Synthesis of bis-(2,3,4,6-tetra-O-acetyl-α-d-mannopyranosyl)-l-serinyl phosphate, as a prodrug of mannose-1-phosphate

Article abstract of DOI:10.1016/j.tetasy.2007.09.023

An efficient synthesis of di-mannosyl serinyl phosphate using a convergent strategy involving a silver assisted nucleophilic substitution of tetra-O-acetyl-α-d-mannopyranosyl bromide by a conveniently protected O-serinyl phosphate is described. The straightforward synthesis of phosphoserine under Mitsunobu conditions is also reported.

Full text of DOI:10.1016/j.tetasy.2007.09.023

Available online at www.sciencedirect.com
Tetrahedron: Asymmetry 18 (2007) 2121–2124
Synthesis of bis-(2,3,4,6-tetra-O-acetyl-a-D-mannopyranosyl)-
L-serinyl phosphate, as a prodrug of mannose-1-phosphate
Amira Khaled, Christine Gravier-Pelletier* and Yves Le Merrer*
Universite´ Paris Descartes, CNRS UMR 8601, Laboratoire de Chimie et Biochimie Pharmacologiques et Toxicologiques,
`
45 rue des Saints-Peres, 75270 Paris cedex 06, France
Received 22 June 2007; accepted 19 September 2007
Abstract—An efficient synthesis of di-mannosyl serinyl phosphate using a convergent strategy involving a silver assisted nucleophilic sub-
stitution of tetra-O-acetyl-a-D-mannopyranosyl bromide by a conveniently protected O-serinyl phosphate is described. The straightfor-
ward synthesis of phosphoserine under Mitsunobu conditions is also reported.
Ó 2007 Published by Elsevier Ltd.
1. Introduction
OAc
OAc
OAc
O
OAc
O
AcO
AcO
The congenital disorders of glycosylation1 (CDG) belong
to a group of inherited multisystem metabolic disorders
characterised by the abnormal glycosylation of a number
of serum glycoproteins.2 Amongst the CDG, CDG-Ia is
most frequently encountered and is related to a deficiency
in the production of guanosine-diphosphate-mannose
(GDP-Man). Its biosynthesis notably requires conversion
of mannose-6-phosphate (M6P) into mannose-1-phosphate
(M1P) by phosphomannose mutase (PMM2). M1P is then
at the origin of GDP-Man under the action of guanosyl-
monophosphate transferase. The enzymatic deficiency in
PMM2 is responsible for CDG-Ia3 and there is actually
no treatment for CDG-Ia affected patients. In this context,
we carried out a programme aiming at generating mem-
brane permeant M1P prodrugs. Targeted prodrugs
(Fig. 1) display either mono, di or tri-mannosyl phosphate
structure and we have recently reported an efficient and
general synthetic route towards these derivatives.4 Preli-
minary biological evaluation of these prodrugs showed
notably that dimannosyl phosphates appear as promising
candidates. Indeed, these molecules have the advantage
that after action by cellular esterases or phosphoesterases,
the probability that they will generate M1P is higher than
with monomannosyl phosphate.5
AcO
AcO
O
O P
O
CO2H
NH2
O
O P OR
m
n
2
Prodrugs of M1P
1
n = 1, m = 2
or n = 2, m = 1
or n = 3, m = 0
Figure 1. Prodrugs of M1P and targeted compound 1.
In order to reduce the toxicity of such prodrugs displaying
a dimannosyl phosphate structure, an emphasis is now
placed upon the introduction of R groups to phosphate,
which after liberation by cellular esterases, would lead to
metabolites exhibiting weak toxicity. For that purpose, a
serinyl moiety, expected to have a reduced toxicity has been
chosen (Fig. 1). Furthermore, such a serinyl moiety should
improve the hydrosolubility and bioavailability of the
resulting prodrugs. Indeed, aminophosphoglycans have al-
ready been used as drug carriers.6 It should be noted that
the synthesis of a-D-monomannosyl-serine phosphate has
already been described7 through the formation of an a-D-
mannosylphosphoramidite intermediate, followed by reac-
tion with a protected serine derivative and subsequent oxi-
dation of the resulting phosphite. However, to the best of
our knowledge this strategy has only been applied to the
synthesis of monomannosyl serinyl phosphate,7 while no
access to the targeted dimannosyl serinyl phosphate has
*
Corresponding authors. Tel.: +33 142862181 (C.G.P.); tel.: +33
142862176; fax: +33 142868387 (Y.L.M.); e-mail addresses: Christine.
Gravier-Pelletier@univ-paris5.fr; Yves.Le-Merrer@univ-paris5.fr
0957-4166/$ - see front matter Ó 2007 Published by Elsevier Ltd.
doi:10.1016/j.tetasy.2007.09.023

2122
A. Khaled et al. / Tetrahedron: Asymmetry 18 (2007) 2121–2124
been described. In order to reach such a compound, a com-
plementary strategy is reported.
new strategy to obtain serinyl phosphate in a single one-
pot reaction is an interesting alternative compared to the
phosphoramidite-oxidation methodology. Furthermore, it
should be noted that on the one hand the hydroxyl activa-
tion of serine could be possible on a N-carbamate serine
derivative in spite of the low nucleophilicity of its hydroxyl
usually claimed for such N-protected serine derivatives,
while on the other hand the b-elimination commonly
observed with O-activated serine derivatives9 was circum-
vented. Having succeeded in the synthesis of the
phosphoserine according to Mitsunobu conditions, we next
turned to the same reaction involving bis-(peracetylmanno-
syl)-phosphate 7. This latter one was readily obtained by
peracetylation of D-mannose followed by bromination
and subsequent silver assisted nucleophilic substitution
with benzyl phosphate as its disilver salt10 followed by
hydrogenolysis of the O-benzyl bond. Unfortunately the
previous Mitsunobu conditions involving 3 and 7 proved
to be unsuccessful. The lack of reactivity of phosphate 7
compared to that of the dibenzyl phosphate can tentatively
be explained by the weak nucleophilicity of mannosyl
phosphate as already reported in other alkylation reactions
where the described yields are generally low.5a
2. Results and discussion
Retrosynthetic analysis of the target compound 1 (Fig. 2)
involves a coupling between bis-(peracetyl mannosyl)-
phosphate 2 as a nucleophile and a conveniently O-acti-
vated L-serine derivative 3. The protective groups of the
serine derivative have been chosen to be cleavable in a single
non-acidic step to preserve the phosphate at the anomeric
position of mannose.
OAc
OAc
O
AcO
AcO
O
OAc
O P OH
2
OAc
O
2
+
AcO
AcO
O
CO2H
NH2
CO2P
O P
O
HO
2
NHP
3
1
To overcome this difficulty, we next explored a new strat-
egy involving the condensation of O-serinyl phosphate 9
as a nucleophile with the peracylated a-D-mannopyranosyl
bromide 5 as an electrophile (Fig. 3).
Figure 2. Retrosynthetic pathway of the targeted compound 1.
Activation of the primary alcohol of serine 3 was intended
to be performed under Mitsunobu conditions. To test this
new reaction, a model study was carried out from benzyl
N-benzyloxycarbonyl serine 3 and the commercially avail-
able dibenzyl phosphate (Scheme 1).
OAc
OAc
O
O
HO P
CO2Bn
NHZ
AcO
AcO
1
O
+
OH
Br
First, the esterification of the N-benzyloxycarbonyl L-ser-
ine by successive treatment with cesium carbonate in meth-
anol followed by concentration and addition of benzyl
bromide in DMF afforded 38 (94% yield). Then, Mitsun-
obu coupling of 3 and dibenzyl phosphate in the presence
of triphenyl phosphine and diethyl azodicarboxylate gave
the expected dibenzylphosphoserine 4 in 50% yield. This
5
9
Figure 3. Retrosynthetic pathway of the targeted compound 1.
The O-serinyl phosphate 9 was prepared in two steps
(Scheme 2) involving first, an N,N0-di-isopropyl di-tert-butyl
N-Z-serine
a
O
P
OBn
O
P
OBn
CO2Bn
b
CO2Bn
NHZ
4
HO
BnO
O
+
BnO
OH
NHZ
3
OAc
OAc
OAc
OAc
O
OAc
O
OAc
O
b
X
d
AcO
AcO
AcO
AcO
AcO
AcO
O
O
CO2Bn
NHZ
2 P OR
O
O
2 P
Br
O
5
6 : R = Bn
7 : R = H
8
e
c
D-Mannose
Scheme 1. Reagents and conditions: (a) (i) Cs2CO3, MeOH, (ii) BnBr, DMF (94%); (b) Ph3P, DEAD, PhCH3, THF (50% for 4); (c) (i) Ac2O, DMAP,
˚
Et3N, CH2Cl2, rt, (ii) HBr, AcOH (86% overall); (d) BnOP(O)O2Ag2, PhCH3, molecular sieves 4 A, rt (70%); (e) H2, Pd/C, THF, MeOH (60%).

A. Khaled et al. / Tetrahedron: Asymmetry 18 (2007) 2121–2124
2123
O
O
CO2Bn
NHZ
CO2Bn
NHZ
CO2Bn
NHZ
a
b
HO
(HO)2P O
(tBuO)2P
O
9
3
10
c
5
OAc
OAc
AcO
AcO
OAc
O
O
AcO
AcO
d
O
P
2
AcO
O
P
CO2Bn
NHZ
CO2H
NH2
O
O
O
O
2
1
8
Scheme 2. Reagents and conditions: (a) (i) (tBuO)2PNiPr2, 1H-tetrazole, CH2Cl2, THF; (ii) H2O2 (60%); (b) TFA, CH2Cl2 (100%); (c) 5 (2.2 equiv),
˚
Ag2CO3 (2.2 equiv), PhCH3, molecular sieves 4 A (45%); (d) H2, Pd/C, THF, MeOH (100%).
phosphoramidite reaction with benzyl N-benzyloxycar-
bonyl-L-serine 3 followed by in situ H2O2 oxidation of the
resulting phosphite to phosphate 10 (60%).11 Next, acido-
lysis by trifluoroacetic acid in dichloromethane afforded
the corresponding phosphate 9 in quantitative yield.12 The
silver assisted nucleophilic substitution of tetra-O-acetyl-
a-D-mannopyranosyl bromide 5 by phosphate 9 was then
carried out. The expected protected di-mannosyl serinyl
phosphate 8 was obtained as an enantiomerically pure
a-anomer after flash chromatographic purification13 (45%
References
1. (a) Grunewald, S.; Matthijs, G.; Jaeken, J. Pediatr. Res. 2002,
52, 618–624; (b) Jaeken, J.; Carchon, H. Curr. Opin. Pediatr.
2004, 16, 434–439.
2. (a) Aebi, M.; Hennet, T. Trends Cell. Biol. 2001, 11, 136–141;
(b) de Lonlay, P.; Seta, N.; Barrot, S.; Chabrol, B.; Drouin,
V.; Gabriel, B. M.; Journel, H.; Kretz, M.; Laurent, J.; Le
Merrer, M.; Leroy, A.; Pedespan, D.; Sarda, P.; Villeneuve,
N.; Schmitz, J.; van Schaftingen, E.; Matthijs, G.; Jaeken, J.;
Korner, C.; Munnich, A.; Saudubray, J. M.; Cormier-Daire,
V. J. Med. Genet. 2001, 38, 14–19.
3. (a) van Schafftingen, E.; Jaeken, J. FEBS Lett. 1995, 337, 318–
320; (b) Jaeken, J.; Artigas, J.; Barone, R.; Fiumara, A.; de
Koening, T. J.; Poll-The, B. T.; de Rijk-van Andel, J. F.;
Hoffman, G. F.; Assmann, B.; Mayatepek, E.; Pineda, M.;
Vilaseca, M. A.; Saudubray, J. M.; Schluˆter, B.; Wevers, R.;
Van Schafftingen, E. J. Inherit. Metab. Dis. 1997, 20, 447–449.
0
0
yield, JH1,H2 = 1.4 Hz, JH1,P = 6.6 Hz, JH1 ;H2 ¼ 1:2 Hz,
0
JH1 ;P ¼ 5:9 Hz). Finally, simultaneous removal of both
benzyl and benzyloxycarbonyl protective groups was per-
formed by hydrogenolysis in the presence of palladium
on charcoal in THF–methanol to afford the targeted
di-mannosyl serinylphosphate 1 in quantitative yield.
´
´
4. Hardre, R.; Khaled, A.; Willemetz, A.; Dupre, T.; Moore, S.;
Gravier-Pelletier, C.; Le Merrer, Y. Bioorg. Med. Chem. Lett.
2007, 17, 152–155.
3. Conclusion
5. For the synthesis and biological evaluation of other M1P
prodrugs displaying mono-mannosyl phosphate structure,
see: (a) Rutschow, S.; Thiem, J.; Kranz, C.; Marquardt, T.
Bioorg. Med. Chem. 2002, 10, 4043–4049; (b) Muus, U.;
Kranz, C.; Marquardt, T.; Meier, C. Eur. J. Org. Chem. 2004,
1228–1235; (c) Eklund, E. A.; Merbouh, N.; Ichikawa, M.;
Nishikawa, A.; Clima, J. M.; Dorman, J. A.; Norberg, T.;
Freeze, H. H. Glycobiology 2005, 15, 1084–1093.
6. de Albuquerque Sylva, A. T.; Chung, M. C.; Castro, L. F.;
Guido, R. V. C.; Ferreira, E. I. Min.-Rev. Med. Chem. 2005,
5, 893–914.
7. For the synthesis of a-D-mannosyl phosphate serine deriva-
tives see: (a) Elsayed, G. A.; Boons, G.-J. Synlett 2003, 1373–
1375; (b) Majumdar, D.; Elsayed, G. A.; Buskas, T.; Boons,
G.-J. J. Org. Chem. 2005, 70, 1691–1697.
8. Ono, N.; Yamada, T.; Saito, T.; Tanaka, K.; Kaji, A. Bull.
Chem. Soc. Jpn. 1978, 51, 2401–2404.
In conclusion, we have described an efficient synthesis of
di-mannosyl serinyl phosphate in a convergent strategy
involving five steps from the commercially available N-benz-
yloxycarbonyl-L-serine and peracetyl a-D-mannopyrano-
syl bromide in 25% overall yield. To the best of the
knowledge, this is the first synthesis of a phosphotriester
involving two sugars linked at their anomeric position
and an aminoacid. Based on this strategy, further work will
focus upon the introduction of various aminoacid or pep-
tide derivatives on the di-mannosyl phosphate to investi-
gate their biological properties related to a potential
CDG-Ia treatment. Furthermore, we also reported the suc-
cessful synthesis of the phosphoserine according to Mitsun-
obu conditions. The versatility and the scope of that direct
O-phosphorylation towards both phosphate reagents and
aminoacids are currently being studied.
9. Polt, R.; Szabo, L.; Treiberg, J.; Li, Y.; Hruby, V. J. J. Am.
Chem. Soc. 1992, 114, 10249–10258.
10. Farquhar, D.; Khan, S.; Srivastva, D. N.; Saunders, P. P.
J. Med. Chem. 1994, 37, 3902–3909.
20
1
11. Selected physical data for 10: ½aꢀD ¼ ꢁ6 (c 1.0, CH2Cl2); H
NMR (CDCl3, 250 MHz): d 7.33–7.26 (m, 10H, 2Ph), 6.03
(d, 1H, NH), 5.19 (s, 2H, CO2CH2Ph), 5.12 (s, 2H,
NHCO2CH2Ph), 4.59–4.56 (m, 1H, Ha), 4.45–4.30 (m, 1H,
CHa–OP), 4.27–4.19 (m, 1H, CHb–OP), 1.48 (18H, C(CH3)3);
13C NMR (CDCl3, 63 MHz): d 169.1 (CO2), 155.9 (CONH),
Acknowledgement
We gratefully acknowledge Orphan Europe for their finan-
cial support and for providing post-doctoral research fel-
lowship to A.K.

2124
A. Khaled et al. / Tetrahedron: Asymmetry 18 (2007) 2121–2124
136.2, 135.0 (2 Car), 128.5, 128.4, 128.1, 128.0 (CHar.), 83.1,
5.37–5.29 (m, 6H, H2, H3, H4), 5.23, 5.13 (2d, 4H,
J = 10.7 Hz, J = 7.8 Hz, CO2CH2Ph, NHCO2CH2Ph), 4.66–
4.58 (m, 2H, Ha,CHa–OP), 4.50–4.47 (m 1H, CHb–OP),
4.36–4.10 (6H, H-5, 2 H-6), 2.15–1.98 (24H, COCH3); 13C
NMR (CDCl3, 63 MHz): d 170.4, 169.4, 168.4, 168.0 (CO),
155.8 (CONH), 134.9, 136.1 (Car.), 128.6, 128.4, 128.3, 128.0,
127.8 (CHar.), 95.6, 96.0 (2d, JC1,P = 3.6 Hz, C1A, C1B), 70.8,
70.7 (C5A, C5B), 68.5, 68.6 (C2A, C2B), 68.4 (CO2CH2Ph),
68.0, 67.9 (C3A, C3B), 67.8 (NHCOCH2Ph), 67.0 (br s, CH2–
OP), 65.0, 64.9 (C4A, C4B), 61.5, 61.7 (C6A, C6B), 54.3 (d,
JC,P = 8.9 Hz, Ca), 20.5 (CH3); 31P NMR (CDCl3, 250 MHz):
d 3.67 ppm; MS (ESI+) m/z 1092 for C46H56NNaO26P
[M+Na]+; 762 for C32H38NNaO17P [Mꢁmannosyl+Na]+;
HRMS (ESI+) calcd for C46H56NO26NaP [M+Na]+:
1092.2726. Found: 1092.2712.
83.0 (2d, JtBu,P = 3.7 Hz, 2 C(CH3)3), 67.4 (CO2CH2Ph), 66.9
(NHCOCH2Ph), 66.4 (d, JC,P = 4.6 Hz, CH2OP), 54.6 (d,
JC,P = 6.9 Hz, Ca), 29.7 (d, JC,P = 3.6 Hz, C(CH3)3); 31P
NMR (CDCl3, 250 MHz): d ꢁ9.19 ppm; MS (ESI+) m/z 522
for C26H37NO8P [M+H]+; 410 for C18H19NO8P
[Mꢁ2tBu+H]+; HRMS (ESI+) m/z calcd for C26H36NNaO8P
[M+Na]+: 544.2076. Found: 544.2050.
12. For the preparation of the N-Boc analogue of 10, see:
Clemens, J. J.; Davis, M. D.; Lynch, K. R.; Macdonald, T. L.
Bioorg. Med. Chem. Lett. 2003, 13, 3401–3404.
20
1
13. Selected physical data for 8: ½aꢀD ¼ þ45 (c 1.0, CH2Cl2); H
NMR (CDCl3, 250 MHz): d 7.30–7.35 (m, 10H, 2Ph), 6.11
(d, 1H, NH), 5.71 (dd, 1H, J1A,2A = 1.4 Hz, J1A,P = 6.6 Hz,
H
1A), 5.66 (dd, 1H, J1B,2B = 1.2 Hz, J1b,P = 5.9 Hz, H1B),