Efficient synthesis of a nucleoside-diphospho-exo-glycal displaying time-dependent inactivation of UDP-galactopyranose mutase.

Article abstract of DOI:10.1039/b402469a

A short and efficient synthesis of UDP-exo-galactofuranosyl-glycal is presented. This molecule displayed an interesting time-dependent inactivation of UDP-galactopyranose mutase, an essential enzyme of the mycobacterial cell wall biosynthesis.

Full text of DOI:10.1039/b402469a

Efficient synthesis of a nucleoside-diphospho-exo-glycal displaying
time-dependent inactivation of UDP-galactopyranose mutase†
Audrey Caravano, Stéphane P. Vincent* and Pierre Sinaÿ*
Ecole Normale Supérieure, Département de Chimie, UMR-8642 du CNRS, 24 rue Lhomond, 75231 Paris
Cedex 05, France. E-mail: pierre.sinay@ens.fr; Fax: 33 (0)1 44 32 33 97
Received (in Cambridge, UK) 18th February 2004, Accepted 29th March 2004
First published as an Advance Article on the web 23rd April 2004
A short and efficient synthesis of UDP-exo-galactofuranosyl-
glycal is presented. This molecule displayed an interesting time-
dependent inactivation of UDP-galactopyranose mutase, an
essential enzyme of the mycobacterial cell wall biosynthesis.
ferase inhibition levels have been observed with nucleotide-exo-
glycals,9 these species have never been prepared in the galactofur-
anose series.10 In this context, we designed Galf-exo-glycal-UDP 1
(Scheme 1) which might be a transition-state analogue inhibitor of
the mutase.
Galactans of the mycobacterial cell walls are comprised of over 25
galactofuranose (Galf) residues and the inhibition of their bio-
synthesis prevents proliferation of mycobacteria, including im-
portant pathogens such as Mycobacterium tuberculosis.1 This
biosynthesis involves several enzymes having uridine-diphospho-
galactofuranose (UDP-Galf) as substrate: a mutase catalysing the
interconversion of UDP-galactopyranose (UDP-Galp) and UDP-
Galf, as well as Galf-transferases. On the basis of the respective
mechanisms of these enzymes, we are seeking and designing
molecules that could be transition state analogues of all enzymes
involved in the galactan biosynthesis. The mechanism by which the
flavoenzyme UDP-Galp mutase catalyses the interconversion of
UDP-Galp into UDP-Galf is still unclear. Two main hypotheses
have been addressed: i) the involvement of 1,4-anhydrogalactose as
intermediate;2,3 ii) a single electron transfer from the flavin cofactor
to the sugar ring leading to an anomeric galactosyl radical.4,5
However, none of these assumptions has been demonstrated yet.
Furthermore, Liu et al. have shown that UDP-Galf analogues
fluorinated at the 2- and the 3-positions were substrates of the
reduced mutase, whereas they displayed time-dependent inactiva-
tion against the native enzyme.6 In a previous study, we probed the
mutase binding site with conformationally constrained analogues
wherein the galactose moiety had been locked in a furanose and
a1,4B boat conformation.7
The synthesis of exo-glycal-NDP has to meet two requirements:
i) an efficient methodology for the construction of exo-vinylphos-
phonate; ii) an adequate protection strategy to allow mild
deprotection steps without modifying the sensitive enol-ether
functionality. Galf-exo-glycals 3 and 4 were synthesised in two
steps from the commercially available 1,4-galactonolactone. Tetra-
silylated galactonolactone 2 was condensed with the lithium anion
of dialkyl methyl phosphonate,11 yielding an intermediate lactol
that was directly subjected to a very efficient elimination procedure
recently developed by Lin et al.12 This procedure exclusively gave
exo-glycals with a (Z)-configuration.7,12 The complete one-pot
deprotection of dimethyl phosphonate 4 with trimethylsilyl iodide
mainly led to a decomposition of the sensitive enol-ether function-
ality. Seeking cleaner conditions, we developed a stepwise
procedure from compound 3 based on a preliminary selective
debenzylation followed by a tetradesilylation. The hydrogenolysis
had to be performed first since the same reaction on the tetraol 5 led
to partial saturation of the double bond and partial hydrolysis.
When performed on the tetrasilylated species 3, standard hydro-
genolysis conditions quantitatively yielded the desired intermediate
6. The tert-butyldimethylsilyl (TBS) groups were then removed by
The design of glycal-nucleoside-diphosphates (NDPs) as in-
hibitors is based on the fact that the transition state of enzymatic
reactions involving a cleavage of the anomeric C1-O-(NDP) bond
might be close to an oxycarbenium species with a flattened
conformation. Due to the presence of sp2-hybridized carbons in the
glycal skeletons, these species might closely mimic the conforma-
tion of oxycarbenium species. The main drawback of the endo-
glycal-NDP structures is that they lack a hydroxyl group at the
2-position which might be involved in tight interactions with
glycosyl processing enzymes.8 Although spectacular sialyl trans-
Fig. 1 (left) Residual enzymatic activity at different incubation times of
inactivator 1 (concentration of 1 < 1 mM, : 2.2 mM); (right) kobs
determination at different inactivator concentrations (2 2.2 mM, : 1 mM,
3 0.8 mM, . 0.6 mM, < 0.4 mM); (inset) double reciprocal plot of kobs
versus concentration of 1.
† Electronic supplementary information (ESI) available: experimental
section. See http://www.rsc.org/suppdata/cc/b4/b402469a/
Scheme 1 Reagents and conditions: a) (RO)2POCH2Li, THF, concentrated then THF/Py, (CF3CO)2O; b) H2, Pd/C, Et3N, CH2Cl2; c) TBAF, THF; d) TMSI;
e) UMP activated by the Bogachev procedure (see text). TBS (tert-butyldimethylsilyl), Bn (benzyl).
1216
C h e m . C o m m u n . , 2 0 0 4 , 1 2 1 6 – 1 2 1 7
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 4

Scheme 2 Hypothetical inactivation pathways.
tetrabutylammonium fluoride, giving phosphonate 7 in 76% yield
explain the inactivation of the mutase by exo-glycal 1 (Scheme 2).
Pathway A describes the usual bis-electronic reactivity of enol-
ethers upon acid activation of the double bond. Taking into account
the possibility of a SET, pathways B and C can also be envisioned:
either via an oxycarbenium intermediate 10 (pathway C) or via a
SET preceding a protonation (pathway B). In the two former cases
a coupling with a radical species leading to a covalent adduct 12
might be invoked. We extensively tried to characterise the
formation of such an adduct by mass spectrometry but all attempts
have failed so far. This may be due to the instability of 12, if
formed, in the conditions required for mass spectrometry.17 Three
conserved tyrosine residues, located in the binding site, have been
shown to play an important role in the mechanism.4 Since a tyrosine
has recently been demonstrated to be the catalytic nucleophile of a
trans-sialidase,18 a tyrosine of the mutase could possibly display a
nucleophilic role in this ring contraction. On the other hand, FADH·
might also be a good candidate for a radical coupling.16
In conclusion, this work presents the first synthesis of an NDP-
exo-furanoglycal that displays an interesting time-dependent in-
activation of UDP-galactopyranose mutase. This study opens the
way to the design and the synthesis of a new generation of UDP-
galactopyranose mutase inactivators with improved binding and
kinetic properties that would ultimately facilitate the character-
isation of the still hypothetical catalytic nucleophile involved in this
intriguing enzymatic transformation.
after ion-exchange chromatography. As outlined in Scheme 1, the
choice of TBS as a protective group happened to be critical: the silyl
ethers were not only compatible with the chemistry of the exo-
glycal construction but also prevented side-reactions occurring
during the key hydrogenolysis step. Phosphonate 7 was then
coupled to UMP following a procedure recently developed by
Bogachev.13 Thus, NDP-exo-glycal 1 was efficiently obtained in 5
steps and 40% overall yield from commercial 1,4-galactono-
lactone.14
The target molecule was assayed against UDP-Galp mutase
following a described procedure.6,7 Figure 1 shows the time
dependency of the native enzyme activity when incubated with 1 at
different times. As already observed with 3F-Galf-UDP,6 the
inactivation was only observed with the native enzyme, and not
when a strong reducing agent such as dithionite was used. As in the
case of Liu’s study, sodium dithionite could restore enzymatic
activity of the inactivated mutase. After an incubation time of 30
minutes at a high inactivator concentration (2.2 mM), the mutase
was almost totally inactivated ( > 95%). The inactivated enzyme
did not recover its activity after extensive dialysis against the assay
buffer (8 hours, with 3 buffer changes). These results suggest that
a covalent intermediate between the inactivator and the enzyme is
involved. Competition experiments such as the ones depicted in
Fig. 1 (left) showed that the residual mutase activity depended on
the ratio UDP-Galf/1, suggesting that the inactivation is active-site-
directed.
We are grateful to Dr Didier Blanot (Orsay University) for mass
spectrometry analysis.
The kinetic characteristics KI and kinact were determined from the
double reciprocal plot of kobs versus concentration of 1 (Fig. 1,
inset), the kobs values being obtained from the natural logarithm of
residual activity versus time at different inactivator concentrations.
The KI value thus determined was 0.9 mM which is significantly
higher than the Km value of UDP-Galf (0.2 mM). This probably
signifies that the relative position of the Galf and the UDP moieties
of 1 does not provide an optimal fit within the enzyme binding site.
Interestingly the kinact value found (0.23 min21) is very similar to
that of UDP-3F-Galf (0.19 min21), the only inactivator of this
enzyme reported to date, despite the major structural difference
between the two molecules. Inactivation of glycosyl-processing
enzymes by fluorinated molecules is well documented, especially
in the glycosidase series,8 wherein the electron-withdrawing
character of the fluorine atom dramatically slows down the
deglycosylation of the covalent glycosyl-enzyme intermediate. On
the contrary, glycals are generally designed as transition state
analogues. Thus, rather than a time-dependent inactivation, a
competitive inhibition with molecule 1 was expected.
The surprising behaviour of exo-glycal 1 has to be directly
related to the unique but still unresolved isomerization mechanism
of UDP-galactopyranose mutase. Although the reaction catalysed
by this flavoenzyme does not imply any redox change, the role of
the flavin cofactor has been addressed.15 Potentiometric titration
experiments have shown the stabilisation of FADH·, the semi-
quinone form of FAD, in presence of the substrate, suggesting the
possibility of a crypto redox process in which a single electron
transfer (SET) from FADH2 to the nucleotide sugar would
generate an anomeric radical.5 A recent study of the catalytic
properties of the mutase reconstituted with 1- and 5-deaza-FAD
supports this hypothesis.16 Thus, several putative pathways might
Notes and references
1 F. Pan, M. Jackson, Y. Ma and M. R. McNeil, J. Bacteriol., 2001, 183,
3991–3998.
2 J. N. Barlow, M. E. Girvin and J. S. Blanchard, J. Am. Chem. Soc., 1999,
121, 6968–6969.
3 J. Kovensky and P. Sinaÿ, Eur. J. Org. Chem., 2000, 3523–3525.
4 D. A. R. Sanders, A. G. Staines, S. A. McMahon, M. R. McNeil, C.
Whitfield and J. H. Naismith, Nat. Struct. Biol., 2001, 8, 858–863.
5 S. W. B. Fullerton, S. Daff, D. A. R. Sanders, W. J. Ingledew, C.
Whitfield, S. K. Chapman and J. H. Naismith, Biochemistry, 2003, 42,
2104–2109.
6 Q. Zhang and H.-w. Liu, J. Am. Chem. Soc., 2001, 123, 6756–6766.
7 A. Caravano, D. Mengin-Lecreulx, J.-M. Brondello, S. P. Vincent and
P. Sinaÿ, Chem. Eur. J., 2003, 23, 5888–5898.
8 S. G. Withers, Carbohydr. Polym., 2000, 44, 325–337.
9 B. Müller, C. Schaub and R. R. Schmidt, Angew. Chem., Int. Ed., 1998,
37, 2893–2896.
10 F. Stolz, M. Reiner, A. Blume, W. Reutter and R. R. Schmidt, J. Org.
Chem., 2004, 69, 665–679.
11 A. Dondoni, A. Marra and C. Pasti, Tetrahedron: Asymmetry, 2000, 11,
305–317.
12 W.-B. Yang, C.-Y. Wu, C.-C. Chang, S.-H. Wang, C.-F. Teo and C.-H.
Lin, Tetrahedron Lett., 2001, 42, 6907–6910.
13 V. S. Bogachev, Russ. J. Bioorg. Chem., 1996, 22, 599–604.
14 For experimental data see supporting information.
15 S. Bornemann, Nat. Prod. Rep., 2002, 19, 761–772.
16 Z. Huang, Q. Zhang and H.-W. Liu, Bioorg. Chem., 2003, 31,
494–502.
17 D. Purich, Methods Enzymol., 2002, 354, 1–27.
18 A. G. Watts, I. Damager, M. L. Amaya, A. Buschiazzo, P. Alzari, A. C.
Frasch and S. G. Withers, J. Am. Chem. Soc., 2003, 125, 7532–7533.
C h e m . C o m m u n . , 2 0 0 4 , 1 2 1 6 – 1 2 1 7
1217