Synthesis of mannose-6-phosphate analogs: Large-scale preparation of isosteric mannose-6-phosphonate via cyclic sulfate precursor

Article abstract of DOI:10.1016/S0040-4039(02)00734-7

A scale-up synthesis of isosteric phosphonate analog of mannose-6-phosphate (M6P) is presented via regioselective nucleophilic displacement of the protected 4,6-cyclic sulfate precursor by dialkyl lithiomethylphosphonates in the key step. The five-step synthesis requires only simple purification procedures after each step excluding chromatography separations. The usage of proper nucleophiles makes the method also applicable for the preparation of other isosteric M6P analogs, e.g. α,β-unsaturated phosphonate, sulfonate, etc., with the potential to expand the method for analogs of other sugar-phosphates.

Full text of DOI:10.1016/S0040-4039(02)00734-7

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 4017–4020
Synthesis of mannose-6-phosphate analogs: large-scale
preparation of isosteric mannose-6-phosphonate via cyclic
sulfate precursor
Nikolai A. Khanjin and Jean-Louis Montero*
Laboratoire de Chimie Biomole´culaire, UM2, ENSCM, 8 Rue de l’Ecole Normale, 34296 Montpellier Cedex 5, France
Received 14 March 2002; accepted 11 April 2002
Abstract—A scale-up synthesis of isosteric phosphonate analog of mannose-6-phosphate (M6P) is presented via regioselective
nucleophilic displacement of the protected 4,6-cyclic sulfate precursor by dialkyl lithiomethylphosphonates in the key step. The
five-step synthesis requires only simple purification procedures after each step excluding chromatography separations. The usage
of proper nucleophiles makes the method also applicable for the preparation of other isosteric M6P analogs, e.g. a,b-unsaturated
phosphonate, sulfonate, etc., with the potential to expand the method for analogs of other sugar-phosphates. © 2002 Elsevier
Science Ltd. All rights reserved.
Mannose-6-phosphate (M6P, 1),1 its isosteric phospho-
nate 2a and other hydrolytically stable M6P analogs
have been claimed to accelerate wound healing and
prevent or mitigate scar tissue formation and other
fibrotic disorders2 (Fig. 1). The phosphonate 2a had the
binding affinity approximately three times that of M6P
for the M6P receptor.2b From several M6P analogs
tested in our laboratory, phosphonate 2b was also better
recognized by the receptor3 (Fig. 1). In addition, breast
cancer cells can be selectively targeted by bioactive
molecules endowed with M6P-like moieties4 due to the
increased level of the M6P receptor in breast cancer
tumors.5 However, for any further practical application
of M6P analogs, their convenient scale-up synthesis is
required.
of the known cyclic sulfate precursor methyl 2,3-iso-
propylidene-a-
-mannopyranoside-4,6-sulfate (5)6 with
the lithiated dialkyl methylphosphonates in the key
homologation step followed by the facile deprotection of
sugar hydroxyls and dealkylation of phosphonate
(Scheme 1).
D
The existing routes to 2b by triflate displacement at C(6)7
or by the Horner–Emmons–Wadsworth approach3a were
only applicable for a small scale preparation due to the
unpractical purification procedures for a scale-up, the
cost of reagents or the length of the synthetic route
complicated in particular by the protection and depro-
tection of sugar hydroxyls. Other general methods of
synthesis of sugar phosphonates have also been consid-
ered but have not led to a practical route to 2b.8
Here we report a novel and efficient five-step synthesis
of 2b (Scheme 1) starting from the commercial methyl
Thus, the cyclic sulfate 5 was prepared on a 100 g scale
a-
D
-mannopyranoside (3) via nucleophilic displacement
according to the modified original procedure6a starting
Figure 1.
* Corresponding author. Tel./fax: 33 (0)4 67 14 43 43; e-mail: montero@univ-montp2.fr
0040-4039/02/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(02)00734-7

4018
N. A. Khanjin, J.-L. Montero / Tetrahedron Letters 43 (2002) 4017–4020
Scheme 1. Reagents and conditions: (a) (i) DMP, acetone, [TsOH], 3 h, (ii) H2O; (b) (i) SOCl2, Et3N, CH2Cl2, 5°C, (ii) NaIO4,
[RuCl3], MeCN/CH2Cl2/H2O 2:2:3; (c) HNu, BuLi (2.5 M in hexanes), DMPU or HMPA, THF, −80°C; (d) Amberlyst-15 (H+),
MeOH/THF 1:1; (e) (i) Ac2O/Py, (ii) TMSCl/NaI/MeCN, (iii) NaOH/MeOH.
from 3 via methyl 2,3-O-isopropylidene-a-
D
-mannopy-
lithiomethylphosphonates. The best results were
obtained with diisopropyl ester 6a (overall yield of 8a
was 85% from 5 after chromatography purification)
compared to 72% yield (8b) with the diethyl ester 6b
and only 20% yield (8c) with dimethyl ester 6c. The
lower separated yield with 6b and especially with 6c was
due perhaps to the competitive monodealkylation of
the less hindered phosphonates 7b and especially 7c by
the excess of nucleophile with the formation of double-
charged monophosphonate–monosulfate salts.
ranoside (4).9 In contrast to reactive protected man-
nose-6-triflates, solid 5 can be stored at 0°C for several
months without any observable decomposition. In
addition, the scale-up preparation of corresponding
benzyl-protected manno-pyranoside-6-triflate used in
the reported synthesis of 2b required four steps from 3.7
The synthesis will be also limited by the higher cost of
triflic acid derivatives.
Our initial attempts with nucleophilic substitution of 5
by lithiated dialkyl methylphosphonates 6a–c showed
low but observable reactivity with 5 at −80°C in THF.
Longer reaction time or slow warming to ambient
temperature did not improve the low yield of 7 (ca.
10%) in case of dimethyl and diethyl lithiomethylphos-
phonates 6b and 6c even with the excess of reagents
(1.5–3 equiv.). On the contrary, with 1.3 equiv. of
diisopropyl lithiomethylphosphonate (6a), the desired
phosphonate product 8a was isolated in a good yield
(50% before optimization) after sugar deprotection step
along with the unreacted 5. This was in agreement with
the stability of diisopropyl ester of lithiomethylphos-
phonate for several hours at 20°C in contrast to diethyl
and dimethyl lithiomethylphosphonates (6b, 6c) which
undergo fast and quantitative self-condensation at
>−50°C.10 Addition of HMPA11a (1 equiv.) significantly
accelerates the reaction which was usually completed
within 5–10 minutes at −80°C similar to the fast sugar–
triflate displacements with lithiomethylphosphonates
catalyzed by HMPA.8e DMPU11b induced the same
catalytic effect and should be used for the scale-up due
to the reported toxicity of HMPA. The optimized yields
at this step were varied with the three dialkyl
Also, for the higher reliability and reproducibility of the
yield at this step, we generated lithiated methylphos-
phonates 6a–c from corresponding methylphosphonates
with BuLi in the presence of 1,1-diphenylethylene (0.01
equiv.) for the first time as the indicator for BuLi which
gave clear end points compared to other indicators
investigated.12 Thus, 100% formation of anions was
secured by the in situ titration despite the uncertain
exact concentration of BuLi and the possible presence
of the air or the moisture in the system. In a control
experiment we were able to separate the product in high
yield even when THF was not anhydrous! However, a
larger amount of BuLi was required in this case.
The pure intermediate monosulfate salt 7a can be read-
ily separated from HMPA, DMPU, unreacted nucle-
ophile, and other uncharged impurities by partitioning
between water and CH2Cl2 prior to the deprotection
step. The formation of easily separable monosulfate
salts after nucleophilic substitution is another advan-
tage of cyclic sulfates over corresponding triflate deriva-
tives. For the simultaneous and quantitative cleavage of
monosulfate and isopropylidene groups we have used

N. A. Khanjin, J.-L. Montero / Tetrahedron Letters 43 (2002) 4017–4020
4019
Scheme 2. Reagents and conditions: (a) (i) (OEt)2P(O)CH2SPh, BuLi (2.5 M in hexanes), DMPU, THF, −80°C to rt, then
Amberlyst-15, 95%; (b) (i) mcpba, 1 equiv., then C6H6, D, 80%, (iii) Ac2O/Py, 100%, (iv) TMSCl/NaI, then NaOH/MeOH, 95%.
Amberlyst-15 (H+) resin for the first time13 which
allows deprotection of monosulfate in 10–30 min and
isopropylidene group in 3–5 h at ambient temperature
in MeOH/THF. The separated 8 after extraction con-
tained less than 5% impurities according to the NMR
and it was used in the next step without further purifi-
cation. Also, in anhydrous THF solution (one-pot after
the reaction with lithiomethylphosphonates), intermedi-
ate monosulfates can be selectively cleaved with
Amberlyst-15 in the presence of equivalent amount of
water in 10–30 min at rt, keeping the isopropylidene
protection intact.
Cyclic sulfate 5 was also reactive towards nucleophilic
displacement with several other nucleophiles including
anions of diethyl difluoromethylphosphonate, phos-
phoroamidate, thiophosphate, tert-butyl acetate and
isopropyl mesylate, which allows preparation of corre-
sponding isosteric M6P analogs where P-O fragment is
replaced with one or two heteroatoms or a carbon
atom. Also, a,b-unsaturation can be introduced via
sulfoxide elimination. In particular, a,b-unsaturated
analog of 2c was prepared by this method as outlined in
Scheme 2. Since many other sugar-derived cyclic sul-
fates were described in the literature including
pyranoside-4,6-sulfates and furanoside-5,6-sulfates, the
method should be applicable for the preparation of
sugar phosphate analogs in general. Currently we are
working on the optimization of the method for the
synthesis of other M6P analogs as well as on the
biological evaluation of these compounds which will be
discussed in a separate publication.
The original procedure of Kim and Sharpless14 for
selective deprotection of monosulfates with catalytic
conc. H2SO4 worked as well, but longer reaction time
was needed. However, the following one-pot acidic
cleavage of isopropylidene protection with an excess of
water or methanol required heating at 50°C and pro-
ceeded with some side reactions. Thus, additional chro-
matography purification to separate pure
8 was
In conclusion, we described the first use of cyclic sulfate
in the preparation of isosteric sugar-phosphonates via
nucleophilic substitution by lithiated methylphospho-
nates.17 They were quantitatively generated from
methylphosphonates for the first time by the in situ
titration with BuLi in the presence of 1,1-
diphenylethylene as an indicator. The pure intermediate
monosulfate salts after cyclic sulfate displacement were
separated in the water phase. Also, Amberlyst-15 (H+-
form) was suggested for the first time for the fast and
quantitative cleavage of monosulfate salts. Overall, the
described short synthesis of isosteric phosphonate
analog of M6P required simple purification procedures
with extraction or precipitation and should be applica-
ble for the scale-up.
necessary in this case compared to quantitative depro-
tection of both groups with Amberlyst-15.
Finally, the obtained sugar-phosphonates 8 were quan-
titatively converted to their peracetylated derivatives
with an excess of Ac2O in pyridine followed by one-pot
phosphonate dealkylation with TMSBr15a or with an in
situ preparation of TMSI (TMSCl/NaI)15b in MeCN at
ambient temperature. The latter method allowed use of
a smaller amount of TMSCl/NaI for the faster phos-
phonate dealkylation relative to the larger excess of the
more expensive TMSBr. Thus, with TMSCl/NaI even
hindered diisopropyl phosphonate 8a can be cleaved
quantitavely with 3 equiv. of the reagent in a few hours.
For the diethyl phosphonate 8b 2.5 equiv. of the
reagent was sufficient to cleave phosphonate in 60 min
at rt. After the NaOH/MeOH work up, the final dis-
odium phosphonate 2b was readily precipitated from
the reaction mixture with ethanol. The product can be
further purified by the additional precipitation from
MeOH with EtOH or from diluted AcOH/MeOH for
better solubility followed by precipitation with diluted
ethanolic NaOH.
Acknowledgements
We gratefully acknowledge the Laboratoire Mayoly-
Spindler and the Poˆle Europe´en Universitaire de Mont-
pellier et du Languedoc Roussillon for the generous
financial support of this work in the forms of postdoc-
toral fellowship (N.A.K.) and consumables.
Additional protection of sugar hydroxyls in 8a with
acetate before the phosphonate dealkylation might not
be necessary,3a however, at this time the procedure is
not optimized for the scale-up. Also, the TMSCl/NaI/
CH3CN system was reported to readily convert alco-
hols into iodides at rt, which might be a problem with
the unprotected sugar moiety.16
References
1. For the biological role of M6P-based molecules see: (a)
Kornfeld, S.; Mellman, I. Annu. Rev. Cell. Biol. 1989, 5,

4020
N. A. Khanjin, J.-L. Montero / Tetrahedron Letters 43 (2002) 4017–4020
483–525; (b) Kornfeld, S. Annu. Rev. Biochem. 1992, 61,
yield. The original procedure of Klein and Boom (addi-
tion of pyridine to solution of 4 and SOCl2 in AcOEt)6a
has led to partial deprotection of 4 and methyl mannopy-
ranoside-2,3:4,6-disulfate was separated after one-pot oxi-
dation step as the side product. The disulfate had also
reacted with LiCH2PO(OR)2 but the major product came
from the base-promoted elimination of the axial C(2)ꢀO
bond similar to the described earlier cleavage of man-
nose-derived 2,3-sulfate with basic fluoride anion: (c)
Tewson, T. J. J. Org. Chem. 1983, 48, 3507–3510.
10. (a) Aboujaoude, E. E.; Collignon, N.; Savignac, P. J.
Organomet. Chem. 1984, 264, 9–17; (b) Durand, T.; Savi-
gnac, P.; Girard, J.-P.; Escale, R.; Rossi, J.-C. Tetra-
hedron Lett. 1990, 31, 4131–4134.
307–330; (c) Fagira, K. Von Hasilik, A. Annu. Rev.
Biochem. 1986, 55, 167–193.
2. (a) Ferguson, M. W. J.; British Technology Group Ltd,
U.S. Patent 5,520,926, 1996 (WO 93/18777, 1993),
Method of Using Mannose Phosphate for the Treatment
of Fibrotic Disorders; (b) Ferguson, M. W. J.; Victoria
University of Manchester, UK, US Patent 6,140,307,
2000 (WO 97/05883, 1997) Pharmaceutical Composition
Containing Analogs of M6P for Promoting Wound Heal-
ing; (c) Ferguson, M. W. J.; BTG Limited, U.S. Patent
6,093,388, 2000, Mannose-6-phosphate Composition and
its Use in Treating Fibrotic Disorders.
3. (a) Vidil, C.; Morere, A.; Garcia, M.; Barragan, V.;
Hamdaoui, B.; Rochefort, H.; Montero, J.-L. Eur. J.
Org. Chem. 1999, 447–450; (b) Vidal, S.; Morere, A.;
Montero, J.-L. Phosphorus, Sulfur and Silicon, 2002,
submitted.
11. Encyclopedia of Reagents for Organic Synthesis; Paquette,
L. A., Ed.; John Wiley: New York, 1995 (a) v. 4, pp.
2668–2675; (b) v. 3, pp. 2123–2127.
12. Biphenyl-4-methanol, 2,2%-bipyridyl, 2,2-biquinoline,
triphenylmethane, and 1,10-phenanthroline were also
evaluated, but they were either unselective to BuLi and
(RO)2OPCH2Li or no sharp end points were observed. In
the optimized procedure to the 10% THF solution of
diisopropyl methylphosphonate 6a (1.3–1.5 equiv.) and
1,1-diphenylethylene (0.01 equiv.) BuLi (2.5 M in hex-
anes) was added at −80°C until the deep red-orange color
of Ph2CLiCH2Bu preserved for at least 5 min (the deep
coloration quickly disappeared after the addition of a
small amount of 6a). Then the solution of 5 (20% in
THF) was added followed by HMPA or DMPU (1–2
equiv.). After complete consumption of 5 (<60 min) the
desired pure monosulfate 7a was separated in the water
phase after extraction with CH2Cl2 as the only
hydrophilic compound.
13. For the use of Amberlyst-15/MeOH for isopropylidene
group deprotection at rt in 98% yield see: Ashton, P. R.;
Brown, C. L.; Menzer, S.; Nepogodiev, S. A.; Stoddart, J.
F.; Williams, D. J. Chem. Eur. J. 1996, 2, 580–591.
14. Kim, B. M.; Sharpless, K. B. Tetrahedron Lett. 1989, 30,
655–658.
15. (a) McKenna, C. E.; Higa, M. T.; Cheung, N. H.;
McKenna, M.-C. Tetrahedron Lett. 1977, 18, 155–158;
(b) Morita, T.; Okamoto, Y.; Sakurai, H. Tetrahedron
Lett. 1978, 19, 2523–2526.
4. (a) Hamdaoui, B.; Dewynter, G.; Capony, F.; Montero,
J.-L.; Toiron, C.; Hnach, M.; Rochefort, H. Bull. Soc.
Chim. Fr. 1994, 131, 854–864; (b) Barragan, V.; Menger,
F. M.; Caran, K. L.; Vidil, C.; Morere, A.; Montero,
J.-L. Chem. Commun. 2001, 85–86.
5. Zhao, Y.; Escot, C.; Maudelonde, T.; Puech, C.;
Rouanet, P.; Rochefort, H. Cancer Res. 1993, 53, 2901–
2905.
6. (a) Van der Klein, P. A. M.; Van Boom, J. H. Carbohydr.
Res. 1992, 224, 193–200; (b) Van der Klein, P. A. M.;
Boons, G. J. P. H.; Veeneman, G. H.; Van der Marel, G.
A.; Van Boom, J. H. Tetrahedron Lett. 1989, 30, 5477–
5480.
7. Berkowitz, D. B.; Bhuniya, D.; Peris, P. Tetrahedron
Lett. 1999, 40, 1869–1872 and references cited therein.
8. Reviews: (a) Engel, R. Chem. Rev. 1977, 77, 349–367; (b)
Weimer, D. F. Tetrahedron 1997, 53, 16609–16644; (c)
McClard, R. W.; Wittel, J. F. In Handbook of
Organophosphorus Chemistry; Engel, R., Ed.; M. Dekker:
New York, 1992; pp. 655–668; (d) Barton, D. H. R.;
Ge´ro, S. D.; Quiclet-Sire, B.; Samadi, M. Tetrahedron
1992, 48, 1627–1636; (e) Shen, Q.; Sloss, D. G.; Berkow-
itz, D. B. Synth. Commun. 1994, 24, 1519–1530 and
references cited therein.
9. The product 4 is commercially available (Sigma), how-
ever, for its scale-up preparation the method of Evans
and Parish ((a) Evans, M. E.; Parrish, P. W. Carbohydr.
Res. 1977, 54, 105–114) was adopted for a large scale
with the minor modifications which allowed the use of
smaller amounts of reagents and solvents. Then 4 was
converted to 5 via cyclic sulfite according to the general
procedure of Gao and Sharpless ((b) Gao, Y.; Sharpless,
K. B. J. Am. Chem. Soc. 1988, 110, 7538–7539) in 85%
16. Olah, G. A.; Narang, S. C.; Gupta, B. G. B.; Malhotra,
R. J. Org. Chem. 1979, 44, 1247–1251.
17. Only one example of the nucleophilic displacement of a
cyclic sulfate (derived from 1,2-dodecanediol) by methyl
lithiomethylphosphonate could be found in the literature:
Hoye, T. R.; Crawford, K. B. J. Org. Chem. 1994, 59,
520–522.