recipient of fellowship from the China Scholarship Council
(grant number 201906380025).
References
[1] a) B. Nidetzky, A. Gutmann, C. Zhong, ACS Catal.
2018, 8, 6283–6300; b) L. Krasnova, C. H. Wong, J. Am.
Chem. Soc. 2019, 141, 3735–3754; c) L. Mestrom, M.
Przypis, D. Kowalczykiewicz, U. Hanefeld, P. L. Hage-
doorn, Int. J. Mol. Sci. 2019, 20, 5263.
[2] a) S. Schelch, C. Zhong, B. Petschacher, B. Nidetzky,
Biotechnol. Adv. 2020, 44, 107613; b) M. M. Nielsen,
C. M. Pedersen, Chem. Rev. 2018, 118, 8285–8358; c) L.
Wen, G. Edmunds, C. Gibbons, Y. Kong, P. G. Wang,
Chem. Rev. 2018, 118, 8151–8187; d) H. Yu, X. Chen,
Org. Biomol. Chem. 2016, 14, 2809–2818.
[3] a) R. W. Gantt, P. Peltier-Pain, J. S. Thorson, Nat. Prod.
Rep. 2011, 28, 1811–1853; b) D. M. Liang, J. H. Liu, H.
Wu, B. B. Wang, H. J. Zhu, J. J. Qiao, Chem. Soc. Rev.
Adv. Synth. Catal. 2021, 363, 2157–2169
2168
© 2021 The Authors. Advanced Synthesis & Catalysis
published by Wiley-VCH GmbH

FULL PAPER
asc.wiley-vch.de
Fernandez-Lorente, J. M. Guisan, R. Fernandez-La- [21] Y. Zheng, S. Anderson, Y. Zhang, R. M. Garavito, J.
fuente, Enzyme Microb. Technol. 2007, 40, 1451–1463; Biol. Chem. 2011, 286, 36108–36118.
e) J. M. Bolivar, B. Nidetzky, Bioch. Biophys. Acta - [22] P. García-García, J. Rocha-Martin, J. M. Guisan, G.
Proteins Proteom. 2020, 1868, 140333; f) J. M. Bolivar, Fernandez-Lorente, Catalysts 2020, 10, 810.
I. Eisl, B. Nidetzky, Catal. Today 2016, 259, 66–80; [23] a) E. M. H. F. Bongers, J. M. Opitz, A. Fryer, P. Sarda,
g) C. Garcia-Galan, Á. Berenguer-Murcia, R. Fernandez-
Lafuente, R. C. Rodrigues, Adv. Synth. Catal. 2011, 353,
2885–2904; h) O. Barbosa, C. Ortiz, Á. Berenguer-
Murcia, R. Torres, R. C. Rodrigues, R. Fernandez-
Lafuente, Biotechnol. Adv. 2015, 33, 435–456.
B. C. J. Hamel, N. V. A. M. Knoers, Am. J. Med. Genet.
2001, 102, 115–124; b) B. Zeuner, C. Jers, J. D.
Mikkelsen, A. S. Meyer, J. Agric. Food Chem. 2014, 62,
9615–9631; c) K. Schmölzer, M. Weingarten, K. Balde-
nius, B. Nidetzky, ACS Catal. 2019, 9, 5503–5514; d) L.
Ruzic, J. M. Bolivar, B. Nidetzky, Biotechnol. Bioeng.
2020, 117, 1597–1602; e) S. W. Jang, C. H. Cho, Y. S.
Jung, H. S. Shin, D. H. Seo, PLoS One 2018, 13,
e0207466; f) Y. Malbert, C. Moulis, I. André, M.
Remaud-Simeon, Sci. Rep. 2018, 8, 15153; g) M.
Benkoulouche, R. Fauré, M. Remaud-Siméon, C. Moulis,
I. André, Interface Focus 2019, 9, 20180069; h) K.
Slámová, J. Kapešová, K. Valentová, Int. J. Mol. Sci.
2018, 19, 1–19; i) I. Bassanini, J. Kapešová, L.
Petrásková, H. Pelantová, K. Markošová, M. Rebroš, K.
Valentová, M. Kotik, K. Káňová, P. Bojarová, J. Cvačka,
L. Turková, E. E. Ferrandi, I. Bayout, S. Riva, V. Křen,
Adv. Synth. Catal. 2019, 361, 2627–2637.
[15] a) H. Kajiwara, M. Tsunashima, T. Mine, Y. Takakura, T.
Yamamoto, J. Biosci. Bioeng. 2016, 121, 390–393; b) R.
Schneider, M. Hammel, E. G. Berger, O. Ghisalba, J.
Nuesch, D. Gygax, Glycoconjugate J. 1990, 7, 589–600;
c) Z. Chen, J. Zhang, S. Singh, P. Peltier-Pain, J. S.
Thorson, B. J. Hinds, ACS Nano. 2014, 8, 8104–8112;
d) L. Trobo-Maseda, A. H. Orrego, S. Moreno-Pérez, G.
Fernández-Lorente, J. M. Guisan, J. Rocha-Martin, Appl.
Microbiol. Biotechnol. 2018, 102, 773–787; e) A. H.
Orrego, L. Trobo-Maseda, J. Rocha-Martin, J. M. Gui-
san, Enzyme Microb. Technol. 2017, 105, 51–58.
[16] S. Arana-Peña, D. Carballares, R. Morellon-Sterlling,
R. C. Rodrigues, R. Fernandez-Lafuente, Biotechnol.
Adv. 2020, doi.org/10.1016/j.biotechadv.2020.107584.
[17] L. Trobo-Maseda, A. H. Orrego, J. M. Guisan, J. Rocha-
Martin, Int. J. Biol. Macromol. 2020, 157, 510–521.
[18] a) T. Gräslund, M. Hedhammar, M. Uhlén, P. Å. Nygren,
S. Hober, J. Biotechnol. 2002, 96, 93–102; b) M.
Hedhammar, S. Hober, J. Chromatogr. A 2007, 1161, 22–
28.
[24] a) T. Desmet, W. Soetaert, P. Bojarová, V. Křen, L.
Dijkhuizen, V. Eastwick-Field, A. Schiller, Chem. Eur. J.
2012, 18, 10786–10801; b) J. Franceus, T. Desmet, Int.
J. Mol. Sci. 2020, 21, 2526; c) K. De Winter, K.
Verlinden, V. Křen, W. Soetaert, T. Desmet, Green Chem.
2013, 15, 1949–1955; d) K. De Winter, G. Dewitte,
M. E. Dirks-Hofmeister, S. De Laet, H. Pelantová, V.
Křen, T. Desmet, J. Agric. Food Chem. 2015, 63, 10131–
10139; e) M. Kraus, C. Grimm, J. Seibel, ChemBioChem
2016, 17, 33–36; f) M. E. Dirks-Hofmeister, T. Ver-
haeghe, K. De Winter, T. Desmet, Angew. Chem. 2015,
127, 9421–9424; Angew. Chem. Int. Ed. 2015, 54, 9289–
9292.
[19] a) J. Wiesbauer, J. M. Bolivar, M. Mueller, M. Schiller,
B. Nidetzky, ChemCatChem 2011, 3, 1299–1303;
b) J. M. Bolivar, B. Nidetzky, Langmuir 2012, 28,
10040–10049; c) J. M. Bolivar, B. Nidetzky, Biotechnol.
Bioeng. 2012, 109, 1490–1498; d) J. M. Bolivar, S.
Schelch, T. Mayr, B. Nidetzky, ACS Catal. 2015, 5,
5984–5993; e) D. Valikhani, J. M. Bolivar, M. Pfeiffer, [25] a) S. Garcia, C. T. Trinh, Biotechnol. Adv. 2019, 37,
B. Nidetzky, ChemCatChem 2017, 9, 161–166; f) J. M.
Bolivar, C. Luley-Goedl, E. Leitner, T. Sawangwan, B.
Nidetzky, J. Biotechnol. 2017, 257, 131–138.
107403; b) C. T. Trinh, B. Mendoza, Curr. Opin. Chem.
Eng. 2016, 14, 18–25; c) K. Yu, C. Liu, B. G. Kim, D. Y.
Lee, Biotechnol. Adv. 2015, 33, 155–164; d) C. J.
Hartley, C. C. Williams, J. A. Scoble, N. J. Turner, C.
Scott, Nat. Catal. 2019, 2, 1006–1015.
[20] a) C. Zhong, B. Duić, J. M. Bolivar, B. Nidetzky,
ChemCatChem 2020, 12, 1350–1358; b) M. B. Buergler,
A. Dennig, B. Nidetzky, Biotechnol. Bioeng. 2020, 117, [26] A. Gutmann, A. Lepak, M. Diricks, T. Desmet, B.
2377–2388; c) D. Valikhani, J. M. Bolivar, A. Dennig, B.
Nidetzky, Biotechnol. J. 2017, 12, 1600557.
Nidetzky, Biotechnol. Bioeng. 2018, 115, 2416–2425.
Adv. Synth. Catal. 2021, 363, 2157–2169
2169
© 2021 The Authors. Advanced Synthesis & Catalysis
published by Wiley-VCH GmbH