A novel three-enzyme reaction cycle for the synthesis of N-acetyllactosamine with in situ regeneration of uridine 5'-diphosphate glucose and uridine 5'-diphosphate galactose

Article abstract of DOI:10.1021/ja953495e

A new three-enzyme reaction cycle consisting of sucrose synthase, UDP glucose 4'-epimerase, and human β-1,4-galactosyltransferase was established for the synthesis of N-acetyllactosamine (LacNAc) with in situ regeneration of UDP galactose. We found that UDP glucose 4'-epimerase is reductively inactivated in the presence of UMP and acceptor substrates of β-1,4-galactosyltransferase. Reactivation of UDP glucose 4'-epimerase by the transition state analogues dUDP or dTDP 6-deoxy-D-xylo-4-hexulose in combination with the repetitive batch technique enabled us to use the native enzymes for 11 days in this cycle. With 10 U of sucrose synthase, 5 U of UDP glucose 4'-epimerase, and 1.25 U of β-1,4-galactosyltransferase, 594 mg of LacNAc could be synthesized. N-Acetyllactosamine was also subsequently converted to Neu5Acα2,6Ga1β1,4GlcNAc with α-2,6-sialyltransferase and CMP-Neu5Ac.

Full text of DOI:10.1021/ja953495e

1836
J. Am. Chem. Soc. 1996, 118, 1836-1840
A Novel Three-Enzyme Reaction Cycle for the Synthesis of
N-Acetyllactosamine with in Situ Regeneration of Uridine
5′-Diphosphate Glucose and Uridine 5′-Diphosphate Galactose
Astrid Zervosen and Lothar Elling*
Contribution from the Institute of Enzyme Technology, Heinrich-Heine-UniVersity of Du¨sseldorf,
Research Center Ju¨lich, P.O. Box 2050, D-52404 Ju¨lich, Germany
ReceiVed October 18, 1995X
Abstract: A new three-enzyme reaction cycle consisting of sucrose synthase, UDP glucose 4′-epimerase, and human
â-1,4-galactosyltransferase was established for the synthesis of N-acetyllactosamine (LacNAc) with in situ regeneration
of UDP galactose. We found that UDP glucose 4′-epimerase is reductively inactivated in the presence of UMP and
acceptor substrates of â-1,4-galactosyltransferase. Reactivation of UDP glucose 4′-epimerase by the transition state
analogues dUDP or dTDP 6-deoxy-D-xylo-4-hexulose in combination with the repetitive batch technique enabled us
to use the native enzymes for 11 days in this cycle. With 10 U of sucrose synthase, 5 U of UDP glucose 4′-
epimerase, and 1.25 U of â-1,4-galactosyltransferase, 594 mg of LacNAc could be synthesized. N-Acetyllactosamine
was also subsequently converted to Neu5AcR2,6Galâl,4GlcNAc with R-2,6-sialyltransferase and CMP-Neu5Ac.
Introduction
of transferases by nucleotides as well as large-scale preparation
and isolation of nucleotide sugars for stoichiometric reactions,
in situ regeneration cycles for nucleotide sugars have been
established.4 Previously published enzymatic syntheses of
LacNAc with in situ regeneration of UDP-Gal comprised six
enzymes.5,6
We present here the synthesis of N-acetyllactosamine (Lac-
NAc, 1) with a three-enzyme reaction cycle using human milk
â-1,4-galactosyltransferase (GalT, EC 2.4.1.38) and the con-
comitant in situ regeneration of UDP-Gal by two enzymes:
sucrose synthase (SuSy, EC 2.4.1.13) and UDP glucose 4′-
epimerase (epimerase, EC 5.1.3.2) (Figure 1).
The unique character of sucrose synthase to generate activated
glucoses by the cleavage of sucrose with nucleoside diphos-
phates makes this plant glycosyltransferase favorable over
pyrophosphorylases.7-9 Compared to previously published in
situ regeneration cycles for UDP-Gal,5,6 our novel three-enzyme
reaction cycle does not afford phospho(enol)pyruvate (PEP) for
nucleoside triphosphate regeneration nor does it generate
phosphate (from pyrophosphate by pyrophosphatase) avoiding
inhibition. To ensure enzyme availability we have purified
sucrose synthase from rice grains in a pilot scale.10
In this paper, we present results following three objectives:
(i) optimized conditions for all three enzymes with respect to
high enzyme productivities (g of 1/U of enzyme), (ii) reuse of
the three native enzymes with the repetitive batch technique in
order to increase enzyme productivity and decrease enzyme
costs, (iii) reactivation of reductively inactivated UDP-Glc 4′-
epimerase appearing in the presence of UMP and free monosac-
charides (acceptor substrates of GalT) in the reaction mixture.
The realization that the oligosaccharide moieties of glyco-
conjugates are involved in important intra- and intercellular
communication events leads to an increasing demand for
oligosaccharides.1a-f Compared to chemical syntheses with
many protection and deprotection steps, enzymatic syntheses
using glycosyltransferases and glycosidases give stereo- and
regioselectively glycosidic bonds in one step without the need
for protection.2a-d
Glycosyltransferases require the very expensive nucleotide
sugars as substrates which can chemically3a-g and enzy-
matically3h-o be prepared. In order to avoid product inhibition
X Abstract published in AdVance ACS Abstracts, February 1, 1996.
(1) (a) Radermacher, T. W.; Parekh, R. B.; Dwek, R. A. Ann. ReV.
Biochem. 1988, 57, 785-838. (b) Varki, A. Glycobiology 1993, 3, 97-
130. (c) Varki, A. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 7390-7397. (d)
Lasky, L. A. Science 1992, 258, 964-969. (e) Hart, G. W. Curr. Opin.
Cell. Biol. 1992, 4, 1017-1023. (f) Feizi, T. Trends Biochem. Sci. 1991,
84.
(2) (a) Wong, C.-H.; Halcomb, R. L.; Ichikawa, Y.; Kajimoto, T. Angew.
Chem. 1995, 107, 453-474. (b) Wong, C.-H.; Halcomb, R. L.; Ichikawa,
Y.; Kajimoto, T. Angew. Chem. 1995, 107, 569-593. (c) De Frees, S. A.;
Kosch, W.; Way, W.; Paulson, J. C.; Sabesan, S.; Halcomb, R. L.; Huang,
D.-H.; Ichikawa, Y.; Wong, C.-H. J. Am. Chem. Soc. 1995, 117, 66-79.
(d) Nilsson, K. G. I. Trends Biotechnol. 1988, 6, 256-264.
(3) Chemical synthesis: (a) Kochetkov, N. K.; Shibaev, V. N. In
AdVances in carbohydrate chemistry and biochemistry; Tipson, R. S.,
Horton, D., Eds.; Academic Press: New York, 1973; Vol. 28, pp 307-
399. (b) Ichikawa, Y.; Sim, M. M.; Wong, C.-H. J. Org. Chem. 1992, 57,
2943-2946. (c) Sim, M. M., Kondo, H.; Wong, C. H. J. Am. Chem. Soc.
1993, 115, 2260-2267. (d) Schmidt, R. R; Wegmann, B.; Jung, K.-H.
Liebigs Ann. Chem. 1991, 2, 121-124. (e) Gokhale, U. B.; Hindsgaul, O.;
Palcic, M. M. Can. J. Chem. 1990, 68, 1063-1071. (f) Klaffke, W.
Carbohydr. Res. 1995, 266, 285-292. (g) Arlt, M.; Hindsgaul, O. J. Org.
Chem. 1995, 60, 14-15. Enzymatic synthesis: (h) Toone, E. J.; Whitesides,
G. M. Am. Chem. Soc. Symp. Ser. 1991, 466, 1-22. (i) Wong, C.-H.;
Whitesides, G. M. Enzymes in Synthetic Organic Chemistry. In Tetrahedron
Organic Chemistry Series, Vol. 12; Baldwin, J. E., Magnus, P. D., Eds.;
Elsevier Science: Oxford, U. K., 1994; pp252-311. (j) Heidlas, J. E.; Lees,
W. J.; Whitesides, G. M. J. Org. Chem. 1992, 57, 152-157. (k) Heidlas,
J. E.; Lees, W. J.; Pale, P.; Whitesides, G. M. J. Org. Chem. 1992, 57,
146-151. (l) Simon, E. S.; Grabowski, S.; Whitesides, G. M. J. Org. Chem.
1990, 55, 1834-1841. (m) Toone, E. J.; Simon, E. S.; Whitesides, G. M.
J. Org. Chem. 1991, 56, 5603-5606. (n) Shames, S. L.; Simon, E. S.;
Christopher, C. W.; Schmid, W.; Whitesides, G. M.; Yang, L.-L. Glyco-
biology 1991, 1, 187-191. (o) Shen, G.-J.; Liu, J. L.-C.; Wong, C.-H.
Biocatalysis 1992, 6, 31-42.
(4) Ichikawa, Y.; Look, G. C.; Wong, C.-H. Anal. Biochem. 1992, 202,
215-238.
(5) Wong, C.-H.; Haynie, S. L.; Whitesides, G. M. J. Org. Chem. 1982,
47, 5416-5418.
(6) Wong, C.-H.; Wang, R.; Ichikawa, Y. J. Org. Chem. 1992, 57, 4343-
4344.
(7) Elling, L.; Kula, M.-R. J. Biotechnol. 1993, 29, 277-286.
(8) Elling, L.; Grothus, M.; Kula, M.-R. Glycobiology 1993, 349-355.
(9) Zervosen, A.; Elling, L. Kula, M.-R. Angew. Chem., Int. Ed. Engl.
1994, 33, 571-572.
(10) Elling, L.; Gu¨ldenberg, B.; Grothus, M.; Zervosen, A.; Peus, M.;
Helfer, A.; Stein, A.; Adrian, H.; Kula, M.-R. Biotechnol. Appl. Biochem.
l995, 29-37.
0002-7863/96/1518-1836$12.00/0 © 1996 American Chemical Society

Reaction Cycle for Synthesis of N-Acetyllactosamine
J. Am. Chem. Soc., Vol. 118, No. 8, 1996 1837
Figure 1. Three-enzyme reaction cycle for LacNAc synthesis with (1) sucrose synthase, (2) UDP-Glc 4′-epimerase, and (3) â-1,4-galactosyl-
transferase.
Table 1. Results for LacNAc Synthesis under Starting and
Optimized Conditions
starting
conditions
optimized
conditions
yield
95.6% after 24 h 100% after 15 h
space-time yield (g L-1 d-1
)
1.8
6.1
10
cycle no.a
4.8
g of LacNAc/U of GalT
g of LacNAc/U of SuSy
g of LacNAc/U of epimerase 0.009
0.018
0.002
0.077
0.010
0.019
a Cycle number means mol of UDP-Glc/mol of 1.
Figure 2. Reductive inactivation of UDP-Glc 4′-epimerase in the
presence of different acceptors and donors of GalT. Epimerase was
incubated in the presence of 1 mM UDP-Glc alone and in the presence
of 0.1 mM UMP with 50 mM GlcNAc, 10 mM O-(6-aminohexyl)-2-
acetamido-2-deoxy-â-D-glucopyranoside, 50 mM O-n-octyl-â-D-glu-
copyranoside, 50 mM D-Glc, 50 mM 2-deoxy-D-Glc, and 50 mM 5-thio-
D-Glc, respectively. The incubation period is indicated in the bars.
Results and Discussion
Optimization of LacNAc Synthesis. The synthesis of 1 was
optimized considering data from literature11-13 which were
included in sequential and parallel strategies.14 Table 1
demonstrates that under optimized conditions a yield of 100%
for 10 mM acceptor substrate could be obtained. The space-
time yield and cycle number of UDP-Glc were increased by
2-fold. The productivity of GalT and SuSy (g of 1/U of enzyme)
could be increased by 4-5-fold.
LacNAc Synthesis with Repetitive Batch Technique. In
an attempt to increase enzyme productivity further, we tried to
reuse the native enzymes in subsequent batches. In the first
batch the yield of 1 was 95% after 12 h. After separation of
the enzymes from the product solution by diafiltration, new
substrates were added to the enzyme solution. The second batch
synthesis resulted in only 6.2% yield after 8 h. Detailed
investigations revealed that the epimerase was no longer active
in the second batch. Addition of fresh epimerase increased the
yield of 1 to 67.1% after 14 h, demonstrating that Susy and
GalT were still active. The repetitive addition of UDP-Glc 4′-
epimerase ($6.3/U) appeared too costly and inefficient for large-
scale synthesis. We therefore investigated the stability of
epimerase in more detail.
by noncovalent forces.13,15 Epimerization at C-4 occurs via a
UDP 4-ketoglucose transition state with concomitant formation
of NADH. The cofactor NAD+ is then regenerated intramo-
lecular by reduction of the C-4 oxo group. The enzyme suffers
from reductive inactivation in the presence of uridine nucleotides
(UMP) and different sugars as well as in the presence of UDP-
Glc and UDP-Gal.15,16 Binding of UMP increases the reactivity
of NAD+ toward reducing sugars by an induced conformational
change of the epimerase. The oxidized sugar quickly leaves
the enzyme, and the enzyme-bound NADH cannot be regener-
ated.15,16
During synthesis of 1, UMP is formed as a byproduct by the
Mn2+-catalyzed decomposition of UDP-Glc and UDP-Gal.17 We
found that in the presence of 0.1 mM UMP and 500 mM sucrose
in 200 mM Hepes-NaOH (pH 7.2) the epimerase activity is
drastically reduced after 8 h (15% residual activity). Buffer,
sucrose, or UMP alone do not affect epimerase activity.
Since UMP is certainly also a byproduct in previously
published LacNAc cycles,5 we investigated whether epimerase
inactivation also occurs in the presence of UMP and monosac-
charides. Figure 2 clearly demonstrates that glucose derivatives,
which can serve as acceptors for GalT, inactivate epimerase in
the presence of UMP. With GlcNAc and its derivates the
epimerase was not inactivated after 7 h. In the presence of
UDP-Glc or UDP-Gal alone, 40% of the epimerase activity was
lost after 24 h.
Stability of UDP-Glc 4′-Epimerase. The reason for the
rapid loss of UDP-Glc 4′-epimerase activity is a “suicide”
mechanism by which the enzyme is inactivated.
UDP-Glc 4′-epimerase from yeast or Escherichia coli is a
homodimeric enzyme with one molecule NAD+ per native
enzyme molecule. The cofactor is tightly bound to the enzyme
(11) Khatra, B. S.; Herries, D. G.; Brew, K. Eur. J. Biochem. 1974, 44,
537.
(12) Auge´, C.; Mathieu, C.; Me´rienne, C. Carbohydr. Res. 1986, 151,
147-156.
(13) Fukusawa, T.; Obonai, K.; Segawa, T.; Nogi, Y. J. Biol. Chem.
1980, 255, 2705-2707.
(14) Weuster-Botz, D.; Wandrey, C. Process Biochem. 1995, 563-571.
For enzyme reaction optimization using a parallel strategy, the computer
program GALOP (Genetic Algorithm for the Optimization of Processes)
was kindly provided by Dr. Weuster-Botz, Institute of Biotechnology,
Research Center Ju¨lich.-
In conclusion, the epimerase is inactivated in LacNAc cycles
where UMP and Glc or derivatives thereof as well as UDP-Glc
(15) Frey, P. A. In Pyridine Nucleotide Coenzymes: Chemical, Bio-
chemical and Medical Aspects; Dolphin, D., et al., Eds.; Wiley: New York,
1987; Vol. 2B, pp 462-511.
(16) Nelestuen, G. L.; Kirkwood, S. J. Biol. Chem. 1971, 4, 7533-7543.
(17) Nunez, H. A.; Barker, R. Biochemistry 1976, 15, 3843-3847.

1838 J. Am. Chem. Soc., Vol. 118, No. 8, 1996
ZerVosen and Elling
Table 2. Comparision of Parallel and Repetitive Batches for
LacNAc Synthesis
parallel batches repetitive batches
batches/repetitions/volume (mL)
GalT (U)
epimerase (U)
11/0/25
13.75
55
1/10/25
1.25
5
SuSy
110
100
6.1
10
10
57.4
2.2
average yield (%)
space-time yield (g L-1 day-1
cycle no.
)
Figure 3. Preparative synthesis of 1 with repetitive batch technique
(11 repeated batches) and with reactivation of epimerase using dUDP
6-deoxy-D-xylo-4-hexulose.
5.7
LacNAc synthesized (g)
g of LacNAc/U of GalT
g of LacNAc/U of epimerase
g of LacNAc/U of SuSy
estimated cost $/g of LacNAc
1.05
0.077
0.019
0.004
4124
0.594
0.475
0.119
0.059
712
and UDP-Gal are present. Since sucrose synthase still contains
0.05% invertase activity, Glc is probably also present in the
three enzyme reaction cycle.
Reactivation of UDP-Glc 4′-epimerase can be achieved by
incubation of reductively inactivated epimerase with transition
state analogues like dTDP 6-deoxy-D-xylo-4-hexulose,16,18 myo-
inosose-2, or 2-ketoglucose.19
We have enzymatically synthesized dUDP and dTDP-6-
deoxy-D-xylo-4-hexulose on a preparative scale (0.1-0.6 g) from
dUMP20 and dTDP glucose21 and used them for reactivation
and stabilization of UDP-Glc 4′-epimerase. Under inactivating
conditions (50 mM Gal, 0.1 mM UMP, 3% residual epimerase
activity) the addition of the transition state analogue (0.1 mM
or 1 mM dUDP 6-deoxy-D-xylo-4-hexulose) induces a quick
reactivation (full activity within 1 h) and a prolonged enzyme
stability over 1-5 days, respectively. With this improved
epimerase stability it is now possible to synthesize lactose
analogues and analogues of 1 with reuse of all enzymes and
improved enzyme productivities.22
Preparative LacNAc Synthesis. With the improved epi-
merase stability we used the three-enzyme reaction cycle for
preparative synthesis of 1 by applying the repetitive-batch-
technique (Figure 3). In 11 batches (11 days) 594 mg (1.55
mmol) of 1 was produced from 2.7 mmol of GlcNAc using 5
U of epimerase, 10 U of SuSy, and 1.25 U of GalT. For the
reactivation of epimerase, 0.027 mmol (14.3 mg) of dUDP
6-deoxy-D-xylo-4-hexulose was consumed. The addition of this
transition state analogue had no influence on the activity of
SuSy, GalT, and other glycosyltransferases. The overall average
yield of 1 was 57.4%. The drop in yield after 3 days is probably
due to the repetitive filtration of the enzyme solution. Table 2
compares the effectiveness of 11 parallel batches (based on
optimized conditions in Table 1) versus one batch with 10
repetitions (sum: 11 batches) as performed in our experiment.
The productivities of all enzymes are significantly increased
when the repetitive batch technique is used. These figures reveal
that for the production of 1 g of 1, 13.1 U of GalT is needed in
parallel batches and only 2.1 U of GalT in the repetitive batch
mode. Although the cycle number, space-time yield, and
Figure 4. Multi-enzyme system for the synthesis of 1 with (1) sucrose
synthase, (2) Gal-l-phosphate uridyltransferase, (3) â-1,4-galactosyl-
transferase, (4) UDP-Glc pyrophosphorylase, (5) pyruvate kinase, (6)
galactokinase, and (7) pyrophosphatase.
average yield are lower, the product costs are reduced by 6-fold
based on consumed chemicals, GalT and epimerase.
Product isolation started with the cleavage of excess sucrose
by invertase [EC 3.2.1.26] from yeast. The protein was
separated from the product solution by ultrafiltration. In order
to destroy traces of LacNAc-cleaving activities the commercial
invertase preparation was preincubated for 2 h at 45 °C.
D-Fructose and most of GlcNAc and Glc could be separated
from the product by chromatography using AG50W-X8 in Ca2+
form. After separation from buffer and polar substances by
anion exchange, 1 was further purified by gel filtration (356
mg of 1, overall yield 34.4%).
Synthesis of Neu5Acr2,6Galâl,4GlcNAc. On a smaller
scale, a solution of 1 obtained after separation of the enzymes
by ultrafiltration was directly used for the synthesis of Neu5AcR2,-
6Galâl,4GlcNAc with R-2,6-sialyltransferase [EC 2.4.99.1] from
rat liver.23a,b The CMP formed was cleaved with calf intestinal
alkaline phosphatase [EC 3.1.3.1].24 The product formation was
analyzed with HPLC. Neu5AcR2,6Galâl,4GlcNAc (7.36 mg)
was isolated (overall yield 64.3%). By this combination the
laborious isolation of 1 can be avoided.25
(18) Adair, W. L.; Gabriel, O.; Ullrey, D.; Kalckar, H. M. J. J. Biol.
Chem. 1973, 248, 4635-4639.
(19) Kang, U. G.; Nolan, L. D.; Frey, P. A. J. Biol. Chem. 1975, 250,
7099-7105.
(20) Zervosen, A.; Stein, A.; Adrian, H.; Elling, L. Tetrahedron 1996,
in press. Preparation of dUDP-6-deoxy-D-xylo-4-hexulose: 10 mM dUDP-
Glc was incubated with 0.7 U/mL dTDP-glucose 4,6-dehydratase in 200
mM Hepes-NaOH (pH 7.2, 1 mM DTT, 500 mM sucrose, 25 mM KCl,
0.1% BSA) 4 h at 30 °C. The reaction was followed by HPLC and stopped
by incubation for 5 min at 95 °C. A 250 µL sample of the solution was
directly used for the reactivation of epimerase in repetitive batch synthesis
of 1.
Replacement of UDP Glucose 4′-Epimerase by Galactose
l-Phosphate Uridyltransferase. An alternative way to over-
come the obvious disadvantages of epimerase utilization is to
replace the enzyme by Gal-l-phosphate uridyltransferase [EC
2.7.7.12] (Figure 4). By starting from Gal-l-phosphate omitting
(21) Stein, A.; Kula, M.-R.; Elling, L.; Verseck, S.; Klaffke, W. Angew.
Chem., Int. Ed. Engl. 1995, 34, 1748-1749.
(22) We have already tested Glc, 5-thio-D-Glc, O-n-octyl-â-D-glucopy-
ranoside, S-n-octyl-â-D-thioglucopyranoside (all in presence of R-lactalbu-
min), and O-(6-aminohexyl)-2-acetamido-2-deoxy-â-D-glucopyranoside as
acceptor substrates of GalT on an analytical scale by HPLC and TLC and
proved the applicability of the reactivation of UDP-Glc 4′-epimerase.
(23) (a) Sabesan, S.; Paulson, J. C. J. Am. Chem. Soc. 1986, 108, 2068-
2080. (b) Unverzagt, C.; Kelm, S.; Paulson, J. C. Carbohydr. Res. 1994,
251, 285-301.
(24) Unverzagt, C.; Kunz, H.; Paulson, J. C. J. Am. Chem. Soc. 1990,
112, 9308-9309.
(25) Work is in progress in our laboratory to combine the novel three-
enzyme reaction cycle with in situ regeneration of CMP-Neu5Ac.

Reaction Cycle for Synthesis of N-Acetyllactosamine
J. Am. Chem. Soc., Vol. 118, No. 8, 1996 1839
under optimized conditions in a 1 mL scale. After 12 h, 750 µL of the
product solution was separated from the enzymes by ultrafiltration using
a centricon 10 cartridge (cutoff 10 kDa, Amicon, Berverly U.S.A.) and
analyzed by HPLC. After diafiltration of the enzyme solution with 3
× 250 µL of buffer (without BSA), a fresh substrate solution containing
UDP-Glc, GlcNAc, MnCl2, and sucrose was added. After adjustment
of the volume to 1 mL with buffer, the second batch was incubated for
8 h and analyzed again by HPLC. To prove that UDP-Glc 4′-epimerase
was not active in the second batch, 0.18 U of epimerase was added
and incubation continued for further 14 h.
Stability of UDP-Glc 4′-Epimerase. Photometrical Enzyme
Assay. The reaction mixture consisted of 893 µL of 0.1 M glycine-
NaOH buffer (pH 8.8), 20 µL of 5 mM UDP-Gal, 20 µL of 50 mM
NAD+, and 33.3 µL of UDP-Glc dehydrogenase (2 U/mL). The
reaction was started by the addition of 33.3 µL epimerase diluted in
buffer. The initial reaction rate was followed by the increase of
absorption at 340 nm (DE/min) at 25 °C. One unit is the amount of
enzyme which forms 1 µmol of UDP-Glc/min.
Inactivation of UDP-Glc 4′-Epimerase in the Presence of Sucrose
and UMP. Epimerase (0.13 mg (0.13 U)) was incubated in 500 µL
of buffer A (200 mM Hepes-NaOH (pH 7.2), 1 mM DTT, 25 mM
KCl, 0.1% BSA), in buffer A with 0.1 mM UMP, in buffer A with
500 mM sucrose, and in buffer A with 500 mM sucrose and 0.1 mM
UMP, respectively. After 8 h of incubation at 30 °C, the epimerase
activity was determined and expressed as residual activity (100%
residual activity at t ) 0).
Inactivation of UDP-Glc 4′-Epimerase in the Presence of Nucle-
otide Sugars and Different Acceptor Substrates of â-1,4-GalT.
Epimerase (0.13 mg (0.13 U)) was incubated at 30 °C in 500 µL of
buffer A containing 0.1 mM UMP with 50 mM GlcNAc, 50 mM
2-deoxy-D-Glc, 50 mM D-Glc, 50 mM 5-thio-D-Glc, 50 mM O-n-octyl-
â-D-glucopyranoside, and 10 mM O-(6-aminohexyl)-2-acetamido-2-
deoxy-â-D-glucopyranoside, respectively. The epimerase activity was
determined after various incubation periods and expressed as residual
activity (100% residual activity at t ) 0).
Reactivation of Reductively Inactivated UDP-Glc 4′-Epimerase.
Epimerase (0.25 mg/mL (0.25 U/mL)) was inactivated in the presence
of 50 mM Gal and 0.1 mM UMP at 30 °C in 2 h, yielding 3% residual
activity. Inactivated epimerase (160 µL) was incubated with 40 µL of
0.1 or 1 mM dUDP and dTDP6-deoxy-D-xylo-4-hexulose, respectively.
dUDP and dTDP 6-deoxy-D-xylo-4-hexulose were prepared from
dUMP- or dUDP-Glc20 and dTDP-Glc9,21 by the combination of sucrose
synthase and recombinant dTDP glucose 4,6-dehydratase [EC 4.2.1.46].
Preparative Synthesis of LacNAc. The repetitive batch technique
(11 repeated batches) was also applied for the synthesis of 1 in a
preparative scale. Each batch (10 with a final volume of 25 mL and
one with 20 mL) contained 1 mM UDP-Glc, 10 mM GlcNAc, 1 mM
MnCl2, 500 mM sucrose, and 0.1 mM dUDP 6-deoxy-D-xylo-4-hexulose
and 0.1% (w/v) BSA in 200 mM Hepes-NaOH (pH 7.2, 1 mM DTT,
25 mM KCl). Starting with the first batch, the solution (25 mL) was
incubated with 1.25 U of GalT (0.05 U/mL), 5 U of epimerase (0.2
U/mL), and 10 U of SuSy (0.4 U/mL) at 30 °C. After 22.5 h, the
product and nonreacted substrates were separated from the enzymes
by ultrafiltration in a stirred ultrafiltration cell (Amicon, model 8050,
equipped with a membrane YM 10, cutoff 10 kDa). The solution was
brought to a volume of 50 mL with buffer and concentrated to 25 mL.
Filtration was repeated three times with 20 mL buffer for the last step.
The second batch (25 mL) was started by addition of concentrated
substrates in 5 mL buffer solution. After sterile filtration the second
batch was incubated at 30 °C for 24 h. The synthesis consisted of 10
batches with a 25 mL and 1 batch with a 20 mL reaction volume
employing incubation times between 21 and 30 h. Each batch was
followed by HPLC to determine the yield of each batch. Overall, 183
mg of UDP-Glc (0.3 mmol), 597 mg of GlcNAc (2.7 mmol), 46.2 mg
of MnCl2, 46.2 g of sucrose (135 mmol), 14.3 mg of dUDP 6-deoxy-
D-xylo-4-hexulose (0.025 mmol), 25 mg of BSA, 1.25 U of GalT, 5 U
of epimerase, and 10 U of SuSy were used. Overall 594 mg of 1 (1.55
mmol) was synthesized from 2.7 mmol of GlcNAc (yield 57.4%).
Product isolation was started by enzymatic cleavage of sucrose.
Invertase (25 000 U/mL) was preincubated in buffer (200 mM Hepes-
NaOH (pH 7.2, 500 mM sucrose, 1 mM DTT, 0.1% BSA, 25 mM
KCl)) 2 h at 45 °C. The reaction was started by addition of 10 µL of
Glc-l-phosphate recycling, 1 was synthesized in 63% yield after
24 h. However, recycling of Glc-l-phosphate to UDP-Glc and
ADP to ATP as well as PEP consumption in both cases makes
this cycle more complex.
Conclusion
In summary, we have developed a novel three-enzyme
reaction cycle with in situ regeneration of UDP-Glc and UDP-
Gal for the synthesis of 1. We have clearly demonstrated that
reductive inactivation of UDP-Glc 4′-epimerase is a serious
problem in all cycles where at least UDP-Glc/UDP-Gal or UMP
and glucose derivatives as acceptor substrates of GalT are
present. The reactivation of UDP-Glc 4′-epimerase by dNDP
6-deoxy-D-xylo-4-hexulose and the utilization of the repetitive
batch technique enabled us to increase the productivity of all
three enzymes. Further work is in progress in our laboratory
to exploit the novel LacNAc cycle for the synthesis of 1, as
well as analogues of lactose and 1, and to combine it with other
glycosyltransferases employing in situ regeneration cycles for
CMP-NeuAc and GDP-Fuc.
Experimental Section
Materials. Sucrose synthase (SuSy) was purified from rice grains.10
CMP-Neu5Ac was kindly provided by U. Kragl. â-1,4-GalT [EC
2.4.1.38] from human milk, calf intestinal alkaline phosphatase [EC
3.1.3.1], and R-2,6-sialyltransferase [EC 2.4.99.1] from rat liver were
from Boehringer Mannheim (Mannheim, Germany). UDP-Glc 4′-
epimerase [EC 5.1.3.2] from Saccharomyces cereVisiae, Gal-l-phosphate
uridyltransferase [EC 2.7.7.12] from yeast, invertase [EC 3.2.1.26] from
yeast, bovine serum albumine BSA, nucleotides, nucleotide sugars,
GlcNAc, and O-n-octyl-â-D-glucopyranoside were supplied by Sigma
(Deisenhofen, Germany). Hepes and NAD+ were from Biomol
(Hamburg, Germany). S-n-Octyl-â-D-thioglucopyranoside was from
Aldrich (Steinheim, Germany). O-(6-Aminohexyl)-2-acetamido-2-
deoxy-â-D-glucopyranoside was a gift from the Hoechst AG (Frankfurt
am Main, Germany). 5-Thio-D-Glc, 2-deoxy-D-Glc, p-anisaldehyde,
and dithiothreitol (DTT) were from Fluka (Neu-Ulm, Germany). All
other chemicals were purchased from Merck (Darmstadt, Germany).
Analytical Methods. The activity of SuSy was determined for the
cleavage reaction with UDP.10 One enzyme unit (U) corresponds to
the formation of 1 µmol of UDP-Glc/min using standard conditions.
Optimization of the LacNAc Synthesis. The concentrations of 1
and GlcNAc were analyzed by HPLC using two connected Aminex
HPX-87H columns (300 × 7.8 mm each, Biorad, Mu¨nchen, Germany).
Compounds were separated using 4 mM H2SO4 as the eluent at a flow
rate of 0.55 mL/min at 65 °C and quantified by UV detection at 205
nm.
Starting conditions for synthesis of 1: 1 mM UDP-Glc, 1 mM
MnCl2, 5 mM GlcNAc, 500 mM sucrose, 0.1 U/mL GalT, 0.2 U/mL
epimerase, and 0.8 U/mL SuSy were incubated for 24 h in 200 mM
Hepes-NaOH (pH 7.2, 25 mM KCl, 0.01% BSA, 1 mM DTT) at 30
°C.
Because of the high stability of SuSy in the buffer we selected 200
mM Hepes-NaOH (pH 7.2) for synthesis of l including 25 mM KCl,
0.01% BSA which activates GalT,11 and 1 mM dithiothreitol (DTT)
which prevents the oxidation of SH groups of cysteine in the
epimerase.12,13 The experiments were carried out with native enzymes
at 30 °C.
In optimization experiments,14 the pH and the concentrations of
GlcNAc, MnCl2, UDP-Glc, SuSy, GalT, epimerase, and BSA were
varied.
Optimized conditions for synthesis of 1: 1 mM UDP-Glc, 1 mM
MnCl2, 10 mM GlcNAc, 500 mM sucrose, 0.05 U/mL GalT, 0.2 U/mL
epimerase, and 0.4 U/mL SuSy were incubated for 15 h in 200 mM
Hepes-NaOH (pH 7.2, 25 mM KCl, 0.1% BSA, 1 mM DTT) at 30 °C.
LacNAc Synthesis with Repetitive Batch Technique. To test the
repetitive use of the native enzymes,26 synthesis of 1 was carried out
(26) Kragl, U.; Go¨dde, A.; Wandrey, C.; Kinzy, W.; Cappon, J. J.;
Lugtenburg, J. Tetrahedron: Asymmetry 1993, 4, 1193-1202.

1840 J. Am. Chem. Soc., Vol. 118, No. 8, 1996
ZerVosen and Elling
invertase solution/mL of product solution at 45 °C. Cleavage of sucrose
was followed by a polarimeter (241 polarimeter, Perkin-Elmer). After
2 h the invertase was separated from the product solution by
ultrafiltration (see above). In order to facilitate product isolation the
final solution was divided into five batches. Each batch was concen-
trated at 20-25 mbar and 30-35 °C to a volume of 30 mL. After
chromatography on AG50W-X8 (Ca2+ form, column 5 × 35 cm, 0.5
mL/min bidistilled water as the eluent), 1 was further purified on an
anion-exchange column (2.6 × 26.5 cm, Dowex lx2, 100-200 mesh,
Cl- form). Prior to loading the pH of the solution was brought to 8.5.
The sugars were eluted with bidest. water (flow rate 3.5 mL/min). The
fractions containing 1 were pooled and subsequently adjusted to pH
7.0. The product solution was concentrated at 20-25 mbar and 30-
35 °C to a volume of 9 mL. After a gel filtration on Biogel P2 (2.6 ×
82 cm, 0.5 mL/min bidest. water as the eluent), the resulting product
solutions of all five batches were combined. After repeating the anion-
exchange chromatography and the gel filtration once and subsequent
lyophilization, 462 mg of product containing 77% of 1 (356 mg, 0.93
mmol, overall yield 34.4%), 4.2% GlcNAc, and 18.7% NaCl according
to HPLC analysis was found. For NMR analysis the residual GlcNAc
and salt was separated from 1 (20 mg) by gel filtration. The NMR
data correspond exactly to those previously reported.5,27
R-2,6-sialyltransferase (0.1 U), 6 µL of alkaline phosphatase (6 U), 2
mg of BSA, and 13.7 mg of CMP-Neu5Ac (10 mM) were incubated
at 37 °C for 24 h. The reaction was analyzed by HPLC as described
above for 1. The concentration of CMP-Neu5Ac during the synthesis
was followed by ion pair HPLC.7
The product was isolated from a 1.7 mL solution as described above.
After cleavage of sucrose and ultrafiltration, Neu5AcR2,6Galâl,-
4GlcNAc was separated off by gel filtration on Biogel P2. After
lyophilization, 7.36 mg of product (overall yield 64.3 %) was isolated
and analyzed by HPLC (100% Neu5AcR2,6Galâl,4GlcNAc) and NMR
spectroscopy.
The NMR spectrum corresponds to that previously reported.28
NMR analysis of NeuAcR2,6Galâl,4GlcNAc: 1H-NMR (300 MHz,
3
D2O) δ 5.25 (d, 1 H, GlcNAc, H-1, R, JH-1a,H-2 ) 2.2 Hz), 4.54 (d, 1
H, Gal, H-1, 3JH-1,H-2 ) 8 Hz), 2.74 (dd, 1 H, 3e-H-Neu5Ac, 2JH-3a,H-3e
) 13 Hz, 3JH-3e,H-4 ) 5 Hz), 2.13 (s, 3 H, NAc, GlcNAc), 2.09 (s, 3 H,
2
3
NAc, Neu5Ac), 1.77 (dd, 3a-H-Neu5Ac, JH-3a,H-3e ≈ JH-3a,H-4 ) 12
Hz).
Replacement of UDP-Glc 4′-Epimerase by Galactose l-Phosphate
Uridyltransferase. Galactose 1-phosphate (10 mM), 10 mM GlcNAc,
1 mM UDP-Glc, 1 mM MnCl2, 0.2 U/mL Gal-1-phosphate uridyl-
transferase, 0.4 U/mL SuSy, and 0.05 U/mL GalT were incubated in
200 mM Hepes-NaOH (pH 7.2, 500 mM sucrose, 1 mM DTT, 25 mM
KCl, 0.1% BSA) at 30 °C. The formation of 1 was followed by HPLC.
NMR analysis of LacNAc: 13C-NMR (75 MHz, D2O) δ 177.6 (CdO
R, â), 105.8, 105.7 (Gal, C-1, (R), (â) anomerization of GlcNAc), 97.7
(GlcNAc, C-1, â), 93.4 (GlcNAc, C-1, R), 81.7 (GlcNAc, C-4, R), 81.3
(GlcNAc, C-4, â), 78.2 (Gal, C-5), 77.7 (GlcNAc, C-5, â), 75.4 (Gal,
C-3), 73.8 (Gal, C-2), 73.1 (GlcNAc,C-5, R), 72.1 (GlcNAc, C-3, R,
â), 71.4 (Gal, C-4), 63.8 (Gal, C-6), 62.8 (GlcNAc, C-6, R, â), 59.1
(GlcNAc, C-2, â), 56.6 (GlcNAc, C-2, R), 25.0 (CH3-, â), 24.8 (CH3-,
Acknowledgment. This project is a part of the Ph.D. thesis
from Dipl.-Ing (FH) Dipl.-Chem. A.Z. The authors thank the
Hoechst AG, Frankfurt am Main, Germany, for financial
support, Dr. Udo Kragl (Institute of Biotechnology, Research
Center Ju¨lich) for providing CMP-Neu5Ac, Dr. Weuster-Botz
(Institute of Biotechnology, Research Center Ju¨lich) for provid-
ing the computer program GALOP, and Prof. Dr. M.-R. Kula
for reading the manuscript and helpful discussions.
1
3
R); H-NMR (300 MHz, D2O) δ 5.21 (d, 1 H, GlcNAc, H-1, R, JH-
1a,H-2 ) 2.2 Hz), 4.73 (d, 1 H, GlcNAc, H-1, â, 3JH-1b,H-2 ) 8 Hz), 4.478
3
(d, 1 H, Gal (R), H-1, JH-1,H-2 ) 8 Hz), 4.476 (d, 1 H, Gal (â), H-1,
3JH-1,H-2 ) 8 Hz), 4.03-3.69 (m, 12 H), 2.055 (s, 3 H, CH3-).
Synthesis of Neu5Acr2,6Galâl,4GlcNAc. Synthesis of 1 was
carried out under optimized conditions (see above) on a 5 mL scale.
After ultrafiltration (see above), 2 mL of product solution, 50 µL of
JA953495E
(27) Nunez, H. A.; Barker, R. Biochemistry 1980, 19, 489-495.
(28) Thiem, J.; Treder, W. Angew. Chem. 1986, 98, 1100-1111.