Chemoenzymatic synthesis of CMP-sialic acid derivatives by a one-pot two-enzyme system: Comparison of substrate flexibility of three microbial CMP-sialic acid synthetases

Article abstract of DOI:10.1016/j.bmc.2004.09.030

Three microbial CMP-sialic acid synthetases were cloned from Neisseria meningitidis, Streptococcus agalactiae, and Escherichia coli, respectively. Their activities in the production of CMP-sialic acid analogs were compared by HPLC analysis. The N. meningitidis synthetase was used in the preparative synthesis of eight CMP-sialic acid derivatives in a one-pot two-enzyme system. Three C terminal His₆-tagged recombinant microbial CMP-sialic acid synthetases [EC 2.7.7.43] cloned from Neisseria meningitidis group B, Streptococcus agalactiae serotype V, and Escherichia coli K1, respectively, were evaluated for their ability in the synthesis of CMP-sialic acid derivatives in a one-pot two-enzyme system. In this system, N-acetylmannosamine or mannose analogs were condensed with pyruvate, catalyzed by a recombinant sialic acid aldolase [EC 4.1.3.3] cloned from E. coli K12 to provide sialic acid analogs as substrates for the CMP-sialic acid synthetases. The substrate flexibility and the reaction efficiency of the three recombinant CMP-sialic acid synthetases were compared, first by qualitative screening using thin layer chromatography, and then by quantitative analysis using high performance liquid chromatography. The N. meningitidis synthetase was shown to have the highest expression level, the most flexible substrate specificity, and the highest catalytic efficiency among the three synthetases. Finally, eight sugar nucleotides, including cytidine 5′-monophosphate N-acetylneuraminic acid (CMP-Neu5Ac) and its derivatives with substitutions at carbon-5, carbon-8, or carbon-9 of Neu5Ac, were synthesized in a preparative (100-200 mg) scale from their 5- or 6-carbon sugar precursors using the N. meningitidis synthetase and the aldolase.

Full text of DOI:10.1016/j.bmc.2004.09.030

Bioorganic & Medicinal Chemistry 12 (2004) 6427–6435
Chemoenzymatic synthesis of CMP–sialic acid derivatives by a
one-pot two-enzyme system: comparison of substrate flexibility
of three microbial CMP–sialic acid synthetases
Hai Yu, Hui Yu, Rebekah Karpel and Xi Chen*
Department of Chemistry, University of California, One Shields Avenue, Davis, CA 95616, USA
Received 2 August 2004; accepted 16 September 2004
Abstract—Three C terminal His6-tagged recombinant microbial CMP–sialic acid synthetases [EC 2.7.7.43] cloned from Neisseria
meningitidis group B, Streptococcus agalactiae serotype V, and Escherichia coli K1, respectively, were evaluated for their ability
in the synthesis of CMP–sialic acid derivatives in a one-pot two-enzyme system. In this system, N-acetylmannosamine or mannose
analogs were condensed with pyruvate, catalyzed by a recombinant sialic acid aldolase [EC 4.1.3.3] cloned from E. coli K12 to pro-
vide sialic acid analogs as substrates for the CMP–sialic acid synthetases. The substrate flexibility and the reaction efficiency of the
three recombinant CMP–sialic acid synthetases were compared, first by qualitative screening using thin layer chromatography, and
then by quantitative analysis using high performance liquid chromatography. The N. meningitidis synthetase was shown to have the
highest expression level, the most flexible substrate specificity, and the highest catalytic efficiency among the three synthetases.
Finally, eight sugar nucleotides, including cytidine 50-monophosphate N-acetylneuraminic acid (CMP–Neu5Ac) and its derivatives
with substitutions at carbon-5, carbon-8, or carbon-9 of Neu5Ac, were synthesized in a preparative (100–200mg) scale from their
5- or 6-carbon sugar precursors using the N. meningitidis synthetase and the aldolase.
ꢀ 2004 Elsevier Ltd. All rights reserved.
1. Introduction
host cell surface carbohydrates and their presence could
account for the ability of the bacteria to evade the host
immune defense mechanisms.3–5
Sialic acids are negatively charged, 9-carbon sugars that
have been predominantly found in vertebrates, a few
higher invertebrates, and certain types of bacteria.1 They
constitute a remarkable diverse family of more than 50
structurally distinct members. In vertebrates, sialic acids
are primarily expressed on the outermost end of carbohy-
drate structures on cell surface glycoproteins and glyco-
lipids or exist as homopolymers of N-acetylneuraminic
acid (Neu5Ac) in a2,8 linkages in mammalian brain tis-
sues.2 Sialic acid containing structures in eukaryotic sys-
tems play important roles in a variety of physiological
and pathological processes, including cell–cell inter-
actions, inflammation, fertilization, viral infection, differ-
entiation, malignancies, and cell signaling etc.1,2 In
bacteria, sialic acids are found as components of extracel-
lular capsular polysaccharides and lipooligosaccharides
(LOS). These structures are believed to mimic sialylated
Glycoconjugates containing either naturally occurring
or structurally modified sialic acids are invaluable tools
in understanding important biological and physiological
properties of sialylated structures. Enzymatic glycosyl-
ation employing sialyltransferases overcomes tedious
and expensive synthetic scheme of chemical synthe-
sis and produces regio- and stereospecific sialates in a
short and economic way. Thus, it has become one of
the most practical and efficient methods for production
of complex sialates and derivatives. Sialyltransferase-
based synthesis requires activation of free sialic acids
or their synthetic analogs to form cytidine 50-monophos-
phate-sialic acid (CMP–sialic acid) analogs.6 Biosyn-
thetically, this monosaccharide activation process is
catalyzed by a class of enzymes named CMP–sialic acid
synthetase (CSS, or sialic acid cytidylyltransferase, EC
2.7.7.43; Scheme 1, Reaction B).7
Keywords: Aldolase; CMP–sialic acid synthetase; CMP–sialic acid;
Enzymatic synthesis.
CSS has previously been isolated from eukaryotic
sources and bacterial cultures and has been used in the
synthesis of CMP–sialic acids.8–15 The enzyme has also
*
Corresponding author. Tel.: +1 530 754 6037; fax: +1 530 752
8995; e-mail: chen@chem.ucdavis.edu
0968-0896/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2004.09.030

6428
H. Yu et al. / Bioorg. Med. Chem. 12 (2004) 6427–6435
Reaction A
Reaction B
Pyruvate
CTP
PPi
R1
O
R2
OH
OH
COO
OH
OCMP
OH
OH
Mg2+
_
_
HO
HO
R2
R2
COO
O
O
R1
R1
CSS
Aldolase
Pyruvate
HO
HO
HO
1
2
3
a. R1 = NH(CO)CH3, R2 = CH2OH
e. R1 = N3, R2 = CH2OH
b. R1 = NH(CO)CH2OH, R2 = CH2OH f. R1 = OH, R2 = CH2OH
c. R1 = NH(CO)CH2N3, R2 = CH2OH
d. R1 = NHCbzGly, R2 = CH2OH
g. R1 = OH, R2 = CH2N3
h. R1 = OH, R2 = H
Scheme 1. One-pot two-enzyme synthesis of CMP–sialic acid derivatives (3a–h) from ManNAc or mannose derivatives (1a–h) with a recombinant
sialic acid aldolase cloned from E. coli K12 and a recombinant microbial CMP–sialic acid synthetase.
been cloned from various eukaryotic and prokaryotic
sources.16–25 The crystal structures of the N-terminal
catalytically active domain of a mouse CSS,26 a full
length Neisseria meningitidis CSS,27 and a related
CMP–KDO synthetase of Escherichia coli28,29 have been
reported. Recombinant CMP–sialic acid synthetases
cloned from bacteria sources are effective in obtaining
natural and unnatural CMP–sialic acid analogs. The
recombinant CSS from E. coli K1 is the first commer-
cially available CSS. This enzyme, however, has narrow
substrate specificity. The enzyme has been found to
accept only a few C-9 substituted derivatives of
Neu5Ac.30,31 Recently, high level expression of N. men-
ingitidis CSS has been reported32,33 and shown to have a
broad substrate specificity.35,36 Both E. coli K1 and N.
meningitidis enzymes have been applied in the enzymatic
synthesis of CMP–sialic acid and sialic acid-containing
glycoconjugates.30–40 However, no direct comparison
of the substrate specificity of these two enzymes has
been reported. CSS has also been isolated41 and cloned19
from Streptococcus agalactiae (group B streptococci),
but its substrate specificity and application in the syn-
thesis of CMP–sialic acid analogs have not been
investigated.
from E. coli K12 in the same reaction mixture. Among
the three recombinant CSSs, the N. meningitidis synthe-
tase was shown to have the highest expression level, the
most flexible substrate specificity, and the highest cata-
lytic efficiency as indicated by quantitative high per-
formance liquid chromatography (HPLC) analysis. A
series of structural analogs of CMP–Neu5Ac were then
synthesized on a preparative scale from ManNAc/Man
analogs, sodium pyruvate, and CTP using the N. menin-
gitidis CSS and the recombinant sialic acid aldolase.
2. Results and discussion
2.1. Cloning, expression, and purification of CMP–sialic
acid synthetases
Three C-terminal hexahistidine-tagged full length CSSs
were cloned in pET vectors (Novagen, Madison, WI)
from N. meningitidis group B, S. agalactiae serotype
V, and E. coli K1, respectively. The expression of the en-
zymes in E. coli BL21 (DE3) cells was induced with
0.1mM isopropyl-1-thio-b-D-galactopyranoside (IPTG)
and carried out in LB rich medium at 20ꢁC for 20h with
vigorous shaking (250rpm). The recombinant enzymes
were purified using an affinity column packed with nickel–
nitrilotriacetic acid–agarose (Ni2+–NTA–agarose) resin,
which bind specifically to the C-terminal hexahistidine
tag in the recombinant proteins. Sodium dodecylsul-
fate-polyacrylamide gel electrophoresis (SDS-PAGE)
data (Fig. 1) indicated that high level expressions
were achieved for both N. meningitidis and S. agalactiae
CSSs with the recombinant CSS presented more than
85% of whole cell proteins (lane 3 in Fig. 1A and B).
More interestingly, the N. meningitidis CSS was very
Here, we report a study on direct comparison of the
ability and efficiency of three recombinant microbial
CSSs, cloned from N. meningitidis group B, S. agalactiae
serotype V, and E. coli K1, respectively, in the synthesis
of CMP–sialic acid analogs in a one-pot two-enzyme
reaction system. In this system, the substrates (sialic acid
analogs) of the CSS were obtained by condensation
reactions of N-acetylmannosamine (ManNAc) or man-
nose (Man) analogs and an excess amount (5equiv) of
pyruvate using a recombinant sialic acid aldolase cloned
Figure 1. SDS-PAGE (12% Tris–glycine gel) for CSS. Lanes: 1, protein standards; 2, whole cells, before induction; 3, whole cells, after induction; 4,
lysate, after induction; 5, CSS purified by Ni2+–NTA column.

H. Yu et al. / Bioorg. Med. Chem. 12 (2004) 6427–6435
6429
soluble and existed in more than 85% of the lysate pro-
teins (lane 4 in Fig. 1A), the soluble portion of total host
proteins. However, the S. agalactiae CSS presented only
about 25% of the lysate proteins (lane 4 in Fig. 1B). The
expression level of the E. coli CSS (lane 3 in Fig. 1C) was
much lower than that of the N. meningitidis or the S.
agalactiae CSS. Nevertheless, a similar percentage of
E. coli (lane 4 in Fig. 1C) and S. agalactiae (lane 4 in
Fig. 1B) CSSs were found in the lysate samples. Overall,
the highest expression level of the soluble protein was
achieved for the N. meningitidis CSS. The SDS-PAGE
also indicated that one step Ni2+–NTA column purifica-
tion was very efficient to provide CSSs in over 95% pur-
ity (lane 5 in Fig. 1A–C).
used for all assays and syntheses. The optimal reaction
time was 2h. Elongating the incubation time resulted
in the breakdown of the CMP–sialic acid products.
2.3. High performance liquid chromatography (HPLC)
analysis for substrate specificity
HPLC analysis was carried out to quantitatively com-
pare the substrate flexibility and reaction efficiency of
the N. meningitidis, S. agalactiae, and E. coli CSSs.
Reactions were carried out with CTP (10mM), sodium
pyruvate (50mM), and ManNAc/Man analogs
(10mM) in a Tris–HCl (100mM, pH8.8) buffer contain-
ing 20mM of MgCl2. The reaction was stopped at 2h by
adding ice cold acetonitrile.
Unlike the N. meningitidis CSS, which contains only 228
amino acid (aa) residues with a molecular weight of
about 25kDa, both S. agalactiae (413 aa, 46kDa) and
E. coli K1 (419 aa, 49kDa) CSSs are longer proteins.
A recent study indicates that only the N-terminal amino
acid residues (1-229 aa) of the E. coli CSS are essential
for CSS activity.42 Due to a high sequence homology
found in the N. meningitidis CSS, and the N-terminus
of the S. agalactiae and E. coli CSSs, only the N-termi-
nus of the S. agalactiase CSS is presumably required for
the synthetase activity. The lower expression level of the
E. coli CSS and the lower proportion of the S. agalactiae
CSS in the lysate compared to those of the N. meningit-
idis CSS indicate that the C-terminal peptide sequences
may interfere with the protein expression and solubility.
Various assay methods have been previously reported
for CMP–Neu5Ac synthetase. The most frequently used
method is the colorimetric periodate-thiobarbituric acid
assay developed by Kean and Roseman.44 However, this
method is not suitable for Neu5Ac derivatives that do
not participate in the periodate–thiobarbituric acid reac-
tion.45 An improved ion-pair high-performance liquid
chromatographic (HPLC) method with gradient elu-
tion46 overcomes the limitations of the colorimetric
method and provides quantitative data. Thus, it was
chosen for our studies. CMP–sialic acid analogs were
efficiently separated from CTP, CMP, and CDP within
35min using methanol-pH gradient elution with tetra-
butylammonium hydrogensulfate as a pairing agent in
a Premier C18 reversed-phase column (5l) monitored
by a UV detector at 270nm.
2.2. Thin layer chromatography (TLC) analysis for
substrate specificity assays
More than 50 naturally occurring sialic acid derivatives
were described.2 The most abundant forms found in ani-
mal glycans were N-acetylneuraminic acid (Neu5Ac)
and N-glycolylneuraminic acid (Neu5Gc), while deamino-
neuraminc acid (KDN) was frequently found in the
lipooligosaccharide (LOS) of various Gram-negative
bacteria.1,2 In agreement with previous reports,34,35 the
N. meningitidis CSS had a broad substrate specificity.
It accepted all the compounds we tested as the sub-
strates, including an eight-carbon sugar 4,6-bis-epi-
KDO (2h), and produced activated CMP–sialic acid
derivatives in high yields (Table 1). In comparison, the
E. coli K1 CSS only recognized the natural substrate
Neu5Ac (N-acetylneuraminic acid, 2a) and its hydroxyl-
and azido-derivatives of the N-acetyl group linked to
carbon 5 (N-glycolylneuraminic acid, Neu5Gc, 2b; and
N-azidoacetylneuraminic acid, Neu5AcN3, 2c). The
activity of the E. coli enzyme was also less effective for
modified substrates. Only about 57–82% yields were
achieved with the same molar concentration of the pro-
tein. The substrate specificity of the S. agalactiae sero-
type V CSS was more narrow than that of the N.
meningitidis CSS but was more flexible than that of the
E. coli synthetase. It showed very good activities for
Neu5Ac and its derivatives, and produced CMP acti-
vated Neu5Ac (2a), Neu5Gc (2b), Neu5AcN3 (2c), and
Neu5CbzGly (N-Cbzglycine neuraminic acid, 2d) in
high yields. It also accepted a KDN derivative KDN9N3
(9-azido-9-deoxy-KDN, 2g) and synthesized CMP–
KDN9N3 (3g) in a relatively good yield (80%).
D-N-Acetylmannosamine (ManNAc) and mannose
(Man) analogs were either obtained from commercial
sources or synthesized using known methods as de-
scribed in the experimental section. These ManNAc/
Man derivatives were used in TLC analysis to test the
substrate flexibility of a recombinant sialic acid aldolase
cloned from E. coli K12. Sialic acid aldolase catalyzes
the reversible condensation of pyruvate and ManNAc
to form Neu5Ac with the equilibrium favoring the aldol
cleavage (Scheme 1, Reaction A).43 In agreement with
the previous reports,43 the aldolase was shown to have
a flexible acceptor specificity. The recombinant E. coli
K12 aldolase was very effective in accepting all the Man-
NAc/Man analogs we tested as the substrates, and pro-
duced corresponding Neu5Ac analogs. Pyruvate was
used in an excess amount (5equiv) to drive the aldolase
reaction toward the formation of the condensation
products. Then, qualitative TLC assays were carried
out as an initial screening step to test the feasibility of
the synthesis of CMP–sialic acid analogs from pyruvate,
CTP, and ManNAc analogs using a combined enzyme
system (Scheme 1) containing both the aldolase and a
microbial CSS cloned from E. coli K1, N. meningitidis,
or S. agalactiae. TLC results provided informative indi-
cation of suitable reaction conditions and reaction
times. The CSSs were shown to require a pH of
8.5–9.0 for best activity. Although the aldolase worked
most efficiently at pH7.5, it also worked well at
pH8.5–9.0. Therefore, a Tris–HCl buffer of pH8.8 was

6430
H. Yu et al. / Bioorg. Med. Chem. 12 (2004) 6427–6435
Table 1. Substrate specificity and catalytic efficiency of three recombinant microbial CMP–sialic acid synthetases in a one-pot two-enzyme reaction
systema
Analogs
1
Staring sugar
3
Product
Overall yield (%) (HPLC analysis)
Synthetic yield (%)
N. meningitidis
serotype B
E. coli
K1
S. agalactiae
serotype V
N. meningitidis
serotype B
OH
OCMP
COONa
O
OH
HO
HO
HN
a
96
98
99
80
O
O
NH
HO
HO
HO
OH
O
OH
OCMP
COONa
O
OH
HO
OH
OH
HO
HN
O
O
b
c
d
e
f
94
95
57
81
85
85
82
88
90
80
80
NH
HO
HO
HO
HO
O
OH
OCMP
COONa
O
OH
HO
N3
HO
HN
O
O
82
90
NH
HO
N3
HO
HO
OH
O
OH
OCMP
COONa
O
OH
HO
CbzHN
NHCbz
OH
HO
HN
O
O
97
<0.1
<0.1
<0.1
<0.1
<0.1
68
NH
HO
HO
HO
O
OH
N3
OCMP
COONa
N3
O
OH
HO
HO
N3
HO
HO
HO
100
99
<0.1
30
O
O
OH
OH
HO
OH
HO
OH
HO
OCMP
COONa
HO
HO
HO
HO
O
OH
OCMP
COONa
OH
N3
HO
HO
HO
O
g
85
80
O
HO
OH
OH
HO
OH
HO
OCMP
COONa
HO
O
h
HO
HO
82
<0.1
HO
O
HO
a The sialic acid analogs, which are the substrates for the CSS, were generated in situ from pyruvate and ManNAc/Man analogs by a sialic acid
aldolase catalyzed condensation reaction.
However, it catalyzed the synthesis of CMP–KDN (2f)
less efficiently with a low yield of 30%. Neu5N3 (5-
azido-5-deoxy-neuraminic acid, 2e) and the eight-carbon
sugar 4,6-bis-epi-KDO (2h) were not acceptable sub-
strates for the enzyme. These results indicated an impor-
tant role of an acylamino group at C-5 of the Neu5Ac in
S. agalactiae CSS activity. The study of the variation of
the substrate specificity of different microbial CSSs may
help us to understand the mechanism of sialic acid
metabolism in bacteria and may help in elucidating the
pathogen host-invading mechanism.
were relatively low for these reactions, with 15% for the
N. meningitidis CSS, and 12% for both E. coli and S. aga-
lactiae CSSs. The low yields may be resulted from the
inhibitory effect of the GlcNAc to the CSS.48
2.4. Preparative synthesis of CMP–Neu5Ac and analogs
Eight compounds, including CMP–Neu5Ac and its
derivatives with substitutions at C-5, C-8, or C-9 at
Neu5Ac, were successfully synthesized in a preparative
scale (100mg scale) using the one-pot two-enzyme sys-
tem with the N. meningitidis CSS and the aldolase.
The products were purified by anion-exchange chroma-
tography and their structures were confirmed by 1H and
13C NMR spectra, as well as High Resolution Mass
Spectra as described in the experimental section. These
sugar nucleotide donors are important intermedi-
ates for future enzymatic synthesis of sialylated
oligosaccharides.
CMP–Neu5Ac can also be synthesized from N-acetyl-D-
glucosamine (GlcNAc) since GlcNAc was able to be epi-
merized to ManNAc chemically under basic condition.47
In this situation, ManNAc was generated in situ from
GlcNAc in a reversible chemical epimerization, and then
reacted with pyruvate and CTP in the one-pot two-
enzyme reaction containing both the aldolase and a
CSS to produce activated Neu5Ac. However, the yields

H. Yu et al. / Bioorg. Med. Chem. 12 (2004) 6427–6435
6431
3. Conclusion
genomic DNA (ATCC#BAA-611D) as a template.
The primers used were: forward primer 50CATG-
CCATGGGCATGAAGCCAATTTGTATTATTCCT-
GCG30 (NcoI restriction site is underlined) and reverse
primer 50CGCGGATCCTTAGTGGTGGTGGTGGTG-
GTGTAAGGTTTTAACTTCGTCTAC30 (BamHI rest-
riction site is underlined, the gene sequence for a hexa-
histindine-tag is in italics). The primers for the E. coli
K1 CSS cloned in pET15b vector were: forward primer
50CTGTCATGAGAACAAAAATTATTGCG30 (BspHI
restriction site is underlined) and reverse primer
50CGCGGATCCTCAGTGGTGGTGGTGGTGGTGTT-
TAACAATCTCCGCTATTTCG30 (BamHI restriction
site is underlined, the gene sequence for a hexahistin-
dine-tag is in italics). Polymerase chain reactions
(PCR) for the amplification of the target genes were per-
formed in a 50lL reaction mixture containing 1lg of
genomic DNA as the template DNA, 1lM each of for-
ward and reverse primers, 5lL of 10 · Herculase buffer,
1mM dNTP mixture, and 5units (1lL) of Herculase
enhanced DNA polymerase. The reaction mixtures were
subjected to 30 cycles of amplifications with an anneal-
ing temperature of 50–60ꢁC. The resulting PCR prod-
ucts were digested with the corresponding restriction
enzymes introduced in the primers, purified, and ligated
with predigested pET15b or pET22b(+) vector. The
ligated product was transformed into electrocompetent
E. coli DH5a cells. Positive plasmids were selected and
subsequently transformed into BL21 (DE3) chemically
competent cells. Selected clones were grown for mini-
preps and characterization by restriction mapping and
DNA sequencing.
The one-pot two-enzyme system containing a recombi-
nant sialic acid aldolase and a microbial CSS, especially
the N. meningitidis CSS, was efficient in producing
CMP–Neu5Ac derivatives from Mannose or ManNAc
analogs. The N. meningitidis synthetase was shown to
have the highest expression level, the most flexible sub-
strate specificity, and the highest catalytic efficiency
among three recombinant synthetases. Together with a
sialic acid aldolase and a sialyltransferase, the S. agalac-
tiae and N. meningitidis CSSs could be invaluable tools
for large-scale synthesis of biologically and medicinally
important sialylated oligosaccharides, polysialic acids,
and glycoconjugates.
4. Experimental
4.1. Materials
Vector plasmids pET15b and pET22b(+) were pur-
chased from Novagen (EMD Biosciences, Inc. Madison,
WI). Ni2+–NTA agarose (nickel–nitrilotriacetic acid–
agarose), QIAprep spin miniprep kit, and QIAEX II
gel extraction kit were from Qiagen (Valencia, CA).
Herculase enhanced DNA polymerase was from Strata-
gene (La Jolla, CA). Electrocompetent E. coli DH5a
cells and chemically competent E. coli BL21 (DE3) cells
were from Invitrogen (Carlsbad, CA). T4 DNA ligase, 1
kb DNA ladder, and restriction enzymes were from Pro-
mega (Madison, WI). BCA protein assay kit was from
Pierce Biotechnology, Inc. (Rockford, IL). D-N-Acetyl-
mannosamine (ManNAc, 1a), D-mannose (1f), and
D-lyxose (1h) were purchased from Sigma (St. Louis,
MO). All other materials were obtained from commer-
cial suppliers and used as received with or without fur-
ther purification.
4.2.2. Expression of the recombinant sialic acid aldolase
and CMP–sialic acid synthetases. All plasmid-bearing
E. coli strains were cultured in LB rich medium (10g/L
tryptone, 5g/L yeast extract, and 10g/L NaCl) supple-
mented with 100lg/mL ampicillin. Generally, overex-
pression of the target protein was achieved by
inducing the E. coli culture with 0.1mM IPTG (isoprop-
yl-1-thio-b-D-galactopyranoside) at OD600 nm = 0.8–1.0
and incubating at 20ꢁC for 20h with vigorous shaking
at 250rpm in a C25KC incubator shaker (New Bruns-
wick Scientific, Edison, NJ).
4.2. Cloning, expression, and purification of a sialic acid
aldolase and three CMP–sialic acid synthetases
4.2.1. Cloning of a sialic acid aldolase and three CMP–
sialic acid synthetases. The C-terminal His6-tagged full
length sialic acid aldolase (NanA) was cloned in pET15b
vector from E. coli K12, substrain MG1655
(ATCC#47076). Primers used were: forward primer
50CATGCCATGGCAACGAATTTACGTGGC30 (NcoI
restriction site is underlined) and reverse primer
50CGCGGATCCTCATTAGTGATGATGATGATGAT-
GCCCGCGCTCTTGCATCAACTG30 (BamHI restric-
tion site is underlined, the gene sequence for a hexahis-
tidine-tage is in italics). Three full length C-terminal
His6-tagged CMP–sialic acid synthetases were cloned
in a similar method. N. meningitidis CSS was cloned in
pET22b(+) vector from group B strain MC58 using
genomic DNA (ATCC#BAA-335D) as a template.
The primers used were: forward primer 50GATCCA-
TATGGAAAAACAAAATATTGCGG30 (NdeI restric-
tion site is underlined) and reverse primer 50CCGCTC-
GAGGCTTTCCTTGTGATTAAG30 (XhoI restriction
site is underlined). S. agalactiae CSS was cloned in
pET15b vector from serotype V strain 2603V/R using
4.2.3. Purification of the recombinant sialic acid aldolase
and CMP–sialic acid synthetases. C-Terminal His6-
tagged target proteins were purified from cell lysate indi-
vidually. To obtain the cell lysate, cells were harvested
by centrifugation at 4000rpm for 30min. The cell pellet
was resuspended at 20mL per liter cell culture in lysis
buffer (pH8.0, 100mM Tris–HCl containing 0.1% Tri-
ton X-100). Lysozyme (1mg/L culture) and DNaseI
(50lg/L culture) were then added to the cell resuspen-
sion. The mixture was incubated at 37ꢁC for 50min with
vigorous shaking. Cell lysate was obtained by centrifug-
ation at 4000 rpm for 60min to separate the cell lysate
from the inclusion bodies and other cellular debris.
The enzyme was purified directly from the lysate using a
Ni2+–NTA affinity column, which binds to the C-termi-
nal hexahistidine sequence of the enzymes. All proce-
dures were performed on ice. After loading the cell

6432
H. Yu et al. / Bioorg. Med. Chem. 12 (2004) 6427–6435
lysate, the Ni2+ column was washed with six column vol-
umes of binding buffer (5mM imidazole, 0.5M NaCl,
20mM Tris–HCl, pH7.5), followed by 8 volumes of
washing buffer (20mM imidazole, 0.5M NaCl, 20mM
Tris–HCl, pH7.5). The overexpressed protein was eluted
with 8 volumes of elution buffer (200mM imidazole,
0.5M NaCl, 20mM Tris–HCl, pH7.5). Detected by a
UV–vis spectrometer at 280nm, fractions containing
the purified enzyme were collected and stored at 4ꢁC.
The purified enzyme in elute solution was dialyzed be-
fore use.
4.4.2. N-Azidoacetyl-D-mannosamine (1c). This com-
pound was synthesized in a similar method described
for 1b. Mannosamine hydrochloride salt (0.5g,
2.32mmol) was dissolved in 10mL of dry MeOH under
argon. To this solution was added 1.3mL of triethyl-
amine, followed by azidoacetic acid NHS ester (0.6g,
1.3equiv). The resulting solution was stirred at room
temperature for overnight until the completion of the
reaction. The reaction mixture was then concentrated
and the product was purified by flash chromatography
(MeOH/CH2Cl2 = 20:1 to 10:1) to give a white solid
(0.498g, 82%). 1H NMR (400MHz, D2O): d 4.97 (s,
0.5H), 4.87 (s, 0.5H), 4.32 (d, 0.5H, J = 4.0Hz), 4.19
(d, 0.5H, J = 4.4Hz), 3.94–3.88 (m, 3H), 3.73–3.59 (m,
4H), 3.43 (t, 0.5H, J = 10.0Hz), 3.33 (t, 0.5H,
J = 10.0Hz).
4.2.4. Sodium dodecylsulfate-polyacrylamide gel electro-
phoresis (SDS-PAGE). SDS-PAGE was performed in a
12% Tris–glycine gel using Bio-Rad Mini-protein III cell
gel electrophoresis unit (Bio-Rad, Hercules, CA) at
DC = 150V. Bio-Rad Precision Plus Protein Standards
(10–250kDa) were used as molecular weight standards.
Gels were stained with Coomassie Blue.
4.4.3. N-Cbz-glycine-D-mannosamine (1d). This com-
pound was synthesized in a similar method described
for 1b and 1c. Mannosamine hydrochloride salt (2.0g,
9.27mmol) was dissolved in 20mL of dry MeOH under
argon. To this solution was added 1.3mL of triethyl-
amine. After stirring the reaction mixture for 10min,
N-Cbz-glycine N-hydroxysuccinimide ester (3.40g,
1.2equiv) was added, and the reaction continued with
stirring at room temperature overnight. The reaction
mixture was concentrated and the product was purified
by flash chromatography (MeOH/CH2Cl2 = 20:1 to
10:1) to give a white solid (3.02g, 88%). 1H NMR
(400MHz, CD3OD): d 7.35–7.27 (m, 5H), 5.05 (s, 2H),
4.44 (d, 0.3H, J = 4.2Hz), 4.32–4.30 (d, 0.7H,
J = 4.5Hz), 4.04–4.0 (dd, 1H, J = 4.6 and 9.7Hz), 3.88
(s, 2H), 3.86–3.68 (m, 4H), 3.62–3.57 (t, 0.7H,
J = 9.5Hz), 3.50–3.46 (t, 0.3H, J = 9.5Hz).
4.2.5. Quantification of purified protein. The concentra-
tion of purified enzymes was obtained using a Bicinch-
oninic acid (BCA) Protein Assay Kit (Pierce
Biotechnology, Rockford, IL) with bovine serum albu-
min as a protein standard.
4.3. General method for compound characterization and
purification
1H NMR and 13C NMR spectra were recorded on an
Mercury (300MHz) and an Inova (400MHz) NMR
spectrometers. High resolution mass spectra (FAB) were
obtained at the mass spectrometry facility at the Univer-
sity of California, Riverside. Gel filtration chromatogra-
phy was performed with a column (100cm · 2.5cm)
packed with BioGel P-2 Fine resins (Bio-Rad, Hercules,
CA). Analytical thin-layer chromatography was per-
formed on silica gel 60 F254 plates (Sorbent Tech.) using
anisaldehyde stain for detection.
4.4.4. 2-Azido-2-deoxy-D-mannose (1e). This compound
was synthesized from 2-amino-2-deoxy-D-mannopyr-
anose hydrochloride with trifluoromethylsulfonyl azide
as reported.50 An aqueous solution (8mL) of NaN3
(3.5g, 54mmol) was mixed with CH2Cl2 (4mL) and
cooled to 0ꢁC in an ice-water bath. To the vigorously
stirred mixture was added Tf2O (2.45g, 1.51mL,
9mmol) in droplets. After stirring for an additional 2h
at 0ꢁC, the organic phase was separated and collected.
The aqueous phase was extracted with CH2Cl2 (30mL)
twice. The organic phases containing triflic azide were
combined and washed with saturated aqueous Na2CO3
by extraction. This freshly prepared triflic azide solution
was added immediately to a solution prepared by adding
mannosamine Æ HCl (0.45g, 2.1mmol), K2CO3 (0.48g),
and CuSO4 Æ 5H2O (8mg) sequentially to a water
(8mL) and MeOH (15mL) mixture. Additional MeOH
was added until the solution became clear. The reaction
mixture was then stirred overnight at room temperature.
The solution was concentrated and the product was
purified by flash chromatography (CH2Cl2/MeOH =
10:1 to 5:1) to give a yellow syrup, which was dissolved
in MeOH and decolorized by activated charcoal. After
filtration, the solution was concentrated to give a white
solid (0.37g, 86%). 1H NMR (400MHz, CD3OD): d
5.09 (d, 0.5H, J = 1.6Hz), 4.89 (d, 0.5H, J = 1.2Hz),
4.00 (dd, 0.5H, J = 5.6 and 9.6Hz), 3.84–3.56 (m,
4.5H), 3.44 (t, 0.5H, J = 9.6Hz), 3.20 (m, 0.5H).
4.4. Synthesis of mannose and N-acetylmannosamine deri-
vatives
4.4.1. N-Glycolyl-D-mannosamine (1b). This compound
was synthesized using a reported method.49 Mannos-
amine hydrochloride salt (0.32g, 1.48mmol) was dis-
solved in 10mL of dry MeOH under argon. To this
solution was added 0.42mL of triethylamine. The mix-
ture was stirred for 10min until the solution turned
clear.
N-hydroxysuccinimidyl
glycolate
(0.38g,
1.5equiv) was then added and the resulting solution
was stirred at room temperature overnight. When TLC
analysis indicated the disappearance of the starting
material, the reaction mixture was concentrated using
a rotavapor (Buchi Rotavapor R-205). The product
was purified by flash chromatography on silica gel by
eluting with MeOH/CH2Cl2 = 20:1 to 10:1. The product
was obtained as a white solid (0.32g, 91%) after concen-
1
tration using a rotavapor. H NMR (300MHz, D2O): d
5.00 (s, 0.6H), 4.92 (s, 0.4H), 4.23 (d, 0.4H, J = 3.3Hz),
4.21 (d, 0.6H, J = 3.6Hz), 4.05–3.90 (m, 3H), 3.79–3.60
(m, 4H), 3.46 (t, 0.6H, J = 9.6Hz), 3.36 (t, 0.4H,
J = 9.6Hz).

H. Yu et al. / Bioorg. Med. Chem. 12 (2004) 6427–6435
6433
4.4.5. 6-Azido-6-deoxy-D-mannose (1g). This compound
was prepared from 6-toluenesulfonate-D-mannopyra-
nose and lithium azide.51 A solution of p-toluenesulfo-
nyl chloride (2.83g, 14.8mmol) in dry pyridine (10mL)
was added in droplets to a solution of D-(+)-mannose
(2.06g, 11.4mmol) dissolved in dry pyridine in an ice-
water bath. The mixture was stirred at 0ꢁC for 2h and
the reaction was stopped by the addition of MeOH
(10mL). The mixture was concentrated and the product
was purified by flash chromatography (CH2Cl2/
MeOH = 10:1) to give a yellow syrup (6-O-tosyl-D-man-
nopyranose) (3.44g, 90%). To a solution of 6-O-tosyl-D-
mannopyranose (610mg, 1.82mmol) in N,N-dimethyl-
formamide (10mL) was added lithium azide (in 20%
water, 1.31g, 3equiv). The mixture was stirred at 65ꢁC
for 12h. The solution was concentrated and the product
was purified by a flash chromatography (CH2Cl2/
MeOH = 10:1 to 5:1) to give a yellow syrup (265mg,
71%), which was then decolorized by activated char-
fer B for 1min, followed by an equilibration phase of
100% buffer A for 4min. Flow-rate was maintained at
1mL/min for all separations. Compounds 3a, 3b, 3c,
3f, 3g, 3h were analyzed by gradient condition 2, 3d
and 3e by gradient condition 1.
4.6. Preparative synthesis of CMP–sialic acid analogs
4.6.1. General reaction procedures. The preparative syn-
thesis of CMP–sialic acid analogs was carried out in a
Tris–HCl buffer (100mM, pH8.8) containing MgCl2
(20mM), mannose or ManNAc derivative (50–100mg),
sodium pyruvate (5equiv), CTP (1equiv), the recombi-
nant sialic acid aldolase, and the N. meningitidis
CMP–Neu5Ac synthetase. The reaction mixture was
incubated at 37ꢁC for 2–5h in an incubator with shak-
ing (140rpm). The reaction was stopped by addition of
MeOH when TLC (elution system: EtOH/1M
NH4HCO3; 7:3 by volume) analysis indicated reaction
completion. The precipitates in the reaction mixture
were removed by centrifugation. The supernatant was
concentrated and purified by a gel filtration column
packed with BioGel P-2 resin.
1
coal. H NMR (400MHz, D2O): d 5.01 (s, 0.6H), 4.74
(s, 0.4H), 3.78–3.56 (m, 3H), 3.52–3.30 (m, 3H). 13C
NMR (100MHz, D2O, a anomer only): d 94.3 (C-1),
71.2, 70.7, 70.2, 67.7, 51.2 (C-6).
4.5. TLC and HPLC analyses for the production of
CMP–sialic acid analogs
4.6.2. Cytidine-50-monophospho-5-acetamido-3,5-dideoxy-
b-D-glycero-D-galacto-2-nonulopyranosonic acid (3a).
HPLC retention time: 6.63min. 1H NMR (400MHz,
D2O): d 7.87 (d, 1H, J = 7.6Hz, H-6 of cytidine), 6.25
(d, 1H, J = 7.6Hz, H-5 of cytidine), 5.91 (d, 1H,
J = 4.4Hz, H-1 of ribose), 4.26–4.20 (m, 2H, H-2 of ri-
bose, H-3 of ribose), 4.17–4.10 (m, 3H, H-5a, H-5b,
H-4 of ribose), 4.09–3.98 (m, 2H), 3.90–3.80 (m, 3H),
3.59–3.40 (m, 2H), 2.42 (dd, 1H, J = 4.8, 13.2Hz), 1.98
(s, 3H), 1.60 (dt, 1H, J = 5.6,12.6Hz); 13C (100MHz):
d 174.9, 174.6, 166.3, 157.9, 141.7, 100.2, 96.8, 89.3,
83.0, 74.4, 71.9, 69.7, 69.2, 68.9, 67.0, 65.1, 63.1, 51.9,
4.5.1. General reaction procedures. All one-pot two-enzy-
matic reactions catalyzed by an aldolase and a CMP–
Neu5Ac synthetase were performed in a Tris–HCl buffer
(100mM, pH8.8) containing MgCl2 (20mM), mannose
or ManNAc derivatives (10mM), sodium pyruvate
(50mM), and CTP (10mM). After the addition of the
sialic acid aldolase and a CMP–sialic acid synthetase,
the reaction was allowed to proceed for 2–5h at 37ꢁC.
The reaction was monitored by TLC analysis until no
increased product formation could be detected. The pre-
cipitates in the reaction mixture were removed by cen-
trifugation and then filtration (0.2lm syringe filter).
Aliquots of 10lL clear solution were injected for HPLC
analysis.
41.3,
22.3.
HRMS
(FAB):
C20H29N4O16PNa
(M2ꢀ + Na+), calcd 635.1214, found 635.1191.
4.6.3. Cytidine-50-monophospho-5-(2-hydroxy)-acetamido-
3,5-dideoxy-b-D-glycero-D-galacto-2-nonulopyranosonic
acid (3b). HPLC retention time: 5.69min. 1H NMR
(400MHz, D2O): d 7.79 (d, 1H, J = 7.6Hz, H-6 of cyt-
idine), 5.95 (d, 1H, J = 7.6Hz, H-5 of cytidine), 5.81
(d, 1H, J = 4.4Hz, H-1 of ribose), 4.19–4.14 (m, 2H,
H-2 of ribose, H-3 of ribose), 4.13–4.04 (m, 3H, H-5a,
H-5b, H-4 of ribose), 4.03–3.98 (m, 1H), 3.97 (s, 2H),
3.86–3.70 (m, 3H), 3.48–3.41 (m, 2H), 3.28 (d, 1H,
J = 9.6Hz, H-700), 2.34 (dd, 1H, J = 4.8, 13.2Hz), 1.51
(dt, 1H, J = 5.6,12.6Hz); 13C (100MHz): d 175.7,
174.5, 166.2, 157.8, 141.7, 100.2, 96.8, 89.3, 83.0, 74.4,
71.7, 69.7, 68.9, 66.7, 65.0, 63.0, 61.2, 59.5, 51.6, 41.3.
HRMS (FAB): C20H29N4O17PNa (M2ꢀ + Na+), calcd
651.1163, found 651.1126.
4.5.2. HPLC conditions. HPLC analysis was carried out
using
a
reverse phase Premier C18 column
(250 · 4.6mm ID, 5lm, particle size) protected with a
C18 guard column cartridge in a Shimadzu LC-2010A
system equipped with a membrane on-line degasser, a
temperature control unit (maintained at 30ꢁC through-
out the experiment), and a UV–vis detector. The com-
pounds were monitored at 270nm and the system was
controlled via Shimadzu EZStrat v7.2 SP1 chromato-
graphy software. Buffer A (0.1M potassium phosphate
buffer supplemented with 8mM tetrabutylammonium
hydrogensulfate, pH5.3) and buffer B (70% buffer A
plus 30% methanol, pH5.9) were prepared and used in
two gradient conditions for separation of different com-
pounds. Gradient condition 1 was: 100% buffer A for
2.5min, 0–40% buffer B for 14min, 40–100% buffer B
for 1min, 100% buffer B for 6min, 100–0% buffer B
for 1min, followed by an equilibration phase of 100%
buffer A for 8min. Gradient condition 2 was: 100% buf-
fer A for 2.5min, 0–40% buffer B for 7.5min, 40–100%
buffer B for 1min, 100% buffer B for 4min, 100–0% buf-
4.6.4. Cytidine-50-monophospho-5-azidoacetamido-3,5-
dideoxy-b-D-glycero-D-galacto-2-nonulopyranosonic acid
(3c). HPLC retention time: 14.48min. 1H NMR
(400MHz, D2O): d 7.70 (d, 1H, J = 7.6Hz, H-6 of cyt-
idine), 5.87 (d, 1H, J = 7.2Hz, H-5 of cytidine), 5.73
(d, 1H, J = 4.0Hz, H-1 of ribose), 4.15–4.05 (m, 2H,
H-2 of ribose, H-3 of ribose), 4.05–4.01 (m, 3H, H-5a,
H-5b, H-4 of ribose), 3.90–3.52 (m, 5H), 3.85 (s, 2H),

6434
H. Yu et al. / Bioorg. Med. Chem. 12 (2004) 6427–6435
3.41–3.36 (m, 1H), 3.22 (d, 1H, J = 9.2Hz, H-700), 2.22
(dd, 1H, J = 4.8, 13.6Hz), 1.42 (dt, 1H,
J = 6.0,12.8Hz); 13C (100MHz): d 174.7, 169.4, 166.2,
157.8, 141.7, 100.1, 96.7, 89.4, 83.0, 74.3, 73.8, 71.9,
69.8, 69.5, 67.9, 65.0, 63.1, 59.6, 48.9, 41.0. HRMS
(FAB): C20H29N7O16P (M2ꢀ + H+), calcd 654.1408,
found 654.1421.
4.6.9. Cytidine-50-monophospho-3-deoxy-b-D-galacto-2-
octulopyranosonic acid (3h). HPLC retention time:
5.94min. 1H NMR (400MHz, D2O) d 7.81 (d, 1H,
J = 7.6Hz, H-6 of cytidine), 5.96 (d, 1H, J = 7.2Hz,
H-5 of cytidine), 5.83 (d, 1H, J = 3.6Hz, H-1 of ribose),
4.20–4.14 (m, 2H, H-2 of ribose, H-3 of ribose), 4.14–
4.05 (m, 3H, H-5a, H-5b, H-4 of ribose), 3.90–3.80 (m,
3H), 3.71 (d, 1H, J = 8.8Hz), 3.67–3.63 (m, 1H), 3.52–
3.48 (m, 1H), 2.27 (dd, 1H, J = 4.8, 13.6Hz), 1.46 (dt,
1H, J = 6.0,12.8Hz); 13C (100MHz): d 174.8, 166.3,
157.9, 141.7, 100.2, 96.7, 89.3, 83.0, 74.4, 73.6, 69.4,
68.4, 64.8, 63.1, 62.1, 59.8, 41.0. HRMS (FAB):
C17H24N3O15PNa (M2ꢀ + Na+), calcd 564.0843, found
564.0867.
4.6.5. Cytidine-50-monophospho-5-(N-benzyloxycarbox-
amido)glycylamido-3,5-dideoxy-b-D-glycero-D-galacto-2-
nonulopyranosonic acid (3d). HPLC retention time:
27.76min. 1H NMR (300MHz, D2O) d 7.77 (d, 1H,
J = 7.5Hz, H-6 of cytidine), 7.24 (m, 5H), 5.92 (d, 1H,
J = 7.68Hz, H-5 of cytidine), 5.79 (d, 1H, J = 4.2Hz,
H-1 of ribose), 4.96 (s, 2H), 4.16–4.05 (m, 5H, H-2 of
ribose, H-3 of ribose, H-5a, H-5b, H-4 of ribose),
4.03–3.14 (m, 9H), 2.31 (dd, 1H), 1.51 (dt, 1H); 13C
(100MHz): d 174.5, 166.3, 158.7, 157.9, 141.7, 136.3,
128.9, 128.5, 127.9, 100.2, 96.7, 89.3, 82.9, 74.4, 72.2,
71.7, 69.8, 69.4, 68.9, 67.5, 64.9, 63.1, 52.0, 43.8, 41.0.
HRMS (FAB): C28H37N5O18P (M2ꢀ + H+), calcd
762.1871, found 762.1777.
Acknowledgements
This work was supported by New Faculty Research
Grant and start-up funds from the University of Califor-
nia at Davis.
References and notes
4.6.6. Cytidine-50-monophospho-5-azido-3,5-dideoxy-b-D-
glycero-D-galacto-2-nonulopyranosonic acid (3e). HPLC
retention time: 20.02min. 1H NMR (400MHz, D2O)
d 7.77 (d, 1H, J = 7.6Hz, H-6 of cytidine), 5.94 (d,
1H, J = 7.6Hz, H-5 of cytidine), 5.80 (d, 1H,
J = 4.0Hz, H-1 of ribose), 4.17 (t, 1H, J = 4.0Hz, H-2
of ribose), 4.13 (t, 1H, J = 4.0Hz, H-3 of ribose),
4.09–4.04 (m, 3H, H-5a, H-5b, H-4 of ribose), 3.97–
3.85 (m, 2H), 3.81–3.70 (m, 2H), 3.53–3.48 (m, 1H),
3.45–3.34 (m, 2H), 2.28 (1H, dd, J = 4.8, 13.2Hz), 1.50
(td, 1H, J = 6.0 and 12.8Hz); 13C (100MHz): d 174.4,
166.2, 157.8, 141.7, 100.1, 96.7, 89.4, 83.0, 74.3, 72.0,
69.8, 69.5, 69.4, 67.9, 65.0, 63.1, 59.6, 41.4. HRMS
1. Schauer, R. Glycoconjugate J. 2000, 17, 485.
2. Angata, T.; Varki, A. Chem. Rev. 2002, 102, 439.
3. Mandrell, R. E.; Apicella, M. A. Immunobiology 1993,
187, 382.
4. Moran, A. P.; Prendergast, M. M.; Appelmelk, B. J.
FEMS Immunol. Med. Microbiol. 1996, 16, 105.
5. Jennings, H. J. Curr. Top. Microbiol. Immunol. 1990, 150,
97.
6. Chappell, M. D.; Halcomb, R. L. J. Am. Chem. Soc. 1997,
119, 3393.
7. Kean, E. L. Glycobiology 1991, 1, 441.
8. Higa, H. H.; Paulson, J. C. J. Biol. Chem. 1985, 260, 8838.
9. Kittelmann, M.; Klein, T.; Kragl, U.; Wandrey, C.;
Ghisalba, O. Ann. N.Y. Acad. Sci. 1992, 672, 444.
10. Schmelter, T.; Ivanov, S.; Wember, M.; Stangier, P.;
Thiem, J.; Schauer, R. Biol. Chem. H–S 1993, 374, 337.
11. Vionnet, J.; Concepcion, N.; Warner, T.; Zapata, G.;
Hanover, J.; Vann, W. F. Glycobiology 1999, 9, 481.
12. Kean, E. L.; Roseman, S. J. Biol. Chem. 1966, 241, 56430.
13. Schauer, R.; Haverkamp, J.; Ehrlich, K. H–S Z. Physiol.
Chem. 1980, 361, 641.
14. Vann, W. F.; Silver, R. P.; Abeijon, C.; Chang, K.;
Aaronson, W.; Sutton, A.; Finn, C. W.; Lindner, W.;
Kotsatos, M. J. Biol. Chem. 1987, 262, 17556.
15. Warren, L.; Blacklow, R. S. J. Biol. Chem. 1962, 237,
3527.
16. Zapata, G.; Vann, W. F.; Aaronson, W.; Lewis, M. S.;
Moos, M. J. Biol. Chem. 1989, 264, 14769.
17. Edwards, U.; Frosch, M. FEMS Microbiol. Lett. 1992, 96,
161.
18. Ganguli, S.; Zapata, G.; Wallis, T.; Reid, C.; Boulnois, G.;
Vann, W. F.; Roberts, I. S. J. Bacteriol. 1994, 176, 4583.
19. Haft, R. F.; Wessels, M. R.; Mebane, M. F.; Conaty, N.;
Rubens, C. E. Mol. Microbiol. 1996, 19, 555.
20. Guerry, P.; Doig, P.; Alm, R. A.; Burr, D. H.; Kinsella,
N.; Trust, T. J. Mol. Microbiol. 1996, 19, 369.
21. Tullius, M. V.; Munson, R. S., Jr.; Wang, J.; Gibson, B.
W. J. Biol. Chem. 1996, 271, 15373.
(FAB): C18H25N6O15PNa
619.1013, found 619.0986.
(M2ꢀ + Na+),
calcd
4.6.7. Cytidine-50-monophospho-3-deoxy-b-D-glycero-D-
galacto-2-nonulopyranosonic acid (3f). HPLC retention
1
time: 5.42min. H NMR (400MHz, D2O) d 7.75 (d, 1H,
J = 7.6Hz, H-6 of cytidine), 5.90 (d, 1H, J = 7.6Hz,
H-5 of cytidine), 5.77 (d, 1H, J = 4.0Hz, H-1 of ribose),
4.16–4.08 (m, 2H, H-2 of ribose, H-3 of ribose), 4.05–
4.01 (m, 3H, H-5a, H-5b, H-4 of ribose), 3.90–3.62 (m,
4H), 3.59–3.35 (m, 3H), 2.22 (dd, 1H, J = 4.8,
13.2Hz), 1.40 (dt, 1H, J = 6.0,12.8Hz); 13C (100MHz):
d 174.7, 166.1, 157.7, 141.7, 100.2, 96.7, 89.4, 83.0,
74.3, 73.0, 69.9, 69.4, 68.6, 68.4, 64.9, 63.1, 59.5, 41.0.
HRMS (FAB): C18H27N3O16P (M2ꢀ + H+), calcd
572.1129, found 572.1152.
4.6.8. Cytidine-50-monophospho-9-azido-3,9-dideoxy-b-D-
glycero-D-galacto-2-nonulopyranosonic acid (3g). HPLC
retention time: 16.74min. 1H NMR (400MHz, D2O)
d 7.76 (d, 1H, J = 7.6Hz, H-6 of cytidine), 5.90 (d, 1H,
J = 7.6Hz, H-5 of cytidine), 5.80 (d, 1H, J = 4.0Hz,
H-1 of ribose), 4.20–4.00 (m, 5H, H-2 of ribose, H-3
of ribose, H-5a, H-5b, H-4 of ribose), 3.98–3.35 (m,
7H), 2.22 (dd, 1H, J = 4.8, 13.2Hz), 1.40 (dt, 1H,
J = 6.0,12.8Hz). HRMS (FAB): C18H25N6O15PNa3
(M2ꢀ + 3Na+), calcd 665.0809, found 665.0807.
22. Gilbert, M.; Watson, D. C.; Wakarchuk, W. W. Biotech-
nol. Lett. 1997, 19, 417.
23. Luneberg, E.; Zetzmann, N.; Alber, D.; Knirel, Y. A.;
Kooistra, O.; Zahringer, U.; Frosch, M. Int. J. Med.
Microbiol. 2000, 290, 37.

H. Yu et al. / Bioorg. Med. Chem. 12 (2004) 6427–6435
6435
24. Xia, G.; Han, X.; Ding, H.; Yang, S.; Zhang, S.; Jin, C. J.
Mol. Catal. B—Enzym. 2000, 10, 199.
25. Nakata, D.; Munster, A. K.; Gerardy-Schahn, R.; Aoki,
N.; Matsuda, T.; Kitajima, K. Glycobiology 2001, 11,
f685.
26. Krapp, S.; Munster-Kuhnel, A. K.; Kaiser, J. T.; Huber,
R.; Tiralongo, J.; Gerardy-Schahn, R.; Jacob, U. J. Mol.
Biol. 2003, 334, 625.
37. Kittelmann, M.; Klein, T.; Kragl, U.; Wandrey, C.;
Ghisalba, O. Appl. Microbiol. Biotechnol. 1995, 44, 59.
38. Lee, S. G.; Lee, J. O.; Yi, J. K.; Kim, B. G. Biotechnol.
Bioeng. 2002, 80, 516.
39. Martin, R.; Witte, K. L.; Wong, C.-H. Bioorg. Med.
Chem. 1998, 6, 1283.
40. Nahalka, J.; Wu, B.; Shao, J.; Gemeiner, P.; Wang, P. G.
Biotechnol. Appl. Biochem. 2004, 40, 101.
27. Mosimann, S. C.; Gilbert, M.; Dombroswki, D.; To, R.;
Wakarchuk, W.; Strynadka, N. C. J. Biol. Chem. 2001,
276, 8190.
28. Jelakovic, S.; Jann, K.; Schulz, G. E. FEBS Lett. 1996,
391, 157.
29. Jelakovic, S.; Schulz, G. E. J. Mol. Biol. 2001, 312, 143.
30. Shames, S. L.; Simon, S. L.; Christopher, C. W.; Schmid,
W.; Whitesides, G. M.; Yang, L. L. Glycobiology 1991, 1,
187.
31. Shen, G. J.; Liu, J. L. C.; Wong, C.-H. Biocatalysis 1992,
6, 31.
32. Liu, J. L. C.; Shen, G. J.; Ichikawa, Y.; Rutan, J. F.;
Zapata, G.; Van, W. F.; Wong, C.-H. J. Am. Chem. Soc.
1992, 114, 3901.
33. Karwaski, M. F.; Wakarchuk, W. W.; Gilbert, M. Protein
Expres. Purif. 2002, 25, 237.
34. Knorst, M.; Fessner, W. D. Adv. Synth. Catal. 2001, 343,
698.
35. Blixt, O.; Paulson, J. C. Adv. Synth. Catal. 2003, 345, 687.
36. Gilbert, M.; Bayer, R.; Cunningham, A. M.; DeFrees, S.;
Gao, Y.; Watson, D. C.; Young, N. M.; Wakarchuk, W.
W. Nat. Biotechnol. 1998, 16, 769.
41. Haft, R. F.; Wessels, M. R. J. Bacteriol. 1994, 176, 7372.
42. Liu, G.; Jin, C. J. Biol. Chem. 2004, 279, 17738.
43. Wong, C.-H.; Whitesides, G. M. In Enzymes in Synthetic
Organic Chemistry; Wong, C.-H., Whitesides, G. M., Eds.;
Elsevier Science: Oxford, 1994; Vol. 12, pp 215–219.
44. Kean, E. L.; Roseman, S. In Methods in Enzymology;
Neufeld, E. F., Ginsburg, V., Eds.; Academic: New York,
1966; Vol. 8, pp 208–215.
45. Petrie, C. R., 3rd; Korytnyk, W. Anal. Biochem. 1983, 131,
153.
46. Ryll, T.; Wagner, R. J. Chromatogr. 1991, 570, 77.
47. Simon, E. S.; Bednarski, M. D.; Whitesides, G. M. J. Am.
Chem. Soc. 1988, 110, 7159.
48. Kuboki, A.; Okazaki, H.; Sugai, T.; Ohta, H. Tetrahedron
1997, 53, 2387.
49. Choi, S. K.; Lee, S.; Whitesides, G. M. J. Org. Chem.
1996, 61, 8739.
50. Izumi, M.; Shen, G. J.; Wacowich-Sgarbi, S.; Nakatani,
T.; Plettenburg, O.; Wong, C.-H. J. Am. Chem. Soc. 2001,
123, 10909.
51. Saxon, E.; Luchansky, S. J.; Hang, H. C.; Yu, C.; Lee,
S. C.; Bertozzi, C. R. J. Am. Chem. Soc. 2002, 124, 14893.