Studies of alkaline mediated phosphate migration in synthetic phosphoethanolamine L-glycero-D-manno-heptoside derivatives

Article abstract of DOI:10.1016/S0008-6215(98)00271-7

Synthetic 2-, 3-, 4- and 6-monophosphate derivatives of methyl α-D-mannopyranosides, the 4-, 6- and 7-monophosphate derivatives of methyl L-glycero-α-D-manno-heptopyranosides and the corresponding phosphoethanolamine derivatives and a 6,7-cyclic phosphate analogue of methyl L-glycero-α-D-manno-heptopyranoside were used to study phosphate migration and hydrolysis when subjected to strong alkaline conditions (4 M KOH, 120°C, 18 h). The resulting products were analyzed by ¹H NMR spectroscopy and electrospray mass spectrometry. It was found that phosphate substituents were stable under these conditions and neither migration nor hydrolysis was observed except for the heptose 7-phosphate, which gave a substantial amount of phosphate hydrolysis. In phosphoethanolamine-substituted compounds migration to adjacent positions with concomitant loss of ethanolamine was found together with hydrolysis. Copyright (C) 1998 Elsevier Science Ltd.

Full text of DOI:10.1016/S0008-6215(98)00271-7

Carbohydrate Research 313 (1998) 193–202
Studies of alkaline mediated phosphate migration in
synthetic phosphoethanolamine
L
-glycero- -manno-heptoside derivatives
D
Andrea Stewart a, Christian Bernlind b, Adele Martin a, Stefan Oscarson b,
James C. Richards a, Elke K.H. Schweda c,*
a Institute for Biological Sciences, National Research Council of Canada, Ottawa, Ontario K1A 0R6, Canada
b Department of Organic Chemistry, Arrhenius Laboratory, Stockholm Uni6ersity, S-106 91 Stockholm, Sweden
c Uni6ersity College of South Stockholm, Clinical Research Center, NOVUM, S-141 86 Huddinge, Sweden
Received 27 April 1998; accepted 10 October 1998
Abstract
Synthetic 2-, 3-, 4- and 6-monophosphate derivatives of methyl a-
monophosphate derivatives of methyl -glycero-a- -manno-heptopyranosides and the corresponding phospho-
ethanolamine derivatives and a 6,7-cyclic phosphate analogue of methyl -manno-heptopyranoside were
D-mannopyranosides, the 4-, 6- and 7-
L
D
L
-glycero-a-
D
used to study phosphate migration and hydrolysis when subjected to strong alkaline conditions (4 M KOH, 120 °C,
1
18 h). The resulting products were analyzed by H NMR spectroscopy and electrospray mass spectrometry. It was
found that phosphate substituents were stable under these conditions and neither migration nor hydrolysis was
observed except for the heptose 7-phosphate, which gave a substantial amount of phosphate hydrolysis. In
phosphoethanolamine-substituted compounds migration to adjacent positions with concomitant loss of ethanolamine
was found together with hydrolysis. © 1998 Elsevier Science Ltd. All rights reserved.
Keywords: Sugar phosphates; Phosphate migration; Structural analysis; Lipopolysaccharides
1. Introduction
Gram-negative bacteria [3], but their role in
biological and immunological functions is not
clear. However, recent studies have implicated
the presence of phosphate substituents as po-
tential triggers that control the biosynthesis of
important oligosaccharide epitopical struc-
tures [4]. Structural analysis of LPS requires
complete identification of the number and
sites of these substituents in order to deter-
mine their role in biological function. It has
been observed that ethanolamine was lost
from PEtn-substituted heptose with concomi-
tant migration of the phosphate group when
LPS was subjected to strong alkaline treat-
ment [2,5,6]. Moreover, choline was observed
to be lost from PChol when LPS of H. influ-
The inner-core region in lipopolysaccha-
rides (LPS) of Gram-negative bacteria most
often contains
residues, which are commonly substituted by
phosphate (P) and phosphoethanolamine
(PEtn) groups [1]. Recently, phosphocholine
(PChol) was also identified as a substituent in
Haemophilus influenzae LPS [2]. Phosphate
groups are known to be important in main-
taining the integrity of outer membranes in
L
-glycero- -manno-heptose
D
* Corresponding author. Tel.: +46-8-58583823; fax: +46-
8-58583820; e-mail: elke.schweda@kfcmail.hs.sll.se
0008-6215/98/$ - see front matter © 1998 Elsevier Science Ltd. All rights reserved.
PII: S0008-6215(98)00271-7

194
A. Stewart et al. / Carbohydrate Research 313 (1998) 193–202
Scheme 1.
enzae was treated under strong alkaline condi-
tions [2]. Migration of appended substituents
due to experimental conditions is a well-
known problem. The migration of acetate es-
ters has been studied, and it was found that
they migrate under basic conditions to the
thermodynamically more favored positions in
a sugar residue, but are more stable under
acidic conditions [7,8]. Strong alkaline treat-
ment of LPS is a standard procedure in struc-
tural analyses to remove N- and O-linked
fatty acids. It was thus of interest to obtain
more insight into the stability of P and PEtn
groups under those conditions and to find
experimental conditions where no degradation
or migration occurred. For this reason
O-benzyl-1,6-dialdo-a- -mannopyranoside. In
D
contrast to earlier similar syntheses by
Grzeszczyk and Zamojski [10], the reactions
were performed using Barbier reaction condi-
tions, i.e., the chloride and the aldehyde were
added simultaneously to the magnesium turn-
ings [11]. The yields and stereoselectivities
were comparable to the previous ones, but in
our hands the Barbier variant made the elon-
gation less sensitive to reaction conditions,
and, furthermore, less of an excess of the toxic
chloride could be used, which also simplified
purification. Thus, heptose derivatives 5 and
10 were prepared in good yields and with high
stereoselectivity from the known mannose pre-
cursor. To allow the synthesis of 7- and cyclic
6,7-phosphorylated derivatives, compound 10
was benzylated and deallylated or directly
deallylated to give intermediates 12 and 13,
respectively. In order to synthesize the 4-phos-
phorylated derivatives, compound 5 was
debenzylated and treated with dimethoxypro-
pane to give 20. To obtain the 6-, 7- and
4-monophosphates, compounds 5, 12 and 20
were treated with PCl3, imidazole and Et3N
followed by benzyl alcohol to give the diben-
zyl phosphites. These compounds were in situ
oxidized by m-chloroperoxybenzoic acid to
yield the dibenzylphosphates 6, 14 and 21,
respectively, which were catalytically hy-
drogenolyzed, and 21 also further treated with
aqueous acid to complete deisopropylidena-
tion, to give the corresponding deprotected
monophosphates 7, 15 and 22. Treatment of
the diol 13 with commercial methyl
dichlorophosphate followed by basic work-up
afforded 18, which was deprotected to give the
cyclic 6,7-phosphate 19. Initially, an approach
using bis(N,N-diisopropylamino)benzyloxyph-
osphine [12] and 1H-tetrazole was employed
for the synthesis of the ethanolaminephos-
phate compounds. This procedure, however,
gave only moderate yields of the secondary 4-
monophosphorylated
were subjected to strong alkaline treatment.
Monophosphorylated -heptose derivatives
D-mannose derivatives
L D
,
with and without ethanolamine and a 6,7-
cyclic phosphate analog were synthesized and
treated in the same manner. The resulting
products were analyzed by NMR spec-
troscopy and electrospray mass spectrometry
(ESMS).
2. Results and discussion
The synthesis of methyl a-
noside 2-phosphate (1), methyl a-
ranoside 3-phosphate (2), methyl a-
D
-mannopyra-
-mannopy-
-manno-
D
D
pyranoside 4-phosphate (3), and methyl a- -
D
mannopyranoside 6-phosphate (4) has been
described earlier [9].
The syntheses of the phosphate and phos-
phoethanolamine substituted methyl
L
, -hep-
D
tosides and the 6,7-cyclic phosphate derivative
are summarized in Schemes 1–3. The one-car-
bon elongations to obtain the variously pro-
tected heptose precursors were accomplished
by a Grignard reaction using benzyloxy- or
allyloxymethyl chloride and methyl 2,3,4-tri-

A. Stewart et al. / Carbohydrate Research 313 (1998) 193–202
195
Scheme 2.
and 6-phosphates 24 (40%) and 9 (48%) and
proved inefficient for the preparation of the
primary 7-phosphoethanolamine derivative
17. Hence, the H-phosphonate method was
tried instead. Pivaloyl chloride-promoted cou-
affected by the alkaline treatment. The result-
ing products were identical with the starting
materials. The 7-monophosphate 15, however,
showed a substantial amount of the phosphate
hydrolysis product 26 (36%), indicating a
mechanism involving a 6,7-anhydro intermedi-
ate [14] (Scheme 4), since the other primary
monophosphate 4 was inert.
pling
of
N-benzyloxycarbonyl-protected
ethanolamine phosphonate [13] to 5, 12 and
20, respectively, followed by oxidation by
iodine in aqueous pyridine gave intermediates
8 (84%), 16 (75%) and 23 (89%), respectively.
The latter compounds were deprotected to
give the target compounds 9 (94%), 17 (82%)
and 24 (79%), respectively.
The heptose 4-, 6- and 7-phosphoethano-
lamine derivatives 24, 9 and 17 gave mixtures
of components after the alkaline treatment as
revealed by NMR spectroscopy. ESMS, run in
the negative mode, of the products showed
pseudomolecular ions at m/z 303 correspond-
ing to monophosphorylated methyl heptosides
(C8H17O10P), which clearly indicated the loss
of ethanolamine from the starting materials.
The products of 24 were identified to be the
4-phosphate 22 and the 3-phosphate 25 as
characterized by NMR. The products of 9
were identified to be the 6-phosphate 7 and
the 7-phosphate 15. The products of 17 were
identified also to be the 6-phosphate 7 and
7-phosphate 15. In addition dephosphoryla-
The monophosphates 1–4, 7, 15 and 22, the
cyclic phosphate 19 and the ethanolamine-
phosphates 9, 17 and 24 were used as model
substances for studying the stability of these
substituents under alkaline conditions. All
samples were subjected to the strong alkaline
treatment that is used in N/O-deacylation of
LPS, i.e., 4 M KOH, 18 h, 120 °C. The reac-
tion mixtures were neutralized, subjected to
gel-permeation chromatography and the ob-
tained products were analyzed by NMR and
ESMS.
tion of 24, 9 and 17 was evident since methyl
1
L,
D
-heptoside 26 could be identified in the H
NMR spectra and inorganic phosphate in the
31P NMR spectra. The relative proportions of
the resulting products are given in Table 1.
Thus, in all the phosphoethanolamine deriva-
tives, the phosphate groups partly migrated to
the adjacent positions with concomitant loss
of ethanolamine. Migration was found to oc-
cur also when compound 9 was treated with
KOH at lower concentration (i.e., 2 M) and
under even milder reaction conditions (2 h,
60–100 °C).
It was evident that the monophosphory-
lated compounds 1–4, 7, and 22 were not

196
A. Stewart et al. / Carbohydrate Research 313 (1998) 193–202
Scheme 3.
experiment. Thus, alternative ways of N/O-
deacylation should be taken into considera-
tion.
Similar migrations are known to proceed
via a cyclic phosphate intermediate, where a
five-membered ring intermediate is preferred
over a six-membered one [14,15]. Our results
were in accordance with this, as only 4“3
and 6l7, and no 4l6 migrations were ob-
served, together with the similarity in the
product distribution of the hydrolysis of 9 and
17. To further verify this, the 6,7-cyclic phos-
phate 19 was synthesized and subjected to
alkaline treatment. The products obtained
were the 6-phosphate 7, the 7-phosphate 15
and the non-phosphorylated 26 in the molar
ratio 50:33:17 (Table 1), the latter product
probably formed through phosphate hydroly-
sis of 15 as found above. Thus, the 6,7-cyclic
phosphate opens up to give about a 1:1 mix-
ture of the 6 and the 7-phosphate of which the
latter then is partly hydrolyzed to give 26. The
product ratios obtained from the alkaline
treatment of 9 and 17 are very similar (Table
1). Hence, the basic hydrolysis of 9 and 17
seems to proceed almost exclusively through
the cyclic intermediate 19 as depicted in
Scheme 4.
A primary objective of this study was to
find experimental conditions of alkaline treat-
ment of LPS in which no degradation of PEtn
and migration of P occurs. For this reason
LPS from a strain of Neisseria meningitidis [5]
was subjected to different N/O-deacylation
conditions (4 M KOH, 2–18 h, 120 °C or 2 M
KOH, 2 h, 60 °C). However, reducing the
KOH concentration, reaction time or temper-
ature did not change the fact that ethanol-
amine is lost from the heptose bearing the
PEtn substituent as revealed by ESMS. It can
be concluded that the strong alkaline condi-
tions used for N/O-deacylation of LPS do not
degrade the material if it is substituted with P
but in case of substitution with PEtn and/or
PChol, information will be lost during that
3. Experimental
General methods.—These were as described
earlier [16]. Benzyloxymethyl chloride was
synthesized according to Connor et al. [17],
dried over CaCl2 and stored without drying
agent in a sealed container at −18 °C. The
same procedure was also applied in the syn-
thesis of allyloxymethyl chloride.
General procedure for the synthesis of the
protected heptose dibenzyl phosphates.—A so-
lution of phosphorus trichloride (350 mL, 4.01
mmol) in CH2Cl2 (5 mL) was added dropwise
to a solution of imidazole (820 mg, 12.0
mmol) in CH2Cl2 (25 mL) at 0 °C, followed by
Et3N (1.76 mL, 12.6 mmol). The mixture was
stirred for 10 min and the suitably protected
monosaccharide (1.00 mmol) in CH2Cl2 (10
mL) was then added dropwise. After stirring
for 1 h, benzyl alcohol (1.30 mL, 12.6 mmol)
in CH2Cl2 (5 mL) was added and stirring was
continued at 0 °C for 40 min. Recrystallized
m-chloroperoxybenzoic acid (1.00 g, 5.79
mmol) in CH2Cl2 (10 mL) was added, and the
mixture was stirred for 1.5 h at room tempera-
ture, then washed with aq 5% Na2S2O3 (20
mL) and water (20 mL), dried (Na2SO4), and
concentrated. The crude products were
purified on silica gel columns.
General procedure for the synthesis of the
protected heptose N-(benzyloxycarbonyl)eth-
anolamine phosphates.—Suitably protected
heptose derivatives (1.00 mmol) and N-(ben-
zyloxycarbonyl)ethanolamine triethylammo-
nium H-phosphonate [12] (432 mg, 1.20
mmol) were concentrated from dry pyridine
and dried on a vacuum pump for 1 h. The

A. Stewart et al. / Carbohydrate Research 313 (1998) 193–202
197
Scheme 4. Suggested pathways for the alkaline treatment of 9, 17 and 19.
residue was dissolved in MeCN (6 mL) and
pyridine (2 mL) under Ar, and cooled to 0 °C.
Freshly distilled pivaloyl chloride (493 mL,
4.00 mmol) was added and the solution was
stirred for 10 min at 0 °C and then for 50 min
at room temperature, when TLC showed a
complete conversion into a more hydrophilic
product. The solution was diluted with
toluene (50 mL) and washed with water (25
mL). After drying (MgSO4) and concentra-
tion, the residue was dissolved in pyridine (5
mL) and cooled (0 °C). Water (100 mL) and
iodine (279 mg, 1.10 mmol) were added and
the solution was stirred for 1 h at 0 °C, diluted
with EtOAc (60 mL) and washed with aq satd
Na2S2O3 (30 mL) and aq 1 M triethylammo-
nium bicarbonate, pH 8 (50 mL). The com-
bined aqueous phases were back-extracted
with CH2Cl2 (2×25 mL for compounds 8 and
16, and 5×25 mL for compound 23). The
organic phases were concentrated without dry-
ing, and purified by column chromatography
using CHCl3–MeOH containing 1% Et3N.
Methyl 2,3,4,7-tetra-O-benzyl-
L
-glycero-h-
D-manno-heptopyranoside (5).—To a flame-
dried flask equipped with an internal
thermometer, an efficient stirrer and two
dropping funnels were added freshly activated
magnesium turnings (1.11 g, 45.7 mmol) and
sublimed HgBr2 (132 mg, 0.37 mmol) under
Ar. Benzyloxymethyl chloride (4.2 mL, 30.3
mmol) was dissolved in THF (30 mL) and a
portion of this solution (approx. 5 mL) was
added to the magnesium at room temperature
via one of the dropping funnels. Once the
exothermic reaction had started (as monitored
by the temperature) the flask was partially
immersed into an ice bath (0 °C) and crude
Table 1
Products obtained for compounds 9, 17, 19 and 24 after
alkaline treatment with 4 M KOH (120 °C, 18 h)
Compound
Productsa
methyl 2,3,4 - tri - O - benzyl - 1,6 - dialdo - a - -
D
9
17
19
24
7 (57)
7 (54)
7 (50)
22 (44)
15 (36)
15 (26)
15 (33)
25 (49)
26 (7)
26 (20)
26 (17)
26 (7)
mannopyranoside [10] (7.52 g, 15.3 mmol) in
THF (30 mL) and the alkyl halide solution
were added simultaneously through the drop-
ping funnels at such a rate that the internal
temperature was kept between 20 and 24 °C
a The ratios measured from signal areas of appropriate
NMR resonances are given in parentheses.

198
A. Stewart et al. / Carbohydrate Research 313 (1998) 193–202
Methyl 6 - O - (2 - aminoethyl)phosphoryl -
L
-
(approx. rate: 2 mL/min of each solution).
After stirring overnight the mixture was di-
luted with Et2O (200 mL) and freshly pre-
pared aq satd NH4Cl (250 mL) was added.
The mixture was stirred for 2 h, when all
the metal turnings had dissolved. The or-
ganic phase was separated, dried (MgSO4),
filtered, and concentrated. To a solution of
the residue in CHCl3 was added silica gel
(ca. 15 g), and the suspension was carefully
concentrated and dried under reduced pres-
sure for at least 1 h. The silica gel was
poured on top of a packed column, and elu-
tion with 6:1 toluene–EtOAc then gave 5
(6.65 g, 57%); [h]D +13° (c 1.0, CHCl3) lit.
[h]D +10.5° [18]; 13C NMR (CDCl3): l 54.7
(MeO), 68.0, 70.8, 71.5, 72.2, 72.8, 73.3,
74.3, 74.5, 75.2, and 80.3 (C-2–7, PhCH2),
99.4 (C-1), 127.6–138.6 (aromatic).
glycero - h - - manno - heptopyranoside (9). —
D
Following the general procedure, compound
5 (331 mg, 0.566 mmol) gave methyl 2,3,4,7-
tetra - O - benzyl - 6 - O - (2 - benzyloxycarbony-
lamino)ethyl - phosphoryl -
L
- glycero - a -
D
-
manno-heptopyranoside, triethylammonium
salt (8; 446 mg, 84%) after purification on
silica gel (19:1 CHCl3 –MeOH+1% Et3N).
NMR: 13C (CDCl3), l 8.4 (Et3N), 42.2 (d,
CH2CH2N), 45.3 (Et3N), 54.5 (MeO), 64.8
(d), 66.3, 68.6, 69.7 (d), 71.1, 72.0, 73.0,
73.8, 74.4, 75.3, and 80.8 (C-2–7, PhCH2,
OCH2CH2), 98.9 (C-1), 127.2–139.2 (aro-
matic), 156.6 (CON); 31P (9:1 EtOAc–
CHCl3), l 0.8 (m). To a solution of 8 (295
mg, 0.313 mmol) in 4:3 THF–H2O (14 mL)
and HOAc (l mL) was added a catalytic
amount of Pd(OH)2 on carbon. The mixture
was hydrogenolyzed at 9 atm for 2 days,
filtered (Celite), concentrated, and dissolved
in H2O (10 mL). The solution was set to
neutral pH and washed once with Et2O.
Size-exclusion chromatography on a Bio-Gel
P-2 column eluted with aq 0.1 mM pyri-
dinium acetate buffer, pH 5.4 gave 9 (102
mg, 94%) after freeze-drying; [h]D +26° (c
1.0, H2O); NMR: 13C (D2O), l 41.0 (d,
CH2CH2N), 55.5 (MeO), 61.8, 63.3, 66.7,
70.6, 71.1, 71.3 (d), and 74.4 (d) (C-2–7,
OCH2CH2), 101.9 (C-1); 31P (H2O), l 0.5
(m). HRMS Calcd for C10H23O10NP [M+
H]: 348.1060. Found: 348.1104.
Methyl
L
- glycero - h - - manno - heptopyra-
D
noside 6-(disodium phosphate) (7).—Phospho-
rylation of 1.03 mg (1.75 mmol) of 5 accord-
ing to the general procedure gave
methyl 2,3,4,7-tetra-O-benzyl-6-O-dibenzy-
loxyphosphoryl-
L
-glycero-a- -manno-hepto-
D
pyranoside (6; 1.30 mg, 88%) after silica gel
chromatography (two columns: 3:2 light
petroleum ether [bp 40–65 °C]–EtOAc and
4:1 toluene–EtOAc). NMR: 13C (CDCl3), l
54.7 (MeO), 69.1, 69.4 (d), 70.1 (d), 71.9,
72.7, 73.2, 73.8, 74.30 (d), 74.34, 74.8,
and 80.7 (C-2–7, PhCH2), 99.1 (C-1),
127.4–138.7 (aromatic); 31P (EtOAc), l −
0.5 (m). Compound 6 (641 mg, 0.759 mmol)
dissolved in a mixture of absolute EtOH (25
mL) and aq 60% HOAc (4 mL) was hy-
drogenolyzed over Pd–C at 9 atm. After 3
days, the mixture was filtered and H2O was
added. The pH was adjusted to 9 by the
addition of aq NaOH, and the aqueous
phase was washed once with Et2O. After a
slight concentration under reduced pressure,
the crude product was desalted on a Bio-Gel
P-2 column eluted with H2O containing 1%
n-BuOH. Freeze-drying gave 7 (256 mg,
97%); [h]D +41° (c 1.0, H2O); NMR: 13C
(D2O), l 55.5 (MeO), 62.5, 66.5, 70.7, 70.8,
72.2 (d), and 72.4 (d) (C-2–7), 102.0 (C-1);
31P (H2O), l 4.8 (s). HRMS Calcd for
C8H17O10P [M+Na]: 327.0457. Found:
327.0467.
Methyl 7-O-allyl-2,3,4-tri-O-benzyl-
ero - h - - manno - heptopyranoside (10).—
Methyl 2,3,4 - tri - O - benzyl - 1,6 - dialdo - a -
L-glyc-
D
D
-
mannopyranoside (5.42 g, 11.7 mmol) was
processed as described for the formation of
5 above, except that allyloxymethyl chloride
(3.2 mL, 30.9 mmol) was used instead of
benzyloxymethyl chloride, to yield, after sil-
ica gel chromatography (4:1 toluene–
EtOAc), 10 (2.82 g, 45%); [h]D +13° (c 1.0,
CHCl3); 13C NMR (CDCl3): l 54.7 (MeO),
68.0, 70.9, 71.5, 72.3, 73.0, 74.4, 74.7, 75.3,
and 80.4 (C-2–7, PhCH2, allyl CH2), 99.5
(C-1), 117.2 (allyl), 127.7–138.7 (aromatic,
allyl). Anal. Calcd for C32H38O7: C, 71.89;
H, 7.16. Found: C, 71.53; H, 7.23.
Methyl 7-O-allyl-2,3,4,6-tetra-O-benzyl-
glycero-h- -manno-heptopyranoside (11).—
Compound 10 (5.44 g, 10.2 mmol) in DMF
L
-
D

A. Stewart et al. / Carbohydrate Research 313 (1998) 193–202
199
was purified on a silica gel column (19:1
CHCl3–MeOH) followed by a silanized SiO2
column (acetone containing a 30“20% H2O
gradient). After concentration, the residue was
dissolved in toluene and washed once with
water to remove polar impurities. The organic
phase was dried (MgSO4), filtered, and con-
centrated to give methyl 2,3,4,6-tetra-O-ben-
(30 mL) was added dropwise to a suspension
of hexane-washed NaH (1.85 g, 55–60% in
mineral oil) and DMF (20 mL). The mixture
was stirred for 1 h at room temperature,
whereafter benzyl bromide (3.6 mL, 30.3
mmol) was added and stirring was continued
for 90 min. MeOH (5 mL) was added care-
fully during 20 min. The mixture was diluted
with toluene (200 mL) and washed with water
(3x100 mL). The organic phase was filtered
through a sintered glass funnel containing dry
MgSO4 and concentrated. Silica gel chro-
matography (9:1 toluene–EtOAc) gave 11
(5.96 g, 94%); [h]D +25° (c 1.0, CHCl3); 13C
NMR (CDCl3): l 54.7 (MeO), 70.0, 71.3, 72.0,
72.3, 72.5, 72.8, 74.3, 74.4, 74.6, 75.1, and 80.8
(C-2–7, PhCH2, allyl CH2), 99.1 (C-1), 117.0
(allyl), 127.5–138.9 (aromatic, allyl). Anal.
Calcd for C39H44O7: C, 74.98; H, 7.10; Found:
C, 74.60; H, 6.85.
zyl-7-O-dibenzyloxyphosphoryl-
L
-glycero-a- -
D
manno-heptopyranoside (14; 856 mg, 75%).
NMR: 13C (CDCl3), l 54.9 (MeO), 66.0 (d),
69.3 (d), 70.6, 71.9, 72.5, 72.8, 73.9, 74.2, 74.4,
74.9 (d), and 80.5 (C-2–7, PhCH2), 99.1 (C-1),
127.4–138.8 (aromatic); 31P (EtOAc), l 0.3
(m). To a solution of compound 14 (772 mg,
0.914 mmol) in 1:1 EtOAc–EtOH (30 mL)
was added Pd–C and H2O (0.5 mL), and the
mixture was stirred overnight under H2 (9
atm). The mixture was filtered through Celite,
diluted with MeOH, and the pH was adjusted
to 9 by aq 1 M NaOH. After concentration,
the residue was dissolved in H2O, washed once
with Et2O, and purified by size-exclusion chro-
matography (Bio-Gel P-2, eluent H2O con-
taining 1% of n-BuOH). Freeze-drying of
appropriate fractions gave 15 (283 mg, 89%);
[h]D +33° (c 1.0, H2O); NMR: 13C (D2O), l
55.6 (MeO), 66.4, 66.8, 68.8 (d), 70.7, 71.6,
and 71.7 (C-2–7), 101.7 (C-1); 31P (H2O), l
4.1 (dd). HRMS Calcd for C8H17O10P [M+
Na]: 327.0457. Found: 327.0447.
Methyl 2,3,4,6-tetra-O-benzyl-
L
-glycero-h-
D-manno-heptopyranoside (12).—A sample of
PdCl2 (60%; 159 mg, 0.54 mmol) was added to
a solution of 11 (2.90 g, 4.64 mmol) in 3:2
EtOH–MeOH (50 mL). The mixture was
stirred overnight, filtered through Celite, con-
centrated, and purified on a silica gel column
(2:1 toluene–EtOAc) to yield 12 (2.47 g, 91%);
[h]D +32° (c 1.0, CHCl3), lit. [h]D +37.8°
[19]; 13C NMR (CDCl3): l 54.9 (MeO), 62.5
(C-7), 71.9, 72.1, 72.5, 73.1, 74.2, 74.3, 74.6,
75.9, and 80.4 (C-2–6, PhCH2), 99.1 (C-1),
127.6–138.6 (aromatic).
Methyl 7 - O - (2 - aminoethyl)phosphoryl -
glycero-h- -manno-heptopyranoside (17).—
L
-
D
Methyl 2,3,4-tri-O-benzyl-
L
-glycero-h-
D-
Following the general phosphoethanolamine-
coupling procedure, compound 12 (766 mg,
1.31 mmol) afforded methyl 2,3,4,6-tetra-O-
benzyl-7-O-(2-benzyloxycarbonylamino)ethyl-
manno-heptopyranoside (13).—Compound 10
(2.10 g, 3.93 mmol) was deallylated as de-
scribed for 11 to give, after silica gel chro-
matography (3:1“1:2 toluene–EtOAc), 13
(1.28 g, 66%); [h]D +21° (c 1.0, CHCl3) lit.
[h]D +30.5° [20]; 13C NMR (CDCl3): l 54.8
(MeO), 65.0, 69.4, 72.2, 72.4, 73.0, 74.3, 74.5,
75.3, and 80.1 (C-2–7, PhCH2), 99.5 (C-1),
127.6–138.4 (aromatic).
phosphoryl-
L
-glycero-a- -manno-heptopyra-
D
noside, triethylammonium salt (16; 923 mg,
75%) after purification on a silica gel column
eluted with CHCl3 containing 10“20%
MeOH and 1% Et3N. NMR: 13C (CDCl3), l
8.4 (Et3N), 42.4 (d, CH2CH2N), 45.4 (Et3N),
55.0 (MeO), 64.2 (d), 64.4 (d), 66.4, 71.8, 72.4,
74.17, 74.22, 74.5, 75.4 (d), and 80.6 (C-2–7,
PhCH2, OCH2CH2), 98.9 (C-1), 127.2–139.2
(aromatic), 156.6 (CON); 31P (H2O), l 0.7 (m).
Compound 16 (780 mg, 0.827 mmol) in 1:1
EtOAc–EtOH (30 mL) and 0.25 mL H2O was
hydrogenolyzed over Pd–C at 9 atm for 5
days. Removal of the catalyst (Celite filtra-
Methyl
L
-glycero-h- -manno-heptopyra-
D
noside 7-(disodium phosphate) (15).—Com-
pound 12 (789 mg, 1.35 mmol) was subjected
to phosphorylation following the general pro-
cedure above. Before silica gel chromatogra-
phy the crude product was acetylated (2:1
pyridine–Ac2O) in order to acetylate excess
benzyl alcohol, and concentrated. The mixture

200
A. Stewart et al. / Carbohydrate Research 313 (1998) 193–202
tion) and concentration followed by gel filtra-
tion (two Bio-Gel P-2 columns eluted with
H2O containing 1% n-BuOH and pyridinium
acetate buffer, respectively) gave 17 (235 mg,
82%); [h]D +30° (c 1.0, H2O); NMR: 13C
(D2O), l 40.7 (d, CH2CH2N), 55.5 (CH3O),
62.5 (d), 66.6, 67.2 (d), 68.0 (d), 70.6, 71.4,
NMR: 13C (D2O), l 55.7 (MeO), 66.4, 67.4,
70.6, 71.3, 71.7 (d), and 74.9 (d) (C-2–7),
101.8 (C-1); 31P (H2O), l 18.3 (m). HRMS
Calcd for C8H14O9P [M−H]: 285.0375.
Found: 285.0420.
Methyl
L
-glycero-h- -manno-heptopyrano-
D
side 4-(disodium phosphate) (22).—Following
the general phosphorylation procedure above,
20 [18] (315 mg, 1.04 mmol) gave, after silica
gel chromatography (2:1 toluene–EtOAc),
methyl 4-O-dibenzyloxyphosphoryl-2,3;6,7-di-
31
and 71.5 (C-2–7, OCH2CH2), 101.7 (C-1); P
(H2O),
l
0.1 (m). HRMS Calcd for
C10H23O10NP [M+H]: 348.1060. Found:
348.1053.
Methyl
L
-glycero-h-
D
-manno-heptopyran-
O-isopropylidene-
L
-glycero-a- -manno-hep-
D
oside 6,7-cyclic phosphate, ammonium salt
(19).—Methyl dichlorophosphate (238 mL,
2.38 mmol) was dissolved in dry pyridine (4
mL) at 0 °C. The mixture was stirred for 30
min, whereafter a solution of 13 (463 mg,
0.936 mmol) in dry pyridine (3 mL) was added
dropwise. After stirring at 0 °C until the con-
version into a hydrophilic product was com-
plete (1 h), the reaction was quenched by the
addition of aq satd (NH4)2CO3 (10 mL). The
solution was concentrated, and the residue
was dissolved and partitioned between EtOAc
and H2O. The pH was adjusted to 1 prior to
separation of the EtOAc phase. The aqueous
phase was additionally washed once with
EtOAc, and the combined organic phases
were concentrated and co-concentrated three
times from EtOH. The crude methyl 2,3,4-tri-
topyranoside (21; 470 mg, 80%); NMR: 13C
(CDCl3), l 25.4, 26.2, 26.3, and 27.7
[(CH3)2C], 55.2 (CH3O), 65.7, 68.2 (d), 69.2
(d), 69.6 (d), 74.2, 75.9, 76.4 (d), and 76.8
(C-2–7, PhCH2), 98.3 (C-1), 109.3 and 110.0
[(CH3)2COO], 127.8–136.1 (aromatic); 31P
(EtOAc), l −0.8 (m). A mixture of 21 (235
mg, 0.416 mmol) and Pd–C in absolute EtOH
(10 mL) and aq 80% HOAc (10 mL) was
hydrogenolyzed overnight at 9 atm. The mix-
ture was filtered through Celite, evaporated
and the residue was stirred with aq 80%
HOAc (15 mL) for 1 h at 40 °C in order to
complete deisopropylidenation. After concen-
tration, the residue was dissolved in H2O (10
mL, containing 1% n-BuOH) and the pH was
adjusted to ꢀ7 using aq NaOH. Freeze-dry-
ing, followed by desalting on a Bio-Gel P-2
column (H2O containing 1% n-BuOH) ren-
dered 22 (119 mg, 82%); [h]D +30° (c 1.0,
H2O); NMR: 13C (D2O), l 55.5 (MeO), 63.2,
69.2, 70.0 (d), 70.2, 71.3 (d), and 71.7 (C-2–7),
101.5 (C-1); 31P (H2O): l 4.7 (d). HRMS
Calcd for C8H17O10P [M+Na]: 327.0457.
Found: 327.0478.
O-benzyl-
L
-glycero-a- -manno-heptopyranos-
D
ide 6,7-cyclic phosphate, ammonium salt (18;
426 mg, 79%) was used in the next step with-
out further purification. NMR: 13C (CDCl3):
55.0 (CH3O), 64.7, 69.9 (d), 72.3, 73.1, 73.5,
73.8, 75.0, 75.4, and 80.0 (C-2–7, PhCH2),
99.2 (C-1), 127.5–138.2 (aromatic); 31P
(EtOAc), l 18.4 (m). Compound 18 (426 mg,
0.743 mmol) was dissolved in 1:1:1 THF–
EtOH–H2O (24 mL) and hydrogenolyzed in a
Parr apparatus (9 atm) using Pd(OH)2 as cata-
lyst. After 3 days the mixture was filtered
through Celite, and concentrated. The Celite
was suspended in H2O and centrifuged. The
pellet was additionally washed and centrifuged
from water. The original residue was dissolved
in the combined aqueous phases and washed
once with Et2O. The water phase was concen-
trated and the residue desalted on a Bio-Gel
P-2 column to give, after lyophilization, 19
(195 mg, 87%); [h]D +41° (c 1.0, H2O);
Methyl 4 - O - (2 - aminoethyl)phosphoryl -
L
-
glycero-h- -manno-heptopyranoside (24).—
D
The general phosphoethanolamine coupling
protocol applied on 20 (384 mg, 1.26 mmol)
gave after silica gel chromatography (CHCl3
containing 10“17% MeOH and 1% Et3N)
methyl 4-O-(2-benzyloxycarbonylaminoethyl)-
phosphoryl - 2,3;6,7 - di - O - isopropylidene- -
L
glycero-a- -manno-heptopyranoside, triethy-
D
lammonium salt (23; 742 mg, 89%). NMR: 13C
(CDCl3), l 8.5 (Et3N), 25.8, 26.2, 26.3, and
27.7 [(CH3)2C], 42.1 (d, CH2CH2N), 45.3
(Et3N), 54.9 (CH3O), 64.7 (d), 66.5, 66.8, 69.9
(d), 74.0 (d), 75.7, and 77.5 (C-2–7, PhCH2,

A. Stewart et al. / Carbohydrate Research 313 (1998) 193–202
201
Table 2
1H NMR data of selected compoundsa obtained at 38 °C relative to internal acetone (lH 2.225)
Compound
H-1
H-2
H-3
H-4
H-5
H-6
H-7a
H-7b
OMe
7b
9
4.76
4.78
4.78
4.78
4.81
4.76
4.78
4.77
4.78
3.92
3.95
3.91
3.92
3.93
3.93
3.95
4.07
ꢀ3.91
3.80
3.81
3.77
3.78
3.75
3.89
3.94
4.18
ꢀ3.74
3.96
3.95
3.87
3.88
3.82
4.25
4.35
3.95
3.88
3.67
3.71
3.62
3.60
3.56
3.63
3.70
3.62
3.58
4.42
4.58
4.18
4.20
4.82
4.17
4.08
4.02
4.05
3.82
3.81
3.88
4.01
4.36
3.75
3.95
3.74
3.78
3.90
3.89
3.98
4.01
4.29
3.75
3.95
3.74
3.78
3.41
3.40
3.41
3.41
3.40
3.41
3.40
3.41
3.41
15b
17
19
22b
24
25
26
a Compounds 25 and 26 were not synthesized for this study but were obtained after alkaline treatment and characterized by the
use of 2D NMR and MS. Signals for ethanolamine protons for compounds 9, 17 and 24 occurred at l 4.11, 4.18, 4.12 (O–CH2),
3.19, 3.30, 3.29 (N–CH2), respectively.
1
b Assignments were made from COSY experiments and the location of phosphate substituents confirmed by H–31P HMQC
experiments.
OCH2CH2), 98.3 (C-1), 108.7 and 109.6
[(CH3)2COO], 128.0–136.8 (aromatic), 156.7
for 2 h at 60, 80 and 100 °C. LPS from N.
meningitidis immunotype L3 lgtB [5] was
treated with 4 M KOH at 120 °C for 18, 8, 4
and 2 h. One experiment was conducted with
2 M KOH for 2 h and 60 °C. Reaction mix-
tures were cooled, neutralized and fraction-
ated on a Bio-Gel P-2 column (43×1.5 cm)
with water as eluent. Products were analyzed
31
(CON); P (9:1 EtOAc–CHCl3), l −0.7 (m).
A solution of 23 (360 mg, 0.543 mmol) in
EtOH (6 mL) and aq 80% HOAc (6 mL) was
hydrogenolyzed over Pd–C at 8 atm. After 4
days the mixture was filtered through Celite,
and concentrated. The residue was dissolved
in 1:1 THF–H2O (12 mL), new catalyst was
added and the hydrogenolysis was continued
for 3 days whereby on the last day aq 25%
NH3 (1 mL) was added. The solution was
neutralized with HOAc (approx. 1 mL),
filtered, and concentrated. The residue was
dissolved in H2O and washed once with Et2O.
Desalting on a Bio-Gel P-2 column (aq 0.1
mM pyridinium acetate buffer, pH 5.4) gave,
after freeze-drying of appropriate fractions, 24
(149 mg, 79%); [h]D +37° (c 1.0, H2O);
NMR: 13C (D2O), l 9.0 (Et3N), 40.8 (d,
CH2CH2N), 47.4 (Et3N), 55.5 (MeO), 62.9 (d),
63.1, 69.3, 70.8, 70.9 (d), and 72.4 (d), (C-2–7,
1
by H NMR and ESMS.
NMR spectroscopy.—13C NMR spectra in
D2O of 7, 15, 19 at pDꢀ9, and 9, 17 and 24
at pDꢀ7 were recorded on a JEOL GSX-270
1
instrument. H NMR data for 7, 9, 15, 17, 19,
22, 24, 25 and 26 are shown in Table 2 and
were obtained with a Bruker AMX 500 spec-
trometer using standard Bruker software. All
1H measurements were made in D2O at 38 °C
after several lyophilizations with D2O. Ace-
tone was used as the internal standard (lH,
2.225 ppm). 2D COSY experiments were per-
formed by using standard Bruker pulse se-
quences. 31P spectra were measured at 202 or
109 MHz and aq 85% phosphoric acid was
used as the internal standard (lP, 0.0 ppm).
1H–31P correlations (HMQC) and 1H–31P
HMQC-TOCSY experiments were made in
31
OCH2CH2), 101.6 (C-1); P (H2O), l 0.4 (m).
HRMS Calcd for C10H23O10NP [M+H]:
348.1060. Found: 348.1089.
Alkaline treatment.—Phosphate and phos-
1
phoethanolamine substituted
L
,
D
-heptose and
the H-detected mode.
D
-mannose methyl glycosides (10 mg) were
Electrospray ionization mass spectrometry
(ESMS).—Samples were analyzed on a VG
Quattro Mass spectrometer (Micromass,
Manchester, UK) fitted with an electrospray
treated with 4 M KOH (2 mL) for 18 h at
120 °C. In addition, compound 9 was sub-
jected to alkaline treatment with 2 M KOH

202
A. Stewart et al. / Carbohydrate Research 313 (1998) 193–202
[3] R.A. Moore, R.E.W. Hancock, Antimicrob. Agents
Chemother., 30 (1986) 923–926.
[4] H. Masoud, E.R. Moxon, A. Martin, D. Krajcarski, J.C.
Richards, Biochemistry, 36 (1997) 2091–2103.
[5] W.W. Wakarchuk, A. Martin, M.P. Jennings, E.R.
Moxon, J.C. Richards, J. Biol. Chem., 271 (1996) 19166–
19173.
[6] W. Brabetz, S. Mu¨ller-Loennies, O. Holst, H. Brade,
Eur. J. Biochem., 247 (1997) 716–724.
[7] J.P. Kamerling, R. Schauer, A.K. Shukla, S. Stoll, H.
van Halbeek, J.F.G. Vliegenthart, Eur. J. Biochem., 162
(1987) 601–607.
[8] E. Schweda, Doctoral Dissertation, Chem. Commun.
Uni6. Stockholm, 3 (1987).
[9] C. Bernlind, S. Oscarson, G. Widmalm, Carbohydr. Res.,
263 (1994) 173–180.
[10] B. Grzeszczyk, A. Zamojski, Carbohydr. Res., 262 (1994)
49–57.
[11] C. Blomberg, F.A. Hartog, Synthesis, (1977) 18–30.
[12] C.E. Dreef, C.J.J. Elie, P. Hoogerhout, G.A. van der
Marel, J.H. van Boom, Tetrahedron Lett., 29 (1988)
6513–6516.
ion source. Samples were dissolved in water,
which was then mixed in a 1:1 ratio with aq
50% acetonitrile containing 1% HOAc for
mass spectral analysis in the negative mode.
High resolution mass spectra (HRMS).—
High resolution mass spectra were obtained
by fast atom bombardment in either the posi-
tive or negative ion mode on a JEOL AX 505
H double focussing mass spectrometer with a
mass resolution of 7000. In the positive ion
mode glycerol was used as the matrix and
polyethyleneglycol (PEG) 200–400 was used
as the internal calibrant. HRMS of methyl
L
-glycero-a-
D
-manno-heptopyranoside
6,7-
cyclic phosphate was obtained in the negative
mode with triethanolamine as the matrix and
PEG 600 diacid as the calibrant.
[13] R.R. King, F.R. Cooper, C.T. Bishop, Carbohydr. Res.,
55 (1977) 83–93.
[14] T.N. Cawley, R. Letters, Carbohydr. Res., 19 (1971)
373–382.
Acknowledgements
[15] H.G. Khorana, G.M. Tener, R.S. Wright, J.G. Moffatt,
J. Am. Chem. Soc., 79 (1957) 430–436.
[16] P.J. Garegg, S. Oscarson, M. Szo¨nyi, Carbohydr. Res.,
205 (1990) 125–132.
[17] D.S. Connor, G.W. Klein, G.N. Taylor, Org. Synth., 52
(1972) 16–19.
Don Krajcarski is acknowledged for ESMS
analyses and Kenneth H.N. Chan for HRMS
determinations.
[18] G.J.P.H. Boons, Doctoral Dissertation, Leiden Univer-
sity, 1991.
[19] G.J.P.H. Boons, R. Steyger, M. Overhand, G.A. van der
Marel, J.H. van Boom, J. Carbohydr. Chem., 10 (1991)
995–1000.
[20] G.J.P.H. Boons, G.A. van der Marel, J.T. Poolman, J.H.
van Boom, Recl. Tra6. Chim. Pays-Bas, 108 (1989) 339–
343.
References
[1] Y.A. Knirel, N.K. Kochetkov, Biochemistry (Moscow),
58 (1993) 166–201.
[2] A.C. Sundstro¨m, E.K.H. Schweda, P.-E. Jansson, Eur. J.
Biochem., 243 (1997) 701–707.
.