Synthesis and biological evaluation of new UDP-GalNAc analogues for the study of polypeptide-α-GalNAc-transferases

Article abstract of DOI:10.1016/S0960-894X(03)00287-7

A series of three O-methylated UDP-GalNAc analogues have been synthesised using a divergent strategy from a 3,6-di-O-pivaloyl GlcNAc derivative. The biological activity of these probes toward polypeptide-α-GalNAc-transferase T1 has been investigated. This study shows that this glycosyltransferase exhibits a very high substrate specificity.

Full text of DOI:10.1016/S0960-894X(03)00287-7

Bioorganic & Medicinal Chemistry Letters 13 (2003) 1853–1856
Synthesis and Biological Evaluation of New UDP-GalNAc
Analogues for the Study of Polypeptide-ꢀ-GalNAc-transferases
Patricia Busca,a Veronique Piller,b Friedrich Pillerb and Olivier R. Martina,*
a
´
´
´
Institut de Chimie Organique et Analytique, Universite d’Orleans and CNRS UMR 6005, BP 6759, 45067 Orleans, France
b
´
Centre de Biophysique Moleculaire, CNRS UPR 4301, 45071 Orleans, France
´
Received 31 January 2003; revised 17 March 2003; accepted 19 March 2003
Abstract—A series of three O-methylated UDP-GalNAc analogues have been synthesised using a divergent strategy from a 3,6-di-
O-pivaloyl GlcNAc derivative. The biological activity of these probes toward polypeptide-a-GalNAc-transferase T1 has been
investigated. This study shows that this glycosyltransferase exhibits a very high substrate specificity.
# 2003 Elsevier Science Ltd. All rights reserved.
It is becoming increasingly evident that glycoproteins,
as markers in cell–cell communication events, are play-
ing crucial regulatory functions in living organisms.
Moreover, alteration of N- or O-linked oligosaccharides
has been shown to be closely related to many diseases,
including for instance the progression of cancer.1
nucleotide substrate, we decided to investigate a series
of mono-O-methylated UDP-GalNAc derivatives (1–3,
Fig. 1). An analogous approach gave relevant results
regarding the glycosyl transfer of b-1,4-GalT.7
The evaluation of these probes as substrates or inhibi-
tors of GalNAc transferases would provide information
about steric allowance as well as possible requirement of
free OH groups for hydrogen bonding. To the best of
our knowledge, this study is the first one dedicated to
the donor substrate specificity of the ppGalNAcTs. In
this communication, we describe the synthesis of UDP-
GalNAc analogues 1–3 and their behaviour towards
UDP-GalNAc transferase T1.
The discovery that tunicamycin is a powerful inhibitor
of the protein N-glycosylation pathway led to significant
progress in understanding both biosynthesis and prop-
erties of N-glycosylated proteins.2 In contrast, despite
its biological importance, many aspects of the O-glyco-
sylation process remain to be elucidated.3,4
Polypeptide-a-GalNAc-transferases (ppGalNAcTs) cat-
alyse the first glycosylation step in the biosynthesis of
O-glycosylated proteins (Scheme 1) by transferring a
GalNAc unit from UDP-GalNAc to the hydroxyl group
of a serine or a threonine residue with retention of the
anomeric configuration.5
Synthesis of Probes
The syntheses of mono-O-substituted sugars are usually
based on selective protection/deprotection strategies
which are often time-consuming. Recently, an efficient
three-step procedure for the preparation of GalNAc
Therefore, in order to further study the O-glycosylation
process and to elucidate the importance of the O-gly-
cans, there is a great interest in finding inhibitors of
ppGalNAcTs. For that purpose, one of the best strategy
in the field of glycosyltransferases is the design of
bisubstrate-type inhibitors.6 In order to determine
where a linker arm could be attached to the sugar
*Corresponding author. Tel.:+33-2-3849-4581; fax:+33-2-3841-7281;
e-mail: olivier.martin@univ-orleans.fr
Scheme 1. First step in the biosynthesis of O-glycosylated proteins.
0960-894X/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00287-7

1854
P. Busca et al. / Bioorg. Med. Chem. Lett. 13 (2003) 1853–1856
Among the various methodologies available to intro-
duce the MTM group (mainly used as a protecting
group), only one is performed under very mild condi-
tions compatible with the ester functions.13 Thus, com-
pounds 6, 7 and 8 were reacted with 4 equivalents Bz2O2
and 8 equivalents Me2S in MeCN at 0 ꢀC to provide
their corresponding MTM derivatives 9, 10 and 11 in
good yields.
Figure 1. Mono-O-methylated analogues of UDP-GalNAc.
units starting from GlcNAc was reported, in which 3,4-
and 4,6-di-O-pivaloyl-d-galacto derivatives 6 and 7 were
the key intermediates.8 Therefore, taking advantage of
these results, the three precursors necessary for a regio-
selective methylation (6, 7 and 8) were efficiently syn-
thesised in four steps starting from GlcNAc (Scheme 2).
Hydrogenolysis with Raney Ni led to the expected C–S
bond cleavage, as well as promoted the deblocking of
the anomeric position, yielding quantitatively mono-O-
methylated derivatives 12, 13 and 14.
Further elaboration to UDP-sugars was achieved fol-
lowing the usual strategy, as outlined in Scheme 4. First
attempts of phosphorylation at O-1 with BuLi and
The starting material was heated at 90 ꢀC in a large
.
excess of benzyl alcohol in the presence of BF3 Et2O to
14
15
ClPO(OBn)2
or DMAP and ClPO(OPh)2
were
afford, after recrystallisation, the benzyl glycoside
(74%) 4.9 Following the previously mentioned proce-
dure, treatment of 4 with 2.2 equivalents PivCl at 0 ꢀC
yielded the 3,6-di-O-pivaloyl derivative 5 (70%).
unsuccessful in our hands. However, the one-pot treat-
ment with (BnO)2PNEt2, followed by oxidation with
hydrogen peroxide at ꢁ78 ꢀC provided the glycosyl-
phosphotriesters 15, 16 and 17 in fairly good yields.
Inversion of configuration at C-4 was achieved by for-
mation of the 4-O-triflate followed by intramolecular
displacement by either of the adjacent pivaloyl groups,
leading to a mixture of 3,4- and 4,6-di-O-protected
GalNAc derivatives 6 and 7 (overall yield 90%). These
regioisomers were easily separated by flash chromato-
graphy, affording 6 and 7 in 32and 58% yield, respec-
tively. Finally, the third required precursor 8 was
obtained through a 4!3 migration induced by treating 7
with DBU.10 However, in spite of a thermodynamically
favourable switch from axial to equatorial position, the
yield of this transformation never exceeded 50%. Since
the starting material can be easily recovered after chro-
matography, the overall sequence, starting from 3 g
GlcNAc, provided nearly 1 g of each precursor.
Conventional hydrogenolysis of the benzyl phosphates
over Pd/C furnished phosphates 18, 19, and 20 in
quantitative yields. After removal of pivaloyl groups
using Zemplen conditions, the resulting sugar-mono-
Scheme 3. (a) Bz2O2 (4 equiv), Me2S (8 equiv), MeCN, 0 ꢀC to rt, 4 h,
9: 91%, 10: 76%, 11: 68%; (b) Raney Ni, EtOH, rt, overnight, 12, 13,
14: 99%.
Methylation of 7 using standard Kuhn procedure (MeI,
Ag2O, DMF) furnished an inseparable mixture of
O-methylated derivatives because, like their acetyl coun-
terparts, pivaloates undergo migration under these con-
ditions.11 This result prompted us to envisage a two-step
strategy which consists in forming a methylthiomethyl
(MTM) derivative followed by its hydrogenolysis with
Raney Ni, thus delivering the expected methyl ether
(Scheme 3).12
Scheme 4. (a) (i) (BnO)2PNEt2 (1.5 equiv), 1H-tetrazole (1.5 equiv),
THF, rt, 3 h; (ii) H2O2 (3 equiv), ꢁ78 ꢀC, 1 h, 15: 74%, 16: 62 %,17:
60% (two steps); (b) H2, Pd/C (10% wt), EtOH, rt, 1 h, 18, 19, 20:
Scheme 2. (a) BF3/Et2O, BnOH, 90 ꢀC, 5 h, 74%; (b) PivCl (2.2
equiv), pyridine, dichloromethane, 0 ꢀC, 2h, 70%; (c) (i) Tf 2O (1.3
equiv), pyridine, dichloroethane, ꢁ15 ꢀC, 2h; (ii) H 2O, reflux, 2h,
90% (two steps); (iii) separation by flash chromatography, 6: 32%, 7:
58%; (d) (i) DBU (5 equiv), dichloromethane, rt, 2h; (ii) separation by
flash chromatography, 7: 50%, 8: 50%.
1
99%; (c) (i) MeONa (2equiv), MeOH, rt, overnight; (ii) Amberlite
IR-120 (Et3NH+ form), 21: 95%, 22: 95%, 23: 96% (two steps); (d)
UMP-morpholidate (2equiv), 1 H-tetrazole (3 equiv), pyridine, rt, 4
days 1: 81%, 2: 36%, 3: 28%.

P. Busca et al. / Bioorg. Med. Chem. Lett. 13 (2003) 1853–1856
1855
phosphates were converted into their triethylammonium
salts 21, 22 and 23 with a cationic exchange resin.
Finally, coupling with UMP was carried out using
Moffatt and Khorana procedure, with UMP-morpholi-
date as the activated UMP source and 1H-tetrazole as
the catalyst.16 After 4 days of reaction, UDP-sugars 1, 2
and 3 were obtained in a yield range of 28–81%.17 For
the purification of those UDP-hexoses, we developed a
new HPLC method on a Hypercarb column which
allowed a convenient separation from the dimeric side-
product UPPU.18
are substrates for the ppGalNAc transferase T1
although another member of the ppGalNAc transferase
family has been shown to use UDP-Gal as a substrate
albeit very poorly.22 Molecular modeling of the UDP-
GalNAc binding site based on structural homologies
with other glycosyltransferases of known three-dimen-
sional structures and tested through site directed muta-
genesis experiments23 indicates that at least two of the
OH groups are involved in the substrate binding and
that the sugar nucleotide fits very tightly into the deep
binding pocket mostly through the GalNAc part
whereas the uridine part remains exposed to the med-
ium at the surface of the protein. These observations
may explain the weak inhibitory activity of the
UDP(OMe)GalNAc derivatives. In the absence of cris-
tallographic data, further studies with the deoxy ana-
logues of UDP-GalNAc will be useful to determine
which of the OH groups may be interacting with amino
acids at the transferase active site. These results will be
reported in due course.
Biological Evaluation
The methylated sugar nucleotides were no donor sub-
strates for the ppGalNAc transferase T1, as revealed by
the mass spectrometric analysis of model acceptor pep-
tides exposed to glycosylation conditions.19 The pep-
tides incubated with UDP-GalNAc as substrate were
completely converted into glycopeptides whereas those
incubated with the UDP(OMe)GalNAc analogues 1–3
remained completely unsubstituted. In addition, the
methylated sugar nucleotides proved to be only weak
competitive inhibitors of the enzyme. At an UDP-Gal-
NAc concentration of 20 mM (about 2times Km) the 3-
OMe derivative 1 showed an inhibition comparable to
UDP-GlcNAc which likewise is not recognised as a
donor substrate (Table 1). The concentrations of UDP-
GlcNAc and UDP(3-OMe)GalNAc necessary for 50%
inhibition (IC50) were very similar at around 300 mM.
The UDP(4-OMe)GalNAc and the UDP(6-OMe)-
GalNAc analogues were weaker inhibitors with IC50’s
around 700 mM. These results lend further support to
the observations that the sugar nucleotide binding site
of the ppGalNAc transferase T1 has a very high affinity
for UDP-GalNAc and a rather low affinity for other
UDP-sugars or UDP and its derivatives. Affinity chro-
matography experiments21 showed that the enzyme does
not bind to UDP-hexanolamine agarose but binds effi-
ciently to mercuri-UDP-GalNAc thiopropyl agarose
where the sugar nucleotide is immobilised through the
Hg on the uracil, thereby making the GalNAc freely
accessible. Furthermore, the same study showed that the
enzyme did not bind to mercuri-UDP-GlcNAc thiopro-
pyl agarose suggesting a strong selectivity of the sugar
nucleotide binding site. The different O-methyl groups
may either prevent the sugar nucleotide from entering
the binding pocket on the protein or free OH groups
may be indispensable for substrate binding via hydro-
gen or water bridges. In particular, the OH group at C-4
and the N-acetyl group may be involved in substrate
recognition since neither UDP-GlcNAc nor UDP-Gal
Conclusion
The initial plans to build bisubstrate analogues by link-
ing acceptor surrogates to one of the free OH group of
the donor UDP-GalNAc appear to be thwarted by the
high substrate specificity of the transferase. Never-
theless, the work reported here provides new and useful
information on structure–activity relationships in an
important enzymatic system for which studies with
substrate analogues are just starting to be undertaken.24
Acknowledgements
The present work was financially supported, in part, by
a grant from the CNRS (PCV program) and in part by
the Ligue Nationale contre le Cancer (F.P.) and the
French network for Recombinant Glycosyltransferases
(GT-rec) (F.P and V.P.). The award of a fellowship to
P.B. by the ‘Ministere de l’Education Nationale, de
l’Enseignement Superieur et de la Recherche’ is also
gratefully acknowledged.
References and Notes
1. (a) Seitz, O. Chembiochem. 2000, 1, 214. (b) Toyokuni, T.;
Singhal, A. K. Chem. Soc. Rev. 1995, 231. (c) Barchi, J. J.
Curr. Pharm. Des. 2000, 6, 485. (d) Dennis, J. W.; Granovsky,
M.; Warren, C. E. Biochim. Biophys. Acta 1999, 1473, 21. (e)
Brockhausen, I. Biochem. Soc. Trans. 1997, 25, 871. (f) Kim,
Y. J.; Varki, A. Glycoconjugate J. 1997, 14, 569. (g) Broc-
khausen, I. Biochim. Biophys. Acta 1999, 1473, 67. (h) Rutish-
auser, U. Curr. Opin. Cell. Biol. 1996, 8, 679. (i) Hakomori, S.-
I. In Glycoproteins and Disease; Montreuil, J., Vliegenthart, J.,
Schachter, H., Eds.; Elsevier: New York, 1996; p 243. (j)
Hakomori, S. Adv. Cancer Res. 1989, 52, 257. (k) Hakomori,
S.; Zhang, Y. Chem. Biol. 1997, 4, 97.
Table 1. Inhibitory activity of 1–3 and of UDP-GlcNAc; relative
inhibition of ppGalNAc transferase T1 activity at various inhibitor
concentrations (%)
Concn (mM)
1
2
3
UDP-GlcNAc
2. (a) Elbein, A. Ann. Rev. Biochem. 1987, 56, 497. (b) Elbein,
2
200
1000
0
30
74
0
0
0
0
A. Methods Enzymol. 1987, 138, 661. (c) Tkacz, J. S.; Lampen,
J. O. Biochem. Biophys. Res. Commun. 1975, 65, 218. (d)
Kornfeld, R.; Kornfeld, S. Ann. Rev. Biochem. 1985, 54, 631.
7
7276
17
39
79

1856
P. Busca et al. / Bioorg. Med. Chem. Lett. 13 (2003) 1853–1856
3. Dwek, R. A. Chem. Rev. 1996, 96, 683.
19. The conditions used were essentially those described earl-
ier20 with modifications. The enzyme source was the bovine
polypeptide: a-N-acetyl-galactosaminyl transferase T1
(ppGalNAc transferase) (EC 2.4.1.41) expressed as a secreted,
6His-tagged protein in insect cells after infection with a
recombinant baculovirus, isolated from the culture medium by
affinity chromatography on Ni-trinitrilotetraacetic acid agar-
ose (Qiagen, Hilden, Germany) to a specific activity of 0.5
mmol GalNAc transferred per min and mg protein (=0.5 U/
mg). Enzyme assays were carried out in a total volume of 25
4. Clausen, H.; Bennett, E. P. Glycobiology 1996, 6, 635.
5. (a) Joziasse, D. H. Glycobiology 1993, 3, 97. (b) Bennett,
E. P.; Hassan, H.; Mandel, U.; Mirgorodskaya, E.; Roep-
storff, P.; Burchell, J.; Taylor-Papadimitriou, J.; Hollings-
worth, M. A.; Merkx, G.; van Kessel, A. G.; Eiberg, H.;
Steffensen, R.; Clausen, H. J. Biol. Chem. 1998, 273, 30472. (c)
Bennett, E. P.; Hassan, H.; Mandel, U.; Hollingsworth, M. A.;
Akisawa, N.; Ikematsu, Y.; Merkx, G.; van Kessel, A. G.;
Olofson, S.; Clausen, H. J. Biol. Chem. 1999, 274, 25362. (d)
Ten Hagen, K. G.; Hagen, F. K.; Balys, M. M.; Beres, T. M.;
Van Wuyckhuyse, B.; Tabak, L. A. J. Biol. Chem. 1998, 273,
27749. (e) Ten Hagen, K. G.; Tetaert, D.; Hagen, F. K.;
Richet, C.; Beres, T. M.; Gagnon, J.; Balys, M. M.; Van
Wuyckhuyse, B.; Bedi, G. S.; Degand, P.; Tabak, L. A. J. Biol.
Chem. 1999, 274, 27867.
6. (a) Waldscheck, B.; Streiff, M.; Notz, W.; Kinzy, W.;
Schmidt, R. R. Angew. Chem. Int. Ed. 2001, 40, 4007. (b)
Khan, S. H.; Matta, K. L. Glycoconjugates 1992, 361. (c)
Palcic, M. M.; Heerze, L. D.; Srivastava, O. P.; Hindsgaul, O.
J. Biol. Chem. 1989, 264, 17174. (d) Hashimoto, H.; Endo, T.;
Kajihara, Y. J. Org. Chem. 1997, 62, 1914. (e) Compain, P.;
Martin, O. R. Bioorg. Med. Chem. 2001, 9, 3077, and refs cited
therein.
mL
containing
17.5
nmol
peptide
substrate
(STPSTPSTPSTPSTP, Sigma-Genosys, St. Quentin Fallavier,
France), 0.5 nmol UDP[3H]GalNAc (Amersham, Orsay,
France) at a specific activity of 20,000 cpm/nmol and 50 ng
enzyme in 50 mM morpholinoethanesulfonic acid buffer (pH
6.5) with 15 mM MnCl2. After 15-min incubation at 37 ꢀC, the
reaction was stopped by the addition of 750 mL of ice-cold
water and the solution applied to a SepPak C18 cartridge
(Waters, St. Quentin en Yvelines, France), washed with 10 mL
of water, the peptide and the labelled glycopeptide eluted with
methanol and the radioactivity of the eluate determined by
scintillation counting. For inhibition studies, the nucleotide
sugar analogues were added at the concentrations indicated
and the tests carried out as above.
7. Endo, T.; Kajihara, Y.; Kodama, H.; Hashimoto, H.
Bioorg. Med. Chem. 1996, 4, 1939.
8. Rochepeau-Jobron, L.; Jacquinet, J.-C. Carbohydr. Res.
1998, 305, 181.
9. Shulman, M. L.; Khorlin, A. Y. Carbohydr. Res. 1973, 27,
141.
10. (a) Becker, B.; Thiem, J. Carbohydr. Res. 1998, 308, 77. (b)
Becker, B.; Thiem, J. Tetrahedron: Asymmetry 1994, 5, 2339.
11. (a) Angyal, S. J.; Melrose, G. J. H. J. Chem. Soc. 1965,
6501. (b) Bonner, W. A. J. Org. Chem. 1959, 24, 1388.
12. Projer, P. M.; Angyal, S. J. Austr. J. Chem. 1978, 31, 1031.
13. Medina, J. C.; Salomon, M.; Kyler, K. S. Tetrahedron
Lett. 1988, 29, 3773.
14. Inage, M.; Chaki, H.; Kusumoto, S.; Shiba, T. Chem.
Lett. 1982, 1281.
15. Sabesan, S.; Neira, S. Carbohydr. Res. 1992, 223, 169.
16. (a) Roseman, S.; Distler, J. J.; Moffatt, J. G.; Khorana,
H. G. J. Am. Chem. Soc. 1961, 83, 659. (b) Wittmann, V.;
Wong, C.-H. J. Org. Chem. 1997, 62, 2144.
In order to determine whether the selectively methylated sugar
nucleotides could be used as donor substrates by the ppGal-
NAc transferase, larger scale transfer reactions were set up.
The reaction mixture contained 70 nmol of peptide, 75 mU of
ppGalNAc transferase, 2.5 U alkaline phosphatase from calf
intestine (New England Biolabs, Beverlay, MA, USA), 100
nmol of nucleotide sugar in 50 mM MES pH 6.5 and 15 mM
MnCl2 in a total volume of 100 mL. After 16 h at 37 ꢀC, a
second aliquot of the enzymes and the nucleotide sugars were
added and the incubation continued for 24 h. The reaction was
terminated by the addition of 750 mL of 50% MeOH, the
precipitated proteins removed by centrifugation and the mix-
ture applied onto a column (0.5ꢂ2.5 cm) of Dowex 1X8 (Clꢁ)
(100–200 mesh) (Bio-Rad, Marne-la-Coquette, France). The
flow-through and washings were combined, evaporated under
reduced pressure, taken up in water and purified as above on a
SepPak C-18 cartridge. After concentration in a Speed-Vac
concentrator the samples were processed for electrospray mass
spectrometry on a Quattro II mass spectrometer (Micromass,
Manchester, UK).
20. Craig, A. G.; Park, M.; Fischer, W. H.; Kang, J.; Com-
pain, P.; Piller, F. Toxicon 2001, 39, 809.
21. Wang, Y.; Abernethy, J. L.; Eckhardt, A. E.; Hill, R. L. J.
Biol. Chem. 1992, 267, 12709.
22. Wandall, H. H.; Hassan, H.; Mirgorodskaya, E.; Kris-
tensen, A. K.; Roepstorff, P.; Bennett, E. P.; Nielsen, P. A.;
Hollingsworth, M. A.; Burchell, J.; Taylor-Papadimitriou, J.;
Clausen, H. J. Biol. Chem. 1997, 272, 23503.
23. Duclos, S., Da Silva, P., Vovelle, F., Piller, F. and Piller,
V. Manuscript in preparation.
24. Lazarevic, D.; Thiem, J. Carbohydr. Res. 2002, 337, 2187.
17. NMR data: (1): 1H (D2O) d 5.59 (dd, J1,2=3.5 Hz, J1,P=7.0
Hz, H-1); 13C (D2O): d 175.0, 171.3, 166.5, 152.7, 142.7, 102.9,
94.8 (JC,P=6.1 Hz, C-1), 89.6, 83.7 (JC,P=9.3 Hz), 76.7, 74.2,
72.2, 70.1, 69.9, 64.4 (JC,P=5.1 Hz), 61.0, 56.3 (OCH3), 49.1
(JC,P=8.3 Hz), 22.8; 31P (D2O): d ꢁ10.64 (d, JP,P=20.5 Hz, Pa),
ꢁ12.24 (d, Pb). (2): d5.56 (dd, J1,2=3.5, J1,P=6.5 Hz, H-1); 31
P
(D2O): d ꢁ10.70 (d, JP,P=20.4 Hz, Pa), ꢁ12.33 (d, Pb). (3); 1H
(D2O) d 5.56 (dd, J1,2=3.5, J1,P=7.5 Hz, H-1); 31P (D2O): d
ꢁ10.63 (d, JP,P=20.8 Hz, Pa), ꢁ12.26 (d, Pb).
18. Column: Hypercarb1 250ꢂ10 mm; elution: 45 mM aq
ammonium formate and gradient MeCN. More details on the
separation procedure will be provided in a separate publication.