β-Stereoselective phosphorylations applied to the synthesis of ADP- and polyprenyl-β-mannopyranosides

Article abstract of DOI:10.1021/ol5026876

An efficient and convenient synthetic route to glycosyl 1-β-phosphates has been developed using diallyl chlorophosphate as a phosphorylating agent with 4-N,N-dimethylaminopyridine under mild conditions. Diallyl-glycosyl 1-β-phosphate triesters of d-manno, l-glycero-d-manno-hepto-, d-gluco-, d-galacto-, and l-fuco-pyranose as well as lactose have been obtained by this strategy in good yields and excellent β-selectivities. Furthermore, the diallyl 6-azido-mannosyl 1-β-phosphate 2 was deprotected under mild conditions and converted into potentially clickable analogues of β-mannosyl phosphoisoprenoids I and ADP-heptose II.

Full text of DOI:10.1021/ol5026876

Letter
pubs.acs.org/OrgLett
β‑Stereoselective Phosphorylations Applied to the Synthesis of ADP-
and Polyprenyl-β-Mannopyranosides
Tianlei Li,† Abdellatif Tikad,† Weidong Pan,‡ and Stephane P. Vincent*,†
́
†Dep
́
artement de Chimie, Laboratoire de Chimie Bio-Organique, University of Namur, rue de Bruxelles 61, B-5000 Namur, Belgium
‡The Key Laboratory of Chemistry for Natural Products of Guizhou Province and Chinese Academy of Sciences 202, Sha-chong
South Road, Guiyang 550002, P. R. China
S
* Supporting Information
ABSTRACT: An efficient and convenient synthetic route to glycosyl 1-
β-phosphates has been developed using diallyl chlorophosphate as a
phosphorylating agent with 4-N,N-dimethylaminopyridine under mild
conditions. Diallyl-glycosyl 1-β-phosphate triesters of D-manno, L-
glycero-D-manno-hepto-, D-gluco-, D-galacto-, and L-fuco-pyranose as well
as lactose have been obtained by this strategy in good yields and
excellent β-selectivities. Furthermore, the diallyl 6-azido-mannosyl 1-β-phosphate 2 was deprotected under mild conditions and
converted into potentially clickable analogues of β-mannosyl phosphoisoprenoids I and ADP-heptose II.
lycosyl-1-phosphates play a central role in glycobiology
because they are involved in carbohydrate metabolism and
scientific value. As detailed below, this study reports a
methodology that allows the β-stereoselective preparation of 1-
phospho mannosides bearing an azido or an alkynyl group as
potentially clickable functionalities. To demonstrate the
applicability of this approach, two clickable analogues of I and
II have been prepared.
The conversion of glycopyranosides into their corresponding
glycosyl 1-phosphates has been accomplished by two main
strategies. The first uses glycosylation reactions that employ
nucleophilic phosphates and electrophilic glycosides such as
glycosyl bromide,9 thioglycosides,10 trichloroacetimidates,11 and
activated glycals.5 The second involves the treatment of
anomeric lactols with electrophilic phosphates P(V) or
phosphites/phosphoramidites P(III).12 In general, most anome-
ric phosphorylations yield α-glycopyranosides as the major
product, especially mannopyranosides.12a
G
the biosynthesis of oligosaccharides.1,2 Glycosyl-1-phosphates
are key intermediates in the enzymatic3 and the chemical4,5
preparations of nucleotide sugars and some important
glycolipids. However, they are generally quite expensive and
rarely commercially available.
Therefore, the stereoselective phosphorylation of carbohy-
drates at the anomeric position has naturally been a topic of
intense research aimed at the preparation of enzyme substrates or
at the synthesis of analogues of natural metabolites.2 Yet, the
synthesis of β-mannosyl phosphates remains a difficult challenge
of considerable importance. For instance, β-mannosyl phosphoi-
soprenoid I is involved as a substrate in the biosynthesis of the
mannan core of the lipoarabinomannan found in the cell wall of
major pathogens such as Mycobacterium tuberculosis (Figure 1).6
Among all the methods that use electrophilic phosphorylating
reagents, Sabesan et al. developed the first efficient β-selective
phosphorylation of mannopyranosyl lactols using diphenyl
chlorophosphate and DMAP as a base.12a More recently, Crich
et al. developed a β-selective phosphorylation of electrophilic
mannosides.10b The high stereoselectivity of the latter approach
requires the presence of a 4,6-O-benzylidene protecting group in
the sugar unit, as is the case for the general β-mannosylation
reaction developed by the same team.13 Unfortunately these two
methods cannot be readily applied for the synthesis of azido/
alkynyl containing mannosides because the deprotection of
diphenyl or benzylidene groups requires hydrogenolytic or harsh
conditions that are frequently not compatible with the presence
of an anomeric phosphate.
Figure 1. Structures of β-manno-configured phosphates I and II.
Another important metabolite that displays a mannose-β-1-
phosphate structure is ADP-L-glycero-β-D-manno-heptopyranose
II, the donor substrate of heptosyl transferases. The latter
catalyze the incorporation of heptose units into the inner core of
lipopolysaccharide (LPS), a key component of the outer
membrane of Gram-negative bacteria (Figure 1).7,5
Furthermore, the synthesis of clickable analogues of
metabolites has become a powerful tool in glycosciences for
applications that include the development of enzyme inhibitors
as well as in vivo cell imaging.8 Therefore, any synthetic
methodology that allows both the preparation of natural
molecules and their clickable analogues is of considerable
To generate potentially clickable glycosides, a protective group
strategy that avoids hydrogenation steps should be followed.
Received: September 11, 2014
Published: October 14, 2014
© 2014 American Chemical Society
5628
dx.doi.org/10.1021/ol5026876 | Org. Lett. 2014, 16, 5628−5631

Organic Letters
Letter
Inspection of the recent literature shows that diallyl phosphates
have been prepared to resolve this problem. For instance,
Macmillan et al. have successfully prepared GDP-2-, 3-, 4-, and 6-
azidomannoses from GTP and the corresponding azidoman-
nose-1-α-phosphate, which were generated after deallylation.14
Recently, Lowary et al. have reported the synthesis of UDP-2-
azido-2-deoxy-α-D-galactofuranose (UDP-GalfN3) using galac-
tofuranosyl diallyl 1-α-phosphate triesters as a precursor.15
However, these phosphorylation techniques give only α-
phosphates.
Compared to the result obtained by diallyl chlorophosphate 7,
the use of diphenyl chlorophosphate yielded product 4 in 75%,
but the selectivity was significantly lower (α/β 1:4, entry 7). With
chlorophosphates as electrophilic reagents, the β-selectivity of
the reaction can be explained by the higher nucleophilicity of β-
lactols compared to their α isomers. Therefore, a slow addition of
chlorophosphate into a solution of mannoside 1 could indeed be
expected to favor formation of the β-phosphate 2. However, we
cannot rule out the involvement of stereoelectronic effects in this
selective phosphorylation.
Therefore, to develop a novel β-selective phosphorylation
methodology, we first decided to screen allyl-protected
phosphorylating reagents on the protected 6-azido-mannose 1
(see Table 1) easily prepared from D-mannose.14 Sabesan’s
After a careful screening of experimental conditions, it was
found that the addition rate (1.6 mL/h) and an increase in the
amount of diallyl chlorophosphate 7 (7.5 equiv) and DMAP
(10.0 equiv) improved the isolated yield of 2 to 61%, while
maintaining an excellent 1:20 α/β selectivity (entry 8). The
moderate yield can be explained by the fact that the
corresponding 1-chloro-mannoside is also generated during the
reaction.17 Consequently, these optimal conditions (Table 1,
entry 8) were selected for further investigation.
Table 1. Optimization of β-Phosphorylation Conditions of 1
From this initial screening, diallyl chlorophosphate 7 gave the
most promising results for the challenging β-stereoselectivity,
and this reagent was selected to study the scope of the reaction.
In contrast to diphenyl chlorophosphate, diallyl chlorophosphate
7 is not commercially available. We thus developed a convenient
and scalable synthetic method for obtaining it; it is depicted in
Scheme 1.
e
entry phosphorylating agent t (h) ratio (α/β) product (yield (%))
ab
f
,
1
2
3
4
5
6
7
8
((iPr)2N)2PCl 5
(AllylO)2PN(iPr)2 6
(BnO)2PN(iPr)2 8
(AllylO)2POCl 7
(BnO)2POCl 9
12
6
4:1
2:1
2:1
1:20
1:5
2 (18)
2 (53)
3 (46)
2 (30)
3 (18)
ac
,
ac
,
10
4
a
a
a
a
4
g
PhO2POCl 10
5
NI
Scheme 1. Synthesis of Diallyl Chlorophosphate 7
(PhO)2POCl 11
(AllylO)2POCl 7
3
1:4
4 (75)
ad
,
4
1:20
2 (61)
a
To a solution of 1 and DMAP (5 equiv) in DCM, 3.0 equiv of
phosphorylating agent was slowly added via a syringe pump.18,12a
b
Allyl alcohol (2.5 equiv) and 1H-tetrazole (2.5 equiv) were added
after 5 h, then t-BuOOH.16 1H-Tetrazole (2.5 equiv) was added after
c
5 h, then t-BuOOH.14 Here, 7.5 equiv of phosphorylating agent and
d
We first tried to obtain 7 in one step, by the direct chlorination
of the commercially available diallyl phosphite 12. Despite
several attempts with sulfuryl chloride,19 trichloroisocyanuric
acid,20 and carbon tetrachloride,21 the purity of the final diallyl
chlorophosphate 7 never exceeded 50% because of the formation
of side products. Therefore, we investigated a stepwise procedure
that involved the oxidation of phosphite 12 followed by
chlorination. The oxidation was realized in the presence of
iodine in pyridine/water (20:1), followed by a reprotonation by a
Dowex 50WX8 (H+) resin that provided diallyl phosphate 13 in
85% yield. The optimized chlorination conditions for 13 entailed
using oxalyl chloride in CH2Cl2 at rt. We found that 5 mol %
equiv of DMF as a catalyst was necessary to minimize the
formation of tetraallyl pyrophosphate (<5%), the usual by-
product of this reaction. The reaction could be easily monitored
by 31P NMR: in CDCl3, the chemical shifts of diallyl
chlorophosphate and its corresponding pyrophosphate were
5.50 and −12.50 ppm, respectively.
Noteworthy, the addition order of reagents and the amount of
DMF happened to be very important parameters for the quality
of the final chlorophosphate (see Supporting Information (SI)
for a detailed procedure). It was also found that diallyl
chlorophosphate 7 was more stable than its dibenzylated
analogue 9, which is advantageous for synthetic procedures
that require long reaction times or long addition times. Indeed,
after 2 days in CDCl3 at rt under argon, the 31P NMR spectrum of
7 showed less than 10% decomposition.
e
f
10.0 equiv of DMAP were used. Isolated yield. In this reaction, the β
anomer could not be isolated in pure form. The products could not
g
be isolated because of a very low conversion of lactol 1 and the
transformation of the chlorophosphate into many side products.
method was selected as a starting strategy, and we thus selected
electrophilic P(III) and P(V) reagents 5−11 for this purpose.
The efficiency and stereoselectivity of the phosphorylations of 1
were evaluated and compared with dibenzyl and diphenyl
chlorophosphates 9 and 11.
The reaction between 1 and chloro-bis(N,N-diisopropyl)-
phosphoramidite 5 in the presence of DMAP, followed by the
addition of allyl alcohol and 1H-tetrazole afforded the
corresponding phosphite with the undesired α-selectivity
(Table 1, entry 1).16 Using (AllylO)2PN(iPr)2 6 instead of
((iPr)2N)2PCl improved the α/β ratio (2:1) and provided the
expected phosphate 2 in 53% yield (for the anomeric mixture),
after oxidation (entry 2). A similar ratio (α/β 2:1) and a slightly
lower yield were obtained with benzyl-protected phosphor-
amidite 8 (entry 3). Interestingly, the treatment of 1 with freshly
prepared diallyl chlorophosphate 7 in the presence of 5 equiv of
DMAP in dichloromethane at room temperature (rt) gave 2 in a
moderate 30% yield but now the selectivity (α/β 1:20) was in
favor of the desired β-anomer (entry 4). Under the same
conditions, the phosphorylation of 1 with dibenzyl chlorophos-
phate 9 (entry 5) and O-phenylene chlorophosphate PhO2POCl
10 (entry 6) improved neither the selectivity nor the isolated
yield.
Having established the optimal conditions for β-phosphoryl-
ation, the scope of the reaction was examined with various
mannopyranosides (compounds 14−18, Table 2). All starting
5629
dx.doi.org/10.1021/ol5026876 | Org. Lett. 2014, 16, 5628−5631

Organic Letters
Letter
Table 2. Phosphorylation of Mannopyranosides
Table 3. β-Phosphorylation of D-Gluco-, D-Galacto-, and L-
Fuco-Pyranosides and Lactose
a
The α/β ratios were determined by 31P NMR of the crude reaction
mixture.
lactols were prepared from the corresponding peracetylated or
perbenzoylated hexopyranoses by selective anomeric depro-
tection using ammonium acetate or methylamine (see SI).
The results presented in Table 2 show that, in all cases, the
phosphorylation is always highly 1,2-cis-stereoselective, whatever
axial group is present at the 2-position. Switching the azido group
from the C-6 (in 1, entry 1) to the C-2 position (in 14, entry 2)
affected neither the overall performance, nor the stereo-
selectivity, of the phosphorylation. The results reported in
entries 3 and 4 indicate that protecting groups may influence the
reaction. Indeed, a good yield (80%) and an excellent selectivity
(α/β 1:30) were observed when benzoates were used as
protecting groups instead of acetates. Furthermore, reaction of
2,3,4,6-tetra-O-benzyl-D-mannopyranose did not produce any
phosphorylation product after 3 h at rt (data not shown).
However, we were pleased to find that D-manno-heptopyr-
anose-β-1-phosphate 17, suitable for the preparation of ADP-
heptose I, was isolated in 55% yield with an excellent selectivity
(Table 2, entry 5). Similar yield and selectivity was obtained in
the preparation of the acetylated mannopyranoside 18 bearing a
triple bond at C-6, indicating the compatibility of this clickable
functional group under these phosphorylation conditions (Table
2, entry 6). For all products 2 and 14−18, the corresponding α-
phosphates could not be isolated. However, from the 31P NMR
spectra of the final crude reaction mixtures, we cannot rule out
that trace amounts of the α-anomers were also formed.
Therefore, in all the cases, we estimated that the α/β selectivity
was at least 1:20.
glucosides and galactosides gave better yields than the
corresponding tetra-acetates (Table 3, entries 1−8).
Gratifyingly, the benzoylated lactoside also gave selectively the
desired β-phosphate 28 in 65% (Table 3, entry 10). The yield of
β-L-fucosyl-1-phosphate 27, a key building block for the synthesis
of GDP-β-fucose, the substrate of fucosyltransferases, was only
36% yield because of some difficulties encountered during the
purification, probably due to the instability of the anomeric
phosphate during prolonged silica gel chromatography (Table 3,
entry 9). Performing this reaction under the conditions reported
by Sabesan et al.,12a using diphenyl chlorophosphate 11,
provided exclusively the α-phosphate. At this stage, we have no
explanation for this surprising observation. Moreover, we have
observed that some 1-β-diallylphosphates (in the gluco- and
galacto-series) can slowly be anomerized, at rt, to the more stable
α-phosphates. Thus, it is recommended to store them at −25 °C
or to deprotect them immediately after their preparation. Indeed,
the corresponding phosphate triethylammonium salts were
found much more stable.
The structures of all products were fully ascertained by NMR
spectroscopy using 1H, 13C, 31P, 1H−1H COSY, 1H−1H NOESY,
and 1H−13C (HSQC, HMBC), including the mannopyranosides
2 and 14−18. In fact, the stereochemistry at the anomeric
position was unambiguously confirmed by NOE experiments
between H-1, H-3, and H-5 for all manno-configured glycosides,
and for the 1,2-trans-pyranosides, by the 3J1−2 coupling constants
showing a 1,2-trans-diaxial relationship.22
Encouraged by the results obtained with the different
mannopyranosides, we explored the synthetic potential of this
selective β-phosphorylation with other glycosides. The results
are gathered in Table 3. Interestingly, all the reactions (entries
1−10) produced exclusively the corresponding β-phosphates
19−28, in moderate-to-good yields in a few hours. Once again, as
for the mannopyranosides, phosphorylations of benzoylated
5630
dx.doi.org/10.1021/ol5026876 | Org. Lett. 2014, 16, 5628−5631

Organic Letters
Letter
Because of the known reactivity of dialkyl chlorophosphates
with carbohydrates’ primary alcohols,23 we did not attempt this
reaction with unprotected sugars.
To demonstrate that this procedure can be applied to the
synthesis of biologically relevant β-phosphates, we prepared
deprotected glycosides 30 and 31 (Scheme 2), two potentially
ACKNOWLEDGMENTS
The authors are grateful to China Scholarship Council (Ph.D.
grant No. 2010667003 to T.L.) and to FNRS (Charge
Recherche for A.T.).
■
́
de
REFERENCES
■
(1) Wagner, G. K.; Pesnot, T.; Field, R. A. Nat. Prod. Rep. 2009, 26,
1172−1194.
Scheme 2. Synthesis of Azido Analogues 30 and 31
(2) Nikolaev, A. V.; Botvinko, I. V.; Ross, A. J. Carbohydr. Res. 2007,
342, 297−344.
(3) Wang, W.; Hu, T.; Frantom, P. A.; Zheng, T.; Gerwe, B.; Del, A. D.
S.; Garret, S.; Seidel, R. D., 3rd; Wu, P. Proc. Natl. Acad. Sci. U.S.A. 2009,
106, 16096−16101.
(4) Plante, O. J.; Palmacci, E. R.; Andrade, R. B.; Seeberger, P. H. J. Am.
Chem. Soc. 2001, 123, 9545−9554.
́ ́
(5) Dohi, H.; Perion, R.; Durka, M.; Bosco, M.; Roue, Y.; Moreau, F.;
Grizot, S.; Ducruix, A.; Escaich, S.; Vincent, S. P. Chem.Eur. J. 2008,
14, 9530−9539.
(6) Guy, M. R.; Illarionov, P. A.; Gurcha, S. S.; Dover, L. G.; Gibson, K.
J. C.; Smith, P. W.; Minnikin, D. E.; Besra, G. S. Biochem. J. 2004, 382,
905−912.
(7) Grizot, S.; Salem, M.; Vongsouthi, V.; Durand, L.; Moreau, F.;
Dohi, H.; Vincent, S.; Escaich, S.; Ducruix, A. J. Mol. Biol. 2006, 363,
383−394.
(8) Sletten, E. M.; Bertozzi, C. R. Acc. Chem. Res. 2011, 44, 666−676.
(9) (a) Arlt, M.; Hindsgaul, O. J. Org. Chem. 1995, 60, 14−15.
(b) Zhang, Q.; Liu, H.-W. J. Am. Chem. Soc. 2000, 122, 9065−9070.
(10) (a) Veeneman, G. H.; Broxterman, H. J. G.; van der Marel, G. A.;
van Boom, J. H. Tetrahedron Lett. 1991, 32, 6175−6178. (b) Crich, D.;
Dudkin, V. Org. Lett. 2000, 2, 3941−3943.
clickable analogues of β-mannosyl phosphoisoprenoids I and
ADP-heptose II (Figure 1). The deallylation of compound 2 was
carried out efficiently in the presence of PdCl2 in DCM/MeOH
at rt, and 29 was isolated in 88% yield (Scheme 2).24 The
coupling between 29 and adenosine 5′-phosphorimidazolide25 in
DMF in the presence of MgCl2 as a Lewis acid catalyst afforded
the desired protected nucleotide azido sugar, which was
deacetylated to lead to 30 in 95% yield. In order to prepare the
analogue of β-mannosyl phosphoisoprenoids I, the reaction
between Nerol and phosphate 29 in the presence of
trichloroacetonitrile in pyridine26 at 65 °C, followed by
deprotection, gave phospholipid 31 in 76% yield over two steps.
In conclusion, we have developed a simple regioselective 1-β-
phosphorylation of acetylated or benzoylated glycopyranosides
that bear an azido or an alkynyl group, using diallyl
chlorophosphate as phosphorylating agent. The scope and
limitations of this reaction have been described and allow various
pyranosides (D-mannose, L-glycero-D-manno-heptopyranose, D-
galactose, D-glycose, L-fucose, and lactose) to afford the
corresponding 1-β-phosphates in moderate-to-good yields and
very high β-selectivity. The interest of this transformation has
been highlighted by the synthesis of a nucleotide sugar and a β-
mannosyl phosphoisoprenoid. Furthermore, the findings
described herein represent a significant advance in selective β-
phosphorylation and will create new opportunities for the design
of other complex glycolipids or nucleotide sugars.
(11) Schmidt, R. R.; Wegmann, B.; Jung, K. H. Liebigs Ann. Chem.
1991, 121−124.
(12) (a) Sabesan, S.; Neira, S. Carbohydr. Res. 1992, 223, 169−185.
(b) van Summeren, R. P.; Moody, D. B.; Feringa, B. L.; Minnaard, A. J. J.
Am. Chem. Soc. 2006, 128, 4546−4547. (c) Kondo, H.; Ichikawa, Y.;
Wong, C. H. J. Am. Chem. Soc. 1992, 114, 8748−8750.
(13) Crich, D.; Sun, S. J. Am. Chem. Soc. 1997, 119, 11217−11223.
(14) Marchesan, S.; Macmillan, D. Chem. Commun. 2008, 4321−4323.
(15) Snitynsky, R. B.; Lowary, T. L. Org. Lett. 2014, 16, 212−215.
(16) Majumdar, D.; Elsayed, G. A.; Buskas, T.; Boons, G. J. J. Org.
Chem. 2005, 70, 1691−1697.
(17) Hung, S.-C.; Wong, C.-H. Tetrahedron Lett. 1996, 37, 4903−
4906.
(18) Danac, R.; Ball, L.; Gurr, S. J.; Fairbanks, A. J. Carbohydr. Res.
2008, 343, 1012−1022.
(19) Maruszewska-Wieczorkowska, E.; Michalski, J.; Zwierzak, A.
Chem. Ind. 1961, 1668.
(20) Acharya, J.; Gupta, A. K.; Shakya, P. D.; Kaushik, M. P.
Tetrahedron Lett. 2005, 46, 5293−5295.
(21) Steinberg, G. M. J. Org. Chem. 1950, 15, 637−647.
(22) The β-cofiguration of mannoside can also be demonstrated by
measuring the 1H−13C 1J coupling constant; see Brand, C.; Ketterholt,
K.; Werz, D. B. Org. Lett. 2012, 14, 5126−5129.
́
(23) Durka, M.; Tikad, A.; Perion, R.; Bosco, M.; Andaloussi, M.;
Floquet, S.; Malacain, E.; Moreau, F.; Oxoby, M.; Gerusz, V.; Vincent, S.
P. Chem.Eur. J. 2011, 17, 11305−11313.
ASSOCIATED CONTENT
■
(24) Gola, G.; Libenson, P.; Gandolfi-Donadio, L.; Gallo-Rodriguez,
C. Arkivoc 2005, 234−242.
S
* Supporting Information
(25) Dabrowski-Tumanski, P.; Kowalska, J.; Jemielity, J. Eur. J. Org.
Chem. 2013, 2147−2154.
Experimental procedures and NMR spectra. This material is
available free of charge via the Internet at http://pubs.acs.org.
(26) Zhang, J.; Angala, S. K.; Pramanik, P. K.; Li, K.; Crick, D. C.; Liav,
A.; Jozwiak, A.; Swiezewska, E.; Jackson, M.; Chatterjee, D. ACS Chem.
Biol. 2011, 6, 819−828.
AUTHOR INFORMATION
■
Corresponding Author
*Fax: +32 81 72 45 17. E-mail: stephane.vincent@unamur.be.
Notes
The authors declare no competing financial interest.
5631
dx.doi.org/10.1021/ol5026876 | Org. Lett. 2014, 16, 5628−5631