Chemoenzymatic synthesis of GDP-L-fucose derivatives as potent and selective α-1,3-fucosyltransferase inhibitors

Article abstract of DOI:10.1002/adsc.201100940

Fucosyltransferases (FucTs) usually catalyze the final step of glycosylation and are critical to many biological processes. High levels of specific FucT activities are often associated with various cancers. Here we report the development of a chemoenzymatic method for synthesizing a library of guanosine diphosphate β-L-fucose (GDP-Fuc) derivatives, followed by in situ screening for inhibitory activity against bacterial and human α-1,3-FucTs. Several compounds incorporating appropriate hydrophobic moieties were identified from the initial screening. These were individually synthesized, purified and characterized in detail for their inhibition kinetics. Compound 5 had a K_i of 29 nM for human FucT-VI, and is 269 and 11 times more selective than for Helicobacter pylori FucT (K_i=7.8 μM) and for human FucT-V (K_i=0.31 μM). Copyright

Full text of DOI:10.1002/adsc.201100940

FULL PAPERS
DOI: 10.1002/adsc.201100940
Chemoenzymatic Synthesis of GDP-l-Fucose Derivatives as
Potent and Selective a-1,3-Fucosyltransferase Inhibitors
Yu-Nong Lin,a,b Daniel Stein,a Sheng-Wei Lin,a,d Sue-Ming Chang,a,b
Ting-Chien Lin,a,c Yu-Ruei Chuang,a,c Jacquelyn Gervay-Hague,e
Hisashi Narimatsu,f and Chun-Hung Lina,b,c,
*
a
Institute of Biological Chemistry, Academia Sinica, No. 128 Academia Road Section 2, Nan-Kang, Taipei 11529, Taiwan
Fax : (+886)-2-2651-4705; phone: (+886)-2-2789-0110; e-mail: chunhung@gate.sinica.edu.tw
Institute of Biochemical Sciences, National Taiwan University, Taipei, 10617, Taiwan
Department of Chemistry, National Taiwan University, Taipei, 10617, Taiwan
Genomics Research Center, Academia Sinica, Taipei, 11529, Taiwan
Department of Chemistry, University of California, Davis, CA 95618, USA
Research Center for Medical Glycoscience, National Institute of Advanced Industrial Science and Technology, Tsukuba,
Japan
b
c
d
e
f
Received: December 2, 2011; Revised: February 23, 2012; Published online: May 16, 2012
Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/adsc.201100940.
Abstract: Fucosyltransferases (FucTs) usually cata- These were individually synthesized, purified and
lyze the final step of glycosylation and are critical to characterized in detail for their inhibition kinetics.
many biological processes. High levels of specific Compound 5 had a Ki of 29 nM for human FucT-VI,
FucT activities are often associated with various can- and is 269 and 11 times more selective than for Heli-
cers. Here we report the development of a chemoen- cobacter pylori FucT (Ki =7.8 mM) and for human
zymatic method for synthesizing a library of guano- FucT-V (Ki =0.31 mM).
sine diphosphate b-l-fucose (GDP-Fuc) derivatives,
followed by in situ screening for inhibitory activity
against bacterial and human a-1,3-FucTs. Several Keywords: enzyme catalysis; enzymes; guanosine di-
compounds incorporating appropriate hydrophobic phosphate b-l-fucose (GDP-Fuc) derivatives; inhibi-
moieties were identified from the initial screening. tors; transferases
Introduction
anti-tumor therapies is therefore a promising strat-
egy.[10]
Cell-surface glycoconjugates play an important role in
In humans, there are 12 FucTs that catalyze the for-
numerous physiological and pathological processes, mation of various fucosyl linkages (a-1,2-, a-1,3-, a-
including cell-cell adhesion, cell differentiation, 1,4- and a-1,6-glycosidic bonds);[11] thus, selective in-
immune response, fertilization, viral and bacterial in- hibition of a specific FucT(s) offers a prospective
fection, and tumor progression.[1–4] Fucosylated gly- strategy for developing a therapeutic agent. FucTs
cans are critical to a variety of cell events, and fucosy- catalyze the transfer of l-fucose (Fuc) from guanosine
lation of cell-surface glycans is often the final step in diphosphate b-l-fucose (GDP-Fuc) to various sugar-
the biosynthesis of an oligosaccharide. For instance, acceptor substrates. Several approaches have been de-
the tetrasaccharide sialyl Lewis x (sLex) is well known veloped for the preparation of FucT inhibitors, such
for its physiological importance in the inflammatory as the design and synthesis of substrate- and transition
response and the recruitment of leukocytes, and in state-based analogues.[12–17] Rational design is mainly
tumor metastasis.[5] a-2,3-Sialyltransferase and a-1,3- impeded by the limited structural information avail-
fucosyltransferase (a-1,3-FucTs) sequentially catalyze able, however. We previously reported the first X-ray
the last two steps in the corresponding biosynthesis. structures of Helicobacter pylori a-1,3-FucT (HP-
Aberrant expression of sLex and enhanced FucT ac- FucT); this remains the only study on the enzyme-
tivities are found in many carcinomas.[6–9] Developing substrate and enzyme-product complex structures to
FucT inhibitors as potential anti-inflammatory and date.[18] Recently, potent FucT inhibitors have been
1750
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2012, 354, 1750 – 1758

Chemoenzymatic Synthesis of GDP-l-Fucose Derivatives
Scheme 1. Three-step preparation of the 94 GDP-Fuc derivatives investigated for FucT inhibition. (A) Chemoenzymatic syn-
thesis of GDP-6-amino-Fuc (2) by FKP. PPase=inorganic pyrophosphatase. (B) Synthesis of GDP-Fuc derivatives by
amide-forming reactions in microtiter plates, followed by the in situ FucT activity assay.
developed separately by Wong and by Nishimura, by Results and Discussion
applying click chemistry to make GDP and GDP-Fuc
derivatives,[19–21] respectively. These were subsequently Chemoenzymatic Synthesis of GDP-6-amino-l-fucose
assayed for inhibition activity. In their studies, GDP- (2)
alkyne or GDP-6-alkynylfucose was prepared to
couple with different azide derivatives via 1,3-dipolar In previous studies, GDP or GDP-Fuc analogues have
cycloaddition. Potent inhibition was thus achieved by mainly been synthesized using click chemistry strat-
incorporation of a suitably hydrophobic group to egies. Instead, we used amide-forming reactions, and
either the phosphate of GDP or the fucose residue of GDP-6-amino-l-fucose (2) was designed as the core
GDP-Fuc.
template to couple to various carboxylic acids. This
Previous studies indicated that sugar nucleotides study allowed us to determine if the amide coupling
and analogues often offer satisfying inhibition potency products are able to produce potent inhibition. The
against glycosyltransferases because they provide the bifunctional enzyme l-fucokinase/GDP-Fuc pyro-
necessary hydrogen bonding and electrostatic interac- phosphorylase (FKP), from Bacteroides fragilis,[26–28]
tions.[22–25] However, their synthesis is not trivial. Par- was used for the synthesis of compound 2, and it
ticularly it usually requires numerous trials for opti- offers the following advantages: (i) FKP is capable of
mizing the yields when introducing the mono- or di- synthesizing GDP-Fuc analogues in one step, directly
phosphate moiety, and purifying the intermediates or/ from the corresponding Fuc analogues; (ii) the opera-
and products. The development of an efficient syn- tion is simple compared to the chemical method;[29,30]
thetic procedure will bring a beneficial impact to the and (iii) FKP is easy to prepare with high yield and
fields of glycosyltransferases and related inhibition purity. Thus, N-terminally His-tagged FKP was
studies.
cloned, overexpressed in Escherichia coli and purified
In this report, we have established a three-step che- in one step, giving a yield of 40 mg of FKP with
moenzymatic method for the rapid assembly of vari- >95% homogeneity from 1 L of culture medium.
ous GDP-Fuc derivatives from 6-azido-l-fucose. To
Despite the flexible substrate tolerance of FKP, 6-
examine the inhibition and selectivity of these deriva- amino-l-fucose turned out to be a very poor sub-
tives, three a-1,3-FucTs were selected, including HP- strate. An alternative two-step reaction was therefore
FucT, human FucT-VI (FucT-VI), and human FucT-V employed in which GDP-6-azido-l-fucose (1) was
(FucT-V). Of these products, compound 5 was found synthesized by the FKP-catalyzed reaction from 6-
to be a potent inhibitor of FucT-VI (Ki =29 nM), with azido-l-fucose, followed by reduction of the azido
highly selective inhibition compared to that of FucT- group to give 2 (Scheme 1). The synthesis can be car-
V and HP-FucT.
ried out at the 100 mg scale and gives 68% overall
yield.
Adv. Synth. Catal. 2012, 354, 1750 – 1758
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1751

Yu-Nong Lin et al.
FULL PAPERS
Preparation of Various GDP-Fuc Derivatives and in
situ Screening for Inhibition Activity
Compound 2 was subjected to diversity-oriented syn-
thesis, i.e., the amide-forming reactions of compound
2 with 94 carboxylic acids (see the Supporting Infor-
mation, Figure S1 for the carboxylic acid structures)
were carried out in the presence of O-(benzotriazol-1-
yl)-N,N,N’,N’-tetramethyluronium
hexafluorophos-
phate (HBTU) and N,N-diisopropylethylamine
(DIEA) in wells of 96-well microtiter plates. The cou-
pling reactions were monitored by thin-layer chroma-
tography (TLC). Assuming the reaction yields to be
quantitative (as judged by TLC), the reaction mix-
tures were diluted and directly assayed for the inhibi-
tion activity of HP-FucT and FucT-VI without prod-
uct purification. We used radioisotope-labeled sub-
strate and TLC analysis for the activity assay because
of the difficulty in preparing human FucT, and be-
cause of the low enzyme activity and high cost of
Figure 1. Inhibition of FucT-VI (closed bars) and HP-FucT
(open bars) by 12 selected compounds. The synthesized
GDP-Fuc derivatives were screened in situ for inhibitory ac-
tivity against the two FucTs. Results for compounds dis-
cussed below or for those showing higher inhibition activi-
ties are shown. Amide products are labeled by adding “z” to
the end of the corresponding acid names. For example, G1z
is the product of coupling the carboxylic acid G1 and com-
pound 2. For the complete screening results, see Figure S2
GDP-Fuc. Furthermore, fluorescence-coupled enzy- in the Supporting Information.
matic assays[31] are not suitable for large-scale screen-
ing owing to the limited volume. Thus, enzyme activi-
ty was quantified by the transfer of l-[14C]Fuc to the minimal inhibition (<5%) under the conditions
acceptor substrate N-acetyllactosamine (LacNAc). tested.
None of the reagents used for the amide-forming re-
actions were found to generate any substantial inhibi-
tion (see the Supporting Information, Figure S3). The Evaluation of Inhibitory Activities of Compounds 3
concentration of each coupling product was kept at through 5
1 mM for preliminary screening. Most of the products
derived from the 48 carboxylic acids in groups A We individually synthesized and purified compounds
through D (see the Supporting Information, Fig- 3, 4 and 5 (Figure 2) for further characterization. The
ure S1) showed no inhibitory effect for either FucT- carboxylic acids of G6, E11 and E12, were converted
VI or HP-FucT (see the Supporting Information, Fig- to the corresponding succinimide esters using 1-ethyl-
ure S2), revealing that simple, aliphatic extensions 3-(3-dimethylaminopropyl)carbodiimide (EDC) and
from C-6 of Fuc may not improve FucT inhibition. In N-hydroxysuccinimide (NHS), followed by the cou-
contrast, the reaction products of several of the 12 pling reaction with 2. The desired products (3–5) were
compounds in group G (see the Supporting Informa- purified by anion exchange chromatography and char-
1
tion, Figure S1) showed good inhibition potency acterized by H and 13C NMR spectroscopy and mass
against FucT-VI, and their inhibitory activities (at the spectrometry. These amide products were stable in
concentration of 1 mM) are shown in Figure 1. In par- the assay condition (50 mM Tris-HCl buffer at
ticular, the amide product of G6 (designated G6z in pH 7.5) for at least three weeks, with no sign of deg-
1
Figure 1 and compound 5 in the following text) and radation according to the observation by H NMR.
G7 resulted in the most potent inhibition against
Lineweaver–Burk plot analysis indicated that com-
FucT-VI (98% and 32% inhibition, respectively), and pounds 3–5 all exhibited competitive inhibition with
those of E11 and E12 (designated E11z and E12z in respect to GDP-Fuc (Figure 3 and Supporting Infor-
Figure 1 and compounds 3 and 4 in the following text, mation, Figure S4–Figure S10). The IC50 and Ki
respectively) showed moderate inhibition (approxi- values for the three FucTs are listed in Table 1. Com-
mately 20%), as shown as the closed bars in Figure 1. pound 5 was found to be an excellent inhibitor for
Interestingly, these products exhibited a different in- FucT-VI with a Ki of 29 nM. It had Ki values of 7.8
hibition pattern for HP-FucT (shown as open bars in and 0.31 mM for HP-FucT and FucT-V, respectively,
Figure 1). Compounds 3 and 4 were the most potent indicating that it is 269 and 11 times more selective
inhibitors of HP-FucT, with 32% and 60% inhibition for FucT-VI than for HP-FucT and FucT-V, respec-
at 1 mM, respectively. Conversely, the coupling prod- tively. In contrast, although compound 4 was a potent
ucts derived from the carboxylic acids in group G inhibitor of HP-FucT (Ki =0.22 mM), it had no sub-
(see the Supporting Information, Figure S1) exhibited stantial selectivity for the three FucTs examined. Sim-
ilar results were obtained for 3. Regarding the inhibi-
1752
asc.wiley-vch.de
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2012, 354, 1750 – 1758

Chemoenzymatic Synthesis of GDP-l-Fucose Derivatives
Figure 2. The structures of compounds 3–5.
Figure 3. Lineweaver–Burk analysis of the steady-state kinetics of FucT-VI and HP-FucT in the presence of compounds 5
and 3, respectively. In both experiments, the concentration of GDP-Fuc was varied while that of LacNAc was fixed. The re-
sulting slopes were plotted against inhibitor concentration (shown as inserts). (A) Kinetics for FucT-VI in the presence of
~
&
*
^
compound 5 at 0.01 ( ), 0.02 ( ), 0.06 ( ) and 0.1 mM ( ). The insert indicates that 5 is a competitive inhibitor of FucT-VI
~
&
*
^
with a Ki of 29 nM. (B) Kinetics for HP-FucT in the presence of compound 4 at 0.1 ( ), 0.2 ( ), 0.3 ( ) and 0.4 mM ( ). The
insert indicates that 4 is a competitive inhibitor of HP-FucT with a Ki of 0.22 mM.
Table 1. IC50 and Ki values (mM) of compounds 3–5 against various FucTs.
3
4
5
Ki
GDP-Fuc
Km
Enzyme
IC50
Ki
IC50
Ki
IC50
FucT-V
FucT-VI
HP-FucT
2.7
5.4
3.9
0.81Æ0.15
1.9Æ0.7
0.9Æ0.2
1.2
2.1
0.78
0.52Æ0.11
1.1Æ0.29
0.22Æ0.06
0.69
0.094
26
0.31Æ0.09
0.029Æ0.009
7.8Æ0.7
13.8Æ2.3
13.2Æ2.5
11.3Æ2.2
tion of HP-FucT, the available X-ray complex struc- eling of HP-FucT (PDB code 2NZY) complexed to 4
ture provides useful insight to account for the differ- (Figure 4) suggests that a deep pocket formed by hy-
ent potencies of compounds 3–5. Computational mod- drophobic residues, including Trp33, Trp34, Phe42,
Adv. Synth. Catal. 2012, 354, 1750 – 1758
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1753

Yu-Nong Lin et al.
FULL PAPERS
Figure 4. Molecular model of compound 4 in the active site of HP-FucT. The molecular structure of 4 is depicted in yellow
(carbon), red (oxygen), blue (nitrogen), green (chlorine) and orange (phosphorus). Hydrophobic residues of HP-FucT (PDB
code 2NZY) are shown in orange and other residues in white. A deep hydrophobic pocket consisting of residues Trp33,
Trp34, Val46, Leu47, Phe42, Phe71, and Tyr92 is proposed to interact with the hydrophobic group of compound 4. Arrows
A–C indicate the guanine, ribose, and diphosphate moieties of compound 4, respectively. Glu95 acts as a general acid/base
in the catalysis. The molecular model was created and displayed using Discovery Studio v2.5 and PyMOL v1.4.
Val46, Leu47, Phe71, and Tyr92, is near the active because compounds 3–5 are all proved to be competi-
site. The benzophenone moiety of 4 (Ki =0.22 mM) tive inhibitors (Figure 3 and Supporting Information,
may fit well into the hydrophobic pocket and thereby Figure S4–Figure S10). A variety of carboxylic acids
enhance the binding. Compound 3 (Ki =0.9 mM), were incorporated as amides and examined to match
however, may be a weaker fit for the pocket owing to the different surfaces of the three a-1,3-FucTs. As
its bulky biphenyl group. The branched cyclopentyl a result, compounds 4 and 5 have highly improved af-
group of 5 (Ki =7.8 mM) was found to be a poor fit finities for HP-FucT and FucT-VI, respectively, com-
for the hydrophobic pocket, and this would explain its pared to GDP-Fuc (Km =11.3 mM for HP-FucT, and
low affinity.
Km =13.2 mM for FucT-VI; Table 1). The additional
We previously had success in developing potent structural motifs of GDP-Fuc and 5 apart from GDP
and selective inhibitors for a-fucosidase[32–34] and b- (i.e., the groups of fucose and hydrophobically C-6-
hexosaminidase.[35,36] In these studies, even though the substituted fucose) contribute a 2.5- and 1100-fold en-
same type of enzymes share high sequence similarity hancement of inhibition against FucT-VI, as com-
and use the same catalytic mechanism, their protein pared to GDP (Ki =33 mM). This large difference in
surfaces are very different. Thus, achieving selective binding affinity underscores the validity of our ap-
inhibition is possible by choosing a suitable group(s) proach of efficiently enhancing inhibition potency.
for optimizing the binding interactions. A similar
Previous studies have suggested that the binding in-
strategy was applied for this study. We expected the teraction between GDP-Fuc and FucT could be en-
GDP-Fuc moiety to occupy the conserved active site hanced by the addition of a suitably hydrophobic
1754
asc.wiley-vch.de
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2012, 354, 1750 – 1758

Chemoenzymatic Synthesis of GDP-l-Fucose Derivatives
drich. Anion exchange chromatography was performed on
a HiTrap Q Sepharose column (GE Healthcare Life Scien-
ces, Uppsala, Sweden).
group to the C-6 of Fuc, or directly to GDP without
the sugar residue.[19,21] This suggests that both the in-
corporated hydrophobic group and the triazole may
contribute to the improved inhibition. Maeda et al.
prepared two GDP-Fuc-derived compounds by apply-
ing either click chemistry or the amide-forming reac-
tion to couple GDP-Fuc with a naphthylmethyl
group.[37] Despite their similar structures, the triazole-
containing compound was found to be an inhibitor of
FucT-VI, whereas the amide-containing compound
turned out to be a good substrate. Instead of being
substrates, compounds 3–5 are potent FucT inhibitors.
Conversely, compound 2 is neither a substrate nor an
inhibitor for the a-1,3-FucTs tested. We observed
a similar phenomenon in the case of FKP, which ac-
cepted a wide range of Fuc analogues but failed to
Plasmid Construction, Protein Expression and
Purification
DNA manipulations were carried out according to standard
procedures. For the construction of pET28-FKP, the 3-kb
coding sequence of FKP was amplified from B. fragilis
NCTC9343 genomic DNA by PCR using the forward
primer,
5’-AACTCTGACCATATGCAAAAACTACTATCTT
TACCGTCC-3’, and reverse primer, 5’-CCAATACTCGAGTTAT
GATCGTGATACTTGGAATCCCTT-3’, to create the underlined
NdeI and XhoI restriction enzyme sites, respectively. After
digestion with NdeI and XhoI, the PCR fragment was first
ligated into the pGEM-T EASY vector (Promega, Fitch-
utilize 6-amino-l-fucose as a substrate. The same mol- burg, WI) and then subcloned into the pET28a vector to
yield pET28-FKP. The desired constructs were verified by
DNA sequencing. E. coli BL21 (DE3) was transformed with
pET28-FKP to overexpress the N-terminal hexahistidine
(His6)-tagged FKP. The clone harboring pET28-FKP was
cultured in lysogeny broth (containing 50 mgmLÀ1 kanamy-
cin) overnight at 378C with vigorous shaking. The overnight
culture was then diluted 1:1000 (v/v) with the same medium
and grown at 378C until cells reached an A600 of 0.7. Protein
expression was induced by addition of isopropyl-1-thio-b-d-
galacto-pyranoside to a final concentration of 0.2 mM, and
the culture was grown for an additional 24 h at 168C. Cells
were harvested by centrifugation at 10000ꢁg for 10 min at
48C, resuspended in phosphate-buffered saline, and lysed
using a French Press. Recombinant FKP was purified by
a single step of Ni2+ affinity chromatography. The desired
protein fractions were collected, concentrated and desalted
(against 50 mM pH 8.0 Tris-HCl) using a 50 kDa molecular
weight cut-off Amicon Ultra-15 centrifugal filter (Millipore,
Taiwan). For the preparation of human FucT-V and -VI, the
DNA fragment encoding the desired FucT was amplified by
PCR, and the sequence corresponding to N-terminal trans-
membrane domain was removed. The amplified fragment
was digested with NdeI and XhoI and ligated into the
pFLAG-CMVꢂ-3 vector (Sigma–Aldrich). HEK293T cells
were transfected with the FucT plasmid using calcium phos-
phate. The stable 293T cell lines expressing FLAG-tagged
FucT were selected and cultured in Dulbeccoꢃs modified
Eagleꢃs medium supplemented with 10% fetal bovine
serum, 2 mM glutamine, 100 U/mL penicillin, and
100 mgmLÀ1 streptomycin at 378C in a humidified atmos-
phere with 5% CO2.
For purification of FucT, cells were harvested by treat-
ment with trypsin/EDTA. After washing with phosphate-
buffered saline twice, the cell pellets were suspended in lysis
buffer containing 0.5% Nonidet P-40, 25 mM hydroxyethyl
piperazineethanesulfonic acid (Hepes, pH 7.9), 200 mM
KCl, 5 mM EDTA, 1 mM phenylmethylsulfonyl fluoride,
5 mM dithiothreitol, 10 mM NaF, 0.1 mM Na3VO4, 5% glyc-
erol and a protease inhibitor cocktail (Sigma–Aldrich).
After 20 min on ice, cell suspensions were homogenized by
sonication then centrifuged at 13,000ꢁg for 10 min at 48C
to remove the nuclear and cellular debris. The resulting su-
pernatant was incubated with an antibody against FLAG
that had been coupled to agarose beads (Sigma–Aldrich) at
ecule is neither an inhibitor of FKP. It is intriguing
that compound 2 showed no activity for the FucTs,
yet it was converted to a potent inhibitor after incor-
poration of the optimal hydrophobic moiety. In addi-
tion, high selectivity was also achieved, which enabled
distinction between different FucTs.
Conclusions
In summary, this chemoenzymatic method was devel-
oped to discover potent and selective FucT inhibitors,
and assess them in situ using an activity assay. Despite
improved inhibition potency, several critical issues
remain to be addressed. For instance, the pyrophos-
phate group needs to be replaced to allow delivery
across the cell membrane. To increase stability, substi-
tution of the glycosidic linkage of compounds 3–5 is
currently being investigated, and our progress will be
reported in due course.
Experimental Section
Materials
Chemical reagents were used as purchased from Sigma–Al-
drich (St. Louis, MO) unless otherwise indicated. Reactions
and assays were monitored by TLC on silica gel 60 F254
(Merck, Darmstadt, Germany) and visualized under UV,
cerium ammonium molybdate stain, or by isotope signal.
GDP-l-[U-14C]fucose (GDP-[14C]Fuc) was purchased from
Perkin–Elmer (Boston, MA). NMR spectra were recorded
routinely on a Bruker AV-400 (400 MHz) or AVII-500
(500 MHz) spectrometer with D2O (dH =4.80) as the inter-
nal standard.
B. fragilis NCTC9343 genomic DNA was obtained from
Bioresource Collection and Research Center (Hsinchu,
Taiwan). Restriction enzymes and alkaline phosphatase
were purchased from New England Biolabs (Ipswich, MA).
Inorganic pyrophosphatase was purchased from Sigma–Al-
Adv. Synth. Catal. 2012, 354, 1750 – 1758
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1755

Yu-Nong Lin et al.
FULL PAPERS
48C for 2 h. After extensive washing with lysis buffer,
FLAG-tagged FucT was eluted with 100 ng/mL 3ꢁ FLAG
peptide (Sigma–Aldrich). HP-FucT was prepared as de-
scribed.[18]
titer plate, 8 mL of reaction buffer containing 50 mM pH 7.5
Tris-HCl buffer, LacNAc (due to the different Km values of
each enzyme, the concentration of LacNAc was 20, 5 and
2 mM for FucT-V, FucT-VI and HP-FucT, respectively),
10 mM MnCl2, 50 munit of FucT were mixed with an aliquot
(1 mL) of the mixture from previous microtiter plate to give
~1 mm inhibitor concentration. Assays were initiated by
adding 1 mL of GDP-Fuc (from a 100 mM stock solution
containing 25% GDP-[14C]Fuc) to each well at 378C.
Enzyme activity was detected by measuring the incorpora-
tion of radioactive label from GDP-[14C]Fuc into the reac-
tion product. Samples were taken at defined time points (30,
90, 120 min) and analyzed by TLC using a 7:1:3 (v/v/v) solu-
tion of i-PrOH:AcOH:H2O. Radioactivity was detected by
phosphoimaging using a BAS-MS 2040 imaging plate and
BAS-1500 PhosphorImager (Fujifilm, Tokyo, Japan). Signals
were evaluated and quantified using the software Image
Gauge, v4.0 (Fujifilm).
Synthesis of Guanosine 5’-Diphospho-6-amino-b-l-
fucopyranoside (2)
Guanosine 5’-diphospho-6-azido-b-l-fucopyranoside, 1, was
enzymatically synthesized by FKP from 6-azido-b-l-fucose
(Scheme 1). Reduction of the azide to an amine was per-
formed over Pd/C to give 2. A 25 mL mixture of 50 mM
Tris-HCl (pH 8.2), 10 mM 6-azido-l-fucose (36 mg, 200
mmol), 11 mM ATP (113 mg, 220 mmol), 5.5 mM GTP
(108 mg, 220 mmol), 10 mM MnCl2, inorganic pyrophospha-
tase (20 units), and FKP (5 units) was incubated at 378C for
12 h. After the reaction was complete, as judged by TLC
using a 7:1:3 (v/v/v) solution of i-PrOH:AcOH:H2O, the
sample was treated with 5 units of alkaline phosphatase for
5 h at 378C to hydrolyze the remaining nucleoside phos-
phates. After removing the precipitated magnesium phos-
phate by centrifugation, the resulting sample was applied to
the HiTrap Q Sepharose column. The column was initially
washed with 10 mL water, then the retained materials were
Synthesis of Compounds 3–5
Compounds displaying strong inhibition activity were select-
ed, synthesized, and purified for further investigation. To
avoid formation of side products that are troublesome to
remove, DIEA was not used, and the NHS-ester was added
in 5-fold excess for the amide-forming reactions. To synthe-
size 5, a mixture of G6 (20 mg, 100 mmol), EDC (172 mg,
150 mmol), and NHS (21 mg, 110 mmol) in 10 mL CH2Cl2
was stirred at room temperature for 1 h. After TLC analysis
with hexane:ethyl acetate (v/v=1:1) to indicate the reaction
was complete, the mixture was washed with a saturated solu-
tion of NaHCO3, then dried over MgSO4, concentrated
under vacuum, and precipitated by the slow addition of di-
ethyl ether at 48C. The precipitate was washed with diethyl
ether to give the NHS-activated G6 (G6-OSu) as a white
powder; yield: 29 mg (96%).
G6-OSu (25 mg, 83 mmol, dissolved in 1 mL of DMF) was
mixed with 2 (10 mg, 16 mmol, dissolved in 0.5 mL of 50 mM
sodium phosphate buffer, pH 8.1) and stirred at 508C. The
reaction was monitored by TLC using a 7:1:3 (v/v/v) solution
of i-PrOH:AcOH:H2O. After the reaction was complete,
the mixture was washed with CH2Cl2 to remove DMF and
the excess G6-OSu. The solution was lyophilized, reconsti-
tuted in water, and the product purified using HiTrap Q Se-
pharose anion exchange chromatography as described above
to give 5 as a white powder; yield: 11.8 mg (90%). Com-
pounds 3 and 4 were synthesized in the same manner; yield:
92 and 83%, respectively.
eluted with
a 40 mL linear gradient of 50–200 mM
NH4HCO3. Eluents were monitored by UV absorption at
260 nm. The desired fractions were pooled, concentrated
under reduced pressure, and lyophilized to yield 101 mg of
1 as a white solid in 72% yield. For the synthesis of 2, a solu-
tion of 1 (30 mg, 300 mmol) and 10% Pd/C (8 mg) in water
(10 mL) was stirred under a hydrogen atmosphere for
15 min (the product decomposed when the reaction was car-
ried out in methanol). The mixture was filtered, and the fil-
trate was lyophilized to afford 2 as a white foam; yield:
1
28 mg (95%). H NMR (400 MHz, D2O): d=3.23–3.35 (m,
2H, H-6a’’, H-6b’’), 3.63 (dd, 1H, J=9.5, 7.8 Hz, H-2’’), 3.71
(dd, 1H, J2’’,3’’ =10.0 Hz, J3’’,4’’ =3.2 Hz, H-3’’), 3.92–3.98 (m,
2H, H-4’’, H-5’’), 4.22 (m, 2H, H-5a’, H-5b’), 4.36 (m, 1H,
H-4’), 4.51-4.53 (m, 1H, H-3’), 4.97 (dd, 1H, J1’’,2’’ =J1’’,P = 7.6
Hz, H-1’’), 5.94 (d, 1H, J1’,2’ =6.0 Hz, H-1’), 8.15 (s, 1H, H-
8); 13C NMR (125 MHz, D2O): d=39.7, 50.2, 65.3, 68.4, 70.3,
71.0, 72.0, 73.6, 83.6, 86.7, 98.4, 116.1, 137.4, 151.7, 153.8,
158.8; HR-MS (ESI): m/z=603.0930, calcd.
for
C16H25N6O15P2 [MÀH]À: 603.0848.
Amide-Forming Reactions of GDP-Fuc Derivatives
and in-situ FucT Activity Assay
Dimethylformamide (DMF, 12 mL) was aliquoted into each
well of a 96-well microtiter plate. The following reagents
were then added into each well: 1.1 mL of each carboxylic
acid (from a 50 mM stock solution in DMF), 1.1 mL of
50 mM O-(benzotriazol-1-yl)-N,N,N’,N’-tetramethyluronium
hexafluorophosphate (HBTU) dissolved in DMF (1.1 equiv-
alents), and 2 mL of 50 mM DIEA dissolved in DMF (2
equivalents). The reaction was initiated by adding 10 mL of
a 5 mM aqueous solution of 2 to each well, and progress
was monitored by TLC using a 7:1:3 (v/v/v) solution of i-
PrOH:AcOH:H2O. After 3 h, a 1 mL aliquot was withdrawn
from the reaction and mixed with 199 mL of water, i.e.,
a final concentration of 10 mM of the compound in each
well of the microtiter plate. To each well of another micro-
Compound 3: 1H NMR (400 MHz, D2O): d=3.42–3.50
(m, 2H, H-6a’’, H-6b’’), 3.60–3.68 (m, 4H, H-2’’, H-3’’, H-4’’,
H-5’’), 4.25–4.28 (m, 2H, H-5a’, H-5b’), 4.39 (m, 1H, H-4’),
4.57 (m, 1H, H-3’), 5.00 (dd, 1H, J1’’,2’’ =J1’’,P =7.7 Hz, H-1’’),
5.95 (d, 1H, J1,2’ = 6.0 Hz, H-1’), 7.46–7.65 (m, 9H, aromat-
ic), 8.13 (s, 1H, H-8); 13C NMR (125 MHz, D2O): d=41.5,
65.2, 68.5, 70.2, 71.2, 72.1, 72.3, 73.7, 83.6, 86.7, 98.3, 161.2,
127.2, 127.3, 128.1, 128.3, 129.1, 130.5, 130.8, 132.4, 137.4,
140.9, 151.7, 154.2, 159.2, 169.7; HR-MS (ESI): m/z=
783.1505, calcd. for C29H33N6O16P2 [MÀH]À: 783.1423.
1
Compound 4: H NMR (400 MHz, D2O): d=3.39-3.46 (m,
2H, H-6a’’, H-6b’’), 3.61–3.72 (m, 2H, H-2’’, H-3’’), 3.82–
3.89 (m, 2H, H-4’’, H-5’’), 4.20–4.21 (m, 2H, H-5a’, H-5b’),
4.34 (m, 1H, H-4’), 4.55–4.59 (m, H-3’), 5.94 (d, 1H, J1’,2’
=
1756
asc.wiley-vch.de
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2012, 354, 1750 – 1758

Chemoenzymatic Synthesis of GDP-l-Fucose Derivatives
6.0 Hz, H-1’), 7.40–7.89 (m, 8H, aromatic), 8.13 (s, 1H, H-
8); 13C NMR (125 MHz, D2O): d=42.5, 65.1, 68.2, 70.1, 70.8,
70.9, 71.8, 73.4, 83.4, 83.6, 86.7, 98.2, 116.5, 127.4, 128.8,
129.8, 130, 130.5, 132.7, 135.8, 137.5, 151.7, 153.9, 159.1,
[5] E. Staudacher, Trends Glycosci. Glycotechnol. 1996, 8,
391–408.
[6] Q. Y. Wang, S. L. Wu, J. H. Chen, F. Liu, H. L. Chen, J.
Exp. Clin. Cancer Res. 2003, 22, 431–440.
[7] E. Mas, E. Pasqualini, N. Caillol, A. El Battari, C.
Crotte, D. Lombardo, M. O. Sadoulet, Glycobiology
1998, 8, 605–613.
[8] S. R. Barthel, G. K. Wiese, J. Cho, M. J. Opperman,
D. L. Hays, J. Siddiqui, K. J. Pienta, B. Furie, C. J. Di-
mitroff, Proc. Natl. Acad. Sci. USA 2009, 106, 19491–
19496.
165.8; HR-MS (ESI): m/z=845.0972, calcd.
for
C30H32ClN6O17P2 [MÀH]À: 845.0982.
Compound 5: 1H NMR (500 MHz, D2O): d=0.93–0.95,
1.19–1.27, 1.46–1.55, 1.76, 2.54–2.56 (m, 9H, cyclopentane),
3.26–3.34 (m, 2H, H-6a’’, H-6b’’), 3.52–3.66 (m, 5H, H-2’’,
H-3’’, H-4’’, H-5’’, CH), 4.20–4.24 (m, 2H, H-5a’, H-5b’),
4.32–4.34 (m, 1H, H-4’), 4.50 (dd, 1H, J2’’,3’’ =4.5, J3’’,4’’
=
3.8 Hz, H-3’), 4.93 (dd, 1H, J1’’,2’’ =J1’’,P =7.9 Hz, H-1’’), 5.90
(d, 1H, J1’,2’ = 6.0 Hz, H-1’), 7.29–7.42 (m, 5H, aromatic),
8.14 (s, 1H, H-8); 13C NMR (125 MHz, D2O): d=24.4, 24.7,
30.0, 31.2, 41.2, 42.8, 57.1, 65.2, 68.4, 70.3, 70.9, 71.0, 72.0,
73.6, 83.6, 83.7, 86.7, 98.3, 116.2, 126.9, 128.1, 128.3, 137.5,
139.5, 151.7, 154.0, 159.1, 170.3; HR-MS (ESI): m/z=
789.1817, calcd. for C29H39N6O16P2 [MÀH]À: 789.1892.
[9] S. R. Barthel, J. D. Gavino, G. K. Wiese, J. M. Jaynes, J.
Siddiqui, C. J. Dimitroff, Glycobiology 2008, 18, 806–
817.
[10] J. J. Listinsky, C. M. Listinsky, V. Alapati, G. P. Siegal,
Adv. Anat. Pathol. 2001, 8, 330–337.
[11] N. Taniguchi, K. Honke, M. Fukuda, Handbook of Gly-
cosyltransferases and Related Genes, Springer, Tokyo,
2002.
[12] M. D. Burkart, S. P. Vincent, A. Dꢄffels, B. W. Murray,
S. V. Ley, C. H. Wong, Bioorg. Med. Chem. 2000, 8,
1937–1946.
[13] P. Compain, O. R. Martin, Bioorg. Med. Chem. 2001, 9,
3077–3092.
[14] M. L. Mitchell, F. Tian, L. V. Lee, C. H. Wong, Angew.
Chem. Int. Ed. Angew Chem. Int. Ed. Engl 2002, 41,
3041–3044.
Measurement of IC50 and Ki Values
To measure IC50 values, the activity assays were carried out
in 10 mL solutions containing 50 mM Tris-HCl (pH 7.5),
10 mM MnCl2, 20 mM GDP-Fuc (containing 25% [GDP-
[14C]Fuc), LacNAc (20 mM, 5 mM, and 2 mM for FucT-V,
FucT-VI, and HP-FucT, respectively), 50 munits of FucT, and
a variable inhibitor concentration (0.01–50 mM). Reactions
were conducted at 378C, and enzyme activity was detected
by monitoring the incorporation of radioactive label from
GDP-[14C]Fuc by TLC as described above.
[15] M. Izumi, S. Kaneko, H. Yuasa, H. Hashimoto, Org.
Biomol. Chem. 2006, 4, 681–690.
The experiments to determine Ki values were performed
with various concentrations of GDP-Fuc when one inhibitor
concentration was fixed. Assays were carried out in 10 mL
solutions containing 50 mM Tris-HCl buffer (pH 7.5),
10 mM MnCl2, LacNAc (20, 5 and 2 mM for FucT-V, FucT-
VI and HP-FucT, respectively), 50 munits of FucT, GDP-Fuc
(within the range of 5–50 mM, each contained 2.5 mM GDP-
[14C]Fuc), and inhibitor (0.01–20 mM). Reactions were con-
ducted at 378C, and enzyme activity was detected by radio-
TLC as described previously. Lineweaver–Burk plot analysis
was then carried out to determine the Ki values.
[16] M. Izumi, H. Yuasa, H. Hashimoto, Curr. Top. Med.
Chem. 2009, 9, 87–105.
[17] T. Kajimoto, M. Node, Synthesis 2009, 2009, 3179–3210.
[18] H. Y. Sun, S. W. Lin, T. P. Ko, J. F. Pan, C. L. Liu, C. N.
Lin, A. H. Wang, C. H. Lin, J. Biol. Chem. 2007, 282,
9973–9982.
[19] L. V. Lee, M. L. Mitchell, S. J. Huang, V. V. Fokin, K. B.
Sharpless, C. H. Wong, J. Am. Chem. Soc. 2003, 125,
9588–9589.
[20] M. C. Bryan, L. V. Lee, C. H. Wong, Bioorg. Med.
Chem. Lett. 2004, 14, 3185–3188.
[21] K. Hosoguchi, T. Maeda, J. Furukawa, Y. Shinohara, H.
Hinou, M. Sekiguchi, H. Togame, H. Takemoto, H.
Kondo, S. Nishimura, J. Med. Chem. 2010, 53, 5607–
5619.
[22] B. Mꢄller, C. Schaub, R. R. Schmidt, Angew. Chem.
1998, 110, 3021–3024; Angew. Chem. Int. Ed. 1998, 37,
2893–2897.
[23] B. Waldscheck, M. Streiff, W. Notz, W. Kinzy, R. R.
Schmidt, Angew. Chem. 2001, 113, 4120–4124; Angew.
Chem. Int. Ed. Angew. Chem. Int. Ed. Engl. 2001, 40,
4007–4011.
Acknowledgements
This work was supported by Academia Sinica (AS-99-TP-
AB4 and 100-AS-12) and the National Science Council of
Taiwan (100-2113M-001-021-MY3). C.H.L. is an Academia
Sinica Investigator Award recipient.
[24] R. Schwçrer, R. R. Schmidt, J. Am. Chem. Soc. 2002,
124, 1632–1637.
References
[25] H. Hinou, X. L. Sun, Y. Ito, J. Org. Chem. 2003, 68,
5602–5613.
[26] M. J. Coyne, B. Reinap, M. M. Lee, L. E. Comstock,
Science 2005, 307, 1778–1781.
[27] W. Wang, T. Hu, P. A. Frantom, T. Zheng, B. Gerwe,
D. S. Del Amo, S. Garret, R. D. Seidel, P. Wu, Proc.
Natl. Acad. Sci. USA 2009, 106, 16096–16101.
[1] D. J. Becker, J. B. Lowe, Glycobiology 2003, 13, 41R-
53R.
[2] T. de Vries, R. M. Knegtel, E. H. Holmes, B. A.
Macher, Glycobiology 2001, 11, 119R-128R.
[3] D. H. Dube, C. R. Bertozzi, Nat. Rev. Drug. Discov.
2005, 4, 477–488.
[4] B. Ma, J. L. Simala-Grant, D. E. Taylor, Glycobiology
2006, 16, 158R-184R.
Adv. Synth. Catal. 2012, 354, 1750 – 1758
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1757

Yu-Nong Lin et al.
FULL PAPERS
[28] D. Soriano del Amo, W. Wang, C. Besanceney, T.
Zheng, Y. He, B. Gerwe, R. D. Seidel, P. Wu, Carbo-
hydr. Res. 2010, 345, 1107–1113.
[29] V. Wittmann, C. H. Wong, J. Org. Chem. 1997, 62,
2144–2147.
[33] C. F. Chang, C. W. Ho, C. Y. Wu, T. A. Chao, C. H.
Wong, C. H. Lin, Chem. Biol. 2004, 11, 1301–1306.
[34] C. W. Ho, Y. N. Lin, C. F. Chang, S. T. Li, Y. T. Wu,
C. Y. Wu, S. W. Liu, Y. K. Li, C. H. Lin, Biochemistry
2006, 45, 5695–5702.
[30] M. Sawa, T. L. Hsu, T. Itoh, M. Sugiyama, S. R.
Hanson, P. K. Vogt, C. H. Wong, Proc. Natl. Acad. Sci.
USA 2006, 103, 12371–12376.
[35] C. W. Ho, S. D. Popat, T. W. Liu, K. C. Tsai, M. J. Ho,
W. H. Chen, A. S. Yang, C. H. Lin, ACS Chem. Biol.
2010, 5, 489–497.
[36] H. J. Wu, C. W. Ho, T. P. Ko, S. D. Popat, C. H. Lin,
A. H. Wang, Angew. Chem. 2010, 122, 347–350; Angew.
Chem. Int. Ed. 2010, 49, 337–340.
[31] P. Gonzalo, B. Sontag, D. Guillot, J. P. Reboud, Anal.
Biochem. 1995, 225, 178–180.
[32] C. Y. Wu, C. F. Chang, J. S. Chen, C. H. Wong, C. H.
Lin, Angew. Chem. 2003, 115, 4809–4812; Angew.
Chem. Int. Ed. 2003, 42, 4661–4664.
[37] T. Maeda, S. Nishimura, Chem. Eur. J. 2008, 14, 478–
487.
1758
asc.wiley-vch.de
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2012, 354, 1750 – 1758