Synthesis and biochemical properties of reversible inhibitors of UDP-N-acetylglucosamine 2-epimerase

Article abstract of DOI:10.1002/anie.200453863

Sialic acids have important biochemical functions in healthy organisms, for example, in embryogenesis, as well as in the spread of diseases like influenza. A key step in the biosynthesis of the sialic acid N-acetylneuraminic acid is the epimerization of 1

Full text of DOI:10.1002/anie.200453863

Communications
Sialic Acids
Synthesis and Biochemical Properties of
Reversible Inhibitors of UDP-N-
Acetylglucosamine 2-Epimerase
Samy Al-Rawi, Stephan Hinderlich,* Werner Reutter,
and Athanassios Giannis*
Sialic acids comprise a family of a-keto acids with a backbone
skeleton consisting of nine carbon atoms. N-Acetylneura-
minic acid (Neu5Ac or NANA; see Scheme 1) is the most
common sialic acid and also the biosynthetic precursor of
[1]
almost all known 50members of this class.
Due to their
exposed position at the distal end of oligosaccharide chains,
sialic acids contribute to the biological properties of glyco-
[*] Dr. S. Hinderlich, Prof. Dr. W. Reutter
CharitØ – Universitätsmedizin Berlin
Campus Benjamin Franklin
Institut für Molekularbiologie und Biochemie
Arnimallee 22, 14195 Berlin-Dahlem (Germany)
Fax : (+49)30-8445-1541
E-mail: stephan.hinderlich@charite.de
Dipl.-Chem. S. Al-Rawi, Prof. Dr. A. Giannis
Institut für Organische Chemie
Universität Leipzig
Johannisallee 29, 04103 Leipzig (Germany)
Fax : (+49)341-973-6599
E-mail: giannis@chemie.uni-leipzig.de
4366
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/anie.200453863
Angew. Chem. Int. Ed. 2004, 43, 4366 –4370

Angewandte
Chemie
Scheme 1. Biosynthesis of Neu5Ac and incorporation on cell-surface glycoconjugates.
conjugate structures in a significant manner. Glycoconjugates
containing sialic acid play an essential role in a variety of
fundamental physiological and pathological processes,[2] for
example, embryogenesis, organogenesis, immune defense,
migration and homing of leucocytes, and metastasis of
neoplastic cells as well as inflammation reactions and invasion
of pathogens in cells. Many tumors bear a large amount of
sialic acids like a protective shield on their cell surface. The
most prominent example of sialic acid mediated cell–virus
interaction is the infection by the influenza virus. The invasion
of cells, as well as the release of newly formed viruses is
mediated by sialic acid containing structures on the host cell
surface. Inhibitors of influenza sialidase, the virus-releasing
enzyme, are currently applied in antiviral therapy (Zanamivir,
Oseltamivir).[3] Although all steps of Neu5Ac biosynthesis
have been known for a long time,[4] only a limited number of
inhibitors have been developed.[5,6] In this paper we report the
conception, synthesis, and biochemical properties of a potent
reversible inhibitor of UDP-N-acetylglucosamine (GlcNAc)
2-epimerase (UDP = uridinediphosphate, GlcNAc = N-ace-
tylglucosamine), the key enzyme in Neu5Ac biosynthesis.
Biosynthesis of Neu5Ac begins with the epimerization of
UDP-GlcNAc under concomitant release of UDP, followed
by phosphorylation of the formed N-acetylmannosamine
(ManNAc) at the 6-hydroxy functionality. Both steps are
catalyzed by the bifunctional enzyme UDP-GlcNAc 2-
epimerase/ManNac kinase (Scheme 1).[7,8] This key enzyme
is regulated in a complex manner.[7–10] Lymphocytes that show
a deficiency in activitiy of this enzyme are not capable of
synthesizing sialic acids by themselves and show serious
defects in sialic acid dependent functions.[11] Inactivation
through targeted mutagenesis in mice causes hyposialylation
of the embryos and induces early lethality on day 8.5 of
embryonic development.[12] These results emphasize the
importance of UDP-GlcNAc 2-epimerase in the biosynthesis
of N-acetylneuraminic acid and make this enzyme an
interesting target for the development of effective inhibitors.
The homologous mammalian UDP-GlcNAc 2-epimerase
has recently become available in a recombinant form and
does not differ in mechanism[13] from the bacterial[14] enzyme.
Both enzymes first generate 2-acetamidoglucal as an inter-
mediate after anti-elimination of UDP. Positional isotope
exchange experiments showed that the reaction proceeds by
ꢀ
C O bond cleavage at the anomeric position. By monitoring
the reaction with coupled NMR experiments, we were able to
determine that the subsequent addition of water occurs from
the bottom side of the 2-acetamidoglucal. Thus, the released
ManNAc is the a-anomer (Scheme 2). The oxocarbenium
intermediate might represent the transition state of the
addition of water on the 2-acetamidoglucal.
Because stable analogues of the transition state are known
to be bound much more tightly to the active site of the enzyme
than analogues of the substrate in its ground state,[15] we
considered the synthesis of iminosugars 2–4 (Figure 1) as
potential inhibitors of UDP-GlcNAc 2-epimerase. At phys-
iological pH such iminosugars are protonated and they should
consequently be good analogues of the proposed transition
state.
The bicyclic oxazolidinylpiperidine of type 9 (Scheme 3)
served as the central, chiral building block in the preparation
Angew. Chem. Int. Ed. 2004, 43, 4366 –4370
www.angewandte.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4367

Communications
Scheme 2. Proposed reaction mechanism of mammalian UDP-GlcNAc
2-epimerase and postulated transition state for the syn-addition of
water at the 2-acetamidoglucal.
of the 1-deoxyiminosugar. According to
the previous work by Martin et al.,[16] the
doubly unsaturated alcohol 5 was stereo-
selectively epoxidized under Sharpless
Figure 1. Transition-
state analogues of
UDP-GlcNAc 2-epi-
merase.
conditions.[17] Subsequent conversion of
Scheme 3. Synthesis of the protected iminosugar 15: a) Ti(OiPr)4,
l-(+)-DIPT, tBuOOH, CH2Cl2, 3-Š molecular sieves, ꢀ158C, 57%;
b) BzNCO, Et2O, RT, 84%; c) K2CO3, (C8H17)3CH3NCl, CH2Cl2/H2O, RT,
84%; d) NaH, allyl bromide, DMF, 08C!RT, 81%; e) Grubbs catalyst,
toluene, RT, 99%; f) K2CO3, MeOH, RT, 85%; g) NaH, BnBr, DMF,
08C!RT, 91%; h) (DHQD)2Phal, K2OsO4, K2CO3, K3[Fe(CN)6],
CH3SO2NH2, tBuOH/H2O, RT, 90%; i) 1. nBu2SnO, toluene, reflux;
2. BnBr, Bu4NBr, 1008C, 69%; j) DMSO, (COCl)2, DIPEA, CH2Cl2,
ꢀ708C!RT, 86%; k) 4-methoxybenzylamine, AcOH, Na(OAc)3BH, 1,2-
dichloroethane, RT, 78%; l) pyridine, DMAP, AcCl, 08C!RT, 86%;
m) 1. CAN, CH3CN/H2O, RT; 2. Ba(OH)2 ”8H2O, EtOH/H2O, reflux,
63%. Bz=benzoyl, Bn=benzyl, CAN= ceric ammonium nitrate,
(DHQD)2Phal=dihydroquinidine 1,4-phthalazinediyl ether,
the (2S,3S)-epoxyalcohol 6 with benzoyl-
isocyanate gave the benzoylcarbamate 7.
Regioselective opening of the epoxide
2: R=H, 3:
under mild conditions in a biphasic
system lead to an intramolecular migra-
tion of the benzoyl group to afford the
substituted oxazolidinone 8. The com-
R=nBu, and 4:
R=CH2CH2Ph.
pound was first allylated at the ring nitrogen then subjected to
a ring-closing metathesis to give the bicyclic derivate 9a in
almost quantitative yield. With regard to the following
synthetic steps, it was necessary to convert the ester group
of the piperidine ring into a benzyl ether. Compound 9c was
synthesized from 9a in two steps. By using catalytic amounts
of potassium osmate and N-methylmorpholine-N-oxide
(NMMO) as the cooxidant in a mixture of tBuOH/water,
we converted compound 9c into the corresponding diol. In
this reaction no acceptable diastereomeric excess could be
achieved; however, with (DHQD)2Phal as chiral auxiliary[18]
and K3[Fe(CN)6] as cooxidant, bishydroxylation of compound
10 resulted with nearly complete diastereoselectivity in very
good yields. The structure was determined by NOE and
COSY experiments.
DIPEA=N,N-diisopropylethylamine, DIPT=diisopropyl tartrate,
DMAP=4-dimethylaminopyridine, PMB= p-methoxybenzyl.
Reaction of compound 10 with dibutyltin oxide gave as
intermediate the corresponding dibutyltin ketal, which was
converted into compound 11 using benzyl bromide in the
presence of tetrabutylammonium bromide[19] with a regiose-
lectivity of 3.5:1. Swern oxidation of 11 yielded ketone 12.
Reductive amination of this derivative with p-methoxyben-
zylamine and sodium triacetoxyborohydride resulted in com-
pound 13. Acetylation of compound 13 gave the amide 14 in
good yield. Oxidative cleavage of the p-methoxybenzyl group
in the presence of cerium(iv) ammonium nitrate (CAN)
followed by opening of the oxazolidinone ring with barium
hydroxide afforded compound 15. After cleavage of the
protecting groups of compound 15 by hydrogenolysis, target
structure 2 was finally formed (Scheme 4).[20] For subsequent
application of the inhibitors in cellular systems and for the
Scheme 4. Derivatization and deprotection: a) Butyraldehyde, AcOH,
Na(OAc)3BH, 1,2-dichloroethane, RT, 95%; b) phenylacetaldehyde,
AcOH, NaCNBH3, CH3CN, RT, 85%; c) H2, Pd/C, HCl, EtOH, RT, 2:
87%, 3: 90%, 4: 90%.
determination of new structure–activity relationships, it could
be helpful to synthesize more lipophilic N-alkylated deriva-
tives like compounds 3 and 4. The synthesis succeeded
without any problems by reductive amination of 15 in the
4368
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 4366 –4370

Angewandte
Chemie
presence of the corresponding aldehydes and gave derivatives
16 and 17, which provided compounds 3 and 4 after hydro-
genolysis (Scheme 4).
nearly 10%. The presence of UDP in the preincubation
samples prevented enzyme inhibition in almost all experi-
ments. Addition of UDP and/or UDP-GlcNAc after an
incubation of 20min had nearly no effect on the inhibition
by 2–4. These data suggest that the inhibitors are bound very
tightly by the enzyme as soon as they have reached the active
site. This observation supports the hypothesis that transition-
state analogues are bound much more tightly by the enzyme
than the physiological substrate. Although the binding con-
stants of the inhibitors were not determined in these experi-
ments, they should be lower than the binding constants of
UDP and UDP-GlcNAc, since the subsequent addition of
both substrates does not displace the inhibitors.[22] The Km
values of UDP-GlcNAc and UDP are about 10 mm; thus the
Ki values of the inhibitors should be lower.
The specificity of the inhibitors was determined by the
lack of inhibition of the natural sugars ManNAc and GlcNAc,
which were tested under the same conditions as the inhibitors.
Neither the activity of the ManNAc kinase of the bifunctional
UDP-GlcNAc 2-epimerase/ManNAc kinase was influenced
by the inhibitors nor was its hexameric oligomeric structure
affected, which serves as an indicator of structural integrity.[7]
By comparing the inhibitors it seems to be noteworthy to
mention that the ring nitrogen can be substituted with
hydrophobic residues without affecting inhibitory potency.
This might be important for subsequent applications in
cellular systems, since hydrophobic compounds permeate
the plasma membrane much more easily. The application of
these inhibitors, for which we propose the name nanastatines,
opens up the possibility of better understanding the role of N-
acetylneuraminic acid containing conjugates in physiological
and pathological processes.
In tests with standard enzyme preparations[21] compounds
2–4 did not show any inhibition of UDP-GlcNAc 2-epimerase
activity in either concentrations equimolar to (1 mm) or in
concentrations up to four times higher of that of the substrate
UDP-GlcNAc. These results show that the 1,2-dideoxy-2-
acetamidomannojirimycin derivatives, if they are effective
inhibitors, are not able to reach the active site of the enzyme,
perhaps because they are blocked by UDP-GlcNAc or UDP,
which is used in standard preparations to stabilize the enzyme.
Further experiments were then performed using enzyme
preparations containing no UDP.[21] Preincubation of these
samples in the presence of the inhibitors 2–4 led to clear
inhibition of enzyme activity (up to 70%). When 0.1 m m UDP
was added to the enzyme samples before addition of the
inhibitors and then preincubation, no significant inhibition
(under 10%) was observed. Furthermore, no inhibition was
achieved when UDP-free enzyme preparations were used in
the enzyme activity experiments without any preincubation,
that is, enzyme, inhibitor, and UDP-GlcNAc were mixed
simultaneously. These experiments show that the synthesized
compounds are readily able to bind to the active site of the
enzyme, if it is not blocked by UDP and/or by UDP-GlcNAc.
In order to compare the effectiveness of the inhibitors 2–4,
they were tested under various concentrations (Figure 2). All
three compounds examined inhibited the activity of UDP-
GlcNAc 2-epimerase in a concentration-dependent manner.
Inhibition of 50% activity was observed at a concentration of
0.5 mm. Concentrations of 2 and 3 up to 2 mm in the
preincubation sample induced almost complete inhibition,
while treatment with 4 resulted in a remaining activity of
Received: January 27, 2004
Revised: May 19, 2004 [Z53863]
Keywords: biosynthesis · carbohydrates · epimerases ·
.
inhibitors · sialic acids
[1] T. Angata, A. Varki, Chem. Rev. 2002, 102, 439.
[2] S. Kelm, R. Schauer, Int. Rev. Cytol. 1997, 175, 137.
[3] M. J. Kiefel, M. von Itzstein, Chem. Rev. 2002, 102, 471.
[4] a) A. P. Corfield, R. Schauer in Sialic Acids—Chemistry, Metab-
olism and Function (Ed.: R. Schauer), Springer, Vienna, 1982,
p. 195; b) A. Varki, Glycobiology 1993, 3, 97.
[5] R. Zeitler, A. Giannis, S. Danneschewski, E. Henk, T. Henk, C.
Bauer, W. Reutter, K. Sandhoff, Eur. J. Biochem. 1992, 204, 1165.
[6] a) A. Blume, H. Chen, W. Reutter, R. R. Schmidt, S. Hinderlich,
FEBS Lett. 2002, 521, 127; b) F. Stolz, M. Reiner, A. Blume, W.
Reutter, R. R. Schmidt, J. Org. Chem. 2004, 69, 665 – 679.
[7] S. Hinderlich, R. Stäsche, R. Zeitler, W. Reutter, J. Biol. Chem.
1997, 272, 24313.
[8] R. Stäsche, S. Hinderlich, C. Weise, K. Effertz, L. Lucka, P.
Moormann, W. Reutter, J. Biol. Chem. 1997, 272, 24319.
[9] S. Kornfeld, R. Kornfeld, E. F. Neufeld, P. J. OꢀBrien, Proc. Natl.
Acad. Sci. USA 1964, 61, 371.
[10] R. Horstkorte, S. Nöhring, K. Danker, K. Effertz, W. Reutter, L.
Lucka, FEBS Lett. 2000, 470, 315.
Figure 2. Inhibition of UDP-GlcNAc 2-epimerase by compounds 2–4.
&
^
^
*
*
: 2, : 2 + UDP, : 3, &: 3 + UDP, : 4, : 4 + UDP. The enzyme
was preincubated with the inhibitors for 20 min at 378C in the absence
or presence of 0.1 mm UDP and subsequently used in the activity
assay. Noted concentrations refer to the final volume in the activity
assay. Data shown above represent mean values ꢁ standard deviation
of three independent experiments.
[11] O. T. Keppler, S. Hinderlich, J. Langner, R. Schwartz-Albiez, W.
Reutter, M. Pawlita, Science 1999, 284, 1372.
Angew. Chem. Int. Ed. 2004, 43, 4366 –4370
www.angewandte.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4369

Communications
[12] M. Schwarzkopf, K.-P. Knobeloch, E. Rohde, S. Hinderlich, N.
[21] Recombinant UDP-GlcNAc-2-epimerase from insect (Sf9) cells
was purified as described by A. Blume, D. Ghaderi, V. Liebig, S.
Hinderlich, P. Donner, W. Reutter, L. Lucka, Protein Expression
Purif. 2004, in press. For purification of UDP-free preparations
the samples were applied to a PD-10column (Amersham) and
the column was eluted with 10m m NaH2PO4 (pH 7.5), 100 mm
NaCl, 1 mm EDTA, 1 mm dithiotreitol. UDP-GlcNAc 2-epimer-
ase assays contained in a final volume 200 mL 45 mm NaH2PO4
(pH 7.5), 10m m MgCl2, 1 mm UDP-GlcNAc, 0.2 mm NADH,
2 mm phosphoenolpyruvat, 2 U pyruvat kinase, 2 U lactate
dehydrogenase, 5 mg purified UDP-GlcNAc 2-epimerase/
ManNAc kinase, and indicated concentrations of inhibitors.
The samples were incubated for 20min at 37 8C and quenched by
addition of 800 mL 10 mm EDTA. Subsequently the extinction
was determined at 340nm. Preincubation samples contained
5 mg purified UDP-GlcNAc 2-epimerase/ManNAc kinase and
variable amounts of inhibitors in a final volume of 20 mL. The
samples were incubated for 20min at 37 8C, and enzyme reaction
was started by addition of remaining components in a final
volume of 200 mL. Inhibitor concentrations noted in the text
refer to the final volume.
Wiechens, L. Lucka, I. Horak, W. Reutter, R. Horstkorte, Proc.
Natl. Acad. Sci. USA 2002, 99, 5267.
[13] W. K. Chou, S. Hinderlich, W. Reutter, M. E. Tanner, J. Am.
Chem. Soc. 2003, 125, 2455.
[14] P. M. Morgan, R. F. Sala, M. E. Tanner, J. Am. Chem. Soc. 1997,
119, 10269.
[15] L. Pauling, Chem. Eng. News 1946, 24, 1375.
[16] R. Martin, A. Moyano, M. A. Pericas, A. Riera, Org. Lett. 2000,
2, 93.
[17] Y. Gao, J. M. Klunder, R. M. Hanson, H. Masamune, S. Y. Ko,
K. B. Sharpless, J. Am. Chem. Soc. 1987, 109, 5765.
[18] K. B. Sharpless, W. Amberg, Y. L. Bennani, G. A. Crispino, J.
Hartung, K. S. Jeong, H. L. Kwong, K. Morikawa, Z. M. Wang,
D. Xu, X. L. Zhang, J. Org. Chem. 1992, 57, 2768.
[19] H. Böshagen, F. R. Heiker, A. M. Schꢁller, Carbohydr. Res.
1987, 164, 141.
[20] Compound 2 was already synthesized earlier using a chemo-
enzymatic strategy as well as starting from suitable functional-
ized carbohydrate structures and tested as an inhibitor for a wide
range of glycosidases. However, no inhibitory effect was
detected. a) G. W. Fleet, L. E. Fellows, P. W. Smith, Tetrahedron
1987, 43, 979; b) T. Kajimoto, K. K. C. Liu, R. L. Pederson, Z.
Zhong, Y. Ichikawa, J. A. Porco, Jr., C. H. Wong, J. Am. Chem.
Soc. 1991, 113, 6187.
[22] In this context it might be interesting to use chimeric com-
pounds, which consist of the iminosugars described here and
specific parts of UDP-GlcNAc. Currently we are synthesizing
such derivatives.
4370
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 4366 –4370