An enzyme module system for in situ regeneration of deoxythymidine 5′-diphosphate (dTDP)-activated deoxy sugars

Article abstract of DOI:10.1002/adsc.200700058

A highly flexible enzyme module system (EMS) was developed which allows for the first time the in situ regeneration of deoxythymidine 5′-diphosphate (dTDP)-activated deoxy sugars and furthermore enables us to produce novel sorangiosides in a combinatorial biocatalytic approach using three enzyme modules. The SuSy module with the recombinant plant enzyme sucrose synthase (SuSy) and the deoxy sugar module consisting of the enzymes RmlB (4,6-dehydratase), RmlC (3,5-epimerase) and RmlD (4-ketoreductase) from the biosynthetic pathway of dTDP-β-L-rhamnose were combined with the glycosyltransferase module containing the promiscuous recombinant glycosyltransferase SorF from Sorangium cellulosum So cel2. Kinetic data and the catalytic efficiency were determined for the donor substrates of SorF: dTDP-α-D-glucose, dTDP-β-L-rhamnose, uridine diphosphate (UDP)-α-D-glucose (Glc), and dTDP-6-deoxy-4-keto-α-D-glucose. The synthesis of glucosyl-sorangioside with in situ regeneration of dTDP-Glc was accomplished by combination of SuSy and SorF. The potential of the EMS is demonstrated by combining SuSy, RmlB, RmlC, RmlD with SorF in one-pot for the in situ regeneration of dTDP-activated (deoxy) sugars. The HPLC/MS analysis revealed the formation of rhamnosyl-sorangioside and glucosyl-sorangioside, demonstrating the in situ regeneration of dTDP-β-L-rhamnose and dTDP-a-D-glucose and a cycle number for dTDP higher than 9. Furthermore, NADH (reduced form of nicotinamdie adenine dinucleotide) regeneration with formate dehydrogenase in the reduction step catalyzed by the 4-ketoreductase RmlD could be integrated in the one-pot synthesis yielding similar conversion rates and cycle numbers. In summary, we have established the first in situ regeneration cycle for dTDP-activated (deoxy) sugars by a highly flexible EMS which allows simple exchange of enzymes in the deoxy sugar module and exchange of glycosyltransferases as well as aglycones in the glycosyltransferase module to synthesize new hybrid glycosylated natural products in one-pot.

Full text of DOI:10.1002/adsc.200700058

FULL PAPERS
DOI: 10.1002/adsc.200700058
An Enzyme Module System for in situ Regeneration of Deoxy-
thymidine 5’-Diphosphate (dTDP)-Activated Deoxy Sugars
Carsten Rupprath,a Maren Kopp,b Dennis Hirtz,a Rolf Müller,b,*
and Lothar Ellinga,*
a
Laboratory for Biomaterials, Institute of Biotechnology and Helmholtz-Institute for Biomedical Engineering, RWTH
Aachen University, Worringer Weg 1, 52074 Aachen, Germany
Fax : (+49)-241-80-22387; e-mail: l.elling@biotec.rwth-aachen.de
Saarland University, Department of Pharmaceutical Biotechnology, P. O. Box 151150, 66041 Saarbrücken, Germany
b
Fax : (+49)-681-302-5473; e-mail: rom@mx.uni-saarland.de
Received: January 24, 2007; Revised: April 6, 2007
Dedicated to Prof. Dr. Dr. h.c. Maria-Regina Kula on the occasion of her 70th birthday (16.03.2007).
Supporting information for this article is available on the WWW under http://asc.wiley-vch.de/home/.
Abstract: A highly flexible enzyme module system RmlC, RmlD with SorF in one-pot for the in situ re-
(EMS) was developed which allows for the first time generation of dTDP-activated (deoxy) sugars. The
the in situ regeneration of deoxythymidine 5’-diphos- HPLC/MS analysis revealed the formation of rham-
phate (dTDP)-activated deoxy sugars and further- nosyl-sorangioside and glucosyl-sorangioside, demon-
more enables us to produce novel sorangiosides in a strating the in situ regeneration of dTDP-b-l-rham-
combinatorial biocatalytic approach using three nose and dTDP-a-d-glucose and a cycle number for
enzyme modules. The SuSy module with the re- dTDP higher than 9. Furthermore, NADH (reduced
combinant plant enzyme sucrose synthase (SuSy) form of nicotinamdie adenine dinucleotide) regener-
and the deoxy sugar module consisting of the en- ation with formate dehydrogenase in the reduction
zymes RmlB (4,6-dehydratase), RmlC (3,5-epimer- step catalyzed by the 4-ketoreductase RmlD could
ase) and RmlD (4-ketoreductase) from the biosyn- be integrated in the one-pot synthesis yielding simi-
thetic pathway of dTDP-b-l-rhamnose were com- lar conversion rates and cycle numbers. In summary,
bined with the glycosyltransferase module containing we have established the first in situ regeneration
the promiscuous recombinant glycosyltransferase cycle for dTDP-activated (deoxy) sugars by a highly
SorF from Sorangium cellulosum So ce12. Kinetic flexible EMS which allows simple exchange of en-
data and the catalytic efficiency were determined for zymes in the deoxy sugar module and exchange of
the donor substrates of SorF: dTDP-a-d-glucose, glycosyltransferases as well as aglycones in the glyco-
dTDP-b-l-rhamnose, uridine diphosphate (UDP)-a- syltransferase module to synthesize new hybrid gly-
d-glucose (Glc), and dTDP-6-deoxy-4-keto-a-d-glu- cosylated natural products in one-pot.
cose. The synthesis of glucosyl-sorangioside with in
situ regeneration of dTDP-Glc was accomplished by Keywords: biocatalysis; deoxythymidine 5’-diphos-
combination of SuSy and SorF. The potential of the phate (dTDP)-activated sugars; glycosylation; glyco-
EMS is demonstrated by combining SuSy, RmlB, syltransferase; natural products; sorangicin
Introduction
iation of the involved sugar compound, and recent
studies have emphasized the role of deoxy sugars as
Biological functions of glycoconjugates in animals, recognition elements in the action mechanism of
plants and microorganisms are mediated by d- and l- drugs.[2] Actinomycetes are the major source of micro-
deoxyhexoses.[1] A widely used method of nature to bial natural products and glycosylation is a very
change surface properties of compounds and organ- common finding in these secondary metabolites. An-
isms by influencing the interaction with the environ- other important source of such compounds are the
ment is the presentation of deoxy sugars. The biologi- myxobacteria.[3] Interestingly, only a very few glycosy-
cal function is often drastically altered by a small var- lated structures have been reported and no informa-
Adv. Synth. Catal. 2007, 349, 1489 – 1496
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1489

Carsten Rupprath et al.
FULL PAPERS
Scheme 1. Enzyme module system (EMS) for the synthesis of novel glycosylated natural products with in situ regeneration
of dTDP-activated deoxy sugars.
tion at all is available today about the biochemistry of kinase, and sucrose synthase generates the precursor
glycosyl transfer to natural product backbones in substrate dTDP-a-d-glucose (dTDP-Glc). The deoxy
myxobacteria. The glycosylation of natural products is sugar module (B) is highly flexible containing biosyn-
catalyzed by glycosyl transferases which requires com- thetic deoxy sugar pathways enzymes. dTDP-l-Rha
plex nucleotide sugars as donor substrates. For in and dTDP-6-deoxy-4-keto-a-d-glucose can now be
vitro glycosylation experiments these nucleotide synthesized in 0.1 and 0.6 g scale (Rupprath et al., un-
sugars can be synthesized by laborious chemical published results), respectively, by the combination of
methods which afford for every single nucleotide both enzyme modules with four enzymes (RmlB,
sugar its own individual synthesis route with 5 to 20 RmlC and RmlD, FDH) in the deoxy sugar module
synthesis steps depending on simple or complex nu- including also NADH cofactor regeneration.
cleotide sugars, for example, trideoxy sugars, and
In situ regeneration of NDP-activated sugars during
often results in low yields.[4] An alternative to chemi- synthesis of glycoconjugates is a versatile approach
cal synthesis is the enzymatic synthesis of nucleotide because isolation procedures of the most often labile
sugars involving subsequent or one-pot multiple cata- NDP-activated sugars and the feedback inhibition of
lytic steps with unprotected sugars in aqueous solu- the glycosyltransferase by-product NDP are avoided
tion.[5–8] The complex dideoxy sugar dTDP-b-l-olivose leading to substantial cost reduction and better reac-
and the dideoxy sugar pathway intermediate dTDP- tion performance.[9] Because of these exceptional ad-
2,6-dideoxy-4-keto-a-d-glucose were thus synthesized vances in situ regeneration systems for UDP-a-d-glu-
starting from 2-deoxyglucose 6-phosphate and dTTP cose (UDP-Glc),[10–14] UDP-a-d-galactose (UDP-
and exploiting the substrate flexibility of the biosyn- Gal),[10–12,15] UDP-a-d-glucuronic acid,[16] CMP-N-ace-
thetic pathway enzymes for dTDP-b-l-rhamnose tylneuraminic acid,[14,17] GDP-a-d-mannose,[18] GDP-
(dTDP-l-Rha).[6]
b-l-fucose,[17] UDP-N-acteyl-a-d-glucosamine[19] and
Recently, we established an enzyme module system UDP-N-acteyl-a-d-galactosamine[19] nucleotide sugars
(EMS) for the synthesis of dTDP-activated deoxy have been developed. In our previous work we com-
sugars starting from dTMP and sucrose (Scheme 1).[7] bined SuSy, UDP-Glc 4’-epimerase and b1,4-galacto-
The SuSy module (A) with dTMP-kinase, pyruvate syltransferase for the production of 0.5 g N-acetyllac-
1490
asc.wiley-vch.de
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 1489 – 1496

Regeneration of Deoxythymidine 5’-Diphosphate (dTDP)-Activated Deoxy Sugars
FULL PAPERS
Table 1. Kinetic data of the donor substrates of the glycosyl-
transferase SorF.
tosamine (LacNAc) in a 3-enzyme in situ regenera-
tion system for UDP-Gal.[11,12] We further utilized this
regeneration system for UDP-Gal for the synthesis of
the Galili-epitope involving b1,4-galactosyltransferase
and a1,3-galactosyltransferase.[15,20]
Despite all the benefits of in situ regeneration of
nucleotide sugars, an in situ regeneration system for
dTDP-activated sugars has not been established so
far. We could recently identify SorF, a glycosyl trans-
ferase from the myxobacterium Sorangium cellulosum
involved in the formation of the antibiotic sorangi-
cin.[21] Applying SorF, we present here the first in situ
Substrate
KM [mM]
vmax [U/mg] vmax/KM
dTDP-a-d-Glc (1)
0.12Æ0.026 1.36Æ0.026 11.33
dTDP-6-deoxy-4-keto- 0.39Æ0.095 0.48Æ0.019 1.23
glucose (2)
dTDP-b-l-Rha (3)
UDP-a-d-Glc (4)
1.01Æ0.44
1.69Æ0.28
0.85Æ0.1
0.84
4.01
6.78Æ0.32
regeneration system for dTDP-activated deoxy sugars Also the deoxy sugar pathway intermediate 2 is ac-
by using a generally applicable enzyme module cepted with a similar catalytic efficiency as 3, but ex-
system (EMS, Scheme 1).
hibits the lowest reaction rate of all tested donor sub-
strates. The enzyme affinity of SorF for 4 is similar to
the described glycosyltransferase GtfB and GtfE (KM
1.3 mM and 0.72 mM).[22]
Results and Discussion
In the present paper we have extended the EMS for
the synthesis of dTDP-deoxy sugars by a third compo- EMS for in situ Regeneration of dTDP-Deoxy Sugars
nent, the glycosyltransferase module (C, Scheme 1).
The combination of all three modules in one-pot es- The established EMS for the synthesis of dTDP-
tablishes the first in situ regeneration cycle for dTDP- deoxy sugars from dTMP and sucrose was extended
(deoxy) sugars. The cycle is started by catalytic by the glycosyltransferase module (C, Scheme 2). The
amounts of dTDP and avoids laborious purification SuSy module starts with catalytic amounts of dTDP
procedures for the often labile dTDP-activated deoxy and is combined with the two other modules in an
sugars. Also feed-back inhibition of deoxy sugars one-pot reaction. Based on the donor substrate flexi-
pathway enzymes[6] in the deoxy sugar module is bility of SorF,[21] two combinations of the modules
minimized by coupling it to the glycosyltransferase were tested to demonstrate in situ regeneration of the
module. The released dTDP from the glycosyltrans- dTDP-activated (deoxy) sugar during synthesis of the
ferase reaction is subsequently utilized to regenerate corresponding sorangiosides.
dTDP-Glc and the dTDP-deoxy sugar via the SuSy
The direct combination of the SuSy module and the
and deoxy sugar modules, respectively. We here dem- glycosyltransferase-module (EMS 1 in Scheme 2) led
onstrate the feasibility of our EMS by utilizing the to the formation of 5 with a conversion rate of 97%
promiscuous donor substrate acceptance of the glyco- (Figure 1) for the acceptor substrate sorangicin A as
syltransferase SorF[21] which leads to the synthesis of determined by HPLC/MS analysis (Supporting Infor-
novel sorangiosides. In order to optimize the EMS mation). Since only catalytic concentrations of dTDP,
the SorF glycosyltransferase was first kinetically char- namely one-tenth of the acceptor concentration, were
acterized for its donor substrate spectrum.
used, the maximum cycle number of 10 for dTDP was
nearly reached (9.7) demonstrating efficient in situ re-
generation of 1.
Kinetic Characterization of the Glycosyltransferase SorF
All three modules were combined to prove in situ
regeneration of 3 (EMS 2 in Scheme 2). Four differ-
The SorF glycosyltransferase exhibits substrate flexi- ent experiments were carried out to optimize the
bility towards dTDP-activated d- and l-sugars.[21] ratios of the enzyme activities. In the first experiment
Table 1 summarizes the kinetic values and the catalyt- the activities of SuSy and the Rml enzymes were
ic efficiencies of SorF for the donor substrates dTDP- equal. Figure 2 illustrate that 5 is the main product
Glc (1), dTDP-6-deoxy-4-keto-a-d-glucose (2), and 6 is only synthesized as a minor product with con-
dTDP-l-Rha (3), and UDP-Glc (4). The affinity of version rates of 59% and 9.2%, respectively (see also
SorF for 1 was significantly higher (KM 0.12 mM) than Supporting Information). This result reflects the ki-
for the proposed natural substrate 4 (KM 1.69 mM). netic properties of SorF (Table 1). dTDP-Glc (1) has
Although the reaction rate for 4 is better than for 1, the highest affinity towards SorF and 13-fold greater
the catalytic efficiency (vmax/KM) for 1 is notably vmax/KM compared to 3 which favours the formation of
higher suggesting that this is the natural donor sub- the 5.
strate of SorF. Interestingly, with 3 SorF shows a rea-
At that time optimization of the enzyme activities
sonable reaction rate and a higher affinity than for 4. in the modules came into focus to direct the synthesis
Adv. Synth. Catal. 2007, 349, 1489 – 1496
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1491

Carsten Rupprath et al.
FULL PAPERS
Scheme 2. Enzyme module systems (EMS) for the in situ regeneration of dTDP-activated deoxysugars. EMS 1: Combination
of the SuSy module (A with i for SuSy) and the glycosyltransferase module (C with v for SorF) for the synthesis of glucosyl-
sorangioside (5) with in situ regeneration of 1. EMS 2: Combination of the SuSy module (A), the deoxy sugar module (B
with ii for RmlB; iii for RmlC; iv for RmlD), and the glycosyltransferase module (C) for the synthesis of rhamnosylsorangio-
side (6), 6-deoxy-4-keto-glucosylsorangioside (7) and 5 with in situ regeneration of 3, 2, and 1.
of the favoured glycosylated compound (Figure 2 and
Supporting Information). In the second experiment
the activities of RmlB (5-fold), RmlC (2.4-fold) and
RmlD (2.4-fold) in the deoxy sugar module were in-
creased to prevent the accumulation of 1 and to force
the formation of 3 as substrate for the transfer reac-
tion. These conditions resulted in the formation of 6
as the main product with a conversion rate of 57%.
However, 5 and 7 were also detected by HPLC/MS
with conversion rates of 32% and 6%, respectively
(Figure 2). The cycle number for dTDP reached 9.5
(maximum 10) demonstrating again efficient in situ
regeneration of the dTDP-(deoxy) sugars. Most im-
Figure 1. HPLC/MS analysis with peak areas for the educt
portantly, dTDP was cycled nearly 6 times throughout
sorangicin A and the product glucosyl-sorangioside 5 in
the whole EMS with five enzymes involved in the re-
enzyme module system 1 (EMS 1, Scheme 2). The SuSy
generation of 3. However, the conversion of dTDP-
module (A) and the glycosyltransferase module (C) with
Glc led still to the formation of 5, because of the high
affinity and high reaction rate of SorF with 1
(Table 1).
SorF were combined for in situ regeneration of 1. The peak
area of product 5 corresponds to the conversion rate of the
educt.
In the third experiment the activity ratio of RmlB/
SorF was further increased to give high excess of
1492
asc.wiley-vch.de
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 1489 – 1496

Regeneration of Deoxythymidine 5’-Diphosphate (dTDP)-Activated Deoxy Sugars
FULL PAPERS
Recently, Eguchi et al. and Thorson et al. reported
on the in situ generation of complex dTDP-deoxy
sugars by the reverse reaction of glycosyltransferases
involved in the formation of vicenistatin, calichamy-
cin, and vancomycin.[25] The reactions resemble on
first sight those of sucrose synthase (SuSy) which we
used to generate dTDP-Glc by the reverse reaction of
this plant glycosyltransferase. However, sucrose is a
high energy substrate and the synthesis of NDP-Glc
with SuSy can be readily optimized towards high
enzyme productivities as demonstrated by our previ-
ous work.[7,8,12,26] In the synthesis of 1 and 3 from su-
crose we reach for the SuSy reaction turnover num-
bers of 33,000 and 440 (mmol product/mmol enzyme),
respectively.[7] In contrast, the formation of dTDP-
deoxy sugars from calichamycins or vancomycins by
the reverse reaction of glycosyltransferase reaches
only turnover numbers between 5 and 10, which is
due to relatively high enzyme concentrations in the
mM range.[25b] Moreover, for the enzymatic aglycone
switch using the reverse reaction of the glycosyltrans-
ferase VinC, Eguchi et al. report a maximum conver-
sion yield of 42% for the acceptor substrate with a 3-
fold molar excess of the enzyme which corresponds to
a turnover number of 0.14.[25a]
Figure 2. HPLC/MS analysis with peak areas for the educt
sorangicin A and the products glucosyl-sorangioside 5,
rhamnosyl-sorangioside 6, and 6-deoxy-4-keto-glucosyl-sor-
angioside 7 in enzyme module system 2 (EMS 2, Scheme 2).
The three modules A, B and C in Scheme 2 were combined
for in situ regeneration of the donor substrates 1, 2, and 3.
Enzyme ratios in the enzyme modules were varied in experi-
ments 1 to 4 as described in the Experimental Section. The
peak areas of the products correspond to the conversion
rate of the educt.
RmlB and to force the synthesis of 3. Figure 2 illus-
trate that only 6 and 5 with conversion rates of 61%
and 30%, respectively, were synthesized (see also Conclusions
Supporting Information). dTDP is recycled more than
6 times in the EMS running through all enzymatic In summary, we have demonstrated that dTDP-acti-
steps for the in situ regeneration of 3. All together vated (deoxy) sugars like 1, 2, and 3 are regenerated
more than 9 cycles of dTDP were achieved for the in in situ using an enzyme module system consisting of
situ regeneration of 1 and 3 as donor substrates of up to 6 enzymes. Other systems where the NDP-sugar
SorF (maximum cycle number is 10). Most interest- is generated by a pyrophosphorylase, for example,
ingly, compound 7 was not formed under these condi- RmlA,[24] may also be used, but are difficult to opti-
tions, which is probably a result of the reduced SorF mize due to the larger number of enzymes needed for
activity.
regeneration of the nucleotide sugar (eight enzymes
Finally, to demonstrate that in situ regeneration of starting from d-glucose) and the sensitive feedback
NADH (Scheme 2) can also be integrated into the inhibition of RmlA (22 mM)[6] by 3. In our EMS we
EMS, formate dehydrogenase (FDH) was added to- could also demonstrate that the use of a promiscuous
gether with the substrate sodium formate to generate glycosyltransferase yielded differently glycosylated
NADH from NAD+ as cofactor for the reaction of products. Extension of our EMS by other deoxy sugar
RmlD (experiment 4 in Figure 2 and Supporting In- pathway enzymes should also yield libraries of glyco-
formation). HPLC/MS analysis clearly demonstrates sylated natural compounds. On first sight, the combi-
that the EMS with 6 enzymes generates 6 as the main nation of many enzymes seems to be a disadvantage.
product with a conversion rate of 53%. Together with However, many dTDP-deoxy sugar pathways have
the products 5 (26%) and 7 (17%) a conversion rate been already set up in vitro.[8,27]
of 96% is reached for the acceptor sorangicin A.
Again dTDP reaches almost the maximum cycle
number of 10 including 5 cycles for the in situ regen-
Experimental Section
eration of 3. Thus, NAD+ also reaches a cycle number
of 5 (maximum 10).
dTDP-6-deoxy-4-keto-a-d-glucose (2) and dTDP-l-Rha (3)
were synthesized as described elsewhere.[7] dTDP, 1 and 4
were purchased from Sigma–Aldrich. Formate dehydrogen-
ase was from Julich Chiral Solutions (Jülich, Germany).
NADH, Tris buffer, MgCl2, ammonium formate, BSA and
IPTG were purchased from Roth (Karlsruhe, Germany);
Besides the improvement of product formation
demonstrated by these four experiments, further opti-
mization of the in situ regeneration of dTDP-activat-
ed deoxy sugars could be facilitated by the utilization
of genetic algorithms.[23]
Adv. Synth. Catal. 2007, 349, 1489 – 1496
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1493

Carsten Rupprath et al.
FULL PAPERS
yeast extract and tryptone peptone were from BD Biosci-
ences (Heidelberg, Germany). Sorangicin A was provided
by Dr. Irschik (Helmholtz Centre for Infection Research,
Braunschweig, Germany) as described before.[21]
Enzyme Module System 2 (EMS 2)
The three modules (Scheme 2) containing the purified en-
zymes SuSy, RmlB (4,6-dehydratase), RmlC (3,5-epimer-
ase), RmlD (4-ketoreductase), and SorF were combined.
Four experiments were performed as described below for
optimization of product formation and demonstration of
NADH in situ regeneration using formate dehydrogenase.
All samples were incubated and worked up as described
above.
Experiment 1: In a total volume of 500 mL Tris-HCl buffer
(50 mM, pH 7.5), MgCl2 (10 mM), sucrose (250 mM), soran-
gicin A (1.4 mM), SuSy (1 UmLÀ1), RmlB (1 UmLÀ1),
RmlC (1 UmLÀ1), RmlD (1 UmLÀ1), NADH (1.4 mM), and
SorF (0.13 UmLÀ1) were mixed. The reaction was started by
adding the dTDP (0.14 mM).
Production and Purification of SorF
The production and purification of SorF was carried out as
described elsewhere.[21] The fermentation was scaled up to
10-L scale. An ED10 B. Braun (Melsungen, Germany) fer-
menter was filled with 10 L LB medium and inoculated with
a 200 mL LB preparatory culture, which was incubated for
18 h at 378C in an HT shaker (Infors, Einsbach, Germany).
The E. coli cells were grown at 378C until an OD600 =0.5–
0.7 was reached. The oxygen partial pressure was regulated
dynamically with the stirrer speed using a threshold concen-
tration of 40% with an air flow of 10 L/min. Production of
SorF was induced with 0.4 mm IPTG, followed by incubation
for 24 h at 168C. The cells were then harvested by centrifu-
gation using a RC5B centrifuge (Thermo Electron, Langen-
selbold, Germany).
Experiment 2: The conditions in experiment 1 were modi-
fied by increasing the enzyme activities in module B to
RmlB (5 UmLÀ1), RmlC (2.4 UmLÀ1), and RmlD (2.4
UmLÀ1). All other parameters remained as described for ex-
periment 1.
Experiment 3: The SorF activity was reduced to 0.065
UmLÀ1 and the RmlB activity increased to 10 UmLÀ1.
RmlC and D activities and conditions remained as described
in experiment 2.
Experiment 4: NADH in situ regeneration by formate de-
hydrogenase should be demonstrated in the EMS. In a total
volume of 500 mL Tris-HCl buffer (50 mM, pH 7.5) MgCl2
(10 mÀM1 ), sucrose (250 mM), sorangicin A (1.4 mM), SuSy (1
UmL ), RmlB (5 UmLÀ1), RmlC (1.2 UmLÀ1), RmlD (1.2
UmLÀ1), NAD+ (0.14 mM), FDH (5 UmLÀ1), ammonium
formate (200 mM), and SorF (0.13 UmLÀ1) were mixed.
The reaction was started by adding dTDP (0.14 mM).
Production and Purification of RmlB, RmlC, RmlD,
and SuSy
The cloning, expression and purification were performed as
described and published elsewhere.[6,7,28]
Kinetic Data of SorF
The maximum reaction rate vmax and the Michaelis constant
KM were determined for the donor substrates dTDP-Glc (1),
dTDP-6-deoxy-4-keto-a-d-glucose (2), dTDP-l-Rha (3), and
UDP-Glc (4). In a total volume of 20 mL the Tris-HCl
buffer (50 mM), MgCl2 (10 mM), sorangicin A (1.4 mM),
BSA (1 mgmLÀ1), SorF (20 mUmLÀ1 and 10 mUmLÀ1)
were mixed with the NDP-activated sugars (0.5, 1, 2, 4, 6, 8
and 10 mM). After starting the reaction with sorangicin A,
the samples were incubated for 5 min at 308C and then stop-
ped by heating (958C) for 30 s. The samples were centri-
fuged for 15 min at 15000 rpm (Rotina 35R, Hettich, Ger-
many) and subsequently analyzed by CE (see below). Spe-
cific enzyme activities were determined from the produced
dTDP, and the corresponding protein concentrations.[29] The
vmax and KM values were calculated by non-linear regression
using the Michaelis–Menten-equation (Sigma Plot, Systat
Software GmbH, Erkrath, Germany).
Analysis of Sorangicin and the Sorangiosides
After stopping the EMS by heating, samples were centri-
fuged for 15 min at 15000 rpm (Rotina 35R, Hettich, Tuttlin-
gen, Germany) and products purified by solid-phase extrac-
tion on Sep-Pak C18 columns (Waters, Eschborn, Germany)
using methanol. The samples were evaporated in a speed-
vac, and the remaining residue was dissolved in 200 mL
methanol for HPLC/MS analysis. A DAD-HPLC (Agilent
1100 Serie) coupled to an electrospray ionization device
connected to an ion trap of an HCT plus MS was employed
(Bruker, Bremen, Germany). HPLC was carried out on a
reversed phase Nucleodur C18 column 125”2 mm/3 mm
(Macherey–Nagel, Düren, Germany) using different gradi-
ents ranging from 95% solvent A (water + 0.1% formic
acid) to 95% solvent B (acetonitrile + 0.1% formic acid)
with a flow rate of 0.4 mL/min.
Enzyme Module System 1 (EMS 1)
The SuSy and the glycosyltransferase modules (Scheme 2)
containing purified SuSy and SorF were combined. In a
total volume of 500 mL Tris-HCl buffer (50 mM, pH 7.5),
MgCl2 (10 mM), sucrose (250 mM), sorangicin A (1.4 mM), Capillary Electrophoresis of NDP and NDP-Sugars
SuSy (1 UmLÀ1), and SorF (0.13 UmLÀ1) were mixed. The
reaction was started by adding dTDP (0.14 mM). The sam-
The samples for the determination of kinetic data (vmax and
KM) of the NDP-activated donor substrates of SorF were
ples were incubated at 308C for 18 h and then stopped by
stopped by heating (958C) for 0.5 min, centrifuged for
heating (958C) for 3 min. Subsequently, the formed soran-
15 min at 15000 rpm (Rotina 35R, Hettich, Tuttlingen, Ger-
giosides were purified by solid-phase extraction and ana-
many) and examined by analysis of the side products. The
lysed by HPLC/MS (see below).
formation of dTDP and UDP was monitored by capillary
electrophoresis on a P/ACE MDQ apparatus from Beckman
Coulter (Krefeld, Germany), equipped with a UV detector.
Separation of NDP and NDP-sugars was accomplished on
1494
asc.wiley-vch.de
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 1489 – 1496

Regeneration of Deoxythymidine 5’-Diphosphate (dTDP)-Activated Deoxy Sugars
FULL PAPERS
an untreated fused-silica capillary (I.D. 75 mm, 57 cm total
capillary length, 50 cm to the detector) with 50 mM
Na2B4O7·10H2O/64 mM boric acid buffer, pH 8.9. Condi-
tions for migration and detection were 25 kV (23 mA) at
258C and UV detection at 254 nm, respectively. Samples
were injected by pressure (5.0 sec at 0.5 psi in the forward
direction). The conversion of NDP-sugars by SorF was cal-
culated by determination of NDP concentrations using stan-
dard calibration curves for dTDP and UDP, respectively.
1995, 60, 14–15; f) G. Baisch, R. Oehrlein, Bioorg Med
Chem. 1997, 5, 383–391.
[5] T. Bülter, L. Elling, Glycoconjugate J. 1999, 16, 147–
159.
[6] S. Amann, G. Dräger, C. Rupprath, A. Kirschning, L.
Elling, Carbohydr. Res. 2001, 335, 23–32.
[7] L. Elling, C. Rupprath, N. Günther, U. Rçmer, S. Ver-
seck, P. Weingarten, G. Dräger, A. Kirschning, W. Pie-
persberg, ChemBioChem. 2005, 6, 1423–1430.
[8] C. Rupprath, T. Schumacher, L. Elling, Curr. Med.
Chem. 2005, 12, 1637–1675.
[9] a) L. Elling, in: Advances in Biochemical Engineering/
Biotechnology, Vol. 58, (Ed.: T. Scheper), Springer-
Verlag, Berlin, 1997, pp 89–144; b) K. M. Koeller,
C. H. Wong, Chem. Rev. 2000, 100, 4465–4493.
[10] a) C. H. Wong, S. L. Haynie, G. M. Whitesides, J. Org.
Chem. 1982, 47, 5416–5418; b) T. Noguchi, T. Shiba,
Biosci. Biotechnol. Biochem. 1998, 62, 1594–1596.
[11] L. Elling, M. Grothus, M.-R. Kula, Glycobiology 1993,
3, 349–355.
Acknowledgements
L.E. and C.R. dedicate this paper to Prof. Dr. Dr. h.c. Maria-
Regina Kula on the occasion of her 70th birthday. L.E. thanks
Prof. Kula for her constant support and fruitful discussions
during research at the former Institute of Enzyme Technology
of the Heinrich-Heine-University Düsseldorf at the Research
Centre Jülich, Germany. Financial support through the EU
grant “GENOVA” (Contract No. QLK3–999–00095) to L.E.
is gratefully acknowledged. The authors thank Daniel Krug
for help with HPLC-MS analysis of sorangicin and its deriv-
atives. Research in the laboratory of R.M. was supported by
the Deutsche Forschungsgemeinschaft (DFG) and the Bun-
desministerium für Bildung und Forschung (BMB+F).
[12] A. Zervosen, L. Elling, J. Am. Chem. Soc. 1996, 118,
1836–1840.
[13] a) M. Ichikawa, R. L. Schnaar, Y. Ichikawa, Tetrahe-
dron Lett. 1995, 36, 8731–8732; b) Y. Ichikawa, R.
Wang, C. H. Wong, Meth. Enzymol. 1994, 247 107.
[14] Y. Ichikawa, J. L.-C. Liu, G.-J. Shen, C. H. Wong, J.
Am. Chem. Soc. 1991, 113, 6300–6302.
[15] C. H. Hokke, A. Zervosen, L. Elling, D. H. Joziasse,
D. H. van den Eijnden, Glycoconj. J. 1996, 13, 687–692.
[16] D. Gygax, P. Spies, T. Winkler, U. Pfaar, Tetrahedron
1991, 47, 5119–5122.
References
[17] Y. Ichikawa, Y. C. Lin, D. P. Dumas, G. J. Shen, E. Gar-
ciajunceda, M. A. Williams, R. Bayer, C. Ketcham,
L. E. Walker, J. C. Paulson, C. H. Wong, J. Am. Chem.
Soc. 1992, 114, 9283–9298.
[1] a) S. Hanessian, Adv. Carbohydr. Chem. 1966, 21, 143–
207; b) T. Reichstein, E. Weiss, Adv. Carbohydr. Chem.
1962, 17, 65–120; c) N. Williams, J. Wander, in: The
Carbohydrates, Chemistry and Biochemistry, Vol. 1B,
(Eds.: W. Pigman, D. Horton), Academic Press, New
York, 1980, pp 761–798; d) J. F. Kennedy, C. A. White,
Bioactive carbohydrates in chemistry, biochemistry and
biology, Ellis Horwood Publishers, Chichester, 1983.
[2] a) S. M. Poulsen, C. Kofoed, B. Vester, J. Mol. Biol.
2000, 304, 471–481; b) J. R. Menninger, J. Basic Clin.
Physiol. Pharmacol. 1995, 6, 229–250; c) F. Schlunzen,
R. Zarivach, J. Harms, A. Bashan, A. Tocilj, R. Al-
brecht, A. Yonath, F. Franceschi, Nature 2001, 413,
814–821; d) J. S. Thorson, T. J. J. Hosted Jr., J. Jiang,
B. J. B. , J. Ahlert, Curr. Org. Chem. 2001, 5, 139–167;
e) J. L. Hansen, J. A. Ippolito, N. Ban, P. Nissen, P. B.
Moore, T. A. Steitz, Mol. Cell 2002, 10, 117–128; f) F.
Schlunzen, J. M. Harms, F. Franceschi, H. A. S. Hansen,
H. Bartels, R. Zarivach, A. Yonath, Structure 2003, 11,
329–338; g) M. Gaynor, A. S. Mankin, Curr. Top. Med.
Chem. 2003, 3, 949–960.
[18] P. Wang, G. J. Shen, Y. F. Wang, Y. Ichikawa, C. H.
Wong, J. Org. Chem. 1993, 58, 3985–3990.
[19] G. C. Look, Y. Ichikawa, G. J. Shen, P. W. Cheng, C. H.
Wong, J. Org. Chem. 1993, 58, 4326–4330.
[20] N. Brinkmann, M. Malissard, M. Ramuz, U. Rçmer, T.
Schumacher, E. G. Berger, L. Elling, C. Wandrey, A.
Liese, Bioorg. Med. Chem. Lett. 2001, 11, 2503–2506.
[21] M. Kopp, C. Rupprath, H. Irschik, A. Bechthold, L.
Elling, R. Müller, ChemBioChem. 2007, 8, 813–819.
[22] A. M. Mulichak, H. C. Losey, C. T. Walsh, R. M. Gara-
vito, Structure (Camb) 2001, 9, 547–557.
[23] C. Hoh, G. Dudziak, A. Liese, Bioorg. Med. Chem.
Lett. 2002, 1031–1034.
[24] Y. B. Kang, Y. H. Yang, K. W. Lee, S. G. Lee, J. K.
Sohng, H. C. Lee, K. Liou, B. G. Kim, Biotechnol.
Bioeng. 2006, 93, 21–27.
[25] a) A. Minami, K. Kakinuma, T. Eguchi, Tetrahedron
Lett. 2005, 46, 6187–6190; b) C. S. Zhang, B. R. Grif-
fith, Q. Fu, C. Albermann, X. Fu, I. K. Lee, L. J. Li,
J. S. Thorson, Science 2006, 313, 1291–1294; c) H. B.
Bode, R. Müller, Angew. Chem. 2007, 119, 2195–2198.
[26] a) A. Zervosen, A. Stein, H. Adrian, L. Elling, Tetrahe-
dron 1996, 52, 2395–2404; b) T. Bülter, C. Wandrey, L.
Elling, Carbohydr. Res. 1997, 305, 469–473; c) A. Stein,
M. R. Kula, L. Elling, Glycoconjugate J. 1998, 15, 139–
145; d) U. Rçmer, N. Nettelstroth, W. Kçckenberger,
L. Elling, Adv. Synth. Catal. 2001, 343, 655–661.
[3] K. Gerth, S. Pradella, O. Perlova, S. Beyer, R. Müller,
J. Biotechnol. 2003, 106, 233–253.
[4] a) J. G. Moffat, in: Methods in Enzymology, Vol. 8,
(Ed.: F. Ginsburg), Academic Press, New York,
London, 1966, pp 136–142; b) V. Wittmann, C. H.
Wong, J. Org. Chem. 1997, 62, 2144–2147; c) F.
Cramer, H. Neuenhoeffer, K. H. Scheit, G. Schneider,
J. Tennigkeit, Angew. Chem. 1962, 74, 387; d) F.
Cramer, H. Schaller, H. A. Staab, Chem. Ber. 1961, 94,
1612–1621; e) M. Arlt, O. Hindsgaul, J. Org. Chem.
Adv. Synth. Catal. 2007, 349, 1489 – 1496
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1495

Carsten Rupprath et al.
FULL PAPERS
[27] a) T. M. Hallis, H. W. Liu, Acc. Chem. Res. 1999, 32,
579–588; b) A. Trefzer, J. A. Salas, A. Bechthold, Nat.
Prod. Rep. 1999, 16, 283–299.
[28] U. Rçmer, H. Schrader, N. Günther, N. Nettelstroth,
W. B. Frommer, L. Elling, J. Biotechnol. 2004, 107,
135–149.
[29] M. M. Bradford, Anal. Biochem. 1976, 72, 248–254.
1496
asc.wiley-vch.de
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2007, 349, 1489 – 1496